Vous êtes sur la page 1sur 284


23, 2012


, Turnhout

All rights reserved.

No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, recording, or otherwise,
without the prior permission of the publisher.
ISBN 978-2-503-54673-5
Printed in the E.U. on acid-free paper

Revue fonde en 1990 par
Jean-Claude P ICARD et Pierre G EOLTRAIN


Directeur de la publication
Secrtaires de rdaction
Comit de lecture
Comit scientifique
Revue publie avec le concours scientifique de
lAssociation pour ltude de la Littrature Apocryphe Chrtienne
et de
la Socit pour ltude de la Littrature Apocryphe Chrtienne

Adresse du secrtariat de la revue:

187, rue Belliard
F-75018 PARIS

Liminaire .................................................................................................

Virtutes apostolorum: Editorial Problems and Principles

par Els ROSE....................................................................................


Reworking the Virtutes apostolorum in the Salzburg Sermon-Collection (1st quarter of the ninth century)
par Maximilian DIESENBERGER ........................................................


A (Socio)linguistic Approach to Hagiographic Text Transmission: the

Torino Collection Biblioteca Nazionale D.v.3. (8th/9th c.)
par Marieke VAN ACKER .................................................................


Editing a Fluid and Unstable Text: The Example of the Acts of Pilate
(or Gospel of Nicodemus)
par Rmi GOUNELLE ........................................................................


On the Evangelium Nicodemi before Print: Towards a New Edition

par Zbigniew IZYDORCZYK ...............................................................


The Transmission of Marian Apocrypha in the Latin Middle Ages

par Rita BEYERS ..............................................................................


Homiletic Texts and the Transmission of Eschatological Apocrypha

in a Medieval Irish Context
par Caitrona DOCHARTAIGH .........................................................


The Digital Edition: New Possibilities and Challenges

par Mariken TEEUWEN .....................................................................


Un rexamen des notices de Thodore Bar Kona sur les mandens
par Jean-Marie DUCHEMIN ...............................................................


Jsus, la lumire et le Pre vivant. Principe de gmellit dans

lvangile selon Thomas
par Andr GAGN ............................................................................


Some Philological Notes on the Mf rft lMaryam Liber Requiei (LR)

par Tedros ABRAHA .........................................................................


RECENSIONS ..............................................................................................


LIVRES REUS ............................................................................................


The Dynamics of Apocryphal Traditions in

Medieval Religious Culture

Proceedings of the Expert Meeting
Utrecht University, 19-20 January 2012

In January 2012 an Expert Meeting took place at Utrecht University,
organized as part of the research project The Dynamics of Apocryphal
Traditions in Medieval Religious Culture, funded by the Netherlands
Organisation for Scientific Research (NWO) and supervised by project
leader Els Rose since 2008. The project deals with the medieval textual transmission of the so-called Virtutes apostolorum, Latin rewritings
of the apocryphal Acts of the Apostles, transmitted as a coherent series
with a section on each individual apostle. While the Virtutes apostolorum
might have circulated as a series earlier, the oldest extant manuscripts
date to the second half of the eighth century. Within the project, the textual approach consists of an examination of relevant manuscripts in order
to find answers to questions concerning coherence, context, and linguistic
features of the Virtutes, to be published together with a new edition cum
translation in CCSA.
During the Expert Meeting, members of the project presented particular problems and possible solutions concerning the Virtutes apostolorum,
while the other speakers contributed with presentations of related text
corpora, either apocryphal texts or genres in combination with which the
Virtutes apostolorum have been transmitted (hagiography, sermons). The
main questions concern the complexity of dealing with a series or collection of texts that are also transmitted individually in earlier ages and
in different (though comparable) contexts; the question how language
can form an instrument to get a better grip on the mutual coherence of
the manuscripts and their descent, and to reconstruct, if not a classic
stemma, a network of interrelations; how modern editorial techniques
(both digital and on paper) can be usefully employed to provide the user
with an adequate insight into the way these texts were transmitted and
used in medieval Christian culture; finally, how a contextual approach
to the transmission of apocryphal texts in medieval religious culture can
enlarge our understanding of these texts and their use.
Thanks are due to the Netherlands Organisation for Scientific Research
(NWO) that sponsored the Expert Meeting. Likewise, the organizer of the
Expert Meeting thanks Utrecht University, particularly its Faculty of the
Humanities and its Research Institute for History and Culture, for hosting
the project as well as the conference.
Els Rose
Editor of the Acts of the Utrecht Expert Meeting

Faculty of Humanities, Utrecht University

The school of New Philology urges us as editors of medieval texts to
take into account variance as a core characteristic of medieval manuscript transmissions. The scope of variance, however, remains unclear
in the programmatic studies presented by this trend, which concentrates
mainly on linguistic variants. Therefore, an elaborate methodology to
incorporate variance in our editorial work still needs to be developed.
It is the aim of this article to present a set of propositions in preparation of an edition of the Latin transmission of the Virtutes apostolorum,
in order to add to the development of this methodology. The spectrum
of variance in the Virtutes apostolorum as a series embraces textual
variants ; variants concerning the presentation of the text at the level of
the individual sections dedicated to one apostle (order of the apostles,
choice of text, choice of redaction) ; variants in (the presence of) paratexts (titles, chapter titles, prologues, epilogues) ; variants caused by the
manual transmission of these texts (corrections, uncorrected errors) ;
and variants inserted by the users/readers of the series (semantic
glosses, both vernacular and Latin, marginal glosses indicating liturgical use, stress marks on words for the benefit of oral performance).
The analyses show the great diversity in the preserved manuscripts. The
main question will be how we can create an edition of such a complex
* This article is based on a paper delivered at the annual conference of
the Association pour ltude de la littrature apocryphe chrtienne (AELAC)
Dole (France), 30 June-2 July 2011, and on a paper presented at the Expert
Meeting The Dynamics of Apocryphal Literature in Medieval Religious Culture, Utrecht (the Netherlands), 19-20 January 2012. It is my pleasure to thank
the participants in both conferences for their helpful remarks. I particularly like
to thank Valentina Covaci, Renske van Nie, Maarten Prot, Tom de Schepper,
vina Steinova, Maril Urbanus, and Giorgia Vocino for their contributions to
the discussions in preparation of the papers and the Expert Meeting, as well as
Peter Schrijver and Rmi Gounelle for their careful reading of an earlier draft.
The article was written within the framework of a research project funded by
the Netherlands Organisation for Scientific Research (NWO) and executed at
Utrecht University. I like to thank the Bibliothque Sainte-Genevive, Paris,
and the Stiftsbibliothek St. Gallen for allowing the reproduction of their manuscripts on p. 34 and 35 of this article.


Apocrypha 23, 2012, p. 11-45



and composite medieval text transmission so that its fluidity and variability become visible, as well as its use in medieval religious practice.
Lcole de la Nouvelle Philologie dfend le retour aux manuscrits
ainsi que lattention aux variantes dans la transmission manuscrite des
textes. Cependant, lide de variante est souvent limite aux variantes
textuelles, tandis quil y a dautres sortes de variantes qui tmoignent
aussi de spcificits ainsi que des diffrents usages qui ont t faits
des manuscrits. Le prsent article examine la transmission latine des
Virtutes apostolorum dans le monde mdival. Dans cet ensemble de
textes, plusieurs sortes de variantes sont analyses : variantes textuelles, variantes dans la prsentation du texte au niveau des sections
individuelles ddies un aptre (ordre des aptres, choix des textes,
choix des rdactions), variantes dans les paratextes (titres, titres des
chapitres, prologues, pilogues), variantes causes par la transmission
(corrections et passages fautifs mais non corrigs), variantes cres
par lusage des textes (gloses vernaculaires et latines, gloses marginales indiquant lusage liturgique, signes daccentuation). Les analyses
attestent de la grande diversit des manuscrits conservs. La question
principale est de savoir comment viter que ldition des Virtutes apostolorum ne fournisse une prsentation trop statique dune tradition textuelle complexe et composite, et comment faire pour quelle atteste la
variabilit qui caractrise ces textes, de mme que lusage de ces textes
dans la pratique religieuse du Moyen Age.

The Virtutes apostolorum : Variety rules

Variety and textual instability have been highlighted positively over
the past decades as key characteristics of handwritten text transmission. Return to the manuscripts is the device of the movement that
identifies itself as the New Philology, or Nouvelle Philologie. This
movement urges scholars to take seriously the variety that characterizes manuscript transmission, instead of chopping ones way through
the jungle of variants in search of a common ancestor or archetype.
In the Introduction to the special issue of Speculum (1990), dedicated to the programme of the New Philology, Stephen Nichols defines
its main objectives. Nichols emphasizes the need to consider diversity
as the principle characteristic of medieval text transmission and to take
account of this in our editorial work : [] not simply to live with it,
but to situate it squarely within our methodology.1 While Nichols call
to make diversity the core of our editorial methodology found response
in broad circles of medievalists (and philologists of other periods), the
way to implement it is not straightforward, nor does Nichols describe
such a method in detail.
1. Stephen G. NICHOLS, Introduction : Philology in a Manuscript Culture,
Speculum 65, 1990 (The New Philology), p. 1-10, at p. 9.



The label New Philology implies that previous philological methodologies are considered insufficient. We find a discussion of the methodology criticized most by the defenders of the new school in one of
the latters principal sources of inspiration : Bernard Cerquiglinis loge
de la variante (1989).2 In this essay, Cerquiglini first reconsiders what
became known as the Lachmannian method, associated with the German philologist Karl Lachmann ( 1851)3 and propagated in France by
Gaston Paris. Lachmann developed a genealogical approach to manuscripts, in line with the nineteenth-century tendency to consider cultural phenomena (not only language, but also, for instance, liturgical
traditions) as biological organisms.4 Family traits, interrelating different manuscripts, became visible in variants that Lachmann considered
scribal errors. Starting with the assumption that two scribes would
never make the same mistake (an assumption called into question by
others5), the Lachmannian method connected manuscripts in which
similar inherited defects were found, and thus created a family tree
of manuscripts : the stemma codicum. The stemma finds its roots in a
reconstructed Ur-text, also called the hypothetical ancestor or, in
less genealogical terms, the archetype not identical with but closely
related to the original. This archetype is, in most cases, the reconstruction of a manuscript now lost.
2. Bernard CERQUIGLINI, loge de la variante. Histoire critique de la philologie, Paris, 1989.
3. And yet, in Michael Driscolls words, Lachmann never produced a
stemma. Michael James DRISCOLL, The Words on the Page : Thoughts on Philology, Old and New, in Judy QUINN and Emily LETHBRIDGE (eds.), Creating the
Medieval Saga : Versions, Variability, and Editorial Interpretations of Old Norse
Saga Literature, Odense, 2010, p. 87-104, at p. 88.
4. For a survey of the Darwinian background to a genealogical approach
to cultural phenomena, particularly texts copied by hand, see Caroline MAC
and Philippe V. BARET, Why Phylogenetic Methods Work : The Theory of Evolution and Textual Criticism, in Caroline MAC e.a. (eds.), The Evolution of
Texts : Confronting Stemmatological and Genetical Methods, Pisa, Istituto editoriali e poligrafici internazionali, 2006, p. 89-108. With regard to liturgy, cf.
Anton BAUMSTARK, Liturgie compare, Chevetogne, 1953. This third edition of
what was first published as a series of articles in Irenikon 11, 1934, was revised
by Bernard BOTTE, who added the subtitle Principes et mthodes pour ltude
historique des liturgies chrtiennes. See also Fritz WEST, The Comparative Liturgy of Anton Baumstark, Bramcote, 1995, part. p. 16-25.
5. CERQUIGLINI, loge de la variante, p. 77 ; DRISCOLL, The Words on the
Page, p. 89. Early in the twentieth century, Joseph Bdier (1864-1938) formulated a critique of Lachmanns method of learned reconstruction. In an essay
dated 1928, Bdier rejected the scientific objectivity of a reconstructed (Ur-)
text, and advocated instead the selection of a best manuscript, to be chosen
among the manuscripts still extant. Whereas the Lachmannian method considered every textual variant a scribal error, Bdier took the scribe more seriously.
Joseph BDIER, La tradition manuscrite du Lai de lombre. Rflexions sur lart
dditer les anciens textes, Paris, 1929. Cf. DRISCOLL, The Words on the Page,
p. 89-90.



The programme of renewal formulated by the adherents of the New

Philology in many cases scholars of medieval vernaculars is aptly
captured by the title of Cerquiglinis essay (labelled polemical by
Michael Driscoll6) : loge de la variante, In Praise of the Variant.
Cerquiglini describes the Lachmannian method as archaeological, its
aim being to reconstruct a historical reality now lost, given the fact
that the Lachmannian archetype exists only in the work of the stemmatologist. Cerquiglini dismisses the search for an archetype, and is
of the opinion that the philologist should not work as an archaeologist
but rather as a comparatist. His or her main objective should be to present variety as the main characteristic of handwritten text transmission,
[] for archaeology reduces to unity that which derives its meaning
or significance from variety.7
In imitation of Cerquiglini, subsequent defenders of the New Philology put the phenomenon of variety centre stage. They underline Cerquiglinis statement that variation in the vernacular literature of the
Middle Ages is its core characteristic, a state of alterity which is so
concrete that it shapes the very object, and which is so primary that
it should be the editors prime concern to show it.8 Variance rules in
the medieval written world perceived through the lens of Cerquiglini
and the New Philologists. Or, even more poignantly : chaos has taken
over, as the Old Norse philologist Michael Driscoll puts it, referring to
Paul Zumthors discussion of the instability of handwritten text transmission.9 Driscoll recalls the introduction of the notion of mobility
(mouvance) as an essential quality of a medieval text in Zumthors
Essai de potique mdivale (1972), a main source of inspiration for
Cerquiglini and, hence, for the New Philologists. Just as Cerquiglini
praises the variant, Driscoll follows Zumthor in his positive approach
to the instability of medieval manuscript transmission : [] textual
instability (variance, mouvance, unfixedness) is so fundamental a feature of chirographically transmitted texts that rather than trying to bring
order to this chaos we should celebrate it.10

6. DRISCOLL, The Words on the Page, p. 90.

7. CERQUIGLINI, loge de la variante, p. 68. We can differ in our opinions
about the extent to which Cerquiglinis comment does justice to the archaeological method.
8. La variance de luvre mdivale romane est son caractre premier, altrit concrte qui fonde cet objet, et que la publication devrait prioritairement
donner voir. CERQUIGLINI, loge de la variante, p. 62. Cf. NICHOLS, Philology in a Manuscript Culture, p. 9 ; Suzanne FLEISCHMAN, Philology, Linguistics, and the Discourse of the Medieval Text, in Speculum 65, 1990, p. 19-37,
at p. 25.
9. Paul ZUMTHOR, Essai de potique mdivale, Paris, 1972/2000, e.g. p. 93
(Luvre est fondamentalement mouvante) and 94 (Luvre, ainsi conue, est
par dfinition dynamique).
10. DRISCOLL, The Words on the Page, p. 104.



Celebrating the chaos of textual instability is familiar to scholars

who study the transmission of apocryphal texts, for which fluidity is
one of the most characteristic properties. In my approach to the Virtutes
apostolorum, a medieval series of Latin rewritings of the apocryphal
Acts of the Apostles, the overall aim is to enlarge our insight into the
use of these texts in medieval religious culture. In this endeavour, variety is an important key, because it can inform us about the differences
in reception, transmission and use per region and period. The present
article concentrates on the methodological implications of Nichols
plea, examining how a modern edition can incorporate this variety in
order to make visible not simply a readable text cut off from its performative context but also its medieval uses. I will not be able to provide
a complete answer to this question within the scope of this article, for
an editorial methodology for such an edition or text presentation still
needs to be developed. The aim here is to identify different kinds of
variance in the medieval transmission of the Virtutes apostolorum as
well as the challenges they present, and to propose a survey of principles and possible strategies for dealing with these problems.

How to define the Virtutes apostolorum

The aim of this article thus formulated hides the most fundamental
problem concerning the Virtutes apostolorum, namely the difficulty of
defining precisely what this label covers. The title Virtutes apostolorum
refers to a series of narrative texts on the apostles, coherent to a certain
degree, but marked at the same time by so much variation that the
legitimacy of the label collection becomes questionable. Moreover, a
variety of (series of) texts on the life and martyrdom of the apostles
circulated in different regions of the Latin Middle Ages under various
titles and variously combining the protagonists and textual accounts of
their deeds. Therefore, before listing the editorial problems and principles more specifically, it is best to clarify first what I mean when I use
the term Virtutes apostolorum, and define what this title can and what
it cannot cover in the context of my own specific research parameters,
earlier scholarly assertions notwithstanding.
It is difficult to establish when and where a coherent series of Virtutes apostolorum occurred for the first time. Previous scholars, particularly Richard Lipsius, pointed to late sixth-century Gaul as the cradle
of the phenomenon, considering Venantius Fortunatus (530-c. 610) and
Gregory of Tours (538-594) as likely candidates to have been involved
in the compilation of the series.11 Elsewhere, I have elaborated on the
problematic character of this assumption, and indicated that the first
convincing piece of evidence for the existence of the narrative Virtutes
11. Richard LIPSIUS, Die apokryphen Apostelgeschichten und Apostellegenden, Braunschweig, 1893-1890, vol. 1, p. 117-178.



apostolorum is found only after the sixth century, and outside continental Gaul.12 Anglo-Saxon authors like Aldhelm (639-709) and Bede
(673-735) certainly knew the narratives about the apostles, but we must
be cautious about equating their sources with the material we find in
continental manuscripts. The case of Aldhelm, in particular, shows
clearly that the historiae or passiones apostolorum that inspired his
poems deviate partly from the narratives we find on the continent from
the eighth century onwards. This observation has received little attention in the scholarly debate so far, which puts its trust in the earlymodern idea that the Virtutes apostolorum constitute a fixed collection,
attributable to a single author usually referred to as Pseudo-Abdias.13 I
shall discuss this further below.
When I use the term Virtutes apostolorum, I refer to the phenomenon whereby (1) separate Latin narratives on individual apostles, partly
rewritings of the ancient Apocryphal Acts, partly new compositions,
occur in manuscripts as a more or less coherent series. This is not to
say that the individual sections did not circulate separately, or could not
occur scattered through a manuscript organized liturgically rather than
hierarchically they did and abundantly so.14 However, I am interested
in the transmission of the texts as a coherent series because of the
information this kind of transmission yields vis--vis the status of and
approach to the apostles in the Middle Ages, more specifically in the
practice of liturgical commemoration.
The series takes different shapes and has different contents in each
manuscript, but my working definition of Virtutes apostolorum requires
that (2) at least the twelve be included, that is, Peter, Paul, Andrew,
John, James the Greater, Thomas, Philip, James the Less, Bartholomew,
Matthew, Simon and Jude. Simon and Jude share a section, as do Peter
and Paul in some instances. To define the series on the grounds of a
12. Els ROSE, Virtutes apostolorum : Origin, Aim, and Use, forthcoming.
13. In scholarly contributions on the theme, the series of Virtutes apostolorum as it is found in continental manuscripts stemming from Frankish Gaul
and Bavaria is generally referred to as the Collection of Pseudo-Abdias, a
problematic notion which suggests far more unity than is accounted for in the
manuscript tradition. The idea of a Collection of Pseudo-Abdias as a uniform
entity is persistent also among modern scholars studying the transmission of the
apostle narratives in the insular world. See e.g. Charles WRIGHT, Apocryphal
Acts, in Frederick M. BIGGS e.a. (eds.), Sources of Anglo-Saxon Literary Culture : A Trial Version, Binghamton NY, 1990, p. 48-63 ; cf. Abdias, in William SMITH and Henry WACE, A Dictionary of Christian Biography, Literature,
Sects and Doctrines, London, 1877-1887, vol. 1, p. 1-4. For an analysis of the
attribution to Pseudo-Abdias and its unsuitability, see Els ROSE, Abdias scriptor
vitarum sanctorum apostolorum ? The Collection of Pseudo-Abdias reconsidered, Revue dhistoire des textes 8, 2013, p. 227-268.
14. The abundance of manuscripts becomes apparent when one checks the
dossiers of individual apostles at the BHL-database at http ://bhlms.fltr.ucl.ac.be
(last consulted 29 February 2012).



fixed list of protagonists is not without difficulties, as we shall see later.

The transmission of these narratives is studied as (3) an uninterrupted
series (but there are exceptions to this rule, as we shall see below) in
manuscripts (4) from the first preserved witnesses (late eighth century)
until the emergence of large hagiographic compilations in the thirteenth
century, such as the Legenda aurea.15
Narrative accounts of the life and martyrdom of the apostles occur
throughout the medieval West, from Rome16 and Italy17 to Anglo-Saxon
England, as we have seen in the previous section. (5) The oldest preserved manuscripts, however, occur on the continent, more precisely in
Frankish Gaul and Bavaria (perhaps West-Germany/Alsace18). In order
to create a manageable corpus of manuscripts for our studies and for
15. This deadline is inspired by the hypothesis that the Legenda aurea
became the most influential source in the later Middle Ages, also with regard
to visual representations of the narratives on the life and martyrdom of the
apostles. Franois Dolbeau, however, has come to the conclusion that the traditional legendaries did not fall into disuse after the composition of the thirteenth-century narrative compilations. Franois DOLBEAU, Typologie et formation des collections hagiographiques daprs les recueils de labbaye de SaintThierry, in Michel BUR (ed.), Saint-Thierry. Une abbaye du VIe au XXe sicle,
Saint-Thierry, 1979, p. 159-182.
16. The earliest manuscripts used in Rome seem to stem from the twelfth
century. Famous are the codices Vat. lat. 5736 (from Santa Caecilia in Trastevere) and Vat. reg. lat. 1272.
17. The earliest Italian manuscripts with a complete series known to me are
Stuttgart, Wrttembergische Landesbibliothek HB XIV 14 (s. IX) and Vat. lat.
5771 (s. IXfin or Xinc, from Bobbio). Cf. Klaus ZELZER (ed.), Die alten lateinischen Thomasakten, Berlin, 1977, p. xlix.
18. The Bavarian Group in my selection of manuscripts (see appendix 1)
includes a manuscript of which the Bavarian origin is not certain : Wolfenbttel
Weissenburg 48. While Bischoff points to Weissenburg (Alsace) as the books
origin (unpublished correspondence ; see ric JUNOD and Jean-Daniel KAESTLI,
Acta Iohannis, CCSA 1-2, Turnhout, 1983, p. 756), Butzmann doubts whether
the manuscript is from Weissenburg or from Sankt Gallen (Hans BUTZMANN,
Althochdeutsche Priscian-Glossen aus Weissenburg, Beitrge zur Geschichte
der deutschen Sprache und Literatur 86, 1964, p. 388-402, at p. 401-402). For
the time being, I classify Wolfenbttel Weissenburg 48 as Bavarian for three
reasons. First, there is some doubt concerning its origin, and reason to think
it is from Sankt Gallen. Second, there are important similarities between this
manuscript and the representatives of the core Bavarian Group, both textually
and concerning matters of text presentation. Third, there are in general close
connections between Weissenburg and Sankt Gallen. I thank Max Diesenberger
who, in an e-mail d.d. 16 November 2011, underlines the strong political links
between Bavaria and Alsace, a.o. in the presence of Bavarian bishops at Louis
the Germans court, such as archchaplain Baturich of Regensburg, and archchaplain and chancellor Grimoald of Weissenburg, who also had strong links
with Bavarian scriptoria through the collection and copying of many books.
It is clear that Alsace is an important region for the early transmission of the
Virtutes apostolorum ; see for example the eighth-century manuscript Wrzburg



the new edition now under way for the Corpus Christianorum series
Apocryphorum, we concentrate on the further development of the transmission in these areas.19 This has resulted in a selection of twenty-five
manuscripts, established with the help of earlier inventories and studies.20
The manuscripts that testify to the transmission of the Virtutes apostolorum as defined above are generally liturgical in character, containing Passions and Lives of saints and martyrs. The texts were obviously
meant to be read aloud during the liturgical office or hours in cathedral and monastic communities, particularly during the night office
(matins).21 Although the initial motive for rewriting the stories of the
apostles was to collect and complete the knowledge about them, as a
number of prologues preceding different sections in the series declare,22
the further incentive must have been connected with the texts liturgical
use. This means that the texts were copied not to provide a standard
reading form, but to offer a lively oral presentation for a performative
context that played an important role in its transmission history. The
variety that characterizes the manuscript transmission of the Virtutes
apostolorum as a series demonstrates that the wish to have a complete
overview of the life and martyrdom of the twelve apostles did not
include the desire to mould the separate narratives into a uniform collection. Efforts to streamline and unify the individual texts into a coherent series were rather limited. This allowed for the continued transmission of a variety of all-inclusive collections of texts on the apostles,
which offered room for adaptation to local needs and preferences.23
The specific character of the manuscript transmission of the Virtutes apostolorum, a composite tradition that takes different shapes and
contents in different manuscripts, demands the consideration of many
kinds of variance. The method of the New Philology, aiming to return
Mp. th.f.78 (not included in the selection because it does not contain a complete
series, missing sections for Peter, Paul, Andrew, and one of the two Jameses).
19. The reader, therefore, should keep in mind that my definition of the
Virtutes apostolorum does not include the complete tradition in the medieval
West, but concentrates on a geographically and chronologically defined wedge
of the cake.
20. Guy PHILIPPART, Les lgendiers latins et autres manuscrits hagiographiques, Turnhout, Typologie des sources du Moyen Age occidental 24-25,
1977, p. 16-18 ; Gisle BESSON, La collection dite du Pseudo-Abdias : un essai
de dfinition partir de ltude des manuscrits, Apocrypha 11, 2000, p. 181194 ; LIPSIUS, Die apokryphen Apostelgeschichten und Apostellegenden, vol. 1,
p. 117-178 ; JUNOD and KAESTLI (eds.), Acta Iohannis, p. 750-795 ; ZELZER (ed.),
Die alten lateinischen Thomasakten, p. xlv-liv. For a brief description of the
selected manuscripts, see ROSE, Abdias scriptor ? appendix.
21. Cf. ROSE, Virtutes apostolorum : Origin, Aim, and Use, forthcoming.
22. Els ROSE, La rcriture des Actes apocryphes des Aptres dans le
Moyen-ge latin, Apocrypha 22, 2011, p. 135-166.
23. Cf. Els ROSE, Paratexts in the Virtutes apostolorum, Viator 44, 2013,



to the manuscripts, often seems to concentrate mainly on the linguistic

variance in the transmission of a work.24 In the present article, I shall
present five different kinds of variance on different levels (from the
detailed level of individual readings to the more encompassing level of
variation in the addition or omission of paratexts and whole passages)
of the manuscript transmission of the Virtutes apostolorum. I shall also
address the problems these kinds of variance present for the editorial
presentation of these texts to a modern reader, and especially for a presentation that would reflect the liturgical or, more broadly, performative
use of these texts.

Virtutes apostolorum : Editorial Problems and Principles

The spectrum of variance in the Virtutes apostolorum embraces variants brought about either by the initial process of composition and
transmission or by the later use of these texts. The main question is
how we can devise an edition-model that accounts for and reflects these
different kinds of variance. While developing this model, I have a paper
edition in mind. Although our project has availed itself of the digital
and online editorial tool eLaborate, developed by the institute Huygens
ING (former Huygens Institute) of the Royal Netherlands Academy of
Arts and Sciences,25 the project is embedded in the apocryphal series
of Corpus Christianorum, and therefore presupposes an edition in that
series. More fundamentally, however, the methodological challenge is
to devise a new approach to text editing that would render the resulting
edition more dynamic and multidimensional, regardless of its eventual
material or virtual form.
For the present discussion, I distinguish five main categories of variants, with various subcategories :
1. Textual variants ;
2. Variants concerning the presentation of the text at the level
of individual sections dedicated to one apostle, especially
the order of the apostles, the choice of protagonists, the
choice of text and redaction ;
3. Variants involving the presence of paratexts, such as titles,
chapter titles, prologues and epilogues ;
4. Variants through corrections and uncorrected errors ; though
generally included in the domain of textual variety, this category deserves special attention ;
5. Variants inserted in the course of use of the series, especially semantic glosses (vernacular as well as Latin), mar24. See e.g. Cerquiglinis own examples in loge de la variante, and
FLEISCHMAN, Philology, Linguistics, and the Discourse of the Medieval Text.
25. eLaborate and other digital editorial tools are further introduced by
Mariken Teeuwen in her contribution to this issue. Some examples are accessible via www.e-laborate.nl.



ginal glosses indicating liturgical use, and stress marks on

words for the benefit of oral performance.
Textual variants
Textual variety in the Virtutes apostolorum is the central topic of a
doctoral thesis prepared by Maarten Prot.26 I use the term textual variance to refer to all variation in the different transmissions of the text
after it was initially written down. The term thus covers a huge number
of variants of different kinds and character, ranging from differences in
spelling and morphology or word order and choice of lexeme, to the
addition or omission of words and phrases and the omission or addition
of entire textual and paratextual passages. Strictly speaking, corrections
and uncorrected errors also belong to this category, but because of the
special position these elements could have in a future edition, I shall
discuss them separately as category (4).
I will not elaborate here on Prots methods and findings, but will
concentrate only on his conclusions concerning the charting of relations between manuscripts. The analysis of textual variety has yielded
some clear results, even if they are based on only two sections of the
Virtutes apostolorum (Bartholomew, Philip). It proves convincingly that
the transalpine continental transmission of the Virtutes apostolorum was
not uniform, but can be divided into a number of different families
or traditions. While earlier scholarship distinguished two traditions, one
Bavarian and one Frankish,27 the analysis of textual variants enables a
further refinement, showing more multiformity than hitherto assumed.
This is depicted in the following two tables, the first of which presents
the Bavarian tradition :
tradition, core
Dublin TC 73728
Vienna NB
lat. 455
Vienna NB
lat. 534
Weissenburg 48

tradition, closely
related to core
Bamberg Msc.
Hist. 139
Paris BnF lat.
Paris BnF lat.

tradition, related
to core
Munich Clm
Vienna NB 497
Vienna NB 560
Helmst. 497

BM 281
Paris BnF
lat. 12602

Paris BnF
lat. 18298

Figure 1. Bavarian tradition (textual analysis)29

26. This section is based on research performed by Maarten Prot and to
be published in Maarten PROT : Textual Variety and Linguistic Context of the
Virtutes apostolorum (PhD diss. Utrecht, forthcoming). The conclusions are presented here with the authors consent.
27. ZELZER (ed.), Die alten lateinischen Thomasakten ; JUNOD and KAESTLI
(ed.), Acta Iohannis.
28 For an explanation of abbreviations in manuscript sigla, see appendix 1.
29. Based on Prot, Textual Variety (forthcoming).



The table makes it clear that four ninth-century manuscripts form

the core of the Bavarian tradition textually, while two groups of manuscripts remain in close or more distant relationship to this core. The
fourth column shows the occurrence of manuscripts that contain much
variety in comparison with the core, while a fifth column accommodates one manuscript belonging not only to the Bavarian, but also to
the Frankish tradition.
A similar graph can be made of the Frankish manuscripts :
Frankish tradition,
group I

group II
Sankt Gallen
UL 55
SB 561
Ste Genevive 557 Paris BnF lat.
Paris BnF
lat. 5273
Graz UL 412
Paris BnF
lat. 527431

tradition, mixed tradition
Paris BnF
Paris BnF
lat. 1175033
lat. 18298
Ste Genevive

Munich Clm

Mainly Frankish I
Ste Genevive

Figure 2. Frankish tradition (textual analysis)35

The analysis of textual variants enables us to distinguish various
clearly defined groups within the Frankish tradition, unknown to earlier
The mixed tradition, indicated in both figures and best represented
by Paris BnF lat. 18298, is potentially revealing. The analysis of the
section on Bartholomew places BnF lat. 18298 in Frankish tradition II,
while the analysis of the section on Philip situates it in the Bavarian
tradition, closely related to core. Thus, the analysis of textual variants
30 A term coined by Rmi Gounelle ; see his contribution to this issue.
31 The classification of Paris BnF lat. 5274 as Frankish tradition I is

based on the section on Philip, since the section on Bartholomew is missing.

32 The classification of Ste Genevive 547 as mainly Frankish tradition
I is based on the section on Bartholomew ; however, the section on Philip
does not fit this category.
33 Textual analysis of Paris BnF lat. 11750 demonstrates that the texts
contain elements of Frankish tradition I (Philip) as well as Frankish tradition II (Bartholomew).
34 Textual analysis of Ste Gen. 558 demonstrates that the texts contain
elements of Frankish tradition I (Philip) as well as Frankish tradition II
35. Based on Prot : Textual Variety (forthcoming).
36. It should be noted, however, that Zelzer has already remarked the
greater degree of variance in the Frankish tradition : ZELZER (ed.), Die lateinischen Thomasakten, p. xxxiv.



makes it clear that, in the case of a composite tradition such as the

Virtutes apostolorum, textual analysis alone cannot answer all questions about manuscript relations. The tables indicate exactly where the
main obstacle is situated : the texts under consideration were not always
transmitted as an integral part of a coherent series. They could also
be transmitted individually, in different kinds of manuscripts and for
different purposes. Prots analysis, further elaborated in his dissertation, reveals that the section on Bartholomew is much more fluid than
the section on Philip. This seems to confirm what we know about the
general state of the different cults of Bartholomew and Philip. While
Bartholomew was highly popular in many areas of the medieval West,
Philip, in contrast, remained a subdued saint, with few personal church
dedications and a liturgy that was hardly characteristic (proprium).37 It
is not surprising to find such differences in the popularity and dissemination of the apostles cults reflected in different degrees of stability in
the transmission of texts relating to them.
Variants concerning the presentation of the text
Variance concerning the presentation of the text at the level of the
individual sections comprises four kinds of variables. The first variable pertains to the presentation of the texts as a coherent series and
concerns the order in which the apostles appear. A second variable concerns the choice of protagonists. One of the criteria used to select manuscripts for our study involved, at the minimum, the presence of the
twelve apostles. In the following, we shall see that manuscripts vary in
the addition of other protagonists, closely related to the twelve apostles,
such as the evangelists Mark and Luke or Pauls disciples Barnabas and
Timothy. The third variable is concerned with the presentation of the
narrative as such. I base my findings on a preliminary examination of
the texts, indicated by the numbers of classification used in the Bibliotheca Hagiographica Latina (henceforth BHL). The fourth variable is
the specific redaction of individual sections at a more detailed level.
This will not be analysed here, simply because the extensive work of
collation and comparison has not been completed yet. An analysis of
the first three variables confirms the outcome of textual analysis, but it
also identifies even more multiformity in the transmission of the Virtutes apostolorum as a series.
Both the order in which the apostles appear in the Virtutes apostolorum as a series and the choice of texts have been decisive issues
in the history of scholarship. Ever since Wolfgang Lazius published his

37. Cf. Els ROSE, Ritual Memory : The Apocryphal Acts and Liturgical Commemoration in the Early Medieval West (c. 500-1215), Leiden-Boston, 2009,
chapters two (Bartholomew) and four (Philip).



edition of the Virtutes apostolorum38 based on two ninth-century manuscripts from the Bavarian tradition, the series has lived on under the
(pseudo-) authorship of Abdias. Abdias plays a role in the final passages of the section on Simon and Jude as a follower of these two
apostles, as the first bishop of Babylon appointed by them, and as the
author of the narrative relating their life, acts, and martyrdom. Lazius
interpreted the epilogue mentioning Abdias (Scripsit autem gesta)39 as
a general epilogue to the series of Virtutes apostolorum as a whole. He
composed a new prologue to introduce his edition, in which he combined parts of Scripsit autem gesta with the prologue to Peter (Licet
plurima).40 In Licet plurima, the author expresses his wish to collect
the scattered knowledge about the acts and martyrdom (and not the
martyrdom alone) of all individual apostles.
The idea of a Collection of Pseudo-Abdias can only survive when
the order of the apostles and the choice of texts meet a certain expectation. The series must start with Peter, so that Licet plurima can function
as a general prologue to the collection, and it must end with Simon
and Jude, so that the epilogue Scripsit autem gesta can act as a general epilogue. Apart from the fact that this is the situation in only one
of the two manuscripts Lazius used for his edition (Vienna NB 455,
but not 53441), there is no reason, either of palaeographical nature or
dictated by content, to consider Scripsit autem gesta as an epilogue to
the series as a whole, as I have argued elsewhere.42 Likewise, I have
doubts about the general function of the prologue Licet plurima, which
can also be read as a prologue to the section on Peter or Peter and Paul
alone.43 Moreover, as we shall see below, the manuscripts differ in their
choice of a specific BHL-text for one apostle. It is clear that Laziuss
idea of a Collection of Pseudo-Abdias must be abandoned, just like
the assumption that a fixed list of specific BHL-texts, as well as one
fixed order of apostles, is characteristic of the transalpine continental
transmission of the Virtutes apostolorum as a series.
Order of apostles and choice of texts
The analysis of two aspects of text presentation (order of apostles
and choice of texts) confirms the outcome of the textual analysis and
38. Wolfgang LAZIUS, Abdiae Babyloniae episcopi et apostolorum discipuli
de historia certaminis apostolici libri decem, Basel, 1552.
39. Scripsit autem gesta sanctorum apostolorum Abdias episcopus Babyloniae, qui ab ipsis ordinatus est, sermone Hebraico BHL 7751.
40. Licet plurima de apostolicis signis sacra euangeliorum uel illa quae ab
ipsis actibus nomen accepit narret historia BHL 6663.
41. Cf. Ernst TRENKLER, Wolfgang Lazius, Humanist und Bchersammler,
Biblos 27, 1978, p. 186-203. On the problems of order in NB 534, see note
44 below.
42. ROSE, Abdias scriptor ?
43. ROSE, La rcriture des Actes apocryphes, p. 148s.



enables a further classification of manuscripts in groups more precisely

defined and typified than the rough distinction between Bavarian and
Frankish manuscripts.
The Bavarian Group, presented in appendix 2, counts thirteen manuscripts, and is typified by seven main characteristics :
1. The series starts with Peter (BHL 6663, Licet plurima)
and ends with Simon and Jude (BHL 7751 : epilogue on
Abdias) ;
2. The section on Paul consists of BHL 6575 ;
3. The hymn Praelata mundi culmina for Peter and Paul is
added to the section on Paul ;
4. The section on Andrew has both Inclita sanctorum and
Conuersante et docente (BHL 430, 429) ;
5. The section on John is Tempore illo sancti apostoli (BHL
4316) ;
6. The section on Thomas is Miracula Thomae (BHL 8140) ;
7. The section on Matthew (BHL 5690) has no prologue or
This Group is called Bavarian because, although the origin of a
number of manuscripts is unknown, the Bavarian origin of two of the
oldest manuscripts (Dublin TC 737 from Regensburg, and Vienna NB
534 from Salzburg, both dated s. IX) can be ascertained. Moreover,
Wolfenbttel Weissenburg 48 is located by some in Skt Gallen (by others in the Alsatian Weissenburg, see footnote 18 above). A considerable number of later manuscripts belonging to this Group are also made
in Bavaria : Vienna NB 560 (Abbey of Rein, Carinthia) ; Bamberg
Msc. Hist. 139 (Skt Emmeram, Regensburg) ; Munich BSB Clm 12641
(Abbey of Ranshofen, Bavaria). Angers BM 281, however, is copied in
Angers, perhaps from a Bavarian model. The same may be valid for
Paris BnF lat. 5563, owned by St. Thierry in Reims, which is close to
Dublin TC 737.
The Bavarian Group consists of five subgroups. Bavarian 1.0 is its
core, and contains six manuscripts that meet all seven criteria. Of this
core group, only Vienna NB 560 and Paris BnF lat. 5563 are identical
both in the order of the apostles and in the choice of texts. For the rest,
even if the differences are minor, no two manuscripts are the same.
Thus, Vienna NB 534, belonging to core group 1.0, deviates only in
its placement of Philip and Matthew.44 The manuscripts in subgroup 1a,
44. Despite this deviation in the order of the apostles, I classify Vienna
NB 534 as a core group 1.0 manuscript, because of its close relation to the
other manuscripts belonging to this group. The codicological situation of Vienna
NB 534 in the final parts of the manuscript is complex. The sequence of narratives (Bartholomew, Philip, Simon and Jude, Matthew) is irregular. In the sec-



1b, and 1c show various degrees of distance from the core group. Some
manuscripts deviate in only one characteristic concerning the order of
the apostles or the choice of texts. Thus, Vienna NB 497, belonging
to subgroup 1a, deviates only in its omission of the hymn for Peter and
Paul, while Paris BnF lat. 18298, belonging to subgroup 1c, inserts the
prologue and epilogue in the section of Matthew.45 At a greater distance
are manuscripts that combine several deviations from the typical Bavarian characteristics. Thus, Munich BSB Clm 12641 has both a different
order of the apostles and two or more different texts ; Angers BM 281
has a different order of the apostles, different texts, and an additional
text, the Passio Marci (BHL 5276). Bamberg Msc. Hist. 139 deviates
only in the order of the apostles, but this order is so different from the
Bavarian core group 1.0 that I classify this manuscript in the same subgroup as Munich BSB Clm 12641 and Angers BM 281, under the title
Other Bavarian, but without suggesting that there is much similarity
between them.
Finally, three manuscripts occupy a special position. Two of them
have been classified as subgroup 1b (the two Wolfenbttel manuscripts)
and one as subgroup 1c (Paris BnF lat. 18298).
The two Wolfenbttel manuscripts are, in the first place, identical
as far as the order of the apostles and content are concerned. That is
noteworthy in a manuscript tradition with so much variety. Moreover,
they include a number of important singular characteristics. Typical of
this subgroup is the addition of the prologue attributed to Melito (Volo
sollicitam esse, part of BHL 4320, the text on John) at the beginning
of the series, rewritten as a prologue to the collection as a whole.46
The palaeographical design of the prologue confirms this function.47
In the second place, the section on John includes a text on the death
of Herod, presented as a just death (digna mors). In this respect, the
two Wolfenbttel manuscripts are linked to Vienna NB 455, part of
Bavarian core group 1.0, since the text (BHL 4318) appears only in
these three Bavarian manuscripts.
Subgroup 1c (Paris BnF lat. 18298) is a remarkable manuscript. It
combines Bavarian properties (the hymn for Peter and Paul, Miracula
Thomae BHL 8140 Bavarian characteristics 3 and 6, respectively)
with features that are typical of the Frankish transmission. Most notable
is the inclusion of the prologue and epilogue in the section on Matthew,
tions on Bartholomew, Simon and Jude, and Matthew hands change repeatedly,
while in the section on Matthew there are two lacunae, one in the middle of the
story and one at the end of the manuscript. These irregularities, even if they are
not fully explained yet, raise the question whether the present organization of
quires and/or sections is the original or planned one.
45. The label Subgroup 1c is a little ponderous since the subgroup contains
only one manuscript.
46. Cf. ROSE, La rcriture des Actes apocryphes, p. 149s.
47. Cf. ROSE, Abdias scriptor ?, p. 251-252.



occurring only in the Frankish tradition (Frankish characteristic 7, see

If we juxtapose the results of the analysis of text presentation in
the Bavarian manuscripts with the results of the analysis of textual
variants, we encounter a number of parallels, even though the relative
distance between manuscripts might differ. The classification of BnF
lat. 18298 as belonging partly to the Bavarian, partly to the Frankish
transmission is the most interesting parallel. However, there are striking deviations as well. The two Wolfenbttel manuscripts, identical in
their text presentation, are at relatively great distance from one another
in terms of textual variants. Obviously, the scribe(s) who copied the
eleventh-century Wolfenbttel Helmstdt 497 from the ninth-century
Wolfenbttel Weissenburg 48 (or from a book very close to it) changed
and adapted the language considerably. The most remarkable case of
contradiction is Paris BnF lat. 12602. The textual analysis identifies this
copy as peripheral in the Bavarian tradition, but it has no place at all
in the Bavarian group when we consider the order of the apostles and
the choice of texts.
The second group defined on the basis of text presentation is called
Frankish, because the origin of its oldest representatives is known or
assumed to be Frankish : Montpellier UL 55 (Autun, though perhaps
Metz) ; Skt Gallen 561 (North-East France) ; Ste Gen. 558 (Ste Genevive in Paris ?). Paris BnF lat. 18298 stems from Paris, and may have
been copied from a Bavarian model adapted to Frankish practice. To
find Graz UL 412 in this group is surprising, given its origin (Aquileia/
Bavaria). Like the Bavarian Group, the Frankish Group exhibits seven
characteristic features :
1. The section on Peter does not contain the prologue Licet plurima, and is not BHL 6663 ;
2. The section on Paul (Peter and Paul) consists of BHL 6657 ;
3. The section on Andrew is Passio Andreae Quam oculis nostris (BHL 428) ;
4. The section on James the Less is BHL 4093 or 4094 ;
5. The section on John is the Melito-text (Secundam post Neronem, BHL 4320) ;
6. The section on Thomas is the Passio Thomae (Thomas qui et
Didimus, BHL 8136) ;
7. The section on Matthew includes the prologue and epilogue
(BHL 5690).
Based on these characteristics, we can distinguish three subgroups
(appendix 2). Subgroup 2a consists of four manuscripts that are very
close, both in the order of the apostles and in choice of texts. There are
minor differences, such as the position and choice of the text for James



the Less in Ste Genevive 547 and 558, and the position of Philip in
Paris BnF lat. 5274. Moreover, Paris BnF lat. 5274 adds a text for Paul.
The representatives of the second subgroup, 2b, are even more
strongly related. This subgroup consists of Graz UL 412, Paris BnF lat.
9737, and Skt Gallen SB 561. The only difference between the three
is the insertion of the Breviarium apostolorum into the section on Bartholomew in Paris BnF lat. 9737. This subgroup is similar to Frankish
tradition II defined by Maarten Prot on the basis of his textual analysis.
The Frankish characteristics indicate that Ste Genevive 557 and
Paris BnF lat. 5273 are also related to the Frankish Group, although
there are important deviations. I classify them as subgroup 2c, even
though these two manuscripts differ in several respects. The order of
apostles in this subgroup differs importantly from subgroups 2a and
2b. Differences in the choice of texts are found in the occurrence of
separate sections for Peter and Paul, which both manuscripts have in
common. Ste Genevive 557 gives BHL 6663 for Peter, not found anywhere else in the Frankish subgroups. The explicit of BHL 428 (Passio Andreae quam oculis nostris) differs in Ste Genevive 557, and
this manuscript adds a text on Mark (BHL 5281). Paris BnF lat. 5273
shows many similarities with subgroup 2a in the choice of texts, but
it also shows some differences : it alone among the twenty-five manuscripts uses a different text for Peter and Paul, and adds the Translatio
Bartholomaei (BHL 1004). The distance between these two manuscripts
and subgroups 2a and 2b in terms of text presentation is even more
striking given the proximity between both manuscripts and Montpellier
UL 55 (subgroup 2a) in terms of textual variants.
Singular manuscripts
Three manuscripts are at much greater distance from both Group 1
(Bavaria) and Group 2 (Francia) : Paris BnF lat. 11750 and 12602, and
Munich BSB Clm 22020. Paris BnF lat. 11750 is, despite many important differences, still related to Group 2 (Frankish) ; the other two stand
out as independent copies.
Paris BnF lat. 11750
The manuscript shares six of the seven characteristics with Group
2 (Frankish), but the order of the apostles is singular. In this respect,
the manuscript does not resemble any other. Moreover, the choice of
texts for Peter and Paul is rather unusual (even though the differences
between BHL 6663-6665 are minor). In its choice for BHL 6665, Paris
BnF lat. 11750 is alone, while it shares its choice of BHL 6664 with
Munich BSB Clm 12641, Ste Genevive 557, and Paris BnF lat. 12604.
Finally, it includes BHL 6574, also found in the later addition (s. XII)
to Dublin TC 737.



Paris BnF lat. 12602

Paris BnF lat. 12602 exhibits a mix of Bavarian and Frankish characteristics, together with some striking singularities. It lacks BHL
6663 (contra Bavarian characteristic 1) and the hymn for Peter and
Paul (contra Bavarian characteristic 3). However, it has the Miracula
Thomae (BHL 8140, Bavarian characteristic 6). It includes, on the other
hand, BHL 6657 for (Peter &) Paul (Frankish characteristic 2), the Passio Andreae quam oculis nostris (BHL 428, Frankish characteristic 3)
and the Melito-text for John (BHL 4320, Frankish characteristic 5). It
stands alone in giving the prologue for Matthew (Frankish characteristic 7) but without giving the epilogue. It is the only witness in the
selection that has a deviant incipit for the text on Philip (BHL 6813
minor difference), and is likewise unique in its addition of an important
series of sermons for Peter & Paul. Finally, it deviates greatly in the
order of the apostles, except for the final three (Bartholomew, Matthew,
Simon & Jude).
Munich BSB Clm 22020
Munich BSB Clm 22020 combines characteristics of Group 1 and
2. In its choice of BHL 6663 (Peter ; Bavarian characteristic 1) and
the Miracula Thomae (BHL 8140 ; Bavarian characteristic 6), it follows the Bavarian tradition. In its choice for the Melito-text on John
(BHL 4320 ; Frankish characteristic 5) and the inclusion of the prologue and epilogue of Matthew (Frankish characteristic 7), it is similar
to the manuscripts belonging to the Frankish tradition. In the order of
the apostles, the manuscript is singular ; even though an epilogue to
Matthew is included, and announces the section on Simon and Jude
(the following libellus), this section does not follow immediately.
The position of Paris BnF lat. 11750 and Munich BSB Clm 22020
in the analysis of text presentation (order of apostles, choice of texts)
is not surprising given the results of the textual analysis. In the latter,
Paris BnF lat. 11750 is part of the Frankish tradition, while Munich
BSB Clm 22020 stands alone, though still on the periphery of the
Frankish Group. The outcome is much more remarkable in the case of
Paris BnF lat. 12602.
From the preceding analyses follows principle 1, which stipulates
that, based on the analyses of textual variance and of variance in the
presentation of the texts as a series, the future edition will distinguish
two main parts. Part one will present the Bavarian transmission and its
subdivisions, part two the Frankish transmission and its subdivisions.
Together, the two parts will present a comprehensive picture of the
Virtutes apostolorum. A third part will accommodate editions of singular manuscripts that occupy a special position as witnesses either in
between the groups, like Paris BnF lat. 18298, or outside the groups,
like Munich BSB Clm 22020.



Choice of protagonists
While the order of the apostles and the choice of texts are, like
the textual analysis, helpful instruments to show the variability of the
transalpine continental transmission of the Virtutes apostolorum and
to refine the earlier rough distinction between the two traditions, the
matter of protagonists in the textual series is much more complex and
problematic. For even if we approach the transmission of the Virtutes apostolorum as a process of constant flux, the question here is not
so much how do we recognize and acknowledge the diversity of the
Virtutes apostolorum ?, but how far can we stretch the concept of
Virtutes apostolorum ? Moreover, the variance in the choice of texts
enables us to classify the manuscripts into groups, while the variance
in the selection of protagonists, together with the addition or omission
of other elements incidental to the series, such as hymns, sermons, and
additional texts about the twelve apostles or other protagonists, is characteristic of individual manuscripts.
Appendix 3 shows the variance in protagonists, and the different
degrees to which additional elements are integrated in the series. The
overview highlights the following phenomena :
Additional texts on the main protagonists, the twelve apostles,
are integrated or added in individual manuscripts, e.g. sermons
on Peter and Paul (BnF 12602) ; the Translatio Bartholomaei
(BnF 5273), the Acta et Obitus Iohannis (Graz UL 412) ;
Breuiarium apostolorum, Divisio apostolorum (BnF 12604) ;
Texts on other biblical saints than the main protagonists are
added or integrated ; some of them are presented as apostles
(Mathias, Barnabas BnF 12602) ; others are evangelists
(Mark, integrated in Angers BM 281 and Ste Gen 557 and
added to BnF 12604 ; Mark and Luke in BnF 12602) ;
Texts associated with the main protagonists are added or integrated (Lucan Acts NB 497, Apocalypse of John NB
455) ;
Texts on other, non-biblical saints are added.
The texts not belonging to the Virtutes apostolorum according to
the working-definition I formulated at the beginning of this article are
inserted in different ways. Sometimes, they are fully integrated, numbered consecutively, and placed between two texts belonging to the Virtutes apostolorum in the strictest sense. This is the case, for example,
in the Bavarian sub-group 1.0, where the Hymn on Peter and Paul is
incorporated after Peter and Paul and before James the Less. Sometimes, added elements follow the Virtutes apostolorum directly, and in
some cases, the numbering is likewise consecutive. Thus in Graz UL
412, the Acta et Obitus Iohannis follows directly on the section devoted
to John. In some cases, they are at greater distance, for instance the
Breuiarium and Diuisio apostolorum in BnF 12604.



The examples raise the question whether we should reconsider the

initial definition of Virtutes apostolorum based on the assumption that
the twelve form a collegium. It is difficult to formulate firm principles
in response to this problem. Here follows an attempt :

We include those texts that are integrated into the series

of sections on the twelve apostles, as well as those added
without the interruption of one or more other texts at the
beginning or at the end when a rubric indicates that they
belong to the Virtutes apostolorum. On the other hand, we
exclude texts not fully integrated but added at the beginning or the end when a rubric indicates that they do not
belong to the Virtutes apostolorum.48
This implies that we include in the edition the Hymn
on Peter and Paul in various manuscripts ; the additional Latin Virtutes Petri in BnF 5273 ; the Translatio
Bartholomaei in BnF 5273 ; the Passio Marci (BHL
5276/5281) in Angers BM 281 and Ste Genevive

48. In many cases, the rubricator helps us by explicitly marking the beginning (and sometimes the end) of the series of Virtutes apostolorum. This is
the case in BnF 12604, where a clear line is drawn between the Vita of Gregory the Great and the Virtutes apostolorum, the latter being titled as the Liber
de miraculis apostolorum (f. 3v). Conversely, a rubric (explicit) at the end of
the Acta et obitus Iohannis in Graz UL 412 includes this text in the Virtutes apostolorum : Expliciunt gesta sanctorum apostolorum et martirum Christi.
Deo gratias. Amen. Obviously, the copyist of this manuscript considered the
Acta et obitus Iohannis as part of the series of texts on apostles. Sometimes,
however, the rubricator makes it more difficult to decide what does and what
does not belong to the Virtutes apostolorum. This is the case in BnF 12602,
where a general incipit above the capitula indicates the codex as a book containing the passions of the apostles together with sermons of the Fathers and
the commemorations of martyrs and holy virgins (passiones apostolorum cum
quibus[dam] patrum sermonibus. Deinde monimenta kalendarum diue[rsorum]
martirum sanctarum quoque uirg[inum], fol. 1v). Here, it is difficult to exclude
Barnabas from the Virtutes apostolorum, as the incipit of his text identifies him
as apostolus, although his section does not follow immediately on the twelve.
It is even more difficult to exclude Barnabas in Paris BnF lat. 12604, an almost
pure apostle codex (with the exception of the Vita Gregorii Magni). The author
of the Passio Barnabae asserts that in the context of the celebration of the
apostles, the apostle Barnabas may not fail : Ergo cum sanctorum apostolorum
natalitia celebramus, dignum equidem atque salutare uidetur, ut beati Barnabe
natalicia deuotissime celebremus, quemquidem, et non est in ordine apostolatus
eorum, non tamen deest de cathalogo praedicationis, quia excellentiam miraculorum sicut Lucas euangelista refert in actibus apostolorum mirabiliter patrauit.
Nam eo fere tempore quo sancta aecclesia gratiam sancti spiritus suscepit, tertio uidelicet post passionem domini anno, cum Paulo apostolo, ad apostolatus celsitudinem, et praedicationem diuinitas electus est. Paris BnF lat. 12604,
f. 89r.



557 ; the Acta et Obitus Iohannis in Graz UL 412.

The position of the Passio Marci in BnF 12602 is
less integrated, as Mark is not presented as an apostle here. At the same time, we exclude the additional
text on Peter and the sermo on Mathias in BnF 9737
because they are added separately from the sections
on the twelve. We exclude further the Acta apostolorum in NB 497 because this text is separate from
the sections on the twelve. The in- or exclusion of the
Passio Barnabae in BnF 12602 and 12604 is dubious.

We exclude those texts that are attributed to the protagonists as authors.

This implies that we exclude the Apocalypse of John
in NB 455.


We include those texts that put forward the protagonists

as subjects (Hymn on Peter and Paul, Translatio Bartholomeae), but exclude those that deal with the protagonists as
objects (sermons on Peter).
a. This implies that we exclude the Sermones in Paris
BnF lat. 12604 and the text Cathedra sancti Petri, as
they are sermons on the apostles, not narratives with
the apostles as subjects.
b. The person of Mathias is a difficult matter in the
context of the Virtutes apostolorum. There are two
texts concerning this apostle in the manuscripts
under consideration. The first, an anonymous sermon
In natale Mathiae apostoli (inc. Cum praeclara)
appears in BnF 9737, 12602 and 12604, and must
be excluded from the edition since it is not a text on
the apostle as a subject but rather a general sermon
for the occasion of his feast-day. The Vita Mathiae
(BHL 5700-5701) appears in the late Bavarian manuscripts NB 497 and 560. It is a twelfth-century
local tradition from Treves, created at the occasion of
the inventio of his relics during the reign of Henry II
and the episcopate of Eberhard. Given the late date
of this tradition compared to the Virtutes apostolorum proper, it is not opportune to include it in the
c. Finally, inclusion or not of the additional texts Breuiarium apostolorum and Diuisio apostolorum in the
apostle codex BnF 12604 is uncertain. These texts
have the apostles as protagonists but they are not
exactly covered by the incipit Liber de miraculis
apostolorum, and they do not follow without interruption on the sections on the twelve.



The foregoing considerations and propositions lead to principle 2,

stipulating that the edition fixes, both for the Bavarian and for the
Frankish Group, a certain order in which the apostles are presented,
following the order that dominates in the manuscripts of the group
under consideration. Moreover, the protagonists presented as apostles
in the manuscripts of each group are included (figure 3). In order to
show the variety among the manuscripts of a single group as far as
order of the apostles, choice of protagonists, and choice of texts are
concerned, a route planner guides the user through the various possible sequences within a group (when following ms. Angers BM 281,
now go to Mark ; when following ms. Dublin TC 737, now go to the
Hymn on Peter and Paul, etc.). The edition does not impose one fixed
order, but offers an order of sections as a point of departure from which
the user of the edition can choose his or her way to study the series as
it occurs in a specific manuscript or group of manuscripts.
Petrus (BHL 6663)
Paulus (BHL 6574)
Hymn Praelata mundi culmina
Iacobus Minor (BHL 4089)
Philippus (BHL 6814)
Andreas (BHL 430, 429)
Iacobus Maior (BHL 4057)
Iohannes (BHL 4316)
Thomas (BHL 8140)
Bartholomaeus (BHL 1002)
Matthaeus (BHL 5690, excl.
Simon et Iudas (BHL 7750, 7751)

Petrus et Paulus (BHL 6657)
Andreas (BHL 428)
Iacobus Minor (BHL 4093)
Iohannes (BHL 4320)
Iacobus Maior (BHL 4057)
Thomas (BHL 8136)
Bartholomaeus (BHL 1002)
Matthaeus (BHL 5690 incl. pro-epilogue)
Simon et Iudas (BHL 7750, 7751)
Philippus (BHL 6814)

Additional Texts

Additional Texts

Passio Marci (BHL 5276)

Virtutes Petri (BHL 6658, 6655, 6650,


Passio Barnabae (BHL 987) ( ?)

Translatio Bartholomaei (BHL 1004)

Breuiarium apostolorum ( ?)

Passio Marci (BHL 5281)

Diuisio apostolorum ( ?)

Acta et Obitus Iohannis

Figure 3 : Division of texts and additional texts in the Bavarian and

the Frankish Groups
Variants in (the presence of) paratexts
Textual variants and the variants in presentation and selection of
texts (and protagonists) are helpful instruments for establishing the
relations among manuscripts. Another instrument that does the same is
found in the paratexts added to the texts proper : titles, chapter titles,



prologues, and epilogues.49 To begin with the latter, we have seen that
the prologue and epilogue accompanying the section on Matthew is a
characteristic feature of the Frankish Group.50 Likewise, the addition
of capitula or chapter titles at the beginning of each section is characteristic of the Bavarian tradition, even if not all sections are preceded
by capitula. The analysis of this paratextual element makes clear that
the presence of capitula depends on the geographical rather than the
chronological origin of the manuscript51 and is, therefore, a helpful
instrument for classifying the manuscripts in a certain tradition.
Thus, principle 3 demands that the future edition of the Virtutes
apostolorum take into account the presence or absence of paratexts by
distinguishing between a Bavarian and a Frankish transmission.
Variants arising from corrections and uncorrected errors
The first three kinds of variants, described above, are dictated by
the initial setup of the texts as a series, which differs per manuscript or
group of manuscripts. The fourth category is related to the particularities of their transmission, and follows from the simple fact that we are
dealing here with a handwritten tradition, for as Marco Mostert pointed
out, Not one copy of any handwritten text is identical to any other
single copy.52 One of the characteristics of a handwritten text is the
presence of a correctors hand, applying corrections either immediately
after the completion of the manuscript or at various later stages. Corrections are a kind of variance that normally takes much space in a
manuscript, but which in a modern edition, when it is made visible at
all, is often hidden in an apparatus. This is unfortunate, because corrections in a Latin text often tell us a lot about the development of the
language, if both the corrected and the uncorrected layers are visible.
Textual corrections can take different forms. Some give insight into
the development of the Latin language, but others do not. In the following example from Ste Gen. 558 (f. 33r), a superfluous a at the end of
the second line is erased to correct a scribal, not a grammatical error :

49. I elaborated on titles and intertitles (chapter titles) in ROSE, Paratexts,

50. The text is not marked in the BHL as a separate item, but it should be,
for given the fact that it is lacking in the Bavarian manuscripts, it is not an
integral part of the section.
51. Cf. ROSE, Paratexts, forthcoming.
52. Marco MOSTERT, Forgery and trust, in Petra SCHULTE et al. (eds.),
Strategies of Writing : Studies on Text and Trust in the Middle Ages, Turnhout,
2008 (Utrecht Studies in Medieval Literacy 13), p. 37-59, at p. 55.



Example 1. BSG Ms 558, f. 33r. Bibliothque Sainte-Genevive, Paris

et ipse aduolens se pedibus eius ueniam postulabat. Dicit eis [a
erased] apostolus : Multum uobis praestitit dominus meus Ihesus Christus, ut osten

This correction is not of the same order as the one found in Skt Gallen SB 561, p. 24, where the correcting hand(s) show a complex transformation in the gender of the word dolor, in the sentence Et si dolores
timendi sunt, illi sunt utique formidandi qui sic incipiunt ut numquam
finiantur. Isti enim dolores aut leue sunt et tollerare possunt, aut graue
sunt et cito anima eiciunt. Illi autem dolores eterni sunt

Example 2. Cod. Sang. 561, p. 24 Stiftsbibliothek St. Gallen

In the first line of the fragment, the gerundive timende has been
corrected to timendi and again back to timende ; the pronoun illi to ille,
the second gerundive formidandi to formidande. In the second line, the
pronoun isti is corrected into iste ; the adjectives leue and graue (line
3) already have feminine endings. In the third line, the pronoun illi is
corrected into ille, and the adjective aeterni into aeterne. While the
initial scribe alternated i and e (ae) in the endings of gerundives,
adjectives and pronouns attributed to the noun dolor[es], the corrector
was sure that the gender of dolor is feminine and corrected the forms
accordingly. The example in Skt Gallen SB 561 is informative about
the development of the Latin language (morphology, orthography), but
the example in Ste Gen. 558 is not. It would, therefore, be useful to
include in the edition corrections like the ones in Skt Gallen in order
to show transformation of the Latin language, but it is not necessary to
show corrections of the kind found in Ste Gen.



Furthermore, a choice must be made as to which layer be made visible in the edition, the first, uncorrected layer (option 1), or the layer
with corrections (option 2). It is in almost every case very difficult
to date the corrections. In certain cases, other palaeographical features
may offer some help. In Ste Gen. 558, for instance, a great number of
interlinear corrections are in the same ink colour (lighter than the main
text) as the stress marks and the marginal notes indicating a division in
lessons, which might be datable (cf. example 3). Much more frequently,
however, corrections are not datable. Although they may still inform us
about the development of the Latin language, that information cannot
be anchored in a specific historical period.

Example 3. BSG Ms 558, f. 29r : One hand corrects, adds stress marks,
and inserts (datable ?) liturgical notes in the margin. Bibliothque
Sainte-Genevive, Paris

If the edition presents the uncorrected layer (option 1) with the corrections added in the text itself between the lines or in the margin, we
get an impression of the state of Latin on the date of copying (which
we can establish approximately). At the same time, the edition gives
us an idea of the use of the manuscripts, made visible in the need to
insert corrections for a proper performance of the texts. If, on the other
hand, the edition gives the correction layer (option 2) and collects the
layer ante correctionem either in the margin or in an apparatus, the text
is presented as it was used, even though it is difficult to date that use.
This would be the most practical solution, because the corrected layer
is usually the most correct version. Previous readings can be added in
the margin with a code per manuscript. An exception will be made for
the manuscripts with many layers of corrections, such as Skt Gallen
561, Montpellier UL 55, and Paris BnF lat. 18298, which deserve a
special position in the edition.
Another issue relating to the choice between a corrected and an
uncorrected layer is the management of obvious errors. In the tradition of text editing, Bdiers plea for the choice of the best text is an
accepted practice. However, if we want to edit a text from the point of
view of its use, the best text is not necessarily the most obvious choice.
Two examples may illustrate this.
In the first place, the section on Simon and Jude illustrates the cult
of the sun and the moon by presenting two four-horse teams in the
temple of the sun that carry the statues of the sun and the moon. One
four-horse team is made of gold, in which the shining sun of cast gold



rests ; the other one is of silver and carries the moon of cast silver. In
most manuscripts of the early Bavarian transmission, through an obvious error both the sun and the moon teams are described as fusilis
ex argento. This is the case in the two Wolfenbttel manuscripts (subgroup 1b) and in all manuscripts belonging to the Bavarian core (subgroup 1.0), apart from BnF 5563, where the corrector is confused and
changes luna fusilis ex argento into fusilis ex auro. Manuscripts further removed from the Bavarian core (BnF 18298 and the three Other
Bavarian witnesses) and the thirteenth-century ms. Vienna NB 497
render the passage as expected, with a golden sun-four and a silver
moon-four. The latter version is also found in all Frankish manuscripts
without exception. Figure 4 shows the two variants as represented by a
Bavarian and a Frankish witness, both from the ninth century :
Vienna NB 455
(145r) Stabat autem in una ede templi ab
oriente quadriga solis fusilis ex argento,
in alia autem ede stabat luna fusilis ex
argento, habens quadrigam similiter
bonam fusilem ex argento.

Graz UL 412
(57r) Stabant autem in una aede templi
ab oriente quadriga equorum fusilis ex
auro in qua raditus sol adaeque ex auro
fusilis consistebat. In alia autem aede
stabat fusilis ex argento, habens bigam
biuum fusiles, similiter ex argento.

Figure 4. The statues of sun and moon in the section on Simon and
Jude, as represented by Vienna NB 455 (Bavarian subgroup 1.0) and
Graz UL 412 (Frankish subgroup 2b).
Even if the Frankish version is obviously the better text, the users
of a large part of the Bavarian tradition lived with a depiction of the
sun horses in silver as if they were equal to the moon horses. If we
want an edition to represent a living tradition as it circulated in certain (important) parts of the medieval transmission, rather than a reconstructed Ur-text, the better choice would be to include this obvious
error in the edition.
The second example is taken from the Virtutes Iohannis. Atticus and
Eugenius, two young men born wealthy but converted to the vita apostolica by John, have sold their properties to follow the apostle. When
they see their servants in splendid attire, they regret their decision to
be poor. John, aware of this change of mind, transforms pieces of
wood into gold staves and simple stones into precious stones in order
to enable the youngsters to return to their previous luxury, even if they
lose the riches of eternal life and run the risk of eternal damnation. In
the next scene, a deceased friend, Stacteus, is raised to life and urged
by John to tell his visions of heaven to Atticus and Eugenius. Stacteus
reports on the regret of the angels and the triumph of the devils about
the twosomes fearful destiny. Atticus and Eugenius hereupon ask John
for forgiveness, which he grants if they fulfil a thirty days penance,
after which the precious stones and the gold bars will return to their
proper natures. The majority of manuscripts reports that the gold bars
and precious stones change after the penance as John has foretold :



(1) Factum est autem ut .XXX. dierum transacto spatio, cum uirgae
mutarentur et petre, uenientes Atticus et Eugenius dicerent apostolo53

The narrative continues with a lament of the two young men that
they have not been forgiven. The apostle responds by promising again
that the young men will be forgiven and the magic articles will return
to their proper natures. The narrative is not fluent here. Why would the
two young men interpret the change of virgae and petrae as a sign of
lack of forgiveness ? Scattered through the Bavarian and the Frankish
Groups we find several witnesses that solve the problem : NB 497,
BnF 18298, Graz UL 412, BnF 9737, and Munich BSB Clm 2202054
give a reading with a negation attached to the verb mutarentur, which
explains the agitated reaction of the two young men :
(2) Factum est autem ut triginta dierum spatio transacto cum neque
uirge mutarentur in lignum, neque gemme mutarentur in lapides, uenientes Atticus et Eugenius, dixerunt apostolo55

If we consider the edition as a representation of the medieval use

rather than of the Ur-text, it appears that, again, presenting only text
version (2) as the best text is not helpful. The majority of manuscripts, including the entire early Bavarian transmission, but also large
parts of the Frankish transmission, give a version that is obviously less
correct, but with which the medieval users could live. Elsewhere in this
issue, Rmi Gounelle argues along similar lines to retain obvious errors
in transmission when these errors are found in several manuscripts,
and thus obviously sufficed for contemporary users.56
Thus, principle 4 stipulates that the edition makes visible those corrections that inform us about the development of the Latin language ;
corrections that do not have this function are not made visible in the
edition. The edition will also show an uncorrected error when this error
dominates an important part of the transmission in a certain group,
even if a better text is presented by other witnesses in the same group
or by another group.
Variants inserted in the course of use of the series
Textual corrections are a characteristic feature of handwritten transmission. They are at the same time indications of use, particularly when
they are inserted at a later stage. The present section deals exclusively
with additions to the text at a stage after the completion of the man53. Vienna NB 455, fol. 82v.
54. There is a lacuna in Munich BSB Clm 12641. Moreover, I was not able
to check the readings in Ste Gen. 547.
55. Paris BnF lat. 18298, fol. 48r.
56. Rmi GOUNELLE, Editing a fluid and unstable Text : The Example of the
Acts of Pilate (of Gospel of Nicodemus), p. 88.



uscript, that is, additions providing indications of use of a particular

manuscript or text. Since I discussed this phenomenon more elaborately
elsewhere,57 I will concentrate here on the implications of users additions (particularly semantic glosses, liturgical glosses, and stress marks)
for a future edition.
Some of the Bavarian manuscripts (Dublin TC 737, Vienna NB
534, and Wolfenbttel Weissenburg 48) contain semantic glosses,
explaining the meaning of a word either in Latin or in Old High German. Maril Urbanus, who studied the Viennese codex 534 as a trainee
in the project, observes a certain pattern in the 34 interlinear glosses
added to the Virtutes apostolorum in this codex, showing a particular
concern for clarifying words from non-Latin (Greek or Hebrew) origin, and uncommon Latin words.58 The same goes for the Dublin and
Wolfenbttel codices, though these codices give only a few glosses for
the Virtutes apostolorum.
Much more numerous are the manuscripts containing liturgical
glosses, marginal additions of a later date indicating the division of
the texts into pericopes to be read during the liturgy of hours. During the night office (Matins), one of the vigils was filled with readings
from the passio or vita of the saint celebrated on that day. The liturgical glosses give a clear indication of the particular use of these texts,
namely in a liturgical context, and a closer analysis shows that, for the
most part, each community applied its own divisions and selected its
own readings.59
Finally, a clear trace of use is found in stress marks, added above the
line to indicate the accentuated syllable in a word, obviously inserted
for the benefit of oral performance.
Principle 5 therefore stipulates that the edition reflect the particular uses of the texts in a performative, liturgical setting, and therefore
makes visible the additions informing us about this use as much as possible. The semantic glosses are added between the lines, while liturgical
glosses are inserted in the margins of the edited text, with a code per
manuscript. It seems to be less feasible to give account of word stresses
in all manuscripts, given the large number of these accents in some of
the manuscripts (e.g. Ste Gen. 558).

57. ROSE, Virtutes apostolorum : Origin, Aim, and Use, forthcoming.

58. Maril URBANUS, unpublished internship report, March 2012. Urbanus
adds to the existing literature (part. Arend QUAK, Unverffentliche Althochdeutsche Glossen aus dem Codex Vindobonensis 534, in Cola MINIS (ed.),
Amsterdamer Beitrge zur lteren Germanistik 4, 1973, p. 113-128) by discussing the Latin glosses next to the vernacular ones.
59. Cf. ROSE, Virtutes apostolorum : Origin, Aim, and Use, forthcoming.



Summary and Conclusions

In the foregoing, I have presented the different kinds of variance
that characterize the medieval transmission of the Virtutes apostolorum
and the problems they present regarding a future edition. I have formulated five editorial principles that aim to make visible the variance of
the Virtutes apostolorum in the edition, with regard to both transmission and use. The principles acknowledge (1) that the transmission of
the Virtutes apostolorum all over Europe, but also in the limited area of
the transalpine continent, is multiform. They respond to the fact (2) that
the Virtutes apostolorum as a series are a composite tradition, consisting of individual sections put together but not edited into a streamlined
collection. They accept the fact (3) that different manuscripts classify
different protagonists as belonging to the apostles provided that the
twelve are among them. They admit (4) that sometimes a reading in
a certain passage is obviously less correct but dominates nonetheless
in an important part of the manuscript tradition. Finally, they take into
account (5) that the Virtutes apostolorum were transmitted in a context
of performative use, more specifically liturgical services in which the
texts were recited in public, and that hence individual manuscripts have
singular characteristics with regard to traces of use and additions for
the benefit of this performative use. With the help of the formulated
principles, an edition is pursued which helps the reader to study the
Virtutes apostolorum as a whole, while at the same time enables one to
follow the specificities of one particular manuscript or group of manuscripts. This edition visualizes the multiform and composite character
of the Virtutes apostolorum as transmitted in the form of a series, giving insight into its diversity in terms of transmission and use rather
than straightjacketing it into a fixed and solidified collection.

Appendix 1 : List of manuscripts
Montpellier UL 55

s. VIIIfin/IXinc (Metz ? North Burgundy ?

Property of Autun, St. tienne)
Dublin Trinity College 737 s. IX (additions s. XII, XIII) (Regensburg)
Sankt Gallen Stiftsbibliothek s. IX2 (additions s. X, XI) (North-East
France ; possession of Skt Gallen)
Vienna NB 455
s. IX2 (West-Germany ; possession of Abbey
of Trudpert, Schwarzwald)
Graz UL 412
s. IXiii (Aquileia/Burgundy ?)
Vienna NB 534
s. IX (Salzburg)
Paris, BNF lat. 18298
s. IXfin/Xinc (Paris ; possession of Paris, Notre



Wolfenbttel Weissenburg 48 s. IX2/3 (Weissenburg/Skt Gallen, possession

of Weissenburg)
Wolfenbttel Helmstdt 497 s. XI (unknown ; possession of Abbey of St
John, Bergen, Magdeburg)
Angers BM 281
s. XI (Angers, St Aubin)
Paris, BNF lat. 5563
s. XI (unknown, property of St Thierry,
Paris, BNF lat. 11750
s. XI (unknown)
Paris, Ste Genevive 557
s. XIfin/XIIinc (unknown)
Paris, BNF lat. 9737
s. XII (unknown ; Lorraine ?)
Paris, BNF lat. 12602
s. XII (unknown ; possession of Corbie)
Paris, BNF lat. 12604
s. XII (unknown ; possession of Corbie)
Bamberg Msc. Hist. 139
s. XII (Regensburg)
Munich Clm 12641
s. XII (Abbey of Ranshofen, Bavaria)
Munich Clm 22020
s. XII (Bavaria, property of Abbey of Wessobrunn)
Paris, Ste Genevive 547
s. XII (unknown)
Vienna NB 560
s. XIIii (Abbey of Rein, Carinthia)
Paris, BNF lat. 5274
s. XII (Limoges ? Metz ?)
Paris, Ste Genevive 558
s. XIII (Paris, property of Ste Genevive ?)
Paris, BNF lat. 5273
s. XIII (unknown)
Vienna NB 497
s. XIII (unknown)

Appendix 2 : The division of the continental transalpine transmission in Bavarian and Frankish Groups and sub-groups based on text
presentation (order of apostles, choice of texts)
Group 1 (Bavarian) (numbers accompanying the names of
apostles refer to the BHL)

Paul Hymn IacMi Phil

6575 Pe&Pa 4089 6814


IacMa Ioh
4057 4316




Sim &

1 Vienna NB

X ( ?)
or 7749

2 Paris BNF lat


Pe&Pa [X,
lacuna] 6574,
(add. s.

Subgroup 1.0

3 Dublin TC

4 Vienna NB

Pe&Pa X


5 Paris BNF lat



& Jud

6 Vienna NB




Subgroup 1a
7 Vienna NB

8 Wolfenbttel- Prol.
Weissenburg Melito

Pe&Pa X


9 Wolfenbttel- Prol.

Pe&Pa X





X with X

Subgroup 1b

Subgroup 1c
10 Paris BNF lat

11 Bamberg Msc
Hist 139

12 Angers BM


13 Munich Clm




And X Ioh

IacMi Phil

IacMa X

& Jud


IacMi Phil


Pe&Pa IacMa X
6657 X

& Jud
no epi


Ioh X IacMa Phil X IacMi Barth




Group 2 (Frankish) (numbers accompanying the names of

apostles refer to the BHL)



IacMa Tho


Matth Sim & Phil

pro/epi 7750,

14 Montp X

15 Ste


16 Ste

17 Paris


but title
in table
of contents]



Matth Sim &

pro/epi 7750,

Sub- Pe&Pa
group 6657

Sub- Pe&Pa
group 6657






IacMa Tho





18 Graz


Acta et

19 Paris

X with

20 Skt G X

Ioh X

21 Ste


6576, (expl. X
6570 differs)



Mark IacMa
5281 X


Matth S&J

22 Paris

X, 6658 Pe


Barth Barth








Manuscripts not classified in Group 1 or 2


24 Paris

Pe&Pa Pe
6657 6655 6570

25 Munich Pe
22020 6663,









IacMi Phil
4089 6813




Matth IacMi Sim &

5690 4089 Jud

IacMa Barth Matth Sim &
6572 4057 1002 5690
pro/epi 7750,

23 Paris

serm. And
P&P 428

Ioh IacMa
4320 4057
6571, 430
6657 (diff.

with prowithout


Appendix 3 : Variance in protagonists and additional texts




Hymn on Peter
and Paul Fulget

Bavarian subgroups 1.0 (six

mss.) and 1c (one
Paris BNF 12602
(ms. that does
not belong to
an identified
one ms.)

Integrated between
Peter, Paul, and
James the Less

Sermons on Peter
(and Paul)

Integrated between
Peter, Paul, and

Marking of
the Virtutes

& Jud




Additional Latin
Virtutes Petri
(BHL 6658, 6655,
6650, unidentified)
(BHL 1004)
Passio Marci
(BHL 5276 in
Angers ; BHL 5281
in Ste Gen 557)

Frankish sub-group Integrated between

2c : Paris BNF lat. Peter and Paul
5273 (one ms.)
Frankish sub-group
2c : Paris BNF lat.
5273 (one ms.)
Bavarian sub-group
Other Bavarian :
Angers BM 281
(one ms.) ;
Frankish sub-group
2c Other Frankish :
Ste Gen. 557 (one


Marking of
the Virtutes

Integrated between
Bartholomew and
Integrated between
Philip and Peter &
Paul (Angers) ;
Integrated between
Philip and James
the Greater (Ste

Iohannis Acta et
Obitus (Eusebius/

Frankish sub-group Following directly

2b : Graz UL 412
on the VA, or
(one ms.)
integrated ?
numbering ; text
continues on
the same rectoside. Passiones
martyrum follow
(ca. 40)

Apocalypse of

Bavarian sub-group Following directly

1.0 : Vienna NB on the VA, or
455 (one ms.)
integrated ?
numbering ; text
continues on the
same verso-side.
No other texts

None at the
beginning (f. 1r) ;
none at the end
of the Virtutes
Iohannis BHL
4320 (f. 67r).
However, at
the end of the
Acta et Obitus
Iohannis (f. 74v)
a rubric marks
the end of the
series : Expliciunt
gesta sanctorum
apostolorum et
martirum Christi.
Deo gratias. Amen.
None at the
beginning ; none
at the end of
Simon et Iudas
(f. 146v) ; none
at the end of St
Johns Apocalypse
(f. 170v)






Sermo in natale
sancti Mathie
apostoli (Inc. Cum
Passio sancti
Marci euangeliste
Incipit de beato
Luca euangelista
Incipit passio
sancti Barnabae
Incipit passio
sancti Timothei
episcopi et

Paris BNF 12602

(ms. that does
not belong to
an identified
subgroup, one ms.)

Following directly
on the VA, or
integrated ?
numbering ;
then follow
other passiones
martyrum (ca. 25)

Bavarian sub-group Following

Sermo in natale
Mathiae (Inc. Cum 1.0 : Paris BNF lat. directly on the
12604 (one ms.)
VA (verso side),
or integrated ?
Sections are not
numbered. Other
texts on the
apostles follow.

Sermo de omnibus
apostolis ; Virtutes
Petri et Pauli
(BHL 6658) ;
Cathedra Petri ;
Sermons on
Peter ; Breuiarium
apostolorum ;
Marci ; Diuisio

Bavarian sub-group Following directly

1.0 : Paris BNF lat. on the VA, or
12604 (one ms.)
integrated ?
Sections are not
numbered. Vitae
sanctorum follow.

Marking of
the Virtutes
At the beginning
of the capitula
(f. 1v) : In
hoc codice
apostolorum cum
Deinde monimenta
sanctarum quoque
At the beginning
of the VA (f. 1ar) :
no general title or
incipit. At the end
of the VA (f. 58v) :
no general explicit.
Lacuna at the
beginning (f. 1r,
Gregory the
At the beginning
of the VA proper
(f. 3v) : Incipit
Liber de miraculis
apostolorum. In
primis de miraculis
sancti Petri.
At the end of the
VA (f. 76r, end of
Simon and Jude)
no general explicit.
See previous item.



Mathiae (BHL
5700, 5701, NB
560 ; BHL 5700,
5701/5705 NB

Bavarian sub-group
1.0 : Vienna NB
560 and NB 497
(two mss.)

Acta Apostolorum
(Lucan Acts)

Additional Latin
Virtutes Petri
(BHL 6655)
Sermo in natale
Mathiae (Cum


Marking of
the Virtutes
NB 560 : Flyleaf
Incipit prologus
in passionibus
apostolorum (Licet
End : lacunae.
NB 497 : Incipit
at the beginning
of Licet plurima
(f. 24v) : Gesta
et passiones
No general
explicit, not after
Simon and Jude
(f. 83r), not after
Mathias (f. 88v).

NB 560 :
Following directly
on the VA, or
integrated ? Later
addition ; no other
texts follow. The
sections are not
NB 497 :
Following directly
on the VA, or
integrated ?
Two passiones
follow and a
second Thomas
text (Miracula
Thomae). The
sections are not
Bavarian sub-group Preceding the VA ; See previous item.
1.0 : Vienna NB sections are not
497 (one ms.)
numbered ; section
on Peter follows
on verso side of
f. 24. However, the
Acta apostolorum
are seperated
from the Virtutes
apostolorum by
the rubric at the
beginning of the
prologue to the
first section on
Peter : Incipit
prologus in
Frankish sub-group Among the
2b : Paris BNF lat. additional texts in
9737 (one ms.)
the ms. ; sections
not numbered.

Institut fr Mittelalterforschung der sterreichischen Akademie der


The oldest extant manuscripts of the Virtutes apostolorum are
Bavarian and date to the ninth century. Elements of the collection,
however, also appear in earlier Bavarian compilations, such as a hagiographical collection from the monastery of Benediktbeuren, dating to
the last quarter of the eighth century, or a Salzburg collection of sermons from the first quarter of the eighth century. The latter is of particular interest, as individual Acts have been abridged and adapted to
fit the sermons in which they appear. This essay examines the various
compilers methods of revision and adaptation, and uses this to identify
their particular individual concerns. The reworking of the Virtutes in
the Salzburg Sermonary can be seen as part of a broader strategy of
the Bavarian metropolitan to control and use hagiography as a tool
for correction. The compiler adapted the texts to serve as a basis for
moral admonition and for the instruction of relatively recently converted Christians. He reshaped or omitted passages to avoid ambiguity
and potentially difficult discussions, for example regarding the exact
degree of consanguinity between Jesus and Jacobus or the heavenly
letters associated with Thaddaeus.
Les plus anciens manuscrits conservs des Virtutes apostolorum
sont bavarois et datent du IXe sicle. Cependant, des lments de la
collection apparaissent aussi dans des compilations bavaroises antrieures, comme dans une collection hagiographique du monastre de
Benediktbeuren, datant du dernier quart du VIIIe sicle, ou une collection de sermons de Salzbourg du premier quart du VIIIe sicle. Cette
dernire prsente un intrt particulier, car des Actes individuels ont
t abrgs et remanis pour sadapter aux semons dans lesquels ils
sont insrs. Cet article examine les diverses mthodes de rvision et
dadaptation des compilateurs, afin didentifier leurs proccupations.
Le remaniement des Virtutes dans le Sermonaire de Salzbourg pourrait relever dune stratgie plus large du mtropolitain bavarois pour
contrler et utiliser lhagiographie comme un outil de correction. Le

Apocrypha 23, 2012, p. 47-64



compilateur a revu les textes pour quils servent de base une exhortation morale et linstruction de chrtiens rcemment convertis. Il a
remodel ou a omis des passages afin dviter des ambiguts et des
discussions qui pourraient tre difficiles, notamment au sujet du degr
de consanguinit entre Jsus et Jacques ou au sujet des lettres clestes
associes Thadde.

