Académique Documents
Professionnel Documents
Culture Documents
500
Cantigas
d'Amigo
Rip Cohen
281
226
218
254
536
141
316
520
Bernal de Bonaval
361
Dinis
586
Estevan Coelho
Estevan da Guarda
Estevan Fernandez d Elvas
Estevan Reimondo
Estevan Travanca
207
252
345
178
209
Fernan do Lago
Fernan Fernandez Cogominho
Fernan Figueira de Lemos
Fernan Froiaz
Fernan Gonalvez de Seavra
Fernan Rodriguez de Calheiros
Fernan Velho
Fernand Esquio
522
187
532
271
225
111
275
528
Galisteu Fernandiz
Garcia Soares
Golparro
Gomez Garcia
Gonal Eanes do Vinhal
490
307
501
326
191
537
457
502
523
304
230
180
313
10
136
151
369
163
256
386
399
Lopo
Loureno
482
494
Martin Campina
Martin Codax
Martin de Caldas
Martin de Giinzo
Martin Padrozelos
Meen Rodriguez Tenoiro
Meen Vaasquiz de Folhente
Meendinho
415
513
426
505
473
203
270
311
126
264
354
433
Pae Calvo
Pae de Cana
Pae Gomez Charinho
Pae Soarez de Taveirs
Pedr Amigo de Sevilha
Pedr Eanes Solaz
Pero d Ambroa
Pero d Armea
Pero d Ornelas
Pero da Ponte
Pero de Berdia
Pero de Veer
Pero Garcia Burgals
Pero Gomez Barroso
Pero Gonalvez de Porto Carreiro
Pero Mafaldo
Pero Meogo
Pero Viviaez
471
334
297
123
441
285
470
437
253
290
349
355
134
220
321
534
417
223
Reimon Gonalvez
Rodrig Eanes de Vasconcelos
306
214
341
533
327
342
396
199
Sancho Sanchez
336
Vaasco Gil
Vaasco Perez Pardal
Vaasco Praga de Sandin
Vaasco Rodrigues de Calvelo
150
276
119
309
11
LIST OF POETS
(as presented in this edition [see Introduction, p. 49])
LISTA DE TROVADORES E JOGRAIS
(segundo a ordem nesta edio [ver Introduo, p. 71])
13
14
BIBLIOGRAPHY1
BIBLIOGRAFIA
1. Manuscripts
Manuscritos
1.2 Other manuscript sources of Galego-Portuguese lyric (excluding CSM; vide infra)3
Outros testemunhos manuscritos da lrica galego-portuguesa (excepto CSM; vide infra)
For a review of work between 1800 and 1899, see Michalis CA (1904), vol. I, pp. 1-98. Pellegrini and Marroni (1981) provide
a relatively complete bibliography up until 1977 (with monographic editions until 1980). Since 1987 this has been supplemented
and extended by the fascicules of the Boletn Bibliogrfico de la Asociacin Hispnica de Literatura Medieval. There are now several
relevant websites on the internet. The University of California hosts http://sunsite.berkeley.edu/PhiloBiblon, which is a frequently
updated version of BITAGAP (Bibliografia de Textos Antigos Galegos e Portugueses) Comp. Arthur L-F. Askins, Harvey L.
Sharrer, Aida Fernanda Dias and Martha E. Schaffer. In PhiloBiblon, 1997-. The Centro Ramn Pieiro has www.cirp.es; and
www.spolia.it, Informazione, studi e ricerche sul Medioevo, provides a provisional bibliography of monographic editions.
2
See Introduction, 1.1.
3
This list does not include the two apographs of B 416 / V 27 (a teno between Afonso Sanchez and Vasco Martinz de
Resende) referred to by the sigla M and P (cf. Pellegrini and Marroni 1981: 29 and the references there).
4
Designated S in the Editio Princeps (Sharrer 1991), the Fragmento Sharrer or Pergaminho Sharrer has been referred to as
T (for Torre do Tombo) by Gonalves (1991) and as L by Oliveira (1993 and elsewhere). As happened with the Pergaminho
Vindel, which has been called PV, R and N (cf. Oliveira 1988: 692), the siglum may, unfortunately, vary for a while.
15
CA =
CEM =
LAPA, Manuel Rodrigues (1970). Cantigas descarnho e de mal dizer dos cancioneiros medievais
galego-portugueses. Edio crtica pelo prof. . 2nd ed. Vigo: Editorial Galaxia (1st. ed.
Coimbra, Editorial Galaxia, 1965); with Vocabulrio.
CSM =
METTMANN, Walter (1959-1972). Afonso X, o Sabio. Cantigas de Santa Maria. 4 vols. [Glossrio in vol. IV]. Coimbra: Por ordem da Universidade (rpt. 2 vols. [Glossrio in vol. II]
Vigo: Edins Xerais de Galicia, 1981; 2nd ed.: Alfonso X, el Sabio, Cantigas de Santa Maria, Edicin, introduccin y notas de . 3 vols. Madrid: Clsicos Castlia, 1986-1989
[cf. Parkinson 1988]).
Rep. =
TAVANI, Giuseppe (1967). Repertorio metrico della lirica galego-portoghese. Roma: Edizioni
dellAteneo (Officina Romanica, collana diretta da Aurelio Roncaglia, 7; Sezione di
studi e testi portoghesi e brasiliani a cura di Luciana Stegagno Picchio, 5).
ACCP
AION
AR
BF
BHS
CN
MLR
RBN
RHI
RL
SMV
ZRPh
16
CINTRA, Lus F. Lindley (1973). (Introduo) Cancioneiro portugus da Biblioteca Vaticana (Cd. 4803).
Reproduo fac-similada. Lisboa: Centro de Estudos Filolgicos & Instituto de Alta Cultura.
CINTRA, Lus F. Lindley (1982). (Apresentao) Cancioneiro da Biblioteca Nacional (Colocci-Brancuti) Cd.
10991. Reproduo fac-similada. Lisboa: Imprensa Nacional-Casa da Moeda.
FERREIRA, Manuel Pedro (1986). O Som de Martin Codax. Sobre a dimenso musical da lrica galego-portuguesa
(sculos XII-XIV). Por . Prefcio de Celso Ferreira da Cunha. Lisboa: UNISYS/ Imprensa Nacional-Casa da Moeda.
RAMOS, M. A. & alii (1994). (Apresentao, Estudos e ndices) Fragmento do Nobilirio do Conde Dom
Pedro & Cancioneiro da Ajuda. Edio Fac-similada do cdice existente na Biblioteca da Ajuda. Apresentao,
Estudos e ndices. Lisboa: Edies Tavola Redonda.
2.2 Editions containing the corpus of cantigas d amigo1 (in order of publication; abbreviations given here are
used throughout the apparatus and notes)
2.2 Edies com o corpus das cantigas d amigo (por ordem de publicao; as abreviaturas aqui indicadas so usadas ao longo do aparato e das notas)
MONACI = Monaci, Ernesto (1875a). Il canzoniere portoghese della Biblioteca Vaticana, messo a stampa da ; con
una prefazione con facsimili e con altre illustrazioni. Halle a.S.: Max Niemeyer Editore
(Communicazioni dalle Biblioteche di Roma e da altre biblioteche per lo studio delle lingue e
delle letterature romanze, a cura di Ernesto Monaci, vol. I).
BRAGA = Braga, Theophilo (1878). Cancioneiro portuguez da Vaticana. Edio critica restituida sobre o
texto diplomatico de Halle, acompanhada de um glossario e de uma introduco sobre os
trovadores e cancioneiros portuguezes, por . Lisboa: Imprensa Nacional.
NUNES = Nunes, Jos Joaquim (1926-1928). Cantigas damigo dos trovadores galego-portugueses. Edio crtica
acompanhada de introduo, comentrio, variantes, e glossrio por . 3 vols. Coimbra: Imprensa da
Universidade (Biblioteca de escritores portugueses) (rpt. Lisboa: Centro do Livro Brasileiro, 1973).
MACHADO = Machado, Elza Pacheco & Machado, Jos Pedro (1949-1964). Cancioneiro da Biblioteca
Nacional, Antigo Colocci-Brancuti. Leitura, Comentrios e Glossrio por . 8 vols. Lisboa: Edio
da Revista de Portugal.
ALVAR, Carlos (1986). Las poesas de Pero Garcia dAmbroa, SMV XXXII: 5-112.
ALVAR, Manuel (1965). Las once cantigas de Juan Zorro. Granada: Universidad de Granada, 1969 (1st. ed.
Malaga; text from Cunha 1949).
1
During the revision of this edition there appeared a self-proclaimed pirate version of the whole corpus of Galego
Portuguese Lyric, which lifts the text (though not the critical apparatus) from previous editions of all kinds with no discernible
philological criteria: Lrica Profana Galego-Portuguesa. Corpus completo das cantigas medievais, con estudio biogrfico, anlise retrica e
bibliografa especfica (coord. M. Brea), 2 vols. (Santiago de Compostela, 1996). While apparently convenient, it is marred by
myriad errors of all kinds and should be used only with the utmost caution.
17
AZEVEDO FILHO, Leodegrio A. de (1974). As Cantigas de Pero Meogo (Establecimento crtico dos textos,
anlise literria, glossrio e reproduo fac-similar dos manuscritos). Rio de Janeiro: Edies Gernasa
e Artes Grficas Ltda. (3rd ed. [revised] Santiago de Compostela: Edicins Laiovento, 1995).
BARBIERI, Mario (1999). Le cantigas damigo de Fernam Froyaz, in (Maria Jos de Lancastre, Silvano
Peloso & Ugo Serani, eds.) Miscellanea in Onore di Luciana Stegagno Picchio. E vs, Tgides minhas.
Viareggio-Lucca: Mauro Baroni editore. Pp. 81-99.
BELL, Aubrey F. G (1920). The Eleven Songs of Joan Zorro, MLR XV: 58-64.
________ (1922). The Hill Songs of Pero Moogo, MLR XVII: 258-262.
BELTRAMI, Pietro G. (1978-1979). Pero Viviaez: Poesie damigo e satiriche, SMV XXVI: 107-124.
BLASCO, Pierre (1984). Les chansons de Pero Garcia Burgals, troubadour galicien-portugais du XIIIe sicle. Introduction,
dition critique, notes et glossaire. Paris: Fondation Calouste Gulbenkian, Centre Culturel Portugais.
COHEN, Rip (1991). Reymon Gonalves: Taking Him Back, in Estudos Portugueses. Homenagem a Luciana
Stegagno Picchio. Lisboa: Difel. Pp. 37-55.
________ (1996a). Johan Garcia de Guilhade. 22 Cantigas dAmigo. Edio crtica. Lisboa: Fundao
Calouste Gulbenkian (suppl. to Colquio-Letras 142).
COTARELO Y VALLEDOR, Armando (1934). Cancionero de Payo Gmez Chario, almirante y poeta (Madrid:
Librera General de Victoriano Surez (ed. facs. con prlogo e apndices de Enrique Monteagudo
Romero. Santiago de Compostela, 1984).
CUNHA, Celso Ferreira da (1945). O cancioneiro de Paay Gmez Charinho, Trovador do sculo XIII. Rio de
Janeiro (typescript; rpt. in Cunha 1999).
________ (1949). O cancioneiro de Joan Zorro. Aspectos lingsticos. Texto crtico / Glossrio. Rio de
Janeiro (rpt. in Cunha 1999).