The Virtutes apostolorum in early medieval Bavaria

The Virtutes apostolorum were quite widely transmitted throughout
early medieval Bavaria. Not only do we find the oldest extant manuscripts of this collection in this region,1 but we can even speak of
several textual variations that are exclusive to the Bavarian cultural
sphere.2 The popularity of this text becomes even more obvious when
one takes into account manuscripts that contain only certain parts of the
Virtutes apostolorum, which enables us to push its transmission and use
even further back in time.3
From the start of the 9th century onwards, these two aspects of the
Virtutes apostolorum transmission and rewriting (rcriture) are
illustrated most clearly by a specific collection of sermons and hagiographical texts known as the Salzburg Sermonary.4 By studying this
collection, we may shed more light on the various strategies of rcriture employed by those working with the Virtutes apostolorum, which
1. Dublin, Trinity College, Cod. Lat. 737 ; Vienna, NB, Cod. Lat. 534. I
would like to thank Els Rose, Maarten Prot, Monique Goullet, Cinzia Grifoni,
Marianne Pollheimer and Rutger Kramer for comments on this text.
2. Els ROSE, Virtutes apostolorum : Editorial Problems and Principles, this
issue, p. 11-46.
3. See Munich, BSB, Clm 4554 from the last quarter of the 8th century, a
manuscript from Benediktbeuern.
4. For the concept of rcriture see Monique GOULLET, criture et rcriture
hagiographiques : Essai sur les rcritures de Vies des saints dans lOccident
latin mdival (VIIIe-XIIIe s.), Turnhout, Brepols, 2005 ; Miracles, Vies et rcritures dans loccident mdival, ed. Monique GOULLET and Martin HEINZELMANN, Paris, J. Thorbecke (Beihefte der Francia 65), 2006 ; Livrets, collections
et textes : tudes sur la tradition hagiographique latine, Martin HEINZELMANN
(ed.), Paris, J. Thorbecke, Beihefte der Francia 63, 2006 ; La Rcriture hagiographique dans loccident mdival : Transformations formelles et idologiques, M. GOULLET and M. HEINZELMANN (ed.), Paris, J. Thorbecke, Beihefte der
Francia 58, 2003 ; Manuscrits hagiographiques et travail des hagiographes,
Martin HEINZELMANN (ed), Sigmaringen, J. Thorbecke, Beihefte der Francia 24,
1992) ; Lhagiographie mrovingienne travers ses rcritures, Monique GOULLET, Martin HEINZELMANN, and Christiane VEYRARD-COSME (ed), Paris, J. Thorbecke, Beihefte der Francia 71, 2010) ; Zwischen Niederschrift und Wiederschrift : Frhmittelalterliche Hagiographie und Historiographie im Spannungsfeld von Kompendienberlieferung und Editionstechnik, Richard CORRADINI,
Maximilian DIESENBERGER and Meta NIEDERKORN-BRUCK (ed), Vienna, sterreichische Akademie der Wissenschaften, 2010.



in turn enables us to document the specific interest in this corpus of

texts in early ninth-century Bavaria.
The Salzburg Sermonary was commissioned by archbishop Arn of
Salzburg (785/98-821) and was most probably addressed to the comes
Audulf, who was at the head of a great number of lay potentes in
Bavaria. The Sermonary is a collection of 145 texts. Most of these are
sermons, but there are also over thirty passions, apocryphal Acts of the
Apostles, and one vita. Originally, it consisted of two books, but only
the second of these has survived, in two distinct sections.5 In the course
of its compilation and transmission in Bavaria, this collection and its
texts left behind exceptionally clear traces of their composition and
revision. This is already visible when looking at the selection of apocrypha in the Salzburg sermon collection : texts concerning Peter and
Paul (BHL 6662), Andrew (BHL 0428), James the Great (BHL 4057),
James the Less (BHL 4089), Philip (BHL 6814), Bartholomew (BHL
1002), Matthew (BHL 5690) and Thaddeus (BHL 8011) are included,
while the Acts of John and Thomas are not.6

A combination of Frankish and Bavarian transmissions

The unknown compiler had access to both traditions of the Virtutes
apostolorum, the Frankish and the Bavarian branch.7 It is more likely,
however, that he drew on a mixed tradition that was already estab-

5. Salzburg, Stiftsbibliothek St. Peter, Cod. a VIII 32. For the manuscript
see Bernhard BISCHOFF, Die sdostdeutschen Schreibschulen und Bibliotheken
in der Karolingerzeit 2 : Die vorwiegend sterreichischen Dizesen, Wiesbaden, 1980 p. 146 ; Karl FORSTNER, Eine frhmittelalterliche Interpretation
der augustinischen Stillehre, Mittellateinisches Jahrbuch 4, 1967, p. 61-71.
Jean-Paul BOUHOT, Un sermonnaire carolingien, Revue dhistoire des textes
4, 1974, p. 181-223. Georges FOLLIET, Deux nouveaux tmoins du sermonnaire carolingien recemment reconstitu, Revue des tudes augustiniennes 33,
1977, p. 155-198. Maximilian DIESENBERGER, How Collections shape the texts :
rewriting and re-arranging Passiones in Carolingian Bavaria, Martin HEINZELMANN (ed.) Livrets, collections et textes. tudes de la tradition hagiographique
latine, Ostfildern, Beihefte der Francia 63, 2006, p. 195-224. James C. MCCUNE,
Four Pseudo-Augustine sermons De concupiscentia fugienda from the Carolingian Sermonary of Wrzburg, Revue dtudes augustiniennes et patristiques
52, 2006, p. 391-431. James C. MCCUNE, The sermons on the Virtues and
Vices for lay potentes in the Carolingian Sermonary of Salzburg, The Journal
of Medieval Latin 19, 2009, p. 250-290. Maximilian DIESENBERGER, Sermones.
Predigt und Politik im frhmittelalterlichen Bayern (Habilitationsschrift, Vienna
2011, in print).
6. Instead of a passion of John the compiler reworked CHROMATIUS OF AQUILEIA, Sermones 21 and 22, ed. Joseph LEMARI, Paris, Sources Chrtiennes 164,
1971, p. 38-57. For an edition of the abbreviated text see PSEUDO-AUGUSTINE,
Sermo 192, ed. Angelo MAI, Roma, Nova patrum bibliotheca 1, 1844, p. 449-50.
7. See E. ROSE, Editorial Problems and Principles, this issue, p. 11-46.



lished.8 Only a few of the texts can clearly be assigned specifically to

one of the branches. For example, the Passion of Andrew is clearly
a part of the Frankish tradition (BHL 428),9 whereas James the Less
comes from the Bavarian one (BHL 4089).10 Similarly, the prologue and
epilogue to the Acts of Matthew, which are present in BHL 5690, are
missing from the Salzburg Version a characteristic usually associated
with the Bavarian tradition. In the cases of Bartholomew, Philip, and
James the Great it cannot be established with certainty which branch
the models come from. The text on Peter and Paul is a stand-alone
version, but the model (BHL 6657) is used by the Frankish branch.
Likewise, Thaddaeus presents a unique version as well.
A complicating matter is the fact that in the Salzburg Sermonary
these texts were organized according to the liturgical calendar, and so
they lost any hierarchical organisation that the original model(s) may
have had. With the dispersal of individual apostles among their respective feast days, the Acts not only lost their position at the beginning
of a hierarchically-organized collection evident, for example, in the
manuscript of Benediktbeuern (Munich BSB 4554) but they also lost
their impact as a depiction of the lives of the first generation of Christians to follow Christ.11 The Sermonary thus broke with the ancient
tradition, described by Guy Philippart, of organising such texts hierarchically, which one can see for instance in Gregory of Tours Liber in
gloria martyrum,12 and later in the Sacramentaries, where the saints are
mentioned according to the hierarchy in the prayers of the Canon, and
in the Litanies.

Rewriting through abbreviation and omission

The compiler of the Salzburg collection was not content to simply
re-arrange the texts. He also interfered in the narratives. As such, all the
hagiographical texts of the Salzburg collection were rewritten according
to different considerations.

8. A similar hybrid transmission tradition is already visible in the older

collection extant in Munich BSB 4554 : Paulus (BHL 6571) ; Petrus (BHL
6657) ; Andreas (BHL 0429) ; Johannes (BHL 4320) ; Jacobus maior (BHL
4057) ; Thomas (BHL 8137) ; Bartholomaeus (BHL 1002) ; Matthaeus (BHL
5690) ; Jacobus minor (BHL 4089).
9. The Bavarian branch gives BHL 429 (Inclita sanctorum) and BHL 430
(Conuersante et docente).
10. The Frankish branch gives BHL 4093 or 4094.
11. Munich, BSB, Clm 4554, fol. 1vb-42va. See also DIESENBERGER, How
Collections shape the texts, p. 205-8.
12. Guy PHILIPPART, Les Lgendiers latins et autres manuscrits hagiographiques, Turnhout, Typologie des sources du Moyen ge 24-25, 1977,
p. 85-7.



Generally, the passion narratives were reduced in size and scope

in order to be included in the Collection. The texts were considerably
shortened in some places and sometimes elaborated in others. In fact,
the models were all quite heavily abbreviated : in the edition, the length
of each story rarely exceeds one and a half pages. Oftentimes, entire
segments of a text, such as the opening remarks and the epilogues,
were removed. The remaining portions were also abridged in various
ways, as may be seen, for instance, in a section of the Acta of the
apostle Philip that describes his missionary activity in Frigia, or in the
discussion of the three Indias in the Acts of Bartholomew.13
All in all, the narratives seem to concentrate on the deeds and faith
of the apostles, and omit the discussions, often presented in the form of
actual dialogues, about the origins of the world, the role of Jesus and
the meaning of his deeds that make up the bulk of the Virtutes apostolorum. Some minor parts remained, however, as the cuts were made
for the sake of a more compelling story.
Interestingly, allusions to the apostles appearance in Scripture have
also been excluded, as it is particularly apparent in the case of James
the Great and Andrew.14 In fact, many actual biblical quotations have
fallen victim to these editorial choices of the compiler. The passion of
James the Great, for example, one of the shortest texts in the Virtutes apostolorum, contains more than 40 references to Scripture, most
of which are contained in the apostles speech to the Jews. James
is shown to expertly use testimonies from the Psalms and prophetic
books, making it an outstanding addition to the collection and setting
this text clearly apart from the others.15 In the Salzburg version, not a
single reference had been adopted.
Similarly, the group of characters mentioned in the source material
has been limited in the reworked version. Although names of kings and
their officials are given in the abbreviated versions, the magi remain
mostly anonymous ; in the model text of the Acts of Matthew, the
demons Zaroes and Arphaxat play an important role, and they appear
again in the Acts of Simon and Jude.16 They are, however, missing
in the Salzburg collection. Likewise, in the Acts of Bartholemew, the
13. Cf. Munich, BSB, Clm 14418, fol. 13r-14r, with Biblioteca apostolica
Vaticana, Vat. lat. 5771, fol. 30(bis)v-31v.
14. See for instance the rcriture of the Passio Andreae (BHL 428), where
cc. 3-5 have totally been omitted and only small amounts of the following
chapters, sometimes only parts of a sentence, have been adopted.
15. See Evina STEINOVA, Biblical Material in the Latin Apocryphal Acts of
the Apostles, unpubl. Research Master Thesis, Utrecht 2011, p. 19-20. Available
through http : //igitur.archive.library.uu.nl.
16. For these demons and the link between the Acts of Simon and Jude
and the Acts of Matthew, see Richard Adelbert LIPSIUS, Die apokryphen Apostelgeschichten und Apostellegenden. Ein Beitrag zur altchristlichen Literaturgeschichte 2, 2, Braunschweig, 1884, p. 169-172.



names of the demons Astaroth and Beroth are missing,17 and the same
applies to the names of the idolaters throughout the text. A notable
exception is Hermogenis in the Passion of James, brother of John, most
likely because he had been converted by the apostle.18 This provides a
fitting illustration of the compilers strategy to omit all detailed descriptions of the evildoers and their allies. For instance, dragons as they
appear in the Acts of Matthew are mentioned but not the way they
killed people with their breath.19 In the Acts of Bartholemew, the long
description of the apostles appearance has been retained, but not that
of his evil counterpart in India, the demon Beroth.20
Some details actually show that the compiler used other sources to
correct his text. In the Acts of Philip, the apostle did not carry out
his missionary work in Scythia as it is said in the Virtutes apostolorum,
but in Frigia as it is reported in the Martyrologium Hieronymianum,
which he perhaps consulted.21 With regard to the Acts of Andrew, the
compiler knew that the Dei famula Maximilla cared for the dead body
of the apostle, a tidbit of information which did not derive from the
model he used, the Passio Andreae Quam oculis nostris (BHL 428).22
It is highly probable that the compiler obtained this information from
the passio-text Conversante et docente (BHL 429) that is to say, from

17. See for instance Passio sancti Bartholomaei apostoli 1, ed. Max BONActa apostolorum apocrypha 2, 1, Leipzig 1898, p. 128-150, at p. 128-9 :
Ingressus est templum in quo erat idolum Astaroth et quasi peregrinus ibi
manere coepit. Cf. Passio Bartholomaei apostoli, ed. DIESENBERGER, Sermones
(forthcoming) : erant ibi idola multa et daemonia in eis
18. Dublin, Trinity College, Cod. Lat. 737, fol. 52v-58r, at fol. 52v :
Accidit autem quondam Hermogenem magum ut discipulum suum Philetum
nomine mitteret ad eum. Cf. Passio Iacobi fratris Iohannis brevior, ed. DIESENBERGER, Sermones (forthcoming) : Erat autem quidam magus Hermogenis, qui
Filetum discipulum suum misit ad Iacobum.
19. Dublin, Trinity College, Cod. Lat. 737, fol. 106v-116v, at fol. 108r-108v :
Dracones autem erant galeati, flatus eorum flammiferum ardorem emanabat,
sulfureas spargebant auras naribus, quorum odor homines interficiebat. Cf.
Passio Matthaei apostoli brevior, ed. DIESENBERGER, Sermones (forthcoming) :
Tunc duo magi venerunt et duos secum dracones habebant, qui omnibus terrorem maximum inponebant.
20. Passio sancti Bartholomaei apostoli 2 and 7, ed. BONNET 131-2 (description of the apostle) and p. 146 (description of the demon).
21. Martyrologium Hieronymianum, ed. DELEHAYE/QUENTIN 222 : Kl. Maias.
Initium praedicationis domini nostri Iesu Christi in Frigia Hierapoli provinciae
Asiae natale sanctorum Philippi apostoli et Iacobi. For the tradition of Philip
in early medieval texts see Els ROSE, Ritual Memory. The Apocryphal Acts and
Liturgical Commemoration in the Early Medieval West (c. 500-1215), Leiden,
2009, p. 140.
22. Passio sancti Andreae apostoli (BHL 428), ed. Max BONNET, Acta apostolorum apocrypha 2, 1, Leipzig, 1898, p. 1-37.



the Bavarian tradition as well.23 As we will see, the compiler followed

a completely different path in the case of the apostle Thaddaeus, in that
he used Eusebius Ecclesiastical History as a model instead of the Acts
of Simon and Jude.

Various modes of preaching

All these abridgements, however, did more than merely simplify the
text. They also made some of its statements more radical. For instance,
in the Passion of James, brother of John, the apostle preaches to the
Jews. While in the model the apostle tried to convince the crowd that
they shared elements of belief which he showed, for example, by
referring to Abraham the Salzburg compiler reworked this section to
have the apostle simply blame the Jews for all the evil they had done
to Christ.24 Using this strategy, a rhetoric which tried to convince an
audience was replaced by threats and attempts to induce fear : if they
were not willing to convert, they would have no chance to escape the
fires of hell, the apostle predicts in the Salzburg version.25 This radical
change of the passage may have been due to an increasing hostility
towards the Jews, but it also illustrates a significant change of modalities of preaching which occurred between Late Antiquity and Carolingian times. From the three modes of preaching presented by Augustine in his De doctrina Christiana only two remained in ninth-century
Bavaria : Augustines genus submissum is transformed to lenis et blanda
oratio or dulcis sermo, and his grande genus is changed to terribilior
uel disciplinatior sermo which either threatens with the punishment
of Hell or castigates the unruly and erring.26 The more complex intermediate style (genus temperatum), which required the audience to be
even more understanding, is wholly absent from the text. Obviously

23. Passio Andreae brevior, ed. DIESENBERGER, Sermones (forthcoming) :

Stratocles vero frater eius tenens corpus sancti apostoli evasit, quem postea
Maximilla Dei famula cum aromatibus sepelivit. See Passio sancti Andreae
apostoli (BHL 429), ed. Max BONNET, Analecta Bollandiana 13, 1894, p. 374378, at p. 378 : Sic quaedam Maximilla senatrix diligens pudicitiam et castitatem et sanctitatem cum reverentia colligens corpusculum conditum aromatibus
optimo in loco sepelivit. But see also GREGORY OF TOURS, Liber de miraculis
beati Andreae apostoli c. 36, ed. Max BONNET, MGH SS rer. Merov. 1, 2,
Hanover 1885, p. 371-396, at 395 : Cuius beatum corpus Maximilla accipiens,
conditum aromatibus recondidit in sepulchro, assidue super illum depraecans
Dominum et, ut ei beatus apostolus memor esset, exorans.
24. Dublin, Trinity College, Cod. Lat. 737, fol. 55r-57r.
25. Salzburg, Stiftsbibliothek St. Peter, Cod. Lat. VIII 32, fol. 102r-106r and
26. FORSTNER, Eine frhmittelalterliche Interpretation, p. 64 : Nam in terrebilibus uerbis, id est quae uel de supplicio comminantur inferni uel etiam
neglegentes castigant et errantes ().



the Salzburg compiler chose, in the passage mentioned above, the terribilior sermo.
A similar concern for the expectations of the audience may be found
in the reworking of a sermon given by Matthew in the city of Naddaver
in Ethiopia. The compiler left out almost every argument of the apostle,
but did adopt a very short passage in which he spans an arch from
paradise up to the regnum caeleste, whose gate was opened through
Christ and which could only be achieved by true believers. Following
the description given by Matthew, we are given a clear view of what
life in paradise looked like, which plants and animals had been there
and which had not. Mice, scorpions and snakes did not exist there, and
the lions, leopards and tigers that populate the Garden of Eden are supposed to be in the service of men.27 This vivid image of a lost world
was of no interest to the Salzburg compiler, however, and he left out
almost everything. Against the model explaining the origin of the world
and the paradysum in detail, the Salzburg compiler essentially redefined
what Matthew might have meant when he talked about the believers
in Christ. He condensed the arguments of his model, and wrote simply

27. Dublin, Trinity College, Cod. Lat. 737, fol. 109r-110v : Propter uestram
salutem misit me deus ad uos, ut relicta uanitate idolorum, ad ipsum conuertamini qui uos creauit. Deus autem quando fecit hominum primum, posuit eum
in paradyso deliciarum cum uxore sua quam de costa eius fecerat. Paradysus
autem deliciarum eminet omnibus montibus, et uicinus est celo. Non habet aliquid in se, quod possit esse saluti hominis inimicum, non aues terrentur ad
sonitum hominis et aspectum, non spinae nascuntur et tribuli, non rosae marcescunt et lilia, non aliqui ibi praetereunt flores, non anni senectutem generant aetatibus, non labor fatigationibus subiacet, non infirmitas aliqua sanitati
succedit, tristicia et fletus et mors ibi penitus locum non habet. Aurae enim
quae ibi sunt blandiuntur potius quam perflant, et aeternitatem naribus inferunt.
Nam sicut tymiamatis fumus excludit putores, ita nares ibi exspirant aeternam,
quem non permittit hominem nec fatigationem incurrere, nec dolorem, sed semper equalem, semper iuuenem, semper laetum, semper immobilem permanere.
Sonant ibi organa angelorum, et uoces melliflue auribus inferuntur. Serpens ibi
locum non habet, non scorpio, non falangus, non musca aliqua saluti hominum inimica, famulantur ibi hominibus leones, et tygrides, et pardi, et quicquid
iusserit homo auibus aut feris, quasi carissimo dilecto dei obtemperant cum
reuerentia iussioni eius. Quattuor autem flumina inde fluunt, unus fluuius dicitur Geon, secundus Physon, tertius Tygris, et quartus Eufrates. Hi omni genere
piscium pleni sunt, nullus ibi est latratus canum, nullus leonum rugitus, omnia
blanda sunt, omnia mansueta, omnia quieta. Numquam ibi facies caeli nubium
tegminibus obscuratur, numquam ibi fulgura coruscant, numquam tonitrua concrepant, sed est ibi exultatio sine fine, et festiuitas quae terminum non habet.
() In die autem nouissima resurgentibus etiam regna caelestia, ut intrare
ualeant patefecit. Iam modo qui uult a morte ad uitam currat, et ad paradysum redeat, unde pater noster carnalis Adam foras missus in isto nos omnes
generauit exilio. Aperuit nobis dominus paradysi ianuas, ut ad patriam illam
reuertamur in qua mors locum non habet, in qua iuge gaudium perseuerat.



that only he who avoided sin altogether would be allowed entrance.28

Clearly the compiler had his Bavarian audience in mind, and wanted to
use this sermon to attract the people with positive imagery, in order to
drive home his admonition to lead a moral life.29
Certainly the compiler aimed to avoid any confusing and ambiguous
passages of the texts. After Matthew raised the son of a heathen king
from the dead, the rumour of this miracle spread rapidly all around
Ethiopia. Crowds hurried to the apostle, who now had to clarify that he
is not a God, but merely the servant of the only true God.30 The compiler allowed no scope for this aberration in his text. For the same reason, all biblical narratives about the temptation of Christ by the devil
were excluded from the revised texts.
The reworking of the Passion of Andrew clearly shows that any doubts
about Christ and about Gods creation in general were suppressed, even
if they were formulated by heathens. When in the model the proconsul Egeas considers the words of Jesus to be superstitiosa et vana, the
Salzburg compiler leaves out this passage, as well as the one in which
Egeas describes Andrews words as insanus sermo.31 Nothing remains
of the mockery of the proconsul, as the anonymous compiler has him
only state : Tell these words to the people, who will believe you.32 It
goes without saying that the compiler has also omitted every passage
in which those who have converted to Christianity are described as isti
28. Passio Mathaei apostoli brevior, ed. DIESENBERGER, Sermones (forthcoming) : Audite fratres qui vultis liberare animas vestras a dracone antiquo diabolo, propter vestram enim salutem misit me dominus ad vos, ut relicta vanitate
idolorum ad ipsum convertamini, qui vos creavit. Dominus enim cum primum
hominem fecit, paradysi eum colonum esse constituit, sed a diabolo temptatus
et superatus exul factus est paradysi propter transgressionem Dei mandati et in
erumnam huius saeculi missus et maledictus. Sed Christus dominus noster, quem
nos predicamus et colimus, omne humanum genus a tormento et supplicio sua
passione et morte redemit aeterno et paradysi iterum delitias et regni caelestis
omnibus sibi credentibus patefecit etiam portas. Iam modo, qui vult ad vitam a
morte venire caveat omne peccatum, ut regni caelestis mereatur introitum.
29. Most of the texts in the Salzburg Sermonary admonish the audience to
good behavior. See DIESENBERGER, Sermones (forthcoming).
30. Dublin, Trinity College, Cod. Lat. 737, fol. 111v : Ego deus non sum
sed sum seruus domini mei Ihesu Christi filii dei omnipotentis, qui me misit
ad uos ut relicto errore simulacrorum uestrorum ad uerum deum conuertamini.
Quod si me hominem similem uobis putatis deum, quanto magis hunc deum
debetis credere, cuius me seruum esse confiteor in cuius nomine ego mortuum
hunc regis filium excitaui ().
31. Passio sancti Andreae apostoli (BHL 428) 3 and 4, ed. BONNET 5 and 9 :
Aegeas dixit : Ista superstitiosa et vana verba Iesus vester dum praedicaret ()
Insanus sermo tuus praedicat ().
32. Passio sancti Andreae apostoli brevior, ed. DIESENBERGER, Sermones
(forthcoming) : Ista verba illis enarra, qui tibi crediderint See Passio sancti
Andreae apostoli (BHL 428) 6, ed. BONNET 13 : Ista verba illis narrare debes
qui tibi credunt ().



per te [Andrew] decepti sunt populi.33 Similarly, when Egeas makes a

plea to Andrew to desist from the stultitia to give his life for his faith,
this remark has been suppressed in the abbreviated version.34 By contrast, every instance where the magician Hermogenes is described as a
stultissimus homo in the passio of James, has been allowed to remain.35

A reform audience ?
Thus, the biblical stories and ambiguous passages have been significantly reduced. But it did not stop there : sometimes, if the compiler
found passages that he deemed wrong or misleading, for example, when
dealing with items of faith, he corrected or simplified them. In the Acts
of Bartholomew, for instance, the apostle notes that Christ is born with
the human being (nasci cum homine). This statement is also recorded
in the version given in BSB, CLM 4554, but the compiler of the Salzburg Sermonary changed it to the general observation that the son of
the Lord is God and that the Christ as human being is just the Son of
God. One may wonder why this passage was required to be included
and corrected at all.36 It would have been easy to take another passage
of Bartholomews speech or to omit all the content concerning questions of faith. After all, it had clearly not been the compilers intention
to teach such matters to his Bavarian audience. To him, making this
compilation did not involve instructing the Bavarian lay potentes, or
repeating articles of doctrine, which often made up the larger part of

33. Passio sancti Andreae apostoli (BHL 428) 6, ed. BONNET 19.
34. Passio sancti Andreae apostoli (BHL 428) 8, ed. BONNET 19 : () revocare animum tuum ab stultitia () stultum est enim ultro velle ad passionem
crucis ire et ignibus ac flammis te ipsum pessimus destinare ().
35. Dublin, Trinity College, Cod. Lat. 737, fol. 53v : Stultissime hominum
cum inimicus generis humani rationem tecum haberet, quare non consideras
quem rogasti ut mitteret ad te angelos suos ad laesionem meam quos ego adhuc
non permitto ut furorem suum tibi ostendant ? (...) Cf. Passio sancti Iacobi
fratris Iohannis brevior, ed. DIESENBERGER, Sermones (forthcoming) : Stultissime hominum cur non consideras te a demonibus seductum et ab eis vexari ?
See also Passio sancti Bartholomaei apostoli 1, ed. BONNET 129. Cf. Passio
Bartholomaei apostoli brevior, ed. DIESENBERGER, Sermones (forthcoming) : ()
erant ibi idola multa et daemonia in eis, quae stultis hominibus videbantur
curare languentes ().
36. Passio sancti Bartholomaei apostoli 4, ed. BONNET 134. : Ego autem nihil
terrenum, nihil carnale desidero. Unde scire te volo quia filius dei dignatus est
per uterum virginis nasci cum homine, ita ut homo in virginis vulva conceptus
secum in ipsa secreta virginis deum qui fecit caelum et terram (). Cf. Passio
Bartholomaei apostoli brevior, ed. DIESENBERGER, Sermones (forthcoming) : Ego
autem nec terrena quaero nec diligo, sed vos, ut cognoscatis Deum verum filium
Dei, etiam hominem verum natum ex virgine, qui sine initio semper ante omnia
saecula deus erat cum patre et ex virgine initium accepit homo esse, et omnibus
viventibus initium dedit.



the original stories. It was assumed that the audience of the Salzburg
Collection was familiar with matters of doctrine, since these passages
were largely left out of the reworkings of the Passions and the Acts
of the Apostles. Perhaps, then, this particular revision was carried out
in connection with the heresy of Adoptionism, which was discussed
vigorously at the Carolingian court at this time.37 Similar concerns with
contemporary discussions can also be seen in the abbreviations and editorial choices made in the reworking of the passion of saint Euphemia
for the Sermon Collection.38
The editor of the hagiographic material in the Salzburg Collection had not only made his texts more compatible with and compliant to the ideology behind the Carolingian religious reforms, but he
also responded to earlier hagiographic traditions which had, for various reasons, been discredited. This is the case, for example, with the
text entitled De Thateo apostolo. Eusebius. V Kl. Novembris. Bedes
martyrology places the feast of SS. Simon and Judas (i.e. the apostle
Judas Thaddaeus) on this date.39 Although the Salzburg Collection duly
includes a text referring to Thaddaeus, this is not the Thaddaeus named
as one of Jesus twelve disciples (cf. Matthew 10 :3), but another Thaddaeus who is mentioned as one of the seventy disciples of Christ. This
latter Thaddaeus, a disciple of St Thomas, is thought to have been
involved in missionary activity in the kingdom of Edessa and to have
met Abgar, the local ruler there. The story of the meeting is related in
Eusebiuss Ecclesiastical History, and a colophon in the text explicitly
shows that the compiler was familiar with that source. Eusebius claims
that Abgar had corresponded with Jesus, and that the Lord had replied
in a letter that he would send an apostle to Edessa. And indeed, after
Thaddaeus had been sent to Edessa by the apostle Thomas subsequent
to the ascension of Christ, he was recognized by the king as a representative of the Lord, who saw a familiar feature appear in his face. He
was healed and subsequently converted by Thaddaeus.40
37. See e.g. Concilium Francofurtense a. 794, ed. Albert WERMINGHOFF,
MGH Leges 3, Hanover/Leipzig, Concilia aevi Carolini 2.1, 1906, p. 165-171,
at 137 : Nihil enim humana nativitas divinae praeiudicavit nativitati : adsumpsit
quod non erat ; permansit in id essentialiter quod erat, non commixtione passus neque divisione, sed in una persona Christi Deus verus et homo permanet
verus ; non duo filii Deus et homo, sed unus filius Deus et homo, non alius filius
hominis et alius Dei, sed unus idemque Dei hominisque filius. See also PAULINUS
OF AQUILEIA, Libellus sacrosyllabus Contra Elipandum, ed. J. P. MIGNE, Patrologia Latina 99, 1864, col. 151-166, at col. 160 : () sed in una persona Christi
Deus verus et homo permanet verus, non duo filii Deus et homo ().
38. See DIESENBERGER, Sermones (forthcoming).
39. See also Das Salzburger historische Martyrolog, in Meta NIEDERKORN,
Das Sanctorale Salzburgs um 800. Liturgie zwischen Norm und Praxis, unpubl.
Habilitationsschrift, Wien 1999, p. 404.
40. EUSEBIUS OF CAESAREA, Church History I, 13, ed. Eduard SCHWARTZ/Theodor MOMMSEN, Eusebius, Werke 2. Band 1. Teil. Die griechischen christlichen



In the centuries that followed Eusebius narration of this story, several problems emerged that vexed the generations after him. For one,
the Decretum Gelasianum qualifies the exchange of letters between
Jesus and Abgar, extensively quoted in the course of the Historia Ecclesiastica, as being apocrypha.41 Although the correspondence between
the king and Christ was supposed to have taken place during the latters lifetime, it was nonetheless placed among the so-called Heavenly
Letters that still popped up from time to time in the eighth century. For
example, the years 744/745 saw the infamous case of the heretic Aldebert, who, among other things, claimed to have received a letter from
Jesus himself.42 Even the Admonitio Generalis of 789 warned against
such pseudographia et dubiae narrationes, as did the Libri Carolini
some time later.43 An especially poignant motif in the eighth century,
however, was that of the Epiphany, which had in the sixth century been
expanded by Evagrius Scholasticus to include real images of Christ,
which turned it into an important element in the discussion on the veneration of images especially during the Second Council of Nicaea of
787.44 This is why the Libri Carolini devote an entire chapter to the
legend of Abgar, attempting to prove that Abgars revelation was never
really an existing image of Jesus.45 But still, the involvement of an
authority like Eusebius in the preservation of parts of this story nourished belief in such dubiae narrationes for a long time all the same.
It could have been for this reason that, instead of the Acts of Simon
and Judas Thaddaeus, the text written by Eusebius was reworked and
put into the Salzburg Collection. The compilers copied the passage
about the meeting between Thaddaeus and the king of Edessa rather
exactly, but duly omitted any references to the Heavenly Letters or to
Schriftsteller der ersten Jahrhunderte, Neue Folge Band 6.1, Berlin, 2. unvernderte Auflage, 1999, p. 83-97. See LIPSIUS, Apostelgeschichten 2.2, p. 154-163.
41. Decretum Gelasianum. (Das Decretum Gelasianum de libris recipiendis
et non recipiendis) 5, 8, ed. Ernst von DOBSCHTZ, Texte und Untersuchungen
zur Geschichte der altchristlichen Literatur 38.4, Leipzig, 1912, p. 57 : Epistula
Iesu ad Abgarum apocrypha. Epistula Abgari ad Iesum apocrypha.
42. Nicole ZEDDIES, Bonifatius und zwei ntzliche Rebellen : die Hretiker
Aldebert und Clemens, in Marie Theres FGEN (ed.), Ordnung und Aufruhr
im Mittelalter. Historische und juristische Studien zur Rebellion, Frankfurt am
Main, Ius Commune Sonderheft 70, 1995, p. 217-263.
43. Admonitio generalis c. 78, ed. Alfred BORETIUS, MGH Leges 2, Capitularia regum Francorum 1, Hanover, 1883, p. 53-62, at p. 60.
44. EVAGRIUS SCHOLASTICUS, Ecclesiastical History IV, 27, trans. Michael
WHITBY, Translated Texts for Historians 33, Liverpool 2000, p. 226 ; Sancta
Synoda Septima Generalis, Nicaena secunda, Anastasio bibliothecario interprete, ed. J.-P. MIGNE, PL 129, Paris, 1879, col. 195-512, at col. 368D ; see also
HADRIANUS I PAPA, Epistola 2, ed. Ernst DMMLER, MGH Epistolae 5, Berlin,
1899, p. 23.
45. Libri Carolini c. IV 10, ed. Ann Freeman, Hanover, MGH Concilia 2,
Supplementum 1, 1998, p. 511.



the manifestation to Abgar. As a result, the Abgar legend was transformed into a harmless missionary story from which any inappropriate
details had simply been omitted. It is not clear if the Salzburg compiler
had access to the Acts of Simon and Jude at all. But if he did, he saw
the need to replace it in favour of the revised Abgar legend, because
it was clear that the foundations of any devotion to the Apostles that
was unlawful or in excess of proper veneration had to be eliminated.
This also explains why the Acts of Thomas are lacking. BHL 8140, the
Bavarian version of Thomas, opens with a reference to Thaddaeus and
the Abgar-legend.46
On the other hand, it is also worthwhile to consider those passages
which, in contrast to the extensive abridgements, were carried word
for word into the newly compiled texts. Thus the passages concerned
with the structure of dioceses and churches in the missionary region
of the Apostles, as may be seen in the Acts of Philip. In this piece,
while the immediately preceding description of Christs works and the
meaning of his passion, resurrection et cetera, are removed without
substitution, we are informed that Philip baptized thousands of people, ordained clerics, bishops, priests and deacons, and built numerous
churches.47 The importance of these elements for the Bavarian compiler
is shown even more clearly in the Passio Jacobi brevior, where the
anonymous compiler expands on the organizational performance of the
apostle beyond what he found in his model : ...when the Apostles of
the Lord preached the Word of the Lord in various places and established churches and episcopal seats (pontificales sedes)...48
This projection of a stable, organized Church into the early Christian missionary areas was intended to point out to the lay potentes,
who were as a group the users of this text, the long-established and
approved ecclesiastical customs and structures from which the much
46. See e.g. Vienna, NB, Cod. Lat. 455, fol. 91r. I owe this observation
to Els Rose.
47. Dublin, Trinity College, Cod. Lat. 737, fol. 23r-24v, at 23r : Haec et his
similia dicente Philippo apostolo crediderunt, et eliso Martis simulacro multa
milia hominum baptizata sunt. Ordinatis autem clericis, sed et ordinato eis
episcopo, et presbiteris et diaconibus atque ecclesiis multis instructis, ipse per
reuelationem ad Asiam reuersus, in ciuitate Hieropolim commoratus, heresim
malignam Hebeonitarum extinxit, qui dicebant non uerum hominem natum ex
uirgine dei filium adsumpsisse.
48. Dublin Trinity College, Cod. Lat. 737, fol. 20r : Tempore quo una annorum post passionem dominicam septimana completa est, ecclesia dei Hierusalem
constituta copiosissime multiplicata crescebat per Iacobum, qui a domino ordinatus est in ea episcopus. Compare with Passio Iacobi fratris Christi brevior, ed. Diesenberger, Sermones (forthcoming) : Non multo post ascensionem
Domini tempore, cum apostoli Domini per diversa loca verbum Domini praedicarent atque ecclesias et pontificales sedes aedificarent, coepit Iacob regere
ecclesiam quae erat Hierosolimis ubi ab omnibus ceteris est apostolis episcopus
ordinatus ().



younger Bavarian Church ultimately derived its authority.49 After all,

one should not forget that the Bavarian church itself is based on the
apostolic tradition. The most powerful sees in Bavaria, Salzburg and
Regensburg, for instance, were dedicated to Saint Peter.

A church for the gentes

By appropriating this heritage of successfully evangelizing entire
gentes, and organizing the administration of the new provinces, the
compiler of the Salzburg Collection indicated to his audience their role
in relation to the Church. This is an important point in the compilation, for the Bavarian lay potentes were entrusted with juridical affairs
emerging from the interaction between the secular and clerical spheres.50
At the same time, this reference to the successful administration of
the early Church expresses in no uncertain terms the need for clear
structuring and organisation of the apostolic Acts. Hence in the Salzburg Collection, the precise position of James in Jerusalem at the time
of the meeting of the Apostles which is somewhat unclear and misunderstood in the biblical account is smoothed over. Whereas the Virtutes apostolorum, in encountering the same problem, state that James
had been ordained as bishop by the Lord, the Bavarian compiler claims
that he had been selected as bishop by all the other apostles.51 This was
a very deft move, smoothing over the conflicts between the apostles
attested in the biblical account and in the apostolic letters.
Here, however, the difficulties with James and his position in relation to his fellow apostles only begin. In the New Testament, the name
James is mentioned numerous times, and not all of these instances can
be clearly identified as referring to a single person.52 Also, the identification of James, the brother of the Lord, with James Alphaeus and
James the Younger a custom already long-established by 800 was,
with closer readings of the Bible, questionable at best ; also, the possibility that Jesus had had a brother at all was a very delicate issue.
The unknown author of the Virtutes Jacobi fratris domini (BHL 4089)
assumes this identification but does not expand on it, although he does
recognize that at the time of Jesus the name James was very com49. See Maximilian DIESENBERGER, Sammeln und Gestalten Erinnern und
Vergessen. Erzbischof Arn von Salzburg und die Ursprnge des Salzburger
Episkopats, in Walter POHL (ed), Die Suche nach den Ursprngen, Wien, Forschungen zur Geschichte des Mittelalters 8, 2004, p. 171-189. IDEM, Hagiographie et rforme en Bavire au tournant des VIIIe-IXe sicles, Mdivales 62,
2012, p. 67-82.
50. See DIESENBERGER, Sermones (forthcoming).
51. See note 48.
52. See Wilhelm PRATSCHER, Der Herrenbruder Jakobus und die Jakobustradition, Gttingen, Forschungen zur Religion und Literatur des Alten und Neuen
Testaments 139, 1987, p. 199-206.



mon.53 Employing another strategy, the compiler of the hagiographical

manuscript from Montpellier, for instance, avoids the problem by not
mentioning at all that the apostle was the brother of the Lord, simply
leaving out the first few paragraphs.54
The Bavarian compiler, however, took an entirely different route : in
contrast to his normal modus operandi, he inserted a passage explaining the nature of the blood relationship between Jesus and James. Via
the explanation used by Jerome contra Helvidium which the Church
Fathers later had to abandon because of extensive opposition James
becomes the son of Marias homonymous sister and Joseph Kloppas,
hence a cousin of Jesus, only referred to as brother because they ex
eadem cognatione sunt geniti.55
In this case, therefore, the compiler was not smoothing over a fantastical story of the kind so prevalent in the apocryphal Acts, but was
explaining a misunderstood position of the New Testament. There had
probably been hard-to-control rumours concerning the dubious virginity
of Mary bound up with the James tradition, and this approach helped
to dispel them.

The relation between sermons and passions

Finally, we can see that the compiler did not limit himself to simply rewriting the hagiographical texts according to his interests or in
order to tighten up the narratives and to remove ambiguities. He also
established specific relations between the sermons and passions. In the
Salzburg collection, one can find sermons and passions for one and the
same feast day. This is the case, for instance, on June 29, when the dies
natalis of Peter and Paul is solemnized. First, the collection offers the
passion of the two apostles. The story starts with the two men arriving
in Rome and preaching there to the people. Because of their great success, the magician Simon accuses them of being false magicians and
seducers. Oddly enough, he starts to show his own abilities to perform
magic. But while he is doing senseless tricks like bringing stone snakes
to life, Peter shows that, through the aid of the Lord, powerful miracles
like healing the sick and curing cripples may be performed. Needless
to say, the contest between magical arts and Christian miracles is won
by the apostle, as is the ensuing fight against Nero, who was initially
53. Dublin, Trinity College, Cod. Lat. 737, fol. 20r-23r.
54. See Montpellier, UB, FM, Cod. 55, fol. 11v-12r, which belongs to
the Frankish branch. This manuscript is the oldest witness of this short version (BHL 4093). Incipit beati Iacobi apostoli quod est Kalendis Aprilibus. On
Jesus kinship, see further Rita Beyers, The Transmission of Marian Apocrypha
in the Latin Middle Ages, this issue, p. 134-135.
55. HIERONYMUS, De virginitate beatae Mariae contra Helvidium, 13-14, ed.
J.-P. MIGNE, PL 23, Paris, 1845, col. 181-206, at col. 195-197. See PRATSCHER,
Herrenbruder, p. 204.



amongst the captivated audience but erupted in anger when Peter

started to teach about the almighty God.56 The unknown Salzburg compiler used the version presented by Pseudo-Marcellus as model for his
abbreviated story and concentrated on the fight between Simon Magus
and Peter as well as on its effects on the audience, and in having done
so, he simply omitted all the theological issues involved.57 On the other
hand, the compiler did emphasize that the people watching were actually choosing sides between Peter and Simon ; many of them believed
in the magical power wielded by Simon, and were therefore unable to
distinguish between true and false, between good and evil.58
Still, as a rule, the abbreviated Passio does not dwell on moral
explanations the many sermons accompanying the passiones in this
compilation, however, provided ample opportunity to do so. The story
of Peter and Paul, for example, is followed by a sermon based on a
sermon of Maximus of Turin, which was reworked more or less in line
with the passion.59
God gave Peter the keys to the kingdom of heaven, while Paul
was the idoneus doctor. Both of them are equally important, states the
preacher, saying that this is apparent from the fact that they died on
the same day through the will of a tyrant. With this statement the compiler of the text departed from the model, and added moral conclusions.
These conclusions have great significance for the whole collection. The
martyrium of Peter and Paul should serve as an exemplum for the audience, the preacher starts, and the text then continues as follows :
If we want to remain in malice, we have to endure even worse
torment, because we knew, saw and heard, that we have to do good
deeds, but we presumptuously despised it. For if the divine commandments through which we are admonished to incite good deeds do not
exist, the exempla of the saints however serve us as a law and the Lord
teaches us with his commandments and the life of the saints provides
us with exempla of good deeds. Now there is no excuse for this accusation because the law of the Lord daily knocks at our ears and the inner
evidence of the good things done provokes our heart, and I know, if we

56. Passio Petri et Pauli brevior [BHL 6662], ed. DIESENBERGER, Sermones
57. PS.-MARCELLUS, Passio sanctorum apostolorum Petri et Pauli [BHL
6657], ed. LIPSIUS 119-177. See also Montpellier, UB, FM Cod. 55, fol. 2r-8v.
Wolfenbttel, Weissenburg, Cod. 48, fol. 22v-32v ; Sankt Gallen, Stiftsbibliothek, Cod. 561, p. 3-20.
58. Passio Petri et Pauli brevior [BHL 6662], ed. DIESENBERGER, Sermones
(forthcoming) : et ita contigit inter se populum esse divisum. Alii cum Petro,
alii cum Symone pergebant.
59. Maximus of Turin, Sermo 1, 1-2, ed. Almut MUTZENBECHER, Turnhout,
CCSL 23,1962, p. 2-3.



follow the exempla of the depraved often, why do we not imitate those
done by the saints worthy and pleasing to the Lord ?60

The preacher established a connection between the supporters of

Simon Magus and Peter on the one hand, and his Bavarian audience on
the other. In this regard, all allusions to misbehaviour in the sermons
are as connected with the evil-doers who appear in the Passions read
on the feast days of the martyrs as the good deeds documented in these
hagiographical texts. In the case of Peter and Paul, some scribes did
see the connections between the evil deeds committed by Simon Magus
and the moral commentary provided in the subsequent sermon, and thus
changed the wording of their exemplar. For example, not many of them
believed in the militia of Simon (the evil words he spoke about Peter),
but opted to read this as referring to his malitia instead.61
In fact, in most of the sermons and passions of this collection, the
audience is called on to follow the principal moral standards and to
refrain from bad behaviour. In order to make this task easier, the compiler modified the texts to bring forth the principal morality underlying
the evil deeds of the judges, kings and other people in the passions.
When it came to providing examples to the Bavarian audience, the
compiler could draw upon the deeds of evildoers just as upon the acts
of the holy men and women that form the basis of his narrative.

In conclusion, the Virtutes apostolorum as a whole is preserved in
two Bavarian manuscripts of the ninth century. Traces can be found
even earlier in the hagiographical manuscript of Benediktbeuren, where
parts of the Frankish, Bavarian and Italian branches had been com-

60. Sermo de natale apostolorum Petri et Pauli, ed. DIESENBERGER, Sermones

(forthcoming) : Si autem in malitia perseverare volumus, graviora etiam tormenta sustinere debemus, quia scimus et vidimus atque audivimus, quid boni
facere debemus, sed hoc presumptiosae contempnimus. Si enim ad boni incitamentum divina quibus admoneremur praecepta deessent pro lege nobis exempla
sanctorum proficerent ; ad contrarium et nos deus praeceptis suis docet et vitam
sanctorum boni operis exempla nobis proponit. Nulla est iam de reatu excusatio
quia et lex Dei aures nostras cottidie pulsat et factorum documenta bonorum
cordis nostri intima provocat nosquoque si pravorum sepe secuti sumus exempla
cur non imitamur sanctorum deo digna et bene placita facta ?
61. Passio Petri et Pauli brevior, ed. DIESENBERGER, Sermones (forthcoming) : Hinc sediciosus populi murmur inolevit et Simon in zelum militiae excitatus de Petro multa mala dicere coepit dicens eum magum esse et seductorem.
Multi autem eius militiae credebant quia multa signa arte magica faciebat. See
for instance Wrzburg, UB, Wrzburg, M p. th. q. 15, fol. 166v : Hinc sediciosus populi murmur inolevit et Symon in zelum malitiae excitatus de Petro multa
mala dicere coepit, dicens eum magum esse et seductorem. Multi autem eius
malitiae credebant quia multa signa arte magica faciebat.



bined. The Bavarian branch had clearly been in existence at that time,
and one can discern when and where it came from. At the beginning
of the ninth century, parts of the Virtutes were used as a basis for sermons. Because of the chronological order of the compilation, a preface
or epilogue was not necessary. Perhaps this is one reason why the text
on Simon and Jude is not preserved in this compilation. The use and
revision of parts of the Virtutes apostolorum show, on the one hand, the
importance of apostolic tradition in Bavaria. On the other, one can see
the attempt to control and manage the access to these traditions. Too
much ambiguous and controversial information was transmitted within
these texts, and they could even be used to clarify positions within contemporary conflicts. Interestingly enough, all these aspects also affected
the lay flock, which gathered regularly in country churches very often
dedicated to an apostle. In this context the reworking of the Apostle
Acts in the Salzburg Sermonary can be seen as part of a broader strategy to control and use hagiography as a tool for correction. To the
compiler, the texts he used should serve as a basis for moral admonitions, rather than evoke or strengthen uncontrollable narratives.


Universiteit Gent

The transmission and reception of hagiographical texts is the central concern of the international research project An Assessment of
Merovingian hagiographical texts in their oldest manuscript versions
directed by Monique Goullet.1 This project advocates a return to the
oldest hagiographical manuscripts in order to free research from the
distorting impact of old editions, and to bring it in line with recent
scholarly advances in both hagiographical and sociolinguistic studies.
A return to the oldest hagiographic manuscripts can, on the one hand,
lead to a better understanding of the way in which the first anonymous
hagiographic collections were constituted, and, on the other, shed light
on the way in which written language functioned in the pre-Carolingian
period. It is within this context that the research team chose to edit
one of these oldest manuscripts, Torino, Biblioteca Nazionale D.V.3.
From a linguistic point of view, this document dating from the end of
the 8th or the beginning of the 9th century is a precious witness to the
management of written language. It also instructs us on the way its
producers handled the inheritance of Latin hagiographical texts written between the 5th and the 8th century. This article takes stock of the
linguistic analysis of the collection and of the interpretation of its linguistic features.
La transmission et la rception de textes hagiographiques sont au
centre de lintrt dans le projet de recherche international Expertise de textes hagiographiques mrovingiens dans leurs plus anciennes
versions manuscrites dirig par Monique Goullet.2 Ce projet prne
un retour aux plus anciens manuscrits afin de saffranchir des ditions

1. Cf. Monique GOULLET, Expertise des textes hagiographiques mrovingiens dans leurs plus anciennes versions manuscrites. Prsentation dun projet
de recherche collectif, Hagiographica 18, 2011, p. 67-88.
2. Ibidem.


Apocrypha 23, 2012, p. 65-80



anciennes et dtre mieux en phase avec les avances scientifiques

actuelles tant sur le plan des tudes en hagiographie quen sociolinguistique rtrospective. Un retour aux manuscrits hagiographiques
les plus anciens permettra de mieux saisir quand et comment se sont
constitus les premiers groupements textuels hagiographiques anonymes dune part, et permettra dautre part de contribuer une meilleure comprhension du fonctionnement de la langue crite lpoque
prcarolingienne. Cest dans ce contexte quil a t choisi dditer lun
de ces manuscrits, le manuscrit Torino, Biblioteca Nazionale D.V.3.
Du point de vue linguistique, ce document datant de la fin du VIIIe ou
du dbut du IXe sicle, est un observatoire intressant de la gestion de
la langue crite. Un tel lgendier nous instruit en effet sur la manire
dont ses producteurs graient un patrimoine de textes hagiographiques
latins crits entre le Ve et le VIIIe sicle. Larticle fait le point sur
lanalyse linguistique du recueil et sur linterprtation des traits linguistiques observs.

Studying written texts of the Early Middle Ages : why (not) ?

The specific ways written language functioned in the Early Middle
Ages, and particularly the relation between (written) Latin and (spoken) vernaculars in the Romance speaking areas, is a research issue
that is far from being resolved.3 If it is true that the attitude towards
the texts has evolved significantly over the past fifty years, a certain
mistrust towards written documents is never completely absent.4 Linguistic research on this period has indeed suffered for a long time from
the impact of an old dream : to retrace the genealogy of the Romance
languages up to their origins. From this viewpoint, most of the written
documents were perceived as if they were only of secondary importance. Their learned nature could not reflect the real state of the spoken
language. They were thought of as a screen that masked the living lan3. See the Acts of the International Conferences on Vulgar and Late Latin
(LVLT), especially the 2009 session : Frdrique BIVILLE, Marie-Karine LHOMM,
Daniel VALLAT (ed.), Latin vulgaire - latin tardif IX, Lyon, Maison de lOrient
et de la Mditerrane, 2012. In the programme of this session, one can read :
La priode cruciale est celle des Ve-IXe sicles, une priode historiquement
trouble et complexe, qui conduit progressivement et insensiblement de la fin
de lEmpire romain et de lAntiquit lmergence du monde mdival. Cette
priode encore mal connue linguistiquement est prcisment celle de lvolution
et de la diversification du latin en langues romanes, et des profondes mutations
historiques et socioculturelles quelle rvle.
4. Cf. Michael MAZZOLA, Lanalyse phonologique au service de la mythologie, in : Frdrique BIVILLE, Latin vulgaire - latin tardif IX (see above, note
3), p. 189-201. The abstract of this paper considered that : ... nous narrivons
savoir que trs peu de la varit parle du roman qui se dtache du latin
surtout cause du manque de tmoignage crit et du dcalage entre les
deux variantes.



guage and that was of limited use to the reconstruction of the steady
metamorphosis of Latin into its Romance daughter languages, except
for occasional Romance structures that could break through here and
Nevertheless, studies in both language philosophy and sociolinguistics show more and more clearly that language should be conceived as
a complex intermixing of language uses.5 Different uses and different
norms coexist, depending on the medium, the context and the discourse
traditions.6 Some uses are fit for some linguistic acts, others for other
ones. All speakers of a language do not necessarily master all of its
expressions. They can master some of them actively, others passively.
This means that the most interesting question concerns the interfaces
between the different uses and norms.
The relation between written and spoken language should also be
considered from this point of view : as part of a continuous, multiform
and changing reality. Written language is particular because its durability, compared to the ephemerality of spoken language, pulls it out of
the present and opens up the gates of time. As such, written language
confronts its users with the possibility and the necessity of promoting and ratifying certain values and condemning others. Some linguistic features are considered as fit for written language, while others are
not.7 In this sense, written language presents a reflection of the relationship between a society and its language, especially through the action
of its literate community. It reflects the way in which a society both
5. Essential readings on this subject include Eugenio COSERIU, Sincrona,
diacrona e historia. El problema del cambio lingstico (secunda edicin,
revisada y corregida), Madrid, 1973 ; William LABOV, Sociolinguistique, prsentation de Pierre ENCREV, traduit de langlais par Alain KIHM, Paris, Les ditions de minuit, 1976 (Original edition : Sociolinguistic patterns, Pennsylvania,
1973) ; James MILROY, Linguistic Variation and Change. On the historical sociolinguistics of English, Oxford, 1992.
6. The concept of discourse traditions (Diskurstraditionen) has been
eleborated by Peter Koch and is also used by Johannes Kabatek. It means
that the creativity of the speakers is marked out by the action of traditions,
i.e. some discourses which have been created on the basis of pragmatics, and
which have subsequently become traditional, get their justification through their
repetition. See Peter KOCH, Diskurstraditionen : zu ihrem sprachtheoretischen
Status und ihrer Dynamik, in FRANK, B., HAYE, Th., TOPHINKE, D. (ed.), Gattungen mittelalterlicher Schriftlichkeit, Tbingen, 1997, ScriptOralia 99, p. 43-79 ;
Johannes KABATEK, Tradiciones discursivas jurdicas y elaboracin lingstica
en la Espaa medieval, Cahiers de linguistique hispanique mdivale 27, 2004,
p. 249-261.
7. On the interaction between linguistic features and symbolic values, see
for instance Nina CATACH, Code, langage et norme, loral et lcrit, La
sociolinguistique. Approches, thories, pratiques (Actes du colloque 1978), t.
II, Paris, 1980, p. 519-538 ; Andr WINTHER, Lcrit et la norme, La sociolinguistique. Approches, thories, pratiques (Actes du colloque 1978), t. II, Paris,
1980, p. 539-544.



conceives language and manages the relationship with its language in

function of its conception of it. It is the impact of written language that
explains why the unstable and changing nature of language is generally
perceived in a negative way while immutability, on the other hand, is
seen as an ideal. Chaos must be avoided, unity must be searched for,
and continuity must be guaranteed. Written language should also reflect
respectable usage. This particular tension between dynamics and durability leads to a tendency, especially in terms of language learning, to
consider spoken language uses as imperfect realizations of language,
which must be avoided in written language.8
But this should not incite us to brush aside written language, quite
on the contrary. If written language does not, as such, reflect spoken
language, its difficult relationship with its spoken counterpart is most
interesting. An important question concerns the degree of discrepancy
between spoken and written language. To what extent does the tension
between the dynamics of the present and the abstract fixity of an ideal
be it latent or bravely defended remain operational ? And how does
it cease to be so ?
With respect to the early Middle Ages, it was Michael Richter9 who
first drew attention to the fact that people not only spoke a language,
but that they also were confronted with its tradition, its history. Consequently, it is not enough to ask when they stopped to speak Latin, as the
question when they stopped to understand it is just as important. Since
then, the communicative approach has been chosen repeatedly in order
to search for the continuities as well as the discontinuities between spoken and written expressions.10 Hagiography, as literary entre-deux11
between the literate and the illiterate appeared to be particularly
interesting because it required both very important discourse traditions12
8. Cf. Tony BEX, Variety in Written English, Texts in Society, Societies in
Text, London/New York, Routledge, 1996.
9. Michael RICHTER, Kommunikationsprobleme im lateinischen Mittelalter,
Historische Zeitschrift 222/1, Mnchen, 1976, p. 43-80.
10. See especially : Roger WRIGHT, Late Latin and early Romance in Spain
and Carolingian France, Liverpool, Cairns, 1982 ; Marc VAN UYTFANGHE,
Lhagiographie et son public lpoque mrovingienne, in LIVINGSTONE, E.A.
(ed.), Studia patristica XVI. (Papers presented at the Seventh International Conference on Patristic Studies held in Oxford, 1975), part 2, Berlin, 1985, p. 54-62 ;
Michel BANNIARD, Viva voce, Communication crite et communication orale du
IVe au IXe sicle en Occident latin, Paris, Institut des tudes Augustiniennes, 1992 ;
Hagiographie et communication verticale au temps des Mrovingiens (VIIe-VIIIe
sicles), Turnhout, Brepols, 2007.
11. Marc Van UYTFANGHE, Lhagiographie antique tardive : une littrature
populaire ?, Antiquit tardive 9, 2001, p. 201-218.
12. Marc VAN UYTFANGHE, Le remploi dans lhagiographie : une loi du
genre qui touffe loriginalit, Ideologiche e pratiche del reimpiego nellalto
medioevo, Spoleto, vol. 1, 1999, p. 359-408.



and a language understandable to the illiterate when read aloud.13 As

has been shown in previous research,14 the interface between past and
present, oral and written, exists largely in the multiplicity and the flexibility of language features. Linguistic structures are not just quickly
replaced by others ; they rather fade away, leaving different traces in
the linguistic memory. A spelling that (more or less) conforms to the
ancient Latin spelling rules can correspond, when read aloud, to a pronunciation that no longer distinguishes between several grammatical
markings (manibus, for instance, becoming equivalent to manus, both
being pronounced [ma%nE s]).15 As a consequence, explicit syntactical
relations can turn into implicit ones, supported only by semantic relations and spatial proximity. Hagiographers manage to reconcile the past
and the present by appealing to the slowness of linguistic evolution :
they use structures that function sometimes in different ways both
in older and more recent layers of language development.
To approach Latin texts not as if they were abstract entities, but
to try to reconstruct their communicational dynamism is of course not
easy, the principal difficulty being that the reflection is very much a
theoretical and a logical one, with little concrete verifiable evidence. It
is always somewhat frustrating to be condemned to study what might
have been possible, without real certitudes. Nevertheless, the path is
fascinating and the complex language landscape it leads to seems to
be in accordance with what can be observed in other languages. It
seems indeed natural and plausible to conceive Latin as a complex
monolingualism until the 9th century, with varying language levels in
its successive chronological stages, proposing different mixtures of
archaic, steady and emerging language features, with different correlations between spoken and written language and different discourse
13. Marc VAN UYTFANGHE, Lhagiographie et son public (see note 10).
14. Marieke VAN ACKER, UT QUIQUE RUSTICI (see note 10).
15. Roger Wright insisted a lot on this aspect (Roger WRIGHT, Late Latin
and early Romance, Liverpool, Francis Cairns, 1982, p. 50sq).
16. I made a short synthesis here of concepts developed by Michel Banniard (latinophonie, niveaux de langue, traits archaques, actuels et mergents), Roger Wright (complex monolinguism) and Peter Koch (Diskurstraditionen). See : Michel BANNIARD, Viva voce ; ID., Le latin mrovingien, tat
de la question, in PARISSE, M. & GOULLET, M. (eds.), Les historiens et le latin
mdival. Colloque tenu la Sorbonne les 9, 10 et 11 septembre 1999, Paris,
Publications de la Sorbonne, 2001, p. 17-30 and elsewhere (latinophonie) ; ID.,
Niveaux de langue et communication latinophone (Ve-VIIIe s.), Settimane di
Studio 52, 2005, vol, 155-168 (niveaux de langue) ; ID., Prrequis de rceptibilit du latin tardif en priode de transition, in KISS, S., MONDIN, L. et SALVI,
G. (eds.), Latin et langues romanes, tudes de linguistique offertes Jzsef Herman, Tbingen, Niemeyer, 2005, p. 105-113 (traits archaques, actuels et mergents) ; Roger WRIGHT, Late Latin and early Romance ; ID., Complex monolingualism in Early Romance, in ASHBY, W. J. et al. (eds.), Linguistic Perspectives



The challenge of reconstructing the mechanisms of such a Latin

complex monolingualism covers many aspects. In fact, if one wants to
understand how a language functioned and evolved at a moment, one
has to take interest in the way speakers in the given period managed
their language, in its spoken as well as in its written form. A lot of
energy has been put into disclosing the continuities between oral and
written language, present and past. However, a lot still has to be done
concerning the transmission of the past, the management of the dead
(or dying) memory of the language : linguistic structures that are no
longer transmitted spontaneously but which require schooling. In this
respect, it is important to be conscious of the fact that a language is not
only a means of communication, but that it is also a support of values17
and that the linguistic inheritance often plays an important role here.

Opportunities offered by the Torino Manuscript

The Torino hagiographic collection is an interesting witness to the
way in which language inheritance was managed in the Early Middle
Ages. This collection of forty saints passions and lives, of which the
hypotexts are predominantly late antique and Merovingian,18 is indeed
likely to instruct us on how its producers managed written texts transmitted from the past. Furthermore, since the collection was copied in
Northern France presumably at the end of the 8th or the beginning of
the 9th century, as the paleographical study suggests,19 it enables us to
work on an original document belonging to a key period within the

on Romance Languages, Amsterdam/Philadelphia, Benjamins, 1992, p. 378-387

and elsewhere (complex monolinguism) ; Peter KOCH, Diskurstraditionen : zu
ihrem sprachtheoretischen Status... (see note 6) (Diskurstraditionen).
The traditional diglossic model has been criticized in Marieke VAN ACKER,
La transition latin / langues romanes et la notion de diglossie, Zeitschrift fr
romanische Philologie 126 :1, 2010, p. 1-34.
17. Marieke VAN ACKER, La transition latin / langues romanes et la notion
de diglossie, 4.
18. Not all texts are dated, but some are : the Passio Luceiae was composed
probably about 500 ; the passions of Anastasia and of Mauricius in the 5th century or earlier ; the passion of Quintinus in the 6th century or earlier ; the Hymnus Eulaliae in the 7th century ; the passion of Nazarius in the 7th-8th centuries ;
the life of Remedius about 750.
19. On the basis of the script style (a-b type) and the method of binding
the volume, David Ganz believes that it can be situated in Soissons, between
790 and 825 (Cf. Monique GOULLET, Expertise des textes hagiographiques
mrovingiens, p. 85). The a-b script is usually associated with the monastery
of Corbie, in northern France, where it was used from the end of the 8th until
the middle of the 9th century (cf. Bernhard BISCHOFF [Hartmut ATSMA & Jean
VEZIN (transl.)], Palographie de lantiquit romaine et du moyen ge occidental, Paris, Picard, 1993 (German edition : Berlin, Schmidt, 1979).



history of Western Latinity. Various elements should indeed be taken

into account here :
1. The tension between written and spoken language reaches a new
milestone after 750 : spoken language displays henceforth a majority of
new language features which remain a minority in written language.
Consequently, the reading aloud of texts before an illiterate audience
vertical communication became more and more problematic.20
2. The tension is brought to a climax by the language policy of the
Carolingian dynasty, which redefines the norms of written language in
a way that makes it even more closed to language innovations coming
from the oral. This policy of language reform is being set up after 750
and fits into a larger set of measures aiming at a general resettlement
of order.21 In parallel, monasteries are being transformed into intellectual centres much more than places of mortification, by pitchforking
insular scholars.22
It is interesting to observe where the Torino collection must be situated with respect to this evolution in language management : does it
reflect the stage that precedes the reforms and their effects displaying
a language in need of purification, or does it rather reflect a stage that
has already undergone these reforms ? And can the linguistic features
help us to characterize both stages ?

Description of the Collection

Being a hagiographical collection, the forty texts of the volume23
have of course a lot of elements in common, but there are also many
factors of heterogeneity. Below, I have listed some specific aspects of
convergence and divergence that might be significant for the understanding of the sociolinguistic reach of the collection.

20. Cf. Michel BANNIARD, Viva voce..., p. 485sq.

21. Philippe DEPREUX, Ambitions et limites des rformes culturelles
lpoque carolingienne, Revue historique 623, 2002, p. 721-753.
22. Cf. Albrecht DIEM, The Emergence of Monastic Schools. The role of
Alcuin, in HOUWEN L.A.J.R., MACDONALD A.A. (eds.), Alcuin of York, Scholar
at the Carolingian Court. Proceedings of the third Germania Latina Conference
held at the University of Groningen (May 1995), Germania Latina III, Groningen, Egbert Forsten, 1998, p. 27-44.
23. The saints that are presented are : 1. Quintinus 2. Christophorus 3.
Longinus 4. Speusippus 5. Symphorianus 6. Mauricius 7. Crispinus & Crispinianus 8. Remedius 9. (H)Adrianus 10. Agatha 11. Lucia 12. Luceia 13. Agnes
14. Eupraxia 15. Eulalia 16. Hymnus Eulalia 17. Christina 18. Cyr 19. Laurentius 20. Theodosia 21. Babylas 22. Afra (conversio) 23. Afra 24. Apollonios
25. Agape 26. Cyrillus puer 27. Genesius 28. Eleutherius 29. Juliana 30. Eugenia 31. Radegundis 32. Marina 33. Anastasia 34. Matrona 35. Euphemia 36.
Saturninus 37. Hilarius 38. Albanus 39. Victor 40. Nazarius.



Converging elements


All texts are prose texts except the Hymnus Eulaliae

Most texts are epic passions with profuse repetitions of
confrontations, tribulations and tortures. There are also
some vitae (esp. texts associated with Venantius Fortunatus). Generally speaking, the contents are of the kind to
hold in suspense an audience, in spite of their repetitive
Most of the saints belong to the 3th and 4th centuries, the
period in which far-reaching persecutions of Christians
took place.
Most of the saints do not originate from Gaul, but from
more remote lands : Asia Minor, Rome, Spain, and Southern Italy.
The background of the stories is often Roman, with some
exotic touches.
Several stories must be connected with Greek texts or traditions.

Diverging elements


Attribution : most texts are anonymous productions. Three

of them (Vita Radegondis, Gesta Hilarii, Vita Remedii
[Remigii]) are generally associated with Venantius Fortunatus (as in Mignes Patrologia Latina), although his authorship has been questioned for the Gesta Hilarii and the Vita
Several scribes worked on the copy. They seem to have
worked from models of different origins.
The length of the pieces is very variable, going from 1
folio up to 20 ; from close to 8000 words to 560.

Editing project and linguistic analysis : preliminary observations

The first purpose of the project Expertise de textes hagiographiques
mrovingiens dans leurs plus anciennes versions manuscrites was to
make accessible one of the oldest still existing hagiographic collections.
Both a printed and a digital edition have been prepared. The printed
version will propose the text ante correctionem. This version should
bring us as close as possible to the initial state of the document, that
is, as it was first copied. It will be provided with an apparatus of later
24. In the MGH collection, these texts figure in the section opuscula
Venantio Fortunato male attributa (Monumenta Germaniae Historica, Auctores
Antiquissimi, 4.2, p. 55sqq).



revisions and corrections. The digital version will permit to pass from
the version ante correctionem to the one post correctionem. All
forty texts will be accompanied by hagiographic introductions, situating
each text within the context of its specific saints cult and text tradition.
Furthermore, a general hagiographic introduction (by Guy Philippart)
will shed light on the position of the Torino collection within the development of hagiographical collections.
As indicated earlier, the central concern of our linguistic approach is
one that tries to better understand the modes of transmission of written
texts and the way in which this transmission reflects language conception and language management. This is why we had a double linguistic approach to the texts. The first approach aims at a characterization
of the hypotexts. The questions here join those with respect to written
text production : in both cases, the central concern is about mapping
the relations between tradition and renewal, and about measuring the
degree of compromise, the degree of plasticity, the frequency, the distribution, and the degree of fluctuations within the physiognomy of the
language. We observed the stylistic and narrative features of the texts
in the collection as well as their lexical richness. A sample of texts has
been subjected to more advanced analyses of a syntactical nature.25 The
aim here was to compare the texts as to their linguistic complexity, and
to observe phenomena likely to inform us about the language of the
original writings (hypotexts), in order to possibly help date them.
In particular, we observed the proportion of text in direct speech, which
is likely to reveal its degree of liveliness. We studied the cutting of the
text and compared the number and the type of subordinate clauses in
order to determine tendencies towards simplicity or rather complexity.
Infinitive clauses, for instance, are usually associated with a complex
style. We observed the frequency of finite verbs and of prepositions,
both indicators if present in large numbers of a plain style, with
short syntactical units and explicit markers of structure. We finally
verified the degree of syntactical contiguity : the extent to which verbs
are separated from their subjects and from their complements. This is
important because the tendency towards an alignment of syntactical
units can be linked to the loss of pertinence of casual endings.
The purpose of the second approach was also to observe the quality of text transmission. To this effect, all the forty texts have been
subjected to a linguistic questionnaire which was created in order
to map a number of phonetic and morpho-syntactic features that are
likely to have undergone pressure from the spoken language. Special
attention was given to phonetic phenomena, such as the confusion of o
and u, e and i, b and v, the instability of the final consonants etc. We
25. These analyses were done on 6 texts integrally (1 Quintinus, 7 Crispinus
& Crispinianus, 12 Luceia, 21 Babylas, 27 Genesius, 34 Matrona) and two texts
partially (8 Remedius, 37 Hilarius).



further focused on morphological and morpho-syntactic phenomena,

such as the shifts in the declension system, the verbal and pronominal
forms. These questionnaires produced exhaustive detailed and comparable data.

Linguistic observations
Stylistically, some important variations appear. Some texts are completely descriptive. Others, on the contrary, are very dramatic, with a
significant number of passages in direct speech. The table below shows,
for a selection of texts, the variations in the percentage of words in
direct speech :
percentage of
words in direct speech
1. Quintinus


7. Crispinus & Crispinianus


8. Remedius

12. Luceia


21. Babylas


27. Genesius

34. Matrona


37. Hilarius

The level of style can also be very different from one text to another.
The following passage is written in plain style :
Riciouarus dixit : Quintini, recede ab hanc stulticia qua teneris et
sacrifica diis. Quintinus respondit : Stultus non sum, sed sapienciam
quero, nam illi stulti sunt, qui tibi obaudiunt. (Passio Quintini, fol. 2r,
l. 2-5)

Besides, the Passio Quintini, as a whole, is characterized by a

very low literary quality. Stylistic elegance is completely supplanted
by spontaneity : short and linear clauses with many structuring words.
The Gesta Hilarii, on the contrary, attributed to Venantius Fortunatus,
proposes clauses that are far more ambitious, testing regularly to his
disadvantage the scribes capacities :
nam poene totum mundum graui errore confusum, factis in galliis
sepe synodis, per hilarium fuisse ad uiam ueritatis adductum confitetur
lingua multorum, quid autem interest, uel sibi pro uita aeterna factum
fuisse martyrem uel amplius uixisset religioni. (Gesta Hilarii, fol. 251v,
l. 19-252r, l. 2)



In the Passio Crispini et Crispiniani, the author clearly searches for

stylistic effects by multiplying the hyperbata :
Riciouarum suae direxit impietatis ministrum (Passio Crispini et
Crispiniani, 44r, l. 5sq)
exquisiti calciamenta sunt in predicta urbe pauperum sarcientes
reperti (Ibidem, 44r, l. 6sq)
quibus erat fidei inmobile supra petram solidissimam fundamentum
(Ibidem, 44r, l. 13sq)

Grammatically, the conservative elements are constantly far more

frequent than the innovative ones. The case endings, in particular, are
not excessively deviant from the grammatical norm, neither are the
verbal endings. It must be observed, however, that hagiographic texts
written between 650 and 750 are not necessarily very different in
this respect.26
The lexicon exhibits many late Latin and Christian features, which
comes as no surprise, but there are very few vulgarisms. On the other
hand, the occurrence of typically classical terms (some of them quite
rare) is striking :
tugurium (= tegurium) : hut (Passio Quintini, fol. 1,7r, l. 8)
bracho (= bracchio) : assistance, protection (Passio Christinae,
fol. 121r, l. 21)
etiam nam et torqueo illos (classical use of nam, to reinforce an
exclamation) (Passio Adriani fol. 55r, l. 20)
the deponent construction (me) contumiliatus est : he affronted
me (Martyrium Babyle, fol. 154v, l. 5)
promum : meat-safe, office (Passio Matronae 34, 232v, l. 9)
Syntactically, the
respect to :

advanced analyses show important variations with

the use of verbs : conjugated verbs vs. participia

the use of prepositions
the place of verb complements (anteposition vs.
the distance between verbs and subjects ; verbs
and complements
These variations confirm the association of the texts with different
stylistic levels. But do they therefore also belong to different chronological stages ? It is in fact very difficult to express clear opinions on
the linguistic basis here. Passages written in a plain style can indeed be
perfectly trans-diachronic : they could be written by an author in the 5th
century as well as in the 9th century. Syntactic simplicity is not a guarantee of late composition. It is rather syntactic complexity that is likely
26. Cf. Marieke VAN ACKER, 2007,





to imply an earlier period of composition. In this respect, the majority

of the texts in the Torino manuscript seem to belong to a period anterior to the 7th and 8th centuries. When compared with hagiographies
written in this period, they exhibit more complex constructions (infinitive clauses for instance) and, above all, a higher degree of syntactic variation (types of subordinate clauses, for instance). At the same
time, they do not show the same rusticitas. Hypercharacterization, for
instance, is rare. And so is the excessive use of pronouns, or the rigidity of the sentence structure.
Transmitted copies
The copies in the collection are frequently defective. These holes
in the texts are totally arbitrary and seem to be due to lacunas in the
manuscripts that were used as models by the scribes, or simply to inattention (e.g. skipping a whole folio). For instance, the beginning of the
Gesta Hilarii :
quasi lucifer inter cetera castra (= astra) refulgens processit. Cuius a
conabulis (= cunabulis) tantae sapientiae primitiua lactabu*ur (= lactabatur) infantia ut iam tunc palam possit intellegi christum in sua causa pro
optinenda (= obtinenda) uictoria militem propagare.

There are frequent problems of comprehension, illustrated by passages of nonsense, especially in the syntactically more difficult texts,
such as those attributed to Venantius Fortunatus for instance, but also
in other texts or passages. In the Hymnus Eulalia, the copyist apparently did not realize that the text was in verse form. The following are
examples from the Passio Christinae :
quia dignatus es me accipere incorruptionis panem should be qui a
me dignatus es accipere incorruptionis panem (fol. 114r, l. 18-19)
beata cristina iussit adsessorem dionis mitti in carcere should be beatam cristinam iussit adsessor dionis mitti in carcerem (fol. 120r, l. 5-6)

Punctuation27 can be defective, illustrating again important lapses in

comprehension :
Et mane facto iussit iudex. Dion adduci eam in praetorium (Passio
Christinae, fol. 118v, l. 20)28 should be Et mane facto, iussit iudex Dion
adduci eam in praetorium

Mostly, however, punctuation is coherent, distinguishing in particular

very clearly the passages in direct speech, as in the Passio Marinae :
27. I only consider the primary punctuation, i.e. signs (mostly simple dots)
that coincide with the copying of the text.
28. In this fragment, unlike in those preceding, only primary punctuation is



Et dixit ab (= ad) eam que genus es. Marina dixit ingenua sum. Praefectus dixit. quis foederis uel quod tibi nomen est. quae dixit. marina
cognominor. Praefectus dixit. quem deum colis uel inuocas uel quem
adoras. Marina dixit. ego inuoco deum omnipotentem et eius uerbum
(221v, l. 7sq).29

Furthermore, the linguistic surveys show that the impact of the

oral language on the copied texts is very limited. There is also a certain
homogeneity within the recurrence and the proportion of certain phenomena, as the following data will show.
There are many cases of confusion between e/i and o/u, the first
one being more prominent. E.g. derixit instead of direxit (250r, l. 9) ;
uincolum instead of uinculo (250r, l. 6). Palatalization is a frequently
recurring phenomenon : t + vowel becomes c + vowel, e.g. senciebat
instead of sentiebat (251v, l. 3). Inital h is frequently unstable (omitted
or added). Final consonants are frequently unstable, e.g. futura speciem
indicans instead of futuram speciem indicans (249r, l. 1). There are rather
few hypercorrections, e.g. praecibus instead of precibus (251r, l. 18).
Morphologic variation takes place at a relatively modest scale and
coincides most of the time with phonetic phenomena, e.g. : in hac provinciam (5r, l. 2) ; in eodem locum (5r, l. 12). There are few changes in
gender, declension or conjugation.
Morpho-syntactic variants are most of the time case confusions
(which again coincide most of the time with phonetic phenomena),
confusion of passive and active infinitives (idem), double compound
pasts, and some periphrastic futures. All of these phenomena are rather
Concerning the distribution of these phenomena over the manuscript
as a whole, the general impression is one of uniformity. Nevertheless,
variations in the distribution appear when the number of occurrences is
observed in proportion with the length of the texts. This is particularly
interesting when observed in a sequence copied by one single hand. In
the table below, the numbers proposed are the result of the following
mathematical operation :
Number of occurrences of the observed phenomenon x 100.
Number of words of the analyzed text
29. Second fragment showing only primary punctuation.



However abstract these numbers may seem, they show the fluctuations in the occurrence of the observed phenomena, also when observing texts written by one and the same copyist (in grey) :





1. Quintinus





2. Christophorus





7. Crispinus





9. Adrianus





31. Radegundis





32. Marina





33. Anastasia





34. Matrona





35. Euphemia





This fact seems to suggest that the scribes copied looking at their
models rather than hearing them, maintaining unchanged errors due
to the oral language of anterior periods, when the language of the texts
was still more accessible to the scribes. This important degree of fidelity
to the models contrasts with the obvious lack of comprehension by the
copyists. But it is not unexpected that the respect for the models grows
proportionally to the decreasing capacity for intervention in the texts.

Temporary assessment
The Torino manuscript contains a range of stylistically very diverse
texts whose composition was for the majority of them not recent
and whose comprehension had obviously become problematic. But
despite these problems of comprehension, the respect shown to the
models suggests that it was important for the scribes to transmit texts
belonging to a cultural heritage. This can appear to be a paradox to the
modern literate audience : what can be the sense of transmitting texts
without fully understanding them ?
First, an observation must be made on the practices of copying in
the Early Middle Ages. It is known that in Late Antiquity, copyists
wrote either from dictation or from looking at written models. The second practice seems nevertheless to have become predominant, not in
the least because the literacy of scribes was limited. Studying the Formulae Marculfi, Alf Uddholm noticed that the scribes probably did not
understand very much of what they were writing.30 And to Petrucci,31
30. Alf UDDHOLM, Formulae Marculfi. tudes sur la langue et le style,
Uppsala, Almqvist & Wiksell, 1954, 170.
31. Armando PETRUCCI, Lire au Moyen ge, Mlanges de lcole franaise
de Rome (Moyen ge, Temps modernes) 96-2, 1984, p. 603-616, esp. p. 605-7.



illiterate scribes were perfectly common in the Early Middle Ages. He

gives the example of Eckhard of St-Gallen, who, in the 10th century,
destined the less intelligent of his fellows, those who were less fit for
intellectual occupations, to the copying activity. In fact, and this was
confirmed by palaeography specialist Mirella Ferrari, writing became
very much an aesthetic matter, and had rather little to do with reading
or literacy.
These observations raise the question of function and usefulness of
the texts. The corrections made by subsequent revisers as well as the
secondary punctuation suggest that the texts were read, but some more
than others, and some fragments more than others. Although some passages are absolutely unfit for any use because they are incomprehensible, a large portion of texts is perfectly usable. There are quite a few
trans-diachronic passages, this is, passages with a majority of stable
language features, a few features that are evanescent or passed out, and
also a few features that are innovative. There are also a lot of passages marked by superposition of linguistic perceptions, especially in
text portions that are stylistically not very elaborate. In such passages,
evanescent linguistic features are present in such a balanced way, that
they are not necessarily problematic, either because the context helps to
interpret them (archaic conjunctions, for instance, need not necessarily
be fully known to the hearers), or because they can easily be perceived
in a modern way. Specific case or verbal endings, for instance, can
be omitted in linguistic perception without the utterance becoming necessarily incomprehensible : the functions simply become implicit.
But does this theoretical possibility correspond to the reality ? What
happened when the beginning of a text was truncated (e.g. the very first
( !) text in the collection, Passio Quintini) or when the story suddenly
becomes incomprehensible ? How to reconcile a concrete use of the text
and punctuation errors ? The latter question becomes all the more important since it is known that reading at the time was a very complicated
exercise,32 where punctuation was of capital importance.
But if the collection was indeed not fit for practical use, what could
possibly have been its purpose ? What uses other than communication
can be envisioned ? A first hypothesis is the following : an act of conservation through transmission. It is known how important memory,
and especially continuity, was in the early Middle Ages. Whether it
concerns buildings,33 political institutions, titles, legislation or administration, the desire to stay connected with the past, and especially with
32. Michel BANNIARD, Le lecteur en Espagne wisigothique daprs Isidore
de Sville : de ses fonctions ltat de la langue, Revue dtudes augustiniennes
21, 1975, p. 112-144 ; ID., Credo et langage : les missions de saint Boniface, in
DIERKENS, A. & SANSTERRE, J.-M. (eds.), Voyages et voyageurs Byzance et en
Occident du VIe au XIe sicle, Genve, 2000.
33. Xavier BARRAL I ALTET, Le dcor des monuments religieux de Neustrie, in ATSMA, H. (d.), La Neustrie. Les pays du nord de la Loire de 650



the Roman Empire, can be found always and everywhere.34 So why not
copy texts as if to refresh their connection with the present, at least to
cultivate their memory ? Besides, such an act of conservation through
transmission could have been perfectly combined with study. A collection like Torino D.V.3. could very well have functioned as an object
of study for scholars developing their literacy, learning to decipher, to
understand, to punctuate, to read aloud.
A second hypothesis, though a less plausible one to me at least
with respect to the period concerned is that the collection, in spite of
its many defects, was still used for public monastic readings. Practices
of this kind seem to have existed later on, during the 11th and 12th centuries : the readings performed during the office of Matins were meant
as ritualistic performances rather than as source of information, as the
cutting of the fragments suggests.35
The Torino collection illustrates very well how multiform Merovingian Latinity was and how it was full of compromises and contradictions. It shows particularly well how the importance accorded to a written legacy, and the respect that was shown to it, produced a distended
linguistic spectrum that eventually came to split into two separate ones.
A document like the Torino collection made sense only to highly literate people, who must have been able to measure both the distance and
the similarities between this Latinity and their own, between the living
and the dying memory of their language.
This might also shed new light on the shifting of language levels :
low level Latin, running into high level proto-Romance, as in the Oaths
of Strasbourg. When late Merovingian authors of hagiographies excuse
themselves for using a rustic language, they situate themselves in the
long history of Latin, the one they know for having studied it in books.
In all likelihood, it seems probable that they were conscious of the fact
that their rusticitas was not rustic at all strictly speaking within
their own chronological section of Latinity.

850. Colloque historique international (Rouen, 7-10 octobre 1985), Sigmaringen, Thorbecke, 2 vol., 1989, p. 209-224.
34. Bertrand LANON, Le monde romain tardif, IIIe-VIIe s., Paris, A. Colin,
35. Tjamke SNIJDERS, Celebrating with dignity : The Purpose of Benedictine
Matins Readings, in VANDERPUTTEN, S. (ed.), Understanding Monastic Practices
of Oral Communication (Western Europe, Tenth-Thirteenth centuries), Turnhout,
Brepols, 2011.

Universit de Strasbourg


The notion of mouvance is valid and thoroughly operational for
analyzing the textual tradition of the Acts of Pilate, as it is a fluid and
profoundly unstable tradition. By applying the tools and meeting the
requirements both of classical philology and of the New Philology,
the editors of the Acts of Pilate seek to organize and interpret this
complex manuscript tradition, while underscoring the proliferation of
variants and the transformative dynamic of the text. However, the current editorial project does not satisfactorily account for the insertion of
the Acts of Pilate in a collection of texts that can be identified, by and
large, as a coherent corpus. The development of certain electronic tools
could offer a solution to this difficult problem.
La notion de mouvance est pleinement valable et oprationnelle
pour analyser la tradition textuelle des Actes de Pilate, qui se caractrise par une grande fluidit et une profonde instabilit. Mlant les
exigences et les outils de la philologie classique et de la New Philology , les diteurs des Actes de Pilate cherchent ordonnancer et
interprter cette tradition manuscrite complexe, tout en mettant en
valeur le foisonnement des variantes et la dynamique transformative
du texte. Le projet ditorial actuel ne permet toutefois pas de rendre
compte de faon satisfaisante de linsertion des Actes de Pilate dans
un ensemble de textes avec lesquels il a constitu peu peu un corpus
peru comme cohrent ; le dveloppement doutils lectroniques pourrait offrir une solution ce difficile problme.
The notion of mouvance, used extensively in Medieval Studies,
has yet to penetrate the field of Byzantine Studies. It offers, however, a
thoroughly valid and operational strategy for analyzing the textual tradi-

* I would like to thank Albert Frey (Universit de Lausane) for his attentive
proofreading and Nathanael Szobody (Strasbourg) for his English translation.


Apocrypha 23, 2012, p. 81-97



tion of the Acts of Pilate.1 Composed in Greek during the 4th century,2
this narrative was transmitted by a number of witnesses that all testify
to a text in a state of constant evolution.3 So much so, in fact, that
while preparing a critical edition for the Series Apocryphorum of the
Corpus Christianorum, the editors have run into problems similar to
those encountered by editors of multi-form medieval texts.4
After briefly discussing the problems which arise from the fluidity
of the Greek manuscript tradition of the Acts of Pilate, I will present
the editorial choices that have followed a period of trial and error concerning various issues, some of which still remain unresolved. In doing

1. Concerning this apocryphon, cf. M. GEERARD, Clavis Apocryphorum

Novi Testamenti, Turnhout, Corpus Christianorum, 1992, p. 43-46 (s. n. 62) ;
R. GOUNELLE, Pilate, Acts of, in The Oxford Bible Dictionary, Oxford, 2011,
in press. Unless otherwise indicated the title Acts of Pilate will refer to the
Greek A recension, not to be confused with the Byzantine Greek recension
edited by R. GOUNELLE, Les recensions byzantines de lvangile de Nicodme,
Turnhout/Lausanne, Corpus Christianorum, Series Apocryphorum, Instrumenta
3 ; Instruments pour ltude des langues de lOrient ancient 7, 2008. The references to the Acts of Pilate follow the new numbering system established by the
research group of the AELAC in its preparation of a new critical edition of the
text. The traditional division of chapters and paragraphs is indicated in parentheses and in Roman numerals whenever they are different.
2. The dating of this texts composition is disputed. For a classical view
point, cf. M. SCHRTL, Nicht das ganze Volk will, dass er sterbe. Die Pilatusakten als historische Quelle der Sptantike, Frankfurt a. M., Apeliotes 8, 2011,
p. 12-28. For a more critical view point of the ancient witnesses to the Acts
of Pilate, cf. R. GOUNELLE, Un nouvel vangile judo-chrtien ? Les Actes de
Pilate, in J. SCHRTER (d.), The Apocryphal Gospels within the Context of
Early Christian Theology, Leuven, in press.
3. These witnesses are presented by C. FURRER, La recension grecque ancienne des Actes de Pilate, Apocrypha 21, 2010, p. 11-30, with a list of their
abbreviations. Since the publication of this article, more witnesses have been
identified : Meteora, Monastery of the Holy Trinity, 14 (olim, Rousanou 12),
fol. 1-10v (the Acts of Pilate are at the beginning of the manuscript, in a section
dated between the XVth and XVIth century by D. Z. SOPHANIOS,
, Athens, 4, 2003,
p. 201) ; Meteora, Monastery of the Holy Trinity, 549, fol. 343-346, XVIIth c.
(N. A. BEES, , Athens,
1, 1992, p. 555). Additionally there is a probable witness of
the Narratio Iosephi Rescripta in a manuscript in the ms. Athous Karakallou
87 (1600) (XVIIIth c.) ; I owe this last indication to Antonio Rigo (Universit
CaFoscari, Venice).
4. In order to make this a valid comparison, I will set aside the multiple
ancient translations of the Acts of Pilate which furnish precious clues as to the
evolution of the Greek text in addition to the preserved Greek manuscripts ; on
this issue, cf. proceedings from the Workshop Editing the Acts of Pilate in Early
Christian Languages. Theory and Practice (Winnipeg, October 3-5, 2010), compiled in Apocrypha 21, 2010, p. 9-176.



so, I will emphasize some methodological concerns that arise from the
whole project.

A Fluid Tradition
After it was composed somewhere between 320 and 380,5 the Acts
of Pilate underwent a significant series of revisions. As the extant title
seems not to be coherent with the narrative, the original text may have
first circulated without a title or with a title now lost, and with a prologue. It received the title Memoirs of Our Lord Jesus the Christ,
composed under Pontius Pilate sometime before 387. A postface with
a colophon was added to the end of the text sometime after 425. It
was then moved and inserted between the title and the prologue, thus
becoming a preface.6 Did the copyists who undertook these modifications revise other parts of the text ? It is not impossible. What is
certain, however, is that the Greek manuscript tradition of the Acts
of Pilate, which is on the whole a later tradition,7 exhibits numerous
variations in its details. Two types of textual practices may account
for this : (a) abbreviations and abridgements (intentional or accidental),
and (b) amplifications or revisions, frequently combined with vertical
contamination (the effect of which is that late manuscripts can contain
very ancient variants).
Traces of Editorial Intervention
One can find numerous traces of accidental modificationsresulting from transliteration,8 confusion between sounds and handwritings,9

5. Cf. footnote 2.
6. For more details on all these issues, cf. R. GOUNELLE, Un nouvel vangile judo-chrtien, op. cit. The testimony of Epiphanius of Salamis, Panarion 50.1.5-8, on the use of some Acts of Pilate by Quartodecimans is generally used to prove that the text was variable as early as 380cf. notably
J.-D. DUBOIS, Lutilisation des Actes de Pilate au quatrime sicle, Apocrypha
2, 1991, p. 85-98but it is uncertain whether the heresiologue is referring to the
surviving Acts of Pilate, as neither the title nor the variants quoted by Epiphanius are attested in the manuscripts of the surviving textcf. R. GOUNELLE, Un
nouvel vangile judo-chrtien, op. cit.
7. The oldest preserved manuscript dates to the XIIth century. Cf. C. FURRER,
op. cit.
8. Thus the confusion between and (Prologue, 14.1, 14.2), and
(16.1.2 [XVI.2]). The vast majority of such misreadings occur among all
the Greek witnesses.
9. Hesitation between aorist verb forms of , to call and , to
order ( as opposed to , as opposed to )
likely belong in this category (cf. 1.2, 2.6).



corruption,10 and homoioteleuta (skipping lines to an identical word)11

as well as other modifications that are likely due to the evolution of
the Greek language.12 In addition to these, there are other corrections
in the Greek manuscripts of the Acts of Pilate, which are more likely
the result of intentional revision on the part of copyists attempting to
modify the content or form of the text.
Such variants are often found both in the syntax and the vocabulary.
As for the syntax, the sentence structure of the Acts of Pilate has often
been simplified,13 as unwieldy expressions have been reworked here
and there to render the text more fluid.14 The changes in vocabulary
are more difficult to explain. For example, why is the piece of cloth
which Pilates messenger spreads out under Jesus feet (1.2) referred to
as both (or ) and , depending on the
codices, without any coherence among the manuscripts on this point ?15
The alternation in some manuscripts between (garden) and
(house) to designate the place where the leaders of the Jews sit
(15.5) is equally surprising.16

10. Thus, in 16.1.3 [XVI.3], several manuscripts read with variant details :
; (The
Sanhedrin said to him : So Levi, is the word that you spoke true ?) while others have : ;
(The Sanhedrin said to Rabbi Levi : Is the word that you spoke true ?) There
is every reason to believe that was originally the complement of
(as is clearly the case in Armenian and certain Latin manuscripts) and that
the interrogative is the result of a corruption of .
11. The numerous homoioteleuta found in all the witnesses are due to the
highly repetitive character of the narrative, where nearly identical expressions
are often placed a few lines apart.
12. Such as the difficulty with certain verb forms (cf. the diverse forms
of to bend, in 1.5 : , , , ),
replacement of datives with genitives, hesitation on augments, sporadic presence of nominative absolutes, the presence of late noun and verb forms etc. In
some manuscripts there are, here and there, some archaic forms, like optatives,
testifying to the tendency of certain copyists to restore a purer form of Greek.
If these linguistic phenomena are relatively easy to identify, they are more difficult to date with any degree of precision, and thus offer little information on
the transmission history of the Acts of Pilate.
13. Thus in 1.2, the modification from (let Jesus be
brought) to (Bring Jesus) in some witnesses.
14. Thus in 15.6, in the majority of the Greek witnesses, water comes down
from ( or ) the head to the feet. Some manuscripts employ the less
obvious expression, which seems more primitive : on ( or ) the head
to the feet. These two forms of the text have parallels in the ancient versions.
15. or is certainly the older of the two.
16. The choice of the majority, reinforced by the Coptic and by a Latin
version, seems to be the lectio facilior, as opposed to , which is
employed by the witnesses of the family f (see infra), and which is confirmed
in the two Armenian versions, in Georgian and in a Latin version.



Layers of Revision
Due to its fluidity and instability, the manuscript tradition allows us
neither to situate such modifications in the text history with any great
degree of precision, nor to arrange them chronologically. Several layers
of revision have nevertheless been identified. These layers, or at least
the oldest ones, likely constitute successive groups of modifications,
having been implemented by more than one copyist. This is in any case
what a comparison between the codices extrauagantes17 and the ancient
translations of the text would suggest. Such a series of modifications
represents periods of significant evolution of the text, rather than a contribution of a single writer implementing a coherent agenda.
The identification of these layers is the reason for distinguishing two
textual families, designated by the abbreviations f and c and, within
the family c, subgroups c and c. This distinction between the textual
forms can be represented by the following stemma which is intended
to serve only as an illustration. The layers of editorial intervention are
designated by the Greek lower case :

It would be overly simplistic to summarize the history of the Acts

of Pilate with only these layers of revision, as is evident from various
modifications that cannot be attributed to the writers , , 1 or 2.
For example, to designate the place where Jesus was buried (cf.
12.1, 14.3 and 15.6), some Greek witnesses use (tomb) and
others use (cave). No manuscript is consistent on this point.
But in each occurrence some of the witnesses use , which is
at times corroborated by the Coptic versionan indication that would
support its ancient origin. It is possible that the primitive text of the
Acts of Pilate used everywhere, but that it was later corrected to be consistent with the New Testament. In any case it is difficult to imagine why Byzantine copyists would replace with
. Yet it is not possible to determine if these corrections were
implemented by a single copyist at a particular moment in the history
of the works transmission or if they were made by several copyists at
different stages. Thus the incoherent character of the preserved manuscripts can be attributed either to the non-systematic manner in which
corrections were carried out at some point in the textual tradition, or to
17. On these codices, cf. infra.



the contamination of both ancient and recent forms of the text, or to a

combination of both phenomena.
Another passage (1.2) preserves three equivalent and concurrent
terms. While in the family f, as well as in several indirect Greek witnesses and in most of the other versions, the messenger simply says to
Jesus enter (), some other manuscripts add either in the praetorium ( or ) or on the tribunal
( ). These variants have also resurfaced in Syriac and in
some Latin witnesses. It is possible that they reflect independent efforts
to provide a complement for the imperative enter, but the groups of
manuscripts that have these variants contain parallels in other passages as well. This could indicate that the additions in question were
implemented by a single writer who intervened in several passages. But
given that these modifications are only details, it is difficult to come to
a clear conclusion on this point.
Some classical philologists would consider such variants to be
unimportant details. And indeed, they do not fundamentally modify the
meaning of the narrative. Proponents of the New Philology, however,
would place particular importance on such details even while their origin remains obscure18 and their raison dtre is not easy to determine.

Editorial Choices
Handling such a versatile and elusive tradition within the constraints
of a critical edition is anything but easy. The reconstitution of a primitive version of the Acts of Pilateattempted by several editors of
the 19th century19is out of the question. The enterprise would not only
be risky, but its outcome would inevitably be a deformed summary of
what is actually a rich manuscript tradition ; that richness illustrates
the vitality of the Acts of Pilate and of its interpretation throughout
the centuries. A diplomatic edition of one or several manuscripts, as
certain adherents to the New Philology would likely propose, would
be no more desirable : such an edition would necessarily give undue
preference to one or several witnesses at the expense of others, based
on rather arbitrary criteria, given the current state of the manuscript
And yet, if the manuscript tradition presents such a fundamentally
unstable text, it would be impractical to give the reader direct access
to such an anarchic proliferation of readings. Indeed, editors have a
18. As mentioned above, some of the modifications likely took place at a
very early date, but that does not prove that they are in any way connected to
the modifications made to the title and to the addition of the postface/preface.
19. Cf. particularly the still commonly used edition by C. VON TISCHENDORF,
Evangelia apocrypha, adhibitis plurimis codicibus Graecis et Latinis maximam
partem nunc primum consultis atque ineditorum copia insignibus, 2nd ed. rev.
Leipzig, 1876 (18531 ; reprinted : Hildesheim, 1966), p. 210-286.



responsibility to organize and interpret this complex manuscript tradition in order to make it more accessible to other scholars. On this
point, the editors of the Acts of Pilate differ from certain currents of
New Philology in that they return to a fundamental aspect of classical philology. They nevertheless distance themselves from the latter by
avoiding flights of philological fancy and strictly adhering the manuscript tradition itself.
The editorial choices proceed from the partition of the manuscripts
into three different categories : first, the witnesses that transmit a similar text are grouped into families (in this case f and c) ; second, the
manuscripts which cannot be grouped into families constitute the codices extrauagantes ; and third, the manuscripts that contain quotes from
the Acts of Pilate,20 constitute indirect witnesses.
The Witnesses Belonging to Families
The witnesses belonging to the two main families could form the
basis of a classical critical edition. Using the Lachmannian method, one
could reconstitute, on the basis of these witnesses, the sort of texts that
likely circulated in Byzantine Christianity. The manuscripts from the
family f are often incomplete ; therefore it would not be possible to
reconstitute an archetype without being eclectic. As for the family c,
any attempt to reconstitute an archetype would be rather hazardous ;21
therefore the editors have focused instead on a sub-archetype of the
family c (c). The reconstituted texts of the ancestor of f and of c
are to be published synoptically with the standard critical apparatuses,
allowing both an enumeration of the un-edited variants as well as a
justification for the reconstituted texts. The planned edition of these
texts is moderately eclectic ; it refers to the various witnesses in order
to access the most primitive form of the text possible. The text is also
standardized, but only when necessary.22 When non-classical forms are
found in most of the manuscripts, the editors preserve them as witnesses of the Greek language as it was spoken and/or written at the
time of the redaction of the edited text.

20. In this category belong several witnesses of the Historia Ecclesiastica

of Pseudo Germanus of Constantinople and the Narratio Iosephi rescripta (cf.
C. FURRER, op. cit., p. 15-16), to which we add a homily on Longinos attributed
to Hesychius of Jerusalem (cf. M. AUBINEAU, Les homlies festales dHsychius
de Jrusalem II, Brussels, Socit des Bollandistes, Subsidia Hagiographica 59,
1980, p. 858-859).
21. In the cases where (inexplicably, except by fortuitous coincidence or by
contamination) c and c are both in agreement with f, the reconstitution of the
archetype of c is impossible within the rules of classical philology.
22. This is particularly the case with proper names where spelling is standardized. However, spelling variants of these names are all indicated in the



A real difficulty arises in cases where all of the witnesses that are
being used in the critical edition appear to contain a corrupted textas
is often the case in c. When an incomprehensible text occurs in an
identical or similar manner in all the preserved witnesses, it is possible to postulate that the copyists did not find it altogether incomprehensible. At least such an explanation allows us to account for the
preservation of the text in its current state, without modification. In
such cases, the critical edition preserves the text and thereby may be
seen as not critical in the traditional sense. Thus at the end of the
prologue, the edition of c reads :
,23 even
though has no subject.
The edition makes it possible to identify the different layers of editorial revisions by the scribes mentioned above. The differences between
f and c emerge from a simple comparison of the edited texts, though
it is not possible to attribute the differences either to 1 or 2.24
The codices extrauagantes
As for the codices extrauagantes, which present isolated forms of
the text, it is virtually impossible to organize the chaos. They sporadically contain very ancient forms of the Acts of Pilate ; in some ways
they are closer to f, but in other ways closer to cin fact, they evidence forms of the text that are largely more primitive than either of
the edited forms. It makes no sense to try to include them in the apparatus of either of the textual families. The editors have preferred to
merge them into a separate apparatus called the textual apparatus.
In this section there is no critical philological perspective ; the variants of the codices extrauagantes are simply enumerated, and the Greek
variants are keyed to the principal variants in the ancient translations
of the text, where appropriate. For reasons of practical organization,
the text of the codices extrauagantes and the ancient translations are
presented as an apparatus whose lemmas follow f.25
23. The manuscript W is lacking here. It contains only <><>
W. The accent on the last word excludes that it might have read
as AM does, and implies that before being corrupted it likely contained , as do O and Q.
24. The divergences between them are identifiable through the critical apparatus of the c recension ; a system of asterisks identifies, whenever possible,
the interventions of 1 and 2. An asterisk placed on cAM signifies agreement
between f and cAM and thus an editorial intervention on the level of 2which
is the most frequent case. In the same way agreement between f and c indicates an intervention at the level of 1. When c and c both agree with the
witnesses of f, asterisks are placed on each.
25. On this issue, cf. R. GOUNELLE, Ldition de la recension grecque ancienne des Actes de Pilate. Perspectives mthodologiques, Apocrypha 21, 2010
(publ. 2011), p. 31-47.



This textual apparatus is far more important than the critical apparatuses of f and c. It actually allows the reader to enter the dynamic
process of evolution of the Acts of Pilate ; the codices extrauagantes
demonstrate the intrinsic instability of the text beyond what the critical
edition of the two Byzantine forms reveals.26
The Indirect Witnesses
The indirect witnesses to the Acts of Pilate present a slightly different scenario. Even though most of them, like the codices extrauagantes, evidence a text that is largely older than the two edited families,
they are not autonomous. Rather, these witnesses are discrete sections
of the Acts of Pilate that have been integrated into another narrative
(Narratio Iosephi rescripta, Homily on Longinos) or explicative texts
(Historia Ecclesiastica of Pseudo Germanus of Constantinople).
Unlike the witnesses presented up to this point, the indirect witnesses interest the editor less for what they are than for what they testify to ; they represent states of the text that have largely disappeared
elsewhere, but that have had repercussions in the Byzantine world. For
this reason, only the sections of these witnesses that actually illustrate
the textual history of the Acts of Pilate are quoted in the edition. The
passages belonging only to these documents are not used, even if they
occur between the quotations of the Acts of Pilate. The proponents of
the New Philology will have reason to protest the violation of the
integrity of such important documents ; nevertheless, reason must prevail on this point. The edition in question concerns the Acts of Pilate
and not all of the documents that quote or use it. The editors of this
text reserve the right to refer to indirect witnesses on a utilitarian basis,
even as they would wish them to one day become the object of separate critical editions and in-depth analyses.27
To avoid any confusion between these indirect witnesses and the
manuscripts of the Acts of Pilate, there is a separate apparatus dedicated to the variants of the Narratio Iosephi rescripta, to the homily
on Longinos, and to the Historia Ecclesiastica of Pseudo Germanus
of Constantinople. This apparatus follows the reconstructed text of c,
26. Bold characters in the textual apparatus identify sporadically the variants
that trace back to a very ancient layer of the textthe textus antiquissimusbut
in the majority of the cases the reader finds a list of variants with no possibility
of giving priority to one over another.
27. The Homily of Pseudo-Hesychius alone, mentioned below, has already
been published (cf. supra). A critical edition of the Narratio Iosephi rescripta is
being prepared by G. Aragione (Universit de Strasbourg). As for the Historia
Ecclesiastica of Pseudo Germanus of Constantinople, it has a rich tradition in
Greek and in Slav (cf. R. BORNERT, Les commentaires byzantins de la divine
liturgie, du VIIe au XVe sicle, Paris, Archives de lOrient chrtien, 9, 1966), in
which only a tiny part appears to quote the Acts of Pilate ; I am not aware of
any other editing projects for the different forms of this text.



to which the indirect witnesses are similar in many respects. All of

the variants of concerned sections are indicated therein, even those that
can be explained by their larger context. Thus developments that are
specific to these works are cited on the condition that they are found in
relevant sections of the Acts of Pilate.
In the foregoing I have showed how the planned edition of the Acts
of Pilate complements a classical philological perspective. It presents
two reconstituted and standardized texts based on various manuscripts,
accompanied by their respective critical apparatus. They respond to the
demands of the New Philology as they present the abundant variety
within the manuscripts, both at the level of content and wording. For
the sake of the reader, the project as a whole is rendered in a way that
is, at the same time, chaotic yet orderlya blend that makes the edition of the Acts of Pilate an original enterprise.
As previously indicated, these editorial choices have been largely
imposed by the state of the manuscript tradition, as other more classical solutions have proven unusable.28 But chance sometimes turns out
quite well ; by their differences, the two edited texts f and c reflect, for
the most part, both the linguistic evolution of the text and the evolving
dynamics of the manuscript tradition as a whole. For instance, in the
passage mentioned above where the messenger introduces Jesus into
the praetorium (1.2), the edited text of f reads enter () while
c reads : enter into the praetorium ( ). In the
same paragraph of both of the edited texts, the two terms and
are present, both designating the piece of cloth that Pilates
messenger spreads out under the feet of Jesus. Likewise in 15.5, f and
c use different words to designate the place where the rulers of the
Jews sit ; one uses (garden), and the other (house).
The differences between the two edited families are not limited to
matters of detail. The edition of chapter 9 (IX.5), devoted to the sentence of Pilate,29 illustrates how the reader can generally recognize the
texts transformative dynamic based on the two edited texts, even without getting into the textual apparatus (see the appendix) :
(a) In f, the sentence is composed of two different elements : the
formulation of the charge addressed directly to Jesus, and then
the sentence proper where Jesus is referred to in the third person ; in c, the entire statement is addressed to Jesus.30 So the
addressee was clearly a variable aspect of this passage.
28. Cf. R. GOUNELLE, Ldition de la recension grecque ancienne des Actes
de Pilate, op. cit.
29. Cf. Annex.
30. The attentive reader will note that the process is largely underway at the
level of , and traces back even earlier, since some manuscripts of f omit the
pronoun .



(b) The sentence is identified by a subtitle in f but not in c.

(c) The style of the sentence is of a more enumerative character
in f than in c.31
(d) As for content, the flagellation is associated with the laws
of pious kings/emperors in f, but only to an imperial/
royal law in c. And the thieves crucified with Jesus, who
are anonymous in f, are named in c and emphasized by the
structure of the sentence.
The reader who refers to the critical apparatus in chapter 9 will find
the same processes at work. With the exception of variations on the
law, all the other modifications identified by a comparison of the two
synoptically edited families are also present in the codices extrauagantes.
Other examples could be furnished, but I will add here only the
issue raised by the beginning of the narrative. Here also, the two edited
forms of the text demonstrate two narratives with two distinct profiles.
In family f, the Acts of Pilate carries the title :
, Memoires of
our Lord Jesus Christ Composed under Pontius Pilate. This is followed
by a preface in which Ananias recounts how he discovered and translated this narrative, and then by a prologue that dates the writing of the
text by Nicodemus. In c, the narrative is attributed to Nicodemus in
the title :

, Memoires Concerning What Happened
to our Lord Jesus Christ under Pontius Pilate, Written by Nicodemus,
the Leader of the Synagogue of the Jews. This is followed by a prologue that dates the writing of the text by Nicodemus. The most important variant of this titlethe attribution of either to Memoires () or to its subjectis equally attested in the critical
These examples demonstrate how the synoptic edition of the two
textual families of the Byzantine period allows the attentive reader
to satisfactorily identify the textual dynamics of the Greek Acts of
Pilate, even if what is available represents only the tip of an iceberg.
Of course, not all the variants are found in the divergences between
the two edited familiesjust as the variation between and
becomes evident only in the critical apparatus.32 But the two

31. f reads : . c (= cAM)

reads : , while c has : . The writer probably modified into .
The ensuing transformation of the text, at the level of 2, is explained as the
corruption of the words , which became or , and
which became .
32. The same is true for the variant mentioned in footnote 8.



edited texts do grant the reader entry into a large portion of the textual
history of the Acts of Pilate without locking the text into a single form.
In this way, the planned edition of the Acts of Pilate fulfils a central
requirement of the New Philology without abandoning the merits of
classical philology. At the risk of seeming provocative, one might even
wonder if it is necessary to do any better. An electronic edition, however, could be helpful on several levels. It would enable a visualization
of manuscript photos and collations of the codices extrauagantes (and
the indirect witnesses) and, in the process, would more easily demonstrate their substantive coherence than does the textual apparatus. In
the same way, using photographs of manuscripts would allow a better
understanding of textual phenomena such as barely readable abbreviations, ambiguous notations, and material lacunae. But who could really
benefit from such resources and exploit them outside of the specialists
in the Acts of Pilate ? Who would know how to interpret them in light
of the whole manuscript tradition ? For lack of the ability to contextualize the information in such an eminently complex tradition, there
is a real danger that users would be lost at best, and led to erroneous
assessment at worst.

Other Unresolved Issues

If the editorial choices described above allow one to go beyond the
limits of a classical edition of the text, they do not satisfactorily resolve
one important problem, that of the end of the text.
The association of the Acts of Pilate with other texts
In its present state, the edition of the Acts of Pilate currently being
prepared for publication in the Series Apocryphorum of the Corpus
Christianorum is based on the delimitation of the text inherited from
the philology of the 19th century. The narrative ends with a hymn sung
by the Jewish people before everyone goes home (16.4). But this is not
the ending in most manuscripts. On the one hand, only six of the sixteen manuscripts contain the final hymn ; the others transmit a shorter
textsometimes due to voluntary abridgments, as in c, sometimes due
to the loss of folios. On the other hand, in nearly two thirds of the
manuscripts, the text is followed by another text. In eight codices, the
correspondence of Pilate and the emperor (Anaphora Pilatou and Paradosis Pilatou33) immediately follows the Acts of Pilate. This is also the
33. C. FURRER, op. cit., p. 30, provides a survey of the manuscripts of the
Acts of Pilate containing the Anaphora and the Paradosis, but the list does
not coincide exactly with the list of manuscripts given in the note 33 : there
is also the mention, p. 13, of the Anaphora and the Paradosis in the description of the Manuscript I. The edition of C. VON TISCHENDORF implies that the
Anaphora and the Paradosis are two independent texts. In reality, they form a



case in several manuscripts of the family f, in all the witnesses of c

and in one codex extrauagans. But in one witness of f and one codex
extrauagans it is followed by the Declaration of Joseph of Arimathaea
(Narratio Iosephi), which is in turn followed by the Epistula Pilati ad
Not only do the majority of the manuscripts of the Acts of Pilate
include other texts relative to Pilate, but this phenomenon has also had
consequences for the text history itself. For example, in the manuscript
Paris, BnF gr. 929 (which is referred to as E in the upcoming edition
of the Acts of Pilate), the Anaphora Pilati and the Paradosis Pilati
are inserted in the text of the Acts of Pilate itself, after the prologue.
It is clearly an intentional insertion, because the title given to the narrative explicitly presents the Memoires of our Lord Jesus Christ (the
traditional title of the Acts of Pilate) as a report that Pilate sends to the
emperor (the traditional title of the Anaphora Pilati).35 The text that we
have entitled Narratio Iosephi Rescripta, for its part, fuses the Acts of
Pilate with the Declaration of Joseph of Arimathaea and with another
source that has not been identified.36 The association of the Acts of
Pilate with these texts is thus an intrinsic element of its history.
Editorial Issues
The editorial choices presented above allow the edition to indicate
the existence of both the long forms of the Acts of Pilatewhich
extend the text all the way to the final hymn (16.4)and the shorter
forms, thanks to the text c. But they are unable to account for the
constitution of a cycle surrounding the figure of Pilate in Byzantine
Yet the presence of the Anaphora Pilati and the Paradosis Pilati
in all the manuscripts of the family c suggests that this cycle is an

homogenous textual ensemble in the manuscripts ; there is no problem of continuity in the vast majority of the manuscripts containing the two narratives.
These texts, which are being critically edited by G. Aragione (Universit de
Strasbourg), represent a rich tradition, which has not yet been studied in detail ;
a first account has been given by A.-C. BAUDOUIN, Tmoins manuscrits connus
et inconnus du Rapport de Pilate Tibre (CANT 65 and 66), Revue dtudes
augustiniennes et patristiques 54, 2008, p. 183-20. They have been translated
by R. GOUNELLE, Rapport de Pilate, Rponse de Tibre Pilate, Comparution
de Pilate, in J.-D. KAESTLI et P. GEOLTRAIN (dir.), crits apocryphes chrtiens
II, Paris, Bibliothque de la Plade 516, 2005, p. 299-327.
34. Cf. C. FURRER, op. cit., p. 12-13. On this Declaration, of which a new
edition by G. Aragione (Universit de Strasbourg) is underway, cf. A. FREYB. OUTTIER, Dclaration de Joseph dArimathe, in J.-D. KAESTLI and P. GEOLTRAIN (dir.), op. cit., p. 331-354.
35. Cf. C. FURRER, op. cit., p. 22.
36. Concerning this text cf. R. GOUNELLE, Les recensions byzantines de
lvangile de Nicodme, op. cit., p. 52-53.



integral part of the process leading to the development of this familyor rather, that it was constituted by writer . All the witnesses of
this family abridge the end of the text, preserving only the following
lines from the chapters 16.3.2 (XVI.7) and 16.4 (XVI.8) : Anna and
Caiaphas said : You have rightly spoken what is written in the law of
Moses, but it is better for us to be silent concerning the rest and for
each one to go to his house. And give glory to God. It is likely that,
given the evolution of the Acts of Pilate in other textual traditions,37 the
abridgement of the chapter 16 and the extension of the narrative with
the help of other texts are two related phenomena. If such is the case,
it would be questionable at the very least to edit the short text without
the Anaphora.
It is therefore important that the edition of the Acts of Pilate itself
reflects this merging of several texts beyond a simple mention in the
introductory description of the manuscripts of family c. But how can
we do this without directly editing, after the Acts of Pilate itself, the
form of the Anaphora Pilati and the Paradosis Pilati found in family
c ? This, in turn, would imply adding to the critical and textual apparatus of both families the other textual extensions found in some manuscripts of f and in some of the codices extrauagantes.38
Such an undertaking would provide information of great interest on
the history of the texts relative to Pilate, helping to establish whether
the modifications made to the various editorial layers of the Acts of
Pilate have parallels in other associated texts. But it would also require
penetrating deeply into the textual tradition of the Anaphora Pilati, the
Paradosis Pilati, and the Declaration of Joseph of Arimathaea, and
would consequently take on such dimensions as to discourage any textual editornot to mention that it would partially repeat the work of
G. Aragione, who is currently preparing an edition of these texts for the
Series Apocryphorum of the Corpus Christianorum.
A digital approach to organizing the identified material with the
help of several different research teams could offer some interesting
solutions. Such an approach would enable the user to reconstitute the
sequences of text from one editorial layer, revealing the coherence or
structure of a collection of documents from several different manuscripts. In the case at hand, a combination of the information provided
by G. Aragione with that provided by the Acts of Pilate research team
could identify the constitution of a cycle of Pilate in Byzantine literature. Such collaboration could also resolve some of the problems

37. Cf. particularly the expansions of the text in the Latin world. Cf. Z. IZYThe Evangelium Nicodemi in the Latin Middle Ages, in Z. IZYDORCZYK (ed.), The Medieval Gospel of Nicodemus, Tempe (AZ), Medieval and
Renaissance Texts and Studies 158, 1997, p. 44-101 (46-66).
38. Not to mention a similar operation in the other textual traditions that
will be edited in the CCSA (Armenian, Georgian, etc.).



mentioned above concerning the indirect witnesses. Once the latter are
edited, one could potentially have access to the context of certain quotations in the textual apparatus of co from within the edition of the Acts
of Pilate. Is this a historians fantasy or a future possibility ? Only time
will tell.

If the current editorial project does not satisfactorily account for the
insertion of the Acts of Pilate into a collection of texts that can be
identified, by and large, as a coherent corpus, it should nevertheless
be able to satisfy two contradictory needs : first, it presents in a logical fashion a rich textual tradition by reconstituting two textual forms
that circulated in Byzantine Christianity. Secondly, the specific textual
apparatus allows access to other preserved textual formsa chaos of
information impossible to organize otherwiseand thereby gives access
to the fabrication of the text.
As this editing project limits itself to the preserved manuscript tradition, we feel that it satisfies the legitimate concerns of the New Philology about the classical method without abandoning the precious
tools developed by the latter. It remains to be seen whether such an
edition will be understandable and useful to its users. The risk is that
the reader, surprised by its complexity, might abandon it entirely and
resort to C. von Tischendorfs edition.39 While the risk is real, we must
continue to seek innovative strategies to faithfully represent rich and
fluid textual traditions.

39. C.





Text of f


{ }


Critical apparatus
1 Z : F C om. GH | GHC : Z om. F | 2 post
add. G H | 3 C : Z
om. GH | 4 GHCZ : F | GZ : HC F | F :
GHC Z | 4-5 ante add. FZ | FGZ : C
H | 5 FGHC : Z cf. c | FGHZ : C | om. C | GHZ : C om. F | 6 FGHZ :
C | om. FH | (om. H) FHCZ : G |
7 FGCZ : H | GHZ : FC

Textual apparatus
Variae lectiones codicum IJBN et loci ad editionem pertinentes ex arm
cop geo syr lat
1 IJ : B
N | (. . . BN) IJBN ut fZ c cf. fFC
arma110 cop geo latB1284, 900 syr : om. IJ ut fGH | B :
IJ om. N | IBN : J | 2
FIJ cf. arma cop syr latA latB1 latC :
B cf. c arms geor latB2d om. IJB geo | :
N titulum habent etiam fCZ cop geo latB1336 cf. latA | I : J N B cf. c |
J : B (- I) IN | post add. J |
4 ante add. IJ | 5 ante add. J | post
add. B . N | geo latB198, 336 latC : (
IJ) IJB cop latArr syr N | IJB
arms syr cf. arma : N | 6 N arms cop geo
latArr latB198, 336 : B arma syr IJ |
BN : I om. J | 6-7 latA latB1 latC :
IJB ut c cop latA179 om. N | :
B ut fFC
N IJ |
post add. B
N latA latB1 cf. c armr cop latB2 latC syr



Text of cqw


Critical apparatus
1 OQM : WA | 2 OQA : M W | om. AM |
OQW : AM | 3 OQW : AM | OQAM :
W | post add. AM* | OQW : AM* |
( add. W) OQW : (om. A) (
add. A) AM* | 4-5 OQ : W
AM | 5 om. AM | om. W | OQ* :
AM | ante add. W | W : O : Q M |
W : OQ A M

University of Winnipeg


The apocryphal Evangelium Nicodemi, a Latin translation of the
Greek Acts of Pilate, is richly attested in over 450 extant medieval
manuscripts. A partial collation and a statistical analysis of the variance data have revealed the existence of three major textual forms of
the Evangelium and of several lesser, often hybridized sub-forms. The
Latin apocryphon must have been subject to horizontal transmission,
to corrections against later forms of its Greek source, and to deliberate expansion of its textual scope. Rather than pursuing the elusive
and, ultimately, hypothetical Ur-text, a new edition of the Evangelium
Nicodemi should attempt to capture its textually diverse and dynamic
nature by representing its key textual performances, the ones that most
profoundly influenced medieval and post-medieval culture. This would
be the first step towards accommodating the medieval condition of the
apocryphon, with its susceptibility to and perhaps even encouragement of textual variance, and a step away from restraining that variance with assumptions born in the age of print.
Lapocryphe Evangelium Nicodemi, traduction latine du texte grec
Actes de Pilate, est abondamment attest dans plus de 450 manuscrits
qui ont t conservs. Une collation partielle et une analyse statistique
de la variance de ces attestations ont rvl lexistence de trois formes
textuelles principales de lEvangelium, et de plusieurs sous-formes,
souvent hybrides. Lapocryphe latin a d subir une transmission horizontale, des corrections fondes sur des formes ultrieures de sa source
grecque, et une expansion intentionnelle de sa porte textuelle. Plutt
que de poursuivre un texte original insaisissable et, en fin de compte,
hypothtique, une nouvelle dition de lEvangelium Nicodemi devrait
tenter de saisir son caractre textuellement divers et dynamique, en
reprsentant ses accomplissements textuels-cls, ceux qui influencrent
le plus profondment la culture mdivale et post-mdivale. Cela
serait une premire tape, qui prendrait en compte le statut mdival
de lapocryphe et sa tendance, voire son incitation, la variation textuelle, et ce serait aussi une tape vers un texte dont la variance textuelle ne serait plus restreinte par des prsupposs ns avec lge de

Apocrypha 23, 2012, p. 99-116



Although the apocryphal Evangelium Nicodemi (EN) has a long print
history, extending all the way back to the fifteenth century, that history is largely self-referential as most of the ENs early modern editors
recycled earlier prints, and even Thilo and Tischendorf had only limited
knowledge of its textual condition before print.1 The first scholar to
appreciate the vastness and diversity of the ENs textual past was Ernst
von Dobschtz, who not only identified dozens of its medieval copies
but also collated a large number of them. Unfortunately, he never completed his project of re-editing the apocryphon, and his papers remained
out of scholars reach for most of the twentieth century.2 Consequently,
the textual scope and complexity of the EN before print began to
emerge only in the 1990s, through the research undertaken under the
auspices of the Social Sciences and Humanities Research Council of
Canada and the Association pour ltude de la littrature apocryphe
chrtienne (AELAC). That research unearthed 424 extant manuscripts
of the complete or partial text of the EN, summarily described in the
Census.3 In close to two decades since that publication, thirty-two
additional witnesses have come to light, mostly thanks to on-line tools
and personal communications from other researchers. With over 450
manuscripts,4 the EN represents one of the largest, if not the largest,
documented Latin apocryphal tradition in Western Europe.
The Latin apocryphon that eventually came to be known as the
Evangelium Nicodemi originated as a translation from the Greek Acts
of Pilate, probably somewhere in the fifth century.5 Over the next three
1. J. C. THILO (ed.), Codex apocryphus Novi Testamenti, vol. 1, Leipzig,
F. C. G. Vogel, 1832, p. 491-802 ; C. VON TISCHENDORF (ed.) Evangelia apocrypha, 2nd rev. ed., Leipzig, H. Mendelssohn, 1876, p. 333-432. On the early editions, see Z. IZYDORCZYK, The Unfamiliar Evangelium Nicodemi, Manuscripta
33 1989, p. 170-176 ; ID., The Earliest Printed Versions of the Evangelium
Nicodemi and Their Manuscript Sources, Apocrypha 21, 2010, p. 129-131.
2. VON DOBSCHTZ announced his intention to edit the EN in Nicodemus,
Gospel of, A Dictionary of the Bible, J. Hastings (ed.), New York, C. Scribners Sons, 1919, vol. 3, p. 544-547. His papers are now in the care of AELAC
at the Library of the Universit de Lausanne.
3. See Z. IZYDORCZYK, Manuscripts of the Evangelium Nicodemi : A Census, Toronto, Pontifical Institute of Mediaeval Studies, Subsidia Mediaevalia 21,
1993. In the remainder of this paper, all manuscripts listed in the Census will
be referred to by the first two letters of their current location and the Census
number ; see the Appendix for a complete list of sigla mentioned in this paper.
4. This number does not include abridged versions of the Evangelium
embedded in the Legenda aurea and the Speculum historiale.
5. The terminus ante quem for the Latin translation is the date of the earliest surviving manuscript, the Vienna palimpsest (Vienna, NB MS 563), dated
by M. DESPINEUX, Une version latine palimpseste du Ve sicle de lvangile de
Nicodme (Vienne, NB MS 563), Scriptorium 42, 1988, p. 178-179, to the



centuries, that early Latin translation was re-written, revised, and reconstituted, acquiring a new title (Gesta Salvatoris), the Descensus
narrative, and a letter of Pilate. Since, except for the Vienna palimpsest,
no manuscripts have survived from before the ninth century and since
early medieval writers are silent about the apocryphon, the reasons for
those changes remain obscure ; we know just as little about the uses
of and attitudes towards it during that period. However, from the ninth
century onwards, extant manuscripts begin to shed light on its textual
forms, on its reception, and on religious and cultural tendencies that
may have encouraged its continued transformations.
The manuscripts that transmit the EN usually carry other texts
as well, and to the extent that similar works were often transmitted
together, the ENs companion works may be indicative of the way
the apocryphon itself was perceived and used. Not unexpectedly,
it is often found in the company of other apocryphal works such as
the infancy narratives, Visio Pauli, or Vita Adae et Evae (e.g., Be015,
Mo178, M180, Pa252),6 suggesting that at least some scribes were
aware of its affinity to the works from the margins of the canon.
Other scribes, however, appear to have been so highly impressed
by it that they moved it from the margins closer to, even into the
midst of, canonical texts (e.g., Lo157, Br033, M190, M206), suggesting perhaps some liturgical interest in the EN. The company of
devotional and passion tracts, such as the Planctus Mariae, Dialogus
beatae Mariae et Anselmi de passione Domini, or Meditationes vitae
Christi,7 suggests that one of its frequent uses, especially in the later
Middle Ages, was as a devotional aid. This seems to be a common
purpose behind many of its vernacular translations as well, which
could serve as sources of private or communal readings ; for example,
a German version of the apocryphon was recommended for meal time

fifth century on paleographical grounds. The terminus post quem is the date
of composition of the Greek Acts of Pilate, which R. GOUNELLE, Un nouvel
judo-chrtien ? Les Actes de Pilate, in J. SCHRTER (ed.), The Apocryphal Gospels within the Context of Early Christian Theology, Louvain, Peeters, in press,
places between 320 and 380 A.D.
6. J. GIJSEL, ed., Libri de nativitate Mariae : Pseudo-Matthaei evangelium,
Turnhout, Brepols, CCSA 9, 1997 ; T. SILVERSTEIN and A. HILHORST, Apocalypse
of Paul : A New Critical Edition of Three Long Latin Versions, Geneva, Cramer, 1997 ; W. MEYER, Vita Adae et Evae, Abhandlungen der philosophischphilologischen Classe der kniglichen Akademie der Wissenschaften 14, 1878,
p. 187-250.
7. C. W. MARX, The Quis dabit of Olgerius of Tridino, Monk and Abbot
of Locedio, Journal of Mediaeval Latin 4, 1994, p. 118-129 ; Oskar SCHADE,
Interrogatio sancti Anshelmi de passione Domini, Knigsberg, 1870 and PL
159, 271-290 ; PSEUDO-BONAVENTURE, Meditationes vitae Christi, in S. R. E. Cardinalis S. Bonaventurae Opera omnia, ed. A. C. PELTIER, vol. 12, Paris, L.
Vivs, 1868, p. 509-630.



readings on Good Friday at the convent of St. Katharina in Nrnberg.8

Also common was its placement in collections of sermons ; in fact, the
apocryphon itself was recast as a sermon on more than one occasion.
For example, version LatB2 opens with a typical homiletic introduction, Audistis fratres carissimi (Ca44), while another version found
in copies of the Carolingian homiliary from Saint-Pre de Chartres
(e.g., Gr102), ends with a homiletic epilogue, Videte fratres quomodo
nos redemit Christus.9 This might also explain why parish priests
and monastic chaplains found it useful and either copied it themselves
or owned copies of it.10 And finally, the EN was often treated as a
historical record and revised as such ; for instance, it was copied and
in places altered by the well-known historian Ademar de Chabannes (c.
988-1034 ; Pa263),11 while a medieval English historian incorporated it
into a large chronicle, Eulogium historiarum (siue temporis).12 One of
the secrets of the ENs propensity for change and of its longevity
lies, no doubt, in the fact that it appealed to so many different interests
of so many different readers.

Filiation of texts
The EN does not give up its secrets easily. The sheer number and
diversity of its extant witnesses pose numerous challenges, the least of
which are the logistics of acquiring reproductions from over 160 libraries and the practicalities of reading and transcribing a variety of scripts
and abbreviation systems. Mapping out genealogical affiliations of
those texts and determining patterns of their transmission are seriously
complicated by ample evidence of horizontal transfer of its readings ;
by its enduring relationship to the Greek Acts of Pilate, its ultimate
8. P. RUF, Mittelalterliche Bibliothekskataloge Deutschlands und der
Schweiz, vol. 3.3.4, Munich, C. H. Becksche Verlagsbuchhandlung, 1939,
p. 644. At least eight Latin manuscripts come from female convents.
9. On these and other homiletic uses of the EN, see Z. IZYDORCZYK, Preaching Nicodemuss Gospel, in Medieval Sermons and Society : Closter, City, University, Louvain-la-Neuve, 1998, p. 9-24. Whether such recastings were true
sermons actually used for preaching is a matter of debate; see Caitrona
Dochartaigh, Homiletic Texts and the Transmission of Eschatological Apocrypha in a Medieval Irish Context, p. 142 in the present volume.
10. See Index 4 : Persons and Places Connected with Manuscripts, in IZYDORCZYK, Census, p. 269-82.
11. In chapter 19.1, Ademar changed, for example, the number of years that
had to be completed before the Incarnation from the traditional 5500 to tria
milia nongenti quinquaginta duo ; that latter, more exact figure he took from
Bede, and in the margin he added a note, echoing Bede, secundum hebraicam
ueritatem ab origine mundi adusque incarnationem Christi sunt anni iii milia
dcccc lii. Secundum lxx sunt anni v milia d (Paris, BN MS Lat. 3784, fol.
12. On this and other historical uses of the Evangelium, see IZYDORCZYK,
The Medieval Gospel of Nicodemus, p. 81-82.



source ; and by fluctuations of its textual boundaries. The feasibility of

constructing a complete stemma textuum that would account for all of
its attestations is rather low, as is the likelihood of recovering its Latin
Ur-text. However, even partial reconstructions of its textual descent and
of the relationships among its forms may suggest, and perhaps even
enable, useful editorial solutions.
Establishing genealogical affiliations among hundreds of texts defies
traditional philological methods based on full collation and detailed
comparison of all texts : limitations of time and resources render such
approach unworkable. That is why, in my own research on profiling
and filiation of the ENs witnesses, I opted for selectivity and efficiency
rather than exhaustiveness. Instead of working with the entire text, I
selected three longer passages from different parts of the EN as edited
by Tischendorf as well as several shorter ones that either had been
known to exhibit interesting variation or corresponded to the fragments
preserved in the oldest manuscript of the EN, the Vienna palimpsest.13 I
collated the selected passages from all pre-twelfth century codices14 and
from a sampling of later ones showing different combinations of the
prologues and epilogues, marked by different incipits and explicits, and
coming from different geographical regions, all together from about one
fifth of all witnesses (approx. 85 texts). On the basis of these test collations, I established a list of approximately 100 sites specific locations
in the text where significant or indicative variation tended to occur.
For every location, I identified anywhere from two to thirty significant
variants ; by a significant variant, I mean a reading that is unlikely to
have been independently produced by different scribes through accidental, palaeographically-inspired,15 or biblically-inspired convergence.
Then I surveyed, in a more rapid manner, only the relevant sites in
approximately 60 additional manuscripts to determine which of the
identified variants those manuscripts supported. All results were coded
and collected in a database for maximum parsimony analysis with bootstrapping statistical support, using the philogenetic software package,
originally designed for evolutionary biology, called PAUP.16
13. G. PHILIPPART, Les fragments palimpsests de lvangile de Nicodme
dans le Vindobonensis 563 (Ve s. ?), Analecta Bollandiana 107, 1989, p. 171188.
14. The number of manuscripts of the EN increased dramatically from the
twelfth century onwards : while approximately 31 manuscripts have survived
from the ninth to the eleventh centuries, 49 are associated with the twelfth
century alone, and close to 70 with the thirteenth. An overwhelming majority of later medieval manuscripts transmit the text types and/or textual variants
attested in the early medieval manuscripts.
15. Palaeographically-inspired convergence involves common misreadings of
an abbreviation ; for example, the alternation between accionibus and accusationibus may easily result from the abbreviated form accbus.
16. D. L. SWOFFORD, PAUP*. Phylogenetic Analysis Using Parsimony (*and
other methods), Sunderland, MA, Sinauer Associates, 2001. The maximum parsimony analysis identifies those manuscript groupings and stemma reconstruc-



The results of that analysis have been encouraging although not

wholly satisfying. The hypothetical trees of textual descent generated
by PAUP not only confirmed my initial intuitions about the complexity
of the ENs tradition but also convincingly resolved some of its main
branches. They confirmed the existence of three major paths of textual
transmission, usually referred to as LatA, LatB, and LatC ;17 they set
LatB apart from LatA and LatC, which appear to have shared a common ancestor ; and they identified and correctly placed some hybrid versions, such as the Bohemian redaction between LatA and LatB, and the
Troyes redaction as an offshoot of C.18 They also identified a number
of discrete manuscript groupings, correctly linking manuscripts already
known to have been copied from one another or from the same exemplar. In fact, these preliminary results have been encouraging enough
that I intend to include the rest of the manuscripts in this analysis. The
classification and filiation of all ENs manuscripts remains a project in
However, while my selective approach to phylogenetic analysis of
the ENs texts has produced helpful results, it has not produced certainty, as my expectations based on manual comparisons are not always
reflected in the stemma reconstructions generated by PAUP. For example, the various trees proposed by the software do not reflect a link I
would expect between one specific family of the main text-type, the
Orlans (or Bassi et Tarquilionis) family of LatA (represented by
Or215), and a subfamily of LatB, designated as LatB1 (represented by
Pa284) ;19 or between the same family of LatA and LatC (represented
by Ba12). The most likely reason for these discrepancies is that, while
my expectations had been formed on the totality of the text, the statistical analysis was conducted on samples only ; moreover, the statistical
analysis was carried out with all variants assigned an equal weight,
whereas my expectations may have involved judgments, conscious or
not, about relative significance of and directionality among variants at
particular sites.20
tions which involve the least amount of convergence and reversal among variants ; in other words, it identifies historical (stemmatic) reconstructions with the
minimum number of evolutionary steps. Bootstrapping is a method for assigning a measure of accuracy to reconstruction estimates by resampling the original
data set and constructing a series of replicates.
17. For a brief description of those textual traditions, see Z. IZYDORCZYK,
The Medieval Gospel of Nicodemus : Texts, Intertexts, and Contexts, Tempe, AZ,
Medieval & Renaissance Texts & Studies, 1997, p. 47-53.
18. On the Bohemian redaction and its manuscripts, see Z. IZYDORCZYK and
W. WYDRA, A Gospel of Nicodemus Preserved in Poland, Turnhout, Brepols,
CCSA Instrumenta 2, 2007, p. 18 ; on the Troyes redaction, see Z. IZYDORCZYK,
The Latin Source of an Old French Gospel of Nicodemus, Revue dhistoire des
textes 25, 1995, p. 265-279.
19. On LatB1, see below.
20. For recent studies on the effectiveness of phylogenetic methods in stemmatic analysis, see H. F. WINDRAM et al., Dantes Monarchia as a Test Case for



Horizontal transmission
The reasons for the discrepancies between the manual and PAUP
results may go beyond the limitations of my procedure. The ENs transmission clearly involved more than just simple vertical copying, with
occasional improvements to Latin accidence, syntax, or lexis. Many
scribes engaged in large scale revisions, often deliberately drawing on
multiple exemplars and merging readings from different lines of transmission. In the process, they changed the texture, shape, and scope of
the apocryphon. We can see traces of such activities in many manuscripts. For example, a late medieval reader carefully compared two
texts of LatA, one in the ninth-century Be023 and another in the tenthcentury M199, against each other and entered interlinear variants and
corrections in both manuscripts. In Pa286, the text of the EN begins
as LatA but gradually changes into the LatB type ; another scribe (or
reader) added some LatB corrections to the beginning of the text. In
Ox230, the text opens as LatB but afterwards drifts towards LatA ;
and in Na210, the scribe appears to have changed his exemplar at the
beginning of the Descensus (fol. 361va). Examples of this sort could
easily be multiplied.
Editorial activities of medieval scribes could involve different texts
from the same tradition or texts from different traditions. In extreme
cases, such horizontal transfer of text led to the emergence of hybrid
versions, two of which became highly influential during the later Middle Ages and well into the Early Modern period, giving rise to numerous vernacular translations. The twelfth-century scribe responsible for
the so-called Troyes redaction (Tr362), the source of several English
and French translations, effectively blended version LatA with LatC.
Through much of the trial and Joseph of Arimathea sections, he followed LatC, retaining many of its idiosyncratic passages, albeit not consistently. When he drew on LatA, it was often (but not always) to supply the passages that had been deleted in LatC ; for instance, he restored
the Descensus, highly abridged in LatC, on the basis of LatA. However,
he ended the text with perhaps the most characteristic passage of LatC,
namely, the discussions between Pilate and the high priests in the temple.
The other popular hybrid form of the EN, the Bohemian redaction, is
preserved in several fifteenth-century manuscripts from Austria, Bohemia, and Poland and appears to have originated in Central Europe. It
stands behind a medieval Czech translation and a number of Early Mod-

the Use of Phylogenetic Methods in Stemmatic Analysis, Literary and Linguistic Computing 23.4, 2008, p. 443-463 ; T. ROOS and T. HEIKKIL, Evaluating
Methods for Computer-Assisted Stemmatology Using Artificial Benchmark Data
Sets, Literary and Linguistic Computing 24.4, 2009, p. 417-433.



ern German prints.21 The Bohemian redaction is compiled partly from

LatA and partly from LatB2. It opens with the LatA prologue (Factum est...), which ends with a single sentence from the LatB preface,
introducing Aeneas as the Greek translator of the Hebrew work written by Nicodemus. The apocryphon is extensively and systematically
rewritten and includes many passages characteristic of LatB2, such as
Pilates question about his suitability to judge a king and his comment
about his wife Proculas sympathy for the Jews, as well as large sections drawn from LatA, such as the Descensus.
Obtaining alternative copies of the EN for the purpose of comparing, correcting, or combining of the text was not as difficult as it might
appear. A quick survey of ownership marks in the manuscripts of the
EN reveals that a number of monastic libraries apparently sponsored or
owned three or even four copies of the apocryphon. For example, four
extant manuscripts bear notes identifying them as belonging to the Carthusian monastery at Michaelsberg, Mainz ; similarly, four manuscripts
come from the Regular Canons at St. Victor in Paris, and four from the
Benedictines at Tegernsee. At least ten other monastic libraries owned
three copies each, all of them still extant today.22 According to medieval
catalogues, which record also copies no longer extant, Christ Church
priory in Canterbury owned no fewer than seven different copies of the
EN in the fourteenth century, and a fifteenth-century booklist from the
Augustinian abbey at Leicester records six, although some probably in
French.23 Revisions involving multiple copies and multiple versions of
the EN were thus relatively easy and relatively frequent.
In fact, scribes did not need to use multiple manuscripts to access
different texts of the EN as some codices contained more than one copy
of the apocryphon. For example, the scribe of Ox238, of the thirteenth
century, first included a short version of the EN, beginning with the
story of Joseph of Arimathea and based on LatA, in a compilation of
materials on the Rood-tree legend, and then he copied a complete text
of LatB2.24 In another manuscript, which has only recently come to
my attention, Krakw, Biblioteka Polskiej Akademii Umiejtnosci MS
1713, a rare version of the EN (the so-called Krakw version), ending
21. On the vernacular legacy of this redaction, see, Z. IZYDORCZYK and
C. FILLMORE-HANDLON, The Modern Life of an Ancient Text : The Gospel of
Nicodemus in Manitoba, Apocrypha 21, 2010, p. 116-117.
22. See Index 4 in Izydorczyk, Census.
23. M. R. JAMES, The Ancient Libraries of Canterbury and Dover, Cambridge, at the University Press, 1903, nos. 541, 1258, 1301, 1373, 1389, 1420,
1542 ; T. WEBBER and A. G. WATSON, The Libraries of the Augustinian Canons,
The British Library, Corpus of British Medieval Library Catalogues 6, 1998,
nos. 148 (2 copies), 159, 160, 305, 1235.
24. A similar situation occurs in Pa259, which also preserves a short version
of the EN in the context of Rood-tree materials, but preceded by a complete
copy of LatA.



without the Descensus,25 is fused (fol. 228v-235r) with the preceding

text of the Historia trium regum, while a more typical copy of LatA
occurs later in the codex (fol. 312r-323v).26 Multiple copies of the EN
can also be found in Ca44, Le135, and Va378 (the last one boasting
three texts of LatA), although further research would be needed to
determine exactly when they were bound into single volumes.
Manuscripts of the EN were abundant, perhaps even ubiquitous,
especially in the later Middle Ages, and horizontal transmission of
readings and of portions of the text was probably equally common.
This may constitute another factor negatively affecting phylogenetic
analysis of the ENs textual tradition, for all phylogenetic methods still
have great difficulties in recognizing contamination, the copying of an
exemplar from multiple sources.27

Renewals by translation
An additional factor complicating the results of my phylogenetic
analysis and of the filiation process in general has to do with the
apocryphons status as a translation. It was first rendered from Greek
into Latin possibly in the middle of the fifth century. However, there
is evidence suggesting that the Latin apocryphon did not sever its ties
with the Greek tradition after its initial translation but continued to be
influenced by it, with the original translation being overlaid with layers
of subsequent selective translations. One such partial translation from a
Greek source appears to have played a role in the separation of textual
traditions LatA and LatB. For example, in chap. 1.6, when Pilate orders
that the Jews choose their own strong men to hold the standards, he
addresses seniores plebis in LatA (La133, fol. 4v), while in LatB he
speaks to sacerdotibus populi (Pa284, fol. 110v). The two contrasting
phrases appear to represent two different renditions, one secular and
the other religious, of the Greek .28 Certain readings
of LatA and LatB even suggest that the redactors Greek exemplar differed from the one used by the original translator. Thus in the opening
chapter of LatA, the Jews come to Pilate aduersus ihesum, accusantes
25. This Latin version was edited from another manuscript, Kr127, in IZYand WYDRA, A Gospel of Nicodemus, p. 45-97.
26. M. PIACENTINI, Un importante contributo allo studio degli apocrifi. Il
Vangelo di Nicodemo in Polonia : tradizione latina e traduzione polacca, Studi
Slavistici 8, 2011, p. 195-201 ; only the latter version is mentioned in J. CZUBEK, Katalog rkopisw Akademii Umiejtnosci w Krakowie, Krakw, Nakadem
Akademii Umiejtnosci, 1912, p. 26-27.
27. ROOS and HEIKKIL, Evaluating Methods, p. 428.
28. All citations of the Greek text and all references to Greek manuscripts
and manuscript families are derived from a draft of a new edition of the Greek
Acts of Pilate, currently being prepared by R. GOUNELLE, C. FURRER, and
A. FREY, under the auspices of AELAC.



eum (La133, fol. 2v), and this is also the reading of the Greek
family . However, in LatB, the phrase aduersus ihesum is missing
(Pa284, fol. 109r), as it is in the Greek family c. In chap. 1.5 of LatA,
the Jews say to Pilate regarding the miracle of standards bowing before
Jesus that they only saw quomodo se inclinauerunt signiferi et adorarent ihesum(La133, fol. 4v) ; but in LatB, they saw homines qui signa
tenent quemadmodum inclinauerunt uultus ut adorauerunt ihesum
(Pa284, fol. 110v). While most Greek manuscripts support the reading
of LatA, Greek manuscript N refers to in the same context,
which seems to have migrated into LatB (here and elsewhere) as uultus. Finally, in 1.6, the Jewish strong men are set ante conspectum
praesidis in LatA (La133, fol. 4v), as in the majority of Greek witnesses, but ante tribunal presidis in LatB (Pa284, fol. 111r), with the
word tribunal corresponding to of Greek manuscripts J, B,
and C. Unfortunately, given the late and highly revised character of
the extant Greek manuscripts, it is not at present possible to determine
which set of readings was found in the original translation and which
in the source used for the revision.
Further examples of continued Greek influence can be found within
tradition LatB itself, which splits into two major families, LatB1,
attested in five manuscripts, and LatB2, attested in thirteen.29 The two
families have enough readings in common to warrant their association
into a single tradition, but they differ in a number of ways, one of
which involves, again, what appear to be traces of different Greek antecedents. For instance, in the episode in which the Jewish council interrogates the three rabbis from Galilee (chap. 14.2 ; this passage is absent
from LatA and LatC), LatB1 poses a short question, which corresponds
to what can be found in Greek family c :
ad hoc uenistis adnuntiare nobis hc aut uenistis adorare deum. aut quid hunc multiloquium fecistis coram omni
populo ? (Pa284, fol. 121r)
(Have you come to announce to us these things or have you
come to adore God ; or why do you make long speeches before the
people ?)
whereas LatB2 asks a longer one, which finds parallels in Greek
family :
Ad hoc uenistis nuntiare nobis an uenistis orationem
deo dare ? Dixerunt autem eis. Venimus orationem dare deo. Dicunt
seniores et principes sacerdotum et leuite ad eos. Et si rationem
uenisti reddere deo. deliramento isto quid murmurastis ante omnem
populum ? (Ca044, fol. 43v-44r)

29. For a brief description of the two families, see IZYDORCZYK and WYDRA,
The Gospel of Nicodemus, p. 17.



(Have you come to announce to us or have you come to offer

a prayer to God ? And they said to them : We have come to offer a
prayer to God. The elders of the Jews and the princes of the priests
and the Levites said to them : And if you have come to render a
prayer to God, why do you murmur this nonsense before the people ?)
Similar differences, going back to different Greek models, are not
difficult to find in the two families of LatB. In chap. 15.5, for instance,
when the leaders of the Jews arrive at Nicodemuss place to speak to
Joseph, Nicodemus leads them, according to LatB1 (Pa284, fol. 123v),
in orto suo, and according to LatB2 (Ca044, fol. 46r), in domum
suam. The reading of LatB1 derives from the Greek phrase
, attested in Greek family , while that of LatB2 from the
Greek phrase found in family c.30 Unfortunately,
none of the existing Greek manuscripts matches either LatB1 or LatB2
in their entirety.
Finally, a revision by fresh translation from Greek may have played
a role in an unexpected reappearance of the apparently ancient text in
certain late medieval, otherwise unremarkable copies of the EN. For
example, an early sixteenth-century manuscript that originated probably
in Southern Germany, Wa391, transmits, on the whole, a common LatA
text, but it exhibits some unquestionably ancient readings, including the
mention of the god in whose name evil spirits are cast out (in deo
scolapii, fol. 1r). Although that god is also named in the fifth-century
Vienna palimpsest, it was subsequently lost to the entire Latin tradition, and the Washington manuscript is its only other witness. Wa391
is probably related to the Krakw EN, preserved in two central European codices (Kr127 and Krakw, Polska Akademia Umiejtnosci MS
1713).31 All of them, for instance, use the phrase scema signorum in
chap. 1.5, reflecting the Greek , whereas the rest
of the Latin tradition writes signa signorum.
chap. 1.5 family
c (f is very similar)
Videntes autem iudei scemo signorum (fol. 2r)
Videntes autem Iudei scema signorum (fol. 89v)
Uidentes autem iudaei signa signorum (La133, fol. 4r)
Videntes autem iudei signa (Sa336, fol. 112v)
Videntes iudei signa (Ca044, fol. 40r)
uidentes autem iudaei (Ba012, fol. 123v).
The Krakw EN is the sole Latin version, again besides the Vienna
palimpsest, to end with the full original conclusion of the Greek Acts of
Pilate, placed exactly as in the Greek original, at the close of chap. 16.
30. Cf. R. GOUNELLE, Editing a Fluid and Unstable Text : The Example of
the Acts of Pilate (or Gospel of Nicodemus), in the present volume, p. 81-97,
esp. p. 84 and 90.
31. Edited in IZYDORCZYK and WYDRA, A Gospel of Nicodemus, from Kr127.



Four other Central European manuscripts antedating the Krakw codex

(Ol213, Pr299, Pr322, Kr129) incorporate a fuller version of the same
conclusion but place it out of context, after the Descensus.
How those ancient fragments jumped over a millennium to resurface
in the fifteenth and early sixteenth-century manuscripts remains a matter of speculation. At least some of them may have been inspired by a
now-lost Latin manuscript, perhaps in Northern Italy, and the Krakw
versions affinity to the Vienna palimpsest lends considerable support
to this hypothesis. However, the possibility that at least some of the
archaic features were taken directly from a Greek source in the later
Middle Ages cannot be ruled out.

If the horizontal transmission and the possibility of repeated translations from Greek, layered over the evolving text, seriously complicate
the textual history of the EN, the frequency of deliberate rewriting, and
especially rewriting that amplified it with additional material, testifies
to little or no concern for preserving the received work intact. Amplification could affect either the core text or its peripheries. The former
can be shown in a twelfth-century manuscript of Italian origin (Ox220)
that preserves a fairly accurate copy of LatA, but with one exception :
its redactor-scribe greatly amplified the section in which witnesses
come before Pilate to testify about Jesus miracles, adding witnesses to
the miracle at Cana in Galilee and three miracles in Capharnaum. This
amplification even found its way into Early Modern printings of the
EN.32 Another example is offered by a pair of manuscripts, one from
the fourteenth century (Pa279) and another from the fifteenth (Pa273),
both of which rearrange and interpolate extensive portions of the text.
After the usual prologue, the redactor interpolated accounts of the Jewish council, of Satan entering Judas, of the Last Supper, and of Jesus
arrest, and only then did he return to the usual accusations against Jesus
brought before Pilate. He also rearranged the entire trial section : he has
Pilate send the cursor to fetch Jesus only after an extended discussion
with the Jews, and he inserts Pilates Wifes dream just before Pilates
offer to release either Jesus or Barabbas.
The scope of the EN fluctuated also through equally intentional
expansion and contraction of its peripheries, its textual boundaries. Still
before the ninth century, the apocryphon acquired a number of peripheral accretions, including a new title, the Descensus, and Pilates letter.
By the later Middle Ages, the boundaries of the apocryphon were easily
adjustable. At one end of the spectrum, the EN was transmitted without
the account of the trial before Pilate, beginning only with the story of

32. See IZYDORCZYK, The Earliest Printed Versions, p. 129-131.



Joseph of Arimathea.33 At the other, it was amplified with prefatory

texts and a range of epilogues.34 A frequent preface in manuscripts of
British provenance, for example, is a pair of satellite texts, one attributed to Gregory of Tours and another to St. Augustine, which were
probably viewed as patristic recommendations for the EN. More typically, the text continued to expand beyond the Descensus and Pilates
letter. A common epilogue was the so-called Cura sanitatis Tiberii,
which recounts the healing of Tiberius by an image of Christ.35 Many
scribes and readers considered the Cura as an integral part of the EN,
judging by its fusion with the main body of the apocryphon and by the
placement of the ENs colophon after the Cura. Such inclusive perception of the EN, Pilates letter and the Cura is also reflected in the 1473
editio princeps published under the single title Euangelium Nichodemi
by an Augsburg printer, Gnther Zainer.36 Yet another extension of the
EN is the Somnium Neronis, a thematically diverse group of texts that
had no autonomous existence in the Middle Ages (its title is a modern
invention, based on its first thematic unit) but was always subsumed
by the EN, whose colophons typically marked its closure.37 Prologues,
epilogues and satellites like these problematize the beginning, the conclusion, and the scope of the Evangelium, all of which remained highly
unstable in the medieval age of manuscripts. The need to delimit and
title these various accretions seems to be a characteristic tendency in
the modern age of print.
The delimited and fixed form of the EN, as we know it today, is a
product of five centuries of printing and scholarship, probably in this
order. In fact, most of its early modern editors knew it primarily from
print : the edition of Guilielmus Montanus (1538), itself derived from
a manuscript, was reused by Vivantius Gaultherot (1545) ; Montanus or
Gaultherot was adapted by Johann Basilius Herold (1555), whose text
was re-issued by Johann Jacob Grynaeus (1569) ; Herold or Grynaeus
formed the basis of the edition by Johann Albert Fabricius, which
influenced Johann Carl Thilo and Constantin von Tischendorf.38 While
33. See, for example, the short versions in Ox238 and in Pa259 mentioned
above ; another abridged version was edited by David J. LEWIS, A Short Latin
Gospel of Nicodemus Written in Ireland, Peritia 5, 1986, p. 262-275.
34. For a fuller discussion of those satellite texts, see IZYDORCZYK, The
Medieval Gospel of Nicodemus, p. 55-68.
35. Edited by E. VON DOBSCHTZ, Christusbilder. Untersuchungen zur christlicher Legende, Leipzig, J. C. Hinrichs, 1899, p. 209-214, 157**-203**.
36. The editio princeps has been reprinted by Achim MASSER and Max SILBER,
(eds.), Das Evangelium Nicodemi in sptmittelalterlicher deutscher Prosa. Texte,
Germanische Bibliothek, 4th Series, Texte und Kommentar, Heidelberg, C. Winter,
1987, p. 448467.
37. Edited by E. von DOBSCHTZ, A Collection of Old Latin Bible Quotations : Somnium Neronis, The Journal of Theological Studies 16, 1915, p. 1-27.
38. See IZYDORCZYK, The Unfamiliar Evangelium Nicodemi and ID., The
Earliest Printed Versions.



Fabricius, Thilo, and Tischendorf had recourse to some manuscripts,

their sense of what the EN entailed and what it contained, of its shape
and scope, had been largely formed by its print tradition, which they
did their best further to enhance.

Editorial consequences
As the above discussion illustrates, the Evangelium Nicodemi did not
have a fixed, neatly delimited form in the Middle Ages ; its textual history is more complex and more involved than a traditional model of
imperfect scribal transmission may account for. Its text did not simply
deteriorate from the pristine, authorial form through a chain of successive, faulty copies ; rather, it was deliberately improved, altered, and
adapted on numerous occasions. In fact, if the Vienna palimpsest is
any indication, the original Latin translation of the Acts of Pilate was a
rather rudimentary affair. Like many early Christian translations, it was
very literal, comme dcalque du grec,39 and reproduced the wording
and syntax of the Greek source ; it also quickly acquired a number of
omissions and corruptions. The copies that circulated between the sixth
and the eighth centuries could not have been much better, judging by
some shared flaws of the earliest, ninth-century manuscripts. However,
starting in that same century the quality of syntax, morphology, and
lexis began to improve. There is a distinct possibility that the original
Latin translation was a much cruder text than its medieval descendants.
The extent and nature of scribal engagement with the EN complicates the process and the outcome of a detailed reconstruction of its
stemma. Although some paths of textual descent are clearly visible,
others remain fuzzy and open to a variety of hypotheses. Reversing
the chain of scribal changes to reach the ultimate source of the tradition, the Ur-text, might be an intellectually stimulating but, ultimately,
largely speculative undertaking, especially in view of multi-exemplar
revisions and repeated translations. Since no manuscripts have survived
from the period between the fifth and the ninth centuries, we can only
speculate about the ENs earliest textual history, working backwards
from the extant texts. We know that it went through various intermediate stages responsible for the presence of the Descensus, for the title
Gesta Salvatoris, and so on, but who, where, when, why, in what
order, and for exactly what purpose introduced those changes remains
largely guesswork. Some copies from the various early stages may have
still been around in the later Middle Ages ; for instance, LatB, although
attested only from the twelfth century, preserves certain features of the
Vienna palimpsest absent from all pre-twelfth century manuscripts ;
similarly, Kr127 shows readings coinciding with those in the Vienna
palimpsest but not preserved anywhere else in the Latin tradition. How39. DESPINEUX, Une version latine palimpseste, p. 180.



ever, re-building an ancient textual form from such scattered fragments

of uncertain provenance (after all, some of them may have been reintroduced directly from Greek) cannot lead to reliable insights into
the apocryphons past and cannot offer an adequate representation of
its protean text. The usefulness of the Greek texts is likewise limited
because they are relatively late and are themselves highly revised. A
comparative study of all early Christian versions may, of course, lead
to hypotheses about the original form of individual readings or even of
short passages, but it is unlikely to reconstitute the majority of the text,
let alone its entirety.
In its long manuscript history, the EN does not seem to have ever
had a single, authoritative textual form. What was read in fifth-century
Italy was dramatically different from what reached ninth-century northern France, and what was read in fourteenth-century England was dramatically different from what was read in Bohemia at the same time.
During the millennium of its transmission in manuscripts, the EN was
read, copied, and re-shaped always in time- and place-bound versions.
While efforts to improve its sense and Latinity were plentiful, there
are no signs of concern about its original, authorial, or translatorial
form. The Latin work in its initial, pure, authoritative form did not
exist either as an idea or as an actual text, and it had no bearing on
the culture which sustained it for over a millennium. Rather, it seems
that textual pliability and plurality were essential dimensions of the EN
before print, and, for that reason, it should probably be conceptualized
differently than classical or even patristic works. To the extent that it
was shaped and reshaped by generations of scribes, it was a communal
creation, changing with the needs and predilections of its reading communities. The initial translatorial impulses that started the process of
its construction, its Ur-text(s), dissipated quickly and left few lasting
impressions of their original form.
That is why I think that no single edited and published text can adequately represent the apocryphal work so multifariously concretized or
capture its dimensions fluctuating over centuries from region to region.
To represent both the dynamism and openness of the apocryphon, one
has no choice but to represent several of its key textual performances,
the ones that most profoundly impacted various aspects of medieval
and post-medieval culture. Even then, with multiple edited texts, we
will not have captured the entire work, but only some of its most
salient textualizations, or some of its texts of reference. However, it
would be a step towards accommodating the medieval condition of the
apocryphon, with its susceptibility to perhaps even encouragement
of textual variance, and a step away from restraining that variance
with assumptions born in the age of print for literary texts of a different kind.
Of the many textual performances of the Evangelium Nicodemi preserved in manuscripts, the one that definitely deserves a new edition



is LatA. It was most frequently copied and represents the most widely
diffused form of the apocryphon. Its earliest, ninth- and tenth-century
copies suggest descent from a single exemplar, whose text may be
reconstructed with a degree of confidence. The earliest LatA texts are
by no means uniform, splitting into two families already in the ninth
century, one best represented by the Laon manuscript (La133) and the
other by the Orlans manuscript (Or215). Although both families provided the substrate for later medieval rewritings, it was the Laon type
that inspired a number of vernacular translations.
The other two major forms of the Latin EN that merit edition are
LatB and LatC. The former survives in two textual types, families
LatB1 and LatB2, which preserve reflexes of readings from two different Greek exemplars and from different portions of the Vienna palimpsest. Both should be made available to textual scholars for further
study, not only because of the varying degrees to which they may be
reflecting the early text but also because, though few in number, they
were translated into European vernaculars. LatB2 also stands behind
the Bohemian redaction, which was used by Czech and German translators, and which, in its German, Czech, and Polish printed versions,
continued to attract readers well into the nineteenth century. That redaction represents an important part of the ENs heritage and should likewise be made available in a modern edition.
LatC represents one of the earliest transformations of the LatA
archetype, and it survives in a ninth-century manuscript (Ba012). It
was probably carried out on the Iberian Peninsula, with which most
of its manuscripts are connected. In its pure form, it does not seem
to have had a wide cultural impact, but it provided material for an
influential hybrid version, the Troyes redaction. The latter is extant in
several manuscripts scattered across Europe, which gave rise to numerous vernacular translations, especially into French and English, still in
print in the late eighteenth century.
These six textualizations of the EN would not, of course, exhaust
the textual forms the apocryphon assumed over the centuries. However,
they would reflect its time- and place-bound diversity and enable both
textual and cultural scholars to pursue and elucidate the many murky
patches in its fifteen-hundred year history.
Re-editing early Christian or medieval apocrypha is a precarious,
perhaps even perilous business. An apocryphons print tradition has
already established a sense of familiarity with the work and a broad
consensus as to its textual shape, scope, and purpose. Accepting the
assumptions underpinning that familiarity, one risks reproducing the
biases and limitations of the earlier editions ; rejecting them, one enters
the unchartered territory that may leave one at the mercy of scribal
and critical elements. And yet it is our responsibility as scholars to
explore the unexplored and to draw maps where there were none before,
even if it means contending with scribal whirlwinds and critical storms.



Manuscript sigla

Location (city, library, ms)



Mnchen, Bayerische Staatsbibliothek, Cgm 276

Barcelona, Archivo de la Corona de Aragn, Ripoll 106
s. IX
Berkeley, University of California, Bancroft Library, UCB
s. XII
Berlin, Staatsbibliothek Preusischer Kulturbesitz, Theol.
lat. oct. 157
s. IX
Brno, Sttn vdeck knihovna, Mk 79
s. XV
Paris, Bibliothque Nationale, Gr. 770
Cambridge, Corpus Christi College, 288
Grenoble, Bibliothque Municipale, 470 (Cat. 278)
s. XII
Andros, Moni Hagias, 46
Krakw, Biblioteka Jagielloska, 1509
s. XV
Krakw, Biblioteka Jagielloska, 2724
s. XV
Krakw, Polska Akademia Umiejtnosci 1713
s. XV
Laon, Bibliothque Municipale, 265
s. IX
London, British Library, Royal 1 E. IX
s. XIV
Montpellier, Bibliothque Interuniversitaire, Sec. Medicine 503
s. XIV
Mnchen, Bayerische Staatsbibliothek, Clm 642
s. XI
Mnchen, Bayerische Staatsbibliothek, Clm 11403
s. XV
Mnchen, Bayerische Staatsbibliothek, Clm 19105
s. X
Mnchen, Bayerische Staatsbibliothek, Clm 28168
s. XIV
Napoli, Biblioteca Nazionale Vitt. Em. III, VII.G.15
s. XV
Olomouc, Kapituln knihovna (at Sttn oblastn archv v
Opav), CO 407, s. XV





Orlans, Bibliothque Municipale, 341 (289)

s. IX
Oxford, Bodleian Library, Addit.A 367
Oxford, Bodleian Library, Canon. Pat. Lat. 117
s. XV
Oxford, Bodleian Library, Rawlinson D. 1236
Paris, Bibliothque Nationale (BN), Lat. 1652
s. XV
Paris, Bibliothque Nationale, Lat. 3784
s. X
Paris, Bibliothque Nationale, Lat. 5559
s. XV
Paris, Bibliothque Nationale, Lat. 5559A
s. XIV
Paris, Bibliothque Nationale, Lat. 10586
s. XIV
Paris, Bibliothque Nationale, Lat. 14864
s. XII
Paris, Bibliothque Nationale, N.a.lat. 503
s. XIV
Praha, Knihovna metropolitn kapituly, N. LIV
s. XV
Praha, Sttn knhovna, XX.A.7
Troyes, Bibliothque Municipale, 1636
s. XII
Vaticano, Citt del, Biblioteca Apostolica VaticanaReg.
Lat 1037
s. XV
Washington, D.C., Library of Congress 95 (Faye and
Bond 114)
s. XVI

Universiteit Antwerpen


in honorem
magistri amicique
This paper discusses the transmission of the apocrypha concerning
the birth and youth of Mary in the Latin West, to point out differences
and resemblances with the transmission of the Virtutes apostolorum. It
focuses especially on the Gospel of Pseudo-Matthew. Concerning the
relation between text transmission and editorial method, it argues that
the traditional presentation of the edition published in SA 9 was the
appropriate ratio edendi for a text that had an intricate but recognizable transmission history.
La transmission des apocryphes latins sur la naissance et la jeunesse
de Marie prsente des diffrences et des correspondances par rapport
la transmission des Virtutes apostolorum. Les spcificits de la tradition textuelle dterminent le choix de la mthode ditoriale. Cet article
montre lexemple du Pseudo-Matthieu publi dans SA 9 comment une
mthode quon qualifierait aujourdhui de traditionelle tait la ratio
edendi qui convenait le mieux un texte qui a connu une transmission
complexe certes, mais dont les mandres peuvent tre retracs.
This paper discusses some editions of Latin apocrypha concerning the birth and youth of Mary which were produced in the traditional way, in a time when New Philology and eLaborate did not
yet exist.1 An outline of the salient characteristics of the transmission
of these Marian apocrypha will reveal the differences and similarities
between them and the Virtutes apostolorum. I argue that the traditional
approach was the appropriate ratio edendi for the former, which have
an intricate but discernible transmission history. I begin by presenting
1. I thank Els Rose for having invited me to participate at the Utrecht
Expert Meeting on The Dynamics of Apocryphal Traditions in Medieval Religious Culture, and Eleanor Roach for reviewing the English of this written version of my contribution. For eLaborate, see the article by Mariken Teeuwen
in this journal.


Apocrypha 23, 2012, p. 117-140



the Latin apocryphal texts concerning Mary in the early Middle Ages,
and then focus particularly on the most influential of them, the Gospel
of Pseudo-Matthew.

Latin apocrypha concerning Mary in the early middle ages

Works mentioned in the Gelasian Decree
The list of apocryphal books mentioned in the fifth chapter of the
so-called Decretum Gelasianum contains four works connected with the
Virgin Mary :
Evangelium nomine Iacobi minoris, apocryphum
Liber de infantia Salvatoris, apocryphus
Liber de nativitate Salvatoris et de Maria vel obstetrice, apocryphus
Liber qui appellatur Transitus sanctae Mariae, apocryphus.2
We may assume that these works have been circulating in the Latin
West by the first half of the sixth century,3 but the extant Latin manuscripts are all considerably later.
Evangelium nomine Iacobi minoris, apocryphum
The Protevangelium of James (CANT 50) recounts the story of
Joachim and Anna, the birth and childhood of their daughter Mary, the
selection of Joseph as Marys ward (rather than husband), the birth of
Jesus with the miraculous catalepsy of all creation witnessed by Joseph,
the role of the Hebrew midwife who diagnosed the virginal birth and
the role of Salome, who at first did not believe in the virginal birth and
was punished for her unbelief, but was ultimately healed ; finally the
murder of the priest Zacharias, the father of John the Baptist, who was
slain at the altar in the temple by order of Herod.
The Greek Protevangelium dates to the second half of the second
century ; the oldest written form of it is the papyrus Bodmer V, from
2. Ernst VON DOBSCHTZ (ed.), Das Decretum Gelasianum de libris recipiendis et non recipiendis (Texte und Untersuchungen 38/4), Leipzig, 1912, p. 50,
l. 271 ; 51, l. 279-280 ; 53, l. 296. This list is considered to be the work of a
private person, probably from the South of Gaul. The two versions of the list
edited by von Dobschtz in appendix I and II reduce the second and third titles
to one : Liber de infantia saluatoris et de Maria uel obstetrice apocrifa in the
pseudo-isidorian compilation De numeris (VON DOBSCHTZ, p. 73, l. 205-6 ; CPL
1193, a Hiberno-Latin composition of the eighth century) and Liber infantia salvatoris et de Maria vel obstetrice eius in the excerpt in the Collectio
Herovalliana (VON DOBSCHTZ, p. 83, l. 88-89).
3. The Decretum Gelasianum is generally dated to the beginning of the
sixth century ; other, more recent datings, however, have been suggested. See
Els ROSE, Ritual Memory. The Apocryphal Acts and Liturgical Commemoration
in the Early Medieval West (c. 500-1215) (Mittellateinische Studien und Texte
4), Leiden/Boston, 2009, p. 25 n. 9.



the first half of the fourth century.4 At least two different Latin translations of the Protevangelium circulated in the early Middle Ages.5 The
oldest extant copy of these translations is a fragment from the beginning of the ninth century,6 the famous manuscript Montpellier, Bibliothque Universitaire de Mdecine, H 55.7
Liber de infantia Salvatoris, apocryphus
The Book on the Infancy of the Saviour mentioned in the Gelasian Decree is most probably the collection of miracles, supposedly
performed by the child Jesus between the ages of five and twelve, and
known as the Infancy Gospel of Thomas,8 though the original title
may have been Paidika Iesou.9 The latest editor of the Greek version of
these childhood tales, Tony Burke, described them as one of the most
neglected texts of Christian Apocrypha (p. 45) ; his recently published
edition and preliminary study will, no doubt, contribute significantly to
changing this situation.
For this text there is no great hiatus between the alleged time of
origin of the tales and the Latin manuscripts. In this case, the earliest
extant manuscript of the whole collection is a Latin palimpsest from
the fifth century, Vienna NB Cod. 563, containing not only chapters

4. mile DE STRYCKER (ed.), La forme la plus ancienne du Protvangile de

Jacques. Recherches sur le papyrus Bodmer 5 (Subsidia Hagiographica 33),
Bruxelles, 1961.
5. Jan GIJSEL, Het Protevangelium Iacobi in het Latijn, Lantiquit classique 50/1-2, 1981, p. 351-366 ; Jean-Daniel KAESTLI, Le Protvangile de
Jacques en latin. tat de la question et perspectives nouvelles, Revue dhistoire
des textes 26, 1996, p. 41-102.
6. Bernhard BISCHOFF, Katalog der festlndischen Handschriften des neunten Jahrhunderts (mit Ausnahme der wisigotischen), Teil I. Aachen Lambach, Teil II. Laon - Paderborn, Wiesbaden, 1998, 2004, nr 2822 : Lothringen
oder Nordburgund, wahrscheinlich unter Mitwirkung eines Metzer Schreibers, IX. Jh., Anfang. See also the rsum of the findings on MS H 55, presented by Martin HEINZELMANN in the context of the Ateliers trilatraux Villa
Vigoni on Merovingian texts and manuscripts : http ://lamop.univ-paris1.fr/spip.
php?rubrique134 (last accessed 2012.03.08).
7. Jean-Daniel KAESTLI, Le Protvangile de Jacques en latin, p. 44-45,
68-72 : M2 (PJ 8-25 : witness of the first, literal translation) and M1 (PJ 1-7 :
fragment of the second, amplified translation).
8. Tony BURKE (ed.), De infantia Iesu euangelium Thomae graece, CCSA
17, Turnhout, 2010, p. 8.
9. CANT 57. Sever J. VOICU, Verso il testo primitivo dei
Racconti dellinfanzia del Signore Ges, Apocrypha 9, 1998,
p. 7-95 : 55-59 ; Frdric AMSLER, Les Paidika Iesou, un nouveau tmoin de la
rencontre entre judasme et christianisme Antioche au IVe sicle ?, in Claire
CLIVAZ, Andreas DETTWILER, Luc DEVILLERS and Enrico NORELLI with the assistance of Benjamin BERTHO (eds), Infancy Gospels. Stories and Identities, Tbingen, 2011, p. 433-458 : 434.