________ (1956). O cancioneiro de Martin Codax. Rio de Janeiro (rpt. in Cunha 1999).
________ (1999). Cancioneiros dos Trovadores do Mar. edio preparada por Elsa Gonalves. Lisboa: Imprensa Nacional-Casa da Moeda (filologia portuguesa, 1) [contains Cunha 1945, 1949, and 1956].
DOMINGUES, Agostinho (1992). Cantigas de Joo Garcia de Guilhade. Subsdios para o seu estudo lingustico
e literrio. Barcelos: Edio da Cmara Municipal de Barcelos.
FERNNDEZ POUSA, Ramon (1959). Cancioneiro gallego de Jon Airas, burgus de Santiago (s. XIIXIII), Compostellanum IV, n. 2: 231-291; and n. 4: 231-291. [Cited from Rodrguez 1980]
FERREIRA, Manuel Pedro (1986). O Som de Martin Codax. Sobre a dimenso musical da lrica galego-portuguesa (sculos XII-XIV). Por . Prefcio de Celso Ferreira da Cunha. Lisboa: UNISYS/ Imprensa Nacional-Casa da Moeda.
FERREIRO, Manuel (1991). A cantiga de Rodrigu Eanes dAlvares. Edio crtica, in Actas. IV Congresso da Associao Hispnica de Literatura Medieval, org. de Aires A. Nascimento e Cristina Almeida
Ribeiro. Lisboa. Cosmos, Vol. III, pp. 319-323.
________ (1992). As cantigas de RodriguEanes de Vasconcelos. Edicin crtica con introducin, notas e
glosrio. Santiago de Compostela: Edicins Laiovento.
FINAZZI-AGR, Ettore (1979). Il canzoniere di Johan Mendiz de Briteyros, edizione critica, con introduzione,
note e glossario. LAquila: Japadre Editore (Romanica Vulgaria, Collezione di testi medievali
diretta da Giuseppe Tavani, 2. [Sezione A. Poesia lirica]).
18
INDINI, Maria Luisa (1979). Bernal de Bonaval. Poesie. Bari: Adriatica Editrice (Biblioteca di Filologia
Romanza diretta da Giuseppe E. Sansone, 32).
JUREZ BLANQUER, Aurora (1988). Cancionero de Pero da Ponte. Granada: Ediciones TAT.
LANCIANI, Giulia (1977). Il canzoniere di Fernan Velho, edizione critica, con introduzione, note e glossario.
LAquila: Japadre Editore (Romanica Vulgaria, Collezione di testi medievali diretta da Giuseppe
Tavani, 1. [Sezione A. Poesia lirica]).
LANG, Henry R. (1894). Das Liederbuch des Knigs Denis von Portugal, zum ersten mal vollstndig
herausgegeben und mit Einleitung, Anmerkungen und Glossar versehen von . Halle a. S.: Max
Niemeyer [rpt. Hildesheim-New York: Georg Olms Verlag, 1972]).
LOPES DE MOURA, Caetano (1847). Cancioneiro dEl Rei D. Diniz, pela primeira vez impresso sobre o
manuscripto da Vaticana, com algumas notas illustrativas, e uma prefao historico-litteraria
pelo Dr. . Paris: J. P. Aillaud.
MACCHI, Giuliano (l966). Le poesie di Roy Martinz do Casal, CN XXVI: 129-157.
MAGNE, Aug. (1931). Um trovador do perodo post-dionisiano. Dom Afonso Sanches [1279-1239],
Revista de Philologia e de Historia I: 58-88.
MAJORANO, Matteo (1979). Il canzoniere di Vasco Perez Pardal. Bari: Adriatica Editrice.
MARRONI, Giovanna (1968). Le poesie di PedrAmigo de Sevilha, AION X: 189-339.
MNDEZ FERRN, X. L. (1966). O cancioneiro de Pero Meogo. Vigo: Galaxia.
MINERVINI, Vincenzo (1974). Le poesie di Ayras Carpancho. Napoli: Pubblicazioni della Sezione Romanza
dell Istituto Universitario Orientale (Testi, vol. VII = AION XVI [1974]: 21-113).
MONACI, Ernesto (1875b). Cantos de ledino tratti dal grande canzoniere portoghese della Biblioteca Vaticana.
Halle a.S.: Tip. dEhrh. Karras. 2nd ed. [cantigas d amigo of Johan Servando].
MORABITO, Maria Teresa (1987). Le due cantigas damigo di Estevan Coelho, RBN s. 2, vol. 2 (1): 7-21.
NOBILING, Oskar (1907a). Die Lieder des Trobadors D. Joan Garcia de Guilhade (13. Jahrhundert) (doct. diss.,
Univ. Bonn 1907) = As Cantigas de D. Joan Garcia de Guilhade / Trovador do Seculo XIII / Edio critica,
com Notas e Introduo. Erlangen: Junge & Sohn (rpt. Romanische Forschungen XXV [1908]: 641-719; Cantigas de D. Joan Garcia de Guilhade / trovador do seculo XIII / escolhidas e annotadas por .
[Erlangen: Junge & Sohn, 1907] has only the first 38 texts).
NUNES, Jos Joaquim (1919-1920). Don Pero Gmez Barroso, trovador portugus do sculo XIII,
Boletn de la Real Academia Gallega XI (1919): 265-268, 321-325 and XII (1920): 7-10.
PAGANI, Walter (1971). Il canzoniere di Estevan da Guarda, SMV XIX: 51-179.
PANUNZIO, Saverio (1967). Pero da Ponte. Poesie. Bari: Adriatica Editrice (Biblioteca di filologia romanza
diretta da Giuseppe E. Sansone, 10) (Galego tr. [by Ramn Mario Paz]: Pero da Ponte. Poesias. Vigo:
Galaxia, 1992).
PICCAT, Marco (1995). Il Canzoniere di Don Vasco Gil. Bari: Adriatica Editrice.
RADULET, Carmen M. (1979). Estevan Fernandez dElvas. Il Canzoniere. Bari: Adriatica Editrice
(Pubblicazioni dei seminari di Portoghese e Brasiliano della Facolt di Lettere dell Universit
di Roma e della Facolt di Lingue dell Universit di Bari Studi e Testi, 2).
REALI, Erilde (1962). Il Canzoniere di Pedro Eanes Solaz, AION IV, 2: 167-195.
19
________ (1964). Le cantigas di Juyo Bolseyro. Napoli: Pubblicazioni della Sezione Romanza
dell Istituto Universitario Orientale (Testi, vol. III = AION VI, 2 [1964]: 237-335).
RODRGUEZ, Jos Luis (1980). El cancionero de Joan Airas de Santiago. Edicin y estudio. Santiago de
Compostela: Universidade de Santiago de Compostela (= Verba, Anuario Galego de Filoloxia, Anexo 12).
SPAGGIARI, Barbara (1980). Il canzoniere di Martim Codax, Studi Medievali, 3 ser. XXI, I: 367-409.
SPINA, Segismundo (1983). As Cantigas de Pero Mafaldo. Rio de Janeiro: Tempo Brasileiro.
STEGAGNO PICCHIO, Luciana (1975). Filtri doggi per testi medievali: Hu~ papagay muy fremoso,
ACCP IX (= Hom. M. Bataillon): 3-41 (Portuguese tr. [by Alberto Pimenta] in Stegagno Picchio
1979: 27-66; French tr. [by Ma Mouniana] in Stegagno Picchio 1982, I: 7-46).
TAVANI, Giuseppe (1960). Spunti narrativi e drammatici nel canzoniere di Joam Nunes Camans,
AION II, 2: 47-70 (Portuguese tr. in Tavani 1988: 244-257).
________ (1964a). Loureno, Poesie e tenzoni, edizione, introduzione e note di . Modena: Societ
Tipografica Editrice Modenese (Istituto di Filologia Romanza dellUniversit di Roma; Collezione
di Testi e Manuali fondata da Giulio Bertoni e diretta da Angelo Monteverdi, 48).
________ (1964b). Le poesie di Ayras Nunez. Edizione critica, con introduzione, note e glossario.
Milano: Ugo-Merendi (Collezione Internazionale di Filologia Romanza, 1) (Galego tr. [by Rosario lvarez Blanco] A Poesa de Airas Nunez. Vigo: Galaxia, 1993).
TORIELLO, Fernanda (1976). Fernand Esquyo. Le poesie. Edizione critica, introduzione, note e glossario
a cura di . Bari: Adriatica Editrice (Pubblicazioni dei seminari di Portoghese e Brasiliano della
Facolt di Lettere dell Universit di Roma e della Facolt di Lingue dell Universit di Bari).
VALLN, Gema (1995). Las Cantigas de Pay Soarez de Taveirs. Estudio histrico y edicin de . Bellaterra
Barcelona: Facultad de Letras, Universidad Autnoma de Barcelona (Publicaciones del
Seminario de Literatura Medieval y Humanstica).
VEZ SANCHEZ, Antnia (1993). Don Gonal Eanes do Vinhal. Reconstruccin histrico-biogrfica y edicin
crtica de su obra (doct. thesis, Univ. Cadiz).
ZILLI, Carmelo (1977). Johan Baveca. Poesie. Bari: Adriatica Editrice (Biblioteca di Filologia Romanza
diretta da Giuseppe E. Sansone, 30).
2.4 Other works (includes: monographic editions not cited in 2.3; works cited in the edition but not listed above;
works considered essential or of special importance, even though not cited in the edition)
2.4 Outros estudos (incluem-se: edies monogrficas que no so indicadas em 2.3; estudos citados na edio, mas no listados em cima;
trabalhos considerados essenciais ou de importncia especial, mesmo que no citados na edio)
ANNICCHIARICO, Annamaria (1974). Per una lettura del canzoniere di Johan Vaasquiz de Talaveyra,
AION XVI: 135-157.
ASENSIO, Eugenio (1970). Potica y realidad en el cancionero peninsular de la Edad Media. 2nd ed. Madrid:
Gredos (1st ed. 1957) (Biblioteca Romnica Hispnica II; Estudios e ensayos, 34).
ASKINS, Arthur L-F. (1991). The Cancioneiro da Bancroft Library (previously, the Cancioneiro de um Grande dHespanha): a copy, ca. 1600, of the Cancioneiro da Vaticana, in Actas. IV Congresso da Associao
Hispnica de Literatura Medieval. vol. I (Sesses Plenrias). Lisboa: Edies Cosmos. Pp. 43-47.
20
BARBIERI, Mario (1980). Le poesie di Roy Paez de Ribela, SMV 27: 7-104.
BELL, Aubrey F. G. (1925). The Oxford Book of Portuguese Verse. Oxford: The Clarendon Press (2nd ed. 1952).
BELL, Aubrey F. G., BOWRA, C. M. & ENTWISTLE, William J. (1947). Da poesia medieval portuguesa. 2nd ed.
ampliada. Lisboa: Edio da revista Ocidente (rpt. Lisboa: 1985).
BELTRAMI, Pietro G. (1974). Pero Viviaez e lamore per udita, SMV XXII: 43-65.
BERTOLUCCI PIZZORUSSO, Valeria (1961). A proposito di una recente edizione di Johan Ayras de Santiago, SMV IX: 71-100.
________ (1962). Le poesie di Martin Soares, SMV XXVI: 9-160 (and separately [Bologna: Libreria
Antiquaria Palmaverde, 1963]; Galego tr. [by Ernesto Xos Gonzlez Seoane]: As Poesias de Martin
Soares. Vigo: Galaxia, 1992.