1-16 of the Evangelium Nicodemi,10 but also fragments of the so-called

short recension of the childhood tales.11 It is generally assumed that
the first collection of these stories was written in Greek12 in the second
half of the second century, or the beginning of the third century.13 The
later manuscript tradition, on the other hand, is extremely complex and
there is such a variety of versions, expansions, interpolations, rewritings, and rearrangements, that it is difficult to determine and classify
the exact content of an individual witness to this collection of tales of
Jesus, the infant miracle worker and teacher.14
Two characteristics are specific to the transmission of the Latin texts,
which will be edited by Sever Voicu.15 First, between the early Vienna
palimpsest and the next manuscript there is a gap of seven centuries ;16
most of the manuscripts are even later and date from the fourteenth and
fifteenth centuries. These collections represent two separate translations
of two different Greek versions (Gs and Gd).17 Secondly, the trans10. Number 393 in Zbigniew IZYDORCZYK, Manuscripts of the Evangelium
Nicodemi. A Census (Subsidia Mediaevalia 21), Toronto, 1994, p. 192-193.
11. Guy PHILIPPART, Fragments palimpsestes latins du Vindobonensis 563
(Ve s. ?) : vangile selon S. Matthieu, vangile de Nicodme, vangile de lenfance selon Thomas, Analecta bollandiana 90, 1972, p. 391-411 ; Tony BURKE,
CCSA 17, p. 145-146 : fragments from chapters 2, 4-5, 7, 8-9, 14, 19 of the
Greek recension A.
12. Tony Burke (CCSA 17, p. 174-188) adheres to this opinion, concluding, after a thorough reassessment of the question : Given the evidence, Greek
composition is a far more economical solution to the problem.
13. CCSA 17, p. 201-205 : 205. The cumulative weight of the evidence
indeed suggests that IGT was composed in the second or perhaps third century.
Its use of Luke provides the text with a terminus a quo of around 90 C.E. and
the few internal indications of its antiquity point to a time of composition that
allows for its possible use by Irenaeus and the Epistula Apostolorum. See also
Sever VOICU, Ways to Survival for the Infancy Apocrypha in Infancy Gospels.
Stories and Identities, p. 401-417 : 405. As for the provenance of the text, Tony
Burke leaves this question unanswered, but considers Antioch or Asia Minor to
be acceptable places of origin (p. 206-212). Frdric Amsler in his contribution
cited in footnote 9 above argues for a fourth-century Antiochian origin.
14. Tony Burke calls the three versions of the story of Jesus and the Teacher
the backbone of the IGT (p. 286).
15. For a discussion of the Latin tradition see Sever VOICU, La tradition
latine des Paidika, Bulletin de lAlac 14, 2004, p. 13-21, and Tony BURKE,
CCSA 17, p. 144-160.
16. MS Paris, BnF lat. 1772 (siglum A4a2 in the CCSA edition of PseudoMatthew by J. Gijsel) dating from the beginning of the twelfth century.
17. The Vienna palimpsest (LV) and the Pars Altera of the Gospel of
Pseudo-Matthew (LM : ca. 40 manuscripts ; ed. TISCHENDORF, Evangelia apocrypha, 2nd rev. ed. 1876, p. 93-112) ultimately represent the same early translation, but the Pars Altera is a much-extended and interpolated form of it. A second,
medieval translation, is the basis of the version called the Evangelium Thomae
(LT : provisional list of manuscripts in CCSA 17, p. 151-153 ; ed. TISCHENDORF,
Evangelia apocrypha, p. 164-180). The oldest manuscript of this version appears



mission of both types is connected to that of the Gospel of PseudoMatthew,18 about which much more will be said in the discussion of
editorial methods.
Liber de nativitate Salvatoris et de Maria vel obstetrice,
The title of the third item in the Gelasian list, the Apocryphal Book
on the Birth of the Saviour and on Mary and the Midwife, is a general
one. The mention of the midwife suggests a nativity story similar to the
Protevangelium, possibly a predecessor of the Gospel of Pseudo-Matthew. But the question of the origin of the theme of the midwife, which
we automatically link to the Protevangelium, has been reassessed by
Enrico Norelli through comparison with early Christian testimonia on
the virginal birth of Jesus which all deny the need of a midwife.19 Concurrently, work on the edition of the different versions of the Infancy
Compilation20 has led Jean-Daniel Kaestli to the conclusion that the
Book on the Birth of the Saviour and on Mary and the Midwife is in
fact the New Source21 of the Infancy Compilation, which gives an
to be Vat. Reg. lat. 648, from Rheims, Saint-Rmi (see Sever Voicu, La tradition
latine des Paidika, p. 17), but the date of it is uncertain : s. XII according to Jan
Gijsel (CCSA 9, p. 151), Zbigniew Izydorczyk (Manuscripts of the Evangelium
Nicodemi, p. 185) and Sever Voicu ; s. XII-XIII according to Albert Poncelet and
Franois Dolbeau (De vita et obitu prophetarum. Une traduction mdiolatine
des vies grecques des prophtes, Revue bndictine 100, 1990, p. 507-531 :
512, cited in CCSA 9, p. 152 n. 1), followed by Simon C. Mimouni (Les Vies
de la Vierge. tat de la question, Apocrypha 5, 1994, p. 211-248 : 227 n. 69).
In his unpublished PhD dissertation J. Gijsel described the hand of Vat. Reg.
lat. 648 as a Northern French subcarolina from the end of the twelfth century.
18. See also Sever VOICU, Ways to Survival for the Infancy Apocrypha,
p. 412.
19. Enrico NORELLI, Avant le canonique et lapocryphe : aux origines des
rcits de la naissance de Jsus, Revue de thologie et de philosophie 126, 1994,
p. 305-324 ; ID., Les formes les plus anciennes des noncs sur la naissance
de Jsus par une Vierge in J. LONGRE (ed.), Marie et la Sainte Famille. Rcits
apocryphes chrtiens II (tudes mariales), Paris, 2006, p. 25-44 ; ID., Marie
des apocryphes. Enqute sur la mre de Jsus dans le christianisme antique
(Christianismes antiques 1), Geneva, 2009, p. 70-78. For his latest views on
the question of the early traditions of the birth of Jesus, see his contribution
Les plus anciennes traditions sur la naissance de Jsus et leur rapport avec les
testimonia, in Infancy Gospels. Stories and Identities, p. 47-66.
20. Catalogued by Maurice Geerard as CANT 53. Martin MCNAMARA e.a.
(eds), Apocrypha Hiberniae. I. Evangelia infantiae (CCSA 13-14), Turnhout,
2001. See also below, p. 138-139.
21. The first editor, Montague Rhodes JAMES, Latin Infancy Gospels. A new
text with a parallel version from Irish, Cambridge, 1927, p. xv, coined the
term New Source to identify the hitherto unknown text used by the Infancy
Compilation for the episode of the birth of Jesus ; Martin McNamara and JeanDaniel Kaestli use the term Special Source (CCSA 13, p. 64).



alternative version of the birth of Jesus. It relates in greater detail first

the journey of Mary and Joseph and his sons (with a special role for
Simeon) to Bethlehem, where they pay the obligatory Roman tax and
search for a lodging place ; it continues with the providential arrival of
one midwife who witnesses the miracle of Christs birth Christ does
not come forth from the womb of Mary but out of the gradual concentration of light from on high into a child ; then there is the visit of the
shepherds and their conversation with Joseph ; this is followed by the
visit of the Magi, who know the true identity of the child from their
older prophetic books and disclose it to Joseph. Joseph is the central
character in this narrative, he is described as a craftsman living in poor
circumstances and returning to his native Bethlehem. He is the one to
whom the mysteries of the birth are disclosed by both the midwife and
the Magi.22
In an illuminating paper published in 2011, Jean-Daniel Kaestli23
gives a comprehensive presentation of this apocryphal narrative, laying
special emphasis on the scene with the midwife. Extending the research
of Brent Landau24 on the relationship between this Book on the Birth
of the Saviour and the Syriac narrative on the Revelation of the Magi
preserved in the Chronicle of Zuqnin, Kaestli convincingly concludes
first, that this Book on the Birth of the Saviour (as mentioned in the
Gelasian Decree) is an early Christian writing, independent of the Protevangelium of James, at least as old as the latter, and, more importantly, that it goes back to the same tradition of the coming into the
world of Christ as the Light, that is also found in the Revelation of the
Magi (possibly of Edessenian origin).25
22. Jean-Daniel KAESTLI and Martin MCNAMARA, CCSA 13, p. 64-134 : The
Irish Infancy Narratives and their relationship with Latin sources.
23. Jean-Daniel KAESTLI, Mapping an Unexplored Second Century Apocryphal Gospel : the Liber de Nativitate Salvatoris (CANT 53), in Infancy Gospels. Stories and Identities, p. 506-559. See also his earlier article Recherches nouvelles sur les vangiles latins de lenfance de M.R. James et sur un
rcit apocryphe mal connu de la naissance de Jsus, tudes thologiques et
religieuses 72, 1997, p. 219-233.
24. Brent LANDAU, The Sages and the Star-Child. An Introduction to the Revelation of the Magi, an Ancient Christian Apocryphon, unpublished dissertation
Harvard Divinity School, Cambridge MA, 2008, directed by Franois Bovon ;
ID., The Revelation of the Magi in the Chronicle of Zuqnin. The Magi from
the East in the Ancient Christian Imagination, Apocrypha 19, 2008, p. 182201 ; ID., Revelation of the Magi. The Lost Tale of the Wise Mens Journey to
Bethlehem, New York, 2010.
25. A trace of this apocryphon is extant in the commentary on Matthew
by the mid-ninth century Irish writer Sedulius Scottus (ed. B. LFSTEDT, 1989),
who cites the dialogue between Joseph and his son Simon upon the arrival of
the Magi (CCSA 13, p. 120-126). Elements of the story of the Magi according
to this apocryphon are present in the fourteenth-century Historia trium regum
by the Carmelite John of Hildesheim (ca. 1310-1375). There is as yet no critical



The textual tradition of this apocryphon is more complex than that

of the preceding ones. The original language of it is unknown (Greek
or Syriac ?), and the story does not survive as an independent text but
has to be reconstructed through Irish and Latin medieval adaptations.26
None of the manuscripts of the adaptations is older than the thirteenth
century.27 Although we can no longer call this narrative entirely new
as M.R. James did in 1927, we are still far from knowing its full and
exact content.
Liber qui appellatur Transitus sanctae Mariae, apocryphus
The fourth text mentioned in the list of the Decretum Gelasianum
concerns the story of Marys death and the fate of her body and soul,28
that is, her dormition and/or assumption. This is the most complicated
of Marian traditions ; the dossier contains more than sixty texts distributed over eight linguistic areas (CANT 100-177). The latest progress
in this field is being made by Simon Claude Mimouni and Stephen
J. Shoemaker.29 The origins of this apocryphon are still open to discussion, but it seems safe to say that the literary tradition concerning
Marys death is more recent than that concerning her birth, and that the
earliest account of the end of Marys life (CANT 154) existed by the
fourth century at the latest.
edition of this very popular Legend of the Three Kings. At least 80 Latin manuscripts are extant, and there are medieval translations into German, English,
French and Dutch. See Sylvia C. HARRIS, Johannes von Hildesheim, Verfasserlexikon 4, Berlin, 1983, col. 642-647.
26. Actually compilations of two (Irish) or three (Latin) apocryphal narratives : Protevangelium and Liber de nativitate Salvatoris ; or Protevangelium,
Liber de nativitate Salvatoris and Pseudo-Matthew.
27. Two manuscripts date from the thirteenth century : Hereford, Cathedral
Library O.3.9 and Alenon, Bibliothque Municipale 17 ; they make up the socalled Hereford version of the J Compilation (edited by Jean-Daniel KAESTLI &
Martin MCNAMARA, in CCSA 14, p. 672-862).
28. Her sort final as Simon Minouni defines it in his Dormition et assomption de Marie. Histoire des traditions anciennes (Thologie historique 98),
Paris, 1995.
29. See Simon C. MINOUNI, Dormition et assomption, and his collected
essays on the subject, Les traditions anciennes sur la Dormition et lAssomption
de Marie. tudes littraires, historiques et doctrinales (Supplements to Vigiliae
Christianae 104), Leiden/Boston, 2011. For a different view on the question
of the dormition and assumption traditions, see the work of Stephen J. SHOEMAKER, Ancient Traditions of the Virgin Marys Dormition and Assumption,
Oxford, 2002, and his recent article From Mother of Mysteries to Mother of
the Church : The Institutionalization of the Dormition Apocrypha, Apocrypha
22, 2011, p. 11-47 (with further bibliography). See also Mary CLAYTON, The
Apocryphal Gospels of Mary in Anglo-Saxon England (Cambridge Studies in
Anglo-Saxon England 26), Cambridge, 1998, p. 24-100, and Enrico NORELLI,
Marie des apocryphes, p. 103-147.



If we concentrate on the Latin tradition, which is still under investigation, nine different texts emerge. They can be divided into two main
classes, the definition of which is still open to discussion. Some are
literal translations, others are free adaptations, but all ultimately derive
from Greek sources.30 Most widely attested are the two versions of the
Transitus of Pseudo-Melito of Sardes (CANT 111a-b).31 But other forms
have survived : the translation of the dormition of Pseudo-John (CANT
110, the most popular form in the Greek area, preserved in only one
Latin manuscript32) ; three anonymous Transitus called A (CANT 112),
W (CANT 114 ; 12 known manuscripts) and C (CANT 115) ; the Transitus of Pseudo-Joseph of Arimathea (CANT 116). The most recent text
is a ninth-century Latin translation of four homilies on the assumption
attributed to Cosmas Vestitor (between 730-850), which have not been
preserved in Greek (CANT 117).33 The first Latin writer to mention the
tradition of Marys death is Gregory of Tours. In his De gloria martyrum 1, 2, 4 (CANT 113), he summarizes the story of Marys dormition, her bodily assumption and reunion with her soul.
Manuscripts of these different Latin texts do not occur before the
end of the eighth or the beginning of the ninth century : the oldest are
Lyons 788, f. 34v + Paris BnF Baluze 270, f. 167-174 for the Transitus
30. See Simon C. MIMOUNI, Dormition et assomption, p. 257-299 (La tradition littraire latine). See also ID., Les traditions anciennes sur la Dormition et
lAssomption de Marie, p. XVI-XVII for a convenient summary, and Mary CLAYTON, The Apocryphal Gospels of Mary in Anglo-Saxon England, p. 66-100 for
an extensive survey of that Latin tradition.
31. There is no reliable list of Pseudo-Melito manuscripts. Monika HAIBACHREINISCH, Ein neuer Transitus Mariae des Pseudo-Melito, Rome 1962, mentions 20 manuscripts for CANT 111a (p. 30-32) and 18 manuscripts for CANT
111b (p. 55-59) ; see also Simon C. MIMOUNI, Dormition et assomption, p. 266
n. 32. Numerous additional Pseudo-Melito witnesses are mentioned in CCSA
9 (19) and CCSA 10 (6) ; other manuscripts may well be present in Zbigniew
IZYDORCZYK, Manuscripts of the Evangelium Nicodemi under Transitus Mariae.
See also the BHLms database (http ://bhlms.fltr.ucl.ac.be/), under BHL 53515352.
32. Florence, Bibl. Med. Laur., Santa Croce, Pl. XV d. 12, fol. 19-20v (miscellaneous codex from s. XIII ; this text is a later addition from s. XIV).
33. Recently re-edited by Arpd ORBN in Sermones in dormitionem Mariae.
Sermones Patrum Graecorum praesertim in Dormitionem Assumptionemque beatae Mariae virginis in latinum translati, ex codice Augiensi LXXX (saec. IX),
CCCM 154, Turnhout, 2000, p. 95-126. The codex Reichenau 80 (late ninth
century, Italy) in the Badische Landesbibliothek of Karlsruhe contains a unique
collection of Latin translations from thirteen Greek Marian homilies, primarily
on the assumption and mostly written in the eighth century : Andreas of Crete
( 740) ; Amphilochius of Iconium (ca 340/345 ca 403) ; Cosmas Vestitor
(between 730-850) ; Germanus of Constantinople ( 733), John of Damascus
( before 753) ; the fourteenth item in the collection is a Latin compilation of
elements from the preceding Greek dormition homilies, probably made by John
bishop of Arezzo in the second half of the ninth century.



W (s. VIIIex.),34 and Montpellier, Mdecine H 55 for the Transitus of

Pseudo-Melito, the same legendary which contains the earliest attestations of the Latin Protevangelium and Pseudo-Matthew.35
I do not intend to go into the details of this very complex tradition ;
I just want to draw attention to the variety of texts connected with
Marys death and assumption. The proliferation of the liturgical feast of
the Assumption of Mary in the Latin West created a need for reading
material for this feast.
Other pre-thirteenth-century Marian apocrypha
The four groups of Marian texts mentioned in the Gelasian Decree
were the building blocks for the medieval Latin apocrypha concerning
Mary. Authors took the texts as a starting point and used them in either
free adaptations or creative continuations or short summaries or more
or less faithful compilations.
Adaptations and compilations are a well-known phenomenon in the
later Middle Ages, but for the sake of clarity and comparison with the
Virtutes apostolorum I confine my discussion to texts prior to the thirteenth century : the Gospel of Pseudo-Matthew (CANT 51), the Little
Book on the Birth of Mary (CANT 52) and the Latin Infancy Compilation of James mentioned above.
As for the introduction of new Marian texts from the Greek, there is
only one example to mention for this period, and that is a Life of the
Virgin Mary. Lives of Mary constitute a literary genre that has a long
history in Greek and oriental languages ; in the West however, it did not
emerge before the twelfth century.36 At the end of the eighth century,
34. The Lyons and Baluze fragments are variously dated as s. VIIIex., s. VIII/
and s. IX. See also Bernhard BISCHOFF, Katalog der festlndischen Handschriften des neunten Jahrhunderts, nr 2583a ; and the description of the Lyons
manuscript, which is part of the digitalised catalogue of Merovingian and Carolingian manuscripts of the Bibliothque municipale de Lyon (http://florus.bmlyon.fr/index.php). For a description of the Baluze manuscript, see L. AUVRAY
& R. POUPARDIN, Catalogue des manuscrits de la collection Baluze, Paris, 1921,
p. 326-328 (Transitus dated erroneously s. IX-X) ; a digitalised reproduction of
Baluze 270 can be found in the Gallica database of the Bibliothque nationale
de France (http://gallica.bnf.fr).
35. See footnote 6 above.
36. CANT 90-96. For the Greek literary tradition of Virgins Lives prior to
the year 1000, see Simon C. MIMOUNI, Les Vies de la Vierge (repr. in Les
traditions anciennes sur la Dormition et lAssomption de Marie, p. 75-116).
The poem in Old French by Robert Wace, Conception Nostre Dame, written
before 1155, may be the first example of a Life of the Virgin in the West. In
my view, it is not a translation but an independent composition based on different Latin sources. See Rita BEYERS, La Conception Nostre Dame de Wace :
premier pome narratif sur la Vierge en ancien franais, in Werner VERBEKE
(ed.), Serta devota in memoriam Guillelmi Lourdaux. Pars posterior, Leuven,
1995, p. 359-400.



Epiphanius, monk of the monastery of Kallistratos in Constantinople,

composed a Life of the Virgin (CANT 91) ; this text was translated into
Latin by Pascalis Romanus during his stay in Constantinople between
1158 and 1169.37 Judging by the limited number of extant manuscripts,
his translation did not exercise great influence.38

Editorial methods : the case of Pseudo-Matthew

The Gospel of Pseudo-Matthew, the Little Book on the Birth of
Mary and the Latin Infancy Compilation of James have all appeared in
critical editions,39 which enables me, in the second part of this article,
to reflect on editorial methods, possibilities and limitations. I shall concentrate especially on Pseudo-Matthew, edited by Jan Gijsel in CCSA
9, because this is a case where we have to deal with four different
problems. It is a text of uncertain origin ; it had a broad geographic
distribution (about 200 manuscripts known and analysed) ; it presents
a great variety of textual forms and its content evolved so much that
it is fair to ask the question what do we mean by Pseudo-Matthew ?
Content and sources
The Gospel of Pseudo-Matthew has long been described as an adaptation of the Protevangelium of James, written in such a way as to
avoid the condemnation pronounced against the latter by the Church
Fathers and the Gelasian Decree. If this were the case, one wonders
why the Pseudo-Matthew retains the two aspects of the story which
gave most offence, i.e. the use of the authority of James the Less,
brother of Jesus, and the roles assigned to the anonymous midwife and
Salome. Since Jean-Daniel Kaestlis study of the Latin transmission
of the Protevangelium (published in 1996), which takes into account
the testimony of the Irish compilations,40 it is clear that the author of
Pseudo-Matthew used, as a basis for his rewriting, neither the Greek
text of the Protevangelium nor a literal Latin translation of it. Instead,
he must have had at his disposal a copy of an amplified Latin trans37. Edited from Oxford, Balliol College 227 (s. XIII-XIV) by Ezio Franceschini, Studi e note di filologia latina medievale (Pubblicazioni della Universit
Cattolica del Sacro Cuore, ser. 4, vol. 30), Milano, 1938, p. 109-128 : Il
di Epifanio nella versione latina medievale
di Pasquale Romano.
38. Only three complete manuscripts are known to date : Vat., Reg. lat. 648,
f. 50-57v ; Oxford, Balliol 227, f. 146v-151v ; Praha, Statni Knihovna SR,
XIV E 10, f. 61-67v. MS London, BL Royal 13 A XIV, f. 270v-272, is not
complete. See Franois DOLBEAU, De vita et obitu prophetarum, p. 511-513 ;
Simon C. MIMOUNI, Les Vies de la Vierge, p. 226 n. 69 and 228 n. 74.
39. CCSA 9, 1997 ; CCSA 10, 1997 ; CCSA 14, 2001.
40. I.e. the infancy narrative as contained in the Liber Flavus Fergusiorum
and in the Leabhar Breac, analysed and edited in CCSA 13, 2001.



lation, i.e. one where some scenes were expanded and others were
added. The most important additions are the description of Marys life
in the Temple and her refusal to marry the son of Abiathar the priest
(chap. 6-7). These episodes were not an invention by Pseudo-Matthew
For the events following the visit of the Magi, Pseudo-Matthew
developed a narrative different from that of the Protevangelium. He
describes the flight to Egypt up to the arrival of the Holy Family in
Egypt, which is characterised by a series of miracles. These are the
four well-known miracles of the wild animals worshipping Jesus
(chap. 18-19), the palm tree bending towards Mary to offer its fruit
(chap. 20-21), the shortening of the road so that thirty days are reduced
to one single day (chap. 22), and the collapse of the idols when the
Holy Family enters the temple of the Egyptian city of Sohennen or
Sotinen41 (chap. 23-24). These chronologically predate the miracle
stories associated with the sojourn of the Holy Family in Egypt, as well
as those connected with their stay at different places in the Holy Land,
which form the content of the Liber de infantia Salvatoris. A question
that remains to be answered is whether Pseudo-Matthew depends here
on only one other source ; and if so, can that source be identified ?
The studies by McNamara and Kaestli of the Irish compilations have
shown that the Irish version of these stories preserved in the Leabhar
Breac 126-129, 133-136, does not depend on the Latin version of
Pseudo-Matthew, but that both go back to a common source.42 In 2001,
the editors suggested that this common source might have something
to do with the Liber de nativitate Salvatoris as discussed on p. 119-121
above.43 It would be interesting to know if subsequent research on this
topic has led them to a more definite conclusion. A positive answer
would imply that Pseudo-Matthew knew the Liber de nativitate Salvatoris but used it selectively, leaving aside entirely the story of the
birth and the visit of the Magi, and only retaining the events on the
way to Egypt. Such an attitude towards his source may be hard for us
to accept, but let us not forget that the Gospel of Pseudo-Matthew is a
compilation and that picking and choosing is what a compiler does. We
note that this change in his source is reflected in the state of the textual
41. Possibly to be identified with the ancient city of Syene (now Assuan ;
see CCSA 13, p. 407 n. 173) or, more likely, with the city of Tanis in the northeastern Nile delta (Hebrew Soan ; now San el-Hagar ; see Jos M. CANAL, En
torno al Evangelio del Pseudo-Mateo, Marianum 60, 1998, p. 197-237 : 208).
42. CCSA 13, p. 60-64 ; edition p. 400, 402, 406, 408.
43. CCSA 13, p. 64 : A question that remains to be clarified is whether or
not this source has something to do with the Special Source used in the nativity, shepherds, and Magi narratives. See also p. 67 : There is a possibility,
however, that the original Special Source went beyond the advent and return
of the Magi to cover events of the journey to Egypt, and that traces of this may
still be found in the Leabhar Breac Infancy Narrative.



transmission : whereas in the first half (ch. 1-12, events up to the birth
of Jesus) the two Carolingian families of Pseudo-Matthew do not differ
very much from one another, it is obvious that, for the miracle stories
(ch. 16-24), we have two different versions, and in both we find eloquent proof of adaptations.
Origin and date
The origin and date of Pseudo-Matthew pose a problem that has
not yet been solved. The manuscript evidence does not allow us to go
back before the Carolingian period. The oldest source is, once again,
the famous Montpellier codex H 55 from the beginning of the ninth
century ;44 it preserves traces of an older, pre-Carolingian form of the
text. The direct transmission of Pseudo-Matthew begins in the second
quarter of the ninth century, with two manuscripts : London, BL Add.
Mss. 11880, from Regensburg about 830 (siglum A2a1 in the CCSA
edition), and Budapest, Clmae 316, from the Salzburg region, second
quarter of the ninth century (siglum A3a1).45 Their stemmatic position,
however, is not right at the top of A, the first of the two Carolingian
textual families. In fact, the Budapest manuscript belongs to a subgroup
(A3) which was already contaminated with a version of the other family (P).
All these observations led the editor, Jan Gijsel, to assume that there
must have been a bifurcation into two textual families at the beginning of the ninth century at the latest.46 A and P can clearly be distinguished : A remains closer to the original text than P, which is the
result of a revision. Oddly enough it is P that has preserved the original
prologue mentioning James the Less as the author, whereas in A this
prologue has been replaced by two fictitious letters purporting to prove
that Jerome was the author of the Latin translation of what was an
unknown Hebrew text by Matthew. This reconstruction is very plausible and still holds.
A fragmentary ninth-century letter suggests that the fictitious correspondence was composed by Paschasius Radbertus, the great Carolingian theologian from Corbie (b. probably before 790 and d. probably after 862), to accompany the story of the birth of Mary that was
repudiated by the Gelasian Decree.47 If this is true I am strongly
44. See footnote 6 above.
45. Bernhard BISCHOFF, Katalog der festlndischen Handschriften des neunten Jahrhunderts, nr 2364 and 759.
46. CCSA 9, p. 63 : la bifurcation de ces deux familles est placer au plus
tard au dbut du IXe sicle ou la fin du VIIIe.
47. The evidence consists of a fragmentary letter of Hincmar of Rheims
to Odo of Beauvais from 868-869, edited by Cyrille LAMBOT, Lhomlie du
Pseudo-Jrme sur lAssomption & lvangile de la Nativit de Marie daprs
une lettre indite dHincmar, Revue bndictine 46, 1934, p. 265-282. The
interpretation of C. Lambot is refuted by Maurice CAPPUYNS, Bulletin de tholo-



inclined to believe the accusation of forgery uttered by Paschasius

fellow monk, Ratramnus of Corbie it would mean that Paschasius
composed these two letters in his youth, at the very beginning of the
ninth century.48 There are scholars who cannot accept the idea that the
great mariological thinker Paschasius would have permitted himself to
be associated with a text of such poor quality and such disputable content as Pseudo-Matthew.49 But this is a matter of taste and sensitivity
as opposed to historical evidence.
Our difficulties increase when we try to go farther back in time.
As things now stand, there are no firm indications that allow us to
determine a date of composition or a place of origin. The recent new
insights into the transmission of the apocryphal material concerning
Mary and the birth of Jesus have actually complicated matters, because
references to such material by late antique and early medieval Latin
authors can no longer automatically be linked to Pseudo-Matthew. The
Decretum Gelasianum may represent a terminus post quem, since there
appears to be no mention of Pseudo-Matthew in the list.
What has commonly been regarded as the most characteristic passage of Pseudo-Matthew, Marys daily life in the Temple as a prefiguration of the monastic way of life, appears not to be an invention by
Pseudo-Matthew, but an adaptation of a scene already present in the
amplified translation from the Protevangelium. Modern scholarship generally accepts that this description reflects the influence of the Benedictine Rule and that, therefore, the text must be more recent than the
middle of the sixth century. But in my view a closer reading of the
chapter shows that the text depicts Mary as a young virgin living in
the Temple according to literary descriptions of female ascetic life. She
leads the life of a consecrated virgin and excels in all virtues, as she
had been portrayed by Ambrose in the second book of his treatise De

gie ancienne et mdivale 3, 1937-1940, nr 168, p. 83-84, and Rita BEYERS, De

Nativitate Mariae. Problmes dorigine, Revue de thologie et de philosophie
122, 1990, p. 171-188.
48. The date of the two oldest manuscripts of Pseudo-Matthew offers a terminus ante quem (about 830 A.D). According to Jozef VAN BANNING, Buchstabe
und Geist. Zur Rezeption der Exegese des Opus imperfectum in Matthaeum
im Mittelalter, Nijmegen, 2006 (expanded version of the authors 1992 habilitation at Innsbruck University), I, p. 147-152, Paschasius could not have written
anything as early as the first decade of the ninth century. [] man kann kaum
an ein so frhes Datum fr ein Werk von Paschasius denken. Abgesehen davon,
ob Paschasius damals schon so eine Kenntnis der Vter gehabt htte, dass er
eine solche Korrespondenz verfassen konnte, muss man doch sagen, dass wir
faktisch keine andere Werke von ihm von vor dem Jahre 823 haben. (p. 152)
49. As Rosamond MCKITTERICK, The Carolingians and the Written Word,
Cambridge, 1989, p. 203-204, stresses, Carolingian scholarship was much preoccupied with authority, orthodoxy and correctness.



virginibus (II, 6-19), which he wrote in 377. The regula Mary is

said to have invented for herself is not necessarily a monastic rule.50
Criteria such as language, literary quality, or style do not allow us to
draw precise conclusions about the time or place of origin of PseudoMatthew. In the oldest form (A), the passages from Scripture that are
quoted or alluded to follow the Vetus Latina ; in the revised form (P)
we see the influence of the Vulgate.51 Pseudo-Matthew has an easy narrative style, but this does not mean that the author wrote poor Latin.
Can one say, as is still claimed in a recent translation, that the author
was not a particularly gifted writer, hence the rough and occasionally slovenly character of the older A recension, in contrast to the later,
more refined P ?52 This statement, which echoes the judgement of others, including the editor Jan Gijsel himself,53 prompted me to go back
to the text and look for specific examples that could justify this evaluation. To be sure, there are examples of clumsy constructions with an
occasional anacoluthon or a nominativus absolutus54 (e.g. chap. 17, 1,
1-3 : Videns autem Herodes ... inflammatum est cor eius [cfr Ps. 72,
21] et misit per omnes uias...), but such lapses are not frequent. From
a lexicographical viewpoint, the text, especially the oldest version A,
contains some remarkable, uncommon words or rare uses. I draw attention to four cases.55
(1) inemissus (8, 5, 12)
A : Non erit iste sermo inemissus
P : Non erit iste sermo in fatigationem missus
When Marys companions laugh at her and call her regina uirginum, the angel rebukes them, saying : This word will not be spoken
in vain (A) / in derision (P), for you have made the truest prophecy.
The adjective inemissus seems to be a hapax,56 at least before the

50. Rita BEYERS, La Rgle de Marie : caractre littraire et inspiration

monastique, Apocrypha 22, 2011, p. 49-86.
51. CCSA 9, p. 90.
52. Bart D. EHRMAN & Zlatko PLEE, The Apocryphal Gospels. Texts and
Translations, Oxford, 2011, p. 75.
53. J. Gijsel qualifies the style of the A version as fruste, rempli de soudures brusques et de maladresses de toutes sortes (CCSA 9, p. 88).
54. See Peter STOTZ, Handbuch zur lateinischen Sprache des Mittelalters, 4.
Formenlehre, Syntax und Stilistik, Mnchen, 1998, p. 238, 3.1 : Der Nominativus absolutus.
55. I made full use of the unique and constantly expanding research possibilities of the Brepolis Latin databases Library of Latin Texts (LLT), Cross
Database Searchtool and Database of Latin Dictionaries (DLD), developed
by the Centre Traditio Litterarum Occidentalium under the direction of Paul
56. CCSA 9, p. 374, n. 2.



ninth century,57 but it is clear that the word must be understood as not
sent, not issued, unspoken, hence spoken in vain.
(2) readorare (16, 2, 10)
A : Illi autem readorauerunt infantem et cum omni gaudio per
alteram uiam ad suam sunt reuersi regionem.
P : Illi autem adorauerunt infantem cum omni gaudio et per aliam
uiam reuersi sunt in regionem suam.
Before the Magi return home by a different way, they worship the
infant again. A has readorauerunt infantem ; P adorauerunt infantem.
As far as I can discover, readorare is a hapax, but it is a rather obvious compound.
(3) exagilium dare (21, 1, 2)
A : Hoc exagilium (variant : priuilegium) do tibi, palma, ut unus ex
ramis tuis transferatur ab angelis meis et plantetur in paradiso patris
P : Hoc ex praecepto meo dico tibi, palma, ut unus ex ramis tuis
transferatur ab angelis meis et plantetur in paradiso patris mei.
The child Jesus rewards the palm tree by promising/ordering that
one of its branches will be taken by the angels and planted in paradise.
The noun exagilium seems to be a variant form for either exagella
(Du Cange and Thesaurus Linguae Latinae58) or exagellium (Blaise)59
or exagellia,60 perhaps derived from exagium, the act of weighing,
ponderation, weight, balance, hence : something balanced, a well-balanced part. The word is first attested in the Confessio Sancti Patricii
(fifth century)61 and means a bequest, a (spiritual) legacy, a kind of
last word : the Confessio being a kind of spiritual legacy left by Patrick

57. The LLT-database records only one other attestation of the word. It
occurs in the Latin translation of Pseudo-Dionysius Areopagita, epistula 9,
by Hilduin abbot of Saint-Denis (between 830-835), where it is used with
the noun disciplina : inemissi et inuisibilis multae disciplinae (for the Greek
(sc. ) . See
Philippe CHEVALLIER e.a. (eds), Dionysiaca. Recueil donnant lensemble des
traductions latines des ouvrages attribus au Denys de lAropage, I, Bruges,
1937, p. 635, col. 4.
58. TLL, s.v. exagella, col. 1145, l. 70-77.
59. Albert BLAISE, Dictionnaire latin-franais des auteurs chrtiens, Turnhout, 1967, as updated in the DLD.
60. Ludwig BIELER, Exagellia, American Journal of Philology 69/3, 1948,
p. 309-312.
61. The Confessio Sancti Patricii can easily be consulted on the internet on
the Hypertextstack Saint Patricks Confessio, a multilayered edition with texts,
different apparatus, commentaries, translations, manuscripts, editions, and other
images and audiomaterial, conceived and overseen by Anthony Harvey : www.



to his brethren.62 It is also found in Ennodius, Vita Epiphanii 19163 (ca

501-504) and in the Virtutes Iohannis, in the last speech of John to his
disciples, where he says : Quid igitur uobis pro exagillo relinquam ?64
What can I leave you as a legacy ? What good advice can I give
you ? The Regula Magistri, chap. 91 (s. VI1) uses the adjective exagiliarius : exagiliario (munus) titulo (derelinquat), under the title of
legitimate portion (transl. Bieler). Bieler remarks that the word always
occurs with the verb relinquere, and almost always in a metaphorical
sense. If we take the form exagilium to mean spiritual legacy, it fits
the context of the palm tree episode, but the word is sufficiently rare
to have caused copyists to introduce variants. In this particular case it
would have been useful if the critical apparatus of Pseudo-Matthew had
given the different spellings of this difficult word.
(4) aequitates (22, 1, 7) :
A : Adhuc eo loquente, ecce prospicientes uiderunt montes Aegypti
et aequitates eius.
P : Adhuc eo loquente, ecce prospicientes uidere coeperunt montes
Aegyptii et ciuitates eius.
In the context of the miracle of the shortening of the journey, the
Holy Family looks up when Jesus speaks and immediately sees the
mountains of Egypt and its plains. That is what should have been the
original reading, but most A manuscripts have equitatus, the cavalry,
and P has changed the wording into they started to see the mountains
of Egypt and its cities. The problem here seems to be the rare use of
aequitas in the sense of plain. I found no example of the combination of mons and aequitas, for mountain and plain, and the Thesaurus Linguae Latinae cites only one instance for aequitas in a geographical sense, in combination with loci, the evenness of the ground
or place or location ;65 it is more often used in the opposite way, the
iniquitas loci, the unevenness of the ground. This juxtaposition may

62. Cf. Anthony HARVEY & Jane POWER (eds), The Non-classical Lexicon
of Celtic Latinity, I. Letters A-H (Dictionary of Medieval Latin from Celtic
Sources 1) (CC MLD 2), Turnhout, 2005, p. 286, s.v. ( ?) exagaelliae (pl.).
63. MGH AA 7, 1885, p. 108, l. 14 : quasi exagellam (var. : exagelliam)
64. Chap. 9, l. 22, ed. Jean-Daniel KAESTLI, CCSA 2, 1983, p. 828 : pro
exagillo is corrected into ex illo by M (Paris, BnF lat. 3779, s. X-XI) and W
(Wolfenbttel, Herz. Aug. Bibl., Weissenburg MS 48, s. IX2/3). Els Rose informs
me that the manuscript tradition is more complex on this locus and that the
reading pro exagillo only seems to be attested in Wien NB 455 and Paris
BnF lat. 5563. The Greek text has : What else could I tell you ? (
;) (AJ chap. 107, l. 12 ; CCSA 1, p. 297.)
65. Bellum Hispanicum 29, 7 : interdum aequitas loci aduersarios efflagitabat ut tali condicione contenderent ad uictoriam.



have led to the meaning of aequitas tout court as plain. In the present state of our knowledge, this meaning is not otherwise attested.
For all their lexicographical interest, these examples do not contain
any clear chronological clues. As for style, it is interesting to compare
passages of the A and the P version to see precisely where the original
version might have caused difficulties. But while we can appreciate the
corrections or adaptations by P, in general the version of A is acceptable as well.
Finally, if we look for links with the liturgical context in which the
Pseudo-Matthew may have been used, results for a possible dating are
inconclusive. There is no evidence for the presence of the apocryphal
Virgin of Pseudo-Matthew in the liturgy prior to the beginning of the
manuscript tradition.66 In contrast to the apocryphal Apostles,67 the
apocryphal Mary of the infancy tradition has left no traces in liturgical
usage before the Carolingian period.68 The Latin West originally had
only one Marian feast, which could be celebrated on different days.
Gradually, the four major Marian feasts (Nativity, Annunciation, Purification, Assumption) found their way from the East into the Latin calendar in the course of the seventh and eighth centuries. The feast of
Marys birth was introduced in Rome by Pope Sergius I (687-701).
In liturgical contexts we find plenty of references to the main mariological facts about the birth of Jesus, with formulas stressing that Mary
conceived as a virgin, gave birth as a virgin and remained a virgin, but
specific apocryphal details are absent. Specialists of medieval liturgy
and specifically Marian liturgy have investigated the available sources.
Els Rose made some interesting observations in her analysis of the feast
of the Assumption in the Missale Gothicum.69 Though this source from
ca. 700 is an important example of how the Gallican liturgy was influ-

66. Manuscript Montpellier H 55, f. 94, refers in the title to the feast of the
Nativity of Mary quod est VI Idus Septembris. See also the Carolingian sermon
Inquirendum est for the same feast at use in Saint-Pre de Chartres, which
contains fragments of the second translation of the Protevangelium, ed. JeanDaniel KAESTLI, Le Protvangile de Jacques latin dans lhomlie Inquirendum
est pour la fte de la Nativit de Marie, Apocrypha 12, 2001, p. 99-153.
67. As Els Roses research on the relation between the Apocryphal Acts and
the liturgical commemoration of the Apostles has made abundly clear. See especially Els ROSE, Ritual Memory.
68. The most important feast in the ninth century was still the Assumptio,
for which much new homiletic material was produced. As Henri BARR, Textes
marials indits du Xe sicle, Marianum 27, 1965, p. 3-71, pointed out, we have
to wait until the tenth century to see sermons for the Nativity of Mary appear
in greater quantities.
69. Els ROSE (ed.), Missale Gothicum e codice Vaticano Reginensi latino
317 editum, CCSL 159D, Turnhout, 2005, p. 223-232.



enced by the apocrypha,70 there is no irrefutable indication of influence

upon it by the story of the nativity of Mary. And while the Assumptio
was in fact a general Marian feast, we cannot conclude from the texts
for the Assumption that the Gallican liturgy observed a separate celebration of the Birth of Mary.71
I did not initiate a detailed investigation of my own, but a reading
of hymns and homilies72 confirms that these texts are silent about narrative elements originating in the apocryphal tradition transmitted by
Pseudo-Matthew. The reason may be that the canonical texts offered
enough material for the liturgical celebrations of the great events in
Marys life up to the birth of Jesus, so it was not necessary to take
anything from this apocryphal source.
Different textual forms and their contents
Once the manuscript transmission of Pseudo-Matthew begins, we
can see what happened to the text and how it developed into different
versions. The A version is closest to the original text ; it shows only
traces of basic grammatical corrections.73 The P version on the other
hand is the result of a revision of the text, independent of A, and can
be situated around 800 A.D.
The two other versions, Q and R, enter the stemma at a later stage
of the tradition. Q derives from P and R derives from Q, which emerges
around the middle of the twelfth century. Q not only modifies the
wording of P, but also changes the presentation of the text : the original
prologue of P (Ego Iacobus) is replaced by the fictitious letters which
constitute the prologue of A, and, most notably, two texts are added to
the apocryphon, one at the beginning and one at the end. The first of
these is a short prose text explaining that the brothers of Jesus were
not by a former marriage of Joseph, but by a triple marriage of Marys
mother, Anna.74 This family tree, known as Trinubium Annae, may go

70. CCSL 159D, p. 232 : Capelle drew attention to prayers influenced by

Pseudo-Melito and by the anonymous Transitus W.
71. Els ROSE, Missale Gothicum, p. 229.
72. A look into the recently published edition of the Hymnodia Hispanica by Jos CASTRO SANCHEZ (CCSL 167, 2010) shows that the hymns used
in Spain from the fourth to the eighth century (with some younger texts up
to the eleventh century) contain disappointingly little information about Mary,
other than her titles, the virginal birth and her assumption into the glory of
heaven. See also J.J. SZVRFFY, Marianische Motivik der Hymnen. Ein Beitrag
zur Geschichte der marianischen Lyrik im Mittelalter, Leiden, 1985. For the
homiletic genre, see for instance the sermons by Ambrosius Autpertus (CCCM
27B, 1979).
73. E.g. in + ablative case corrected into in + accusative case, agreement in
tenses or modes. See CCSA 9, p. 88-89.
74. After the death of Joachim, Anna remarried twice and had two other
daughters, also called Mary. The first married Alphaeus and was the mother of



back to Haimo of Auxerre ( ca 855) ; it certainly circulated in different forms, both in prose and verse, and in varying degrees of complexity as to the names included in the genealogy, from the eleventh
century onwards.75 The version of the Q family integrates Elisabeth
in the genealogy, stating that Anna and Emeria, mother of Elisabeth,
were sisters : Anna et Emeria sorores erant. The second, more dramatic
change is the addition, at the end of the story, of the series of miracle
tales known as the Infancy Gospel of Thomas. The manuscript evidence
shows very clearly that this apocryphon occasionally followed PseudoMatthew as an independent text, but that gradually, and such was the
case in the Q family, the distinctions between the two texts disappeared
and they came to constitute one single narrative. The specific version
of the Infancy Gospel of Thomas that is transmitted with the Q family
(chap. 25-42) is known as the Pars Altera.76
The youngest family, R, came into existence about 1200. It is a further adaptation and rewriting of the Q text (Pseudo-Matthew and Pars
Altera), with its own peculiar characteristics, of which the presence of
subtitles throughout the text is the most striking.77
Ratio edendi
As Pseudo-Matthew is a relatively short text (A : 5766 words ; P :
5630 words) Gijsel was able to collate all the manuscripts (about 200),
and to situate every one of them in the general frame of the transmission ; the position of each manuscript is reflected in its siglum. Within
the four main forms of the text, he distinguished different subgroups
and further branches.
As the two Carolingian textual forms, A and P, both go back to the
archetype independently, the editor chose to print both texts synoptically, i.e. one beneath the other.78 The critical apparatus of A and P
gives the readings of the subgroups and branches of these versions.
Copyists, however, did not follow their model slavishly ; instead, they
James the Less and Joseph ; the second married Zebedaeus and was the mother
of James the Great and John the evangelist.
75. BHL 505e-505zw. The oldest Latin text that I know of is preserved in
Cambridge, St. Johns College, MS 35, from Bury St Edmunds, possibly late
in the abbacy of Baldwin (1097). See Thomas N. HALL, The Earliest AngloLatin Text of the Trinubium Annae (BHL 505zl), in Th. HALL (ed.), Via Crucis : Essays on Early Medieval Sources and Ideas in Memory of J. E. Cross,
Morgantown, 2002, p. 104-137.
76. Title introduced by Konstantin VON TISCHENDORF in his edition in Evangelia Apocrypha, Leipzig, 18762, p. 93-112. See above, footnote 17.
77. See CCSA 9, p. 96-97 ; 179 ; 262-264.
78. CCSA 9, p. 218. When Pseudo-Matthew is cited, it is always with the
indication : text A or text P. The reader, however, may be struck by the resemblances between both textual forms rather than by their differences, especially
in the first part. Both textual forms are integrated in the LLT-database.



felt free to adapt and expand or rewrite whenever they deemed appropriate Gijsel calls the result une tradition textuelle sauvage.79 In
less than half of the cases was it possible to choose a specific manuscript as a representative of its branch.80 In the other cases he either
reconstructed the common model of the branch by applying traditional
stemmatic criteria, or he supplied additional evidence if the reconstruction was uncertain.81 As a result, the variants reflect a mixed system of
individual manuscripts and reconstructed readings. The critical apparatus concentrates on transformations of the text, conscious or otherwise ;
the editor omitted purely scribal errors, orthographic variations and
grammatical inconsistencies that do not affect the meaning.82
The result is a synoptic edition that usually gives a convincing text.
In cases where the reader may have doubts, the apparatus supplies sufficient information about the variations in the tradition. More difficult
to follow is the precise application of the stemmatic criteria that led
Gijsel to make his choices for the text as he reconstructed it. By the
very fact that the edition reconstructs two subarchetypes, the model for
A and the model for P, we are still some steps removed from the original text. The reader might occasionally want to take those steps and
decide what the original must have been, but obviously this exercise
cannot and should not be done systematically for the whole text.
As for the younger families, the textual form Q appears in a special
comparative apparatus added to the text of P. Gijsel cited only variants
he considered to be interesting, without further differentiation into the
subgroups of Q, of which there are six. As he had proved that the Pars
Altera was not part of the original Pseudo-Matthew, he did not include
its miracle stories in his edition. He did not make use of R at all, as it
has no independent value for establishing the text, nor did he attempt to
make a separate edition of it, because, as he says, nearly every sentence
in it has been changed.83
Using the edition
The decision to edit the two Carolingian textual families with an
apparatus that shows the transformations on the level of the subgroups
and branches, which may be represented either by a single manuscript
or by a reconstructed common model of the branch or subgroup, was
dictated by two considerations. First, the reconstruction of the Carolingian families cannot lead to the recovery of the original text. There is
79. CCSA 9, p. 267.
80. Especially for A : eight out of twelve branches could be represented by
a single manuscript (CCSA 9, p. 271).
81. CCSA 9, p. 268-271.
82. CCSA 9, p. 273-274.
83. CCSA 9, p. 97 : Dans cet tat final, presque aucune phrase na chapp
au zle correcteur des copistes et rviseurs.



no possibility of going back beyond the year 800. Secondly, while the
state of Pseudo-Matthew in pre-Carolingian times is enveloped in mystery, its influence from the ninth century onwards on other Latin and
vernacular texts, on literature, art and devotions, is manifold. Indeed,
Pseudo-Matthew was arguably the second most popular apocryphal text
in the western Middle Ages, yielding first place only to the Evangelium Nicodemi, of which over 450 extant Latin manuscripts have been
identified and described by Zbigniew Izydorczyk.84 So the edition of
Pseudo-Matthew is also meant to be a tool for comparative research
in many fields, especially literary studies. It should enable scholars to
establish which Latin text was used for whatever translation or adaptation they are studying.
There are limitations to what can be provided for scholars in other
fields. It would have been ideal if all the manuscripts of the PseudoMatthew could have been described in minute detail, with a full exposition of the codicological context, the history of the manuscript and
any paratextual material, but this was not practicable. We must bear
in mind that the roughly 200 manuscripts in which the text occurs are
complex : the Pseudo-Matthew, as other Marian apocrypha, is found in
collections of hagiographical texts (legendaries, lectionaries), sermons
and homilies, sometimes with other Marian texts forming a corpus
marianum or in miscellaneous manuscripts. The most vital part of the
information about the individual manuscripts concerns the content of
the apocryphon, i.e. its constituent parts, titles, divisions, indications of
liturgical use, any special characteristics, salient variants, omissions or
additions, marginal annotations, illustrations, etc., as well as characteristic parts of the text which figure in the manuscripts but are not necessarily found in the edition or its apparatus.85 Here the edition offers
much information but it could have offered more. As discussed in the
last paragraph on the ratio edendi, for the textual forms Q and R of
Pseudo-Matthew a different treatment was entirely justified. Nonetheless, a user of CCSA 9 interested in the miracle tales might wish he
could find more details about the content and wording of these stories
in the Q and R traditions.86
84. See the contribution of Z. Izydorczyk in this issue, p. 99-116. When we
add the manuscripts of the Libellus de nativitate Mariae to those of PseudoMatthew, the total number of manuscripts that transmit the Latin apocryphal
tradition of the birth and youth of Mary is about 330.
85. See the text of the Trinubium Annae, characteristic for the Q version,
which was edited separately in the introduction to CCSA 9, p. 93.
86. I give a few other examples of information that was not included in the
preliminary study of the two Libri de nativitate Mariae edited in CCSA 9 and
10. Information about the Lives of the Virgin as found in manuscripts Florence,
Bibl. Med. Laur., Santa Croce, Pl. XV d. 12 (CCSA 9, p. 176 and CCSA 10,
p. 52), Vienna, NB 930 (CCSA 9, p. 176) and Oxford, Bodl., Rawlinson D
1236 (CCSA 9, p. 114 and CCSA 10, p. 131-132) is too concise. This is also



Let us address the case of manuscripts which appear to be marginal, or not readily classifiable. For Pseudo-Matthew Jan Gijsel discovered many such manuscripts and he catalogued some thirty-two of
them as isolated witnesses of a particular textual form.87 These manuscripts, although they were of no value for the edition, are interesting
because they can demonstrate how the apocryphal traditions were used
and adapted.
A particularly interesting example of the importance of marginal
witnesses are the manuscripts containing fragments of Pseudo-Matthew
transmitted in the Latin Infancy Compilation, which was mentioned
above p. 119-121. This compilation was first edited by M.R. James
in two versions on the basis of two manuscripts,88 and re-edited in 2001
by Jean-Daniel Kaestli and Martin McNamara in the same two versions
but on the basis of nine manuscripts seven for the Arundel version
and two for the Hereford version. It is an intriguing example of how
three different apocryphal texts were compiled and blended together to
form one single long new narrative, something considerably more than
a mere compilation.89 The complicated genesis of this work was unravelled by Jean-Daniel Kaestli.90 From his analysis it becomes clear that
at two different points, elements of Pseudo-Matthew were integrated in
an already existing compilation which contained elements of the Protevangelium and of the Liber de nativitate Salvatoris. The first stage
occurred prior to the emergence of the manuscripts of Pseudo-Matthew, in the pre-Carolingian period before the end of the eighth century ; once again, Montpellier H 55 is the oldest witness. The second
stage involved a manuscript of Pseudo-Matthew that contained the Pars
Altera, so it probably occurred in the course of the twelfth century.
As the manuscript transmission of the final text in its most developed
form was not overly complex, a synoptic edition with critical apparatus giving the readings of the individual manuscripts was feasible.91
The new texts correct James on many occasions, but the greatest value
the case for the sermon De Nativitate Mariae, preserved in manuscripts London,
BL, Add. Mss. 41069 (CCSA 9, p. 198-199), Basel B IX 32 (ibid., p. 199),
and Basel B V 32 (CCSA 10, p. 128). As for the versions of the Trinubium
Annae present in manuscripts of the Libellus de nativitate Mariae (see CCSA
10, p. 171), I only identified them but did not reproduce their wording.
87. Forty-one if we include the manuscripts of the Latin Infancy Compilation. See CCSA 9, p. 186-217 ; see also CCSA 10, p. 93-101 ; 127-138.
88. The Arundel version from MS London, BL Arundel 404, and the Hereford version from MS Hereford O 3 9.
89. Rita BEYERS, Le charme discret de la Compilation latine de lenfance
in Camille FOCANT & Andr WNIN (eds), Analyse narrative et Bible, LouvainParis-Dudley, 2005, p. 351-368.
90. See footnote 5 above.
91. As for Pseudo-Matthew, the two versions of the Latin Infancy Compilation (Arundel : 16361 words, and Hereford : 14235 words) have been integrated
separately in the LLT-database.



of the edition lies in its source apparatus, where the reader finds the
results of all the detective work of the editors. It is a truly remarkable
achievement, revealing all the richness and complexity of this text. For
Pseudo-Matthew, it was just one of the fragmentary, less interesting
witnesses of the narrative, but as an independent text the Compilation
is a gold-mine for apocryphal traditions, a mine, as Jean-Daniel Kaestli
has recently shown, that still needs to be further explored.
I turn briefly to the resum of the first part of Pseudo-Matthew
known as the Libellus de nativitate Mariae or Little Book on the Birth
of Mary. Here the editorial problems were less daunting, and an edition
based on a selection of individual manuscripts was an acceptable way
to treat the text. The presentation of the manuscript tradition is essentially the same as for Pseudo-Matthew.92 The only major unanswered
question about the Little Book on the Birth of Mary is its authorship.
The work (2341 words, less than half the length of Pseudo-Matthew)
has been attributed by Cyrille Lambot to Paschasius Radbertus and
dated between 846 and 849, on the strength of a fragmentary letter
by Hincmar archbishop of Rheims to Odo bishop of Beauvais from
868-869. But this evidence is not conclusive,93 so we have to turn to
authorship analysis. The traditional way to do this is to look for notable resemblances on the level of language, style, content. After making
such an investigation some years ago, I concluded that there are no
convincing resemblances, and I rejected the attribution. But there are
still scholars who prefer to see Paschasius as the author of the Little
Book on the Birth of Mary.94 Maybe it is time to revisit the question
with the help of new computational techniques and let specialists in
stylometric research carry out tests, in the hope of shedding new light
on the matter.95

The transmission of the Pseudo-Matthew appears to be more complicated than that of the Virtutes apostolorum, if we consider how many
extant manuscripts there are, for which the editor had to establish the
correct position in the framework of the textual transmission. On the
other hand, the three apocrypha concerning the birth and youth of Mary,
even in the highly complicated form of the Latin Infancy Compilation,
cannot rival the intricacies of the genesis of the Virtutes apostolorum.
For the edition of Pseudo-Matthew a workable solution was found

92. See footnote 86 above.

93. See footnote 47 above.
94. On the principle of decuit, potuit, ergo fecit freely translated : he had
the motive, the means and the opportunity, so he did it.
95. At the University of Antwerp, Mike Kestemont specialises in research
that deals with stylometric authorship attribution in medieval literature.



through a combination of traditional stemmatic thinking, two selective

critical apparatus, and a thorough analysis of all available manuscripts.
The edition opens the way for further investigation and progress in the
field of apocryphal studies, especially concerning the metamorphoses of
the apocryphon in the later Middle Ages.
Every edition is, by definition, unfinished. In the context of the
question about the appropriate way to edit a text as fluid and protean
as the Virtutes apostolorum, it is useful to bear in mind the words of
the French classical scholar and editor of Greek texts, Alphonse Dain :
En fait, ldition vraiment critique, celle qui repose sur un choix et
un jugement, sera toujours leditio minor.96 More recently, the Dutch
classicist and mediolatinist Haijo Jan Westra summarized our task thus :
But in the end, the user/reader still requires a text and an editor to
make an informed choice. At some point, the textual flux is arrested
and fixed in time for the modern reader, just as the medieval scribe did
for his patron and public.97

96. Alphonse DAIN (ed.), Sophocle Tome I (Collection des Universits de

France), Paris, 19551, 19672, p. LI.
97. New Philology and the Editing of Medieval Latin Texts in Keith
BUSBY (ed.), Towards a Synthesis ? Essays on The New Philology (Faux titre.
tudes de langue et littrature franaises, No. 68), Amsterdam Atlanta GA,
1993, p. 49-58 : 57. H.J. Westra is now emeritus professor of Greek and Roman
Studies at the University of Calgary.

Department of Early and Medieval Irish,
University College Cork


Within the corpus of extant vernacular medieval Irish sermons, the
dominance of eschatological themes borrowed from apocrypha is striking. This raises the question whether this phenomenon is accidental or
symptomatic of Irish preaching in the period. The answer may lie in the
early medieval Latin collections of homilies which demonstrate affiliations with Insular material. Three eschatological apocrypha which were
particularly popular in medieval Ireland and England were chosen as
test cases to investigate the nature of this influence : The Three Utterances, The Apocalypse of Thomas and The Seven Heavens apocryphon. The interrelated manuscript sources in which these texts survive
are discussed, as well as the possible context of their dissemination.
Parmi le corpus des homlies mdivales irlandaises vernaculaires
encore existantes, la dominance des thmes eschatologiques emprunts
aux apocryphes est frappante. Cela soulve la question de savoir si
ce phnomne est accidentel ou symptomatique du prdicateur irlandais de lpoque. La rponse pourrait rsider dans les collections des
homlies latines mdivales qui dmontrent des affiliations avec des
crits insulaires. Trois apocryphes eschatologiques qui taient particulirement populaires dans lIrlande mdivale et en Angleterre ont t
choisis comme spcimen pour tudier la nature de ces influences : Les
Trois Dclarations , LApocalypse de Thomas et Les Sept Cieux .
La corrlation des sources manuscrites dans lesquelles ces textes survivent est analyse, ainsi que le contexte de leur dissmination.
The dominance of eschatological themes in the surviving corpus
of medieval Irish sermons has been noted on a number of occasions.1
1. Cf. OLOUGHLIN, T., The Celtic Homily : Creeds and Eschatology, Milltown Studies 41, 1998, p. 99-115 ; MAC EOIN, G., Na Crocha Danacha i Seanmir na Sean-agus na Men-Ghaeilge, Diagacht/Theology : Bulletin of the
Western Theology Research Association 2, 1997, p. 20-6 ; MAC DONNCHA, F.,


Apocrypha 23, 2012, p. 141-153



Brian Murdoch, in particular, drew attention to the concentration upon

the eschaton as one of the most noticeable features of Irish sermons.2
In addition, the eschatological references employed often draw upon
apocryphal sources, a fact which was highlighted in the list of medieval
Irish sermons drafted by H.L.C. Tristram.3 This convergence of apocrypha and sermons presents a number of interesting issues with regard to
the dynamics of the survival and dissemination of apocrypha through
homily collections. It also, however, presents a number of challenges,
not least of which is defining what we mean by the term sermon and
homily in an Insular medieval context. The fluidity of the sources has
been underlined by the work of Thomas OLoughlin who cautioned
against the imposition of a strict demarcation between catechetics and
homiletics in his attempt to supply a working definition of the category
homily.4 There is the added problem that certain texts include some
homiletic features such as the use of first plural exhortations and homiletic formulae but the work as a whole may not necessarily have been
used as a sermon.5 This is a pertinent problem with regard to theological tracts or exemplum-type narratives which may have been used to
construct homilies and in some cases their inclusion in homiletic collections is presumably for that purpose but the dynamics of the process
remain unclear. Ambiguity in this regard explains the hesitancy with
which the term sermon is applied to important Irish vernacular eschatological works drawing on apocryphal sources even where such works
include a number of homiletic features.6 The overlap between the study
Medieval Irish Homilies, in MCNAMARA, M., (ed.), Biblical Studies : the Medieval Irish Contribution, Dublin, 1976, p. 59-71.
2. MURDOCH, B., Preaching in medieval Ireland : the Irish tradition, in
FLETCHER, A., and R. GILLESPIE, (ed.), Irish Preaching 700-1700, Dublin, 2001,
p. 40-55 (p. 54).
3. TRISTRAM, H.L.C., Sex Aetates Mundi. Die Weltzeitalter bei den Angelsachsen und den Iren. Untersuchungen und Texte, Heidelberg, 1985 (p. 13352) ; and Early Insular Preaching : Verbal Artistry and Method of Composition,
Vienna, Sitzungsberichte der sterreichischen Akademie, Phil. Hist. Kl. 623,
1995. An earlier list of vernacular sermons (many of which are in fact macaronic) was supplied by STOKES, W., A Middle-Irish homily on S. Martin of
Tours, Revue Celtique 2, 1873-5, p. 381-402.
4. OLOUGHLIN, T., Irish Preaching before the end of the ninth century :
assessing the extent of our evidence, in FLETCHER, A., and R. GILLESPIE, (ed.),
Irish Preaching 700-1700, Dublin, 2001, p. 18-39 (p. 19).
5. Brian Murdoch discussed this problem with regard to Saints lives with
appended homiletic formulae but little attempt to create an integrated sermon,
see Preaching in medieval Ireland : the Irish tradition, p. 46-8.
6. For a summary of the debate, see MURDOCH, ibid., p. 43-5. Prof. Tristram tentatively categorised four significant eschatological works as sermons :
D Brn Flatha Nime (Two Sorrows of Heaven), Fs Adomnn (Vision of
Adomnn), Scla Li Brtha (Tidings of Doomsday) and Scla ha hEsrgi
(Tidings of the Resurrection). The editor of the last two texts considered at
least one of them to be a homily, STOKES, W., The Tidings of Doomsday,



of apocrypha and of sermons in an Insular context has again been

brought into focus by John Careys recent edition of the Tenga Bithnua (Evernew Tongue), a long work in a question-and-answer format
which is thought to be based on a lost apocryphon connected with the
apostle Philip. In his edition, Carey argues convincingly that this significant text is a homily for the vigil of Easter, opening as it does with
a pericope from Genesis, 1, 1.7 The fact that one of the most important
texts for the study of medieval Irish eschatological apocrypha survives
as a homily again underlines the importance of an understanding of
the homiletic tradition in order to grasp the complexities of apocrypha
transmission. It also raises some interesting questions with regard to
the reasons for the high percentage of apocrypha surviving in the relatively circumscribed corpus of medieval Irish sermons. The abundance
of apocrypha cannot simply be an accident of transmission and must in
some way be symptomatic of early Irish preaching in general.
Moreover, the homiletic tradition in the Irish language cannot have
come from a vacuum ; there were presumably Latin antecedents or
models. The principal challenge, however, in identifying sources for the
vernacular sermons is that the surviving corpus of early medieval Irish
homiletic material in Latin is relatively small.8 Defining the corpus is
further complicated by the difficulties, outlined above, in delimiting the
terms homily or sermon in the early Middle Ages, and the perennial
challenge of establishing the possible Irish connections of early medieval texts in Latin preserved in Continental manuscripts. Even where
we have a named author who is clearly Irish, extrapolating from the
two surviving authorial examples to trends in medieval Irish preaching
as a whole is problematic, as pointed out by Thomas OLoughlin in his
discussion of the extant material.9 In the case of the first such author,
Eriugena, his sole surviving homily10 displays a level of sophistication
and erudition that cannot be representative of preaching outside of elite,
intellectual circles. In the second instance, the origin of the Instructiones11 attributed to St Columbanus is highly contested,12 and again

Revue Celtique 4, 1879-80, p. 245-57 and Tidings of the Resurrection, Revue

Celtique 25, 1904, p. 232-59.
7. CAREY, J., (ed.), In Tenga Bithnua : The Ever-New Tongue, Turnhout,
2009, Apocrypha Hiberniae II : Apocalyptica 1, CCSA 16, p. 233-6.
8. For a comprehensive survey see OLOUGHLIN, Irish Preaching, art. cit.,
p. 18-39. Cf. also OLOUGHLIN, T., The Celtic Homily, art. cit.
9. OLOUGHLIN, Irish Preaching, art. cit., p. 20.
10. JEAUNEAU, E., (ed.), Jean Scot, Homlie sur le prologue de Jean, Paris,
SC 151, 1969. LAPIDGE, M., and R. SHARPE, A Bibliography of Celtic-Latin Literature 400-1200, Dublin, 1985, 702.
11. WALKER, G.S.M., (ed.), Sancti Columbani opera, Dublin, Scriptores
Latini Hiberniae 2, 1957, p. 60-121.
12. For a survey of the debate see OLOUGHLIN, Irish Preaching, art. cit.,
p. 20-22, and STANCLIFFE, C., The Thirteen Sermons attributed to Columbanus



the literate style, in addition to the subject matter aimed at a monastic audience is not, in any case, representative of medieval preaching
in general. Moreover, neither author displayed the same concentration
on Doomsday and the Afterlife as is evident in vernacular preaching.
Collections of anonymous Latin sermons with Insular affiliations would
seem to be both more representative of preaching to a general audience and also provide more fruitful parallels with vernacular sources
given their inclusion of catechetical and eschatological themes. Thomas
OLoughlin suggested that representative examples of medieval Irish
preaching might be found in collections such as the so-called Catechesis Celtica, although he also pointed out that the attribution of this
work is problematic.13 The Catechesis Celtica refers to a series of biblical commentaries which may have been employed in the preparation of
homilies and which survive in an early tenth-century manuscript probably from Brittany.14 It has been argued that this collection has strong
Irish affiliations based on certain literary and thematic features of the
material, although the glosses which it contains are in Welsh or Cornish, not Irish.15 Two further collections have also been highlighted as
texts with possible Irish connections, although the attribution is tentative. The first such item is a collection of homiletic material, evidently
aimed at a monastic community, and sometimes referred to by the title
Catechesis Cracoviensis because the sole extant manuscript is in Cracow.16 The manuscript is believed to date from the late eighth or early
ninth centuries and to originate from northern Italy.17 The homilies, or
more correctly monastic conferences, which the manuscript contains,
concentrate on the fasting periods of the liturgical calendar. Arguments
have been proposed, on stylistic and palaeographic grounds, for an Irish
monastic centre possibly in France or Italy as the original home of
and the question of their authorship, in LAPIDGE, M., (ed.), Columbanus : studies
in the Latin writings, Woodbridge, 1997, p. 93-202.
13. OLOUGHLIN, Irish Preaching, art. cit., p. 22-25.
14. Biblioteca Apostolica Vaticana, Reg. lat. 49. LAPIDGE and SHARPE, op.
cit., 974. For a partial edition, see WILMART, A., Catchses celtiques, Analecta Reginensia : Extraits des manuscrits latins de la reine Christine conservs
au Vatican, Vatican City, Studi e Testi 59, 1933, p. 29-112.
15. MCNAMARA, M., The Irish Affiliations of the Catechesis Celtica, Celtica 21, 1990, p. 291-334 ; and The affiliations and origins of the Catechesis
Celtica : an ongoing quest, in OLOUGHLIN, T., (ed.), The Scriptures and Early
Medieval Ireland, Turnhout, 1999, p. 179-98.
16. Cracow, Cathedral Library, MS 140 KP. 43 (VIIIex/IXin). CLA 1593 ;
LAPIDGE and SHARPE, op. cit., 802. The problems associated with the titles Catechesis Cracoviensis and Catechesis Veronensis are discussed in OLOUGHLIN,
Irish Preaching, art. cit., p. 25, n. 40. Both terms were coined on the model
of the title Catechesis Celtica but they are not catecheses but homilaria as
OLoughlin pointed out.
17. DAVID, P. , Un recueil de confrences monastiques irlandaises du VIIIe
sicle, Revue Bndictine 49, 1937, p. 62-89.



the collection.18 The Continental, non-Insular features of the language,

however, have also been highlighted.19 The second collection with possible Insular affiliations is a series of eleven homilies preserved in a
manuscript in Verona, and hence occasionally referred to as the Catechesis Veronensis.20 As is the case with the Cracow collection, the
homilies are believed to originate from northern Italy and may demonstrate Insular influence in certain palaeographic features and the choice
of contents, although the evidence is difficult to assess.

The Three Utterances

There is less doubt, however, in regard to the Irish affiliations of a
series of seven homilies, which are believed to date to the late eighth
century and which were collectively labelled the In nomine Dei summi
sermons by their editor.21 Nor are there the same typological uncertainties in this case, since the texts are clearly homiletic and aimed at a
lay audience.22 The title In nomine Dei summi comes from the fact
that this is the opening formula employed at the beginning of each
text in the Vatican manuscript (Pal. lat. 220), which the editor McNally
judged to represent the purest tradition. The manuscript is in an AngloSaxon script and is thought to have been compiled in ninth-century
Middle or Upper Rhineland. McNally based his edition on this text
and another early ninth-century Vatican manuscript.23 Charles Wright
subsequently unearthed a third copy of the sermons in a Berlin manuscript.24 McNally viewed the seven texts as a coherent collection created by a sole author but this theory has been challenged by Thomas
18. BISCHOFF, B., Wendepunkte in der Geschichte der lateinischen Exegese
im Frhmittelalter, Sacris Erudiri 6, 1954, p. 189-279 (p. 221). For a summation of the arguments, see OLOUGHLIN, Irish Preaching, art. cit., p. 26.
19. COCCIA, E., La cultura irlandese precarolingia : miracolo o mito ?, Studi
Medievali 8, 1967, p. 257-42 (p. 347).
20. Verona, Biblioteca Capitolare, 67 (64). MARTIN, L.T., The Catechesis
Veronensis, in OLOUGHLIN, T., (ed.), The Scriptures and Early Medieval Ireland, p. 151-61. LAPIDGE and SHARPE, op. cit., 804. For an assessment of
the problems associated with the possible Irish affiliation of the collection, see
OLOUGHLIN, Irish Preaching, art. cit., p. 27.
21. MCNALLY, R.E., In Nomine Dei Summi : Seven Hiberno-Latin Sermons, Traditio 35, 1979, p. 121-43. LAPIDGE and SHARPE, op. cit., 803. CPL
22. For a discussion of the texts, see : OLOUGHLIN, Irish Preaching, art.
cit. p. 30-38. The texts are also discussed and translated in OLOUGHLIN, The
Celtic homily, art. cit.
23. Vatican City, BAV, Pal. lat. 212.
24. Berlin, Staatsbibliothek, Phillipps 1716. See WRIGHT, C., Apocryphal
Lore and Insular Tradition in St Gall, Stiftsbibliothek MS 908, in N CHATHIN,
P., and M. RICHTER, (ed.), Irland und die Christenheit : Bibelstudien und Mission, Stuttgart, 1987, p. 124-45 (p. 135).



OLoughlin, who questioned the existence of a discernable authorial

unity.25 He noted, however, that in terms of content, length, and tone
there are clear similarities between texts 3-7 which all have a strong
catechetical tone. Sermons 1 and 2 he argued, however, are quite different in style and tone ; in addition they are both eschatological in
theme. This challenge to the concept of the seven sermons as a unified
collection was recently reinforced by the work of Toms OSullivan as
part of his doctoral dissertation on the manuscript, in which he stressed
the Insular affiliations of the codex.26 He also highlighted the important influence exerted by such Insular homiletic collections in early
medieval Germany, a fact reinforced by the reports of contemporary
commentators praising the sermons of medieval Irish missionaries.27
OSullivan demonstrated that the seven sermons are part of a larger
early medieval collection which he entitled the Predicationes Palatinae
and moreover, he identified a fourth partial copy of the work.28 Also,
as Wright, OLoughlin and OSullivan pointed out, the first of the In
nomine Dei summi sermons was widely disseminated independently of
the other texts, and is better known to scholars of Insular apocrypha as
the Three Utterances apocryphon. The three utterances referred to in
the title are those of either damned souls or righteous souls when they
learn their ultimate fate after a struggle between angels and demons for
the souls. The Three Utterances enjoyed remarkable popularity in the
early Middle Ages and to date Charles Wright has identified some 39
manuscripts containing the Latin text.29 Wright also noted that many
of the earliest manuscripts display Insular palaeographical symptoms,
or are transmitted in manuscripts which contain other Hiberno-Latin
texts.30 One such example is the so-called Florilegium Frisingense, a
late eighth-century manuscript compiled in Freising by an Anglo-Saxon
25. OLOUGHLIN, Irish Preaching, art. cit. p. 31-35.
26. OSULLIVAN, T., Predicationes Palatinae : The Sermons in Vat. Pal. lat.
220 as an Insular Resource for the Christianization of Early Medieval Germany, Ph.D. dissertation, Saint Louis University, 2011. I am very grateful to
the author for allowing me access to this work.
27. Cf. MURDOCH, art. cit., p. 42.
28. Munich, Bayerische Staatsbibliothek, Clm 6293.
29. WRIGHT, C., Three Utterances Apocryphon, in BIGGS, F.M., Sources of
Anglo-Saxon Literary Culture : The Apocrypha, Kalamazoo, Instrumenta Anglistica Mediaevalia 1, 2007, p. 80-83. Cf. also WRIGHT, C., The Irish Tradition in
Old English Literature, Cambridge, Cambridge Studies in Anglo-Saxon England
6, 1993, p. 215-8. For editions of the Latin text, other than that produced by
McNally from the Vatican manuscripts, see : WILLARD, R., The Latin Texts of
The Three Utterances of the Soul, Speculum 12, 1937, p. 147-66 ; GRGOIRE,
R., Les homliaires du Moyen ge : Inventaire et analyse de manuscrits, Rome,
Rerum Ecclesiasticarum Documenta, Series Maior, Fontes 6, 1966, p. 224-5 ;
GARCA LARRAGUETA, S.A., Las Glosas Emilianenses : Edicon y Estudio,
Logroo, Biblioteca de Temas Riojanos 54, 1984, p. 139-46.
30. WRIGHT, The Irish Tradition, op. cit., p. 217.



scribe who styled himself Peregrinus.31 Significantly, another Freising

manuscript contains the standard long version of the apocryphon along
with an abbreviated version of the text and also in the same manuscript
we find a slightly elaborated version of sermon 4 from the In nomine
Dei summi collection, thus demonstrating that the Three Utterances
also circulated independently in conjunction with other sermons from
the same collection. Other manuscripts in which the Three Utterances
sermon survives are closely associated with Irish monastic centres on
the Continent such as St Gallen, Stiftsbibliothek, 908.32 McNally and
Wright, among others, considered that the account originated in an Irish
milieu and in addition to the Latin texts, an Old Irish adaptation of the
apocryphon survives.33 There are also three Old English translations of
the sermon which survive, testament to the popularity of the text in
medieval England.34

The Apocalypse of Thomas

As stated above, the Three Utterances apocryphon is the first of
the In nomine Dei summi homilies ; however, there is another sermon
which is found just before the seven which McNally edited in three out
of the four surviving manuscripts containing the Predicationes Palatinae collection. This homily, which predates the year 800, is entitled
Predicatio carere tormenta and it should probably be viewed as part
of the In nomine Dei summi collection ; in addition, its position just
before the Three Utterances apocryphon may not be accidental given
its eschatological tone. The sermon in question is also of significance to
apocrypha specialists because after exploring a number of eschatological themes, the homilist includes towards the end of the text an abbre-

31. Munich, Bayerische Staatsbibliothek, Clm. 6433 (s. VIIIex). LEHNER,

A., (ed.) Florilegia : Florilegium Frisingense (Clm 6433), Testimonia Divinae
Scripturae, Turnhout, CCSL 108D, 1987. CLA 9, 1283.
32. For the other Insular texts which this manuscript contains, see WRIGHT,
Apocryphal Lore and Insular Tradition in St Gall, Stiftsbibliothek MS 908,
art. cit.
33. MARSTRANDER, C., The Two Deaths, riu 5, 1911, p. 120-5 ; SEYMOUR, St. J.D., The Bringing Forth of the Soul in Irish Literature, Journal of
Theological Studies 22, 1920, p. 16-20 ; MCNAMARA, M., The Apocrypha in the
Irish Church, Dublin, 1984, p. 127-8. A new edition of the text has been prepared by Katja Ritari and will be published shortly in CAREY, John, Emma NIC
CRTHAIGH, and Caitrona DOCHARTAIGH (ed.), The End and Beyond : Medieval
Irish Eschatology, Aberystwyth, 2013.
34. WILLARD, R., Two Apocrypha in Old English Homilies, Leipzig, Beitrge
zur englischen Philologie 30, 1935, and The Address of the Soul to the Body,
Publications of the Modern Language Association of America 50, 1935, p. 95783. For a survey of the literature on these Old English homilies, see WRIGHT,
Three Utterances Apocryphon, art. cit.



viated three-day list of signs35 before Doomsday evidently dependent

upon the Apocalypse of Thomas.36 This apocryphal work purports to
be an account of the signs leading up to the end of time as recounted
by Christ to his doubting apostle. The text supplied at the end of the
sermon is relatively short but it displays strong resemblances to the
Apocalypse of Thomas such as the fact that the eschatological signs are
in a series of days organised by ordinal numbers leading up to Judgement Day. In addition, there are some direct verbal parallels with the
source text.37 Instead of the standard seven or eight days of signs in
the Apocalypse, the sermon, however, only supplies three days. It is
possible that the text is deliberately truncated ; as if the homilist were
quoting from the larger work in order to evoke images of Doomsday in
his listeners minds. In this context, it is interesting to note, as highlighted by OSullivan, that the sermon opens with the words Peccatoribus autem, which seems to imply that the work is extracted from a
longer text.38 On the other hand, it is equally possible that the redactor
of this homily only possessed a fragment of the Apocalypse. A possibility which again raises interesting questions about how apocrypha
were truncated, extracted, abbreviated and adapted for homiletic dissemination. It is interesting to note, however, that the manuscript which
McNally used as his base text (Pal. lat. 220) also contains a full copy
of the Latin Apocalypse of Thomas some twenty-one folios after the
In nomine Dei summi sermons.39 Therefore, the manuscript copyist,
if not the original homilist, could certainly have included the full text
should he have wished. The Apocalypse of Thomas enjoyed great popularity in medieval England where it was used in sermons and Charles
Wright has written in detail about the English homiletic adaptations of
the work.40 There is also considerable evidence of the influence of the

35. Wrights edition of the sermon (Catechetical Text Containing Some

Parallels with the Apocalypse of Thomas) can be found as an appendix to an
article by Martin MCNAMARA on The (Fifteen) Signs before Doomsday in Irish
Tradition, Miscellanea Patristica, Warszawskie Studia Teologiczne 20.2, 2007,
p. 252-4.
36. GEERARD, M., (ed.), Clavis apocryphorum Novi Testamenti, Turnhout,
1992, p. 210 (no. 326) ; CPL 796a ; BOVON, F., P. GEOLTRAIN, and J.-D. KAESTLI,
(ed.), crits apocryphes chrtiens, Bibliothque de la Pliade, Paris, 1997,
vol. 2, p. 1021-43 ; BIHLMEYER, P., Un texte non interpol de lApocalypse de
Thomas, Revue bndictine 28, 1911, p. 270-82.
37. See DOCHARTAIGH, C., The Judgement and its Signs : Introduction, in
CAREY, et al., The End and Beyond, op. cit.
38 OSULLIVAN, Predicationes Palatinae, op. cit.
39. The text of the Apocalypse from the Vatican manuscript has been edited
for the first time by Toms OSullivan, and will appear shortly in CAREY, et al.,
The End and Beyond, op. cit.
40. WRIGHT, C.D., The Apocalypse of Thomas : Some New Latin Texts and
their Significance for the Old English versions, POWELL, K., and D. SCRAGG



Apocalypse in Medieval Irish texts.41 In addition, a Latin copy of the

Apocalypse of Thomas is also extant in the Catechesis Celtica collection discussed above.42
The Predicationes Palatinae is not, however, the only context in
which we find this homily quoting the Apocalypse of Thomas. The
fourth extant copy is found in an expanded form under the title Sanctus
Augustinus episcopus dixit,43 within a collection of texts in a Karlsruhe
manuscript44 sometimes referred to as the Apocrypha Priscillianistica.45 This group of six sermons, including the most complete surviving
copy of the homily (3), only survives as a unit in this manuscript.
M. R. James46 was the first to argue in favour of significant Irish affiliation with regard to the sermons in the Karlsruhe manuscript, and since
then the texts have received a good deal of attention from Insular apocrypha scholars47 although they have rarely been discussed in the context of Insular homiletics. The manuscript itself is thought to date from
the late eighth or early ninth century and on palaeographic grounds
has been connected with Novara in Northern Italy.48 Shortly after its
compilation, however, the manuscript found its way to the monastery
of Reichenau where many other important Irish manuscripts were also
housed.49 The manuscript itself consists of three parts which have been
bound together but which are thought to have originally circulated separately. The sermon attributed to St Augustine is part of a collection at
(ed.), Apocryphal Texts and Traditions in Anglo-Saxon England, Woodbridge,
2003, p. 27-64.
41. MCNAMARA, The (Fifteen) Signs before Doomsday and DOCHARTAIGH, art. cit.
42. For the text of the Apocalypse in the Catechesis Celtica and its relationship with other Insular versions of the text, see WRIGHT, art. cit.
43. CLA 792.
44. Karlsruhe, Badische Landesbibliothek, Aug. perg. 254 (fol. 157v-160r).
45. Charles Wright coined this title based on the the title of De Bruynes
edition, see WRIGHT, C., Apocrypha Priscillianistica, BIGGS, op. cit., p. 73-4 ;
DE BRUYNE, (ed.), Fragments retrouvs dapocryphes priscillianistes, Revue
bndictine 24, 1907, p. 318-35 ( III p. 324-5, ApocTh citation lines 54-64).
46. JAMES, M. R., Irish Apocrypha, Journal of Theological Studies 20,
1918-9, p. 9-16 (p. 16).
47. LAPIDGE and SHARPE, op. cit., 1252 ; DUMVILLE, D., Biblical Apocrypha
and the Early Irish : A Preliminary Investigation, Proceedings of the Royal
Irish Academy 73C, 1973, p. 299-338 (p. 325-8) ; WRIGHT, C.D., The Irish Tradition in Old English Literature, Cambridge, Cambridge Studies in Anglo-Saxon
England 6, 1993.
48. CAU, E., Scrittura e cultura a Novara (secoli VIII-IX) , Ricerche medievali 6-9, 1971-4, p. 1-87 (p. 29-31). Cf. HOLDER, A., Die Reichenauer Handschriften, erster Band : Die Pergamenthandschriften, Leipzig, 1906, p. 573-9
(p. 575-7) ; CLA 1110.
49. Cf. LWE, H., (ed.), Die Iren und Europa im frheren Mittelalter, Stuttgart, Verffentlichungen des Europa Zentrums Tbingen, Kulturwissenschaftliche Reihe, 1982.



the beginning of the third section entitled Incipit collectario de diuersis

sententiis,50 many of the items in which have Insular associations.51

The Seven Heavens

As stated above, the fullest and most developed version of the homily containing the extract from the Apocalypse of Thomas and attributed to St Augustine is to be found in the Karlsruhe manuscript. This
fact alone guarantees an important place for this collection in the study
of medieval Insular apocrypha. Moreover, immediately preceding this
pseudo-Augustinian homily in the manuscript there is a short sermon52
( 2) containing a partial account of an apocryphon which has also
been much debated by Insular medievalists.53 The account describes
the journey of the soul through seven heavens : the lower heavens are
hell-like and full of torments, and the upper heavens become steadily
more paradisaical. After a description of the judgement of souls in the
seventh heaven, the fate of the damned in hell is then described. M.R.
James was also the first to recognise the relationship between this text
and certain Insular vision tales and sermons.54 The description of the
Seven Heavens had a profound effect on Insular eschatology, and elements of the imagery from the account can be found in many other
texts. Its relationship to important Irish eschatological texts such as Fs
Adomnin (Vision of Adomnn)55 and In Tenga Bithnua (Evernew
Tongue) has been analysed, in addition to the close verbal parallels
between the text and an English homily preserved in Cambridge, Corpus Christi College MS 41.56 Although the account of the Seven Heavens preserved in the Karlsruhe manuscript is the earliest such account
50. fol. 153-213.
51. For instance there are many striking parallels between the material in
the manuscript and the Irish Liber de Numeris, see MCNALLY, R.E., Der irische
Liber de numeris : Eine Quellenanalyse des pseudo-isidorischen Liber de
numeris, Munich, 1957.
52. A new edition of the homily, as well as an analysis of its contents
has been prepared by CAREY, J., The Reichenau Seven Heavens Homily, in
CAREY, et al., op. cit.
53. SEYMOUR, St J.D., The Seven Heavens in Irish Literature, Zeitschrift
fr celtische Philologie 14, 1923, p. 18-30 (p. 22-3), and The Vision of Adamnan, Proceedings of the Royal Irish Academy 37C, 1927, p. 304-12 (p. 3069). DUMVILLE, art. cit., p. 326-8 ; BAUCKHAM, R., The Apocalypse of the Seven
Heavens : The Latin Version, Apocrypha 4, 1993, p. 141-73.
54. JAMES, Irish Apocrypha, art. cit., p. 13-16.
55. A new edition of the vision has been prepared by John Carey for inclusion in Apocrypha Hiberniae II : Apocalyptica 2, CCSA.
56. WILLARD, op. cit. ; WRIGHT, C., Seven Heavens Apocryphon, in BIGGS,
Sources of Anglo-Saxon Literary Culture, op. cit., p. 78-9 ; BAUCKHAM, art. cit.,
p. 154-5. A new edition of this Old English text has been prepared by Nicole
Volmering for inclusion in CAREY, et al., op. cit. The volume also includes new



with Insular affiliations, analysing its exact relationship with the Irish
and Old English texts is complicated by the fragmentary nature of the
homily. As was also the case with the Apocalypse of Thomas fragment
in homily 3, the Seven Heavens description in homily 2 is incomplete, beginning as it does with the third heaven. Much of the debate
with regard to the Seven Heavens homily, however, has concentrated
on its origins and its possible relationship with a number of early Coptic Gnostic texts.57 This debate has been re-ignited recently by the work
of Charlotte Touati58 who has re-visited De Bruynes original assessment that the collection of sermons in the Karlsruhe manuscript are
Priscillianist in origin. This hypothesis had long been dismissed and
the Insular aspects of the work had been highlighted instead. However,
Touati points to the Gnostic influence on Priscillianism, the accusations of over-reliance on apocrypha made against the Priscillianists in
conjunction with the contacts between Ireland and Spain in the early
medieval period,59 as factors which, when combined, would provide a
plausible context for the dissemination of our sermons.60 Certainly, the
large number of apocryphal texts circulating in early Medieval Irish circles has been remarked upon on a number of occasions.61 With regard
to possible Gnostic influence on some of the texts discussed here, it
is interesting to note that Toms OSullivan in his recently completed
edition of the Apocalypse of Thomas text from the Vatican manuscript
Pal. lat. 220, which also contains the Predicationes Palatinae collection
discussed above, lays particular emphasis on the concluding section of

editions of the Irish Seven Heavens accounts and a discussion of the relationship between the Insular texts.
57. DANDO, M., Les Gnostiques dgypte, les Priscillianistes dEspagne
et lglise primitive dIrlande, Cahiers dtudes cathares 56, 1972, p. 3-34
(p. 8-12) ; STEVENSON, J., Ascent through the Heavens, from Egypt to Ireland,
Cambridge Medieval Celtic Studies 4, 1982, p. 21-35 ; CAREY, J., The Seven
Heavens and the Twelve Dragons in Insular Apocalyptic, MCNAMARA, M.,
(ed.), Apocalyptic and Eschatological Heritage : The Middle East and Celtic
Realms, Dublin, 2003, p. 121-36.
58. TOUATI, C., The Apocalypse of the Seven Heavens : from Egypt to
Ireland, CAREY, et al., op. cit.
59. J.N. Hillgarth has written extensively on the connections between Ireland and Spain in the period : The East, Visigothic Spain and the Irish, Studia Patristica 4, Texte und Untersuchungen 79, 1961, p. 442-56 ; Visigothic
Spain and Early Christian Ireland, Proceedings of the Royal Irish Academy
62C, 1962, p. 167-94 ; Ireland and Spain in the Seventh Century, Peritia 3,
1984, p. 1-16.
60. It is interesting to note that this suggested route of dissemination is the
same as that which has been proposed for the development of the Creed, i.e.
from Spain to Ireland, to Anglo-Saxon England and then to the Continent, cf.
OLOUGHLIN, Irish Preaching, art. cit., p. 33.
61. See for instance DUMVILLE, art. cit ; MCNAMARA, The Apocrypha, op. cit.
and the volumes of Apocrypha Hiberniae, CCSA.



the work.62 OSullivan argues that the concluding words of the Apocalypse in this manuscript contain some striking dualist overtones and
other unorthodox references which may point to Priscillianist influence.
This may explain why this concluding passage is not included in the
other surviving versions of the text. De Bruyne, the original editor of
the so-called Apocrypha Priscillianistica, believed that the Karlsruhe
sermons were drawing on a number of different apocryphal sources. In
our discussion here, only the material in homilies 2 and 3 has been
touched upon. Perhaps an investigation of the sources of the other four
sermons would shed more light on the Priscillianist origins or otherwise
of this collection.
It is clear, in any case, that the collection in the Karlsruhe manuscript plays an important role in understanding the dissemination of
apocrypha in the early medieval period. In this regard it is interesting
to note that the Three Utterances account is also found in this same
manuscript in a section not too far from the collection of sermons we
have been discussing. Therefore, what is striking is the extent to which
the three apocryphal texts which are the focus of this article, circulated
in parallel. In summation, both the Seven Heavens account and the
abbreviated Apocalypse of Thomas text are included in the Karlsruhe
sermons and the Three Utterances account is in a different section
of the same manuscript. The Three Utterances text is also the first of
the In nomine Dei summi sermons, and another text of the abbreviated Apocalypse of Thomas homily comes just before it in three out
of the four extant manuscripts, while the full text of the Apocalypse is
also found in a separate section in one of these codices (Vat. Pal. lat.
220). These many points of intersection demonstrate amply why these
two homiletic collections are fundamental to our understanding both of
how apocrypha circulated, but also why so many are found in Insular
sources. The other feature which is striking is the limited geographical
area from which many of the manuscripts discussed ultimately derive,
with a particular emphasis on Southern Germany and Northern Italy,
perhaps best epitomised by the transfer of the Karlsruhe manuscript
from Northern Italy to Reichenau.63 The missionary activity of Irish
monks in that area in the early medieval period coupled with the abundance of apocryphal material circulating in Irish circles may help to
explain why so many of these texts are extant in homiletic form. It
seems that there were two principal ways in which apocrypha were
employed in homilies : either a fragment was cited and then elaborated
upon, much like a Biblical pericope, or a detail of apocryphal lore is
inserted into a sermon in the same fashion as an exemplum. We also
have details of apocryphal traditions cited in texts in a less systematic
62. OSULLIVAN, T., The Apocalypse of Thomas, in CAREY, et. al., op. cit.
63. For the missionary activity of Irish monks in Southern Germany in the
early medieval period see LOEWE, op. cit.



manner. The importance of homilies as sources for the transmission of

apocrypha in Insular realms has been discussed in detail by Old English scholars,64 but perhaps the issue needs more attention in an Irish
context, and also the relationship between the Irish material, Old English material and Latin texts would benefit from closer examination.
Such an investigation would aid our understanding of the dissemination
of this material, but it may not fully explain the choice of material.
Reasons for the abundance of apocryphal texts in medieval Insular circles can be suggested, and the dominance of the homiletic form can be
linked to missionary activity, but this does not necessarily account for
the dominance of eschatological themes, except perhaps the persuasive
force of a passionate hell-fire sermon. However, recent work by Bernard McGinn on the eschatological themes which dominated early Irish
Christianity in conjunction with a pervasive expectation of imminent
Doom may go some way to answering this question.65

64. See for example the work of WRIGHT, The Irish Tradition, op. cit. and
BIGGS, F., The Fourfold Division of Souls : The Old English Christ III and the
Insular Homiletic Tradition, Traditio 45, 1989-90, p. 69-85 ; Vercelli Homily
6 and the Apocryphal Gospel of Pseudo-Matthew, Notes and Queries 49, 2002,
p. 176-78, inter alia.
65. MCGINN, B., Medieval Visions of the End : The Irish Contribution, in
CAREY, et. al., op. cit.

Huygens Institute for the History of the Netherlands


Traditional philology has not yet found good solutions for the edition of texts that have loose boundaries and a fuzzy transmission
history, such as marginal texts. In this paper, it is explored how the
digital edition could perhaps help to find new strategies to tackle
their complexities.
La philologie classique na pas encore trouv de bonnes solutions
pour diter des textes qui sont fluides et ont une transmission obscure,
par exemple les textes inscrits dans les marges du manuscrit. Dans cet
article, on explore comment ldition numrique pourrait aider trouver de nouvelles stratgies pour faire face de telles difficults.
The digital scholarly edition is no longer a distant future : more
and more experiments are available on the World Wide Web, and they
often show a promising array of new possibilities for the research and
analysis of larger corpora of texts. The web makes it possible to share
resources and research, which has already boosted the creation of communities of scholars collaborating in online projects.1 Moreover, a digital presentation and exploration of medieval written sources free us
from the constraints of the traditional diplomatic or critical edition, in
which they often do not quite fit. In this paper I will explore the new
possibilities this might offer to students of medieval texts.
In order to do this, I start with the example of eLaborate, a digital
editorial tool developed at the Huygens Institute for the History of the
Netherlands (Huygens ING KNAW). The core business of Huygens
ING is the editing of sources relevant to the literary and socio-political
history of the Netherlands, ranging from the correspondence of Hugo
Grotius, the famous seventeenth-century diplomat and scholar who fun1. J. VAN ZUNDERT and P. BOOT, The Digital Edition 2.0 and the Digital
Library : Services, not Resources, in L. Spiro, Examples of Collaborative
Digital Humanities Projects, Houston 2009. Accessed online, January 2012,
http://digitalscholarship.wordpress.com/2009/06/01/examples-of-collaborative-digitalhumanities-projects/. G. CRANE, Give us editors ! Re-inventing the edition
and re-thinking humanities, in Online Humanities Scholarship : The Shape of
Things to Come, ed. by J. MCGANN, Houston 2010, p. 81-97. Accessed online,
January 2012, http://cnx.org/content/col11199/1.1/pdf.


Apocrypha 23, 2012, p. 155-168



damentally changed international law, to the diary of Anne Frank, the

iconic Jewish teenage girl who kept a diary during her days of hiding
from the Nazi regime. In its mission statement, Huygens ING explicitly states that it strives to give the digital humanities a primary role :
it wants to be at the fore front of the development of new editorial
and analytical tools for textual scholarship, and invests accordingly in
a comparatively large department of IT specialists working in the area
of humanities computing.2

eLaborates first prototypes sprang from the wish to collaborate in
online projects in order to create quick transcriptions of large volumes
of text. The first version of eLaborate was developed in 2003-2005, and
turned out as a digital tool that can be used to set up an online working
environment for the transcription of a manuscript or early-printed text.
Its lay-out is deliberately simple, offering the bare essentials for an
effective, collective transcription process, with only a minimal amount
of instruction or IT-knowledge.3 It allows a scholar to upload photos
or scans of a manuscript or printed book, and shows a set of windows underneath : one in which the user can make transcriptions, and
another in which he can make annotations. The environment is online,
hosted by a local server of the Huygens Institute, and accessible to anyone who is invited by the moderator of the project. This online accessibility is crucial, for it makes it possible to collaborate with a group
of researchers, perhaps because one is working on a set of material
which is too large to tackle on ones own, or, as was the case in my
own project, when one wishes to involve an international network of
scholars and to profit from its collective specialized knowledge, even
when they happen to be spread all over the world. Anyone with internet
access can be asked to contribute, even for a small contribution to the
2. See http://www.huygens.knaw.nl/en/over-ons/missiecorporate-story/ (accessed January 2012) : Huygens ING researches texts and sources from the past
with the aid of new methods and techniques. The Institute champions innovation in research methodology () researchers engaged in the humanities
collaborate closely with a completely different type of expert, such as specialists
in informatics, authorities in digital humanities, and a large team of software
developers, all under one roof on a daily basis. As such, we regard ourselves
as a humanities laboratory in which we develop, test and apply new methods in
order to extract more and different information from the sources than has been
possible until now.
3. K.H. VAN DALEN-OSKAM. Please elaborate !, Madoc 23, 2009, 1,
tradition and Web 2.0 : eLaborate as social experiment in humanities scholarship, in T. TAKEVA (ed.), Social Software and the Evolution of User Expertise :
Future Trends in Knowledge Creation and Dissemination, Hershey, Idea Group,
2013 (forthcoming).



project. Moreover, the whole community of scholars thus created can

immediately profit from the work done by its individual members. It is
possible to exchange questions, post comments, make observations, et
One of the first projects to experiment with eLaborate was the one
I coordinated myself, focussing on the oldest commentary tradition on
Martianus Capellas De nuptiis Philologiae et Mercurii. De nuptiis is a
late antique treatise on the seven liberal arts, dressed in the rich clothes
of ancient mythology. Two introductory books set the stage for the
allegory of the marriage between Mercury and Philology, Wisdom and
Knowledge, Language and Number, and then seven books for the seven
liberal arts (the three arts of language, and the four arts of number)
follow. This means that the making of an edition requires a profound
knowledge of Greek mythology, philosophy, Pythagoreanism, number
symbolism, Neoplatonic ideas, technical terminology in the fields of
the arts, and at the same time a deep familiarity with both the ancient
and the medieval learned tradition on the artes liberales, ranging from
grammar, rhetoric and logic, to geometry, arithmetic, astronomy and
music. The oldest commentary tradition dates back to the first half of
the ninth century, possibly incorporating even older layers. It is so large
and varied that since its discovery in the 1930s no scholar succeeded
in delivering a complete edition, even though several great scholars
played with the idea.
In order to overcome this problem I brought together a team of
editors with enough specialist knowledge to tackle the interpretative
problems. Together with the software developers of the Huygens ING,
I filled an online working environment with material. A single manuscript (Leiden, University Library, VLF 48), central to the transmission of the oldest commentary tradition on De nuptiis, was shown, page
by page. A full transcription of the main text of that manuscript was
added, and as a team we began to complement this transcription with
a full transcription of the marginal and interlinear annotations found in
that manuscript. For correction and comparison, full sets of photos of
four additional manuscripts were also made available in the working
environment. The working environment was secure, and each person
involved needed to register at Huygens ING and was issued a password. The project coordinator (me) decided who would have rights to
just read in specific parts of the environment, and who could write
and overwrite material in specific parts.
The team that was thus put together consisted of scholars who had
already studied parts of the commentary tradition or similar texts, and
published on them. In several cases they already had collected notes on
and transcriptions of part of the material, and used those in papers and
articles. However, since these notes and transcriptions were just loose
fragments of a large commentary tradition, they didnt have the right
framework in which to publish them. Thus the benefits of our collabo-



ration were mutual : the scholars involved contributed to my project,

and I offered them a framework in which to publish their study materials. Besides myself, the team included five scholars, and I involved two
students to check and correct the material. Sinad OSullivan, familiar
with commentary traditions because of her edition of a tenth-century
commentary tradition on Prudentius Psychomachia, started to work on
the edition of the commentary attached to the two introductory books.4
I provided first transcriptions of the glosses on most of the remaining
books on the arts, and involved the specialists in my network in different ways. Mary Garrison was involved in the transcription of the
glosses added to the books on the trivial arts : Grammar, Dialectic and
Rhetoric (books 3-5).5 Natalia Lozovsky had already explored Martianuss commentary tradition for her book on geographical knowledge
in the early medieval West, and used her experience to transcribe and
interpret the glosses added to book 6, on Geometry.6 Jean-Yves Guillaumin, a specialist in the ancient learned tradition on arithmetic, commented extensively on my initial transcription of the large amount of
commentary added to Martianuss book on arithmetic (book 7).7 Bruce
Eastwood, the great specialist on Carolingian astronomy, focused on
the book on astronomy (book 8).8 I myself took care of the edition of
the glosses added to the ninth and final book on Music.9 All these specialists thus contributed material to the transcription-environment : some
4. S. OSULLIVAN, Early Medieval Glosses on Prudentius Psychomachia,
The Weitz Tradition, Leiden-Boston, Brill, Mittellateinische Studien und Texte
31, 2004.
5. Mary Garrison published several articles on intellectual life in the Carolingian period, e.g. The Emergence of Carolingian Latin Literature and the
Court of Charlemagne, in R. MCKITTERICK (ed.), Carolingian Culture : emulation and innovation, Cambridge, CUP, 1994, p. 111-140 ; The social world of
Alcuin. Nicknames at York and at the Carolingian court, in L.A.J.R. HOUWEN
and A.A. MACDONALD (eds.), Alcuin of York, scholar at the Carolingian court,
Groningen, Egbert Forsten, Germania Latina 3, p. 59-79 ; Alcuin and Tibullus,
in M.C. DAZ Y DAZ and J.J. DAZ DE BUSTAMANTE (eds.), Poesa Latina Medieval (Siglos V-XV), Florence, SISMEL Editione del Galluzzo, 2005, p. 749-759.
6. N. LOZOVSKY, The Earth is Our Book. Geographical Knowledge in the
Latin West ca. 400-1000, Michigan, University of Michigan Press, 2000.
7. J.-Y. GUILLAUMIN (ed., transl., comm.), Martianus Capella, Les noces de
Philologie et de Mercure, Livre VII, LArithmtique, Paris, Les Belles Lettres,
8. Many articles by Bruce Eastwood deal with Martianus teaching on
astronomy and its reception in the Middle Ages, but the book that brings many
of his earlier findings together is Ordering the Heavens. Roman Astronomy and
Cosmology in the Carolingian Renaissance, Leiden, Boston, Brill, History of
Science and Medicine Library 4, Medieval and Early Modern Science 8, 2007.
9. I had already published an edition and study of the commentary tradition added to this book : M. TEEUWEN, Harmony and the Music of the Spheres.
The ars musica in ninth-century commentaries on Martianus Capella, LeidenBoston-Kln, Brill, Mittellateinische Studien und Texte 30, 2002.



provided transcriptions, others comments, or corrections, or interpretations and translations.

When all this work was done, and the edition more or less complete,
a new version of eLaborate was developed : eLaborate2. This included
an online publication platform for text-editions, accessible to anyone.
In the case of the Martianus Capella project, a lay-out in vertical panels was developed, with one panel for each set of data of the edition.
In standard view, there are four panels : 1. for navigation, 2. for the
photo (facsimile), 3. for the transcription of the main text, and 4. for
the transcription of the annotations. Since the splitting up of a normalsized computer screen into four horizontal slices results in very narrow
views of the information involved, this basic view can be customized
according to ones own wishes. Panels can be clicked shut, so that the
screen can be used to see more of the photo, or more of the transcription of the annotations, and so on. Its lay-out, in other words, is up
to the user to a certain extent. The site shows all the photos from the
central Leiden manuscript with zooming and navigating functionalities,
and a basic search-function has been developed. A search produces a
list of results which acts as a navigation-list, so that one can click from
one result to another. It also has the clever option of fuzzy search,
with which one can easily work around the problem of orthographical variation, declension and case-endings. It is thus possible to search
the whole commentary tradition for a name (Pythagoras), or a concept
(sapientia), and the result is a fine list of material on the subject at
The first result was thus a first complete edition of the oldest commentary tradition of De nuptiis : http://martianus.huygens.knaw.nl,
published in November 2008, and regularly updated with corrections
and new material since then. A second, equally important result was
a symposium organised towards the end of the project, in which the
network gave an account of what they found when working with the
material. The papers of this symposium have been published.10 Moreover, Sinad published a cumulative edition of the commentary on the
first two books of De nuptiis,11 and I plan to make the book-edition
complete with another volume in the series, containing an edition of
the commentary added to books 3-9, but more or less restricted to the
Leiden VLF 48 manuscript.

10. M. TEEUWEN and S. OSULLIVAN (eds.), Carolingian Scholarship and

Martianus Capella : Ninth-Century Commentary Traditions on De nuptiis in
Context, Turnhout, Brepols, CELAMA 12, 2011.
11. Glossae Aevi Carolini in Libros I-II Martiani Capellae De nuptiis Philologiae et Mercurii, ed. by S. OSullivan, Turnhout, Brepols, CCCM 237, 2010.



Flexible text and variation

Apart from the content of the commentary, which requires a wide
array of specialist input, there is another aspect to this material which
makes the making of an edition very complicated. It would be stating
the obvious to say that no two manuscripts offer the same text, that
there is always a certain amount of variation, through scribal errors, the
use of different abbreviations or orthographic practices ; however, in the
case of individual glosses and commentary traditions, the truth of this
statement gains a whole new dimension. The contours of a commentary
text are vague, and the status of a commentary text is fundamentally
different from that of a regular text. The copyists have much more
freedom to leave glosses out, to change them, or to add new ones.
They even feel compelled, so it seems, to complement an earlier layer
of annotations with a new one, if a manuscript which has new material comes into their hands. A commentary text was never seen as a
finished, closed text : one could and should always add new insights,
newly found glosses from other manuscripts, or newly composed annotations. Layers of annotations are added to the main text, and these are
moulded together to form new commentary traditions. Variation is not
an accident, but an essence.12
How to edit these flexible texts is a problem to which the scholarship of editing has not found answers yet. It may be clear that a traditional critical edition, based on a stemma which points out the oldest
or most reliable manuscripts, is impossible. Even if one would want
to approach a commentary text in such a way, then in most cases the
apparatus would grow disproportionately large, drowning its student in
endless lists of variants. But at the same time the comparison of manuscripts and the study of variants is hardly ever more useful than in the
case of commentary texts. Often commentary texts are characterized by
a certain lack of care, a certain disarray. The text is often cramped into
a space which is too small, written in a hand which is tiny, in a manner which is not preoccupied with grammar and style. It is not difficult
to find examples of glosses which have been so mangled in the trans-

12. During the Apocrypha-meeting it became clear that this statement is not
only true for commentary texts, but in fact for many others as well : collections
of saints stories, liturgical text collections or sermons, for example, are as prone
to variation as commentary texts. Yet marginal commentary does have a unique
aspect of inviting others to add more, new or contrary material, which the
project I am currently running, Marginal Scholarship : the Practice of Learning in the Early Middle Ages (c 800-c 1000), wishes to explore. See Marginal Scholarship : Rethinking the Function of Latin Glosses in Early Medieval
Manuscripts, in Rethinking and Recontextualizing Glosses. New Perspectives in
the Study of Late Anglo-Saxon Glossography, ed. by P. LENDINARA, L. LAZZARI,
C. Di SCIACCA,, Porto, FIDEM, TEMA 54, 2011, p. 19-37.



mission process that they lost their meaning entirely.13 Careful comparison with other, better sources is often the only way to reconstruct
their meaning and their error. Moreover, an analysis of variants offers
a good starting point for grasping the relations among manuscripts, and
the contours of the commentary tradition as a whole. There are glosses,
for example, which clearly form part of a corpus of glosses, and there
are individual ones, unique for a single manuscript. How would one be
able to tell the difference, if not by studying multiple manuscripts of
the same text with their marginalia ?
So, even if we (that is, we scholars who are interested in aspects of
medieval texts that traditional textual edition methods cannot accommodate) do not want to phrase our observations in Lachmannian terms of
originality, sincerity, error and corruption,14 we still want to use
the method of text comparison in order to gain insight in our texts. But
is it even possible to make an edition of a text which has such vague
contours ? How can we edit these texts in such a way that we do justice
both to the individual sources, and to the aspect of comparison between
sources, relations between them and overlap between them ? Preparing
a traditional critical edition for these flexible, fuzzy texts, it seems to
me, is not only barely possible but also undesirable in some respects.
What is interesting about these texts is not always their content, or the
solution of their philological problems. Sometimes, the more interesting questions are the ones that concern the shape of a text at a certain
point in time, or in a certain manuscript used by a certain community,
or its place in the setting of a collection of texts, or the mapping of
the changes a certain text underwent in the course of its transmission.
The glosses, attached to the main text in layers, more or less loosely
assembled into set corpora but always open for improvement and additions, are an excellent point in case.
Again, the digital environment offers new possibilities here as
was signalled already twenty years ago by the French textual scholar
Bernard Cerquiglini in his Praise of the Variant.15 In eLaborate 1 and
13. M. LAPIDGE, The Study of Latin Texts in Late Anglo-Saxon England :
the Evidence of Latin Glosses, in N. Brooks (ed.), Latin and the Vernacular
Languages in Early Medieval Britain, Leicester, University Press, 1982, 99-140 ;
J. ZETZEL, Marginal Scholarship and Textual Deviance : The Commentum Cornuti and the Early Scholia on Persius, London, Institute of Classical Studies,
Bulletin of the Institute of Classical Studies Supplement 84, 2005, 144-161.
14. ZETZEL, Marginal Scholarship, 161 : The assumptions of Lachmannite
textual criticism are not wrong, but they are a part of the culture that produced
them Christian, romantic, industrial while the ancient books we read as
well as the medieval copies in which we read them came from a very different
15. B. CERQUILIGNI, In Praise of the Variant. A Critical History of Philology, transl. B. Wing, Baltimore, John Hopkins University Press 1999. Originally
published as loge de la variante : Histoire critique de la philologie, Paris, ditions du Seuil, 1989.



2, it was not possible to work with multiple manuscripts. I chose, for

practical reasons, to work with a single, central manuscript. Three other
manuscripts and editions of other commentary traditions were used to
complete and correct the readings of this single manuscript, and to create a critical apparatus for them. The result is a hybrid text which is
neither a proper critical edition, nor a bare diplomatic transcription. We
could perhaps term it a critical transcription. The ideal would be for
a digital environment to show the user multiple manuscripts, and to
allow the user to set them next to each other to his own liking, and not
only according to the ranking and order chosen by the philologist. In
this comparison of sources, the visualisation of similarity and deviance
would be crucial.
Experiments with this kind of textual editing have in fact been
carried out at Huygens ING for a project involving an edition of the
complete works of Willem Frederik Hermans, a modern Dutch literary
author, famous for his continuous fiddling with his text for each print
run. An electronic edition of one of Hermans short stories, Paranoia,
was made, in which all the variants, encoded with a programme specifically developed for text comparison, are shown.16 The programme
Collate, developed by Peter Robinson for his work on the Canterbury
Tales, was used to encode more or less automatically the variation.17
This was converted to a XML-TEI file, and this file was imported in
the eLaborate edition-tool, so that it could be structured in the architecture of eLaborate. It makes an interlinear presentation of variants
possible, so that all the variants from all the different prints can be
viewed simultaneously. The user of the digital edition can thus choose,
for example, the first edition, and can still see all the variants from the
second, third, and later editions, just by ticking a single box. It is possible to switch views by choosing a different edition as ones central
Of course, variants in modern printed books are a different matter
than variants in hand-written codices, and the kind of variation medievalists encounter is certainly very different from what the Hermansscholars may encounter in a print run. Still, along similar lines, it will
be possible to show multiple manuscripts, multiple texts, and their relations to one another in a more dynamic way than the traditional critical
16. Online, http://hermansdigitaal.huygensinstituut.nl/path, unfortunately in
a secure environment which is password accessible only (because of copyright issues which have to be settled with the inheritors of the W.F. Hermans
material) : Paranoia, Willem Frederik Hermans, electronic edition. Ed. by
J. GIELKENS, P. KEGEL, with assistance of E. KAMP, B. OOSTVEEN and M. VAN
ZOGGEL, and the software developers of the Huygens ING. First published
online in 2009, last visited on 17 February 2012.
17. CollateX, the successor of Collate, is in development at Huygens ING
and will be used in future editions of Hermans. For more information, see the
description at http://collatex.sourceforge.net/ (visited 17 February 2012).



edition apparatus allows, possibly also in a clearer way. The traditional

critical apparatus, moreover, was not designed to show the relations
among the sources, but rather to account for the choices made by the
philologist, or to provide a proof of an editors work.18 It hides the
differences rather than pointing them out, by measuring them against
the yard stick of proper Latin, or a pure text. If presented differently, however, the apparatus has the potential to show the overlap and
the deviance among different sources, and then we could give it a more
satisfactory place in the whole of an edition. And if we combine the
new ideas about presentation with the ideas of flexibility and potential
customization of the material for individual readers, we could produce
a much more informative apparatus which to my mind is vital to
our understanding of text and context.

The nature of glosses

Another important observation about the form of an edition of marginal texts is prompted by the nature of these texts. To grasp the flexibility, the variation and the overlap of commentary traditions or corpora
of glosses, it is not enough to analyse the relations among several manuscripts of a single text tradition. Rather, each gloss should be viewed
in the light of a much larger cluster of texts. The oldest commentary
tradition on De nuptiis will serve as a good example to explain this.
All too often manuscripts with glosses are easily classified as school
texts,19 but in the case of Martianus Capellas De nuptiis, the classification does not fit so well. De nuptiis is not a text which one would
expect in the curriculum of a monastic school. First, the text is full of
subjects which are problematic in a Christian context : Greek myths,
full of love, lust and adultery, ideas about the creation which do not
concur with Christian doctrine, about life after death, or about fortune
telling.20 Second, Martianus Latin is not the kind of Latin one would
try on average students. It has been characterized as the most difficult
Latin there is, full of neologisms, odd grammatical structures, Graecisms, and exploring as many poetic meters as one can possibly do in
one work (and then a few more). All these aspects, which make the
text so excessive in modern eyes, were attractions in the eyes of our
Carolingian scholars, but it must have been read only by the brightest
18. Anthony Grafton makes this point about footnotes, but it fits the critical
apparatus very well. See A. GRAFTON, The Footnote. A Curious History, Cambridge, MA, Harvard University Press, 1999 (first published in 1997), p. 15-16.
19. G. WIELAND, The Glossed Manuscript : Classbook or Library Book ?,
Anglo-Saxon England 14, 1985, p. 153-173.
20. M. TEEUWEN, Seduced by Pagan Poets and Philosophers. Suspicious
Learning in the Early Middle Ages, in C. GILIBERTO and L. TERESI (eds.), The
Limits of Learning, Paris, Louvain and Dudley, MA, Peeters, Storehouses of
Wholesome Learning III, Medievalia Groningana, forthcoming.



minds, at the highest level of study.21 A third point is the lay-out of the
manuscripts themselves. These do not at all strike us as books made
for the schoolroom. Glosses tumble over each other ; the seams of the
lay-out are coming apart because of the overload of annotations. Both
in lay-out and in content, these texts do not come across as suitable for
teaching purposes because the annotations are very often highly associative. A schoolmaster would need oceans of time just to read one
page with students in a structured way, even if he addressed only one
or two students at a time.22 Fourth, some of the annotations are written
in Tironian notes, the medieval shorthand system. These notes not only
suggest that these texts are perhaps personal, and unfinished (still to be
revised), but they could also indicate, perhaps, a deliberate screening
off of material that was meant to be read only by the best educated
Putting all these observations together, I suggested that these manuscripts of Martianus Capellas De nuptiis, with their extensive glossing should not be characterized as schoolbooks, but rather as scholarly
works books used for scholarship.24 Time and again, the glosses show
how De nuptiis was used by the Carolingian readers as a stepping stone
to other authorities. When they explain something about the nature of
the number three, they step to Augustines De doctrina Christiana, or
to the last of his six books On music. When the theme of the harmony
of the spheres, the music of planets, surfaces in Martianuss text, they
bring in Chalcidius and Macrobius, and their views on Platos Timaeus
and Ciceros Dream of Scipio.25 In this way a new, medieval encyclopedia of learning was created, wrapped around the late-ancient one, and
this new layer of learning deeply influenced medieval learned traditions.
This new collection is fed by all kinds of texts (Augustine, Boethius,
Macrobius, Pliny, Isidore, etcetera), and in turn it feeds commentary
traditions on other texts for example Prudentiuss Psychomachia, or
21. M. TEEUWEN, The Pursuit of Secular Learning. The Oldest Commentary Tradition on Martianus Capella, Journal of Medieval Latin 18, 2008,
p. 36-51, esp. p. 42-43 ; S. OSULLIVAN, Obscurity, Pagan Lore, and Secrecy
in Glosses on Books I-II from the Oldest Gloss Tradition, in M. TEEUWEN and
S. OSULLIVAN, Carolingian Scholarship, p. 99-120.
22. M. TEEUWEN, Glossing in Close Co-operation : Examples from NinthCentury Martianus Capella Manuscripts, in R.H. BREMMER Jr. and K. DEKKER
(eds.), Practice in Learning : The Transfer of Encyclopaedic Learning in the
Early Middle Ages, Paris, Leuven and Walpole, MA, Peeters, Storehouses of
Wholesome Learning II, Medievalia Groningana, 2010, p. 85-100.
23. D. GANZ, On the History of Tironian Notes, in P. GANZ (ed.), Tironische Noten, Wiesbaden, Harrassowitz, Wolfenbtteler Mittelalter-Studien 1,
1990, p. 35-51.
24. M. TEEUWEN, Writing between the Lines : Reflections of Scholarly
Debate in a Carolingian Commentary Tradition, in Carolingian Scholarship,
p. 11-34.
25. Ibid., p. 22-28.



something as far removed from ancient myth as Arators Historia Apostolica.26 They use the same sources, they use each other as sources,
and they react upon each other. In order to get a grip on what really
happens in these margins, we should try to map these connections to
other texts. We would have to create a carefully chosen corpus of texts,
which would give us the contours of the backbone of learning of the
Carolingian learned elite. In that corpus we would have to include, for
example, the early medieval glossaries, and the overlap between these
and commentary texts would have to be analysed. This will become
possible, of course, only when more editions are available to us, preferably in easily searchable formats that allow for complex comparisons
and the detection of parallels.27

Future horizons
The technique with which we worked to shape the edition of the
oldest commentary tradition is not powerful enough to enable such
complex undertakings. eLaborate in its present shape cannot be used
for text comparison in that sense, or for the drafting of complicated relationships among texts. But the technologies that could help us address
these questions are already out there in the digital world.28 The basic
workings of programmes that screen theses on plagiarism, for example,
could perhaps be used to find relationships and parallels between texts.
This may not be as far-fetched as it sounds ; programmes have been
developed to map relations among manuscripts using computer techniques from evolutionary biology and exploiting the analogy between
the transmission of manuscripts and the transmission of genetic information and the development of species.29

26. Ibid., p. 28-31.

27. M. TEEUWEN, The Impossible Task of Editing a Ninth-Century Commentary. The Case of Martianus Capella, Variants. Annual Journal for the
European Society of Textual Scholarship 6, 2007, ed. W. VAN MIERLO, p. 191208.
28. On the idea that the digital edition of the future will be using data
sources and functionalities from different origins, see J. VAN ZUNDERT and
P. BOOT, The Digital Edition 2.0.
29. P. ROBINSON and R.J. OHARA, Report on the textual criticism challenge, Bryn Mawr Classical Review 3.4, 1992, p. 331-337 ; B. SALEMANS, Building stemmas with the computer in a Cladistic, Neo-Lachmannian, way : the case
of fourteen text versions of Lanseloet van Denemerken, diss. 2000, http://www.
neder-l.nl/salemans/diss/salemans-diss-2000.pdf ; C. MAC, P. BARET, A. BOZZI
and L. CIGNONI (eds.), The Evolution of Texts : Confronting Stemmatological and
Genetical Methods, Pisa, Roma : Istituti Editoriali e Poligrafici Internazionali,
Linguistica computazionale XXIV-XXV, 2006 ; T. ROOS, T. HEIKKIL, Evaluating methods for computer-assisted stemmatology using artificial benchmark data
sets, Literary and Linguistic Computing 24.4, 2009, p. 417-433.



At present, the Huygens Institute is developing a new version of

eLaborate (eLaborate 3), and one of its aims is to bridge the gap
between the diplomatic transcription and the critical edition. The lines
along which eLaborate is adapted are dictated by the notion that the
material in the edition can be enriched with structured metadata, at different levels of the edition. An experiment with categories of annotations has already been carried out, for example in the digital edition of the Arthurian romance Walewein ende Keye, published online
by Marjolein Hogenbirk in 2009.30 In this edition, three categories of
annotations have been distinguished : 1. (green) annotations concerning
codicological and/or paleographical remarks, 2. (pink) annotations on
vocabulary and language, and 3. (lilac) observations of a literary nature,
or sketching a (literary-)historical background. The splitting up of the
mass of observations into these three categories makes it possible for
the visitor of the online edition to see the text according to his own
interests : he can read the text with some help in terms of language and
translation, and select a view in which only the annotations of category
2 are visible. He can also sift out all observations on codicology and
paleography, to figure out where the text stands in that respect, and perhaps to match these data to a different manuscript on which he himself
is working.
In eLaborate 3, this basic functionality of creating the possibility
to sort through layers of material in the digital edition of a historical
document by assigning them different categories, is further developed.31
The scholar will be able to add metadata to the project in general, to
the objects (transcriptions, facsimiles) within the project, and to the
annotations on these objects. In this way, it will be possible to create
a more complex web of objects that are linked to each other in some
relational way : the codicological and paleographical data of one object
could be linked to those of another one, Latin texts can be related to
vernacular translations, main texts can be related to commentaries or
interpretations, or to sources or parallel texts. The more complex the
mark-up, the more different kinds of representations can be generated
in the end, and the more connections and comparisons between the
material at hand can be made. But while complexity is the key here, it
will also be one of the pitfalls of the (future) digital edition : the more
complex the material is, the harder web designers will have to work
to reduce it so as to avoid visual overload without losing efficiency.
Types of metadata can be shown or hidden as a group, according to the
wishes of the scholar using the edition. As long as they are recorded,
30. M. HOGENBIRK (ed.), in cooperation with W.P. GERRITSEN, Walewein ende
Keye. Een dertiende-eeuwse Arturroman overgeleverd in de Lancelotcompilatie,
first published online in November 2009, at http://www.waleweinendekeye.huygens.knaw.nl (last accessed on 23 February 2012).
31. www.elaborate.huygens.knaw.nl (last accessed on 23 February 2012).



however, they are available for a further pursuit of research on aspects

of the edition, or to compare them to other editions or to subject them
to specific searches.32
A problematic side to all these promising panoramic views into
the future is the reality of the work floor. At Huygens ING, we are
blessed with a particularly large department of software developers, but
even they are always struggling with an overload of work. Also from
the technical point of view, the problems are large all the photos
are saved on our own servers, and this creates capacity problems and
financial problems, for contrary to what one might think digital space
is not without cost. It has become urgent, therefore, to make arrangements with image providers and agree on a protocol that would allow
a scholar to display photos from a different server, instead of saving
them on his own. More and more libraries have websites with databases of photos from their special collections the e-codices project is
an exemplary model.33 But in order to accomplish this, deals have to be
made with manuscript holding institutions and universities or research
institutions. These are a matter of years rather than months. Nevertheless, further possibilities for cooperation have to be explored, and now
is the time, so it seems, as there are indeed a number of initiatives in
that field : they create European online collections of manuscripts,34 and
they explore, in a European collective, the possibilities for cooperation
in preparing digital tools to facilitate text editions, text comparison, and
text analysis.35 These initiatives will perhaps offer good opportunities
32. K. VAN DALEN-OSKAM, All in one, one for all ? A possible world of
digital editions, unpublished lecture held at the LECTIO Round Table Series in
Leuven, 21 November 2011 : Digital or critical/Digital and critical ?.
33. See the website e-codices. Virtual Manuscript Library of Switzerland
(http://www.e-codices.unifr.ch/en), but also, for example, the recent project to
reconstruct the medieval library of Lorsch online : http://www.bibliotheca-laureshamensis-digital.de/de/index.html, For a good inventory of similar projects, see
Albrecht Diems site Monastic Studies Resources Internet Resources for the
Study of Early Medieval Monasticism, http://albrechtdiem.org/research/mmp/
listoflinks.html#Digital (last accessed on 23 February 2012).
34. See, for example, the project Europeana Regia (http://www.europeanaregia.eu/en, last accessed 23 February 2012). The project runs from January 2010
to June 2012, and its goal is to digitise 874 rare and precious manuscripts from
the Middle Ages and the Renaissance, with the collaboration of five major
libraries : the Bibliothque nationale de France, the Bayerische Staatsbibliothek,
the Historic Library of the University of Valencia, the Herzog August Bibliothek in Wolfenbttel, and the Royal Library of Belgium. The project focuses on
three distinct periods in history : the Bibliotheca Carolina (8th and 9th centuries),
the Library of Charles V and Family (14th century) and the Library of the Aragonese Kings of Naples (15th and 16th centuries). These manuscripts will be
fully accessible on the websites of the partner libraries.
35. See, for example, the project Interedition (http://www.interedition.eu/,
last accessed 23 February 2012). This COST project (coordinated by Joris van
Zundert, funded by the ESF from 2008-2012) aims at promoting the interoper-



for the philologists of the 21st century, but they demand a rather drastic
shift in the organisation of our work. Editors of historical texts used
to work alone, bent over some manuscript in some remote library, as
modern monks doing the slaves work of transcription. Now our projects have blossomed into co-operative ventures, involving not only several scholars, but also software developers and technical staffs. It is
not always easy to communicate our research questions to complete
outsiders in order to make them understand what technology we need,
what will help us and what stands in our way. To take such a road is
not like taking the highway to success, but rather to travel along muddy
off-road routes at a discouragingly slow speed. And time is of essence
when one is living on external funding with just a few years to show
There are, in fact, more downsides to digital editions than I care
to mention here : a digital edition dooms someone to a lifetime of
work, just to keep it up and running ; it is more vulnerable than one
would perhaps think, and will easily die from neglect ; the digital world
changes so fast that one would not only need a lot of time but also a
considerable amount of money just to keep up ; a digital publication
is still not acknowledged as a proper edition, even though its usefulness clearly exceeds that of a paper edition, and therefore it does not
fully count in ones publication list. However, I do not see how we can
choose not to take this digital road when our texts are what they are :
flexible texts, with a material reality that tells us just as much about
our history and culture as about the meaning of their words, structured
not along the lines of the modern day book, but often along completely
different lines that reflect their genesis, their transmission history and
their function in medieval intellectual life. These aspects were hidden
by traditional textual criticism in apparatus and foreword in Latin, but
the digital approach offers new opportunities to explore these aspects of
medieval texts. I suggest we rise to the occasion.

ability of the tools that are developed in the field of digital scholarly editing
and research.





Theodore Bar Konais paragraphs about the Kantaeans are of great
importance for the history of Mandaeans. The complex historiographical construct where various sects are mentionned and many grammatical forms deviant from literary Syriac make their interpretation
difficult. The analysis of the authors polemical intent and quotations
shows that Theodore assembles two different text corpora within a
linguistically inconsistent montage and uses stereotyped heresiological fictions (an anti-Samaritan Kuthaean narrative and an anti-Baptist Dosithean one) in order to write the history of the only group he
knows : the Mandaeans.
Les notices de Thodore Bar Kona sur les kantens sont une source
de premire importance pour lhistoire des mandens. La complexit de
la construction historiographique de lauteur, qui met en scne plusieurs
sectes diffrentes, et lobscurit de certaines formes grammaticales
loignes du syriaque littraire rendent cependant leur interprtation
difficile. La dconstruction de lintention polmique de lhrsiologue
et un rexamen de ses citations permettent de montrer que lauteur
combine, dans un montage incohrent dun point de vue linguistique,
deux corpus distincts et se nourrit de fictions hrsiologiques strotypes (un rcit kuten anti-samaritain et un rcit dosithen anti-baptiste)
pour crire lhistoire du seul groupe quil connat : les mandens.

Le Livre des scolies de Thodore Bar Kona et son apport

lhistoire des religions
Il revient Henri Pognon, dans sa monographie sur les coupes
magiques mandennes1, davoir fait connatre le Livre des scolies de
Thodore Bar Kona. Pognon a si bien peru limportance de louvrage
pour lhistoire des religions le manichisme et le mandisme, notamment que trois appendices de son livre y sont spcialement consa1. Henri POGNON, Inscriptions mandates des coupes de Khouabir, Paris,
1898 (cit ci-aprs Pognon).


Apocrypha 23, 2012, p. 171-207



crs (p. 105-255), avec une premire dition de larges extraits du texte,
accompagns dune traduction franaise et dun relev des parallles
avec la littrature mandenne. Cest, du reste, sur ce travail que sappuient les traducteurs du CSCO, ngligeant les corrections proposes
par Nldeke dans sa recension de louvrage de Pognon2. Par la suite,
le Livre des scolies, transmis par des manuscrits qui se rpartissent en
deux familles (recension de Sert et recension dUrmiah), a fait lobjet
dun travail ddition prsent complet : dition du texte syriaque de la
recension de Sert (CSCO 55, 1910 et CSCO 69, 1912) par Adda Scher,
enrichie par le complment lappareil de Scher (CSCO 431, 1981,
p. 23-54) ; traduction franaise de la recension de Sert (CSCO 431,
1981 et CSCO 432, 1982) par Robert Hespel et Ren Draguet ; dition
des passages diffrents dans la recension dUrmiah (CSCO 447, 1983)
et leur traduction franaise (CSCO 448, 1983) par Robert Hespel3.
Lauteur, Thodore Bar Kona4, fut vraisemblablement, la fin du
VIIIe sicle, vque syro-oriental de Kakar, ville situe sur un bras du
Tigre aujourdhui assch, en face de la ville musulmane dal-Wsi5,
cest--dire au nord de la rgion des Marais (al-Baa ou al-Bai)
habite par les mandens6. Lditeur du CSCO 431 ne retient de labondante discussion autour de lauteur que la date dachvement de luvre
prcise dans lexplicit du chapitre IX : 1103 de lan des Grecs, cest-dire 791/792. Lditeur mentionne que lattribution des chapitres X et
XI a t conteste en raison de la place de cet explicit ; mais dans lhypothse o ces deux derniers chapitres auraient t ajouts la composition de Thodore Bar Kona, force est de constater que cet ajout a
eu lieu relativement tt puisque le manuscrit le plus ancien conserv,
qui date du Xe ou du XIe sicle7, les contient dj. Hespel donne un bon
rsum de la composition de louvrage : Les Scolies sont en somme
un expos de la foi chrtienne en onze Mimr sous la forme dun grand
catchisme procdant par questions et rponses. Les Mimr I-V ont trait
lAncien Testament et les Mimr VI-IX au Nouveau. [] Le Mimr X
2. Theodor NLDEKE, H. Pognon, Inscriptions mandates des coupes de
Khouabir , Wiener Zeitschrift fr die Kunde des Morgenlandes 12, Vienne,
1898, p. 141-147, 353-361 (cit ci-aprs Nldeke).
3. Les rfrences au texte syriaque renverront au CSCO 69 ; pour la traduction franaise, voir lannexe.
4. Cette vocalisation parat tre la plus conforme lonomastique syriaque.
Sur Thodore, voir Anton BAUMSTARK, Geschichte der syrischen Literatur,
Bonn, 1922, p. 218-219 (cit ci-aprs Baumstark).
5. Voir M. STRECK et Jakob LASSNER, Kaskar , Encyclopdie de lIslam,
t. 4, Leiden/Paris, 1978, p. 753-754, et Mondher SAKLY, Robert DARLEY-DORAN,
Wsi , EI2, t. 11, 2005, p. 180-186.
6. Et, peut-tre jusquau Xe s., par les elkasates si lon suit la notice du
Fihrist dIbn al-Nadm qui leur est consacre, voir Gustav FLGEL, Mani, seine
Lehre und seine Schriften, Leipzig, 1862, texte arabe p. 133, trad. allemande
p. 133-134.
7. Manuscrit Syr. 366 de la Bibliothque Nationale, voir CSCO 431, p. 7.



est une longue discussion sur la foi chrtienne avec un paen, dont
le point de vue exprime en ralit celui de lIslam. Le Mimra XI enfin
fait un expos plus ou moins tendu des diffrentes sectes ou hrsies
apparues en Grce, en Chalde et en Perse8.
Les notices 84 87 consacres aux mandens se situent entre un
dveloppement sur lhistoire des hrsies dun point de vue syro-oriental ( 80-83 contre Cyrille dAlexandrie, Eutychs de Constantinople,
Svre dAntioche et Julien dHalicarnasse) et des notices consacres
des vestiges cultuels plus qu des mouvements religieux constitus
( 88-94), savoir des noms de divinits (Ramt, 88 ; Dan,
90 ; Nanay, 93 ; Barq des Gouznens, 94), des pratiques de mendicit (Bim, 89), ou encore des repas rituels (Banquet des eaux,
92 ; Banquet des morts, 92). En sappuyant prcisment sur les
titres qui structurent les notices et que les diteurs du CSCO ont ngligs, ainsi que sur les effets de symtrie observables dans les notices 84
87, D. Kruisheer9 a montr lunit de cet ensemble en en dgageant
la structure suivante :
1 : Sur les kantens ( 84 du CSCO 432, p. 255-256). Le propos gnral de lauteur, suggr par la premire phrase, est de rfuter lattribution par les kantens de leur doctrine Abel, rfutation qui
ne sera effectue quau 6. Ce paragraphe, pour lheure, retrace une
histoire en deux tapes qui constitue une antithse parfaite lhistoire
sainte :
les prtres de Dagon, appels prtres de Goliath, accomplissaient
chaque anne un rite pour commmorer la mort de leur hros autour de
sa statue quils renversaient ;
aprs la dportation par Nabuchodonosor des Philistins Babylone, la statue de Goliath est brise et les prtres de Dagon recrent un
rite autour de la mort de Goliath.
La fin du paragraphe donne deux dtails : lappellation de ce premier groupe par les Chaldens (secte de Nergal , qui fait donc le lien
avec le 6 sur les nrigens ), lexistence inchange de ce groupe
jusquau roi Yazdgerd (439-45710) et sa rforme sous limpulsion de
Baay lpoque du roi Prz (459-484).
2 : Sur Baay, do il est [issu] (premire partie du 85 du
CSCO, p. 256). Lauteur dveloppe le deuxime temps de son histoire
de la secte : Baay, esclave du chef de la secte prcdemment dcrite
(Papa, fils des Klly de Gawkay), rforme la secte en reprenant de la
religion des mages ladoration des luminaires et le culte du feu, ainsi
quen empruntant un pseudonyme perse (Yazdaniz), pour viter quelle
8. CSCO 431, p. 2-3.
9. Dirk KRUISHEER, Theodore Bar Konis Ketb d-eskolyon as a Source
for the Study of Early Mandaeism , Jaarbericht van het vooraziatisch-egyptisch
genootschap 33, Leiden, 1995, p. 151-169 (cit ci-aprs Kruisheer).
10. Il sagit de Yazdgerd II, puisque celui-ci prcde le roi Prz.



ne fasse lobjet de rpression de leur part. Lauteur numre les autres

traits disparates qui la caractrisent : des traits juifs (interdit alimentaire
du porc, nom divin tir du Pentateuque) et chrtiens (signe de la croix).
3 : Un peu de sa doctrine (seconde partie du 85 du CSCO,
p. 256-257). Sont voqus, avec le plus souvent des citations, dabord
la cosmogonie de la secte, puis le nom des dix cieux, une offrande
au jardin dAdam et enfin un texte sur le Fils de la lumire , cette
dernire pice se retrouvant presque intgralement dans le Ginza rba
4 : Hrsie des dostens quenseigna le mendiant Ado (premire
partie du 86 du CSCO, p. 257-258). Lauteur y fait la biographie du
fondateur de la secte, issu dune famille de mendiants dAdiabne installs aux bords du fleuve lay auprs dun autre Papa (fils de Tinis)
qui lui construit un abri o se rassemblent dautres mendiants. Aprs
avoir numr les diffrentes appellations locales de la secte, dont celle
de mandens , lauteur affirme quelle a emprunt aux marcionites,
aux manichens et aux kantens, ce qui constitue le seul lien explicite
avec la notice sur les kantens.
5 : Un peu de leur doctrine (seconde partie du 86 du CSCO,
p. 258-259). La cosmogonie et leschatologie de la secte sont voques
par une srie de citations mettant en scne les figures mandennes
dAbtr et de Ptael. Puis un texte intitul Acheminement vers les
magiciennes est cit pour expliquer lorigine de leur baptme. Enfin,
deux figures apparaissant dans le Ginza sont mentionnes (le scribe
Dnns et le petit D).
6 : Hrsie des nrigens ( 87 du CSCO, p. 259). Ce paragraphe a pour fonction de donner la rfutation, promise au premier
paragraphe, de la revendication par les kantens de la filiation dAbel :
lauteur imagine que les descendants de Can auraient rendu celui-ci
un culte sous le nom de Nrig, et que les descendants de Seth, de leur
ct, auraient conu pour Abel un tombeau quils auraient appel kant,
mot qui aurait donn son nom la secte des kantens.

Quel crdit apporter aux notices de Thodore Bar Kona sur

les mandens ?
Les critiques modernes, au premier rang desquels Pognon, ont commenc par se poser la question du rapport entre les textes cits par
Thodore Bar Kona et les textes mandens qui nous sont parvenus,
dans lespoir de donner des lments de rponse la question mandenne , cest--dire la question de lorigine iranienne (avec, au
tournant du XXe sicle, une confusion encore assez grande entre zoroastrisme et manichisme), juive, chrtienne ou seulement baptiste des
mandens, et, au moins, de donner quelques lumires sur le milieu de
naissance du mandisme.



Lanalyse de lintention hrsiologique qui a prsid lagencement

de ces citations a mis plus de temps se faire et est loin de faire lobjet dun consensus. Deux positions se font jour : lune, reprsente par
H. H. Schaeder11 et appele minimaliste par son principal dtracteur,
met en doute la construction historique de Thodore et refuse daccorder le moindre crdit lexistence dune secte kantenne ayant prcd les mandens ; lautre, que nous appellerions volontiers maximaliste, reprsente avant tout par W. Madelung12, sappuie sur les sources
arabes pour affirmer lexistence dau moins une des deux sectes. vrai
dire, une solution intermdiaire fait petit petit son chemin, qui met
au jour, dune part, une matire baptiste (elkasate) quon observe dans
la littrature syro-orientale13 et qui a pu galement servir de source
Thodore et, dautre part, une hrsiologie mandenne qui mentionne
des sectes proches du portrait que lauteur dresse des kantens14.
Il semble ncessaire de dgager du texte du chapitre XI les biais
propres lcriture hrsiologique avant dexploiter le document pour
lhistoire du mandisme et des ventuelles sectes proches de ce mouvement. Or lhrsiologie nest pas seulement un biais mais un genre littraire qui sappuie, ce titre, sur une histoire littraire. Cest pourquoi
il nous parat pertinent de tenter, en partant du texte syriaque lui-mme,
de distinguer les diffrentes sources dont dpend Thodore Bar Kona,
savoir aussi bien les sources internes aux sectes en sa possession que
les sources hrsiologiques, afin de comprendre comment lauteur procde au montage de ces diffrentes sources et surtout quelle fin. Ce
nest quune fois lintention hrsiologique de lauteur mise au jour que
son propos sur lhistoire de ces sectes pourra tre valu.

tude des citations de Thodore Bar Kona

Avant dtudier les citations elles-mmes, indiques dans lannexe
par les lettres A F, il nous parat indispensable dexaminer de prs
la structure des deux paragraphes o elles apparaissent ( 3 et 5 de
la numrotation de D. Kruisheer, que nous suivrons dsormais) afin de
pouvoir les distinguer des passages rdigs par Thodore lui-mme.
11. Hans Heinrich SCHAEDER, Die Kanter , Welt des Orients 1, 1947-52,
p. 288-298 (cit ci-aprs Schaeder).
12. Wilferd MADELUNG, Ab s al-Warrq ber die Bardesaniten, Marcioniten und Kanter , in Hans R. ROEMER., Albrecht NOTH (d.), Studien zur
Geschichte und Kultur des Vorderen Orients : Festschrift fr Bertold Spuler zum
siebzigsten Geburtstag, Leiden, 1981, p. 221-224 (cit ci-aprs Madelung).
13. Voir Christelle et Florence JULLIEN, Aptres des confins. Processus missionnaires chrtiens dans lempire iranien, Leuven, 2002, notamment chap. III,
p. 137-151.
14. Dossier runi par Dan SHAPIRA, Manichaeans (marmanaiia), Zoroastrians (Iazuqaiia), Jews, Christians and other Heretics : a Study in the Redaction
of Mandaic Texts , Le Muson 117.3-4, 2004, p. 243-280.



Les diffrents points de doctrine sont introduits par Thodore de

faon tout fait semblable dans les 3 ( Un peu de sa doctrine , sur
Baay) et 5 ( Un peu de leur doctrine , sur les dostens) :
Les premiers points de doctrine de chacune des deux sectes (A et
D) sont introduits de manire identique : Il dit/Ils disent en effet que .
Les points suivants sont introduits en B, C, E et G par lexpression Et ils disent encore que . Le dcoupage en paragraphes des diteurs du CSCO ne nous parat donc pas correct : certes, les points D et
E mettent en scne les mmes personnages, mais ils ne traitent pas du
mme moment de lhistoire (cosmologie en D, eschatologie en E). La
seule exception (F, o la citation nest pas introduite par la conjonction
d que ) sexplique par la mention unique du titre du texte cit.
Les autres expressions utilises nintroduisent pas des sujets diffrents :
B ( en disant propos de lui que ) est la suite de la phrase commence en B de sorte que le passage nest pas ncessairement lacunaire15,
D ( et ils disent que ) est la suite du rcit cosmogonique.
La dernire phrase conclut sur la folie de ces doctrines ( 3
impit de Baay, 5 [choses] dlirantes des dostens) avec,
dans les deux cas, une insistance sur le nombre de leurs vices ( multitude de son impit 3, aussi 5). Lidentit de structure et la
construction spciale du verbe emar avec la prposition meul ( dire
quelque chose propos de quelque chose ) rendent improbable lhypothse dune lacune dans le dernier paragraphe qua mise Pognon (n. 4,
p. 177), considrant que rien ntait dit des deux personnages mentionns (Dnns et D) : on peut estimer au contraire que la seule
mention de noms sans correspondant dans la littrature chrtienne suffit
discrditer en conclusion la doctrine dj abondamment documente.
Pour marquer ses citations, cest--dire, dun point de vue grammatical, le passage au discours direct, Thodore utilise les nombreux
moyens linguistiques dont dispose le syriaque ; mais la plupart restent
ambigus : le verbe de dclaration (emar) suivi de la conjonction d peut
introduire le discours direct ou indirect16 ; la particule enclitique lam
(que nous avons traduite en annexe dit-il ou dirent-ils avec inversion du sujet pour mieux la distinguer de emar) indique le discours
direct, mais sa rptition est difficile interprter : sagit-il de sutures
entre diffrentes citations ou du rappel que le discours direct continue ?
En outre, ces deux procds, de mme que la mention de la source ou
lusage des premire et deuxime personnes grammaticales, indiquent
seulement que lauteur introduit un discours direct, non quil cite mot
pour mot un texte quil a sous les yeux. Nous devrons donc tre attentifs, pour chaque passage, ces diffrents procds afin de dterminer
15. POGNON (n. 3, p. 223), pour expliquer lobscurit du texte, supposait une
16. Ainsi, dans le 2, il introduit un rsum et non une citation : et la
croix, ils disent quelle est le mystre de la limite [...] .



si lon a affaire de vritables citations (ventuellement tronques) ou

des rsums de lauteur.
Seul lusage de mots nappartenant pas au syriaque littraire signale
donc coup sr une citation. On observe plusieurs types dcart par
rapport la norme du syriaque : certains mots, relevs par les lexicographes syriens, sont inusits dans les textes littraires (mabt
dans la citation F) ; dautres mots, bien syriaques, sont incorrects dun
point de vue morphologique ou orthographique (impratif msy pour ms
attendu en D) ; dautres nont pas de sens satisfaisant en contexte (utr
en E) ; dautres, enfin, sont des hapax (qw en D). Dans la majeure
partie des cas, lincomprhension est leve si on lit ces mots insolites
la mandenne. Les textes o abondent ces formes atypiques ne sont
cependant pas rdigs entirement en manden ni simplement translittrs en syriaque : il sagit comme nous le verrons, de textes quon peut
raisonnablement penser crits en manden et adapts le plus souvent
suivant les rgles phonologiques, morphologiques et orthographiques du
syriaque ; nous parlerons donc, pour les formes qui ont chapp cette
adaptation, de mandismes .
Dans le prsent article, les rfrences aux textes mandens seront
indiques comme suit :
GS : Ginza smala (partie de gauche du Ginza rba, voir GY).
GY XV, 1 (302, 18, trad. p. 300) : Ginza iamina (partie de droite
du Ginza rba), XVe livre et 1er chapitre (dans la numrotation de la
traduction de Lidzbarski), p. 302 et l. 18 de ldition de Heinrich
PETERMANN, Thesaurus s. Liber Adami, Leipzig, 1867, p. 300 de la
traduction de Mark LIDZBARSKI, Ginz. Der Schatz oder Das groe
Buch der Mander, Gttingen, 1925.
JB XLIX (179, 8, trad. p. 177) : Draia -Iahia (Livre de Jean),
p. 179 et l. 8 de ldition de Mark LIDZBARSKI, Das Johannesbuch
der Mander, Gieen 1915, t. 1, p. 177 de sa traduction (t. 2).
Le passage sur le Pre de la grandeur
Le paragraphe sur le Pre de la grandeur , premier des points de
doctrine dont Thodore rend compte, a une forme tout fait originale :
il est la fois le seul tre introduit par un verbe au singulier et le
seul ne comprendre ni mandisme ni lenclitique lam. Il sagit donc
soit dune citation dun auteur crivant en syriaque ( Baay ), soit
dun rsum doctrinal.
La terminologie utilise na pas de parallle direct dans la littrature
mandenne, pas plus que les personnages mis en scne : en manden,
les noms alahuta et alaha renvoient, une seule exception prs17, aux
fausses divinits (lexpression *maria alaha nest de toute faon pas
17. Voir Ethel Stefana DROWER, Rudolf MACUCH, A Mandaic Dictionary,
Oxford, 1963, alaha , p. 18 (cit ci-aprs Mandaic Dictionary).



atteste), lexpression aba -rabuta nest pas atteste, aba et bia, personnifications du bien et du mal, ne renvoient pas des acteurs de la
cosmogonie. Seule la mtaphore de la voix pour rendre compte de la
cration, provenant dune exgse de Gen 1, perceptible dans ce texte
(ql la voix , qr appeler ) est commune au mandisme et au
manichisme18. Autant lchec de la premire cration dAdam a des
parallles (GY III : 100 s., trad. p. 107sq.), autant la ligature de Dieu
ne semble pas tre une ide mandenne, moins que ne soit voque
par-l la ligature du dmiurge (voir texte E).
Si ce passage est une citation, il ne sagit pas dun texte manden,
mais dun texte manichen (personnification du Bon et du Mauvais,
guerre primordiale, Pre de la grandeur, vocabulaire de la cration) ou
proprement kanten .
Le nom des dix cieux et loffrande au jardin dAdam
Hormis lnumration des noms des cieux, le passage est entirement rdig en syriaque : le mot perd graines na mme pas dquivalent phontique proche en manden.
Le nombre de dix cieux nappartient pas la tradition mandenne
qui en voque soit un19, soit sept20. Il est attest en revanche dans le
manichisme et est voqu par Thodore dans sa notice sur la doctrine
manichenne. Les noms eux-mmes sont inintelligibles, sauf le dernier nom de la liste, y, qui peut sinterprter comme le mot syriaque
ayy vivant (et non le pluriel ayy vie , puisque laccord est
au singulier) de sorte quil faut comprendre que cest cette terre qui
a offert des fruits Adam. Ce motif narratif est sans parallle dans la
littrature mandenne, o le motif du jardin dAdam21 et les spculations sur les arbres du Paradis22 sont pourtant bien attests. Selon toute
vraisemblance, le texte nest donc pas manden.
Sur le Fils de la lumire
Ce passage se prsente comme lillustration de ce qui est dit dun
personnage appel Fils de la lumire . La citation, truffe de mandismes, se retrouve, quelques variantes prs, dans le GS III,11
18. Voir qala , Mandaic Dictionary, p. 400, et qra 1 , ibid., p. 414.
19. Voir Mark LIDZBARSKI, Das Johannesbuch der Mander, t. 2, Gieen,
1915, n. 8, p. 56-57.
20. Cf. une reprsentation mandenne du monde dans Ethel Stephana
DROWER, The Mandaeans of Iraq and Iran, Oxford, 1937, n. 5, p. 254-255 (cit
ci-aprs The Mandaeans of Iraq and Iran).
21. Cf. notamment JB LXVI (trad. p. 221-222), GS I, 2 (trad. p. 431), III,
30 (trad. p. 557).
22. Cf. notamment les superbes illustrations du rouleau Alma riaia rba
( le grand premier monde ), reproduit en fac-simil par Ethel Stefana DROWER,
A Pair of Naoraean Commentaries (Two Priesterly Documents) : The Great
First World and the Lesser First World , Leiden, 1963.



(87,22 s., trad. p. 524-525) o le pome est plus long. Voici les deux
textes :
Thodore Bar Kona

Ginza smala III,11

(1) dr-n w-masgn al nemt. (1) darina u-masgina el nimata

Je rpands et je multiplie les mes. kulhin
Je chante et marche vers toutes
les mes.
(2) w-kad zayn knap l-appay :
Et lorsquelles me virent, elles
volrent vers moi

(2) mihzia -hiziun knap alanpai

Quand ils me virent, ils se
rassemblrent vers moi en sortant.

w-lep lm allmn. w-aylel

w-mar l.
et mille saluts elles madressrent,
elles gmirent et me dirent.

(3) knap napqia alanpai u-kulhun

lama almun u-amria
Ils se rassemblrent en sortant vers
moi et tous madressrent un salut
en disant :

(3) bar nuhr.

Fils de la lumire !

(4) br abia
Fils des [hommes] bons !

(4) zel emar leh l-bn.

Va dire notre pre :

(5) l-abuk amar-lx

ton pre, dis-lui :

(5) l-emtay asr metrn :

Quand les captifs seront-ils

(6) l-emat esiria mitrin esiria

Quand les captifs seront-ils dlivrs,
les captifs,

wa-rw lhn l-ayyq d-ayyiq

lhn :
et (quand sera) le repos pour les
affligs qui (sont) dans laffliction

u-ruaha-lun l-aiqia -aiqa-lun

et [quand] le repos23 des anxieux
qui sont anxieux ?

wa-rw lhn l-nemt :

et le repos pour les mes
d-uln b-tbel sbln.
qui endurent la ncessit dans le
monde ?

[Expression contenue dans le

Ginza, deux phrases aprs la fin de
la citation de Thodore.]

(6) w-malllet-n w-emret lhn.

Moi, jai parl et leur ai dit :

(7) mn qudam -l-ab emar-lx

emarlkun ana
Avant que je ne parle mon pre, je
vous dis :
(8) kul -abad eubadia sainia haka
b-sadadia iatbia
Quiconque a commis des actions
dtestables, ici, dans les fers, il

23. LIDZBARSKI traduit Wann wird es den Beengten [...] weit werden ?
comme sil sagissait du verbe rwh ; mais on aurait tarualun et pas ruahalun.



Thodore Bar Kona

Ginza smala III,11

(7) kad prt yibbe pmeh :

Lorsque lEuphrate sera sec
son embouchure

alma -pra iabi mn pumx

jusqu ce que lEuphrate soit
sec son embouchure

w-deqlat men aleh yibbe

et [que] le Tigre, par son
coulement, sera sec

u-diglat ahla mn ihlx

et [que] le Tigre coule hors de son

kulhn zabn. w-metapn kulhn

nal :
tout le temps et que tous les
torrents seront clarifis,

alma -iabix kulhun iamamia

u-kulhun zibia nahlia u-ainaniata
jusqu ce que soient sec toutes
les mers et [que] tous les fleuves,
rivires et sources dbordent25,


u-haizak -rmia b-bira saliq u--esir

b-esura mitria
et alors celui qui a t jet dans le
puits (re)montera et celui qui a t captur dans la
captivit sera dlivr.

l-nemt rw.
pour les mes [sera] le repos.

(9) ruaha-lun l-nimata -haka

daira bit qirsa
Le repos [sera] aux mes qui
ici habitent dans la maison du

Les deux textes sont si proches que les diffrents savants se sont
pos la question de leur anciennet relative. Plusieurs scnarios sont en
effet possibles : a) Thodore Bar Kona a eu sous les yeux un texte diffrent du texte manden actuel, et on peut se demander si les mandens
ne lauraient pas emprunt aux kantens ; b) Thodore, en adaptant
un texte crit dans un aramen proche mais diffrent du syriaque, a
adapt et donc chang la lettre de ses sources ; c) les copistes ont mal
transmis ces citations crites dans une langue parfois obscure pour un
Le nombre important de formes aberrantes en syriaque dans le texte
de Thodore montre quil sagit bel et bien dune citation du manden.
Mais il faut observer le soin et ce, dans tous les manuscrits avec
lequel les copistes ont recopi la ponctuation et les signes diacritiques,
ce qui atteste dune volont de rendre le texte tranger intelligible pour
un lecteur syriacisant. Mme lhapax msgn peut se comprendre ais24. LIDZBARSKI hsite sur le sens de lexpression (expression parallle und
der Tigris seinen Lauf verlegt (?) ), mais le Mandaic Dictionary ( HL I ,
p. 450) comprend comme POGNON (p. 237).
25. POGNON (p. 237) dcoupe tort autrement la phrase ( jusqu ce que
toutes les mers, tous les cours deau soient desschs, que les torrents et les
sources dbordent ) : le verbe mitapian est au fminin pluriel et ne peut donc
saccorder quavec ainaiata sources .



ment comme la graphie phontique en un seul mot de msg n grce

la graphie du mot prcdent dr n. Cest pourquoi nous prfrons
penser que la seule autre graphie vraiment difficile du texte tient une
erreur dans la transmission : la graphie zny dans la deuxime phrase
na pas de sens, le copiste ny a pas, du reste, ajout de sym ; la
correction de Pognon (p. 234) zyny (zayn, ie. acc. 3 f. pl. + suffixe
1 sg.) parat acceptable.
La plupart des lexmes sont communs, mme quand le sens diffre
entre le manden et le syriaque : prposition l (mand. el sur, vers ,
syr. al sur ), peal knp (mand. se rassembler , syr. voler ), etpeel
p (mand. dborder , syr. tre clarifi , la racine syriaque correspondant tymologiquement et smantiquement au manden p est p).
Le premier mot dr n diffre du premier mot manden (darina) : il
sagit soit dune erreur dans la transmission du texte, soit dun essai
pour rendre le texte comprhensible ; mais la correction de Pognon
(n. 1 p. 236), qui proposait damender la fois le texte de Thodore et
celui du Ginza en proposant radina je mavance , parat de mauvaise
mthode : le texte du Ginza prsente un texte satisfaisant ; ainsi, Lidzbarski traduit Ich singe Hymnen (p. 524), ce qui est cohrent avec
lusage liturgique observ du Ginza smala26.
Quand les formes syriaques et mandennes sont lgrement diffrentes, le syriaque offre une graphie proprement syriaque. En premier
lieu, l o le manden, en raison de certaines de ses volutions phontiques particulires ( > h, , > 0), ne distingue pas *h et ni
* et *, le syriaque fait la distinction et note pour hiziun, ruaha,
ihlx, nahlia et h pour kulhun, pour aiqia et pour esiria. Les formes
grammaticales (conjugaison et indices personnels) sont proprement
syriaques, comme lhwn pour le manden lun, et crites en deux mots
conformment aux habitudes graphiques en usage dans cette langue.
Les lexmes sont parfois lgrement diffrents entre le manden et le
syriaque : mand. alanpai (tymon *l-appay- avec prothse vocalique et
traitement nasal de la gmine)/syr. l-appay, mand. l-emat/syr. l-emtay,
mand. diglat/syr. dqlt, mand. pra (le manden prat est galement
attest)/syr. prt.
On a donc affaire la syriacisation dun texte crit en manden ou
dans un dialecte proche. Plus intressantes sont les vritables divergences textuelles. Ce sont en effet les passages qui comprennent le plus
de traits caractristiques de la syntaxe et de la stylistique mandennes
qui sont les plus diffrents dans le texte de Thodore :
Linfinitif sans prposition, ventuellement suivi dune subordonne de mme sens, est une tournure courante et remplit la fonction de

26. Voir The Mandaeans of Iraq and Iran, p. 185.



complment de temps27 ; le syriaque ne connaissant pas cette construction, Thodore a peut-tre rcrit ici le texte.
La rptition est un des traits les plus marquants du style manden. La 2e phrase de Thodore ne comprend pas la rptition prsente
dans le Ginza et comporte le mot lep (lm) mille (saluts) , absent
du texte du Ginza. Ce dernier mot nous semble une erreur de comprhension sur le mot manden alanpai (compris la premire fois l-appay)
qui fait partie du segment de phrase rpt en manden ; auquel cas le
texte de Thodore natteste pas dune recension divergente. De mme,
dans la 5e phrase du syriaque (6e phrase du manden), le mot esiria
rpt en manden ne lest pas en syriaque. En revanche, le dveloppement de la question dans le texte syriaque (5e phrase), typiquement
manden, peut tre la trace dune recension divergente dans le texte
consult par Thodore.
La rponse fort longue dans le texte manden (8e phrase) est beaucoup plus brve dans le texte syriaque. Elle comprend en outre le mot
zabn temps qui, certes, se comprend bien, mais qui pourrait tre le
signe dune incomprhension du mot manden zibia fleuves , absent
du syriaque o zb ne dsigne que le fleuve Zb et qui, dans le texte
manden, dveloppe parfaitement le champ lexical des eaux (pra,
diglat, iamamia, nahlia, ainanata). Le texte diffrent de Thodore nous
parat donc rsulter dune faute de comprhension.
Le changement de personnes grammaticales (et donc de situations
dnonciation) est extrmement frquent dans les textes mandens. Ici,
le moi raconte sa rencontre avec les mes ( la 3e personne), puis
leur parle ( la 2e personne). En syriaque, on reste la 3e personne
(cest--dire sur le plan du rcit de la rencontre, sans entrer dans le dialogue). La situation est plus complique encore dans le texte manden,
puisque le moi anticipe le dialogue quil aura avec le pre. Le texte
de Thodore nous parat donc simplifier une situation de communication complexe qui na pas dintrt pour lexpos doctrinal vis.
Une autre divergence notable est labsence du dbut et de la fin
de la 8e phrase mandenne chez Thodore. Or, le dbut de la phrase
est la proposition principale, et ces deux propositions sont celles qui
donnent la porte morale du texte en manden : Quiconque a commis
des actions dtestables, ici, dans les fers, il reste [] (punition des
pcheurs), et alors celui qui a t jet dans le puits (re)montera et
celui qui a t captur dans la captivit28 sera dlivr (rtribution des
justes ayant souffert ici-bas). Or il nest pas dans le propos de Thodore daccorder le moindre brevet de moralit la secte dcrite.
27. Voir Theodor NLDEKE, Mandische Grammatik, Halle, 1875, 268,
p. 388-389 (ci-aprs cit Mandische Grammatik) o un exemple avec la mme
expression est analys : mihzia -hiziun. Le dcoupage de la phrase par POGNON
( toutes, en me voyant, (....) p. 236) ne se justifie donc pas.
28. Cest--dire le captif attach dans des liens.



La divergence principale tient au nom du personnage qui parle

la premire personne : Fils de la lumire pour Thodore, Fils des
[hommes] bons dans le texte du Ginza. Cest largument principal de
Pognon pour dire que le texte manden est un emprunt au texte kanten recopi par Thodore. Or, les [hommes] bons apparaissent
un autre endroit dans le texte de Thodore : lorsquil voque les noms
par lesquels les dostens sont connus : En Msne, on les appelle
mandens, makanens, gens de Celui qui fait le bien ( 4 ; bt bed
bt). Or ce nom est parfaitement inconnu de la tradition mandenne,
tandis que les autres noms correspondent, comme nous le verrons,
la tradition mandenne ou la tradition hrsiologique syriaque. Est-il
raisonnable de penser que cest partir de cette appellation de br
abia que Thodore a cr une nouvelle tiquette pour ses dostens ?
Dautre part, lintention de Thodore, donne dans la phrase qui introduit la citation, est dillustrer le personnage du Fils de la lumire ,
connu de lui dans sa notice sur Mani comme lpithte des cinq fils de
lEsprit vivant et comme une pithte anglique (texte E) ; peut-tre a-t-il
simplement, et de bonne foi, chang le nom figurant dans le pome,
pour faire correspondre le pome son propos.
Seules les divergences qui ne peuvent sexpliquer par lintention
hrsiologique de Thodore nous semblent pouvoir sexpliquer par des
variantes entre la recension mandenne utilise par lauteur et le texte
reu du Ginza, savoir les mots w-aylel (2e phrase cite) et zel (4e
phrase), ainsi que la place des mots wa-rw lhn l-nemt d-uln
b-tbel sbln (5e phrase), situe plus loin dans le Ginza.
Le texte cit par Thodore est donc le mme texte que celui du
Ginza smala, les divergences qui ne tiennent pas une rcriture du
texte pour son intgration dans la notice tant trs peu nombreuses. Il
nous parat ds lors impossible de soutenir que le texte du Ginza soit
un emprunt un texte antrieur, puisquil sagit du mme : lhypothse
dun emprunt des mandens la littrature kantenne ne peut pas
sappuyer sur ce passage.
La cration du monde par Ptael, fils dAbtr
En reprant lemploi du verbe emar et de la particule enclitique
lam qui ponctuent rgulirement le texte, on peut diviser celui-ci en
quatre passages (numrots dans lannexe de a d) qui senchanent
parfaitement sur le plan narratif et qui sont proches de nombreux rcits
cosmogoniques mandens sans tre identiques eux.
Lintroduction (a) est entirement rdige en syriaque, lexception
de ladjectif masculin pluriel rabb grands , qui est manden (rbia,
syr. rawrb29). La premire phrase reprend, sans en utiliser les termes
29. Mme si pour les substantifs la forme est atteste en syriaque, voir
NLDEKE, p. 359.



exacts, le premier verset de la Gense : la place de lesprit au-dessus

des eaux, des grandes puissances qui ont pour fils Abtr, lequel a
pour fils Ptael.
Le deuxime passage (b) contient de nombreux mandismes : msy est
limpratif manden de msa (syr. ms), de mme que wy pour w30 ;
rbyt na pas de sens en syriaque mais signifie ocan en manden
(rabita31) ; le verbe manden kl, homonyme du syr. kl manger , a,
de plus, le sens de durer 32.
Il dcrit lacte de cration avec le vocabulaire dmiurgique couramment utilis dans les textes mandens : la cration de la terre est dcrite
comme une opration de condensation (syr. zel ms [] mst = mand.
zil msia msuta33), celle du ciel comme ltendage dun drap sur un pilier
(mand. ngad umia, esuna). Le dtail du texte rserve nanmoins une
difficult : les expressions arnqy et mynqy34. Le seul terme proche
qui permet dviter de corriger le texte est ladjectif manden naqia
qui signifie pur et correspond en syriaque au verbe nq sacrifier ;
peut-tre Thodore a-t-il compris le manden arqa naqia comme *ar
naqy (p. passif f. sg.) et *umia naqiia comme mayy nqn (p. passif
m. pl., avec une erreur de transmission par la suite).
Pognon (p. 238) a rapproch la description de lhumanit de celle
de Ptahil dans un passage du GY VI (210,1-2, trad. p. 20935) : hiuar
rix mn haupia mia u-ziqnx mn upia -aqamra hiuara ( plus blanche
[est] sa tte que lcume de leau et sa barbe que les touffes de laine
blanche ), ce qui la oblig faire de nombreuses corrections dans
le texte de Thodore et chafauder une transmission du texte trop
complique pour tre vraisemblable. Le plus simple est de supposer
que Thodore a recopi un texte manden aujourdhui perdu qui joue
sur lide que le dmiurge a plong sa main dans les eaux noires originelles pour donner sa couleur lhomme ; la mention des poissons de
locan pour noircir la barbe reste sans parallle.
La dure de vie de lhomme de deux cent soixante-dix ans est galement sans parallle.
Le texte (c) continue sur lacte de cration manqu, thme central
des rcits cosmogoniques mandens. Il comprend lui aussi un mot man30. Voir NLDEKE, p. 359.
31. Voir NLDEKE, p. 358.
32. Voir POGNON, p. 239.
33. Cest exactement le texte du Diwan Abatur (2e partie, l. 1, voir Ethel
Stefana DROWER, Diwan Abatur, Vatican, 1950). Msuta est le complment
dobjet interne du verbe msa, voir Mandische Grammatik, 271, spcialement
p. 399 avec lexemple : mia msuta la-msun das Wasser verdichtete sich nicht .
34. Mauvaise lecture nqny par POGNON, juge suspecte par NLDEKE (p. 358359) et rectifie par SCHER. La correction de Pognon (p. 238) adinqia sans ,
suivie par lensemble des traducteurs, est lourde et ne donne pas de sens
35. Voir POGNONp. 238.



den (bbay) et un hapax (repst). Baba, qui nest attest en syriaque

que dans le sens de pupille (de lil) 36, est un terme polysmique
en manden : il signifie porte , peuple/religion et dsigne souvent
chacun des douze signes du zodiaque associs une religion37. Le
terme rpst est un hapax en syriaque et en manden. Pognon (p. 240)
la rapproch du manden ripsa (quil comprenait comme moment,
instant et sorte de sortilge ), Nldeke, en sappuyant sur le sens
du verbe rps fouler , a corrig le sens en zitternde Bewegung 38.
Nous comprenons donc que Ptael, au lieu de crer lhumanit, cre les
constellations zodiacales et que celles-ci sont dpourvues de tout esprit
et de toute me ; le tremblement ou le coup de pied dsignerait
alors la mise en mouvement de la machinerie cleste.
Les reproches dAbtr Ptael viennent clore la narration (d).
Une difficult textuelle : qw est un hapax la fois en syriaque et en
manden ; Pognon (p. 241) la rapproch du manden giua maldiction , Nldeke (p. 783) du syriaque qwd lien , sens quil a probablement dans le texte E. Il est remarquer que le sous-texte manden
dit quAbtr, assis sur ses sept firmaments, comme dans le reste de
la littrature mandenne, lve les yeux (sens du manden dla) et non
quil les baisse (sens du syriaque), ce qui confirme que Ptael est bien
en train de crer, en hauteur, les constellations du zodiaque, au lieu de
crer en bas les hommes.
Le sort de Ptael et leschatologie
Les mandismes abondent dans ce passage. Outre lhapax qw
dj mentionn, on y trouve une expression dsignant la fin des temps
(adm l-ym dn : w-nay purqn), qui est trs vraisemblablement
une dformation de lexpression mandenne alma l-ium dina u-alma
l-ita u-ita aiia -purqana jusquau jour du jugement, jusqu
lheure, savoir les heures du salut 39. Lvocation en syriaque de la
rsurrection (w-kad hwy ayyat mt quand sera [venue] la vie des
morts ) suit mme, avec un mot mot diffrent, cette expression en
GS I, 2 (19, 2, trad. p. 437 : alma l-iuma rba -qaiamta jusquau
grand jour de la rsurrection ).
Limage de la brique du mur40 confessant le Christ, repre par
Pognon (p. 7sq.), se trouve galement dans le GY XVIII (386,2436. Voir Robert PAYNE SMITH, Thesaurus syriacus, Oxford, 1879, col. 442
(ci-aprs cit Payne Smith).
37. Voir baba 1 , Mandaic Dictionary, p. 45.
38. Voir NLDEKE, p. 359. Sens repris par le Mandaic Dictionary ( ripsa ,
p. 433).
39. Voir Mandische Grammatik, p. 320-321 ; pour les rfrences, limites
au GY, voir purqana , Mandaic Dictionary, p. 369.
40. La traduction libna mn aita par la brique [parle] avec le mur a t
rejete par NLDEKE (p. 356) et dans le Mandaic Dictionary ( libna , p. 235).
Lexpression de Thodore nappartient pas au syriaque littraire : lbt est soit la



25), quoique dans un autre contexte : celui de lapparition dun messie

(mensonger) qui suivra une catastrophe naturelle de type inondation.
Le verbe nsg est lire (comme dans la pice C) comme le manden
nasgia il savancera et non il chantera . Les utr ( richesses )
du syriaque renvoient en fait aux utri (eutria), appellation mandenne
courante pour les entits angliques, galement qualifies de Fils de
la lumire .
Lexpression qualifiant la chane qui est la totalit du monde
(en syriaque mlteh d-lm) est une expression mandenne mise
au jour par Pognon : malaiia -alma, laquelle signifie que la chane
remplit la totalit du monde41 et est aussi atteste dans la littrature
manichenne42. Le sens du mot llt, en revanche, qui signifie bien
pointe en syriaque (Payne Smith, col. 1910), nest pas assur en
manden : il pourrait tre un mot demprunt en manden, mal compris
et interprt comme embryon 43. Si le premier semble mieux convenir limage du chtiment, le second nest cependant pas exclure, le
dmiurge transperc, qui semble avoir entran dans sa dsobissance
dautres anges, apparaissant comme lembryon de sa propre cration.
Acheminement vers les magiciennes et tiologie du baptme
Cette pice, la seule dont Thodore donne le genre (zmrt) et le titre,
se compose de deux passages assez diffrents : un premier qui consiste
en une ou plusieurs citations, un deuxime, entirement syriaque, qui
pourrait tre un rsum.
Le premier mot du titre drkt l rt nest pas syriaque, mais manden. Lintuition de Pognon ( acheminement vers les sorcires , p. 226
et 244) peut tre dsormais taye : adrakta est attest comme variante
de drakta acheminement (correspondant au syriaque durkt)44. Ds
lors, la prposition al peut se comprendre comme la prposition mandenne el (qui correspond au syriaque al sur et l pour, vers ). La
traduction de Rudolph (p. 259) (Abwehr)Mittel gegen Zauberinnen
est tentante, mais sans justification philologique.
Le premier passage, invocatoire, peut avoir pour source un texte
magique manden perdu ; en raison de la mention dun titre, il sagirait plus vraisemblablement dun livre que dune coupe. Il numre les
appellations de lesprit mauvais45 selon diffrentes rgions (la chane
graphie phontique du syriaque lbnt (lbett) (voir NLDEKE, p. 356), soit une
transcription du manden libta (POGNON, p. 242) ; la graphie du pluriel yt,
contrairement ce quavance POGNON, est bien syriaque (voir PAYNE SMITH,
col. 411).
41. Occurrences reprises par LIDZBARSKI, Johannesbuch 2, n. 6, p. 137.
42. Voir les rfrences chez POGNON, p. 243.
43. Voir lulita 1 , Mandaic Dictionary, p. 233.
44. Voir adrakta , Mandaic Dictionary, p. 8.
45. POGNON (p. 244) qui comprend un pluriel les fes mchantes est
oblig de corriger en consquence est appel , ce qui parat tre dune mau-



lpreuse , la terre des Romains, lInde, lOccident, lOrient et la ville

dal-ra). La traduction, depuis Pognon, de lexpression chane
lpreuse (lt grbnyt) par chane septentrionale suppose une petite
correction (lt grbyyt46) qui ne nous semble pas lgitime : si chane
septentrionale convient bien pour le texte original cit par Thodore
dans la mesure o sont mentionns deux autres points cardinaux, Thodore a sans doute compris lpreuse , puisquil dit ensuite des magiciennes quAdam les a guries de leur lpre (garbhn). Le motif de la
chane rappelle une phrase atteste dans le corpus magique manden :
esira atuat haraata b-ulata -parzla lies [soient] les femmes
magiciennes par des chanes de fer (DC 40,110247).
Quelques-uns des noms des esprits malfiques sexpliquent par le
manden : Hamgay et Hamgagay (mand. hamgai u-hamgagai48) renvoient un esprit de lumire et son pre ; la Nanay (mand. nanai49),
divinit msoptamienne, est aussi atteste dans la littrature magique
mandenne et Thodore y consacre sa notice 93 ; Bil est le nom manden et syriaque (bl) de Jupiter50 ; m peut se lire en syriaque emm
la mre (mand. ema) et peut renvoyer diffrents personnages des
tnbres (Namrus, mre du monde , et Qin, mre des tnbres 51) ;
Mamania est un nom manden de femme52. Trois de ces noms sont
donc plutt positifs dans le mandisme (Hamgay et Hamgagay,
Mamani). Pour la comprhension des autres noms, on se rapportera
la longue note de Kurt Rudolph53.
La suite du texte ne comporte aucun mandisme (sauf le mot dsignant le baptme) ni marque de citation : il sagit vraisemblablement
dun rsum, voire dune fiction. Les seuls parallles avec la littrature mandenne sont lpithte dAdam, gabra qadmaia le premier
homme , et la mention de son baptme administr par Abel (mand.
Hibil Ziwa). La fonction du passage est de donner le type du baptme manden : il reproduirait la purification des magiciennes lpreuses
par Adam. Les mots renvoyant au baptme chrtien sont soigneusement
vits par lauteur, qui utilise le verbe aphel as purifier, baigner
(le peal manden sa renvoie aux ablutions) et le nom manden du baptme mabuta syriacis ici en mabt, forme qui nappartient pas au
vaise mthode. Traduction correcte dans le CSCO (p. 258) et chez Kruisheer
(p. 167).
46. Voir PAYNE SMITH, col. 772 et 773.
47. Voir ualta , Mandaic Dictionary, p. 457.
48. Voir POGNON, p. 244, et Mark LIDZBARSKI, Mandische Liturgien, Berlin,
1920, n. 2, p. 7.
49. Voir nanai , Mandaic Dictionary, p. 283.
50. Voir bil 1 , Mandaic Dictionary, p. 60 et PAYNE SMITH, col. 518.
51. Voir les rfrences du Ginza ema 1 , Mandaic Dictionary, p. 352.
52. Voir mamania , Mandaic Dictionary, p. 245.
53. Voir Kurt RUDOLPH, Die Mander, t. 1, Gttingen, 1960, n. 8, p. 32-33
(ci-aprs cit Rudolph).



syriaque littraire et nest atteste que dans les lexiques54. Lintention

polmique de lauteur un endroit stratgique peu avant la fin de la
notice est nette : il sagit de dnigrer le rite central et le plus visible
du mandisme, pratiqu tous les dimanches, en ayant recours aux procds classiques de lhrsiologie (numration de noms barbares, vocation de pratiques magiques, mise en scne de personnages bibliques
dans les situations les plus loignes du texte biblique), de sorte que
lauthenticit du rcit dun Adam gurisseur, en labsence de parallle
manden, reste douteuse.
La mention de Dnns et de D
Thodore se contente dvoquer les deux noms avec leur qualiticatif : Dnns spar delt Dnns, scribe des sectes (qui peut aussi
se lire spar delt livre des sectes ) et D zaw le petit D .
Ces deux personnages ne sont mentionns que dans le GY VI (204sq.,
trad. p. 205sq.) o ils sont appels Dinanukt sapra hakima spar diuta
-alahia Dinanukt, le scribe sage, le livre dencre des (faux) dieux
(204, 23) et Diai Zua le petit Diai (205, 5) ; il est donc vraisemblable que Thodore a eu sous les yeux une version du GY VI. Dans
ce texte, Dinanukt cherche dans ses livres des rponses ses questions
mtaphysiques et tombe sur le livre Diai Zua dont il finit par rejeter
le message pour la vraie religion quil dcouvre lors dune traverse
des postes de garde stellaires guid par linstructeur anglique DinMlik-Utra ; aprs tre redescendu dans le monde dici-bas, il dcide de
brler tous ses livres. La lecture de Thodore montre tout lintrt
vident pour un hrsiologue quil a port lattitude des mandens
lgard des autres religions. La mention tant trs brve, loriginalit de
lpithte du premier personnage chez Thodore est peut-tre due une
reformulation plutt qu une recension diffrente du texte de GY VI.
Conclusion sur linsertion des citations
Le tableau suivant rsume les conclusions auxquelles notre analyse
nous a conduits :



A) Pre de la


nom des
entits divines,
terminologie de
la cration

de la cration
partage avec le

B) Dix cieux
et offrande au
jardin dAdam

mention de noms
barbares, rcit en

cosmologie : dix

ciel (ayy) //
nom manden de
Dieu (Hiia)

54. Voir PAYNE SMITH, col. 3358.



C) Fils de la


D) Cration du
dbut syriaque,
monde par Abtr narration en
et Ptael



Fils de la
lumire :
qualificatif des
fils de lEsprit

texte du GS III,
11 avec quelques

personnages et
mythes mandens

E) Sort de Ptael


vers les

dbut manden,
narration en
syriaque, nom
manden du

genre invocatoire,
cits ; nom du

G) Dnns et

mention de noms

personnages du

La cohrence de la notice sur les dostens est vidente : tous les

textes cits ou voqus (D-G) le sont, au moins en partie, en manden
et renvoient des figures et des thmes mandens, mme si on ne
trouve pas dans la littrature mandenne conserve les textes correspondant exactement aux citations. En revanche, il faut souligner le caractre composite de la notice sur la doctrine de Baay. Or on peut tablir
une corrlation entre la langue utilise et le contenu thologique des
textes : les passages rdigs en syriaque, ainsi que lnumration des
noncs barbares, contiennent des traits manichens quon ne retrouve
pas dans le mandisme ; le passage rdig en manden est quasiment le
mme quun texte du Ginza smala. Or, si les textes qui composent la
notice sur la doctrine de Baay sont des textes authentiquement kantens , on ne sexplique pas comment la secte pouvait faire usage de
deux langues diffrentes. Il nous semble que cette observation taie
lhypothse dun montage de la part de Thodore : les textes runis
par lui appartiennent deux corpus diffrents un corpus en langue
syriaque coloration manichenne ( le pre de la grandeur ) et un
corpus manden ; la notice sur la doctrine des dostens rassemble les
textes mandens, celle sur la doctrine de Baay est un assemblage de
textes syriaques (sur le Pre de la grandeur, sur les dix cieux ?) et de
textes mandens (sur le Fils de la lumire) lgrement corrigs afin
dtre insrs dans un paragraphe tonalit manichenne (le Fils des
[hommes] bons manden devenant le Fils de la lumire mis en
vedette avant la citation mme du texte). Lintention hrsiologique qui
prside ce montage consisterait alors remplir un vide documentaire
sur les kantens : puisque Thodore croit savoir que les kantens ont
exist (mais ont disparu puisquils ont donn naissance aux dostens) et



que leur doctrine tait un compos de judasme, de manichisme et de

mazdisme, le texte de caractre manichen sur le Pre de la grandeur
peut ses yeux leur tre attribu, de mme que les spculations sur
les noms des dix cieux et un texte manden mettant en scne un personnage que Thodore croit identifier au Fils de la lumire manichen.
Dans cette hypothse, Thodore aurait eu sous les yeux des documents
mandens, identifis par lui aux dostens, et des documents manichens
ou manichisants 55.
La manire dintroduire les sources cites nous parat rvlatrice du
type de sources partir desquelles Thodore a compos ses notices.
La plupart des textes sont introduits par le verbe emar au pluriel ( ils
disent ), ce qui suggre que les textes consults sont anonymes. Le
seul texte introduit par le verbe parler au singulier est le premier
texte du paragraphe sur la doctrine de Baay : soit le texte comportait la
mention dun auteur (Baay), ce qui correspond la pratique de Thodore dans son paragraphe sur la doctrine des manichens ; soit lusage
du singulier drive du titre du paragraphe qui prcde immdiatement
ce dernier ( Un peu de sa doctrine ), de mme que le retour au singulier la fin du paragraphe ( son impit ) sexplique par la rfrence
au titre.
Voici donc les sources que Thodore peut avoir eues sous les yeux :
des ouvrages de Mani (pour sa notice sur les manichens) ;
un texte syriaque de caractre manichen, peut-tre d un certain
Baay (texte A), avec peut-tre le nom des dix cieux (texte B) ;
un texte manden anonyme (texte C = GS III, 11) quil recopie en
coupant les passages sur la rtribution dans lau-del ;
un texte cosmogonique et eschatologique manden anonyme dont
loriginal est perdu (textes D et E) quil cite sans doute par bribes ;
un texte manden anonyme dont le titre ( Acheminement vers les
magiciennes ) suggre quau moins pour ce passage le titre tait inscrit
sur la copie consulte (texte F) ;
un texte manden anonyme sur Dinanukt (G = GY VI).
Le montage des citations semble command par ce que lauteur
croit savoir des diffrentes hrsies : sont cits dans sa notice sur les
manichens les textes syriaques de Mani ou transmis de faon sre par
des manichens ; sont cits dans la notice sur les dostens des textes
mandens (D-G) ; se retrouvent dans la notice sur la doctrine de Baay
les textes correspondant au compos doctrinal suppos de la secte
kantenne , anctre des dostens , savoir un texte syriaque manichisant avec peut-tre un nom dauteur, ce qui justifierait sa place ici
et non dans la notice sur les manichens (A et B), et un texte manden
qui justifierait laffirmation de lorigine kantenne des dostens (C).
55. On sait que Thodore utilise dans sa notice sur lenseignement de Mani
une ample documentation manichenne, voir Michel TARDIEU, Le manichisme,
Paris, 19972, p. 93-94 (cit ci-aprs Tardieu).