BREA, Mercedes & GRADN, Pilar Lorenzo (1998). A Cantiga de Amigo. Vigo: Edicins Xerais de Galicia.
CARTER, Henry H. (1941). Cancioneiro da Ajuda. A Diplomatic Edition. New York-London: Modern Language
Association of America/Oxford University Press (rpt. Milwood, N.J., 1975: Klaus Reprint Co.).
CASTRO, Ivo & RAMOS, Maria Ana (1986). Estratgia e tctica da transcrio, in Critique Textuelle Portugaise.
Actes du Colloque. Paris: Fondation Calouste Gulbenkian, Centre Culturel Portugais. Pp. 99-122.
CINTRA, Lus F. Lindley (1951-1990). Crnica Geral de Espanha de 1344. Edio crtica do texto portugus por . Lisboa: Imprensa Nacional-Casa da Moeda (vol. I 1951 [1952; rpt. 1983]; vol. II 1954
[rpt. 1984]; vol. III 1961 [rpt. 1984], vol. IV 1990) (Academia Portuguesa da Histria. Fontes
Narrativas da Histria Portuguesa).
________ (1987). Sobre o mais antigo texto no-literrio portugus: A Notcia de Torto (leitura crtica, data, lugar de redaco e comentrio lingustico), BF XXXI (1986-87): 21-77.
COHEN, Rip (1987). Thirty-two Cantigas damigo of Dom Dinis: Typology of a Portuguese Renunciation.
Madison: Hispanic Seminary of Medieval Studies (Portuguese Series, 1).
________ (1996b). Dana Jurdica. I: A Potica da Sanhuda nas Cantigas dAmigo; II: 22 Cantigas dAmigo de Johan Garcia de Guilhade: Vingana de Uma Sanhuda Virtuosa, ColquioLetras 142: 5-50.
________ & CORRIENTE, Federico (2002). Lelia doura Revisited. La cornica 31.1: 19-40.
CORREIA, ngela (1992). O refram nos cancioneiros galego-portugueses (Escrita e tipologia).
4 vols. (masters thesis, Univ. Clssica de Lisboa [I: Dissertao; II: Transcrio; III Repertrio:
Lista por tipos; IV Repertrio: Lista por autores]).
________ (1993). Sobre a especificidade da cantiga de romaria, RBN s. 2, 8 (2): 7-22.
________ (1995). O sistema das coblas doblas na lrica galego-portuguesa, in (Juan Paredes, ed.)
Medieoevo y Literatura. Actas del V Congreso de la Asociacin Hispnica de Literatura Medieval. Granada:
Universidad de Granada. Pp. 75-90.
COTARELO VALLEDOR, Armando (1933). Los hermanos Eans Mario, poetas gallegos del siglo XIII,
Boletn de la Academia Espaola XX: 5-32.
CUNHA, Celso Ferreira de (1982). Estudos de Versificao Portuguesa (Sculos XIII a XVI). Paris: Fundao Calouste Gulbenkian, Centro Cultural Portugus (Civilizao Portuguesa, VI).
________ (1985). Significncia e Movncia na Poesia Trovadoresca. Questes de Crtica Textual. Rio de Janeiro:
Tempo Brasileiro.
________ (1986). Sobre o texto e a interpretao das Cantigas de Martin Codax, in Critique Textuelle Portugaise. Actes du Colloque. Paris: Fondation Calouste Gulbenkian, Centre Culturel Portugais. Pp. 65-83.
________ (1991). Valor das grafias -eu e -eo do sculo XIII ao sculo XVI, in Estudos Portugueses.
Homenagem a Luciana Stegagno Picchio. Lisboa: Difel. Pp. 913-927.
21
DE LOLLIS, Cesare (1887). Cantigas de amor e de maldizer di Alfonso El Sabio re di Castigla, Studj
di Filologia Romanza II: 31-66.
________ (1925). Dalle cantigas de amor a quelle de amigo, in Homenaje a Menndez Pidal. Madrid:
Librera y Editorial Hernando. Vol. I, pp. 617-626.
DHEUR, Jean-Marie (1973). Nomenclature des troubadours galiciens-portugais (XIIe-XIVe sicles). Table
de concordance de leurs chansonniers, et liste des incipit de leurs compositions, ACCP VII: 17-100.
________ (1974). Sur la tradition manucrite des chansonniers galiciens-portugais (Contribution
la Bibliographie gnrale et au Corpus des Troubadours), ACCP VIII: 3-43.
________ (1975a). LArt de Trouver du Chansonnier Colocci-Brancuti. dition et analyse, ACCP
IX: 321-398.
________ (1975b). Recherches internes sur la lyrique amoureuse des troubadours galiciens-portugais (XIIe-XIVe
sicles). Lige: Publications de la Facult de Philosophie et Lettres de lUniversit de Lige.
_______ (1984). Sur la gnalogie des chansonniers portugais d Ange Colocci, BF XXIX (= Hom.
a Manuel Rodrigues Lapa): II, 23-43.
DIAS , E. (1887). Beitrge zu einer kritischen Ausgabe des vatikanischen portugiesischen
Liederbuches, ZRPh XI: 42-55.
DIEZ, Friedrich (1863). ber die erste portugiesische Kunst- und Hofpoesie. Bonn: Eduard Weber.
DIONSIO, Joo (1992). As cantigas de Fernan Soarez de Quinhones. Edio Crtica com introduo,
notas e glossrio (masters thesis, Univ. Clssica de Lisboa).
________ (1994). Levad, amigo, que dormides as manhanas frias de Nuno Fernandes Torneol, RBN s. 2,
9 (1): 7-22.