Lapport des citations de Thodore Bar Kona est indniable pour

lhistoire des mandens. Que peut-on dire en revanche de son scnario
hrsiologique ? Permet-il dlucider, comme laffirment Henri Pognon
et Wilferd Madelung, le lien entre kantens et dostens ?

Lhrsie-mre : les kantens

Comme nous lavons vu, Dirk Kruisheer a montr que la notice de
Thodore porte sur les kantens. Ce que Thodore prsente, cest donc
une histoire des kantens dont les dostens ne sont que les descendants. On peut aller plus loin en reprenant prcisment les occurrences
du mot kantens dans lensemble des notices. En dehors du titre et
de la premire phrase de la premire notice et de la dernire phrase
de la dernire notice, les kantens napparaissent que dans lanalyse
du compos doctrinal des dostens ( 4), aux cts des marcionites et
des manichens, et dans le dernier paragraphe o une tymologie de
leur nom est propose. Le terme kantens apparat donc comme une
simple tiquette hrsiologique sans contenu prcis.
Comme dans cinq de ses notices finales56, Thodore situe lorigine
de la secte des kantens dans lhistoire sainte, et mme en terre sainte :
les Philistins partisans de Goliath auraient divinis leur hros et invent
un rite autour de lui ; cest l le type dexplication quon retrouve dans
les notices mentionnant des noms divins babyloniens57. Puis les Philistins, envoys en captivit en Babylonie, de mme que les Judens
ont t dports (2 R 25), auraient rig une nouvelle statue et rinvent un rite tout aussi improbable que le premier. chaque moment
de leur histoire, les pr-kantens apparaissent donc comme les ennemis
du peuple dIsral.
Que ces rcits et ces rites relvent de la fiction hrsiologique ne fait
gure de doute, encore que la recension que Nldeke a faite du livre
de Pognon58 soulve une nouvelle difficult. Nldeke propose en effet
de comprendre les termes rz qylyn (traduits par Pognon, p. 221, les
mystres sont tus ) comme arz qln les cdres sont abattus , en
sappuyant sur le deuxime sens de la racine ql en judo-aramen et
en manden59, lequel correspond larabe qaala abattre (un arbre)
(vs. qatala tuer ) et en justifiant son hypothse par lexpression paral56. La notice 89 sur Bim met en scne un Cananen refaisant le chemin des patriarches bibliques vers lgypte, la notice 91 sur le festin des eaux
reprend lhistoire de la sortie dgypte, la notice 92 sur le festin des morts le
combat des Hbreux contre Balaq, la notice 93 sur Nanay reprend la matire
du premier chapitre de 2 Mac, la notice 94 sur Barq cite le nom syriaque de
la ville de Qadesh-Barna.
57. Notice 88 sur Ramt, 90 sur Dan, 93 sur Nanay et 94 sur Barq.
58. Voir NLDEKE, n. 1, p. 358.
59. Ce sens du manden gl se retrouve dans le Ginza, prcisment dans les
rcits voquant la construction de larche de No en bois de cdre.



lle gnbrmwbdyn les hros ont t anantis . Or, si on accepte les

arguments convaincants de Nldeke, la ressemblance avec Zacharie
11,2 est frappante : Gmis de douleur, cyprs, parce que le cdre est
tomb, parce que les puissants ont t abattus. Gmissez, chnes du
Bashn, car elle est terre, la fort impntrable. Mais le verbe qui
traduit tomber (hbreu massortique npal) na le deuxime sens
d abattre ni dans la Septante, ni dans la Pt, ni dans les Targoums.
Peut-tre le choix de situer lorigine des kantens dans laffrontement
entre David et Goliath vise-t-il expliciter une phrase mandenne
dans son vocabulaire quoiquabsente de la littrature mandenne , le
hros tomb compar au cdre abattu tant identifi au hros philistin
Goliath abattu par David. Nous aurions alors affaire un rcit de fantaisie visant expliquer une phrase trouve dans une source sectaire
rdige en manden.
Que les kantens ne soient que des Philistins se comprend du point
de vue de lhrsiologue. Mais pourquoi donc continuer le jeu des associations jusqu identifier Goliath-Dagon Nergal ?
La seule mention du nom de la divinit babylonienne Nergal (syr.
nergel) dans la littrature syriaque est 2 Rois 17,3060, qui voque le
peuplement de la Samarie par des paens : Les gens de Babylone firent
un Soukkoth-Benoth ; ceux de Kouth, un Nergal ; ceux de Hamath, une
Ashima. Ce verset, qui appartient la polmique juive contre les
samaritains, a t utilis par les juifs pour dsigner sous ltiquette de
kutens les samaritains accuss dorigines paennes et dimpit61,
mais il ne semble pas avoir t utilis par les hrsiologues chrtiens.
Hans Schaeder62 en est galement venu citer ce verset dans son article
par un autre cheminement : si ltiquette dostens renvoie bien aux
dosithens samaritains de lhrsiologie grecque, ltiquette kantens
peut trs bien renvoyer une autre dnomination des samaritains ; le
passage de kuten kanten peut sexpliquer par la proximit des lettres
< n > et < w > en hbreu carr (Schaeder, p. 295) ou par une mauvaise criture de ce mot en syriaque (Schaeder, p. 298) ; Kurt Rudolph63
signale que Kut dans la Septante est orthographi (en 2 R 17,
24, mais pas au v. 30), suggrant ainsi que ltiquette hrsiologique a
des sources grecques.
Avec cette allusion probable 2 R 17, 30, nous avons affaire
un rcit tiologique sur le nom des kutens/kantens. Or Thodore,
contrairement ce quil fait au 6, nexplicite pas ltymologie du mot
kantens , de sorte quon peut penser quil ne fait l que rapporter
60. Voir PAYNE SMITH, col. 2468 (nergel) ; la forme courante en syriaque est
nerg, voir PAYNE SMITH, col. 2469.
61. Voir Lawrence SCHIFFMAN, Cuthaeans , Alan David CROWN, Reinhard
PUMMER, Abraham TAL (d.), A Companion to Samaritan Studies, Tbingen,
1993, p. 63-64.
62. Voir SCHAEDER, p. 295-296.
63. Voir RUDOLPH, n. 2, p. 32.



une tradition hrsiologique plus ancienne o le nom de la secte tait

expliqu par le culte rendu Nergal dans la ville de Kut. Cette tradition a pu natre soit dans le domaine grec (forme ), soit dans
le domaine syro-oriental (o se situe la ville de Kut) et devait faire un
parallle entre deux groupes issus dun mlange judo-paen : les samaritains louest et les kutens lest. Ce paralllisme se retrouve de
faon plus explicite encore dans le Kitb al-r al-bqiya de Brn,
crit en lan 1000. la fin dun dveloppement sur les faux prophtes
des sabens de arrn, qui sappuie sur lapologie du christianisme de
Abd al-Mas al-Kind64, lauteur cherche distinguer ces derniers des
autres sabens qui, comme il le prcise aprs ce passage, habitent dans
la rgion de Wsi65 :
Les sabens sont ceux qui restrent en Babylonie parmi lensemble des tribus qui quittrent lendroit au temps de Cyrus et
dArtaxerxs pour Jrusalem. Ils inclinrent vers les rites des mages
et, ainsi, se penchrent (abaw) vers la religion de Nabuchodonosor et laborrent une doctrine mlange de magisme et de judasme
comme les samaritains en Syrie.66
Les sources prcises de Brn pour ce passage restent inconnues,
mais il nest pas impossible quelles drivent dune tradition hrsiologique semblable celle utilise par Thodore.
Ce nest quau 6 que Thodore Bar Kona donne sa propre tymologie du nom kanten : la mort dAbel, les descendants de Can
dcident dhonorer celui-ci en tant que Nrig et ceux de Seth lorigine des kantens honorent Abel en alignant des nourritures devant
ce qui devient un sanctuaire ( 6 sur les nrigens , tiquette invente pour les besoins du propos). Le paragraphe comprend un double
jeu de mots uniquement possible en syriaque : ny reg il dsire
le repos et nerg Nergal, Mars ; kuwwn blme et kant, nom
prsum du lieu de culte des kantens. Or kant/kent est bien un mot
syriaque, attest, semble-t-il, seulement par les lexicographes (comme
le mabt du texte F), signifiant corbeille fruits ; fondement,

64. Voir Georges TARTAR, Dialogue islamo-chrtien sous le calife Al-Mamn

(813-834), Paris, 1985, p. 39.
65. Les spcialistes du domaine manden y voient les mandens (voir, par
exemple, inasi GNDZ, Problems on the Muslim Understanding of the Mandaeans , ARAM 11-12, 1999-2000, p. 269-279) ; la suggestion que ma faite
Simon C. MIMOUNI selon laquelle il pourrait sagir des elkasates (voir note 6)
mriterait que la documentation islamique sur les sabens des Marais soit
entirement reprise dans cette perspective.
66. Traduit de larabe dit par Edward SACHAU, Chronologie orientalischer
Vlker von Albrn, Leipzig, 1878, p. 206 (voir la traduction anglaise par ID.,
The Chronology of Ancient Nations, Londres, 1879, p. 188). Le passage est
simplement mentionn par RUDOLPH, n. 2, p. 32.



base 67. Lexistence dune secte dite de la corbeille est atteste dans
la documentation manichenne directe, plus prcisment dans le Kephalaion 121 (288, 19 290, 28) qui met en scne un prtre de la secte
de la corbeille comme contradicteur de Mani68 : le texte, lacunaire,
laisse apparatre que lhapax copte nobe est bel et bien comprendre
comme une corbeille fruits 69 (vocation des pommes et du raisin
dans la question du prtre, allgorie de larbre et des fruits dans la
rponse de Mani) ; il parat cependant difficile, mme dans ce kephalaion, de donner un contenu prcis ce groupe religieux, de sorte que
lidentification de cette secte msopotamienne du IIIe sicle aux kantens des sources syriaques reste ltat dhypothse.
Le rapprochement effectu en premier par Dirk Kruisheer (p. 158)
entre le rite des kantens dcrit par Thodore et le laupa, repas rituel
manden pour les morts, est plausible ; on peut dailleurs aller plus
loin. Les ingrdients du repas comprennent, actuellement tout au
moins, outre de la viande, du poisson et du sel, des noix de coco, des
amandes, des noix, des grenades, des coings, des oignons, des grappes
de raisin ainsi que des fruits et lgumes de saison70. Les gawz (syr.
noix , mand. anguza amande ) sont mentionns dans les rites
bizarres dcrits au 1 et, avec lvocation de Nrig/Nergal, on pourrait
avoir affaire un nouveau jeu de mots avec nargila noix de coco
(syr. nargl). Un deuxime jeu de mot prsent dans le texte confirme
lhypothse dune description du laupa manden : kant/kent est bien
un mot syriaque, attest, semble-t-il, seulement par les lexicographes
(comme le mabt du texte F), signifiant corbeille fruits ; fondement, base 71. Lautre tymologie possible passerait par le manden
kinta qui peut signifier communaut (sans renvoyer spcifiquement
aux mandens), mais ne dsigne jamais un lieu de culte72. On aurait
donc affaire un rcit tiologique sur le nom des kantens dont ltymologie, peut-tre atteste par les Kephalaia manichens, est connue
de Thodore , inspir du rite manden du laupa : le nom de kantens
driverait alors de la corbeille de fruits au centre du rite, mais rinterprte comme le nom de leur lieu de culte par analogie avec le nom
des mandens que Thodore croit provenir de manda, nom de leur lieu
de culte, comme nous le verrons plus loin.
67. Voir PAYNE SMITH, col. 1761.
68. Voir la traduction anglaise du chapitre chez Iain GARDNER, The Kephalaia of the Teacher, Leiden, New York, Cologne, 1995, p. 290-291.
69. Voire une treille (communication orale de Jean-Daniel DUBOIS, EPHE,
mars 2012). Pour une interprtation radicalement diffrente, voyant dans le
terme nobe une transcription du nom du dieu babylonien Nab, voir TARDIEU,
p. 6.
70. Voir The Mandaeans of Iraq and Iran, chap. XI Death and Rites for
the Souls of the Dead , p. 178-203, et spcialement p. 187-188.
71. Voir PAYNE SMITH, col. 1761.
72. Voir kinta 2 , Mandaic Dictionary, p. 214-215.



Un autre texte affirme de la mme faon que le nom kantens a

voir avec le lieu de leur culte. Wilferd Madelung cite en traduction
seulement les quelques phrases dAb s al-Warrq sur les kantens (vers 864) : Was die Kantnya, das heit die iymya, angeht,
so rechnen manche Leute sie zu den Christen. Es wird behauptet, da
sie einen Kant genannten Tempel (ba) haben, nach dem sie benannt
werden73. Wilferd Madelung prcise dans la note 19 : In der Handschrift : m-k-n (unpunktiert)-. Mm ist offensichtlich fehlerhaft von
einem Kopisten hinzugefgt. Die Handschrift ist allgemein nur sprlich und hufig falsch punktiert. Il peut cependant ne pas sagir dune
erreur, le prfixe ma- en arabe servant faire des noms de lieu sur
des racines verbales (cf. les noms ddifices cultuels masid, mahad,
mabad). Pour Wilferd Madelung, cette source suffit confirmer ltymologie de Thodore Bar Kona. Il nous semble au contraire que ce
que cet auteur musulman atteste, cest une tradition hrsiologique
commune de driver le nom des kantens de leur lieu de culte : la
dmonstration de lexistence dun groupe confessionnel appel kantens lpoque musulmane ne nous semble donc pas encore faite.

La biographie des hrsiarques

Les deux notices biographiques sur Baay ( 2) et sur Ado ( 4) sont
strotypes : les hrsiarques y sont dpeints qui comme un esclave
paresseux en fuite, qui comme un mendiant paresseux et malade au
point de ne pas mme pouvoir mendier ; leur vocation est prsente
comme une chance de survie mnageant leur paresse (flatterie auprs
des mages, financement dun abri et mendicit), leurs doctrines ne sont
quun compos de dogmes emprunts aux juifs, aux chrtiens et cela
nest que suggr aux manichens pour Baay ; aux marcionites, aux
manichens et aux kantens pour Ado.
Derrire les strotypes de la biographie dAdo, Christelle et Florence Jullien ont su mettre au jour des parallles saisissants avec un
passage des Actes de Mr Mri74 (dates par elles du Ve-VIe sicle), o
laptre gurit un certain Dustay. Elles ont rsum ces parallles dans
le tableau suivant :
Thodore Bar Kona, Livre des
Scholies XI

Actes de Mar Mari 27

Ado venu au pays de Msne, au

bord du fleuve Karoun.

Zab moyen.

Rencontre avec un homme.

Rencontre avec Dousthi.

73. MADELUNG, p. 221.

74. Christelle et Florence JULLIEN, Les Actes de Mr Mri, Turnhout, 2001,
spcialement p. 43-46.



Ado est infirme des mains.

Dousthi est atteint par la gangrne

aux mains et aux pieds.

Sa famille le considre comme


Laptre gurit limpotent inutile

par son disciple Onsime (nom qui
signifie utile ).

Nazarens et partisans de Dostha.

Gurison au nom de Jsus-Christ le


Elles en ont conclu que les Actes de Mr Mri sont un rcit d vanglisation en milieu baptiste 76. Le texte des Actes de Mr Mri claire
son tour la notice de Thodore Bar Kona : cest en effet par celui-l
seul que sexplique le nom de dostens donn dans celle-ci. Il y a
donc dans les deux textes une matire dostenne, savoir un rcit
anti-baptiste circulant dans lglise syro-orientale, centr sur Dustay,
fondateur dune secte baptiste agissant en Msne qui se nomme ellemme nazarens 77. Cest cette matire que Thodore a puis, en la
ramnageant : le nom de lhrsiarque nest plus Dustay, comme dans
le rcit originel, mais Ado sur lequel une nouvelle appellation sectaire
est propose ( adonens ) ; y sont ajouts les noms des membres de la
famille dAdo et la mention de son origine adiabnite.
Le seul lment chronologique dans toute la notice sur les kantens est la mention des rgnes de Yazdgerd (II, 439-457) et de
Prz (459-484) ainsi que de son dit rprimant les religions autres
que le mazdisme ( 1 et 2). Pognon (p. 12) a signal que Michel
le Syrien faisait galement mention des kantens et des dostens dans
le mme cadre chronologique : cette mme poque, Prz, roi des
Perses, et mchant perscuteur des chrtiens, mourut ; et Balou, frre
de Prz, rgna, en 17e lieu, pendant 4 ans. Du temps de ce Balou,
roi des Perses, les chrtiens qui habitaient dans lempire des Perses
jouirent de la tranquillit ; car ce roi ntait pas port la perscution des chrtiens. cette poque parut en Perse lhrsie des kantens et des immondes dosithens78. Contrairement ce que suggre
Pognon, il ne nous semble pas que ce texte de Michel le Syrien soit
une source indpendante de Thodore : le nom eress hrsie et le
verbe dont il est sujet sont au singulier aussi bien dans le texte dont
75. Ibid., p. 46.
76. Titre de leur paragraphe, p. 43.
77. Un hrsiarque du nom de Dosithe est lobjet dun des ouvrages cits
par Abd dans son catalogue : Thophile le Persan a crit contre Dosithe
(dustews) (texte syr. et trad. lat. chez Joseph Simonius ASSEMANI, Bibliotheca
Orientalis, III, 1, Rome, 1725, p. 42). Comme louvrage est perdu, il est difficile de conclure lidentit des deux figures.
78. dition dont ne disposait pas Pognon : Jean-Baptiste CHABOT, Chronique
de Michel le Syrien, t. II, Paris, 1901, p. 151. Les mots en italique, au pluriel
dans la traduction de Chabot, sont corrigs par nous.



disposait Pognon (cit en n. 1, p. 12) que dans celui publi par Chabot
(t. 4, p. 255)79 ; le rgne de Bal est prcisment situ entre ceux de
Yazdgerd et de Prz ; tout cela suggre bien plutt que Thodore est
la source, directe ou indirecte, de Michel le Syrien, qui, pour lhistoire
de la Perse, a bien d avoir recours une documentation non jacobite,
et qui rsume ici la notice de Thodore. Chez Thodore, cette mention
chronologique nest pas tant attache au ministre de Baay qu son
nouveau nom de Yazdnaniz. Elle peut trs bien venir dune tentative
de donner un contexte possible lmergence dune secte juive et/ou
chrtienne, voire manichenne, qui aurait emprunt des traits mazdens,
auquel cas elle napporte aucune information sre sur Baay, Yazdnaniz
ou les kantens.
Les deux notices biographiques fourmillent de noms de personnes.
Les plus cohrents, voire les plus authentiques, concernent la famille
Le caractre manden de certains des noms des membres de la
famille dAdo a t relev par Pognon. On retrouve dans les noms
mkw et kwy le traitement manden (kua), avec dissimilation des
emphatiques q et , de laramen commun, conserv en syriaque, qu
vrit, sincrit, foi .
Parmi ces noms, quatre peuvent tre rapprochs de noms mandens :
ilmai et Nidbai pour mly w-ndby, qui sont un couple dutri prposs
aux eaux vives, Br-Hiia pour bry le fils de la Vie , personnage
mythologique et anthroponyme, itil pour tyl, le nom de Seth, galement utilis comme anthroponyme80. En outre, le nom byzk peut
ventuellement tre rapproch des anthroponymes Ska-Iauar, SkaManda, Sku-Hiia, dont le premier lment est le verbe ska il a veill
(*abizka signifierait donc mon pre a veill )81. Pognon, pour qui les
quatre premiers noms renvoyaient quatre personnages clestes de la
religion mandenne se demandait si, comme le faisaient souvent les
auteurs chrtiens, il [Thodore] na pas pris des divinits ou des gnies
pour des personnages historiques et si dw ne serait pas le nom de
quelque gnie dfigur par lui ou par les copistes 82. Cette hypothse
concorderait assez avec le caractre symbolique du nom de la mre :
emku la mre de la vrit . Si, au contraire, on considre que la
plupart de ces noms peuvent tre des anthroponymes, on peut rapprocher la liste de la pratique des copistes mandens, dans linvocation
situe en tte de leurs manuscrits, de donner ensemble le nom du pre,

79. Observation note par Joseph THOMAS, Le mouvement baptiste en Palestine et Syrie (150 av. J.-C. 300 ap. J.-C.), Gembloux, 1935, n. 2, p. 216.
80. Pour toutes les rfrences, voir nidbai , Mandaic Dictionary, p. 297,
br hiia , p. 69, itil , p. 464.
81. Voir ska iauar , ska manda , skuhia , Mandaic Dictionary, p. 330.
82. POGNON, p. 11.



de la mre, des frres et surs et, ventuellement, des enfants83 ; ce

serait peut-tre l la source de Thodore Bar Kona.
Jean-Daniel Dubois nous a suggr que le nom dw pouvait tre
rapproch du manichen Add. En suivant cette piste, ce ne sont pas
moins de cinq des neuf noms des membres de la famille dAdo ( lire
Addo) quon retrouve parmi les proches de Mani cits dans le Codex
manichen de Cologne : (CMC 165, 6), (CMC 6, 1),
(CMC 14, 3), (CMC 106, 19), (CMC
114, 6), auxquels on peut ajouter le nom du patriarche Seth ().
Quatre de ces noms tant communs aux sources mandennes et manichennes, le rapprochement est dabord rvlateur de la proximit entre
le milieu elkasate dorigine de Mani et de son pre et du mandisme.
Le couple ilmai et Nidbai tant typiquement manden, nous prfrons
penser que les noms attests galement dans les sources manichennes
(avec un doute sur Ado/Addo, trop connot, Emku, trop symbolique,
et Dabd, sans tymologie satisfaisante) ont pu tre ports par des mandens, mais que les hasards de la transmission ne nous les ont pas
encore livrs.
Pour le nom de lhrsiarque, il convient toutefois de tenir compte
dun jeu de mots de Thodore sur les diffrentes dnominations de
lhrsie : Mais le nom qui leur convient est celui dadonens. Le
sens littral (vu que leur fondateur sappelle Ado, ils devraient tre
appels adonens) nest pas satisfaisant puisquon sait prsent que,
dans lhrsiologie syriaque, le nom du fondateur devait tre Dustay,
do leur nom dostens . Ce que Thodore nous dit est bien plutt : ils naimeraient pas quon les appelle adonens, mais cest adonens quil faut les appeler ; ds lors cest le nom dAdo qui drive de
ltiquette polmique adonens . Or, si on ne disposait pas du nom
dAdo, on pourrait se dire que le syriaque adny drive du nom
Adnay, cest--dire la transcription syriaque dun des noms de Dieu
dans le judasme. Voici enfin la trace de ce que Thodore disait de
Baay : Il avait vol aux juifs [linterdiction] de manger de la viande
de porc, le nom du Seigneur Dieu au Pentateuque. Le nologisme
de Thodore sinsre bien dans la polmique anti-mandenne, dans la
mesure o les mandens, dans leurs nombreuses polmiques contre les
autres religions, reprochent aux juifs dadorer en Adona nul autre que
ami84, le Soleil, ainsi par eux divinis.
son tour, le ddoublement du nom de lhrsiarque du 2 parat
suspect : le nouveau nom, iranien, de Baay, Yazdaniz, peut trs bien
renvoyer une autre figure. Le dossier a t repris par Dan Shapira
et nous en donnons ici les grandes lignes. Les mandens polmiquent
contre un groupe appel iazuqaiia dans quelques passages du Ginza
iamina, dont lhistoire rdactionnelle est complique. Les membres de
83. Voir, par exemple, GY I (1).
84. Voir adunai , Mandaic Dictionary, p. 7.



ce groupe semblent suivre des rites zoroastriens (ils adorent le feu, par
exemple), sont dcrits comme issus de ami, de mme que les juifs,
invoquent Jsus sous son nom syriaque et ont un rite qui consiste
attacher la crosse mazdenne (baresma) sur les paules la manire
dune croix ; leur nom est rapprocher de Yazdaniz, mme si, dans le
dtail, les deux formes semblent prsenter des difficults dinterprtation en moyen-perse. Au terme de son analyse, Dan Shapira conclut :
The Christian features of the Iazuqaiia were most probably added
at the latest stage of the redaction, when Zoroastrianism was irrelevant
for the Mandaeans and no longer posed a threat to them. We should
therefore identify these Iazuqaiia with Zoroastrianism, orthodox or
not, and with them alone85.
Si lon suit ces conclusions, il apparat donc que le nom de Yazdaniz, le compos doctrinal de Baay et la mention de la croix mise sur
lpaule gauche des auditeurs ont une source hrsiologique mandenne,
ce qui nest pas tout fait improbable, puisque nous avons vu que
Thodore mentionnait Dinanukt, le personnage central dun chapitre du
Ginza consacr la polmique contre les autres confessions. Si lon ne
suit pas ces conclusions, il faut supposer que les iazuqaiia dsignent,
dans les premiers sicles de la priode musulmane en basse Msopotamie, chez les mandens comme chez les chrtiens, un groupe religieux
syncrtique empruntant la fois au zoroastrisme, au judasme et / ou au
christianisme. Quoi quil en soit, le lien entre Baay et Yazdaniz savre
tre une cration de Thodore, et le rapport avec les kantens parat l
encore artificiel.

Les noms actuels de la secte

Quand il cite les noms des sectes donnes pour actuelles, ce nest
pas des kantens que Thodore parle, mais des dostens. Ceux-ci apparaissent en trois endroits diffrents : dans le titre de la notice ( dostens ), dans lnumration de leurs noms suivant les rgions ( mandens , makanens , gens de Celui qui fait le bien d-bt bed
bt en Msne ; nazorens , gens de Dustay d-bt dustay en
Bt-rmy) et, la fin de lnumration, dans une suggestion faite
par lauteur de la notice ( le nom qui leur convient est celui dadonens ).
Lnumration a t si mal comprise par les copistes que, dans lensemble des manuscrits, la rgion du Bt-rmy a t prise pour un
des noms de la secte : w-d-bt rmy, avec la prposition d, sur le
modle des qualifications des sectes d-bt bed bt et d-bt dustay, cest--dire ceux des gens de [] ; les diteurs, sans corriger le

85. SHAPIRA, art. cit., p. 264. Voir aussi les notes explicatives de LIDZBARSKI,
Ginza, p. 225.



texte, ont raison traduit sans tenir compte du d86. Ce passage ayant
t mal transmis, on peut se demander si les copistes nont pas galement ajout un w entre mandy et makny, dans la mesure o le
terme manden manda dsigne la fois le personnage de Manda -Hiia
et, par mtonymie, le lieu de culte des mandens, galement appel
makna87. Nous aurions donc affaire deux noms pour la Msne
( mandens , glos templiers , et gens de Celui qui fait le bien )
et deux pour le Bt-rmy ( nazorens et gens de Dustay ). La
tendance multiplier les noms des sectes, dans le but de montrer leur
peu dunit doctrinale, est un trait propre lhrsiologie et na donc
t quaccentue par les copistes.
Tous ces noms ont des sources quil sagit de distinguer : les noms
dostens et gens de Dustay , celui-ci explicitant celui-l, proviennent de la littrature hrsiologique ; mandens (peut-tre glos
templiers ) et nazorens sont les noms que se donnent les mandens eux-mmes. Les mandens nont en fait pas de nom autre que
mtaphorique pour qualifier lensemble de leur communaut : manden (sg. mandaia, pl. mandaiia) dsigne en effet les lacs, nazoren (sg. nauraia, pl. nauraiia) les membres du clerg et les gens
instruits88. Pour le nom dadonens, nous avons propos dy voir un jeu
de mot avec Adona.
Lexpression Celui qui fait le bien se retrouve dans la littrature
mandenne, comme, par exemple, dans le GY XV, 18 (375,15sq., trad.
p. 377) : ubx l-man -abid ab uai uai l-man -abid bi Heureux celui
qui fait le bien ! Malheur, malheur celui qui fait le mal ! Peut-tre
les mandens se prsentaient-ils sous ce nom auprs des autres, ou bien
Thodore connaissait-il un texte de plus, comprenant une expression du
type man -abid ab, quil na cependant pas cit. Une autre interprtation un peu mieux taye est possible ; nous avons vu que la principale diffrence entre les deux recensions du texte sur le Fils de la
lumire rside dans lexpression bar nuhr chez Thodore Bar Kona
et br abia Fils des [hommes] bons dans le Ginza. Peut-tre avonsnous ici la trace de lintgration dans lanthologie de Thodore de cette
pice : la lecture du document manden, Thodore identifie le personnage manden aux Fils de la lumire dj voqus dans sa notice
sur les manichens et quil sait attests chez les mandens comme pithte anglique (cf. texte E) ; au moment de citer le texte lui-mme, il
adapte alors lgrement lexpression de br abia en bar nuhr.

86. La prposition b dans nest pas ncessaire. Voir Louis COSTAZ, Grammaire syriaque, Beyrouth, 19974, 720, p. 189.
87. Voir manda 2 , Mandaic Dictionary, p. 247, et makna , p. 255.
88. Voir The Mandaeans of Iraq and Iran, p. 4. En dernier lieu, il pourrait
sagir dun hritage elkasate, selon la suggestion de Simon C. MIMOUNI, Les
nazorens. Recherche tymologique et historique , Revue biblique 105, 1998,
p. 208-262, en particulier n. 9, p. 211.



Lintention hrsiologique de Thodore Bar Kona est claire : il

sagit pour lui de regrouper en une histoire cohrente des hrsies
dont il connat les noms et les acteurs par la tradition hrsiologique
et des sources externes. Certains de ces noms sont manifestement le
produit dobservations personnelles : les dsignations mandens et
nazorens , qui sont encore en usage chez les mandens mais dont
la rpartition dans des zones gographiques diffrentes trahit une rcriture hrsiologique ; Baay, dont le nom, jusque-l sans parallle, na
sans doute pas t invent de toutes pices. Dautres ont toute chance
davoir t puis des sources hrsiologiques : les kantens (attests
dans la littrature syriaque et musulmane), les dostens (hrsiologie
syriaque), Yazdaniz (peut-tre attest par ltiquette iazuqaiia de la
littrature hrsiologique mandenne). Quelques noms sont des inventions : la secte de Nergal, les nrigens, les adonens.

Bilan de ltude
Au terme de cette tude, on peut tenter le bilan suivant.
Des mandens, Thodore sait quils habitent en Msne et en Betrmy, quils se nomment dau moins deux noms diffrents (mandens et nazorens), que leur lieu de culte (syr. makn) sappelle
manda, que leurs rites principaux sont le baptme (dont il connat le
nom manden mabuta) et, peut-tre, le repas rituel pour les morts. Il
possde nombre de leurs textes dont il comprend peu prs la langue89.
Cest pour lui la seule secte encore vivante. Il lui accole ltiquette de
dostens , hrite de la tradition syriaque, reprend ce qui se dit de
leur fondateur Dustay, mais en change le nom pour les affubler dun
nouveau sobriquet infamant adonens , ce qui suppose quil connat
aussi la place dAdona dans la pense mandenne. Il a trouv dans un
texte manden (ou ventuellement manichen) une liste de noms de
personnes (il peut sagir soit de noms tirs dune invocation de copiste
soit dun matriau mythologique dont on a perdu la trace). Cest le seul
groupe dont on peut dire avec certitude quil connat deux plus quil
nen cite.
Des kantens, en revanche, il ne sait pas grand-chose : il connat le
nom de leur groupe et son tymologie (la religion de la corbeille ),
sait quil faut les rattacher Nergal (2 R 17, 30), ce qui est peuttre un hritage douvrages hrsiologiques perdus, comme le suggre
89. Les sources mandennes font aussi tat de livres mandens possds
par des membres dautres confessions : un copiste du nom de Nukraia, fils de
itil, originaire de la ville de b, dans la rgion de Wsi, prcise dans le premier colophon du Qulasta quil a consult sept manuscrits dont lun tait dans
la librairie dune maison de gens du livre (voir Jorunn Jacobsen BUCKLEY,
Conversions and Other VIIIth Century Community Issues , Le Muson 121,
3-4, Leuven, 2008, p. 291, et ID, The Great Stem of Souls, Piscataway, 2005,
p. 189).



le dveloppement de Brn sur les sabens ; le reste de ce quil dit

deux sapplique en fait aux mandens (place dAbel, rle du baptme,
texte C, laupa).
Il connat le nom de Yazdaniz qui, pour linstant, na de parallle
que dans lhrsiologie mandenne (les iazuqaiia du Ginza iamina), o
ils sont prsents comme un groupe judo-chrtien avec des traits mazdens, peut-tre concurrent des mandens eux-mmes. Peut-tre est-ce
cette matire que Thodore a intgre dans sa notice sur Baay (pratiques alimentaires judasantes, croix sur lpaule, culte du feu, adoration dAdona-ami). Peut-tre le nom de Yazdaniz est-il driv de
iazuqaiia, de mme quAdo est tir dadonens.
De Baay, peut-tre connat-il un texte manichisant sign par lui
(le texte A), ce qui le pousse en faire le fondateur du groupe prcdent.
Son travail consiste historiciser lensemble des tiquettes quil
connat et leur donner une paisseur, en inventant une histoire
ancre dans lhistoire sainte, en donnant des noms aux hrsiarques
et en rpartissant dans deux notices distinctes un ensemble de textes
authentiques qui ne se rpartissent pas, ne serait-ce que pour des
raisons linguistiques, de cette faon.
On ne peut pas donc conclure de ces notices lexistence, lpoque
de Thodore, dun groupe concurrent des mandens appel iazuqaiia par
les mandens et kantens dans la tradition hrsiologique syriaque.
Pour les priodes plus anciennes, les quelques attestations de ces derniers dans les sources syriaques (Actes de emn Bar abb90, dats
du Ve sicle, Six explications sur les ftes liturgiques de Cyr ddesse91,
de la premire moiti du VIe sicle, et Contre les svriens, les manichens, les kantens et les mandens de Nathanal92, vque de Srzr
entre 580 et 628), voire manichennes (Kephalaia, crits la fin du
IIIe, dbut du IVe sicle), sont trop allusives pour tre exploitables. Les
attestations des kantens dans les sources musulmanes ultrieures, fort
strotypes, napportent au final que peu dinformations authentiques

90. Voir Michael KMOSKO, De Beato Simeone bar Sabbae , Ren GRAF(d.), Patrologia syriaca I/2, Paris, 1907, col. 823-824. Ce texte cite deux
sectes, dont lune pourrait tre les mandens, lautre les kantens, mais les
variantes sont telles dun manuscrit lautre que le passage ncessiterait un
rexamen approfondi (cf. RUDOLPH, n. 4 et 5, p. 35).
91. William F. MACOMBER (d., trad.), Six Eplanations of the Liturgical
Feasts by Cyrus of Edessa, Louvain, 1974, CSCO 355, p. 11 et CSCO 355,
p. 10.
92. Telle est la description dun ouvrage, perdu, cit par Abd dans
son catalogue (ASSEMANI, op. cit., p. 224). Lidentification de ce Nathanal
lvque de Srzr du mme nom a t propose par Adda SCHER, tude supplmentaire sur les crivains syriens orientaux , Revue de lOrient chrtien 11,
Paris, 1906, p. 12, et par BAUMSTARK, p. 129-130.



sur les kantens93. Du foisonnement suppos du mouvement baptiste,

il ne reste donc quasiment plus aucune trace lpoque de Thodore
Bar Kona.

Annexe : traduction des notices sur les mandens

Dans la traduction suivante, le dcoupage en paragraphes sappuie
sur le texte, notamment sur les enclitiques. La numrotation des paragraphes (1-6) est emprunte ltude de Dirk Kruisheer, les citations
ont t indiques par des lettres majuscules (A-F) et les diffrentes parties du texte D par des minuscules (a-d). La vocalisation de la plupart
des noms propres est bien entendu hypothtique.
[ 1 = 84 CSCO 432] Sur les kantens.
Comme les insenss kantens clament que leur doctrine drive
dAbel, il faut indiquer do elle est [issue].
Aprs que Goliath, le hros des Philistins, eut t tu par David,
gns de dire que cest [frapp] par la pierre dune fronde que leur
hros tait mort, ils affabulrent et dirent : Un guerrier, disaient-ils,
arm dun bton de fer, est venu du camp des Hbreux, la frapp et la
tu. Ils lui firent une statue et clbraient chaque anne le meurtre la
faon dune guerre. Se mettant en rang et se tenant les uns en face des
autres en colonnes, les pontifes de Dagon, qui taient les pontifes de
Goliath, lacraient leurs corps avec des [armes en] fer, se frappaient les
uns les autres avec des btons et couraient les uns aprs les autres en
simulant un combat. Alors, lun deux sapprochait, arm dun bton de
fer, frappait la statue et la renversait, limage de la chute de Goliath ;
ils criaient au moment de sa chute : Ainsi lhumble a tu le hros, le
faible [a tu] le fort ! Cest ainsi quils firent dans leur rgion pendant
Or, quand Nabuchodonosor emmena les Philistins en captivit, il
brisa la statue de Goliath. Quand les pontifes de Dagon furent arrivs
Babylone, ils fabriqurent une grande [statue en] bois limage de
Goliath et placrent son sommet du fer comme le casque de Goliath.
Ils se mettaient en rangs, et lun des pontifes sagenouillait devant
elle et faisait semblant de se frapper lui-mme dun poignard ; ils coupaient un grand rameau pais et y suspendaient des noix et des produits comestibles ; un des pontifes le portait aprs stre dshabill et
avoir mis une ceinture dtoffes teintes autour de ses reins ; ils sortaient
dans le dsert, lanaient des flches en criant et en disant : La flche
a vol ! Les hommes criaient avec les femmes : Les cdres ont t
abattus, et moi je suis tranquille ; les hros ont t anantis et moi je

93. Voir les rfrences chez MADELUNG, p. 221-224.



suis tranquille ! , comme sils pleuraient le meurtre de Goliath. Ils faisaient cette folie au mois db et [au mois de] tern.
Or, les Chaldens, ayant trouv dans leur zodiaque un vieux diable
agit appel par eux Nergal, appelrent cette secte de son nom ; cette
secte resta ainsi jusquau roi Yazdgerd. Aux jours de Prz, Baay y
introduisit un autre changement horrible.
[ 2 = 85a CSCO] Sur Baay, do il est [issu].
Les gens de cette secte avaient un chef nomm Papa, fils des Klly
de Gawkay. Ce Papa avait un esclave du nom de Baay. Celui-ci, par
paresse, senfuit de sa servitude ; il se cacha chez les juifs et de l
partit chez les disciples de Mani. Il avait rassembl et arrang un peu
de leurs paroles et des mystres de leur magie. Aux jours du roi Prz,
lorsquune sentence sortit contre les idoles et leurs pontifes selon
laquelle seule resterait la secte des mages, quand Baay vit que sa secte
prissait, il flatta les mages et adora les luminaires ils reurent mme
du feu quils installrent dans leurs demeures et changea son nom de
Baay en Yazdaniz94, ce qui sinterprte il y a des dieux . En effet,
il avait vol aux juifs [linterdiction] de manger de la viande de porc,
le nom du Seigneur Dieu au Pentateuque, aux chrtiens le signe de la
croix, quil mettait sur lpaule gauche de ses auditeurs ; la croix, ils
disent quelle est le mystre de la limite entre le Pre de la grandeur et
la rgion infrieure.
[ 3 = 85b CSCO] Un peu de sa doctrine.
[A] Il dit en effet : avant toute chose il y avait une seule divinit ;
celle-ci se divisa en deux, et cest delle que furent le Bon et le Mauvais ; le Bon rassembla les luminaires, le Mauvais les tnbres. Alors
le Mauvais sentit ( ?)95 et monta faire la guerre au Pre de la grandeur,
et le Pre de la grandeur sut quil y aurait combat ; il appela de lui
une voix et de cette voix fut cr par lui le Seigneur Dieu. son tour,
le Seigneur Dieu appela sept voix et sept puissances sortirent de lui.
Alors sept dmons montrent et lirent le Seigneur Dieu et les sept
puissances [issues] de lui, et ils emmenrent captive la nature de lme
[loin] du Pre de la grandeur. Des diables et des dmons [au nombre
de] sept et [de] douze se mirent faire Adam, le premier homme. Le
Seigneur Dieu vint, dtruisit Adam et le reconstitua.
[B] Et ils disent encore quil y a dix cieux et ils les nomment de
noms absurdes : rdy, disent-ils, Mrdy, rdbly, Sprsgl, Hrbbl, Qwdy,
Mqdy, Lsy, Msy et ayy, [B] en disant propos de lui que cest
lui qui apporta une offrande du jardin dAdam, des ppins de la bouche
des grenadiers et des fleurs de figuiers et de dattiers.
94. Yzdnnyz, var. POGNON yzdny, Yzdnny.
95. Arge, peut-tre pour rga sexcita .



[C] Et ils disent encore du personnage quils appellent Fils de la

lumire : Je ( ?) et je mavance vers les mes ; et lorsquelles me
virent, elles volrent vers moi et madressrent mille saluts ; elles
gmirent et me dirent : Fils de la lumire ! Va dire notre pre : Quand
les captifs seront-ils dlivrs, et le repos pour les affligs qui [sont]
dans laffliction, et le repos pour les mes qui endurent la ncessit
dans le monde ? Moi, jai parl et leur ai dit : Lorsque lEuphrate sera
sec son embouchure et [que] le Tigre, par son coulement, sera
sec en tout temps et que les torrents seront plat, alors pour les mes
[sera] le repos.
Cen est assez de la multitude de son impit.
[ 4 = 86a CSCO] Hrsie des dostens quenseigna le mendiant
Ado, ce quon dit, tait Adiabnite et vint, en tant que mendiant,
avec sa parent dans la rgion de la Msne. Le nom de son pre tait
Dabd, celui de sa mre Emku ; ses frres taient emlay, Nedbay,
Barayy, Abzk, Kuay et tel. tant venu au fleuve lay, ils
trouvrent un homme du nom de Papa, fils de Tinis, lui demandrent
laumne selon leur coutume et lui demandrent de recevoir chez lui
le paresseux Ado qui, cause de sa maladie, ne pouvait pas mendier.
Papa le remit des gardiens de dattiers ; mais quand les gardiens de
dattiers se furent plaints [en disant] Il nous est inutile , Papa lui
construisit un abri ct de la route, pour quil mendit sa nourriture
auprs des passants de la route. Ses compagnons finirent par se rassembler et venir auprs de lui et l ils frappaient des cymbales, selon la
coutume des mendiants.
En Msne, on les appelle mandens, makanens [= templiers],
gens de Celui qui fait le bien96 ; dans le Bt-rmy97, nazorens, gens
de Dustay ; mais le nom qui leur convient est celui dadonens. Leur
doctrine est un compos [emprunt] aux marcionites, aux manichens
et aux kantens.
[ 5 = 86b CSCO] Un peu de leur doctrine.
[D] [a] Ils disent en effet quavant que ne fussent le ciel et la terre,
il y avait de grandes puissances qui reposaient sur les eaux. Elles
avaient un fils, quelles appelrent Abtr. Abtr eut un fils auquel il
donna le nom de Ptael. [D] [b] Et ils disent quAbtr lui commanda :
Va, condense la terre pure, tends le ciel pur ( ?) la colonne, cre
et dispose lhumanit, lun [comme] fils de lautre, noircis leur tte

96. Comprendre peut-tre (en supprimant le w de w-makny) : mandens

templiers et gens de Celui qui fait le bien .
97. W-d-bt, corriger : w-bt.



dune paume98 qui est dans les eaux, et leur barbe avec les poissons
de locan ; quils vivent, dit-il, et durent deux cent soixante-douze ans,
dit-il. [c] Ptael alla, disent-ils, et ne fit pas selon ce que lui avait
command son pre, mais il cra et disposa les dix Portes et les douze
Portes ; il leur jeta un coup de pied, mais il ny jeta pas desprit ni
dme. [d] Alors, disent-ils, quAbtr tait assis dans les sept firmaments, il leva les yeux, vit Ptael et lui dit : Que des liens ( ?) soient
sur toi, Ptael ; javais dit : Va, dispose lun [comme] fils de lautre !
et il na pas cout ce que je lui ai command.
[E] Et ils disent encore que les anges [utri] et Ptael se levrent et
quils dirent Abtr : Ne dispose pas des liens ( ?) sur Ptael, ton
fils. Il leur dit : Que ces liens ( ?) soient sur Ptael jusquau jour du
jugement et aux heures du salut, quand sera [venue] la vie des morts,
un jour et demi, que le Christ savancera et viendra dans le monde, que
la brique des fondations parlera et dira : Je confesse le Christ. [Abtr]
nayant pas cout ni lui ni les anges [utri], fils de lumire, [Ptael]
sen alla, se tut et reut les liens de son pre. Il dit99 quil jeta sur lui
la chane qui remplit le monde entier et enfona en lui la pointe [ou :
lembryon ?] qui va depuis la terre jusquau ciel, et voici quil est assis
maintenant dans les liens, jusquau jour du jugement et aux heures du
salut, jusqu ce que la brique des fondations parle et dise : Je suis le
[F] Ils disent, dans leur cantique quils appellent Acheminement
vers les magiciennes : Ainsi, disent-ils, lesprit mauvais qui est appel
Hamgay et Hamgagay par la chane lpreuse, Mrdyq100, Lbrnyt et Tty
Hzyt101, ny, Nanay, Bl [= Jupiter] et Belat [= Vnus]102 par la
terre des Romains, Dyq, Mrdyq et Gwztny par lInde, rn et Aphrodite
par lOccident, Mgard alyteh [lintranquille] par lOrient, Emm [la
mre] et Mamani par rt d-ayyy [= al-r], la tte desquelles,
disent-ils, sige la vieille mbyw ; telles, disent-ils, sont toutes les magiciennes. Elles allrent tuer par leurs tours de magie les taureaux, les
bliers, les chevaux, les chameaux et les brebis ; elles desschrent les

98. Il ny a aucune raison de douter du sens paume , comme lont fait

les traducteurs (POGNON, p. 238), mme lorsquils ont compris ainsi (RUDOLPH,
p. 258, CSCO, p. 258, KRUISHEER, p. 166) : le mot upn est attest aussi bien
en syriaque (PAYNE SMITH, col. 1347) quen manden ( hupna , Mandaic Dictionary, p. 136-137) avec le sens de contenu dune paume .
99. POGNON (p. 2, p. 242) propose de corriger par un pluriel : ils disent
que... . La confusion entre singulier et pluriel peut se comprendre comme une
faute de copiste, vu les nombreuses rptitions des deux formes dans tout le
100. Rapproch par POGNON (p. 244) du dieu babylonien Marduk.
101. Hwzyt : gentilice fminin syriaque du Bt Hzy (zistn) (voir
KRUISHEER, p. 167).
102. W-blty (var. D : w-blt) : syr. blat, belat, belet Vnus (plante) (voir
PAYNE SMITH, col. 519).



semences et les plantes jusque chez Adam, le premier homme ; elles

excitrent Adam, qui [les] avait purifies de leur lpre, et elles firent
Adam des tours de magie et le jetrent dans de dures tribulations,
jusqu ce que vnt Abel, qui attacha Adam et le purifia, ce par quoi ils
donnent le type de ce quils appellent le baptme.
[G] Et ils disent encore au sujet de Dnns, scribe des sectes103, et
du petit D, aussi des [choses] dlirantes jusqu ce jour.
[ 6 = 87 CSCO] Hrsie des nrigens.
Cest de Can que drive lhrsie de ceux-ci. Aprs la mort de Can,
ses fils se rassemblrent et dirent : Lesprit de notre pre Can na pas
de repos sur la terre, car il a tremblement et crainte [= Gen 4, 14]
cause du meurtre dAbel. Ils firent un sanctuaire et y alignrent des
nourritures, afin que peut-tre lesprit de Can vnt sjourner sur lui.
Les fils de Can se rassemblrent auprs de ce sanctuaire comme une
maison de pleurs et ils donnrent Can, leur pre, le nom de Nrig,
en disant : Notre pre dsire le repos [ny reg] ; cest Can que
ses fils appelrent Nrig. Les fils de Seth se rassemblrent eux aussi et
dirent : Faisons, nous aussi, une maison de runion pour Abel, frre
de notre pre ; ils la firent et la nommrent kuwwn [= blme], en
disant : Il a t pour nous un blme. Cest le kant [= la corbeille
fruits] des insenss kantens.

103. Ou livre des sectes .

Universit Concordia, Montral


Very few studies have engaged in a synchronic reading of the Gospel
according to Thomas. But such a perspective contributes to a better
understanding of many of the Thomasine logia, as well as an appreciation of the doctrinal particularities of such an enigmatic text. This article is a test case which presents an analysis of the analogous characterization of Jesus, light, and the living Father in the Gos. Thom. The
Thomasine tradition portrays Jesus and his Father in terms of twinship.
This is what lies behind the similar characteristics of both figures.
Trs peu dtudes entreprennent une lecture synchronique de lvangile selon Thomas. Pourtant une telle approche permet de mieux comprendre le sens de plusieurs logia thomasiens, ainsi que de saisir les
particularits doctrinales de ce texte nigmatique. Par voie dexploration, cet article propose lexamen de la caractrisation analogue des
figures de Jsus, de la lumire et du Pre vivant dans lEvTh. Ce qui
est luvre dans la manire dont la tradition thomasienne traite de
ces figures est un principe de gmellit. Jsus est en quelque sorte le
jumeau de son Pre, cest pourquoi il a les mmes caractristiques.
Dans sa monographie sur lvangile selon Thomas (EvTh) publie
il y plus dune trentaine dannes, Jacques-mile Mnard reconnat
que les correspondances entre la figure du Pre et celle de Jsus de
lEvTh pourraient sapparenter une sorte de modalisme primitif2.
Curieusement, aucun commentaire rcent ne fait mention du point de
vue de Mnard. La plupart des approches sur lEvTh sintressent

1. Cet article est une version modifie dune confrence prononce lors de
la runion annuelle de lAssociation pour ltude de la littrature apocryphe
chrtienne (ALAC), tenue du 24-26 juin, 2010 Dole, France. Je remercie
Jean-Michel Roessli et le comit de lALAC pour leur invitation. Ma participation cette rencontre a t rendu possible par une gnreuse subvention de
recherche scientifique du Fonds qubcois pour la recherche sur la socit et
la culture (FQRSC).
2. J.-. MNARD, Lvangile selon Thomas (Nag Hammadi Studies 5), Leiden, Brill, 1975, p. 28 et 125.


Apocrypha 23, 2012, p. 209-221



plutt lhistoire de la rdaction de la collection qu linterprtation

des paroles caches. Dans cette tude, nous explorerons les rapports
entre trois figures convergentes de lEvTh : Jsus, la lumire et le Pre
vivant. Certains aspects de la christologie thomasienne seront examins sous langle de lintra-textualit, en considrant la manire dont les
logia de la collection peuvent fournir une piste dinterprtation nouvelle
pour ce qui a trait ces figures de lvangile.

Une lecture synchronique et intra-textuelle

Lapproche adopte dans cet article sera principalement synchronique, une perspective qui prend la collection des dits thomasiens
comme un tout. Cela ne signifie pas quil faille ignorer larrire-plan
de lcrit, mais lanalyse tentera de suivre la recommandation de JeanMarie Sevrin lorsquil crit quil faut :
() aborder le texte en lui-mme et tenter son exgse, sans ignorer
certes les horizons littraires ou doctrinaux sur lesquels il se dcoupe,
mais en sattachant en premier lieu ses cohrences propres ; comment
dit-il, que dit-il ?3

Cela signifie quil faut avant tout chercher comprendre la cohrence de Thomas comme texte en soi. Il arrive, cependant, que des
rfrences extra-textuelles soient utiles et peuvent jeter un clairage
supplmentaire sur le sens de la collection. Deux autres spcialistes de
lEvTh, des Amricains, partagent un avis semblable celui de JeanMarie Sevrin quant lexgse du texte en question :
I suggest that we should now seek literary questions and literary
answers about the Gospel of Thomas. () The text obviously must have
meant something, () perhaps the arrangement or sequence of statements and groups of statements does indeed convey meaning, though
not necessarily the sort of meaning that we see even in other gospel sayings or in wisdom books. To explore this possibility requires adopting a
more literary sensibility, a focusing of attention on reading the text in its
own terms, searching out its hermeneutical soteriology. The task is difficult, and the meanings provided by stark juxtapositions are not always
obvious. Perhaps that obscurity is part of the point4.

3. J.-M. SEVRIN, Un groupement de trois paraboles contre les richesses

dans lvangile selon Thomas. EvTh 63, 64, 65 , dans J. DELORME (d.), Les
Paraboles vangliques. Perspectives nouvelles. XIIe Congrs de lACFEB, Lyon
1987 (LD 135), Paris, Cerf, 1989, p. 426.
4. P. SELLEW, The Gospel of Thomas : Prospects for Future Research ,
dans J. D. TURNER et A. MCGUIRE (ds), The Nag Hammadi Library After Fifty
Years. Proceedings of the 1995 Society of Biblical Literature Commemoration
(NHMS 44), Leiden, Brill, 1997, p. 335.



Pendant des dcennies, laccent a t mis sur la reconstruction de

lhistoire du texte. Il est clair que la recherche historique est utile pour
comprendre le milieu de production dune uvre, mais trop souvent,
les chercheurs ont nglig lexgse du texte en soi. Il semble que nous
sommes un point o il faut miser davantage sur lexgse littraire du
texte que sur lhistoire de sa production.
Trs peu de spcialistes ont donc t attentifs au programme de lecture de lEvTh dans leur interprtation des dits cachs5. Cela exige
une sensibilit pistmologique particulire, car lexgte doit sengager lui-mme dans la programmation de lecture sil souhaite trouver
un sens aux logia thomasiens. Aprs plusieurs annes consacres une
recherche purement historique, Stephen Patterson livre en ces termes
ses impressions sur linterprtation de lEvTh :
Thomas, interestingly, operates on a heuristic model that might
be regarded as thoroughly post-modern : the real meaning of the text
resides not in the text itself, but in the reader, the seeker after wisdom
and insight. () When the implied author indicates that the meaning of
the text will not be obvious, our approach must be completely different. Interpreting Thomas must be a matter not of disclosing the intended
meaning of the implied author, but rather, exploring the possibilities of
meaning that a particular saying might hold for an ancient reader/hearer.
() One should probably assume that the author, or more properly, the
collector, is content not to give the seeker too much of an agenda, but
to leave more room for thought6.

Qui souhaite trouver linterprtation des paroles caches doit sinvestir dans une qute de sens. Cest dailleurs ce quaffirme le logion
premier : Il a dit : Celui qui trouvera linterprtation de ces paroles
ne gotera pas la mort. Au lieu de voir lEvTh comme un tissu de
paroles dsorganises, relies tout au plus par des mots crochets, mais
sans aucune cohrence interne7, le logion premier semble dire quil y
a un rapport entre les dits de la collection. Derrire lapparence de
5. R. CAMERON a bien identifi la cl hermneutique de lEvTh dans les premires lignes de luvre ; voir Ancient Myths and Modern Theories of the
Gospel of Thomas and Christian Origins , dans R. CAMERON et M. P. MILLER
(ds), Redescribing Christian Origins (SBLSymS 28), Atlanta, Society of Biblical Literature, 2004, p. 105-106.
6. S. J. PATTERSON, The Gospel of Thomas and Historical Jesus Research ,
dans L. PAINCHAUD et P.-H. POIRIER (ds), Coptica Gnostica Manichaica.
Mlanges offerts Wolf-Peter Funk (BCNH Section tudes 7), Qubec/Louvain/Paris, Presses de lUniversit Laval/Peeters, 2006, p. 680.
7. S. J. PATTERSON, The Gospel of Thomas and Jesus, Sonoma, CA, Polebridge Press, 1993, p. 99-102 ; M. MEYER, The Beginning of the Gospel of
Thomas , dans Secret Gospels. Essays on Thomas and the Secret Gospel of
Mark, New York, Trinity Press, 2003, p. 39-41 ; S. L. DAVIES, The Gospel of
Thomas and Christian Wisdom Second Edition, Oregon House, CA, Bardic
Press, 2005, p. 149.



dsorganisation il y aurait peut-tre une certaine cohrence dcouvrir8.

En supposant une volont rdactionnelle dans la mise en forme de la
collection, le dsordre et la difficult des paroles sur le plan interprtatif auraient possiblement une fonction pdagogique. Il ny a aucune
raison de douter dune intention rdactionnelle derrire lagencement
des dits en vue de produire un sens thologique et un effet chez le
lecteur. Il est possible de comprendre lEvTh comme un regroupement
de paroles nigmatiques, poussant le lecteur une qute de sens et de
sagesse, qui le mne la connaissance9. Or, cette qute entrane le lecteur dans un chass-crois de diffrents thmes quil doit correctement
mettre en rapport les uns avec les autres. Par exemple, il est parfois
possible, partir dun logion, de dceler une chane de thmes reliant
les logia entre eux, o dans certains cas, le sens du texte se dploie
par enchanements et glissements successifs, comme en cascade dun
logion lautre. Chaque logion reprend une part du contenu prcdant
pour ensuite laborer dautres lments qui sont repris dans le logion
suivant et ainsi de suite. Nous avons limpression de toujours courir
pour saisir le sens10, dtre dans une inlassable qute de sagesse. Le
lecteur se trouve donc devant lobligation de dcoder un rseau com-

8. En ce sens, J.-M. SEVRIN, Linterprtation de lvangile selon Thomas,

entre tradition et rdaction , p. 353-358 ; Ce que lil na pas vu 1 Co
2,9 comme parole de Jsus , dans J.-M. AUWERS et A. WNIN (ds), Lectures et
relectures de la Bible. Festschrift P.-M. Bogaert (BETL 144), Leuven, Peeters,
1999, p. 314-315 ; on retrouve quelque peu cette approche dans le commentaire
de R. VALANTASIS, The Gospel of Thomas (New Testament Readings), New
York, Routledge, 1997, p. 24-27.
9. Puisque linterprtation des dits doit tre activement recherche, on peut
supposer linstar de J. S. KLOPPENBORG que le programme de lecture de Thomas est une sorte de hermeneutics of penetration ; voir The Formation of Q :
Trajectories in Ancient Wisdom Collections (SAC), Philadelphia, Fortress, 1987,
p. 305-306 ; voir galement J.-M. SEVRIN, Remarque sur le genre littraire de
lvangile selon Thomas (II,2) , dans L. PAINCHAUD et A. PASQUIER (ds), Les
Textes de Nag Hammadi et le problme de leur classification. Actes du colloque tenu Qubec du 15 au 19 septembre 1993 (BCNH Section tudes
3), Qubec/Louvain, Presses de lUniversit Laval/Peeters, 1995, p. 272-278,
ainsi que Thomas, Q et le Jsus de lhistoire , dans A. LINDEMANN (d.), The
Sayings Source Q and the Historical Jesus (BETL 158). Leuven, University
Press - Peeters, 2001, p. 463 et 469.
10. Concernant lEvTh, la qute interprtative laquelle se livre lexgte
ressemble ce quU. Eco qualifie de semiosis hermtique. La pense hermtique, que lon voit merger au IIe sicle, se manifeste sous les traits dune
sapience secrte o la vrit sidentifie au non-dit ou ce qui est dit de
manire obscure et doit tre compris au-del de lapparence et de la lettre. Les
dieux parlent (nous dirions aujourdhui : ltre parle) travers des messages
hiroglyphiques ou nigmatiques (U. ECO, Les Limites de linterprtation,
Paris, Grasset, 1992, p. 54).



plexe de significations11 et de participer en quelque sorte la construction du sens du texte12. Pour trouver linterprtation des paroles caches
de Jsus, le lecteur construit un rseau de significations par renvois
intra-textuels. Lintra-textualit est le fruit de la rflexion du lecteur,
cest une sorte dintentio lectoris13. Dans cette perspective, lintra-textualit se construit partir du rapport quun lecteur tablit entre les
diffrents logia de lEvTh. Ce rapport entre diffrents textes se produit lorsque la mmoire est alerte par un mot, une impression, un
thme () comme un souvenir circulaire 14. Comme nous venons de
le dire, le logion premier parle de la ncessit dinterprter les paroles
caches pour accder au salut. En un certain sens, lEvTh cache ce
quil dvoile. Thomas donne le contenu des paroles caches, sans toutefois en rvler le sens15. Le sens rside donc du ct du lecteur.
Lide de la fonction pdagogique de lcrit nest pas sans fondement ; elle est bien atteste dans lAntiquit. Par exemple, Charles
Meunier rsume la mthode dcriture de Justin en ces termes :
(Cette mthode) consiste conduire insensiblement le lecteur dun
thme lautre, en lui proposant des aspects toujours nouveaux dans le
dveloppement de largumentation. Le progrs se fait surtout par association dides ou par des variations partir de mots-cls ou de synonymes, de membres de phrases entires qui annoncent dune manire
voile le thme nouveau (). Cette mthode, qui rpond des intentions prcises dordre essentiellement pdagogique, a t fort pratique
dans lantiquit, notamment par les philosophes. Il sagit moins dexpo-

11. Jai ailleurs entrepris ce type dexgse dans mon article Connaissance,
identit et androgynit. Conditions du salut dans lvangile selon Thomas ,
dans M. ALLARD, D. COUTURE et J.-G. NADEAU (ds), Pratiques et constructions du corps en christianisme (Hritage et Projet 75), Montral, Fides, 2009,
p. 131-147.
12. Le lecteur participe la construction du sens de bien des manires.
Par exemple, on peut contribuer au sens dun texte en dveloppant un rapport
intertextuel constitu partir de lexprience de lecture de lindividu ; voir
A. GAGN, De lintentio operis lintentio lectoris. Essai hermneutique
partir de lpisode du dmoniaque de Grasa (Mc 5,1-20) , Thologiques 12/12, 2004, p. 213-232.
13. Pour A. Compagnon, lintertextualit ou dans ce cas lintratextualit
peut se trouver du ct du lecteur do son association avec lintentio lectoris ;
voir A. COMPAGNON, Le Dmon de la thorie. Littrature et sens commun (PEs
352), Paris, Seuil, 1998, p. 130-131.
14. Voir N. PIGAY-GROS, Introduction lintertextualit, Paris, Nathan,
2002, p. 19.
15. Cet aspect de lcriture de Thomas peut mieux se comprendre partir
dune analogie avec le secret hermtique dans le discours alchimique o la
force dun secret rside dans le fait dtre toujours annonc, mais jamais
nonc. Sil tait nonc, il perdrait de sa fascination (U. ECO, Les Limites de
linterprtation, p. 105).



ser un systme tout fait que dveiller lattention de lauditeur ou du

lecteur une doctrine de vie16.

On peut aussi relever la fonction pdagogique de lcriture comme

voile17 dans les Stromates de Clment dAlexandrie. Lauteur insiste sur
limportance du travail dinterprtation o seul le vrai gnostique arrive
comprendre le sens des secrets divins. Clment explique en quelques
lignes la manire dont il faut saisir le sens de son uvre :
Il va sans dire que ces Stromates, ptris drudition, singnient
cacher les semences de la connaissance. De mme que le chasseur passionn aime quter, fouiller, pister, lancer ses chiens avant de prendre
la bte, le vrai se rvle plein de douceur quand on la qut et obtenu
grand travail18.
Dieu a annonc par cette parole (Mt 13,12) que les secrets seront
rvls quiconque les coute en secret, et que les choses caches
seront dvoiles, comme la vrit, quiconque est capable de recevoir
les traditions sous un voile ; et que ce qui est secret pour la foule sera
manifest au petit nombre19.
Il y a aussi des choses que mon livre nindiquera que par allusion ;
il insistera sur les unes, il mentionnera seulement les autres, il tchera
de parler sans en avoir lair, de montrer sous le voile, de signifier sans
mot dire20.

Avec une telle perspective en vue, nous allons maintenant traiter de

quelques logia qui semblent premire vue contenir des propos que
lon pourrait peut-tre qualifier de modaliste. Le modalisme classique
est celui que Sabellius enseignait au IIIe sicle Rome, cest--dire,
une doctrine anti-trinitaire o le Pre, le Fils, et lEsprit-Saint sont les
manifestations dune seule ralit divine, une hypostase. La formule
sabellienne : Le Fils est la mme hypostase que le Pre , fut rejete
au concile dAntioche en 272 de notre re. Est-il possible que lEvTh
contienne des traces dune doctrine modaliste primitive ?

16. C. MUNIER, La mthode apologtique de Justin le martyr , Revue des

sciences religieuses 62, 1998, p. 228.
17. Pour plus de dtails sur la mtaphore du voile comme dsignation de
lcriture, voir A. Le BOULLUEC, Voile et ornement : le texte et laddition de
sens selon Clment dAlexandrie , dans P. SALAT (d.), Questions de sens.
Homre, Eschyle, Sophocle, Aristote, Virgile, Apule, Clment (tudes de littrature ancienne. 2), Paris, Presses de lcole Normale Suprieure, 1982, p. 53-64.
18. Stromates I, 2.20-21. Les citations sont tires de ldition de C. MONDSERT, Clment dAlexandrie. Les Stromates : Stromate I (Sources chrtiennes
30), Paris, Cerf, 1951.
19. Stromates I, 1.13.3.
20. Stromates I, 1.15.1.



Les lus, fils du Pre vivant (EvTh 49-50)

Les logia 49-50 de lEvTh constituent le point de dpart de notre
analyse de paroles qui, premire vue, semblent contenir des traces
de modalisme. Ces deux dits parlent des solitaires et des lus
comme des trangers venus du Royaume et de la lumire, qui un jour
retourneront leur lieu dorigine.
(49) 1Jsus a dit : Heureux les solitaires, et lus (), car vous
trouverez () le Royaume, 2car vous () tes issus de
lui, (et) vous y retournerez.
(50) 1Jsus a dit : Sils vous () disent : Do venez-vous ? ,
dites-leur : Nous sommes issus de la lumire (), l o la
lumire est advenue delle-mme et sest tenue debout ( ),
et elle est apparue () dans leur image ( ) .2
Sils vous disent : Est-ce vous ? , dites-leur : Nous sommes ses fils
et nous sommes les lus () du Pre vivant ( ) .
Sils vous demandent : Quel est le signe de votre Pre qui est en
vous ? , dites-leur : Cest un mouvement et un repos.

Ces deux logia font partie dune maxi-structure contenant un paralllisme concentrique [A, B, C // C, B-1, A]21 allant des logia 49 5422.
On peut reprsenter la structure comme suit :
A Batitude sur le Royaume (EvTh 49)
B Questions sur lidentit des lus et le signe de votre Pre
(EvTh 50)
C Le repos des morts et lignorance du monde nouveau
(EvTh 51)
C La parole des morts et loubli du vivant (EvTh 52)
B-1 Question sur lutilit de la circoncision ; la marque de
leur Pre (EvTh 53)
A Batitude sur le Royaume (EvTh 54)
Les logia 49 et 54 indiquent linclusion [A // A], qui est clairement
dlimite par deux batitudes traitant de la trouvaille et de laccs du
Royaume pour les solitaires / lus et les pauvres. On peut supposer une
quivalence entre les solitaires (lus) et les pauvres avec lemploi de la
2e personne du pluriel ( ; ). Le paralllisme antithtique
21. Pour plus de dtails sur les conventions de lanalyse structurelle utilises dans cette analyse, voir M. GIRARD, Les Psaumes Redcouverts. De la
structure au sens. 1-50, Montral, Bellarmin, 1996, p. 65-66.
22. Jai aussi entrepris une tude de la structure des logia 49-54 sous langle
de la stratgie pdagogique de lerotapokriseis dans un article intitul : Structure and Meaning in Gos. Thom. 49-53. An Erotapokritic Teaching on Identity
and Eschatology , dans J. SCHRTER (d.), The Apocryphal Gospels within the
Context of Early Christian Theology (Bibliotheca Ephemeridum Theologicarum
Lovaniensium 260), Leuven, Peeters, 2013, p. 23-31.



[B // B-1] contient des questions sur lidentit. Le signe de votre

Pre qui est en vous (EvTh 50) et la circoncision sont en opposition
(EvTh 53) ; ils reprsentent en quelque sorte la diffrence entre les
vrais et les faux lus. Le paralllisme synonymique [C // C] est caractris par lemploi rptitif de (ceux qui sont morts ;
cest--dire les morts). Le repos des morts viendra au jour du monde
nouveau (EvTh 51). Les disciples ignorent, cependant, que ce jour est
dj prsent. Au logion 52, cest la parole des morts qui empche les
disciples de percevoir le vivant parmi eux.
En revenant maintenant aux logia 49-50, il importe de signaler que
le substantif grco-copte , traduit ici par solitaire , a fait
lobjet de plusieurs analyses philologiques et de commentaires que nous
nous contentons de signaler en note23. Le logion 49 compte parmi les
23. E. BECK, Ein Beitrag zur Terminologie des ltesten syrischen
Mnchtums , Studia Anselmina 38, 1956, p. 254-267 ; D. F. BUMAZHNOV,
Some Ecclesiological Patterns of the Early Christian Period and Their Implications for the History of the Term (monk) , dans A. ANATOLY, C. KARAKOLIS et U. LUZ (ds), Einheit der Kirche im Neuen Testament : dritte europische
orthodox-westliche Exegetenkonferenz in Sankt Petersburg, 24-31 August,
2005, Tbingen, Mohr Siebeck, 2008, p. 251-264 ; R. CHARRON, propos
des et de la solitude divine dans les textes de Nag Hammadi ,
dans L. PAINCHAUD et P.-H. POIRIER (ds), Coptica Gnostica Manichaica.
Mlanges offerts Wolf-Peter Funk (Bibliothque Copte de Nag Hammadi
Section tudes 7), Qubec/Louvain/Paris, Presses de lUniversit Laval/Peeters, 2006, p. 109-133 ; M. CHOAT, The Development and Usage of Terms for
Monk in Late Antique Egypt , Jahrbuch fr Antike und Christentum 45, 2002,
p. 5-23 ; A. D. DECONICK, The Original Gospel of Thomas in Translation, New
York, T&T Clark, 2007, p. 98-99 ; E. DEKKERS, . Solitaire, unanime,
recueilli , dans A. A. BASTIAENSEN, A. HILHORST et C. H. KNEEPKENS (ds), Fructus centesimus : mlanges offerts Grard J. M. Bartelink, Dordrecht, Kluwer,
1989, p. 91-104 ; S. H. GRIFFITH, Asceticism in the Church of Syria : The
Hermenuetics of Early Monasticism , dans V. L. WIMBUSH et R. VALANTASIS
(ds), Asceticism, Oxford, Oxford University Press, 1995, p. 220-245 ; M. HARL,
propos des de Jsus : Le sens du mot , Revue des tudes
grecques 73, 1960, p. 464-474 ; E. JUDGE, The Earliest Use of Monachos for
Monk (P. Coll. Youtie 77) and the Origins of Monasticism , Jahrbuch fr
Antike und Christentum 20, 1977, p. 72-89 ; A. F. J. KLIJN, The Single One
in the Gospel of Thomas , Journal of Biblical Literature 81, 1962, p. 271-278 ;
F.-E. MORARD, Monachos, Moine. Histoire du terme grec jusquau 4e sicle.
Influences bibliques et gnostiques , Freiberger Zeitschrift fr Philosophie
und Theologie 20, 1973, p. 332-411 ; Monachos : une importation smitique
dgypte ? Quelques aperus nouveaux , dans E. A. LIVINGSTONE (d), Papers
Presented at the 6th International Conference on Patristic Studies held in Oxford
1971 (Texte und Untersuchungen 115), Berlin, Akademie-Verlag, 1975, p. 242246 et Encore quelques rflexions sur Monachos , Vigiliae Christianae 34,
1980, p. 395-401 ; R. MURRAY, Symbols of the Church and Kingdom. A Study in
Early Syriac Tradition, Piscataway, NJ, Gorgias Press, 2004, p. 13-16 ; G. QUISPEL, Lvangile selon Thomas et les origines de lascse chrtienne , dans
Aspects du judo-christianisme. Colloque de Strasbourg, 23-25 avril 1964 (Tra-



dix batitudes de lEvTh24. La mention des aux cts des

(lus) suscite un questionnement. Sagit-il de deux classes
dindividus ? Dans son commentaire sur lEvTh, Richard Valantasis est
davis que les deux termes dsignent deux groupes diffrents, et que la
batitude concerne ceux qui sont solitaires et lus. Il serait donc possible, selon lui, dtre solitaire sans tre lu et vice-versa. Seuls ceux
qui possdent le double statut accderont au Royaume25. Mais dans
un rcent commentaire sur lvangile thomasien, Uwe-Karsten Plisch
montre quil nen est pas ainsi. Il a reconnu la fonction pexgtique de
la conjonction linstar du grec26. Et avec le pronom relatif
on devrait lire la premire partie du logion comme suit : Heureux
les solitaires, cest--dire, les lus.
Il importe de noter le passage de la 3e personne la 2e personne du
pluriel () dans la deuxime partie du logion27. Lemploi de la
2e personne du pluriel et la rptition du substantif permettent
dtablir un lien clair avec le logion 50. Ce texte faonne lidentit
des lecteurs / auditeurs, les forant en quelque sorte sidentifier ou
se percevoir comme les solitaires / lus28. La fin du logion 49 parle
du Royaume () comme lorigine et la destination finale des
solitaires / lus.
En prenant les deux logia (49-50), on remarque une certaine correspondance entre le Royaume () et la lumire ()
comme le lieu dorigine des solitaires / lus. Au logion 50, lidentit
de ceux qui questionnent les lus nest pas rvle. La plupart des
chercheurs comparent le questionnement des opposants en EvTh 50
vaux du centre dtudes suprieures spcialis dhistoire des religions de Strasbourg), Paris, Presses universitaires de France, 1965, p. 37-45 ; Makarius, das
Thomasevangelium, und das Lied von der Perle (NovTSup 15), Leiden, Brill,
1967, p. 108 ; The Gospel of Thomas Revisited , dans B. BARC (d.), Colloque International sur les textes de Nag Hammadi. Qubec, 22-25 aot 1978
(Bibliothque Copte de Nag Hammadi Section tudes 1), Qubec/Louvain/
Paris, Les Presses de lUniversit Laval/Peeters, 1981, p. 218-266 ; R. URO, Is
Thomas an encratite gospel ? , dans Thomas at the Crossroads. Essays on the
Gospel of Thomas (Studies of the New Testament and Its World), Edinburgh,
T&T Clark, p. 140-162 ; A. VBUS, History of Asceticism in the Syrian Orient,
vol. 1 : The Origins of Asceticism, Early Monasticism in Persia (Corpus scriptorum christianorum orientalium 184), Louvain, Peeters, 1959, p. 106-108.
24. On fait rfrence aux en EvTh 7 ; 18 ; 19 ; 49 ; 54 ; 58 ; 68 ;
69 (2x) ; 103.
25. R. VALANTASIS, The Gospel of Thomas, p. 126.
26. U.-K. PLISCH, The Gospel of Thomas. Original Text with Commentary,
traduit de lallemand par G. S. ROBINSON, Stuttgart, Deutsche Bibelgesellschaft, 2008, p. 128. Concernant la fonction pexgtique de , voir BLASS et
27. PLISCH a aussi remarqu ce phnomne en EvTh 54, voir The Gospel of
Thomas, p. 127.
28. R. VALANTASIS, The Gospel of Thomas, p. 126.



linterrogation des archontes lendroit des gnostiques lors du retour

leur demeure cleste (voir ApocJac V 33,11-34,20). La prsence des
archontes dans ce logion de lEvTh nest cependant pas explicite. Les
interrogateurs reprsentent certes des opposants quelconques. Mais ce
qui est dun intrt certain est la prcision concernant lorigine des solitaires / lus : la lumire. On apprend que la lumire est advenue dellemme, sest tenue debout et quelle est apparue dans leur image. La
formule copte // (se tenir debout) est aussi employe pour
qualifier la venue de Jsus dans le monde au logion 28, ainsi que le
verbe (apparatre)29 :
Jsus a dit : Je me suis tenu debout ( ) au milieu du
monde, et dans la chair je leur suis apparu ().

Au logion 50, la lumire sest tenue debout et elle est apparue dans
leur image. Au logion 28, Jsus sest tenu debout au milieu du monde
et il est apparu dans la chair. Les correspondances sont frappantes. On
remarquera aussi la prsence du pronom personnel singulier rattach au substantif (leur image). Sil tait question de limage
des rpondants, cest--dire de celle des solitaires / lus, on aurait
(nos images ; pluriel). On pourrait peut-tre comprendre
leur image (au singulier) comme une rfrence la chair du logion 28.
Jsus est apparu au milieu du monde dans la chair, une manifestation
qui fut la ressemblance de lhumanit.
Un autre lment significatif est lide que les solitaires sont fils de
la lumire ( ; masculin pluriel saccordant avec ).
la manire de la fonction pexgtique de au logion 49, il en
est de mme avec , qui pourrait se traduire : cest--dire, nous sommes les lus du Pre vivant. Il y
aurait donc un parallle entre la lumire et le Pre vivant ; les deux sont
lorigine des solitaires / lus.
Mais que reprsente la lumire et qui est le Pre des lus ? En gardant en tte lide que Thomas sinterprte par renvois intratextuels,
nous pourrions y trouver une rponse dans la premire partie du
logion 77 o il est dit :
Jsus a dit : Je suis la lumire () qui est sur tous, je suis le
tout ; le tout est sorti de moi, le tout est arriv jusqu moi.

Ce logion identifie Jsus la lumire et le tout provient de lui. Cela

sapparente, certes, lide que les solitaires / lus sont issus de la
lumire. Il y a aussi la mention dun retour la lumire. On pourrait
se demander si cela correspond au retour des solitaires / lus dans le
Royaume dont ils sont issus (logion 49). Si tel est le cas, on pourrait y
voir une quivalence entre la figure de Jsus, la lumire et le Royaume.

29. R. VALANTASIS, The Gospel of Thomas, p. 128.



Les trois seraient en quelque sorte lorigine de lexistence des solitaires / lus.
Le logion 50 tablissait aussi un lien entre la lumire et le Pre
vivant. Qui donc est ce Pre vivant ? Il y a un dit intressant qui
pourrait peut-tre jeter un clairage sur cette question :
(15) Jsus a dit : Quand vous verrez celui qui na pas t engendr de
la femme (), prosternez-vous sur votre face et adorez-le : celuil est votre pre ().

Il est trs probable que celui qui na pas t engendr de la femme

soit une rfrence Jsus lui-mme. Plusieurs raisons soutiennent une
telle interprtation. Dabord, il y a toujours une indiffrence ou une
valuation ngative de Jsus lorsquil est question de sa mre terrestre.
La batitude sur le ventre et les seins qui lont nourri (logion 74) nest
pas ncessairement applaudie de la part de Jsus (Luc 11,27-28). De
plus, le logion 99 contient la parole sur les frres et la mre de Jsus
qui se tiennent lextrieur. La rponse donne est la mme que celle
des synoptiques : ceux qui font la volont de mon Pre sont mes frres
et ma mre. Enfin, au logion 101, Jsus demande ses disciples de har
pre et mre, comme lui-mme en fait laction. De plus, dans ce mme
logion, il y a une opposition entre sa mre et sa mre vritable qui
dans une perspective intertextuelle pourrait tre une rfrence lEsprit
(dans lvangile des Hbreux). On peut aussi ajouter cela le fait que
Jsus est caractris comme tant la lumire (EvTh 50) ; celle-ci, souvenons-nous, est advenue delle-mme. En conclusion, on peut dire que
le seul que les disciples peuvent visiblement voir est Jsus lui-mme !
Celui qui est appel votre pre au logion 15 pourrait donc tre le
mme que les solitaires / lus appellent le Pre vivant au logion 50.
Il est dailleurs intressant de constater que ladjectif (vivant) est
donn Jsus ainsi quau Pre. On retrouve plusieurs mentions de cet
adjectif ailleurs dans lEvTh :
(incipit) Voici les paroles caches que Jsus le vivant ( ) a
dites et qua crites Didyme Jude Thomas.
(52) 1 Ses disciples lui dirent : Vingt-quatre prophtes ont parl en
Isral et tous ont parl de toi.2 Il leur dit : Vous avez dlaiss celui qui
est vivant () devant vous et vous avez parl des morts.
(59) Jsus a dit : Fixez le vivant () pendant que vous tes
vivants de peur que vous ne mouriez et ne cherchiez le voir, et vous
ne pourrez pas le voir30.

30. Voir galement EvTh 38 : 1Jsus a dit : Bien des fois vous avez
dsir entendre ces paroles que je vous dis et vous navez nul autre de qui les
entendre ;2 viendront des jours o vous me chercherez et ne me trouverez pas.



Pour ce qui est des rfrences au Pre, on peut aussi ajouter les
logia 3, 37 et 5031. Mais rsumons la manire dont Jsus est dcrit dans
lchantillon de textes que nous avons examin jusqu prsent :
1. La lumire () : EvTh 50 (// 28) ; 77,1
- les lus sont issus ( ) de la lumire : EvTh 50,1
- le tout est issu ( ) de Jsus, la lumire : EvTh 77,1
- la lumire sest tenue debout ( ) : EvTh 50,1
- Jsus sest tenu debout ( ) au milieu du monde :
EvTh 28,1
- la lumire est apparue () dans leur image : EvTh 50,1
- Jsus est apparu () eux dans la chair : EvTh 28,1
2. Le Pre () : EvTh 15 ; 101
- Celui qui na pas t engendr de la femme : EvTh 15
- Ma mre vritable : EvTh 101
3. Le Vivant () : EvTh incipit ; 52 ; 59 (// 38)
- Ce titre est souvent associ au Pre (EvTh 3,4 [// 50,2] ; 37)

Conclusion : Modalisme primitif ou principe de gmellit ?

Ces noncs, qui semblent tablir une correspondance entre Jsus et
le Pre, sont-ils pour autant des traces de modalisme comme le supposait Mnard ? Il y a pourtant dautres dits que lon pourrait qualifier
de trinitaires, au sens o ils prsentent des distinctions marques entre
le Pre, le Fils et lEsprit. Par exemple, les logia 44 et 61 parlent de
ces trois figures de manire distincte :
(44) 1Jsus a dit : Celui qui blasphmera contre le Pre, il lui sera
pardonn ; 2celui qui blasphmera contre le Fils, il lui sera pardonn ;
mais celui qui blasphmera contre lEsprit Saint, il ne lui sera pardonn
ni sur terre ni au ciel.
(61)3Jsus lui dit : Cest moi qui suis issu de celui qui est gal ; il
ma t donn des choses de mon Pre (). (Cf. 64 ; 99 : mon
Pre )

Clairement, Jsus nest pas le Pre. Il ny a donc pas de modalisme

dans ces noncs. Comment donc comprendre deux perspectives qui
semblent premire vue contradictoires ? Doit-on simplement renoncer
chercher une certaine cohrence dans les dits thomasiens ? Thomas
ne serait quune collection de paroles disparates sans signification ; ou
peut-il y avoir quelque chose dautre derrire la caractrisation de Jsus
comme lumire et Pre vivant ? Y aurait-il un autre principe luvre
dans lEvTh pouvant expliquer une telle caractrisation de Jsus ?
Depuis longtemps, les chercheurs ont reconnu la notion de gmellit dans lEvTh, particulirement en lien avec la formule
31. Les Actes de Jean contiennent aussi des rfrences qui qualifient Jsus
comme vivant ou Pre ; AcJn 70 ; 97 ; 169.



(faire des deux un). Lunit est aussi troitement lie au retour
lorigine et au salut. Un bon nombre de textes parlent de ce principe
en ces termes :

(108) 1Jsus a dit : Celui qui boira () ma bouche deviendra comme moi ; 2moi aussi je deviendrai lui 3et les choses caches
() se dvoileront lui.
(incipit + 1) Voici les paroles caches ( ) que Jsus le
vivant a dites et qua crites Didyme Jude Thomas. Il a dit : Celui qui
trouvera linterprtation de ces paroles () ne gotera pas la mort.
(106) 1 Jsus a dit : Si de deux vous faites un ( ), vous
deviendrez Fils de lHomme ( )

Au logion 108, on promet la rvlation () des choses

caches celui qui boit () la bouche () de Jsus. De toute
vidence, il y a une correspondance entre ce qui est cach () et les
paroles caches ( ) dans les premires lignes de lEvTh.
Le disciple est transform au moyen des paroles de Jsus. Lorsquil
reoit les paroles (= boit la bouche de Jsus), cest alors que sa propre
identit se perd dans celle de son matre. linstar de Didyme Jude
Thomas, qui est reconnu pour tre le vritable hermneute de Jsus, le
disciple, lui aussi, devient en quelque sorte le jumeau du matre32.
Le logion 106 parle de la manire dont ceux qui sadonnent au principe dunit ( ) deviendront Fils de lHomme (
). Cela correspond sans contredit aux donnes du logion 108
o les fils de lhomme deviendront jumeaux de Jsus, le Fils de
La notion de gmellit est peut-tre ce qui se cache derrire la
caractrisation de Jsus comme Pre vivant. Jsus est en quelque sorte
le jumeau de son Pre, cest pourquoi il a les mmes caractristiques.
linstar de leur matre, les disciples en viendront eux aussi manifester les caractristiques de Jsus, au point mme de porter le titre quil
lui est normalement attribu : fils de lhomme .

32. On peut aussi signaler le rapport vident entre les logia 13 et 108 o il
est question dans les deux cas de boire () la bouche () de Jsus
(EvTh 108) ou la source bouillonnante (EvTh 13). On remarque galement
que les rsultats sont les mmes dans les deux logia : (1) devenir comme Jsus
[EvTh 108] = je ne suis pas ton matre [EvTh 13], et (2) lintelligence des
choses caches [EvTh 108] = il lui dit trois paroles [EvTh 13]. En somme,
lidentit du disciple est modifie lorsquil assimile les paroles de Jsus. Le disciple uni son matre au moyen des paroles caches pourrait donc correspondre
la condition exige pour entrer dans le Royaume (cf. logia 22 et 106) : faire
des deux un ( ) ; voir S. L. DAVIES, The Gospel of Thomas
and Christian Wisdom Second Edition, p. 91-94 et A. MARJANEN, The Woman
that Jesus Loved : Mary Magdalene in the Nag Hammadi Library and Related
Documents (NHMS 40),Leiden, Brill, 1996, p. 42-43.
33. Jsus est galement dsign comme le fils de lhomme au logion 86.

Collegio Internazionale San Lorenzo da Brindisi, Rome


THE Maf rft lMaryam
The article deals with the Gz version of the Liber Requiei (The
Book ofMarys Rest), the Maf rft lMaryam. There is
little doubt that the Gz translation was performed on a lost Greek
text, an aspect which together with the paucity of witnesses of the
Gz makes this work a privileged research item. The Gz version of
the Liber Requiei is an indispensable avenue for the reconstruction of
the lost Greek original, a term of comparison for the Syriac versions
that descended from it and an important example of archaic translation
practice from Greek into Gz. The linguistic quality of the Gz text
is so harsh that the work is virtually unknown to the vast majority of
Ethiopian church men and scholars. The pioneering edition and Latin
version of Victor Arras (upon which two Italian translations were produced) has the merit of unearthing a text which otherwise would have
probably fallen into oblivion. A re-edition of the manuscripts with a
fresh faithful translation of the Maf rft lMaryam would be
very useful.
Larticle traite de la version Gz du Requiei Liber (Le Livre du
Repos de Marie), le Maf rft lMaryam. Nul doute que la
traduction Gz a t ralise sur un texte grec perdu, un aspect
qui, avec la raret des tmoins Gz, fait de ce travail un thme de
recherche privilgi. La version Gz du Liber Requiei est indispensable pour la reconstruction de loriginal grec perdu et pour une comparaison avec les versions syriaques qui en drivent. En outre, elle
illustre remarquablement la pratique archaque de la traduction du
grec en Gz. La qualit linguistique du texte Gz est si impntrable que le travail est pratiquement inconnu par la grande majorit
des hommes dglise et des chercheurs thiopiens. Ldition (et la version latine de Victor Arras), sur laquelle deux traductions italiennes ont
t produites, a le mrite pionnier de dterrer un texte qui, autrement,
serait probablement tomb dans loubli. Une rdition des manuscrits
avec une nouvelle traduction fidle au Maf rft lMaryam
serait trs utile.

Apocrypha 23, 2012, p. 223-245



According to several scholars who have dealt with the Liber Requiei,
the Ethiopic LR derives from a lost Greek original.1 Such a conclusion
is the outcome of a painstaking and extensive search for the sources of
the Dormition literature. Victor Arras, the editor of the LR highlights
the significance of the Gz (Ethiopic) text in enthusiastic terms. In
the foreword to his Latin translation, he declares that the LR : Est fons
e quo, aqua ut par est percolata, omnes illi auctores hausierunt ; omnes
sequaces excerpserunt, vario modo breviantes, vario modo expurgantes,
aut a textu integro aut a textu iam partim breviato.2 A considerable
number of internal elements points to Greek as the language from which
the LR has been translated into Gz,3 plausibly during the so-called
Aksumite period (IV-VII/VIIIth cent.)4. It would be fair to ask : Why
from Greek and not, for instance, from Syriac ?. After the unsuccessful
Abbreviations : AM = Ethiopian Calendar ; ARRAS 1973 = V. ARRAS, De Transitu Apocrypha Aethiopice, CSCO 342 Aeth. 66, Louvain, 1973 ; ARRAS 1974 =
V. ARRAS, De Transitu Apocrypha Aethiopice, CSCO 351 Aeth. 68, Louvain,
1974 ; DillLex = A. DILLMANN, Lexicon Linguae Aethiopicae, Lipsia, 1865 ;
EA = Encyclopaedia Aethiopica, Wiesbaden, 2003-2007 ; ERBETTA, Apocrifi =
M. ERBETTA (ed.), Gli apocrifi del Nuovo Testamento, I/2. Infanzia e Passione
di Cristo Assunzione di Maria. Versione e Commento, Torino, 1981 ; KWKDict
= KIDAN-WLD KFLE, Maf Swasw Wgs Wmzgb Qalat addis [A
Book of Grammar and Verb, and a new Dictionary, Addis Abba 1948] AM ;
F. MANNS, Le rcit = F. MANNS, Le rcit de la Dormition de Marie : (Vatican
grec 1982). Contribution ltude des origines de lexgse chrtienne, Jerusalem, Franciscan Printing Press, 1989 ; MS(S) = manuscript(s) ; OCP = Orientalia Christiana Periodica ; SHOEMAKER, Traditions = S. J. SHOEMAKER, Ancient
Traditions of the Virgin Marys Dormition and Assumption, Oxford, Oxford
Early Christian Studies, 2002 ; T1 = M. JUGIE, Saint Jean, Archevque de Thessalonique (+ vers 630), Homlies mariales byzantines : textes grecs dits et
traduits en latin, PO 19-3, Paris, 1926, coll. 257-287.
1. E. TESTA affirms : Pare che questa versione etiopica sia stata fatta, nel
V/VI sec., su un testo greco precedente che riportava una lezione molto vicina
alloriginale : difatti ci sono errori che si spiegano soltanto con una falsa lettura di termini greci, cf. Lo sviluppo della Dormitio Mariae nella letteratura,
nella teologia e nella archeologia, Marianum Ephemerides Mariologiae, 44/34, n. 129, 1982, p. 317. Testa does not provide any concrete example in support
of his claim.
2. ARRAS 1973, p. VII.
3. Gz also known as classical Ethiopic has been classified as a Semitic
language. Entering into the details of the origin, evolution of this language as
well as with the past and current debate on these issues is beyond the (limited)
scope of this paper. For a summary on the origin and evolution of the Gz
language, cf. S. Weninger, Gz, EA, p. 732-735.
4. Cf. For a list of Greek and Syriac Church Fathers who have influenced
Ethiopic Mariology, cf. C. HAYOZ, Portrait de Marie. Complainte de la Vierge,
Deux posies mariales thiopiennes indites. Texte, traduction, commentaire,
Universit de Fribourg (Suisse), 1956, p. 14-22.



attempts, especially by Arthur Vbus5 to get to Syriac ancestry as the

matrix of Ethiopic versions of biblical and extra-biblical literature, this
route has been abandoned.6 There is no compelling evidence that allows
to conjecture translations from Syriac or Coptic into Gz during the
Aksumite era. There is unanimity among scholars that Greek was at
that time the resource language for writing, as witnessed by epigraphy,
numismatics and a good number of books.7 The LR contains passages
which betray an erroneous understanding of Greek terms.8 There are

5. Cf. A. VBUS, Die Spuren eines lteren thiopischen Evangelientextes im

Lichte der literarischen Monumente, Stockholm, Papers of the Estonian Theological Society in Exile 2, 1951 ; ID., Early Versions of the New Testament, Stockholm 1954 ; ID., Temera Iyasus. Zeuge eines lteren thiopischen Evangelientypus, OCP 17, 1951, p. 462-467. On the same topic, cf. W. WITAKOWSKY,
Syrian Influences in Ethiopian Culture, Orientalia Suecana 38-39, 1989-1990,
p. 191-202.
6. Cf. R. ZUURMOND, Novum Testamentum Aethiopice : The Synoptic Gospels :
General Introduction. Edition of the Gospel of Mark, Stuttgart, Aethiopistische
Forschungen 27, 1989, p. 114-118, see, especially the Excursus on the theses
of A. Vbus, in p. 119-123 ; M.A. KNIBB, Translating the Bible. The Ethiopic
Version of the Old Testament, Oxford Universty Press, The British Academy,
The Schweich Lectures of the British Academy 1995, 1999, p. 1-54 ; P. MARRASSINI, Ancora sul problema degli influssi siriaci in et aksumita, Biblica et
Semitica. Studia in memoria di Francesco Vattioni, a cura di L. Cagni, Napoli,
1999, p. 325-337 ; P. SINISCALCO (a cura di), La chiesa dEtiopia, Le antiche
Chiese Orientali. Storia e letteratura, Roma, 2005, p. 123-127.
7. Cf. GETATCHEW HAILE, Gz literature, EA 2, p. 736-737 ; I. GUIDI, Storia della letteratura etiopica, Roma, 1932, p. 11. In the following page, Guidi
adds : La letteratura geez, come la copta, larmena, la georgiana e, quasi totalmente, la siriaca, comincia con lintroduzione del cristianesimo.
8. Manns supplies a number of instances that reflect a misunderstanding
of a Greek text. His comparison between the Gz and the Greek is based
mainly on the text of Vatican grec. 1982 (= R). The following are some of the
examples : LR 54 words for kids, instead of words for instruction R 30. The
Greek was read as ; LR 37 takes path for suffering (R 12), and the verb to give birth was taken for to fall,
to descend. was read as in LR 49, therefore : You have
planted for us your Messiah, instead of : You have planted in us your goodness : R 24. LR 48 confuses vocation with assembly in
T1. LR 52 understands the in R 29. LR 60 (mid. fut. of
) rejoice for (mid. fut. of ) use T1. LR 75 : So that
you may build up this people, R 41 So that you may build up this people.
The confusion of the Greek text between and is preserved in the
Gz text as well. LR 76 has hidden for crucified
R 43, cf. MANNS, Le rcit, p. 80-81. For an analogous treatment of the issue,
see also, ERBETTA, Apocrifi, p. 422-424 (footnote 2) ; M. VAN ESBROECK, Bild
und Begriff in der Transitus-Literatur, der Palmbaum und der Tempel, in MARGOT SCHMIDT (ed.), Typus, Symbol, Allegorie bei den stlichen Vtern und ihren
Parallelen in Mittelalter, Regensburg, 1982, p. 333-351, especially, p. 340-341.



loan-words from Greek like X, derived from , first fruit9.

Moreover, the claim for a Greek Vorlage is supported by a comparison
of the LR with the earliest witnesses that belong to the family of the
Palm10, such as the Irish and Georgian traditions11.
The scanty number of witnesses so far available, the thorny, often in
tractable vocabulary and style of the LR and more significantly its
exclusion from a direct liturgical use can be taken as plausible clues
that this is one of the texts in which it is virtually impossible to see
revisions mediated by Arabic,12 as has happened with many other
books, starting from the Bible. For instance, the Narratio Brevior of the
Dormitio ascribed to John the Evangelist, uses the term
msyiyin13, to designate the Christians, an Arabic loan-word which
tells clearly that the source of the text is Arabic. In Gz, but also in
several modern languages spoken in Eritrea and in Ethiopia, Christians
are known as krstyan and not as msiin. Commenting on the adjective msiawi, Dillmann observes: adj. rel. christianus (potissimum in
libris ex Arabico sermone translatis usitatum).14 Within the frame of
9. LR n. 98. The entry, which has not been commented upon by the editors and translators of the LR (e.g. SHOEMAKER, Traditions, p. 345), appears in
DillLex col. 806. Dillmann does not register X but X (like Arrass
variant, i.e. MS B) and X with the qualification nomen peregrinum scilicet , , id est primitiae oblatae.
10. For a (complex) genealogical tree of this family, cf. MANNS, Le rcit,
p. 35.
11. SHOEMAKER, Traditions, p. 39, footnote 78. MANNS, Le rcit (cf. especially p. 81-82), has indicated linguistic elements of Semitic origin in the LR.
Some of the examples he provides seem to be cogent. Shoemaker rehearses,
apparently without being fully persuaded, Manns hypothesis which seems to
favour a Semitic linguistic background. Shoemaker, who does not exclude categorically the Semitic track, including Arabic, seems to be rather undecided as
far as the original language at the root of the LR. For his various argumentations which put together Greek, Syriac and Arabic as the source of the Gz,
cf. Traditions, p. 38-39.
12. Manns refers around twenty examples that presume a Semitic original
shared by RL and R, cf. Le rcit, p. 81-82. The cases Manns puts forward,
which by his own admission are hypothetical, do not look particularly cogent.
Moreover, the so-called Semitisms, do not by themselves represent a decisive
proof to presume a Vorlage composed in a Semitic language. The books of the
New Testament are an important example of authors of Semitic background that
can write in Greek, reflecting in the process their Semitic forma mentis, syntax
and sometimes even vocabulary.
13. Cf. ARRAS 1973, p. 97.
14. Cf. DillLex coll. 177-178, where the root msa and the related semantic field is explained. Expressions such as Haymanot Msiawit, do exist, cf.
KWKDict p. 612 ; GETATCHEW HAILE, The Faith of the Unctionists in the Ethiopian Church (Haymanot Msiawit), CSCO 517 Aeth 91 (text) CSCO 518 Aeth
92 (trad.), Louvain, 1990, but did never establish themselves, nor enjoy wide
circulation in Ethiopia.



this general conclusion, the passage in LR 49, which reads: }o :

o : soF : D{ : K[G tu es qui plantasti nobis Messiam
tuum should be remembered. The question as to why the translator has
opted for K[I, and not ek is inevitable.
Sentences starting with the preposition ]K, frequent constructions
with the preposition ED and with the verbal structure { : coD,15
introduced massively in post-aksumite Gz literature and especially
in works translated from Arabic16 from the end of the XIIIth century
onwards,17 do not feature in the LR. The LR is instead a meaningful
example of primitive classical Ethiopic, out of which, together with
works of the same kind it would be possible to draw a grammar of
archaic Gz. The description of the linguistic features of the canonico-liturgical body named by Bausi Collezione aksumita can be, by
and large, applied to the LR too. The following is what he has to say :
letiopico della Collezione aksumita estremamente arduo, la sintassi
calcata, il greco reso parola per parola, come possibile verificare in
presenza del testo originale, senza il cui ausilio il testo risulta spesso
pressoch incomprensibile ; i nomi propri denunciano esiti fonetici del
tutto incompatibili con passaggi attraverso larabo.18

15. Cf. S. WENINGER, Kona qatala zum Ausdruck der Vorvergangenheit

im Gz ?, N. NEBES (ed.), Tempus und Aspekt in den semitischen Sprachen.
Jenaer Kolloquium zur semitischen Sprachwissenschaft, Wiesbaden, Jenaer Beitrge zum Vorderen Orient 1,1999, p. 171-183.
16.One of the most important translations from Arabic which displays the
aforementioned new linguistic phenomena is the Haymanot Abw, a florilegium of patristic works, rendered from the Arabic text ( Itirf
al-ab), The Confession of the Fathers, cf. G. GRAF, Geschichte der cristlichen arabischen Literatur, Studi e Testi 133, 1947, p. 321-323. The text has
been partially published in a bilingual (Gz-Amharic) edition, Haymanot
Abw (Faith of the Fathers), Addis Abba, 1967 (AM); reprinted in 1982 (AM)
with an additional sentence in Amharic to the original title, Haymanot Abw.
Ynt Abbatoan Haymanot (Faith of the Fathers. Faith of our ancient
Fathers), Tsfa Gbr-llase (ed.), Addis Ababa 1982 (AM).
17. With regard to the timing and circumstances of the process of revisions,
Bausi has an interesting proposal. He says : Diversi elementi suggeriscono
che il processo di revisione non ebbe inizio prima del XIII sec. Un ruolo non
secondario vi avr probabilmente svolto il potente monastero di Dabra ayq
Esifnos, nel Wallo, che avr mediato tra preservazione della antica eredit
ricevuta dai monasteri settentrionali, ed elaborazione di una nuova cultura, nel
contesto del rinnovato legame con il patriarcato di Alessandria e di una forte
spinta espansiva verso meridione, cf. A. BAUSI, La Collezione aksumita canonica-liturgica, Adamantius 12, 2006, p. 53.
18. Ibidem, p. 47.



Manuscript attestation of LR
The oldest known manuscripts of LR are two, and would belong
to the XVth century.They are : BrOr. 692, f. 47r-83v19 and fragments in
Vat. Aeth. 26820. The patchy21 London manuscript that Victor Arras has
reproduced22 as his base text (A) exhibits a spelling, morphemes,
morphology and syntax that reflect a rudimentary Gz. For his edition, Arras has used the XVIIIth century manuscript, BN Paris Abb. 158
(B)23, fol. 87r-115v in the critical apparatus. It has a more tamed text
than the previous two : nonetheless, Bs language too does not coincide
with post-medieval Gz.
The LR, as we have it now, cannot be qualified as the Urtext, but
a descendant that displays a close kinship with an Aksumite ancestor.
The lack of total homogeneity in the spelling and in some syntactic
constructions suggests that the oldest text available to us has been subjected to some revision. Nevertheless, the few changes have taken place
without any perceptible role of Arabic, which made its way into Ethiopia as a language of mediation in fresh translations and revisions since
the XIII/XIVth centuries.24

Overview of the edition and translations of the LR

The Ethiopic LR is a text difficult to read. It is clear from the first
lines that it is a translation, and for that matter a servile one.25 There
are frequent paratactic constructions, an important clue that the ver19. Cf. S. WRIGHT, Catalogue of Ethiopian Manuscripts acquired since the
year 1847, London, 1877, p. 164-165.
20. Cf. A. VAN LANTSCHOOT, Inventaire sommaire des Vaticans thiopiens
251-299, Studi e Testi 219, 1962, p. 469-470. The text of Vat. Aeth. 268, ascribed to the period of Slama II (1348-1388), has been edited together with a
French translation by A. VAN LANTSCHOOT, Le Muson 68, 1955, p. 17-46, as
Contribution aux Actes de S. Pierre et de S. Paul.
21. The various lacunae are often retrieved, chiefly by codex B.
22. The method Arras employed for his critical edition does not look totally
clear. For example, some of the passages in the body of the text seem to be
emendations, with a reference to the A reading in the critical apparatus.
23. Cf. C. CONTI ROSSINI, Notice sur les manuscrits thiopiens de la collection dAbbadie. Journal asiatique 1912-1914, Paris, 1914, p. 100. In this
catalogue, the LR is number 56.
24. For a summary on the period and works of translation from Arabic into
Gz, cf. GETATCHEW HAILE, Gz literature, Arabic period, EA 2, p. 737.
25. J. RUSSIANO MILES says : The ancient translators were satisfied with
their work at a point where a modern translator would begin to manipulate his,
not to change the meaning but to create the proper aesthetic effect, Retroversion and Text Criticism. The Predictability of Syntax in an Ancient Translation



sion focuses exclusively on the text it translates, without worrying as

to whether the user would be able to understand it. They resemble the
Syrian translations known as mirror or mechanical renderings26 and
the Yownaban dproc or Hellenistic School (between the Vth and VIth
centuries), in which the targmanik translators, translated from
Greek to Armenian adhering strictly to the language of the original.27
The Editio Princeps
There is no doubt that it is a contribution which has marked a turning point in the knowledge of the Dormition literature. Arrass edition
has rightly earned wide appreciation. It is a land-mark, but also an
incentive to carry on the investigation of this apocryphon. In the process it is possible to question the following aspects of Arras edition :
1. Selection of variants. In the preface of the text, the editor
declares that he has not registered every variant ne apparatus plus satis
excresceret (p. V). The choice of excluding some variants in order
not swamp the critical apparatus28 is a measure taken by a number
of editors, which however, in my opinion perhaps needs to be reconsidered. This is because even the so-called orthographic variants can
be useful to reconstruct the ancestry of a text. Arras does not tell the
reasons for including or leaving out the variants.
2. In the Latin there are tacit corrections of the Gz, e.g. n. 41 :
SrK : DLA she carried her relatives, has been rendered vocavit cognatos suos. The Latin would have corresponded
to rK : DLA. As a matter of fact, emendations like the
present one, do make sense and help to clarify the text.
3. Arbitrary translations. There are passages in the Latin version
which differ from the Gz. The following are token examples : n. 75 :
@DK : { : kDK they were due to call us but he told
them. The Latin, which has apparently chosen the text of B kD{,
reads : dicebant nobis. Moreover, in the apparatus of the translation
from Greek to Ethiopic, Chico, Society of Biblical Literature Septuagint and
Cognate Studies Series 17, 1985, p. 2.
26. Cf. S. BROCK, Towards a History of Syriac Translation Technique,
Orientalia Christiana Analecta 221, 1983, p. 1-14.
27. R. SGARBI, Traduzione armene di testi greci fra linguistica e filologia,
in Autori classici in Lingue del Vicino e Medio Oriente. Atti del VI, VII e VIII
Seminario sul tema : Recupero di testi classici attraverso recezioni in lingue
del Vicino e Medio Oriente (Milano, 5-6 ottobre 1987 ; Napoli, 5-6 dicembre
1988 ; Bologna, 13-14 ottobre 1989), G. FIACCADORI (d.), Roma, 2001, p. 115122. On the same topic, see, in the same volume (p. 315-323) the contribution
of G. ULUHOGIAN, Ricerche filologico-linguistiche su antiche versioni armene
di testi greci.
28. The expression belongs to H. CRASWALL GLEAVE, The Ethiopic Version
of the Song of Songs. Critically edited, Taylors Foreign Press, London 1951,
p. xxvi-xxvii.



Arras writes : AB om. patres nostri. N. 133, in the body the text says :
Kx{{ : : }s : KX : K : D : XK our
Saviour stroke, with the intention that everyone sees you. B has a different reading : ]K : Kx{{ : }s : {X : : K :D
: XK because our Saviour while admiring made haste so that
everyone sees you. Arrass translation, apparently of the basic text,
reads : Salvator noster, festinavit, suadens ut omnia ostendat vobis.
The footnote of his translation of the B variant reads : quia Salvator
noster miratus festinavit ut omnia ostenderet vobis. Erbettas version
is : Il salvator nostro si affrett, volendo mostrarci ogni cosa29.
The problem is the methodology which seems not to give a full and
realistic account of what the Gz says. In the specific cases above,
the readings of manuscript A have been ignored and mis-represented,
without an explanation.
4. At the end of the Latin version of the Transitus second volume
(Arras 1974, pp. 75-105), there is an appendix (II) meant to address
difficulties in some parts of LR. There are thirty pages which do not
apparently unravel all the hurdles of the text. Some obscure terms have
been left unsolved or treated inadequately.
The few and limited attempts by subsequent scholars to decipher
and untie the textual knots have not brought solutions to the linguistic
problems that bedevil the LR.
Both Chane30 and Arras have carried out the daunting task of translating the Ethiopic into Latin. Even though the Latin version does not
always reproduce faithfully the letter and the spirit of the original,
Arrass edition is an important achievement. The translations of the LR
into modern European languages are based on the Latin, with the consequence that the slips in the Latin version are transferred and sometimes
amplified in the new translations. The treatment of a passage in paragraph 41 of LR is an eloquent example of how things can run out of control. The original reads : ]K : K} : ks because I have
believed in my Lord - which as a matter of fact is not a grammatically
perfect sentence31. Arras inexplicably renders the Gz : quia non credidi domino meo. The variant in the critical apparatus is ]K},
which is apparently a fusion of ]K and K}. ]K} does
not mean anything. In the critical apparatus of the Latin, Arras says that
manuscript B omits the non, a non which is nowhere to be found in
the Gz text. It is a behaviour that does not render justice to the text,
which speaks clearly about Marys belief and not about her unbelief.

29. The note at the end of Erbettas translation (p. 455) does not discuss the
differences between the readings.
30. M. CHANE, Apocrypha de beata Maria Virgine, CSCO 39/40 Aeth.
ser. I, t. VII, Paris 1909.
31. It should read, either ]K : K} : ks or ]K : K}
: Ds.



Obviously, it is difficult to reconcile the reading I have believed him

with the context, but, do editors have the authority to tacitly change
a text, even a problematic one, to harmonize it with the context ? The
texts do not concede room for any benefit of the doubt with regard to
Arrass operation, upon which, unfortunately, the unbelief of Mary
will be constructed. In fact, in the problematic reading ]K},
there is no trace of , the particle of negation. Arrass argument for
turning upside down the text is stated in the following terms : Eorum
quae proxime sequuntur nihil aliis in narrationibus servatum est et non
mirum est amanuenses aut codicum lectores quaedam mutanda esse
censuisse. Ubi Maria se non credidisse fatetur, A particulam negationis
omittit et B mutilat esma iamankewo (because I did not believe him
in emenkwo).32 Can it be taken as a justification to change the text ?
Subsequent translators have followed Arras, some of them without
mentioning what the text says. Shoemakers translation reads : Because
I did not believe in my God, which as a matter of fact substitutes the
original s Lord with God. In the footnote, Shoemaker refers to
Arrass discussion of this passage in De Transitu, i. 88 (Lat.), adding that the Georgian version confirms the presence of the negative33.
Again, can the reading of the Georgian version be a reason to intervene
on the Gz, which in this case is the lectio difficilior ?
Erbetta34 introduces paragraph 41 with the title incredulit di
Maria35. Mary Clayton goes even much further in her elaboration
based on the Ethiopic reading created by Arras. She says : The Ethiopic LR gives a passage virtually unknown elsewhere, which furnishes
an explanation of why Mary was afraid of death. This refers back to an
episode on the flight into Egypt when Mary lacked faith and regarded
herself as having sinned () this must surely be a very early episode,
dating from before the period when Marys sinlessness could be taken
for granted, and later versions, including R, presumably omitted it
because it offended the sensibilities of their redactors. The only other
reference to this is in a Georgian apocryphon which contains only the
beginning of an allusion to it.36
32. Appendix II De Libro Requiei, p. 88.
33. SHOEMAKER, Traditions, p. 314.
34. ERBETTA, Apocrifi, p. 430. Gribomont has a sympathetic view of Erbettas work : M. Erbetta en revanche donne une bonne traduction des textes
majeurs, avec une introduction et une annotation solides, J. GRIBOMONT, Le
plus ancien Transitus Marial, Augustinianum, 23 fasc. 1 et 2, 1983, p. 238.
Shoemakers assessment is quite different : () this version largely reproduces
Arrass Latin translation and comments, also summarizing or omitting the many
very difficult passages of this text, Traditions, p. 290, footnote 1. The handicap
of Erbettas version is that it is the translation of a translation.
35. Apocrifi, p. 443.
36. M. CLAYTON, The Transitus Mariae : The Tradition and Its Origin, Apocrypha 10, 1999, p. 79, footnote 18, which refers to, M. VAN ESBROECK, Apo-



LR, n. 108 reads o : s : D : }Q : Xl :

Kw@...Erbetta renders : Il re invi quattro ancelle, disregarding
Arrass Latin translation : Et tunc misit rex quatuor servos suos.
The issues tackled but still unanswered, and those that have been
simply ignored call for further investigation. While paying tribute to the
praiseworthy achievements in the research on the Ethiopic LR, perhaps
it would be useful to go back to the manuscripts once again in order
to produce a new edition and translation. This applies to Shoemakers
translation as well37, which is itself critical indeed towards the versions
of Arras and especially of Erbetta38. A new edition would be an interesting challenge, an investment worthy making, especially if there is
agreement, as it seems to be the case, that the Ethiopic is the cornerstone for the exact understanding of many developments39.
Finally : would a retroversion of some passages, a translation of the
Ethiopic into Greek be of help ?

cryphes gorgiens de la Dormition, Analecta Bollandiana 91, 1973, p. 55-75,

at 58 and 65.
37. Concerning his English version, Shoemaker says : I have omitted an
extensive section here (105-31), which accounts for approximately 25% of the
Liber Requiei as it has been preserved. This section details the contendings of
Peter and Paul with the devil, in an account that circulated separately during the
Middle Ages and has been preserved in an Arabic version entitled The History
of Peter and Paul, edited by Agnes Smith Lewis, Acta Mythologica Apostolorum (Horae Semiticae 3-4), London, C.J. Clay and Sons, 1904, English translation, p. 175-92, cf. Traditions, p. 347, footnote 167.
38. Shoemaker complains about the treatment of the term LW] by Arras
(101) : Arras fails to translate, instead transcribing it as if it were a proper
name. Nor does he provide any explanation, as he usually does with unusual
words that are present in the text. In the translation published by Erbetta, however, the term is rendered i magi, also without any explanation. Presumably,
LW] is understood to be a corrupt form (hardly infrequent in this text) of
LU, LX, LX (plural : LX}, KSq, KVq), meaning magician,
pagan priest, seer, or spiritual leader. This seems to me a likely solution, particularly given the relative instability of this word even in the lexical entry,
Traditions, p. 346, footnote 165. The entries LU, LUq, are regularly registered
in DillLex coll. 168-169 ; KWKDict, 608. In fact LU is attested in an Aksumite
inscription (n. 2.28) with the meaning of pagan priest, cf. E. LITTMANN, Deutsche Aksum-Expedition herausgegeben von der Generalerwaltung der kniglichen Museen zu Berlin. Band IV. Sabische, griechische, und altabessinische
Inschriften, Berlin 1913, p. 34. In Ardt (Disciples), LU, LUq, appear in
what seems to be an apposition with q idol, giving the impression that, at
least according to this text, they are proper names of a male and female idol,
cf. E. LITTMANN, Ardeet : The Magic Book of the Disciples, Journal of the
American Oriental Society 25, 1904, p. 11 (text), p. 31 (English translation).
39. ERBETTA, Apocrifi, p. 422, n. 2, which is a philological discussion in 12



The text of LR
The main purpose of the present paper is to concentrate on the
nature of the Gz which is one of the aspects of the LR that needs to
be dealt with further.
The language of the LR is often unfathomable. Beside the clear syntactic difficulities, the text is bedevilled by the overall character of the
spelling. The high number of misspellings is susceptible to more than
one interpretation. A single mis-written, missing or additional letter has
proved fatal in the handling of a whole sentence. Some of the misspellings have barred access to the text altogether or resulted in fanciful
explanations. There is no doubt that it takes much more than relying
on an Ethiopic lexicon and grammar to go about texts like the LR.41
Token examples will follow which try to illustrate briefly some of the
orthographic difficulties in the LR.
a) Promiscuity between the first and the fourth order, a common
feature of older manuscripts LXM@ for KXM,
n. 23 ; |@X for {GX, n. 80 ; KK|} for MK|} ; L|q
for K|q, n. 89.
b) Exasperating exchange between the gutturals : o for o
ns. ; ED for ED, n. 55 ; F for F, n. 56. Such
exchanges are not totally consistent, for instance, n. 56, first reads
F and one line after is F. It is a phenomenon which sometimes
affects the meaning of the term, e.g. kq (great things) and mkq
(houses), n. 8042 ; KGS or MIS impf. MIX (forgive), and K@S
impf. LBX (teach), n. 100 : Arrass translation of kDK :
o : PMK : sKGTK, Ubi posuistis quae docuerunt vos,
n. 100 is therefore questionable. Gk{ : ~o : ks : for
40. The texts of reference for the remarks regarding the spelling are two
classical lexica : DillLex and KWKDict. In the current unprecedented season of
many publications of religious texts in Ethiopia, especially in Gz and Amharic, there is a lamentable regression to bad old practices, namely, chronic carelessness in the spelling. The list of recent editions marred with brazen spelling
promiscuity, which is perhaps exacerbated by the heavy influence of the editors
phonetic background (due to regional inflexion), would be too long. To quote
an example : Tsfa Gbr-llase (ed.), Maf Ardt. Wwatw lot.
Blssan Gz (The Book of the Disciples and various prayers in Gz),
Addis Ababa, 1988 (AM).
41. I would dare to say that doing so would be tantamount to reading Euripides or Virgil with a glossary at the end of touristic guides to ancient Greece
or Rome with all the consequences of the case.
42. This is an example which calls for caution not to dismiss lightly variants,
which are apparently solely orthographic, in most of the cases a measure taken
to spare space in the critical apparatus.



@k{ : ~o : ks : /, n. 129. The form which

appears also as , means prevail, win, DillLex col. 943.
The following are extreme cases of orthographic mediocrity, if not
poor taste : x for x, n. 40 ; FD instead of FD,
n. 58 ; }kYK for }kYK, n. 65 ; sosX for sosX, n. 65 ;
MGk{ for Mwk{, n. 86 ; mk}o : | for mk}o : |,
n. 108. However three lines below, it writes mk}o : |.
c) Outlandish and problematic orthography :m~S for JS, n. 41 ;
q[M for q[M ; }o[ : q[M for }o[ : q[M
n. 2 ; qmk for qmk n. 40.
d) Arbitrary punctuation : }k
D, n. 22 ; @DMT for
@DK : MT, n. 25 ; X : Dq : K : q ; MQA ;
wkA : X : QA ; n. 10 ; DK : K : k\{q : Dq ;
qmk :... n. 50 ; KM}q for K : M}q, n. 63 ;
qKK : , n. 120.
e) Metatheses. There is an impressive occurrence of terms with
misplaced letters, for example : XyM (non-existing entry) for SMI
(n. 73) ; for (n. 79). B has corrected both readings.
a - Typically archaic verbal forms : @mkK for @mkK, n. 1 ;
sk{ for smk{, like in the apparatus, n. 45 ; JU for GU, n. 2 ;
o for o, n. 47 ; ] : K : k : xq. There is
some margin for doubt as to whether the k is to be rejected out of
hand as an error or rather, as it looks more likely, an antiquated compressed form. In B it has been replaced by the more obvious ] :
K : ks : xq, n. 65 ; K : }JX for K : }GX, n. 78 ;
g, a contracted form for g, n. 121 ; }gwK for }
gIK, n. 127.
b - Voces vacuae : @mkK : DGgXq : K : ^k : Oqg
: }mkmkg, n. 1 ; K : Xc : mkTS : kA : D{] :
}, n. 82. The with }mkmkg and Xc is unnecessary.
Rare terms
n. 7. - mk : }o : [ : ]K : IS : D :
kcFq : QM : K : No : lIX : kk : MX ;
You and Joseph were satisfied because the dough of this palm tree
is elevated as the wave of the sea over the earth.
Instead of No B has the improved reading N.
IS, in B rightly IV, is the plural form of GU powder, flour, dough, DillLex col. 90 ; KWKDic p. 464 ; Leslau p. 241.
The form IV is not attested elsewhere but it can be legitimately
obtained from GU following the pattern : X : X : X. The
translation branches (Shoemaker p. 294) cannot stand because the
term for branches in Ethiopic is e (DillLex col. 1019) and never



IV. The context of the sentence too, which speaks about food does
not encourage an eventual hypothesis of branches to be eaten.
Commenting on IS, Arras says: Forsitan hoc nomen idem est
ac nomen arabicum ramus, Appendix II De Libro Requiei, p. 77.
The entry , is indeed registered in some Arabic lexica and explained
as branch of palm tree43. The presumed resemblance between the Arabic and the Ethiopic, a good intuition of the editor, cannot be dismissed outright but it remains difficult to defend. The case for
a loanword would be acceptable if other occurrences of IV branches were supplied. The Arabic words for branch are: , .
For PK to place, elevate, cf. Gen 9 :13 c]q : OM :
]o : K|q I will place my bow in the clouds.
The verb PK, to put, elevate, etc. occurs again in a peculiar way
in n. 70. The text reads : qSmk : KdmkS : G[ : B : PL :
QA ;and you find a new grave and there put her body.
n. 15. - DK : SmkG : mk[ : D : DK : oS : ED :
[k : M}q : SlG ;
And if a man gains the whole world and remains (is stuck) over
evil what will his gain be ?.
The form [k snake, viper is attested only here. It seems to
be the singular form of [lq, which is used in Lk 3 :7 qF :
[lq, DillLex col. 364. It is noteworthy that the textual tradition
of the synoptic parallel, namely Mt 3 :7 reads unanimously Vgo :
MX44. The Physiologuss reference to John the Baptists address to
the Pharisees with the term [kq is plausibly an echo of Lk 3 :7. Yet
immediately after, in the paragraph dedicated to the viper, snake, the
Physiologus choses Xg : MX45. KWKDict 844, registers the entry
[mkq and as its first meaning would be viper, evil snake. Then,
quoting from the Kbr Ngst and from Zena skndr, Kidan-Wld
Kfle says that it is the name of a tribe whose people have the features
of humans and of animals.
The sentence is extremely elliptic and it is unclear where Shoemakers translation (p. 299) comes from : And even if a person has gained
the whole world, and he has been abandoned to the beast with the body

43.Cf. The Arabic Encyclopaedia (muy al-muy ocean of oceans)

edited by Buros AL-BUSTANI: -
.524 1870 -.
44. Cf. R. ZUURMOND, Novum Testamentum Aethiopice, Part III. The Gospel of Matthew, Aethiopistische Forschungen 55, Wiesbaden 2001, p. 48.49 ;
45. F. HOMMEL, Die Aethiopische Uebersetzung des Physiologusnach einer
Londoner, Parisier und Wiener Handschrift herausgegeben, verdeutscht und mit
einer historischen Einleitung versehen, Leipzig 1877, p. 10-11.



of a lion and the tail of a snake, what is his profit ?46 As it stands,
by employing the figure of the viper, the text does not seem to mean
anything more than evil.
n. 41. - M : DKUX : K : Ke : K}. The variant (B) is slightly different M : DKUX : K : Me :
K} ;
Arras rightly translates : quia gustasti amarum sicut dulce. Both
manuscripts present an objectionable spelling and syntax. The first verb
is usually written K and more rarely K, DillLex col. 1241.
Ke and Me derive from the same root Kc, DillLex I, coll.
220-221, meaning, lift up, to be high. DillLex, col. 221 has a second
meaning (under II) for Kc : to be sweet. Kidan-Wld spells the
root of the second meaning differently, namely, Koc. He explains
the adjective Moe& M# # KXX& E}| DYM :
] : L[} : {X : Moe : lEF ; sweet, tasty (Gz), tasty
(older Amharic), delicious (Gz), something pleasant to the eye and
to the ear is called mtuq. Kidan-Wld presents the same (but full)
testimonium provided by Dillmann, a passage from Cyril of Alexandria.
It reads : sMFo : S : I & Moc : mkXA{ : mkcD :
Do # DK : }T (cX : ]q) ; who healed the one who
was blind from his birth, and let a sweet light sprout for him, and
all of them applauded/were marvelled (Cyril st)47. Next, KidanWld presents the verbal substantive, Mcq &sweetness, K
&{q ; to be tasty, delight (Amharic). Again his witness is
Dillmanns biblical verse, Judges 9 :11, but while the former gives only
the concerned entry in Latin translation, dulcedo, here we have the full
text kFg : : DkD] : } : }P : ED{ & qkDK :
kD] : y{ : Mqcq : W : mkT : }Q : ED :
(K\ : ) ; the trees told to the fig : Come ! Reign
over us !. But the fig told them : Should I abandon my sweetness and
my blessed fruit so that I reign over the trees ? (Judges 9 :11). Then,
Kidan-Wld adds the following precious note : KIq : } :
ku : ]Ioq : {}M : h}M : Me : Mcq : ED&
KMBV}M : QT} : ]EE{hq : kMe : sk : opF ;
Nevertheless, the books because of a mistake of the scribes call this
one (mtuq) and the other one : muq,mqt. The teachers were not
able to retrieve its root, so they explain it in the same was as muq,
KWKDict p. 616. The form Ke in the manuscript A of the LR (duly
corrected by B), does not exist and it is perhaps an illustrative case of
the difficulty around the entry, described by Kidan-Wld. The asyn46. The source of Shoemakers expanded version seems to be Leslaus
Comparative Dictionary, p. 483. Shoemakers elaboration of the term [k in
footnote 33 too appears to need at least one or two witnesses to prove its claim.
47. st is an abbreviation for stgbu (Compilation) is a collection in
which we find the De Recta Fide and the Prosphoneticus ad Reginas.



deton K} emended by B (K}) contributes to the harshness of the basic text.

n. 75. - }o : K : q}c : s}o : Ik : K : q}Q
: M{CK : QVo : scL : D : kS : s : IM]
And this, so that you may build up this people and so that from
them you take the tenth of firstlings and every firstborn that the womb
brings forth. Although the passage does not present errors, the choice
of vocabulary deserves some comments. {c to build up, to make a
wall, DillLex col. 680, Leslau p. 386, is normally connected to material buildings48, when predicated to humans though, it sounds awkward.
K : qI} : s}o : Ik, would have been a more usual language.
The word IMQ uterus, womb, pubes, Leslau p. 234, is
attested. sqI : D : IMO the one that opens every womb,
Ex 13 : ; SK : DFCK : Do : mk[ : ]
VEg : Do : mk[q : K|gq : }o : K}D : IM\
; and he pierced both of them, the Israelite man and that Midianite
woman towards her womb, Nm 25 :8. These are well known witnesses
to this entry which is connected to other Semitic languages in DillLex
col. 76. According to Ethiopian traditional grammarians and lexicographers though, IMQ/yM], would belong to the same semantic field
of wM] five and would mean the fifth of the organs that have
a hole, namely : ears, nose, mouth, anus, KWKDict p. 480. The most
usual terms for womb are LI} and X]. The word IMQ and
especially its Amharic cognate M] are considered foul language.
n. 81. - @nK : DD : wFmkg : D{]...
And give those who are milking my soul (). The entry wFmkg
needs to be re-written as GFmkg as far as orthography is concerned.
As the expression milking the soul = exploiting somebody is unheard
of in Ethiopic, GmkFmkg : D{], would be a better option. DillLex col. 101. GmkD means to spoil, plunder..., Leslau p. 224.
n. 125. - oK : }Q : wk : Do : o|V : D :
s}o : }s : mkF : D : D : Dq : JU : }s : q :
K : kK : ss : GgXq : lIo : qq|

48. There are a few exceptions of {c predicated to elements different

from buildings. Gen 2 :22 is an example. It reads : {d : smkGX :
Do : K : n : }o : {Q : MM and the Lord built that bone
from the side that he had taken from Adam. {d and he built is a faithful
rendering of the Greek (LXX) and of the Hebrew of Gen
2 :22. Cf. also Ez 39 :15 which reproduces literally the Greek and Hebrew verb
to build. The text says : {c : ED@ : qMXo they will build a sign
upon it.



And the king came back to his daughter and told her all of this
saying : Woe ! Woe ! My daughter, go making haste and bring them out
as the apostles have ordered but do not delay.
D : D is not familiar. The more common expression of grief
or anxiety is [ : [, Leslau pp. 521.522.
JU is the older form for GU.
The form o| in this passage, is the intensive of } to
loiter, delay, Leslau p. 197. It appears in DillLex col. 1181 but not in
KWKDict p. 322. In fact, it is not a commonly used form.
n. 126. - L : cDK : IKo : qI Water and ink and
soot, charcoal, and you shall write : this is a lectio difficilior49. The
variant reads cDK : L : IKq : qI n. 126. IKq is
a verbal noun from the rare root GKo cause to char, KWKDict
p. 450 ; Leslau p. 235.
n. 49. - The esoteric divine names : LTF ; LTF ; LX|p ;
mkpX ; have been considered in Arrass appendix (p. 91). Subsequent editors refer to Arrass note.50 His bibliographical references can
be supplemented with additional material. Jesus reveals a long list of
his names with their explanation in the text known as Ardt Disciples. Among the thirty five hidden names of Jesus and disclosed
to his holy Disciples, feature : Abyater which means Merciful ;
Mnyater which means Chastiser ; Stnawi which means Creator of
all ; Maryon which means Preserver of all.51
n. 113. - Y] : ss : M : Do : L : @nK :
Xeo ; Arras, renders Xeq with aspersionem. The translation is
not wrong but diminishes the weight of the entry which denotes the
prayer of exorcism, recited either directly on the person or on the
water, cf. KWKDict p. 837. Dillmann translates it as : 1) incantatio ; 2)
exorcismus, DillLex coll. 284-285.

49. For a similar construction cf. the variant of 2 Jn 12, cf. J. HOFMANN und
S. UHLIG, Novum Testamentum Aethiopische : Die Katholischen Briefe, Aethiopistische Forschungen 29, Stuttgart 1993, p. 204.
50. E.g. SHOEMAKER, Traditions, p. 319, a footnote to Maruyal which refers
to Arras.
51. Cf. E. LITTMANN, Ardeet : The Magic Book of the Disciples, Journal
of the American Oriental Society 25, 1904, p. 1-48, especially, p. 9-10 (text)
and p. 28-29 (English translation) ; Tsfa Gbr-llase (ed.), Maf Ardt.
Wwatw lot.Blssan Gz (The Book of the Disciples and various
prayers in Gz), Addis Ababa 1988 (AM), especially, p. 83-85. For lists of
the esoteric names in apocryphal literature, cf. Enoch 6 :7. On the cultural background and value of divine/esoteric names in Ethiopia, cf. a summary in Tedros
ABRAHA, Effigie di Maria III o Specchio della Visione Mlkaa Maryam
alsit - Met Ray, Marianum 71, 2009, p. 429-431.



Obscure terms
n. 20.- K : ]omkc : D[] : K : mkcF :
o : s : W : keLCK : kDK : GgSK : ]o
: : : keLK : kK : e}\@ : mksy : L@ :
And they came and pleaded with Jesus that it sprouts and bear fruit
before them at that time. He told them : Go to this tree in front of you
and as its branch is manifold so are its leaves.
The passage displays a number of so-far unsolved textual problems. K : mkcF : so that it may sprout would be referred to
the plant, but the subject of the action here is explicitly Jesus. Consequently the clause needs to be K : mkcF : that he (Jesus) might
cause (the plant) to sprout. There is a discrepancy in number between
the demonstrative and the plural . The translators render the
unknown entry e}\ which occurs twice in this paragraph, with frondes
and branches without explaining how. e}\(@), in B e}P(), can
be conjectured as a defective form52 of e}[ from the quadriliteral
verb c}[, DillLex col. 449 ; KWKDict p. 698 ; Leslau p. 436 reads :
gather vegetables by nipping off, break/pluck off vegetables.
n. 25. - ]K : }o : o : }F ; K e :
The form K e which appears to be commonly shared by the
witnesses of the edition, does not exist in Gz. If the reading behind
our text is chosen, it is possible to envisage the verbal substantive
e friend, familiar from c, DillLex col. 998, KWKDict 689,
Leslau p. 78. Psa 54 :14(LXX) reads. }o[ : mk[ : sK : }] :
LMX : e :
. The text can be therefore read : ]K : }o : o :
}F : e because you are virgin and (a close) friend.
n. 97. - k : X{ : F : sk} : [X : : FKq :
emk : kmk} : kmkq : K : }o : Anq : :
And again we saw another one who was condemned in fetters, two
were keeping him (in) darkness and were slapping him on his face with
round stones like soldiers.
For mk} : kmkq cf. DillLex col. 848 ; KWKDict p. 518.
The enigmatic entry Anq is very likely a corruption of Gk, or
Gk, Gmk, pl. Gkq, army officer, chief, manager, cf. DillLex
col. 104 ; KWKDict p. 431 ; Leslau p. 225. The form Gk from Gk
to give security, give bail, be a bondsman, to guard is attested in the
New Testament, e.g. Gk : Mq Acts 10 :1.

52. The LR has several defective terms. There is no need to list them as
the phenomenon of scriptio defectiva is part and parcel of any writing activity,
including our present computer era.



n. 108. - o : s : D : }Q : Xl : Kw@...
And at that time the king sent four of his healers. Arras text reads
Kw@ and has the variant {@. Both readings are a corruption
of Ky|{@, pl. of Ky}, healer, saviour, DillLex col. 1113.
n. 123. - QV : sqD with the variant sq : Dx.
There is no record of D. There is DG/[k/SK to insult,
curse. Other explanations of DG are : to lay flat, to bow down in
reverence, to pay tributes, KWKDict p. 564. None of these meanings
is of help to understand our passage.
a - Disagreement in gender. It is true that in Gz, the definition of
gender of a good number of subjects is not clear cut. This is especially
the case with inanimate things53, but the gender of animate subjects too,
like plants, can be fluid with the result of passages such as : }o{ :
kcFq : {gy, n. 6.
b - Disagreement in number, e.g. kDK : [] :... JX,
n. 20 ; S : K|, n. 50 ; kDK : D] : DY],
n. 78 ; qqGK : DqF : G] : ]K, n. 120 ; o :
o : : K[E}, n. 121.
c - The employment of the accusative reflects older Gz syntax, for
example with n, Fn, n. 61 and . Elsewhere : ]K : MFo :
KX : } for ]K : MFo : KS : }, n. 4 ;
kD : D}o : kcFq : }} : X] per kD : D}o :
kcFq : }} : X[, n. 7 ; oMw : kk|oCK : o :
s for oMw : kk|oCK : o : s, n. 50 ; o :
s for o : s, n. 77.
The use of the preposition D as accusative modifier is noteworthy :
]K : DDK : GgXq : { : wk for ]K : DK
: GgXq : { : wk, n. 1 ; o@g : D{[ : F for
o@g : {[ : F n. 82 (not emended in B).
d - Missing constructs : D}F : smkGX for D}D :
smkGX, n. 40 ; DKkDpq : Imk for DKkDpo : Imk,
n. 75 ; Xo : Fmk for Xo : Fmk, n. 82.

53. The term G sun, is generally masculine, but in some cases it is

referred to as a feminine, e.g. ckr : G : Dq : L} : p :
M{ the sun made a sign to the thief and the thief truly believed, om
Dggwa. Z Qddst arb (Hymnary for Lenten Time), Addis Ababa 1959
(AM), p. 23 col. 1. MX earth which is always feminine, appears exceptionally in the masculine in Mt 5 :13, }qK : o : : DMX You are
the salt of the earth. For a scholarly treatment of the issue of gender in Gz,
cf. A. DILLMANN, Ethiopic Grammar. Second Edition enlarged and improved
(1899) by Carl Bezold, translated by James A. Crichton, Williams & Norgate,
London 1907, p. 271-273.



e - Improbable constructions. @D : : ]o : @{M :

M[mk : qF. The variant has a slightly different reading
@D : ]o : @{M : : M[mk : qF n. 125. The
word order in both texts must be corrected as, MqF : [mk.
f - Misplaced word order K : ck : s@k : DK : qs
for K : ck : qs : s@k : DK, n. 114 ; L]} : @S :
D for L]} : D : @S, n. 122.
As has been already mentioned B has made some changes
which include re-ordering the sequence of the words, for instance :
] : So : W in B reads ] : So : W :
DsmkGX, n. 86.
In the sentence ]K : MLFq : KF, n. 85, the preposition M is a redundancy at the expense of Gz syntax and as
such has been epuratedby codex B, which reads : ]K : LFo :
KF. Non doubt that these passages, which in Gz make little
sense, together with intractable sentences like : : M{ : V] :
sk : sD} : FD : dD n. 119, are important specimen, for
the reconstruction of the Vorlage of the LR.
g - Trivial errors : kD :... : ML : D@X for kD
:... : ML : D@X, n. 70 ; k : soyn for k :
soyl, n. 76 ; n : wk : Dq I entered into my daughter which sounds somehow obscene. B reads : l : ]o : kq
: Dq I introduced into my house and my daughter..., n. 109 ;
kDK : }qK{ : Y] : D] : FCK, n. 123.
h - Lighthearted/messy transitions from the third to the first singular
and plural persons.
i - Proper names : very often the text reads LX54, ns.,
instead of the more usual LXM which is also present in the LR, e.g.
n. 40. The name of the Seraphim appears in two forms : [V} and
[VF, ns. 132.133.
Style of the LR
The style of the LR is apparently nave. The passage recounting the
story of Peter and Paul in Rome standing before the king reproaching them is a good example of this characteristic. Peter says : Paul,
who shall speak ? You or myself ?. Paul replies : It is you that should
speak, my father Peter !, n. 106.
One of the most outstanding features of the LR is the overwhelming
volume of direct speeches. From the beginning to the end, the book is
54. This is how Jesus calls his mother, cf. 98. John and the others call
Mary : Our sister and she calls John : My/our father. Peter though addresses
Mary in a different way : Lover of mankind, Mary, our mother (n. 55).



a succession of speeches, composed by an omniscient author55 who

knows the protagonists, registers theirs conversations as though he was
an eyewitness and can even read their thoughts. Among other reasons,
this stylistic aspect is perhaps one of the causes at the root of the
exclusion of the LR from the liturgical life in Ethiopia.
No doubt that any interlocutor of LR will find huge difficulties in
finding out its way in the thread of the argumentation and in the overall
construction of the work. One of the stumbling blocks is the chronic
wanderings from a given topic. Plausibly, the author has had his own
literary strategy and objectives, which have proved hard to understand
for subsequent generations.
Symbolic language
The LR has a wealth of symbolic language. N. 52 says : I am in
the midst of the apostles like a vineyard that bears fruits in its days.
N. 120 illustrates prayer through military images. The text reads :
Prayer is our shield, its words our arrows, our double-edged sword is
faith and (good) works.
Rudimentary theological language
The Christology of LR features high among its peculiarities. It is a
raw, embryonic Christology. Manuscript reads : { : o : PF] :
omkS : F : Fn : sk{ ; literally : I am the third
made, and there is no son who refuses me. Beside the extravagant
doctrinal formulation, this is one of the examples in which orthography is a casualty with consequences on the meaning of the proposition.
Such a rendering is a slavish translation where the sense is difficult to
be pinpointed. B tries to iron out the harshness of the text. In the process the reading of A is enlarged and modified. B reads : { : o :
sQF[ : omkS : kKDq : { : F : Fn : smk{ ;
I am the one who was made three time in the divinity and I am his
son and there is nobody who is greater than me, n. 25. One of most
noteworthy elements in both readings is the verb mkS make, create,
predicated to the Son. It is the verb employed in Gen 1 :1 : kcL :
mkS : smkGX : [Lo : MS In the beginning God made
the heavens and the earth. The verb was condemned by the Council

55. The expression omniscient narrator was coined by P. LUBBOCK, The

Craft of Fiction, London 1921, p. 115, 120, 197-198, 255.



of Nicea (325) which has professed that the Son was begotten and not
The absence of the technical term QE[ in the whole of LR cannot
fail to be noticed. It is equally noteworthy the employment of totally
inadequate expressions, like : PF] : omkS : the third made, and
sQF[ : omkS : kKDq : the one that was made three times in
the divinity. These readings are sufficient to explain why the LR did
not know wide circulation. Its langague is simply unfit for public reading, especially in the context of the liturgy. After Nicaea nobody would
have dared to utter these expressions.
Angel Christology
At the beginning of the book, there is an imperceptible transition
from the angel who was talking to Mary to Jesus, in fact the impression is that Jesus splits himself from the angel, n. 1. There is a similar phenomenon at the beginning of book five which starts with the
story of a huge horse, bigger than a legion. The equine pronounces
a speech to discourage the troups of Peragmos from pursuing Peter
and Paul. The text reads : After saying this the horse went up. Jesus
ascended to heaven.
What we observe in the LR is a wavering Christology, undecided in
its affirmation of the divinity of the Son.
There is no clue of anti-arian and anti-nestorian polemics, otherwise always part and parcel of the majority of liturgical texts in Gz.
These are all elements that play in favour of the hypothesis of a very
primitive text.
Other theological aspects pointing to the antiquity of the LR
Despite the account of an attack of some Jews against the litter carrying the body of Mary and the funeral cortege, the LR, differs from
the overwhelming majority of Ethiopic religious literature which is normally very critical towards the Jews. The LR is a loud exception, the
book is anything but hostile to the Jews. In fact ns. 25-35 are a midrash
of the first chapters of Exodus. The sympathetic attitude towards the
Jews is expressed also in the marked anti-paulinism. Paul is often
referred to as a new plant, n. 47, always in an inferior position with
respect with to Peter, John and the other apostles, n. 78. At the end of
n. 84, Paul begs again Peter to introduce him to the magnalia Christi,
so that he too, like Peter and the others could preach their doctrine. In
56. G. ALBERIGO et alii, Conciliorum Oecumenicorum Generaliumque
Decreta. Editio Critica. I. The Oecumenical Councils. From Nicaea I to Nicaea
II (325-787), Brepols, Turnhout 2006, p. 19.



n. 103, God himself comes in a cloud and calls Paul on his side. The
devil utters cries of protest : Jesus, Son of God... why do you elevate
to your greatness this (sic !) who is called Paul, who (unlike the twelve)
did not fight against me. Let him first come and face me ; if he wins
he will ascend to you.
Legacy of the LR in the Ethiopian Church
Some of the data of LR are recognizable in the doctrine and in the
liturgical practice of the Ethiopian Orthodox Church.
a. Parts of the text of LR have survived in the liturgy. The yaredian
hymn for the feast of the Assumption on August 22,57 is a florilegium
of passages drawn from LR, ns. 46-47.52. The hyms reads :
okK : kK| : [L : DK : GgXo@ : ol@D :
kk|oCK : M}q{K : ]omk{ : s{58 : ]o : o :
}c : kG & kD : Y] : DD] : D : w :
D] : q|e : eK : smkGX & kD : D] :
DY] " : { : D : sMeLK : ]K : G] :
oF : { ; oQG : DK : GgXq : k}o : skDK :
D59 : mkg : kp : DLXM : : qk : LXM : }Qg :
}mkmkg : Do : KG : : qk : LXM : : M}
o : Q : GX : ]o : Iq : sDDM60& : qk :
LXM : osX : G}] : skD : De : K : qSe : k :
}@61 ;

Gathering themselves together through a heavenly cloud, his62

apostles told to one another : What is the reason why we gathered
together63 in unity before this one door ? Peter said to Paul : Come on,
my brother Paul, pray and do not doubt before God ! Paul told Peter :
How can I, a new plant, pray in front of you ?64. All the apostles
rejoiced because of the words that Paul told them. They introduced
Mary in her house. Mary said : Oh ! Take and read this book !65. Mary
said : Oh ! I will go out from this body of mine and go to eternal

57. Cf. Maf Dggwa zQddus Yared. Bmraq Goam Hagr

Sbkt ft Bet asattamint tattm, 1999 AM Hymn Book of Saint Yared.
Published by the diocese of eastern Goam, 1999 (AM), s.l, p. 519-520.
58. LR, n. 46.
59. LR, n. 47.
60. LR, ns. 51-52.
61. LR, n. 42.
62. His is referred to the Son.
63. LR, n. 46.
64. This dialogue between Peter and Paul appears almost ad litteram in LR,
n. 47.
65. In LR, ns. 44-45, Mary asks the other apostles to leave her alone in her
chamber with John and only to him addresses the present call to take the book.



life !66. Mary said : Oh John remember what the Supreme told you
when you were dining leaning in his side.67
The LR states that Jesus resurrected on the third day after his death,
Mary was taken to heaven on the fourth, n. 11. The believer that keep
the word of the Saviour will share the same destiny in accordance with
the promise given by him : I will come to them and let them enter
into sweet-smelling Paradise, their flesh will remain fresh and will not
exhale bad smell, because they have dominated themselves while they
were living on earth and will live there until the day of resurrection.
It is the officially sanctioned belief in the intermediate eschatology,
according to which the soul of the upright, after death will go to Gnnt (garden of rest, paradise), waiting for the resurrection of the body,
while the unrepentant sinner will descend into Sheol68.
Insistence on the secret
Attachment to and high esteem of the hidden, fascination by the
mysteries to be participated to a restricted circle only is one the more
visible peculiarities of LR. It is a vision that will leave an indelible
mark in the mens, in the spirituality and religious practice in Christian
Ethiopia. In the LR, the name of the great angel is hidden and Mary
can reveal it to the apostles only and not to others, so that they may
not know my power and my great power, n. 1. The mysteries are hidden from the scribes and from the fool, n. 17.
In the past, in Christian Ethiopia people were extremely reluctant
to reveal their baptismal name. The sacraments, the doctrines are protected by discretion. It entails even barring strangers from entering
church premises and monasteries. The guideline of Mt 7 :6, but literature like LR are the source, for such an attitude which is meticulously

66. LR, ns. 51-52.

67. LR, n. 42. Cf. Jn 13 :23.
68. The official diglot document of the Ethiopian Orthodox Church issued
in January 1995, p. 20, puts forward again this position. The text reads : All
those who have departed from the time of Adam and will pass away until the
second coming of Christ, will arise in the union of body and soul. The souls
of the righteous shall abide in paradise, and the souls of sinners in Hades until
the end of this world. But on the last day, when our Lord and Saviour Jesus
Christ shall come in His glory, to judge the living and the dead, the souls shall
be united to their bodies on the day of resurrection and they shall arise from
the dust of the earth.

Paradise in Antiquity. Jewish and Christian Views, d. Markus
BOCKMUEHL, Guy G. STROUMSA, Cambridge, Cambridge University Press, 2010, 272 p., ISBN: 978-0-521-11786-9.
Ce volume, issu dune longue confrence tenue du 3 au 31 mars
2008 au Centre dtude du Christianisme de lUniversit Hbraque de
Jrusalem, est un collectif dirig par deux minentes personnalits de la
recherche en judasme et christianisme antiques. Paradise in Antiquity
runit des articles sur les conceptions et descriptions du sjour des
bienheureux, depuis les premiers essais dans la littrature du second
Temple jusquaux systmatisations des IVe-Ve sicles.
Dans son Introduction Guy G. Stroumsa rappelle les multiples
interprtations dun objet qui, de prime abord, pourrait paratre vident,
le paradis. En effet, celui-ci est mobile tant au niveau chronologique,
puisquil est le lieu des origines et peut-tre celui de la fin, mobile
gographiquement sil est terrestre, quand il nest pas atopique et
La reprsentation du paradis est en fait intimement lie celle de
lhumain, puisque ce dernier y est n et est destin y retourner, dans
le meilleur des cas. Cest justement dans le cadre dune anthropologie ou dune sotriologie reposant sur la Gense quon le retrouve
lpoque des rabbis et des premiers chrtiens. Cest le socle commun qui permet la comparaison des deux religions. Dans ce volume,
judasme et christianisme sont toujours interrogs conjointement, et les
deux parties qui structurent Paradise in Antiquity se fondent essentiellement sur des critres pistmologiques et chronologiques. Selon le programme annonc, la premire partie regroupe les contributions consacres linterprtation de la Gense dans les crits juifs et chrtiens au
tournant de lre courante, tandis que la seconde comprend celles sur
les communauts porteuses de ces traditions et le contexte dans lequel
elles ont dvelopp leurs croyances. vrai dire, la sparation est plutt
artificielle, ce qui nenlve rien la qualit des contributions, quelles
soient affectes lune ou lautre partie.
Avec The Messiah in the Garden: John 19.38-41, (Royal) Gardens,
and Messianic Concepts, Joachim Schaper ouvre lensemble du
volume en sinterrogeant sur le mot mme de paradis dans la Bible
hbraque et la Vulgate. Il se tourne ensuite vers le rfrent, savoir
le jardin et surtout sa symbolique dans la littrature juive hellnistique, avant de remonter aux traditions smitiques plus anciennes ayant
trait notamment au motif du roi jardinier. Il revient finalement la

Apocrypha 23, 2012, p. 247-279



charge symbolique du jardin dans le Nouveau Testament en particulier

dans lvangile de Jean.
Philos scholarly inquiries into the story of paradise de Maren
R. Niehoff rappelle que lexgse alexandrine suit toujours deux lignes
interprtatives, lune littrale et lautre allgorique. Loriginalit de Philon dans le cas prcis de son interprtation de la Gense est dutiliser
des techniques de la lecture littrale pour prparer la lecture allgorique
au lieu de les tenir toutes deux pour correctes et dimportance gale.
Richard Bauckham sintresse au Paradise in the Biblical Antiquities of the Pseudo-Philo, un texte du Ier ap. J.-C., sans doute compos
en hbreu, traduit en grec, puis en latin, langue dans laquelle il est
conserv aujourdhui. linstar dautres crits juifs contemporains,
ces Antiquits bibliques situent le sjour post-mortem des justes dans
lespace arien. Mais doit-on reconnatre en ce paradis lden dont ont
t chasss les humains? travers lexemple des pierres prcieuses de
Havilah (Ant.Bib. 25, 11), R. Bauckham tisse tout un rseau de rfrences dans la littrature apocryphe juive et chrtienne qui atteste de
la diffusion de la croyance en lidentit du paradis perdu et du lieu de
repos des justes.
Dans Paradise, Gardens, and the Afterlife in the First Century CE,
Martin Goodman part du double constat que Flavius Josphe appelle
paradeisos le lieu do furent expulss Adam et ve, mais quil ne
fait pas tat de la croyance en la survie de lme dans ce paradis. Dans
le Nouveau Testament en revanche, le terme revt dj une dimension
eschatologique. M. Goodman va alors sonder la littrature juive antique
pour expliquer le silence de Josphe.
Grant Macaskill se concentre sur les trois occurrences de Paradise
in the New Testament, savoir Lc 23, 43, 2 Cor 12, 4 et Ap 2, 7,
mais considrant quil ne sagit que de la partie visible dun iceberg
de croyances, il largit le champ aux contextes thologiques qui voient
apparatre le terme, savoir la christologie adamique de Paul dune
part et de lautre, linstauration du Royaume.
Avec Quis et unde? Heavenly Obstacles in Gos. Thom. 50 and
Related Literature, Simon Gathercole sintresse aux gnostiques, pour
lesquels le chemin menant au paradis est presque plus important que
la destination elle-mme. Tirant leurs racines du judasme du second
Temple, les ascensions clestes se sont dveloppes comme un vritable genre littraire parmi les gnostiques, au cours des trois premiers
sicles de lre chrtienne. S. Gathercole examine les types dembches
et les figures de gardiens qui mettent lpreuve les mes, limage
des chrubins verrouillant lentre du paradis (Gen 3, 24). Il parvient
la conclusion quun texte aussi prcoce et allusif que lvangile de
Thomas rpond aux mmes interrogations que les ascensions, quoique
nen partageant pas la forme littraire.
Comme lexplique Sabrina Inowlocki dans Tertullians Law of
Paradise (Adversus Judaeos 2): Reflections on a Shared Motif in



Jewish and Christian Literature, linjonction de Gen 2, 16-17 a t

perue chez les juifs de mme que chez les chrtiens comme une
annonce des commandements venir, et cest afin de polmiquer avec
les juifs sur un terrain commun que Tertullien sempare de la question.
Les arguments avancs par le Carthaginois trahissent la connaissance
de dbats internes au judasme portant sur larticulation entre cette loi
dite naturelle et la Loi mosaque. Pour faire pice au judasme, Tertullien aurait tent de ractiver la premire lusage des chrtiens.
Dans The Language of Paradise: Hebrew or Syriac? Linguistic Speculations and Linguistic Realities in Late Antiquity, Yonatan
Moss met en avant une question qui piqua la curiosit des commentateurs anciens: quelle pouvait-tre la langue du paradis? Bien
entendu, Y. Moss ne cherche pas rpondre la question, mais analyse les rponses des rabbins et des premiers chrtiens. Ses conclusions
montrent combien le point de vue est important et comment le sujet
peut tre instrumentalis: selon les rabbins, le langage utilis par Adam
et dans lequel il a nomm les animaux tait laramen queux-mmes
parlaient encore. Pour les juifs comme pour les chrtiens, lidentit de
la langue parle au paradis et celle dans laquelle fut donne la Loi
nest pas vidente. Lhbreu est parfois envisag comme un langage
artificiel mis par Dieu pour encoder la Torah. Tel est lavis dun chrtien dOrient comme Thodore de Mopsueste, pour qui cest le syriaque
qui rsonnait au paradis.
Menahem Kister extrait deux motifs de Gen 3, 22-24 dans The
Tree of Life and the Turning Sword: Jewish Biblical Interpretation,
Symbols, and Theological Patterns and Their Christian Counterparts,
puis les replace dans le tissu de leurs commentaires juifs et chrtiens.
Les deux traditions se sont dveloppes de manire dialectique, chacune pour servir son camp dans la controverse judo-chrtienne. Ainsi
le chemin de larbre de vie sera tantt la Torah ou leucharistie et le
glaive des chrubins, la circoncision ou le baptme.
Erotic Eden: A Rabbinic Nostalgia for Paradise de Galit HasanRoken sinterroge sur lutopie des origines, bien connue des folkloristes,
parmi les juifs de Palestine aux IVe et Ve sicles.
Comme le souligne Gillian Clark dans Paradise for Pagans?
Augustine on Virgil, Cicero, and Plato, Augustin opre le choix stratgique de citer les mmes rfrences que ses adversaires pour mieux
les rfuter. Aussi, pour rpondre son contemporain Macrobe qui commente le Songe de Scipion selon le dernier livre de la Rpublique de
Cicron sur le mode platonicien tout en invoquant Virgile, il va lui
aussi user de lhritage des trois grands hommes pour dfendre le point
de vue chrtien.
Augustin est galement au cur de la contribution dmile PerreauSaussine Heaven as a Political Theme in Augustines City of God
qui commence par dfinir la notion de cit chez lvque dHippone,
puisque celle-ci est essentiellement et tymologiquement lorigine



de la politique. . Perrau-Saussine conclut que, dans une certaine

mesure, il est possible dassimiler la Cit de Dieu au paradis, mais
aussi peut-tre lglise. Nanmoins le succs terrestre de cette dernire ne doit pas faire conclure au succs inconditionnel de la Cit
de Dieu. La politique a permis lglise de triompher, mais dans sa
manifestation visible, et la vritable Cit de Dieu est intrieure. Elle est
transcendante et par essence chappe aux rgles de la cit terrestre et
donc la politique
Si pour Augustin le paradis est la fois partout et nulle part, on a
longtemps cherch un paradis terrestre trs concret au point de linscrire sur les cartes, ainsi que le rappelle Markus Bockmuehl dans
Locating Paradise. Les grands explorateurs comme les imprimeurs
humanistes y font encore rfrence quand bien mme ils savaient distinguer lallgorie du fait historique ou gographique. M. Bockmuehl
remonte dans le temps et constate que jamais le paradis na t conu
comme strictement spirituel, mme chez Philon et Origne, et quil se
doublait toujours au moins dune interface terrestre. Celle-ci avait peuttre pour fonction de rappeler que le royaume divin est attenant au
monde des hommes.
Il revient finalement Alessandro Scafi de signer lEpilogue:
A Heaven on Earth dans le prolongement direct des deux derniers
articles. Auteur dune somme intitule Mapping paradise, A. Scafi livre
une synthse des reprsentations du paradis tout en passant en revue
lensemble des contributions du volume dirig par M. Bockmuehl et
G. G. Stroumsa.
Paradise in Antiquity nest pas une systmatisation de toutes les
conceptions sur le paradis une priode donne ou selon un point de
vue. Il sagit dactes de colloque. Quoique lis par la thmatique densemble, les sujets sont en dfinitive laisss au bon vouloir des auteurs.
Ceux-ci tant dminents spcialistes, Paradise in Antiquity se trouve
tre un ouvrage la fois spcialis et lisible, de trs haute qualit.
Charlotte TOUATI
Universit de Lausanne

EHRMAN, Bart D. PLEE, Zlatko, The Apocryphal Gospels.