DONATI, Cesarina (1979). Pero Goterres, cavaliere e trovatore, in Studi francesi e portoghesi 79 (Romanica
Vulgaria Quaderni). LAquila: Japadre Editore. Pp. 67-92.
DUTTON, Brian (1964). Lelia Doura, Edoy Lelia Doura, An Arabic Refrain in a Thirteenth Century
Galician Poem?, BHS XLI 1/3: 1-9.
FERNNDEZ REI, Francisco (1991). Dialectoloxa da Lingua Galega. Santiago de Compostela: Edicins
Xerais de Galicia (Universitaria / Manuais) (1st ed. 1990).
FERRARI, Anna (1979). Formazione e struttura del canzoniere portoghese della Biblioteca Nazionale
di Lisbona (cod. 10991: Colocci-Brancuti). Premesse codicologiche alla critica del testo (Materiali
e note problematiche), ACCP XIV: 27-142.
________ (1991). Le Chansonnier et son double, in (Madeleine Tyssens, ed.) Lyrique romane
mdivale: la tradition des chansonniers. Actes du Colloque de Lige, 1989. Bibliothque de la Facult de
Philosophie et Lettres de lUniversit de Lige Fascicule CCLVIII, pp. 303-327.
________ (1993a). Parola-rima, in O Cantar dos Trobadores. Santiago de Compostela: Xunta de
Galicia. Pp. 121-136.
_______ (1993b). Cancioneiro da Biblioteca Nacional, in Lanciani & Tavani 1993: 119-123.
________ (1993c). Cancioneiro da Vaticana, in Lanciani & Tavani 1993: 123-126.
FERREIRA, Jos de Azevedo (1987). Afonso X. Foro Real. 2 vols. (Volume 1: Edio e Estudo Lingustico;
Volume II: Glossrio). Lisboa: Instituto Nacional de Investigao Cientfica.
FERREIRA, Manuel Pedro (1991). Relatrio Preliminar sobre o Contedo Musical do Fragmento
Sharrer, in Actas. IV Congresso da Associao Hispnica de Literatura Medieval. Lisboa: Edies Cosmos. Vol. I, pp. 35-42.
________ (1998a). A msica das cantigas galego-portuguesas: balano de duas dcadas de investigao (1977-1997), in (Derek W. Flitter & Patricia O. Baubeta, eds.) Ondas do Mar de Vigo. Actas
do Simposio Internacional sobre a Lrica Medieval Galego-Portuguesa. Birmingham: Seminario de Estudios
22
Galegos, Department of Hispanic Studies, University of Birmingham. Pp. 58-71 (also in Congreso
O Mar das Cantigas. Santiago de Compostela: Xunta de Galicia, 1998. Pp. 235-50).
________ (1998b). Codax Revisitado, Anuario de Estudios Literarios Galegos: 157-168.
________ (2000). Musik und Betonung in cantigas damigo, in (Thomas Cramer, John Greenfield,
Ingrid Kasten and Erwin Koller, eds.) Frauenlieder / Cantigas damigo. Internationale Kolloquien
des Centro de Estudos Humansticos (Universidade do Minho), der Faculdade de Letras (Universidade do Porto) und des Fachbereichs Germanistik (Freie Universitt Berlin), Berlin 6.11.1998
/ Aplia 28.30.3.1999. Stuttgart: S. Hirzel Verlag. Pp. 247-257 [= (forthcoming) Msica e Acentuao nas Cantigas dAmigo].
________ (2001). Afinidades Musicais: As Cantigas de Loor e a lrica profana galego-portuguesa, in
Memria dos Afectos. Homenagem da Cultura Portuguesa a Giuseppe Tavani. Lisboa: Colibri. Pp. 187-205.
________ (2002). Indcios de contactos potico-musicais entre a cultura trovadoresca e a cultura
rabo-andaluza, in Xarajib. Revista do Centro Luso-rabe de Silves 2: 107-113.
________ (forthcoming). The Layout of the Cantigas: A Musicological Overview, in The Galician Review.
FERREIRO FERNNDEZ, Manuel & MARTNEZ PEREIRO, Carlos Paulo (1996). Cantigas d Amigo. Antoloxa.
Vigo: Asociacin Scio-Pedaggica Galega.
FILGUEIRA VALVERDE, Xos (1992). Estudios sobre lrica medieval. Traballos dispersos (1925-1987). Vigo: Editorial Galaxia.
FINAZZI-AGR, Ettore (1976). Le due cantigas di Roy Gomez de Briteyros, Estudos Italianos em Portugal
38-39 (1975-1976): 183-206.
FONTES, Manuel da Costa (ed.) (1987). Romanceiro da Provncia de Trs-os-Montes. 2 vols. Coimbra: Por
ordem da Universidade.
FRENK, Margit (1987). Corpus de la antigua lrica popular hispnica (siglos XV a XVII). Madrid: Editorial
Castlia (Nueva Biblioteca de Erudicin y Crtica) (2nd ed. 1990).
GANGUTIA ELCEGUI, Elvira (1972). Poesa griega de amigo y poesa arbigo-espaola, Emerita XL, 2:
329-396.
GARCA MOUTON, Pilar & MORENO FERNNDEZ, Francisco (1994). El Atlas Lingstico y etnogrfico de Castilla-La Mancha. Materiales fonticos de Ciudad Real y Toledo, in (Pilar Garca Mouton, ed.)
Geolingstica. Trabajos Europeos. Madrid: Consejo Superior de Investigaciones Cientficas (Biblioteca de Filologa Hispnica: 12). Pp. 111-153.
GASSNER, Armin (1933). Zwanzig Lieder des Joan Ayras de Santiago, Revista da Universidade de Coimbra
XI (= Miscelnea de estudos em honra de Carolina Michalis de Vasconcellos): 385-418.