Texts and Translations, Oxford, Oxford University Press,
2011, XII + 609 p., ISBN: 978-0-19-973210-4.
Cet pais ouvrage vise combler labsence de recueil prsentant
un public anglophone les vangiles apocryphes, avec un accs au texte
original. Le projet qui le sous-tend est par l-mme proche de celui
de Aurelio de Santos Otero, dont le recueil, intitul Los Evangelios
Apcrifos, nest plus prsenter (6e d., Madrid, 1988). Une comparaison entre ces deux recueils fait toutefois apparatre dimportantes diffrences. En premier lieu, A. de Santos Otero avait consacr quelques
pages discuter de la notion dapocryphes du Nouveau Testament



et prciser ce quil entendait par vangile, avant de donner une

bibliographie gnrale sur les apocryphes. Rien de tel ne figure chez
Ehrman Plee, qui se contentent dune trs brve prface, o les apocryphes sont dfinis de faon assez caricaturale comme les livres qui
ressemblent plus ou moins ceux qui ont t inclus dans le canon du
Nouveau Testament (vangiles, Actes, ptres, Apocalypses), mais qui
en ont t exclus (p. VII), une dfinition traditionnelle, mais dont le
caractre problmatique est bien connu. De ce fait, les lecteurs nont
droit aucune information synthtique sur la gense des textes cits, ni
sur leur statut et leur rception; aucune bibliographie gnrale ne leur
donne accs la littrature secondaire sur ces sujets pourtant essentiels.
Les deux ouvrages se distinguent galement par leur contenu : A. de
Santos Otero avait donn droit de cit aux vangiles tardifs, brivement
prsents au terme de lintroduction gnrale, ainsi qu de nombreux
textes perdus ou fragmentaires ; il ne stait en outre pas limit aux
christianismes grecs, latins et coptes. Sur ce point aussi, louvrage de
Ehrman Plee se diffrencie de celui de Santos Otero en ce que le
corpus pris en compte est moins large on notera notamment labsence des rcits sur la mort de Marie. Les limites du corpus pris en
compte par Ehrman Plee ne sont dailleurs pas clairement dfinies :
le volume veut prsenter tous les matriaux vangliques non-canoniques du christianisme primitif (p. VIII), mais il ne parle pas dun
texte important comme les Questions de Barthlemy et laisse de ct
une partie des textes coptes non attests Nag Hammadi mais dont
lexistence est avre (comme lvangile dve) ; on peut en outre
regretter labsence de lvangile du Sauveur, justifie p. VIII par le
fait que ldition critique dfinitive na pas encore paru mais bien
dautres textes auraient d tre laisss de ct si cet argument avait
vritablement jou. Dautre part, ce volume inclut des textes tardifs,
comme la Mort de Pilate, qui nappartient pas au christianisme primitif les auteurs prcisent bien quil sagit dun extrait de la Lgende
dore de Jacques de Voragine , mais alors pourquoi ne pas avoir
donn droit de cit aux vangiles arabes et armniens de lenfance, la
Vie arabe de Jsus ou la Lettre de Jsus-Christ sur le dimanche ? Il
est en outre surprenant de trouver dans ce recueil un possible fragment
du Diatessaron de Tatien.
Si, sous ces divers aspects, louvrage de Ehrman Plee est en
retrait par rapport celui de Santos Otero, il nen rendra pas moins
un certain nombre de services. Les auteurs ont en effet utilis les ditions les plus rcentes des textes (mais pour la recension grecque B
de lvangile de Nicodme, ldition du soussign, parue en 2008, a
chapp aux auteurs). Ils les reproduisent avec exactitude (ce qui nest
pas toujours le cas dans le recueil de Santos Otero, o les textes grecs
sont parfois fautifs), souvent avec des extraits de lapparat critique. Les
traductions proposes sont prcises, mais il est dommage que les restitutions ne soient pas indiques par des crochets, comme dans les textes



en langue originale. On trouvera ainsi dans cet ouvrage la premire traduction en langue moderne de ldition que J.-D. Kaestli a donne de la
Compilation J (forme Arundel); certains choix sont en outre originaux,
comme le fait de traduire, pour la recension byzantine de lvangile
de Nicodme, les chapitres XIII-XXVII (mme si lomission du rcit
de la crucifixion ne fait pas apparatre le rle jou par Marie dans ce
texte). Pour plusieurs textes, les auteurs citent plusieurs formes (ainsi
les fragments grecs de lvangile selon Thomas, ou pour les Paidika,
les trois chapitres additionnels du ms. dit par Delatte; on peut par
contre regretter labsence de la Pars Altera du Pseudo-Matthieu).
Les notes aux traductions sont peu abondantes il sagit dans la
majorit des cas de rfrences bibliques; les variantes releves ne sont
hlas! pas traduites.
Les introductions prsentent avec prcision les documents, sans taire
les dbats dont ils font lobjet. Les auteurs y luttent juste titre contre
lide que les apocryphes ne contiendraient que des historiettes superficielles (cf. ainsi p. 6-7, 421). Certains points auraient toutefois pu tre
plus arguments (comme p. 510, laffirmation que les Actes de Pierre et
de Paul ont repris la Lettre de Pilate Claude dune source antrieure)
et on notera quelques renvois trop rapides (ainsi p. 5, sur le titre originel de lvangile de lenfance du Pseudo-Thomas) et donnes discutables, comme laffirmation que les bionites dfendaient une christologie anglique (p. 211) ou le rle donn I P 3, 19 dans llaboration
de la troisime partie de lvangile de Nicodme (p. 465); enfin, pour
certains textes, comme lvangile des bionites, il aurait t utile dtre
plus explicite sur la question du titre des uvres.
Les bibliographies sont courtes mais dans lensemble bien informes;
il y est largement fait droit aux recueils publis dans la Bibliothque de
La Pliade, mais rares sont les autres publications francophones cites.
On note nanmoins des absences dautant plus surprenantes quil sagit
douvrages publis en anglais. Ainsi ltude que C. Tuckett a publie
en 2007 chez le mme diteur sur The Gospel of Mary ne figure pas
dans la bibliographie, et le chapitre consacr lvangile selon Thomas
ne mentionne pas les travaux dApril DeConick, notamment son The
Original Gospel of Thomas in Translation. With a Commentary and
New English Translation of the Complete Gospel, Londres New-York,
T&T Clark, 2006.
La prsentation de louvrage est soigne, mais on notera, p. 559, le
fautif Mars Pilati. Labsence dindex (une autre diffrence par rapport
Santos Otero) est regrettable. Enfin, limage choisie pour illustrer la
jaquette est tout fait surprenante, puisquil sagit de la scne, canonique, des Rameaux!
Facult de thologie protestante, Strasbourg




NORELLI, Enr ico (Hgg.), Infancy Gospels: Stories and Identities (Wissenschaftliche Untersuchungen zum Neuen Testament 281), Tbingen, Mohr Siebeck, 2011, XXX+755 p.,
ISBN: 978-3-16-150833-2.
Dieser internationale Band dokumentiert drei Tagungen, die im
Herbst 2010 von den Schweizer Universitten von Freiburg i.., Genf,
Lausanne und Neuenburg veranstaltet wurden. In der anonym verfassten Introduction (XV-XXX) erlutert das Herausgeberteam (XV:
we) bzw. ein Mitglied desselben (vgl. XVIII: I am not sure;
XXIV Anm. 50: I thank) Claire Clivaz vermutlich (vgl. XX Anm.
33; XXIV Anm. 52) , dass Infancy Gospels die franko- und anglophone Begrifflichkeit miteinander verbindet und somit als Sammelbezeichnung fr kanonische und apokryphe Kindheitsgeschichten ber
Jesus verwendet wird. Die Introduction bietet desweiteren englische
Zusammenfassungen der 29, v.a. in Englisch, z.T. aber auch in Franzsisch und Deutsch geschriebenen Beitrge. Das Buch gliedert sich in
drei Teile, deren nichtssagende berschriften hier nicht wiedergegeben
werden brauchen.
Der aus neun Artikeln bestehende erste Teil (1-186) ist dem Kontext
der Kindheitsgeschichten ber Jesus gewidmet und enthlt z.B. Beitrge
zu paganen (3-14) und frhjdischen (15-32) Kindheitsgeschichten. Der
zweite Teil (187-398) setzt sich aus zehn Artikeln zusammen, die sich
v.a. mit den kanonischen Kindheitsgeschichten in Mt 1-2 und Lk 1-2
beschftigen. Die zehn Artikel des dritten Teils (399-640) konzentrieren
sich dann auf apokryphe Kindheitsgeschichten ber Jesus. Die vorliegende Rezension beschrnkt sich im Wesentlichen auf die Besprechung
eben dieses dritten Teils, da es sich bei Apocrypha um eine speziell der
apokryphen Literatur gewidmeten Zeitschrift handelt.
In seinem Beitrag Ways to Survival for the Infancy Apocrypha
(401-417) bietet Sever J. Voicu ein hilfreiches Inventar von apokryphen
Kindheitsgeschichten ber Jesus und untersucht dann v.a. die auch unter
der Bezeichnung Kindheitserzhlung des Thomas bekannten Paidika
sowie das Protevangelium des Jakobus, zwei Schriften, die er als Quellen der spteren apokryphen Kindheitsgeschichten ber Jesus betrachtet.
In Let Us Bring Her Up to the Temple of the Lord: Exploring the
Boundaries of Jewish and Christian Relations through the Presentation
of Mary in the Protevangelium of James (418-432) zeigt Lily Vuong
berzeugend auf, dass das Protevangelium des Jakobus der spezifisch
christlichen Figur der Maria einen hohen Stellenwert gibt und gleichzeitig
dem jdischen Tempel in Jerusalem eine wichtige Bedeutung beimisst.
Vuongs Einschtzung, dass das Protevangelium mit der positiven Bewertung des Tempels im Gegensatz zur brigen frhchristlichen Literatur
steht (424), ist m.E. jedoch zu undifferenziert (vgl. z.B. 2Thess 2,3-4).



Die drei folgenden Beitrge stellen gelungene und innovative Studien zu den Paidika dar: Frdric Amsler, Les Paidika Iesou, un nouveau tmoin de la rencontre entre judasme et christianisme Antioche
au IVe sicle? (433-458); Ursula Ulrike Kaiser, Die sogenannte
,Kindheitserzhlung des Thomas: berlegungen zur Darstellung Jesu
als Kind, deren Intention und Rezeption (459-481); Geert van Oyen,
Rereading the Rewriting of the Biblical Traditions in the Infancy Gospel of Thomas (Paidika) (482-505).
In seinem relativ langen und sorgfltig geschriebenen Artikel Mapping an Unexplored Second Century Apocryphal Gospel: the Liber de
Nativitate Salvatoris (CANT 53) (506-559) prsentiert und untersucht
Jean-Daniel Kaestli das von ihm sog. Liber de Nativitate Salvatoris
(LNS), ein bisher zu wenig erforschtes apokryphes Kindheitsevangelium ber Jesus, welches im 2. Jh. n. Chr. auf Griechisch oder Syrisch
verfasst wurde, jedoch lediglich auf Lateinisch und Irisch berliefert ist.
Kaestli wendet sich dabei v.a. der sog. Hebammenepisode zu und versucht u.a., diese zu rekonstruieren und deren Quellen zu bestimmen. In
Bezug auf letztere nennt er beilufig die alttestamentliche Perikope ber
die hebrischen Hebammen bei der Geburt des Mose (2Mose 1,15-22)
und dann v.a. das im 24. Kapitel der apokryphen Petrusakten zitierte
testimonium ber das Nicht-Hinzukommen einer Hebamme bei der
Geburt Jesu. Es ist jedoch zu fragen, ob es nicht auch denkbar wre,
dass darber hinaus die von Kaestli gar nicht bercksichtigte Gynkologie des im frhen Christentum bekannten und geschtzten Soranos
von Ephesos auf die Darstellung der Hebammen im LNS eingewirkt
haben knnte. Zwischen den von Soranos geforderten Eigenschaften
von (knftigen) Hebammen (vgl. Soranos, Gynkologie I.1-4) und der
Darstellung der Hebammen im LNS lassen sich jedenfalls zahlreiche
Parallelen ausmachen.
In ihrem Artikel prsentiert Valentina Calzolari [l]es rcits apocryphes de lenfance dans la tradition armnienne (560-587). Der fr
Leser(innen) der Zeitschrift Apocrypha wohl interessanteste Gedanke
von Philip Alexanders Beitrag Jesus and his Mother in the Jewish AntiGospel (the Toledot Yeshu) (588-616) findet sich dessen Nachwort: [I]t
may be questioned whether, in an academic context, it is any longer
acceptable to define corpora of Gospels or of New Testament Apocrypha purely on confessional lines. The Toledot Yeshu has as much claim
to the title ,Apocryphal Gospel as the Protevangelium of James or the
Gospel of Thomas (604). In der entsprechenden Funote fhrt Alexander dann sogar noch fort: One wonders whether the category should
not be widened further to include Islamic traditions about Jesus. Daniel
Barbus Beitrag Voltaire and the Toledoth Yeshu: A Response to Philip
Alexander (617-627) wurde bereits andernorts auf Franzsisch verffentlicht (617) und ist nicht als Response zu betrachten, da Alexander und seine Ausfhrungen weder erwhnt, geschweige denn diskutiert
werden. Beschlossen wird der dritte Buchteil dann mit Franois Ros-



sets Artikel ,False and ,True: Infancy and Apocryphal Gospels in the
Century of Voltaire (628-640), der warum auch immer als einziger
Beitrag des Sammelbandes eine Bewertung durch den Herausgeberkreis
erfhrt (XXIX: The Enlightenments spirit and involvement in apocryphal infancy gospels is masterfully analyzed by Franois Rosset).
Nachdem die den apokryphen Kindheitsgeschichten ber Jesus
gewidmeten Artikel nun detaillierter besprochen wurden, sind noch
einige Bemerkungen zum vorliegenden Sammelband als Ganzem angebracht. Gut ist, dass dieser mit Stellen-, Autoren- und Sachregister ausgestattet ist. Als hilfreich erweist sich auch die reichhaltige und fr
Sammelbnde leider nicht selbstverstndliche Bibliografie (641-695).
Die zahlreichen Schaubilder, Tabellen usw. tragen zur ohnehin guten
Lesbarkeit und Verstndlichkeit des Buches bei. Darauf, dass kunsthistorische und islamwissenschaftliche Beitrge leider fehlen, weisen
bereits die Herausgeber hin (XXX). Es ist nicht berraschend, dass ein
755 Seiten umfassender Band nicht frei von Druckfehlern ist. Dennoch
ist kritisch anzumerken, dass die Herausgeber sorgfltiger zu Werke htten gehen knnen. So werden sogar die Namen von beteiligten Autoren
sowie die Titel von Beitrgen z.T. falsch und uneinheitlich geschrieben.
Alles in allem jedoch haben die Herausgeber einen beeindruckenden
und lehrreichen Sammelband vorgelegt, der mit seinem Ansatz, kanonische und apokryphe Kindheitsgeschichten gemeinsam zu erforschen,
tatschlich a new step in research (XVI) darstellt, ber den die knftige Forschung zu befinden hat: It belongs to research to evaluate the
effect of joining the French and English scholarly usage by promoting
the label of Infancy Gospels to designate the entire corpus, in an academic field where scholars give canonical or apocryphal accents to this
terminololgy (XXX). Bereits an dieser Stelle soll der im hier rezensierten Sammelband gewhlte und erfolgreich verwirklichte Ansatz ausdrcklich als positiv und Gewinn bringend bewertet werden.
Fonds de la Recherche Scientifique Flandre (FWO)

SCHRTL, Monika, Nicht das ganze Vol k wil l , dass er

sterbe. Die Pilatusakten als historische Quelle der Sptantike (Apeliotes. Studien zur Kulturgeschichte und Theologie
8), Frankfurt am Main, Peter Lang, 2001, 361 p., ISBN: 9783-631-61043-5.
La thse de Monika Schrtl, soutenue la Freie Universitt Berlin
en 2010, est une tude historique des Acta Pilati (dsormais AcPil).
LA. fonde son argumentation sur la datation des AcPil au IVe sicle et
propose danalyser diffrents lments du texte en les confrontant au
contexte historique. Le texte de rfrence est ldition de Tischendorf
dans les Evangelia apocrypha, dont la traduction en allemand est pro-



pose en annexe (chapitre 7, p. 323-340). Luvre se compose de chapitres de longueur et dimportance ingales. Les trois premiers constituent une introduction aux AcPil: outre lintroduction proprement dite
(chapitre 1, p. 11-16), qui propose un tat de la question, et une prsentation fort rapide de la christianisation de lEmpire et des relations entre
juifs et chrtiens jusquau Ve sicle (chapitre 3, p. 47-64), une analyse formelle (chapitre 2, p. 17-46) joint la prsentation du paratexte
(titres et prologues) un rappel sommaire de la tradition manuscrite,
fond sur lintroduction de Tischendorf. Le chapitre 4 rsume le rcit
des AcPil, chapitre par chapitre, en soulignant des points remarquables
et en reprenant la bibliographie antrieure sans proposer danalyses
nouveaux frais (p. 65-165). Cest seulement au chapitre 5, qui porte
le mme titre que lensemble, Die Acta Pilati als historische Quelle
der Sptantike, que ces lments sont repris pour tre confronts
aux ralits politiques et religieuses de la seconde moiti du IVe sicle
(p. 167-316). Dans sa conclusion (chapitre 7, p. 317-321), lA. se flicite davoir trouv dans les AcPil la confirmation de son hypothse de
dpart sur la datation du texte au IVe sicle et propose den situer la
composition aprs la mort de Julien lApostat. Une bibliographie et un
index (p. 341-361) compltent le volume.
La perspective adopte par lA., qui cherche confronter diffrents
pisodes caractristiques des AcPil avec des realia de lAntiquit tardive, est intressante. Il tait peut-tre un peu htif de se lancer dans
ce travail sans remettre en question le prsuppos fixant la rdaction
des AcPil au IVe sicle. Un autre prsuppos est admis comme tel par
lA.: le texte reflterait le contexte historique de sa composition. Or le
fil narratif des AcPil, au moins dans les onze premiers chapitres, est le
procs de Jsus, rapport par les vangiles canoniques. Sur ce point,
lA. cherche distinguer les lments du Ier sicle sans proposer de
distinction entre la ralit historique et les rcits vangliques et ceux
des AcPil, qui seraient unanimement du IVe sicle. Lorsquelle propose
des rapprochements avec les sicles intermdiaires (les analyses de laccusation de magie porte contre Jsus, p. 267-268, ou ltude du rapprochement entre le Christ mdecin et Asclpios, p. 283-289), elle ne
fait jamais lhypothse dun premier tat du texte qui pourrait pourtant
expliquer la persistance de thmes clairement attests ds le IIe sicle.
Pour lA., seule cette premire partie (chapitres I XI, lexclusion
de IV, 2 et de V), refltant une polmique anti-juive issue dun milieu
paen (p. 171), et rpondant aux actes publis sous Maximin Daa,
serait dater des annes qui suivent le bref rgne de Julien lApostat. La dmonstration de la datation nest gure convaincante. LA. se
fonde pour cela sur trois lments, un rapprochement improbable entre
la mention du veau dor en IX, 2 au sein dun discours de Pilate qui
reprend un ensemble de motifs de lExode et le culte de Mithra, la
menace de destruction du Temple de Jrusalem quelle lie lordre de
reconstruction donn par Julien et abandonn aprs 363, et la racti-



vation du culte des divinits antiques, dont Asclpios, par ce mme

empereur. Le rcit de Nicodme sur Joseph dArimathe (chapitres XII
XVI, lexclusion de XII, 1 et 3), sadressant un lectorat juif quil
chercherait persuader du bien-fond du christianisme, serait du dbut
du Ve sicle (p. 173-175). Le prologue des AcPil serait le tmoin dune
troisime tape de llaboration du texte, consistant en lunification de
ces deux lments par Ananias les informations du paratexte sont
toutes prises au premier degr.
Dans sa conclusion, lA. souligne la ncessit dune nouvelle dition critique des AcPil. Celle-ci est entreprise depuis assez longtemps
par lALAC pour quune collaboratrice des Neutestamentlichen Apokryphen ait pu en entendre parler. Il est bien comprhensible que sa
connaissance du texte repose sur le travail de Tischendorf (de 1853,
alors quest connue la rdition de 1876, cf. p. 310; la seule mention
de la rimpression, Hildesheim 1966, n. 5, est une distraction) et
que les divergences entre les manuscrits ne soient connues que grce
lapparat critique mais, dans un souci de ne pas dpendre des choix
ditoriaux du savant allemand, il aurait pu tre intressant de confronter
de manire plus systmatique la recension grecque A de Tischendorf et
la version copte (lA. nignore pas la traduction latine propose en note
dans les Evangelia apocrypha, cf. p. 316, mais aurait pu tirer profit
des traductions intgrales italienne ou franaise). La version gorgienne
nest pas mentionne (p. 17). Les traditions vernaculaires ne sont voques quen passant, et lA. na pas consult leurs diffrentes ditions
(p. 309).
LA. a constitu une abondante bibliographie, notamment en langue
allemande. Son ouvrage propose un pendant intressant ltude de
R.A. Lipsius, Die Pilatusacten kritisch untersucht (1871; cest la rdition qui est de 1886, cf. p. 352) et fait le point sur cent quarante annes
de recherche. Trois aspects sont donc particulirement regrettables. Tout
dabord, si les sources sont cites en traduction selon, sans doute, les
exigences de la collection, elles sont trop souvent tires des Patrologies grecque et latine (avec parfois des erreurs de numrotation). Pour
la plupart des uvres grecques et latines existant en traduction, lA.
a rarement recours au texte original, et les quelques citations en grec
sont souvent fautives (e.g. p. 30, 31, 76, 93, 161, 205; p. 24, confusion
entre livre et volume pour lHistoire ecclsiastique dEusbe). Les rfrences de seconde main sont frquentes.
Seconde lacune, les tudes. Lintroduction Lvangile de Nicodme de Rmi Gounelle et Zbigniew Izydorczyk (collection de poche
Apocryphe de lALAC, 1997) est abondamment cite par lA.; en
revanche, il ny a pas dentre Izydorczyk dans la bibliographie, ce
qui tmoigne de lignorance du Medieval Gospel of Nicodemus, Tempe
(AZ), 1997, dont les prcisions auraient pourtant utilement complt
lintroduction du livre de poche. Lauteur aurait ainsi eu connaissance
de la Thematic Bibliography of the Acts of Pilate de Gounelle et



Izydorczyk (p. 419-532), complte dans Apocrypha 11 (2000), p. 259292. Ltude sur les enseignes ignore larticle de Jean-Daniel Dubois,
Laffaire des tendards de Pilate dans le premier chapitre des Actes de
Pilate, Studia Patristica 19, Louvain, 1989, p. 351-358, de mme que
ne sont pas mentionns louvrage fondamental de Jean-Pierre Lmonon,
Ponce Pilate, Paris, 1981, 2007, ni, ne serait-ce que pour sen distancer,
Xavier Levieils, La polmique paenne des Actes de Pilate, RHPR
79.3 (1999), p. 291-314, sans parler des travaux sur les manuscrits.
Enfin, dans lanalyse des AcPil, lA. sintresse chaque pisode
dans son ensemble mais pas au dtail (cf. lpisode du cursor p. 77-78,
dont est retenu le personnage mais pas les changes verbaux, ou le
songe de la femme de Pilate p. 83-84, o est tudi le rve mais pas le
philojudasme du personnage). Cette attitude est peut-tre due la
volont dutiliser le texte comme source historique mais se heurte la
prsence de plusieurs moments de composition, reprs cependant.
linverse, des dtails sont sur-interprts: la dmonstration dune allusion au culte de Mithra dans la mention de lpisode du Veau dor fait
un mauvais usage dun passage dphrem (p. 263-266).
Ltude a le mrite de proposer une synthse utile sur les AcPil dans
lAntiquit, de rassembler les recherches sur les realia et de proposer
des hypothses sur les processus de composition. Les multiples soustitres et la rdaction dans un allemand clair en font pour le chercheur
un ouvrage maniable et utile consulter, mais dont les conclusions ne
sont, mon sens, pas reprendre.
Anne-Catherine BAUDOIN
ENS, Paris

HOVHANESSIAN, Vahan S. (ed.), The Canon of the Bible and the

Apocrypha in the Churches of the East (Bible in the Christian
Orthodox Tradition 2), New York, Peter Lang, 2012, VIII+113 p. ,
ISBN: 978-1-4331-1035-1.
Dans ce recueil, le groupe dtudes Bible in the Eastern and Oriental Orthodox Traditions de la Society of Biblical Literature (SBL)
prsente sept confrences anglaises qui staient tenues dans le cadre
des runions internationales et annuelles de la SBL Rome et La
Nouvelle-Orlans en juillet et novembre 2009. Dans lavant-propos,
lditeur Vahan S. Hovhanessian constate: The papers published
in this book represent the latest scholarly findings in the field of
Apocrypha and the New Testament canon from the perspective of the
churches of the East (p. vii).
E.S. Constantinou a contribu au livre avec deux essais. Le premier The Canon of Scripture in the Orthodox Church (p. 1-6) est
trs court et parat sadresser un lectorat peu averti (voir p. ex. la
remarque p. 2: Originally the Bible did not exist as a single volume



but was more like a library of books). Il nest en fait quune introduction au deuxime article, comme en tmoigneraient les derniers mots:
A New Testament canon was needed (p. 6). Dans ce second essai,
Banned from the Lectionary: Excluding the Apocalypse of John from
the Orthodox New Testament Canon (p. 51-61), Constantinou montre
de manire instructive que la canonicit de lApocalypse de Jean, trs
rpandue dans les communauts chrtiennes des IIe et IIIe sicles, a
dabord t accepte, puis de plus en plus rejete par lglise orientale
au dbut du IVe sicle sous linfluence dEusbe de Csare. Il en rsulte
quaujourdhui the lectionary of the Eastern Orthodox Church, fixed
for centuries now, continues to exclude Revelation (p. 61).
Dans The Prayer of Manasses: Orthodox Tradition and Modern
Studies in Dialogue (p. 7-20), D. Alberto Ayuch tudie la dimension
canonique de la Prire de Manass, un document apocryphe issu
dAntioche au Ier sicle avant J.-C. partir de lexgse des 15 versets
de cette prire, Ayuch conclut: Given all this evidence of canonicity
the scholar would pose the question why the prayer of Manasses has
not become a part of the Old Testament canon (p. 19).
Dans son trs intressant Testament of Solomon and Other Pseudepigraphical Material in Ahkm Sulaymn (Judgment of Solomon)
(p. 21-37), S. pl livre la premire tude sur les prtendus Ahkm
Sulaymn, c.--d. Les jugements de Salomon (abrviation: JSol;
pl nexplicite pas pourquoi il traduit Judgment of Solomon, au
singulier, alors que le mot arabe / ahkm est pluriel), un document
arabe que les chercheurs ont considr jusqu prsent comme une traduction arabe du Testament de Salomon (abrviation: TSol). Selon
pln, le JSol na pas encore t examin: [S]o far, no attempt
has been made to study this work (p. 21). La contribution de pln
comble donc une importante lacune et prsente the latest scholarly
findings promis dans lavant-propos. Voici en quels termes pln
dcrit le but et les rsultats de son article: It is the purpose of this
paper [] to show that JSol does not in fact contain a translation of
any known recension of TSol, although it does draw heavily from it.
We will further show that JSol is a compilation of canonical, pseudepigraphical and legendary material relating to King Solomon, some of
it known previously, other not [] Along with bringing attention to
previously unknown manuscript witnesses, we will briefly examine the
structure and the content of this work, analyze its substantial connection to the Testament of Solomon and see what insights JSol can provide into pseudepigrapha and their place in Eastern and Oriental versions of the Bible (p. 21-22).
Dans The Book of Wisdom of Solomon in the Armenian Church
Literature and Liturgy (p. 39-42), A. Tanielian tudie brivement le
rle de la Sagesse de Salomon dans la littrature et la liturgie de
lglise armnienne, avec des rsultats assez vagues : In conclusion,
the Book of Wisdom of Solomon, regardless of its classification as deu-



terocanonical book has been beneficiary in the Armenian Church life

in different aspects (p. 42). Pour un article scientifique, la langue de
Tanielian est trop personnelle et pieuse (voir p. 40 : our Lord Jesus
Christ ) et manque de la sobrit ncessaire : We should be always
grateful to the Early Church Fathers who have preserved and saved this
precious pearl Wisdom of Solomon (p. 42).
Dans son intressant article Visul Maicii Domnului (The Dream
of the Mother of the Lord) : New Testament Romanian Amulet Text
(p. 43-49), N. Roddy prsente le texte dune amulette roumaine qui
sappuie sur la littrature apocryphe et qui raconte un rve de la
mre du Seigneur. Malheureusement ce nest pas Marie, la mre de
Jsus, mais Mrs. Untu (48), la mre dune collgue de Roddy, qui
reoit beaucoup (c.--d. trop) dattention dans cet article : [T]he most
surprising thing to occur during the course of otherwise uneventful
research on the topic came when I mentioned to a Romanian colleague,
Alexandra Untu, that I was working on a Romanian amulet text
called Visul Maicii Domnului, to which she remarked quite matter-offactly that her mother happens to carry a type-set copy of this very text
with her always in her purse (p. 46 ; voir aussi p. 93 note 14).
Sous le titre New Testament Apocrypha and the Armenian Version of the Bible (p. 63-87), V. S. Hovhanessian donne une bonne
introduction au corpus des documents apocryphes qui sont associs au
Nouveau Testament. Lauteur regrette quil ny ait pas beaucoup de
chercheurs qui travaillent sur ce corpus (p. 64 : Unfortunately []
not many scholarly works have been published concerning this body
of Armenian literature ). Il avance une raison majeure pour expliquer
cette ngligence : Obviously, the unfamiliarity of scholars in the West
with the Armenian language remains a major obstacle in the path of
their exploration of this body of literature (p. 66). La List of New
Testament Apocrypha in Armenian (p. 84-86) qui compte 54 uvres est
certainement la contribution la plus utile de larticle de Hovhanessian.
Les notes de tous les articles viennent ensuite (p. 89-105). Le livre
sachve par une bibliographie (p. 107-109) et un index qui est la fois
thmatique, scripturaire et onomastique (p. 111-113).
Au vu des remarques parfois critiques mises ci-dessus, la dclaration gnrale de lditeur [t]he papers published in this book represent
the latest scholarly findings (p. vii) parat exagre. Nanmoins, le
groupe dtudes Bible in the Eastern and Oriental Orthodox Traditions de la SBL livre la communaut des chercheurs travaillant sur
le canon de la Bible et les documents dits apocryphes un petit livre
instructif et intressant.
Fonds de la Recherche Scientifique Flandre (FWO)



BERMEJO RUBIO, Fernando El Evangelio de Judas, Texto bilinge y commentario (Biblioteca de estudios bblicos minor
19), Salamanca, Ediciones Sgueme, 2012, 174 p., ISBN:
Lauteur est connu pour sa thse en philosophie sur le gnosticisme
valentinien, La escisin impossible, Lectura del gnosticismo valentiniano, 1998, et son double volume, dintroduction et de prsentation de
textes manichens (El maniquesmo, Madrid, Trotta, 2008, 299 + 562
p.), co-dit pour le second avec J. Montserrat Torrents. On le connat
aussi pour diverses contributions sur lvangile de Judas. Voici une
nouvelle dition commente de cet vangile apocryphe qui tient compte
de lvolution de la recherche depuis la publication initiale du texte
copte en 2006. Le texte est prsent dans une traduction continue avant
une longue introduction de plus de cinquante pages qui brossent le
tableau des questions habituelles dintroduction ce texte copte ancien.
crite dans un langage clair et visant un large public, cette introduction
est suivie par une dition critique du texte copte, une traduction qui
suit ligne ligne le texte dit, et des notes abondantes qui pointent
les difficults de la traduction ou du commentaire. Une bibliographie
trs jour accompagne ces bonnes pages. Lintroduction situe le texte
de lvangile de Judas dans le cadre du codex Tchacos (ou plutt alMinya, du nom probable du lieu de sa dcouverte, selon J. Monserrat
Torrents) et par rapport lhistoire de sa dcouverte. Si, comme le rappelle F. Bermejo Rubio, le codex lui-mme a t dat par les mthodes
du Carbone 14 entre les annes 220 et 340, loriginal grec de ce texte
date vraisemblablement du milieu du deuxime sicle et sa traduction copte de la fin du IIIe sicle. F. Bermejo Rubio note au passage
des indices de la priode ancienne de traduction des textes grecs en
copte (p. 30-31). Si la version de cet apocryphe ne suit pas le portrait
de Judas tel quil apparat dans les textes du Nouveau Testament, la
figure de Judas et un texte sy rapportant est connu dIrne (Contre les
hrsies I, 31, 1) et de quelques hrsiologues postrieurs. Mais tant
donn la difficult dinterprtation du texte conserv en copte, on peut
se demander si Irne a bien compris lvangile auquel il fait rfrence
ou sil ne sagit pas dun vangile diffrent. F. Bermejo nhsite pas
mettre en valeur autant les spcificits du texte que les points de vue
contradictoires son sujet. Tout en discutant les diverses propositions
de dcoupage du texte en squences, il passe en revue les positions qui
valorisent la figure de Judas ou qui prennent Judas pour le pire des disciples. Les dernires pages de lintroduction montrent avec beaucoup de
rserve comment cet apocryphe participe la critique de la hirarchie
ecclsiastique qui se dveloppe ds la fin du second sicle, et comment
le sens de la mort de Jsus est objet de rflexions thologiques. Parmi
les difficults dinterprtation du texte copte, F. Bermejo Rubio expose
lexemple clbre dune crux: Judas, qualifi de treizime dmon,



peut recevoir plusieurs sortes dinterprtation. Quoiquon pense de

ce texte apocryphe, et des discours de rvlation de Jsus Judas,
il contribue grandement interpeler les commentateurs sur des questions majeures des premiers sicles chrtiens. F. Bermejo Rubio adopte
chaque fois une position rserve, mme si ce texte manifeste des liens
certains avec plusieurs textes de la collection copte de Nag Hammadi,
et des textes que lon qualifie habituellement de sthiens.
Lun des mrites explicite de cette nouvelle dition et traduction du
texte copte, cest de tenir compte des nouveaux fragments rcuprs en
2009; on trouvera p. 75 et 76 la liste des nouvelles lectures proposes
qui ne suivent pas toujours celles de ldition critique de 2007. On
pourra se reporter aux notes en bas de pages de ldition critique et de
la traduction en fin de volume, pour voir comment ce nouveau commentaire permet de se faire une ide de la varit des solutions proposes
par les divers commentateurs. En dialogue avec les uvres majeures
sur ce texte depuis 2006, cest un commentaire plaisant lire qui ne
pourra pas laisser indiffrent. F. Bermejo Rubio apporte une contribution solide et bien documente pour permettre au lecteur de se faire une
opinion sur tel ou tel dtail du texte, mais aussi sur la figure de Judas
qui explique que la mort de Jsus fait partie dun mythe cosmique o
luvre dun dmon peut servir faire avancer le plan divin du salut.
Jean-Daniel DUBOIS

MAH, Jean-Pierre POIRIER, Paul-Hubert SCOPELLO, Madeleine (ed.), Les Textes de Nag Hammadi: histoire des religions et approches contemporaines. Actes du colloque international runi les 11 et 12 dcembre 2008 la fondation
Simone et Cino del Duca et lAcadmie des Inscriptions
et Belles Lettres (Palais de lInstitut de France), Paris, AIBL
Diffusion De Boccard, 2010, 280 p., ISBN: 978-2-87754250-0.
Le volume, suite concrte du colloque international qui sest tenu
en 2008 la fondation Simone et Cino del Duca (premier jour) et
lAcadmie des Inscriptions et Belles Lettres (second jour), se veut un
tmoignage la fois de lintrt de lAcadmie pour les tudes gnostiques et de limportance du volume crits gnostiques, publi sous la
direction de J.-P. Mah et de P.-H. Poirier, Paris en 2007 dans la
collection de la Pliade.
Louvrage souvre par un article sur Hermtisme et judasme: de
la rubrique de la Prire daction de grces aux banquets des thrapeutes, d Marc Philonenko, qui remet sur le devant de la scne
scientifique les liens avec le judasme. Lauteur souhaite rvaluer les
influences juives dans les traits hermtiques et il met en relation la



Prire daction de grces copte avec les thrapeutes de Philon et le

livre VIII des Oracles Sibyllins, pour conclure lorigine probablement
juive de la prire et lide que des thrapeutes se seraient convertis
lhermtisme. Il sagit dune hypothse sduisante, qui mrite dtre
taye, car largumentation se limite juxtaposer des citations. Soulignons galement que lauteur se contente simplement de dire en note
p. 19 que la traduction franaise du livre VIII des Oracles Sibyllins
quil cite (p. 16, 18 et 19) est celle de Jean-Michel Roessli, sans indiquer ni en note ni en bibliographie la rfrence cette traduction, la
premire en langue franaise, qui a paru dans Oracles Sibyllins,
dans Pierre GEOLTRAIN - Jean-Daniel KAESTLI, crits apocryphes chrtiens, vol. II, Paris, 2005, p. 1047-1083.
Suit une srie darticles sur les relations que le chercheur peut
tablir entre les gnostiques et des penseurs ou communauts appartenant une autre tradition culturelle et religieuse. Dans Plotin et les
gnostiques, Luc Brisson revient sur la controverse de Plotin contre
les gnostiques dans le trait 33 et sur les distinctions que lon peut
tablir entre le premier et les seconds. On notera linterprtation que
L. Brisson donne de lthique des gnostiques telle quelle est expose par Plotin: le philosophe accuserait les gnostiques non pas davoir
une vie dvoye, mais de ne laisser aucune place la pratique de la
vertu; quand on connat limportance que Plotin accorde aux vertus, on
comprend la gravit de cette accusation. Les articles suivants visent
montrer limportance du volume Pliade qui permet des savants non
spcialistes des gnostiques davoir un accs plus facile leurs crits
et qui facilite une recherche croise. Ainsi Christian Jambet tudie les
Similitudes gnostiques dans lislam shite et Fabrice Midal soumet un article Gnose et tantra bouddhique indo-tibtain. Les deux
auteurs reconnaissent chacun les limites de leur tude et la difficult
tablir des filiations. Il sagit pour eux avant tout de relever des similitudes formelles que seule une recherche collective ultrieure pourra
approfondir; lune dentre elles est dans les deux cas limportance
de la connaissance. La remarque de Christian Jambet, selon laquelle
ces similitudes stablissent avec les sectes musulmanes les plus favorables lsotrisme est mettre en rapport avec le fait que la plupart des tudes sintressant aux rapports entre gnostiques et islam (et
aussi entre crits hermtiques et islam) concerne lislam shite. Trois
des quatre articles suivants voquent limpact des crits gnostiques sur
la pense de savants europens du XXe sicle. Les articles Derrida et
la gnose de Jean-Franois Matti et Jung et la gnose de Michel
Cazenave tudient la manire dont la pense de ces deux savants a pu
tre influence par ce quils savaient des crits gnostiques. En revanche,
dans Les Nouvelles rvlations gnostiques: une non-rception gunonienne, Jrme Rousse-Lacordaire analyse le compte rendu que fait
Ren Gunon de la dcouverte des codices de Nag Hammadi. Il montre
ainsi comment cet auteur est conscient de lintrt de la dcouverte,



tout en considrant que le contenu des codices ne bouleverse pas ce

que lon savait dj cette poque des gnostiques; do lide dune
non-rception. Les pages dAntoine Faivre sur Le terme et la notion
de gnose dans les courants sotriques occidentaux modernes (essai
de priodisation) sont suggestives, la fois pour le spcialiste des
courants sotriques modernes et pour les antiquistes. En effet, il relve
les diffrents emplois de gnose dans ces courants pour conclure,
avec justesse, quil est difficile de dfinir en soi ce terme (p. 107): le
sens de ce mot nest jamais que celui quune personne dcide de lui
confrer un moment donn. Une conclusion retenir.
Les articles de la seconde journe sont plus spcifiquement consacrs linterprtation de passages ou de thmatiques gnostiques. Lensemble est introduit par un article trs bref mais trs utile d Nicolas
Grimal, De Louqsor Nag Hammadi: il prsente le cadre gographique et culturel des crits de Nag Hammadi un cadre que lon a
trop souvent tendance oublier , voquant des analogies travers le
temps, tout en rendant compte des difficults les faire avec le risque
de lanachronisme.
Deux articles sintressent la manire dont des gnostiques lisent et
interprtent des rcits bibliques, nouvelle pierre dans les dbats sur les
origines juives dau moins certains groupes gnostiques et sur lidentit
chrtienne. Dans Seth et sa race dans la Bible et dans le Livre des
secrets de Jean, Bernard Barc adopte une dmarche intressante: partir dune lecture littrale des rcits bibliques pour voir comment lauteur
du Livre des secrets de Jean aurait pu le lire et pour rendre compte des
rapprochements entre ces rcits et lcrit gnostique. Cela supposerait
une bonne connaissance de lhbreu, do la raffirmation de lorigine
juive de la gnose sthienne et do lhypothse dune origine sadducenne. Une telle hypothse mriterait des recherches ultrieures.
Dans Le mythe gnostique du blasphme de larchonte, Steve
Johnston tudie les diffrentes interprtations gnostiques de Dt 43 et
46, distinguant deux versions gnostiques du blasphme et deux attitudes
face au Dieu crateur des critures juives (dmonisation ou rhabilitation). Cependant, bien au-del dune question exgtique, le mythe du
blasphme pose la question de lidentit chrtienne de certaines communauts, avec la mise en place dune distinction plus ou moins nette
davec les juifs.
Jean-Marie Sevrin reprend une notion dsormais trs connue de
Pierre Hadot, exercice spirituel pour lappliquer, judicieusement
lvangile selon Thomas. Il met ainsi en avant lusage anthropologique et la cohrence de cet crit qui fonctionne, pour le lecteur assidu,
comme un guide vers lillumination et lunit. Louis Painchaud, dans
Le codex Tchacos et les codices de Nag Hammadi, tudie lapport
du premier dans la transmission et la rception des crits gnostiques.
Cet apport parat plus important dans la phase la plus rcente de la
transmission (de la fabrication du codex son enfouissement). Louis



Painchaud propose, contre Johanna Brankaer et Hans-Gebhard Bethge,

de voir dans la fabrication du codex Tchacos la volont de jeter le discrdit sur les douze aptres et sur largument dapostolicit avanc par
de nombreux chrtiens. Einar Thomassen fait le point sur Le valentinisme Nag Hammadi, en montrant lapport des crits gnostiques
de ces codices par rapport Irne: ils seraient plus proches du valentinisme original quIrne, qui tmoignerait dune variante du valentinisme. Avec Les milieux gnostiques: du mythe la ralit sociale,
Madeleine Scopello clt en quelque sorte le volume en faisant le point
sur certaines questions: les auteurs gnostiques, le lien avec la socit,
lhistoire des genres. Elle termine son article par quelques prospectives,
dessinant les lignes de recherche ultrieures. Le mot de la fin revient
Jean-Pierre Mah, qui dans une conclusion, rcapitule lapport de
chaque contribution.
Ce volume intressera assurment un large public de chercheurs, vu
la varit des sujets abords et des spcialits des contributeurs. Il dessine des pistes de recherche pour les annes venir, en particulier pour
ce qui concerne lislam shite o beaucoup reste faire. Un projet
est dailleurs en prparation ce sujet au sein du Laboratoire dtudes
sur les monothismes, UMR 8584, par Mohammad Ali Amir-Moezzi
(EPHE-LEM), Daniel de Smet (LEM) et Maria de Cillis (Institute of
Ismaeli Studies, Londres).

BULL, Christian H. LIED, Liv Ingeborg TURNER, John D.,

Mystery and Secrecy in the Nag Hammadi Collection and
Other Ancient Literature : Ideas and Practices. Studies for
Einar Thomassen at Sixty (Nag Hammadi and Manichean
Studies, 76) Leiden, Brill, 2012, 540 p., ISBN : 978-90-0421207-7.
Composed in honor of Einar Thomassen on his sixtieth birthday,
this anthology takes up two subjectsmystery and secrecywhich
dominate the religious landscape of late antiquity. The twenty-three
contributions (appearing in English and French) focus on these topics
in a variety of late antique contexts : ancient mystery religions, gnosticChristian and Hermetic traditions, monastic literature, as well as ancient
Jewish texts. The editors define mystery as a revealed secret, a divine
truth communicated through divine messengers, rites or other practices. Secrecy, on the other hand, signifies an attempt to exert control
over divine knowledge ; it can be a social phenomenon or a rhetorical strategy (ix). Broad subjects, mystery and secrecy reflect themes
in Thomassens own work, particularly his interest in the connections
between text, myth, and ritual in studies like The Spiritual Seed : The



Church of Valentinians (published in 2005 in the same Brill series,

Nag Hammadi and Manichean Studies, of which Thomassen is an editor) (x). At the same time, they offer rich subjects for his former students (University of Bergen) and noted international colleagues who
contribute to the volume. Divided into three sections, Nag Hammadi
and Related Gnostic Writings, Other Christian Practices, Texts, and
Traditions, and Non-Christian Practices, Text Traditions, and Material
Culture, the anthology offers geographic and chronological coverage
as well as diversity (in terms of how the concepts mystery and secrecy
are applied), over synthesis. Essays also range in size, with some contributions being substantive considerations of a topic, and others offering more succinct treatments and syntheses of earlier research.
It is not surprising that a volume in honor of Thomassen, an expert in
Valentinian gnosticism, would focus on the Nag Hammadi and related
texts. These occupy the first section and most substantive portion of
this anthology where we find contributions by April DeConick, JeanDaniel Dubois and Flavia Ruani, Karen King, Antti Marjanen, Louis
Painchaud, Anne Pasquier, Nils Arne Pederesen, Madeleine Scopello,
and John D. Turner. Essays provide broad analyses of a set of treatises
and corpora. Scopello, for instance, considers secrecy and revelation
in the tractates that comprise the Tchacos Codex ; Marjanen addresses
four Sethian treatises in light of their claim to be secret books, while
Turner offers a more philosophical investigation that explores the links
between hidden and revealed revelation and epistemology in four
Sethian tractates. Others look outside of the Nag Hammadi and related
Coptic codices ; Pedersen surveys the term mysterion in the CopticManichaean materials, while DeConick reads Hippolytus Refutation of
all Heresies for evidence of astrology and ritual ascent among a gnostic group called the Peratics.
We also find essays that focus more narrowly on particular texts or
passages within them. Pasquier, for example, considers the enunciation of the divine name in The Holy Book of the Great Invisible Spirit
(also known as The Gospel of the Egyptians). Pasquier points out that
the seemingly strange repetition of vowels, I E O U, in this treatise
is a reflection on Exodus 3 :14. She suggests that in this Coptic book
the name Jesus is integrated into that of YHWH (125) and would be
uttered during the baptismal rite, confirming a new soteriological status
on those who pronounced it. This conclusion leads her to consider the
potency of the divine name reflected in other late ancient traditions,
including Clement of Alexandria, Hermetic as well as Jewish traditions.
The contribution by Jean-Daniel Dubois and Flavia Ruani explores a
similar topic. Examining a passage from Irenaeus Against the Heresies (1.23.3) on the Valentinians, Dubois and Ruani argue that behind
the heresiologists polemical account can be read hints at the baptismal formulas employed by this group, which included the enunciation
of divine names. A rich consideration of the Semitic origins of these



enunciations, this essay also addresses how they functioned as part of

the baptismal rite in Valentinian circles to secure eternal life and affiliation with the divine for the initiates. Louis Painchaud, alternatively,
attends to a single saying in the Gospel of Philip in which Joseph is
said to have planted a garden (Gos. Phil. 73,8-15). Here he considers
secrecy as part of the rhetorical character of this text. Meaning, concludes Painchaud, is not inherent, but hidden, in The Gospel of Philip,
and drawn out through the process of reading. Thus, the identity of
Joseph the enigmatic gardener would emerge through the interpretative
process, as readers contemplated the association between the tree in the
Garden of Eden, and Christs cross. Perhaps, for the gnostic reader,
Joseph would be identified (as Thomassen himself suggested) as the
Valentinian demiurge (113-114).
The most rewarding essay in this section comes from Karen King.
Her analysis of The Secret Book of John (SRJ) illustrates how versatile the terms mystery and secrecy can be for understanding a texts
rhetorical strategies, and the social and historical functions of them.
Employing the early sociological approach of Georg Simmel (1906)
to secrecy, King demonstrates that secrecy is a fact of social life,
which functions to formulate a second world and lend it plausibility (73-74). Employing secrecy in a variety of ways, SRJ does not
point to the esoteric religiosity thought to mark gnostic writings,
but shares much in common with early Christian writings, which commonly deploy secrecy in myriad ways (81). In SRJ secrecy is a rhetorical strategy that is also a social practice, actualized, King suggests,
through a material artifact (a book) and embodied practice (74
and 83). As such, secrecy promotes the prestige of this group ; it offers
them protection against persecution or ridicule from outsiders (an especially compelling rhetorical strategy in a context where social-political
critique was neither welcome nor tolerated) (77). SRJs deployment
of secrecy consolidated group identity in order, ultimately, to promote
its vision of a second world, that of the true god, available to whoever would hear and recite its vision (83).
The second section of this volume features essays from David
Brakke, Ismo Dunderberg, Dimitris Kyrtatas, Hugo Lundhag, Birger
Pearson, Paul-Hubert Poirier, and Tuomas Rasimus. Contributions
address secrecy and mystery in a variety of Christian materials : the
writings of Paul (Pearson), the Gospel of John (Dunderberg), monastic
literature of the Egyptian desert (Brakke, Kyrtatas, and Lundhag), and
Christian Apocrypha, specifically The Acts of Thomas in their Greek
and Syriac recensions (Poirier). Rasimus broadens the discussion to
include material culture. Exploring earlier scholarly treatments of the
early Christian images of the Ichthys, the fish, Rasimus suggests that
this symbol was not a secret or coded one, intended as a critique
of imperial power. Instead, he posits that it was a polyvalent symbol,
which likely held rich significance and even apotropaic function for



Christians. Ultimately, Rasimus argues that Christians appropriated the

8-spoked wheel as a monogram for ICHTH (Iesous Christos Theos),
which then inspired the use of the acronym Ichthys (342). Having
developed this acronym Christians could proliferate existing Christological fish symbolism (sic, 342), which included rich connections
with the Eucharist and to Jesus miraculous feeding of the multitudes
(328 and 346-347).
Three essays on Egyptian monasticism by Brakke, Lundhag, and
Kyrtatas, make interesting companion pieces in that they demonstrate
that monastics relied variously on mystery to constitute authority and
affirm the goals of ascetic practice and monastic life. The great theorizer
of Egyptian monasticism, Evagrius Ponticus, suggests Brakke, employs
mystery as part of his pedagogical strategy. Mystery, according to
Evagrius, reflects the very fabric of Gods creation in which knowledge
is concealed, discovered through monastic discipline and obtained as a
sign of spiritual progress (218). Shenoute of Atripe, however, warns his
monks that mystery is dangerousthat it should be limited to matters
of faith. Where Evagrius enhances the quest for knowledge, for the
uncovering of deeper and richer mysteries of the divine logoi, Shenoute
admonishes his monks against such intellectual pursuits, reminding his
audience about the constraints of human knowing. Divine mystery is
revealed in scripture (true scripture, correctly interpreted). Those who
attempt to go beyond it, Shenoute warns, will fall prey to heresy and
lose, not gain salvation (281-282). Where Brakke and Lundhag focus
on epistemology and salvation in Egyptian monastic writings, Kyrtatas
considers the practice of living in tombs among the desert fathers.
He examines a range of early monastic literature but focuses primarily on Athanasius Life of Anthony. Considering this practice as a kind
of mystery initiation, Kyrtatas highlights how this environment enabled
confrontation and struggle (temptations with demons and Satan), establishing the expertise of monastic leaders in a new religious horizon
The final section features studies by Jonas Bjrnebye, Jan Bremmer,
Christian Bull, Liv Ingebord Lied, the late Marvin Meyer, Jorgen Podemann Srensen, and Ritso Uro cover various non-Christian literature
and groups : the Hermetic corpus (Bull and Srenson), Jewish Pseudipigrapha (2 Baruch, Lied), mystery cults, notably Mithras (Bjrnebye
and Meyer) and the Eleusinian Mysteries (Bremmer). Uros contribution sits least comfortably in this volume in that it does not take up
the themes of mystery and secrecy explicitly. Yet he offers a concise,
and valuable, review of ritual theory, focusing on cognitive models and
suggests a heuristic approach to categorizing ancient mystery religions.
Bjrnebyes and Lieds contributions nicely complement Kings in part
one of the volume. Bjrnebye addresses the initiatory practices of the
cult of Mithras in fourth-century Rome. In this cult, secrecy operated
together with public exposure, just as communal structure retained



elements of vertical organization (in the use of familial language) and

horizontal, or hierarchal structure (in the use of grades for its members). Bjrnebye draws a conclusion similar to Kings own (though he
does not rely on Simmels work) that secrecy constituted the Mithraic
community and perpetuated it (in the progressive sharing of secrets
that comprised the cults practices) (371-372). Lied, on the other hand,
explores secrecy as a didactic strategy in the Jewish text, 2 Baruch.
Secrecy authenticates the texts vision of the world and of Gods plan
for it. Here she traces a tension in the idea that knowledge of Gods
plan is imperative for all of his people, and the notion that this knowledge is imperfectly distributed among them. The tension, she argues,
supports Baruchs apocalyptic vision, just as it confirms the value of
Torah observance (444-445).
Overall there is much here for the scholar of late antiquity. Essays
can be read as isolated contributions. A reader might naturally draw
points of commonality between them, or select essays on literature of
particular interest to him or her. A familiar theme, as I have elaborated,
is the connection between rhetoric and social practice. Many contributors draw out how mystery and secrecy served as rhetorical strategies
that could be confirmed in a variety of practices and even in material
artifacts. They could be deployed in ways that constituted a groups
identity as well as demarcated social structure and hierarchy within it.
Unfortunately, the absence of an introduction (there is only a short forward) makes the process of reading this 540-page volume unwieldy.
It seems unlikely that most readers will consume the volume cover
to cover, and it may be the editors did not necessarily intend readers to do so. However, the lack of synthesis undercuts some of the
potential value of this anthology, both methodologically and in terms
of the connections between ancient religious communities and texts.
How do we determine, for instance, which application of the term,
mystery or secrecy, is salient to analyzing a particular text, social practice, or material artifact ? What about the merits of an emic versus an
etic approach to these topics ? If mystery and secrecy are ubiquitous
features of ancient rhetoric, practice, and social structure, as this volume clearly illustrates, can lines of contrast and comparison be drawn
among our sources ? Might they indicate competition and controversy
within or between communities ? These intriguing issues are raised
in the volume, but only elaborated by individual authors in terms the
materials that he or she examines. Contributors do not cross-reference
each others essays, and the editors do not draw conclusions or points
of comparison across them. The editors have opted instead to provide
a rich number of discussions from a variety of scholars. In the end,
we encounter a diversity of approaches to mystery and secrecy in the
literature and practices of late antiquity, with some intriguing results.
Concordia University, Montreal



DUNDERBERG, Ismo, Beyond Gnosticism. Myth, Lifestyle, and

Society in the School of Valentinus, New York, Columbia
University Press, 2008, 307 p., ISBN: 978-0-231-14172-7.
Dans cet ouvrage, Ismo Dunderberg propose une tude captivante
sur lcole de Valentin. Il apporte un souffle nouveau dans les tudes
sur Valentin, le fondateur du mouvement, et ses disciples. Il met laccent non pas tant sur la thologie, que sur le mythe en lien avec la
pratique et il cherche tout au long de son tude replacer Valentin et
son cole dans leur contexte historique et culturel, qui est chrtien et,
plus gnralement, philosophique. En effet, les thmatiques discutes
par les Valentiniens sont communes dautres coles de pense, parmi
lesquelles les coles philosophiques. Une telle option mthodologique
et heuristique permet galement lauteur de ne pas se laisser enfermer
dans les catgories orthodoxe , hrsie et gnostique . Il prend
ses distances par rapport aux deux premires (voir en particulier le
chap. II) ; quant la troisime, il invite en revoir lusage, quil souhaite plus restreint, non pas la faon de David Brakke pour qui les
seuls gnostiques sont les sthiens (The Gnostics : Myth, Ritual, and
Diversity in Early Christianity, Cambridge, MA-Londres : Harvard University Press, 2010), mais comme un outil permettant dindiquer une
certaine parent (au-del des diffrences) entre plusieurs traditions.
Dans son introduction, lauteur pose un certain nombre dlments
utiles pour la suite de son propos. Trois points peuvent tre relevs.
Lauteur met en relief limportance du caractre scolaire de la majorit
des valentiniens et le fait que la pense de Valentin na pas toujours
t respecte au sein de lcole. Il propose une liste de ses sources qui
diverge au moins en deux points de celle des autres savants tudiant
lcole valentinienne. Elle inclut en effet le Sur la volont de Methodius ; et les sources hrsiologiques sont mentionnes sous la dnomination sources hostiles (p. 8). En effet, Ismo Dunderberg considre quelles demeurent hostiles mme une fois dbarrasse de la
gangue hrsiologique et quelles ne permettent donc pas daccder
un rcit fidle. Une telle option a des consquences sur la manire
dont le chercheur envisage lapport de ces sources une meilleure
connaissance des valentiniens et sur le degr de dformation dont il
pense que ces sources sont affectes. Enfin, lauteur voque la faon
dont il conoit le mythe et comment il le lie lthique et la pratique,
avec lide dune rsonnance du mythe avec lexprience humaine ;
on retrouve la notion dexercices spirituels de Pierre Hadot.
La premire partie est consacre lenseignement de Valentin et
ses dimensions la fois thologique, pratique et pdagogique. Ismo
Dunderberg tudie plus particulirement les fragments n 4 (une interprtation de Gense avec laccent mis sur limmortalit comme voie de
vie et sur la parrhsia dAdam) et n 8 (une vision positive de lunivers
qui tranche avec les autres fragments). Fidle son option heuristique,



il met constamment linterprtation de Valentin en regard dautres crits

contemporains, quelle que soit la communaut religieuse et culturelle dont
ils relvent. Sont ainsi convoqus, parmi dautres, Philon, lApocryphe
de Jean, les crits hermtiques, des crits philosophiques et Methodius.
Les deuxime et troisime parties sont consacres aux successeurs
de Valentin et lusage pdagogique de leurs crits et plus particulirement des mythes, en lien avec ltude des relations au sein des
communauts valentiniennes et entre celles-ci et les autres communauts, chrtiennes ou non chrtiennes. Ismo Dunderberg tudie en premier
lieu la Lettre Flora de Ptolme et sa stratgie ducative en lien avec
la proposition dun mode de vie chrtien et une prise de distance par
rapport Marcion. Il sintresse ensuite deux mythes, celui de Sophia
considr comme un rcit tiologique des motions et celui du dieu
crateur qui lamne conclure que, contrairement lopinion dominante, le monde nest en fait pas peru comme effrayant. Dans les deux
derniers chapitres de la deuxime partie, lauteur aborde la question de
lthique des valentiniens, en lien avec les passages sur la tripartition
de lhumanit et la division entre bien et mal. Il souligne limportance
de lide de dveloppement moral et du choix quont les hommes,
avant de consacrer plusieurs pages une analyse de lInterprtation de
la gnose dont la description dun conflit au sein de la communaut est
comprendre en fonction de la pratique des coles de pense contemporaines de proposer une instruction diffrents niveaux.
Dans la troisime partie, lauteur montre, partir dune analyse du
Trait tripartite, comment le mythe est utilis pour dcrire la socit
contemporaine, les Grecs et les juifs, et pour expliquer la ncessit de
la souffrance de lglise dans une telle socit.
Dans la conclusion, lauteur ne revient pas sur les rsultats de son
tude, exposs chaque fois dans les conclusions particulires de
chaque chapitre. Dans un premier temps, il prfre exposer ses opinions
sur lsotrisme suppos des valentiniens, quil relativise en le mettant
en relation avec la pratique du secret et llitisme des coles de pense
contemporaine. Dans un second temps, il insiste sur la diversit au sein
de lcole valentinienne.
Louvrage se clt par une bibliographie et deux index (sources et
ct du Spiritual Seed dEinar Thomassen (Leyde, 2006), la lecture de cet ouvrage nous semble galement indispensable pour une
meilleure comprhension des valentiniens et des dbats autour de
cette glise. Laccent mis sur la dimension pdagogique (et sociale)
des crits et sur linsertion dans le monde culturel contemporain nous
semble tre un apport majeur de ce livre, qui sadresse aussi bien aux
spcialistes des gnostiques et des valentiniens quaux spcialistes de
la philosophie.



RASIMUS, Tuomas, Paradise reconsidered in Gnostic Mythmaking. Rethinking Sethianism in Light of the Ophite Evidence
(Nag Hammadi and Manichaean Studies 68), Leyde, Brill,
2009, xx + 356 p., ISBN : 978-90-04-17323-1.
Spcialiste du christianisme ancien et de la philosophie hellnistique (il a contribu ldition de Stoicism in Early Christianity, Grand
Rapids, Baker Academic, 2010 et de The legacy of John. SecondCentury Reception of the Fourth Gospel, Leyde, Brill, 2010), Tuomas
Rasimus a obtenu le grade de docteur dans le cadre dune convention
passe entre les universits dHelsinki (Finlande) et de Laval (Qubec,
Canada). Paradise reconsidered in Gnostic Mythmaking. Rethinking
Sethianism in Light of the Ophite Evidence est louvrage directement
issu de sa thse de doctorat dirige par les professeurs Ismo Dunderberg (Helsinki) et Louis Painchaud (Laval). Le premier est lun des
animateurs du groupe de travail The Nordic Nag Hammadi and Gnosticism Network, tandis que le second est responsable de projet dans
lentreprise de traduction et ddition de la Bibliothque copte de Nag
Hammadi (BCNH). Les travaux de chercheurs tels que Tuomas Rasimus sont les heureux produits dune collaboration transatlantique dynamique et de trs haute qualit.
Louvrage de Tuomas Rasimus souvre par une ncessaire mise
au point sur les termes et les notions qui seront manies par la suite.
Nombreuses ont t les publications sur le gnosticisme et de nouvelles
coles se forment. Ce champ de recherche est extrmement actif et en
constante volution. Au chapitre premier, le propos de T. Rasimus, sa
dfinition du gnosticisme de mme que lexplicitation de lobjet de sa
thse, prend la forme dun tat de la recherche. Lintroduction (prologue et chapitre premier), qui forme en fait la premire des trois parties du livre, est consacre au concept de sthianisme et son corollaire
lophitisme, deux catgories qualifies de problmatiques , car directement issues du discours hrsiologique antique.
T. Rasimus admet nanmoins lexistence dun groupe social ophite
qui aurait possd sa littrature et des pratiques cultuelles. Il en trouve
la trace dans la littrature hrsiologique (Irne, Contre les Hrsies
I, 30 ; Origne, Contre Celse VI, 224-38 ; piphane, Panarion 26) et
les sources dans certains crits coptes de Nag Hammadi (Sur lOrigine
du monde, Hypostase des archontes, Apocryphe de Jean, Eugnoste,
Sagesse de Jsus-Christ).
Lauteur postule qutudier ce corpus ophite permet de revisiter et
de mieux connatre le sthianisme, lune des formes les plus anciennes
de gnosticisme, quil ne faudrait pas totalement confondre avec lophitisme. En effet, la recherche sest longtemps concentre sur le sthianisme tel que la dfini Hans-Martin Schenke, forant des textes entrer
dans le cadre de son Sethian System tout en en ngligeant dautres.
Selon Schenke, les crits sthiens seraient prchrtiens, rdigs par des



juifs se prtendant de la race lue descendant de Seth, pratiquant le

baptme et une forme de contemplation. Leurs textes partageraient tous
une mythologie et un langage communs.
T. Rasimus avance davantage la manire du dialectologue, relevant
les traits caractristiques de trois mythologies, respectivement sthite,
barbliote et ophite. Comme les isoglosses linguistiques, certains traits
sont communs deux ensembles, ce qui implique des zones de superposition. Les trois domaines (mythologies) ne sexcluent pas. Les traits
se combinent diffremment selon les textes qui peuvent ensuite tre
situs sur un nuancier (uniquement ophite, galement barbliote et
ophite ou sthien et barbliote, etc.) illustr par un diagramme de Venn.
Lensemble de la surface recouvre donc davantage que le matriel
sthien au sens troit du terme et permet galement de comprendre
comment un texte tel que lApocryphe de Jean, considr comme larchtype de lcrit sthien, contient des motifs qui nentrent pas dans
le systme de Schenke. Le mythe de Seth ntant plus le foyer unique
autour duquel gravite lensemble du corpus, ce dernier ne peut tre
sthien proprement parler. T. Rasimus introduit alors la notion tout
fait intressante de Classic Gnostic , qui va tre dveloppe dans
la suite de la premire partie et exemplifie partir dcrits de Nag
La seconde partie est consacre aux motifs mythiques et leur
volution. Quatre font lobjet de chapitres particuliers : le serpent, le
dmiurge et ses archontes, Sophia et la race de Seth.
Sagissant du serpent, T. Rasimus commence par une mise au point.
Certes le reptile est au centre de la mythologie ophite, puisquil lui
donne son nom, mais quun texte le mentionne ne suffit pas le
rendre ophite pour autant. Cest pourquoi lauteur dfinit le motif et
son rle dans les tmoignages sur la gnose classique (Apocryphe de
Jean, Hypostase des archontes, Sur lOrigine du monde, Tmoignage
de vrit et la source dIrne), puis largit le champ dautres documents gnostiques (Paraphrase de Sem et la source dHippolyte sur les
Naasnes, Pistis Sophia, vangile de Thomas, etc.) pour finalement traquer les indices pars dans la littrature orientale, ancienne et mdivale, y compris arabe.
Le chapitre consacr au dmiurge ou crateur et ses archontes
sinscrit dans le prolongement thmatique du chapitre prcdent puisque
dans le systme ophite, le serpent peut tre la forme du crateur Ialdabaoth. Justement, T. Rasimus cherche comprendre la cosmologie
gnostique classique de mme que ses origines en se concentrant sur son
expression, cest--dire un crateur et des archontes thriomorphes, par
lonomastique, ainsi que sur la comparaison des formes et attributs des
archontes, ces puissances subalternes. La gnose classique puise largement langlologie juive.
T. Rasimus livre alors une premire analyse du diagramme ophite
dcrit par Origne dans le Contre Celse VI. Il sagit l dune carte



cleste qui doit avoir probablement exist et dont le philosophe Celse

(fin IIe sicle) usait pour discrditer les chrtiens. Origne reprend le
discours de Celse et le rfute point par point dmontrant que le schma
est utilis par les ophites et non par les vrais chrtiens. Cest ainsi
que lon peut reconstituer les arguments de Celse de mme que le diagramme et les savants ne sen sont pas priv. T. Rasimus dispose en
annexe plusieurs tentatives sous forme de schmas, notamment ceux
de Andrew J. Welburn, Bernd Witte, Attilio Mastrocinque ou Alastair
H. B. Logan. lissue de ce chapitre, lauteur voque de possibles
liens avec le judasme du Ier sicle, une piste dveloppe au chapitre
suivant sur Sophia et la connaissance.
Selon T. Rasimus, le systme gnostique classique peut tre rapproch de spculations sapientiales, dont Philon est le meilleur tmoin et
contre lesquelles Paul tente de slever dans sa Premire ptre aux
Corinthiens : Sophia est la pardre fminine de Dieu, mais non du crateur qui, lui, est une figure ngative. Bien au contraire elle uvre pour
que lhomme se libre de lemprise de la matire et atteigne la connaissance (gnosis). Elle est par ailleurs fortement attache au personnage
dve. On comprend ds lors comment le comportement de la premire
femme peut tre rinterprt. Ialdabaoth est le nom du dmiurge chez
les sthiens et correspond au YHWH des juifs. Il retient lme prisonnire de la Cration. En dsobissant son crateur, ve-Sophia guide
lhomme vers la connaissance et donc le salut. Selon la notice dIrne
Sophia est la sur du Christ et tous deux uvrent de concert.
Aprs ve et Sophia, T. Rasimus consacre un chapitre lautre duo :
Adam et le Christ. En 1 Cor 15, 45-47, le Christ est appel nouvel
Adam . Le contexte confirmerait ce qui a t tabli au chapitre prcdent, savoir que Paul se trouverait face des rivaux, juifs comme
lui, mais aux tendances philoniennes. Cest en explorant ce milieu que
T. Rasimus commence le chapitre 5. Il sarrte plus particulirement
sur les variations autour de la cration dAdam telles quelles peuvent
tre apprhendes chez Philon et en 1 Corinthiens. Ces rcritures de
Gense 1-3 sont ensuite compares celles issues de textes ophites.
Pour T. Rasimus la doctrine sthienne a d voluer pour rpondre
ceux qui professent la divinit du Christ et spculent sur le monothisme dans le cadre des controverses trinitaires.
Le sixime chapitre est aussi le dernier de la seconde partie, consacre aux mythes. Lintrt se dplace des motifs typiquement ophites
vers des thmes plus sthiens, et en premier lieu celui de la descendance ou semence de Seth. Le mythe est toujours reli au paradis et
la geste dAdam et ve, mais il ny a plus de serpent. Parmi les
textes coptes, la semence de Seth est mentionne dans lApocryphe
de Jean, Zostrien, les Trois Stles de Seth, Melchisdech, lApocalypse
dAdam ou encore lvangile de Judas. Il existe plusieurs variantes
du mythe sthien, la plus courante tant celle qui justifie la prtention tre de la race de Seth . Seth biblique est le troisime enfant



dAdam et ve, le survivant aprs que la maldiction se soit abattue

sur ses deux ans, Can et Abel. Ses parents vont alors lui confier une
connaissance secrte. Seth va son tour la transmettre, initiant une tradition denseignement sotrique.
T. Rasimus cherche non pas la version premire du mythe, mais
tente de cerner les raisons socio-historiques qui favorisrent la focalisation dun groupe sur la figure de Seth. Lauteur rvle peu peu les
tres humains derrire les textes et postule par consquent lexistence
dune communaut sthienne, avec ses crits et ses rites. Cest ces
derniers quest consacre la troisime partie du volume.
Le premier chapitre de la troisime partie (le huitime en notation
continue) voit la collecte des traces et indices dun culte du serpent
parmi les ophites. Les tmoignages littraires sont gnralement le fait
de leurs dtracteurs, ce qui implique de les pondrer. ces donnes
plus ou moins fiables, T. Rasimus adjoint quelques amulettes leffigie du serpent. Lauteur estime quil sagit toujours du serpent de la
Gense, mme si le serpent tait un animal frquemment reprsent
dans le monde grco-romain. Il se tourne ensuite vers les sources gnostiques de premire main, savoir les crits coptes de Nag Hammadi
ainsi que dautres sources et artefacts pouvant leur tre associs.
La vnration pour le serpent trouve une justification interne selon
lexgse ophite, mais pour un chrtien non gnostique, cela passe pour
une abomination, une inversion des valeurs. Laccusation porte contre
eux de maudire le Christ au cours de leurs rituels initiatiques sinscrit dans le mme discours. Nanmoins il est tout fait probable que,
comme pour le culte du serpent, il y ait un fond de vrit, mais mal
interprt. En effet, selon T. Rasimus, ce nest pas le Sauveur des chrtiens que renient les ophites, mais lhomme Jsus (cf. Second Trait
du grand Seth) ou alors le Christ, mais tel quil est dpeint dans les
vangiles canoniques (cf. vangile de Judas).
Le dernier chapitre du volume porte sur le sceau des ophites et le
baptme sthien. linterne, aucun texte natteste positivement de la
pratique effective du baptme parmi les gnostiques. Ce silence peut
nanmoins sexpliquer par la nature des sources, trs peu descriptives.
Pour traiter du sujet, qui devrait tre le point culminent de ltude
du rituel sthien, il faut se contenter de sources externes. Toutefois il
existe des biais. Selon Celse, les ophites auraient pratiqu une onction
ou sceau , un terme gnralement utilis par les chrtiens pour dsigner le baptme. En complment, de nombreux textes provenant de Nag
Hammadi (Protennoia trimorphe, Apocryphe de Jean, Sur lOrigine du
monde, Hypostase des archontes, etc.) font tat de ce sceau, ce qui
plaiderait en faveur de la vracit des tmoignages externes. T. Rasimus reprend les conclusions de ses prdcesseur (Schenke, Turner) et
rassemble les lments dune pratique liturgique quil rapproche ensuite
du prologue de lvangile de Jean. Il conclut non pas une dpendance
directe, mais au rattachement une mme tradition sapientiale juive,



esquissant dun mme mouvement ce qui pourrait tre la gnalogie du

Lensemble de louvrage est remarquablement clair malgr la complexit de la littrature gnostique, par ailleurs fort bien matrise et abondamment cite. noter toutefois quelques difficults mthodologiques.
Pour caractriser les spculations ophites sur le serpent, T. Rasimus
commence par la notice dIrne (Contre les Hrsies I, 30, 1-10). En
effet, il nexiste pas dautodfinition de la part des auteurs ophites .
Lunique lien entre cette appellation et une matire thologique et littraire est offert par les hrsiologues. Le risque est alors de procder
une lecture circulaire des sources, cest--dire de venir appuyer ce
qucrivent les hrsiologues par des sources primaires lues, quon le
veuille ou non, travers le prisme tendu par ces mmes hrsiologues.
Actuellement, des chercheurs tels que Jean-Daniel Dubois ou Christoph Markschies privilgient une approche par les sources internes, idalement sans le conditionnement impos par les sources extrieures, mais
est-ce vraiment possible ? Et est-ce mme souhaitable ? Car la littrature
hrsiologique est aussi dun apport utile. La meilleure mthode repose
peut-tre dans la mesure et lhumilit, cest--dire viter les conclusions
tranches et dfinitives pour privilgier une lecture nuances et croise
de toutes les sources, mais aussi viter lesprit de systme et admettre
de faire voluer tout moment ses propres conclusions.
T. Rasimus semble presque gn dutiliser la littrature hrsiologique, il sen excuse au prix dune lgre contradiction au dbut du
chapitre IX : This final chapter examines what appears to be the only
reliable piece of information in the heresiological literature concerning
Ophite rituals. Pourquoi croire piphane ou Celse dans ce cas uniquement ? Et de fait, T. Rasimus suit les hrsiologues sur dautres points.
Cette aporie montre la difficult dtudier le gnosticisme et limportance accorde aux dfinitions dans ce champ de recherche. Toutefois,
lhyper-pistmologie dbouche parfois sur une guerre des mots mene
au dtriment de ltude des textes. T. Rasimus dpasse ce stade et ose
une interprtation, quitte se tromper, mais ce ne semble pas tre le cas.
Sur le plan formel, on apprciera les nombreux tableaux et schmas qui apportent quelque clart dans ce matriel dense en donnant
par exemple voir sous forme de tableaux synoptiques les rfrences
de tous les passages tudis et les traits mythiques dont ils tmoignent.
noter encore que plusieurs chapitres ont dj paru sous forme
Charlotte TOUATI
Universit de Lausanne



PETTIPIECE, Timothy, Pentadic Redaction in the Manichaean

Kephalaia (Nag Hammadi & Manichaean Studies 66), Leyde,
Brill, 2009, 242 p., ISBN: 978-90-04-17436-8.
Avec cette recherche sur lusage du nombre cinq dans les Kephalaia
manichens, Timothy Pettipiece livre ici une des rares tudes sur la
numrologie dans le manichisme (et plus gnralement dans les religions antiques). Lanalyse quil propose du nombre cinq repose sur une
nouvelle approche des Kephalaia, expose dans son introduction. Se
dmarquant de lopinion dominante jusquil y a peu parmi les spcialistes, il considre que ces crits ne tmoignent pas dun enseignement
de Mani quasiment inchang; au contraire, ils seraient luvre dun
compilateur qui cherche, dune part, clarifier le message de Mani
(sous-entendant que contrairement ce que la tradition manichenne
laisse entendre Mani ntait peut-tre pas aussi clair) et, dautre part,
rpondre des dfis internes et externes. Lusage plus extensif du
nombre cinq dans les Kephalaia, par rapport lenseignement mme de
Mani qui utilise plutt le nombre trois, serait d au compilateur, dans le
cadre de ce double objectif.
Dans le premier chapitre, trs court (21-24), il expose les aspects
ontologiques de lusage du nombre cinq, en relation avec les deux
arbres du deuxime Kephalaion. Le compilateur reprend la liste des
cinq membres du bon arbre pour lappliquer au mauvais arbre, mais
dans une perspective ontologique diffrente et pour renforcer le message de ce Kephalaion (mettre en garde contre les apparences).
Dans les deux chapitres suivants, plus longs, Timothy Pettipiece dtaille les aspects thologiques du nombre cinq, en sintressant dabord au royaume de la lumire (chap. 2), puis au royaume
des tnbres (chap. 3). Il relve, sous forme de listes, les diffrentes
pentades qui parcourent les Kephalaia: cinq Pres de la Grandeur, cinq
formulations diffrentes de ces cinq Pres dans diffrents Kephalaia,
cinq rois des Tnbres, etc. Dans le dernier chapitre, il relve dautres
emplois du nombre cinq, en relation avec la sotriologie, lthique,
lecclsiologie, la polmique et ltiologie.
Lauteur montre que le compilateur a effectu un effort important
de rcriture pour organiser le matriau scripturaire quil avait sa
disposition en de nombreuses pentades qui concernent aussi bien le
contenu doctrinal (cinq Pres, etc.) que la forme littraire (liste de cinq
lments, schma rhtorique pentadique, etc.). Le compilateur nhsiterait donc pas prendre une certaine distance par rapport aux versions
canoniques de certains mythes, en tendant des listes de deux ou de
trois lments en cinq. Une telle faon de concevoir le travail du compilateur invite revenir sur la notion de canon et sur les liens entre
canon, tradition et transmission du message de Mani dans des
contextes diffrents.



Louvrage se clt par la traduction anglaise de plusieurs Kephalaia,

une bibliographie (notons que sur le dbat autour du canon manichen,
lauteur aurait pu citer larticle de Jean-Daniel Dubois, Un Kephalaion copte sur le canon des critures manichennes [Keph. Berlin
148], dans Pierre Geoltrain ou Comment faire lhistoire des religions?, d. Simon Mimouni Isabelle Ullern-Weit, Turnhout 2006,
p. 215-221) et deux index, lun gnral (qui recense en particulier les
diffrentes pentades prsentes dans les Kephalaia) et lautre des auteurs
On peut juste regretter que lauteur nexplique pas plus les raisons
du choix du chiffre cinq en relation avec les conceptions numrologiques contemporaines; mais il faut dire que les tudes sont rares sur
le sujet. En tout cas, cette analyse relance lintrt pour la numrologie
manichenne en particulier et religieuse en gnral et pour la constitution de canon et de tradition au sein du manichisme.

Greek and Latin Sources on Manichaean Cosmogony and

Ethics, translated by Dr. Greg FOX and John SHELDON, Compiled with introduction and commentary by Prof. Samuel
N. C. LIEU (Corpus Fontium Manichaeorum, series subsidia
VI), Turnhout, Brepols, 2012, 256 p., ISBN: 978-2-50351247-1.
Ce nouveau volume de la collection Corpus Fontium Manichaeorum est une anthologie de textes grecs et latins relatifs la cosmogonie et lthique manichennes. Samuel N. C. Lieu introduit lensemble
en resituant ces textes dans lensemble des sources et en proposant une
prsentation du mythe cosmogonique en mettant laccent sur la croyance
aux deux principes et aux trois temps et avec ses implications thiques
et communautaires. En quelques pages, Samuel N. C. Lieu expose ainsi
les lments principaux des croyances et de lorganisation des communauts manichennes. Il termine par une liste des sources indirectes
et par lhistoire de la dcouverte des sources directes, en relevant la
faible quantit de citations ou de parallles de textes manichens chez
les Pres de lglise.
Le volume est divis en quatre parties. Les deux premires rassemblent les extraits relatifs la cosmogonie et lthique manichenne,
issus des crits manichens dans la premire partie et des sources chrtiennes et non chrtiennes dans la deuxime. Chaque extrait est donn
dans sa langue originale, grecque ou latine ( lexception dun crit
syriaque qui est en fait la traduction dun crit grec), et avec une traduction anglaise, due Greg Fox et John Sheldon.



La troisime partie, due Samuel N. C. Lieu, intitule Commentarium, est en fait plutt une prsentation, pour chaque extrait,
de luvre do il provient et de son auteur, avec une bibliographie;
Samuel N. C. Lieu y adjoint parfois des notes philologiques, des indications qui relvent de lapparat critique et de brefs commentaires. Le
lecteur ne doit donc pas sattendre un commentaire du contenu de
chaque extrait; ce nest pas le propos de la collection.
Dans la quatrime partie sont runis trois index: des termes (grecs,
syriaques, latins), des sujets et des noms propres manichens, des
sources antiques.
En dpit dun prix exorbitant, ce volume constitue un bon instrument de travail pour tout chercheur sintressant la cosmogonie et
lthique manichenne, notamment parce quil a lavantage de rassembler en un volume des textes qui sont connus et accessibles et dautres
qui le sont moins. Un autre volume sur les sources dautres langues est
en prparation et compltera trs utilement celui-ci.

ARAGIONE, Gabrielle NORELLI, Enrico NUVOLONE Flavio G. (d.),
Diognte. Visions chrtiennes face lempire romain. Actes de la
journe dtude du GSEP du 24 novembre 2007 (Cahiers du Groupe
suisse dtudes patristiques 1), Prahins, ditions du Zbre, 2012, 129 p.
ISBN : 978-2-940351-14-5.
ARRUZA, Cinzia, Les Msaventures de la thodice. Plotin, Origne,
Grgoire de Nysse (Nutrix. Studies in Late Antique Medieval and
Renaissance Thought 6), Turnhout, Brepols, 2011, 314 p. ISBN : 9782-503-53422-0.
BACKUS, Irena BTTGEN, Philippe POUDERON, Bernard (dir.),
LArgument hrsiologique, lglise ancienne et les Rformes, XVIe-XVIIe
sicles (Thologie historique 121), Paris, Beauchesne, 2012, 379 p.
ISBN : 978-2-7010-1604-7.
BARC, Bernard FUNK, Wolf-Peter, Le Livre des secrets de Jean.
Recension Brve (NH III, 1 et BG, 2) (Bibliothque copte de Nag
Hammadi, Section Textes 35), Qubec Louvain Paris, Les
presses de lUniversit Laval Peeters, 2012, xii + 392 p. ISBN : 97890-429-2321-8.
BOVON, Franois, Dans latelier de lexgte. Du canon aux apocryphes (Christianismes antiques), Genve, Labor et Fides, 2012, 399 p.
ISBN : 978-2-8309-1451-1.
BROCCA, Nicoletta, Lattanzio, Agostino e la Sibylla maga. Ricerche
sulla fortuna degli Oracula Sibyllina nellOccidente latino (Studi e
Testi TardoAntichi 11), Roma, Herder, 2011, 437 p. ISBN : 978-8889670-65-1.
BULL, Christian LIED, Liv I. TURNER, John D. (d.), Mystery and
Secrecy in the Nag Hammadi Collection and Other Ancient Literature:
Ideas and Pratices (Nag Hammadi and Manichaean Studies 76), Leyde,
Brill, 2011, 540 p. ISBN : 978-9-0042-1207-7.
Enrico (d.), Infancy Gospels: Stories and Identities (Wissenschaftliche
Untersuchungen zum Neuen Testament 281), Tbingen, Mohr Siebeck,
2011, 755 p. ISBN : 978-3-16-150833-2.
FOSTER, Paul, The Gospel of Peter. Introduction, Critical Edition and
Commentary (Texts and Editions for New Testament Study 4), Leiden,
Brill, 2010, 556 p. ISBN : 978-90-04-18094-9.
GAGN, Andr RACINE, Jean-Franois (dir.), En marge du canon.
tudes sur les crits apocryphes juifs et chrtiens (Lcriture de la
Bible 2), Paris, Cerf, 2012, 288 p. ISBN : 978-2-204-09609-6.



HOVHANESSIAN, Vahan S. (d.), The Canon of the Bible and the Apocrypha in the Churches of the East (Bible in the Christian Orthodox Tradition 2), Bern, Peter Lang, 2012, 113 p. ISBN : 978-1-4331-1035-1.
JENOTT, Lance W., The Gospel of Judas. Coptic Text, Translation,
and Historical Interpretaion of the Betrayers Gospel (Studien und
Texte zu Antike und Christentum 64), Tbingen, Mohr-Siebeck, 2011,
256 p. ISBN : 978-3-16-150978-0.
KAMPEN, John, Wisdom Literature (Eerdmans Commentaries on the
Dead Sea Scrolls), Grand Rapids - Cambridge, Eerdmans, 2011, 390 p.
ISBN : 978-0-8028-4384-5.
LAHE, Jaan, Gnosis und Judentum (Nag Hammadi and Manichaean
Studies 75), Leyde, Brill, 2011, 448 p. ISBN : 978-9-0042-0618-2.
MCNAMARA, Martin, Targum and New Testament (Wissenschaftliche
Untersuchungen zu m Neuen Testament 279), Tbingen, Mohr-Siebeck,
2011, 615 p. ISBN : 978-3-16-150836-3.
MORLET, Sbastien (d.), Le Trait de Porphyre contre les chrtiens.
Un sicle de recherches, nouvelles questions. Actes du colloque international organis les 8 et 9 septembre 2009 lUniversit de Paris
IV-Sorbonne (Collection des tudes Augustiniennes - Srie Antiquit
190), Paris, Institut dtudes Augustiniennes, 2011, 462 p. ISBN : 9782-85121-245-0.
ORBE, Antonio, Introduction la thologie des IIe et IIIe sicles, traduction de lespagnol par Joseph M. LPEZ DE CASTRO, revue et complte par Agns BASTIT et Jean-Michel ROESSLI, avec la collaboration
de Bernard JACOB et Pierre MOLINI, avant-propos de Jean-Michel
ROESSLI, liminaire de Mgr Luis F. LADARIA, s.j. (Patrimoines Christianisme), Paris, Cerf, 2012, 2 vol., 1672 p., ISBN: 978-2-204-09575-4 et
PESTHY-SIMON, Monika, Die Theologie der Versuchung im frhen
Christentum (Traditio Christiana 15), Bern, Peter Lang, 2011, 281 p.
ISBN : 978-3-03910-603-5.
SCHFER, Peter MEERSON, Michael DEUTSCH, Yaacov (d.), Toledot
Yeshu(The Life Story of Jesus) Revisited. A Princeton Conference
(Texts and Studies in Ancient Judaism 143), Tbingen, Mohr Siebeck,
2011, 316 p. ISBN : 978-3-16-150948-3.

Les manuscrits sont soumis la rdaction sous la forme de feuilles recto simples
(ou dun fichier PDF) et dun fichier informatique. Les notes sont numrotes en continu
sur lensemble de larticle (chacune delles ne formant quun paragraphe et sachevant
par un point).
Le prnom et le nom de lauteur, ainsi que son rattachement institutionnel, seront
indiqus en tte du manuscrit.
Le contenu de larticle sera prsent dans deux rsums dau maximum 700 signes
chacun, placs en tte de larticle. Le premier sera rdig en anglais, le second en franais. Lauteur indi-quera galement son rattachement institutionnel, sept mots clefs destins au reprage de son article dans les bases bibliographiques ainsi que son adresse
Les rfrences bibliographiques seront indiques dans les notes, sauf ncessit contraire. Lauteur prendra bien soin duniformiser les rfrences dans lensemble de son
article. Les noms dauteurs doivent tre en petite capitale, avec initiale majuscule ; les
intituls de collections et de revues seront indiqus in extenso. Les rfrences obiront
de prfrence au schma suivant :
Pour un article de revue : NOM Prnom, Titre de larticle , Nom de la revue [numro], anne, p. [pagination] ;
Pour un article douvrage collectif : NOM Prnom, Titre de larticle , NOM DE
LDITEUR Prnom (d), Titre de louvrage (Nom de la collection [numro]), lieu
ddition, anne, p. [pagi-nation] ;
Pour un livre : NOM Prnom, Titre de louvrage (Nom de la collection [numro]),
lieu ddition, anne, p. [pagination].
En franais, les capitales doivent tre accentues (, , ) et les sicles prsents en
petites capitales selon le modle suivant : XIIIe s.
Une citation de plus de quatre lignes doit figurer dans un paragraphe en retrait, sans
guille-mets. Les textes anciens seront systmatiquement traduits. Un texte moderne sera
de prfrence cit dans sa langue originale ; on en proposera, le cas chant, une traduction en note.
Les titres et les sous-titres facilitent la lecture. Leur emploi est vivement encourag.
Pour indiquer les citations bibliques, lauteur doit se conformer aux normes les plus
reues dans son pays (pour les articles en franais, utiliser le modle de la TOB). Les
doubles numro-tations doivent tre systmatiquement indiques. Le soulignement doit
tre proscrit.
Dans le cas de mises en pages complexes, lauteur soumettra la rdaction un projet prenant en compte la faible largeur des pages de la revue (9,8 cm despace dcriture). Il se gardera de faire sur son ordinateur un projet labor avant davoir sollicit les conseils
du secrtariat de rdaction.
Si larticle doit tre accompagn dillustrations, lauteur devra les fournir sous
forme numrique (format JPEG, TIFF ou BMP). Il doit effectuer les dmarches
ncessaires pour obtenir les droits de reproduction.
Remise du manuscrit dnitif
Le tirage papier, ou, dfaut, un chier PDF, sera envoy la rdaction.
Il sera accompagn du texte sous forme numrique en format RTF. Les caractres non latins (grec, syriaque, gorgien, hbreu...) doivent tre cods, dans la
mesure du possible, selon la norme Unicode.
Les ventuelles illustrations seront fournies en JPG, TIFF ou BMP.
Lensemble sera envoy la rdaction. Lauteur recevra pour correction les
premires preuves. Un numro de la revue et vingt-cinq tirs--part lui seront
envoys aprs parution.
Tout courrier doit tre adress au secrtariat de la revue
Adresse du secrtariat de la revue :
187 rue Belliard
F-75018 PARIS

Vous aimerez peut-être aussi