GONALVES, Elsa (1976a). La Tavola Colocciana. Autori Portughesi. Paris: Fundao Calouste Gulbenkian,
Centro Cultural Portugus (= ACCP X [1976]: 387-448).
________ (1976b). [Review of Minervini 1974] CN XXXVI: 121-128.
________ (1983). [Review of A. Ferrari 1979] Romania 104: 403-412.
________ (1986). Pressupostos histricos e geogrficos crtica textual no mbito da lrica medieval
galego-portuguesa: (1) Quel da Ribera; (2) A romaria de San Servando, in Critique Textuelle Portugaise.
Actes du Colloque. Paris: Fondation Calouste Gulbenkian, Centre Culturel Portugais. Pp. 41-53.
________ (1989). Filologia literria e terminologia musical. Martin Codaz esta non acho pontada, in
Miscellanea di studi in onore di Aurelio Roncaglia a cinquant anni dalla sua laurea. Modena: Mucchi
Editore. Pp. 623-635.
________ (1991). Poesia de Rei. Trs notas Dionisinas. Lisboa: Edies Cosmos.
________ (1993a). Ateudas ata a fiinda, in O Cantar dos Trobadores. Santiago de Compostela: Xunta
de Galicia. Pp. 67-186.
23
________ (1993b). Tradio manuscrita da poesia lrica, in Lanciani & Tavani 1993: 627-632.
________ (1993c). Tavola Colocciana, in Lanciani & Tavani 1993: 615-618.
_______ (1993d). RodriguEanes de Vasconcelos, in Lanciani & Tavani 1993: 580-582.
GRZMACHER, W. (1865). Zur galicischen Liederpoesie, in Jahrbuch fr romanische und englische Litteratur.
Band VI: 351-361.
GUERRA, Antnio J. R. (1991). Contributos para a Anlise Material e Paleogrfica do Fragmento
Sharrer, in Actas. IV Congresso da Associao Hispnica de Literatura Medieval. vol. I (Sesses Plenrias). Lisboa: Edies Cosmos. Pp. 31-33.
HUBER, Joseph (1933). Altportugiesisches Elementarbuch. Heidelberg: Carl Winter (Sammlung romanischer
Elementar- und Hndbucher, I, 8) (Port tr. [by Maria Manuela Gouveia Delille] Gramtica do Portugus Antigo. Lisboa: Fundao Calouste Gulbenkian, 1986).
JAKOBSON, Roman (1973). Lettre Haroldo de Campos sur la texture potique de Martin Codax, in
Questions de potique. Paris: Seuil. Pp. 293-298 (Portuguese tr. in Reckert & Macedo [Lisboa n. d.],
pp. 35-47).
JENSEN, Frede (1978). The Earliest Portuguese Lyrics. Odense: Odense University Press.
LANCIANI, Giulia (1974). Ayras Veaz o il trovatore dimezzato, CN XXXIV: 99-115.
_______ & TAVANI, Giuseppe, eds. (1993). Dicionrio da Literatura Medieval Galega e Portuguesa. Lisboa:
Caminho.
LANG, Henry Roseman (1905). Old Portuguese Songs, in Bausteine zur romanischen Philologie. Festschrift
fr Adolfo Mussafia. Halle a. S. Pp. 27-45.
________ (1908). Zum Cancioneiro da Ajuda, ZRPh XXXII: 129-160, 290-311, 385-399.
________ (1934). A repetio de palavras rimantes na fiinda dos trovadores galaico-portugueses, in Miscelnea scientfica e literria dedicada ao doutor J. Leite de Vasconcellos. Coimbra: Imprensa da
Universidade.Vol. I, pp. 27-43.
LAPA, Manuel Rodrigues (1929a). Das Origens da poesia lrica em Portugal na Idade-Mdia. Lisboa: Edio
do Autor.
________ (1929b). O texto das cantigas damigo, in A Lingua Portuguesa I: 12-21, 56-66, 77-88,105-112 [cited here from rpt. in Lapa 1982].
________ (1970). Cantigas descarnho e de mal dizer dos cancioneiros medievais galego-portugueses. Edio crtica pelo
prof. . 2nd ed. Vigo: Editorial Galaxia [1st. ed. Coimbra: Editorial Galaxia, 1965] with Vocabulrio).
________ (1982). Miscelnea de lngua e literatura portugusa medieval. Rio de Janeiro: Instituto Nacional
do Livro / Ministrio da Educao e Cultura: 1965 (Biblioteca Cientfica Brasileira, Coleco de
Filologia, VII) (rpt. Coimbra: Por Ordem da Universidade, 1982).
LEITE DE VASCONCELLOS, Jos (1905). [Review of Lang 1905] RL VIII: 223-225.
LORENZO, Ramn (1977). La traduccion gallega de la Cronica General y de la Cronica de Castilla. Edicin
crtica anotada, con introduccion, ndice onomstico y glosario. Vol. II, Glosario. Orense: Instituto de Estudios Orensanos Padre Feijoo.
MARRONI, Giovanna (1971). Sullentit del canzoniere di Men Rodrigues Tenoiro, in Studi di Filologia
romanza offerti a Silvio Pellegrini. Padova: Liviana. Pp. 267-277.
MARTNEZ PEREIRO, Carlos Paulo (1992). As cantigas de Fernan Paez de Tamalancos. Edicin crtica con
introducin, notas e glosrio. Santiago de Compostela: Edicins Laiovento.
24
MATTOSO, Jos (1984). Joo Soares Coelho e a gesta de Egas Moniz, in J. Mattoso, Portugal Medieval.
Novas Interpretaes. Lisboa: Imprensa Nacional-Casa da Moeda.
METTMANN, Walter (1959-1972). Afonso X, o Sabio. Cantigas de Santa Maria. 4 vols [Glossrio, in vol. 4].
Coimbra: Por ordem da Universidade (rpt. 2 vols. [Glossrio in vol. 2] Vigo: Edicins Xerais
de Galicia, 1981; 2nd ed.: Alfonso X, el Sabio, Cantigas de Santa Maria, Edicin, introduccin y notas
de . 3 vols. Madrid: Clsicos Castlia, 1986-1989 [cf. Parkinson 1988]).
MICHALIS DE VASCONCELLOS, Carolina (1895a). Zum Liederbuch des Knigs Denis von Portugal,
ZRPh XIX: 513-541.
________ (1895b). Fragmentos etymolgicos, RL III: 129-190.
________ (1904). Cancioneiro da Ajuda. Edio critica e commentada por . 2 vols. Halle a. S., Max Niemeyer
(rpt. with Glossrio [= Michalis 1920], Lisboa: Imprensa Nacional-Casa da Moeda, 1990).
________ (1911-1913). Lies de Filologia Portuguesa, segundo as preleces feitas aos cursos de 1911/12 e
de 1912/13. Seguidas das Lies Prticas de Portugus Arcaico. Lisboa: Dinalivro, n.d. (rpt.).
________ (1920). Glossrio do Cancioneiro da Ajuda, RL XXIII: 1-95.
MOLTENI, Enrico (1880). Il canzoniere portoghese Colocci-Brancuti, pubblicato nelle parti che completano il codice vaticano
4803. Halle a. S.: Max Niemeyer (Communicazioni dalle Biblioteche di Roma e da altre biblioteche
per lo studio delle lingue e delle letterature romanze, a cura di Ernesto Monaci, vol. II).
MONACI, Ernesto (1873). Canti antichi portoghesi tratti dal codice vaticano 4803 con traduzione e note a cura
di . Imola: Tip. DIgnazio Galeati e F.
MONTERO, Xess Alonso (1991). Fortuna literaria de Meendio, in Estudos Portugueses. Homenagem a
Luciana Stegagno Picchio. Lisboa: Difel. Pp. 86-109.
MONTERO SANTALHA, Jos-Martinho (1992). A cantiga Dissrom-m hoj, ai amiga, que nom (B 838, V [424]),
de Paai Gmez Charinho ( ca. 1295), e a expressom jogar bem/mal (a algum), Aglia 29: 61-102.
NOBILING, Oskar (1903). Zur Interpretation des Dionysischen Liederbuchs, ZRPh XXVII: 186-192.
________ (1907b). Zu Text und Interpretation des Cancioneiro da Ajuda, Romanische Forschungen
XXIII (= Mlanges Chabaneau): 339-385.
________ (1908). [Review of Michalis 1904] Archiv fr das Studium der neueren Sprachen und Literaturen
LXII, 121, n.s. XXI: 197-208.
NUNES, Jos Joaquim (1959). Crestomatia arcaica. Excertos da literatura portuguesa desde o que de
mais antigo se conhece at ao sculo XVI, acompanhados de introduo gramatical, notas e
glossrio. 5th ed. Lisboa: Livraria Clssica Editora (1st ed. 1906).
________ (1932). Cantigas damor dos trovadores galego-portugueses. Edio crtica acompanhada de introduo, comentrio, variantes, e glossrio por . Coimbra: Imprensa da Universidade (Biblioteca de escritores portugueses) (rpt. Lisboa: Centro do Livro Brasileiro, 1972).
________ (1975). Compndio de gramtica histrica portuguesa (Fontica e morfologia). 8th ed. Lisboa: Livraria Clssica Editora.
OLIVEIRA, Antnio Resende de (1987). A cultura trovadoresca no ocidente peninsular: trovadores e
jograis galegos, Biblos LXIII: 1-22.
________ (1988). Do Cancioneiro da Ajuda ao Livro das Cantigas do Conde D. Pedro. Anlise do
acrescento seco das cantigas de amigo de , RHI 10: 691-751.
________ (1989). A Galiza e a cultura trovadoresca peninsular, RHI 11: 7-36.
________ (1993). A caminho de Galiza. Sobre as primeiras composies em galego-portugus,
in O Cantar dos Trobadores. Santiago de Compostela: Xunta de Galicia. Pp. 249-260 (rpt. in Oliveira 2001b: 65-78).
25
26
27
28
[by Rosario lvarez Blanco e Henrique Monteagudo]: A poesia lrica galego-portuguesa. Vigo: Editorial
Galaxia, 1986; Portuguese tr. [by Isabel Tom and Emdio Ferreira]: A Poesia Lrica Galego-Portuguesa. Lisboa: Comunicao, 1990).
________ (1988a). Ensaios Portugueses. Filologia e Lingustica. Lisboa: Imprensa Nacional-Casa da Moeda.
________ (1988b). Proposta para unha nova lectura da cantiga de Mendinho, Grial XXVI: 59-61.
________ (1999). Arte de Trovar do Cancioneiro da Biblioteca Nacional de Lisboa. Introduo, Edio crtica
e Fac-smile. Lisboa: Edies Colibri.
VARNHAGEN, F. A. de (1872). Cancioneirinho de trovas antigas colligidas de um grande cancioneiro da Bibliotheca
do Vaticano, precedido de uma noticia critica do mesmo grande cancioneiro, com a lista de
todos os trovadores que comprehende, pela maior parte portuguezes e gallegos. Vienna:
[Typographia I. E. R. do E. E. da Corte] 1870; 2nd ed. Vienna, 1872.
VATTERONE, Sergio (1988). Appunti sul testo delle due Cantigas di Caldeyron, SMV XXXIV: 155-181.
WEISS, Julian (1988). Lyric Sequences in the Cantigas damigo, BHS LXV: 21-37.
WILLIAMS, Edwin B. (1962). From Latin to Portuguese. 2nd ed. Philadelphia: University of Pennsylvania
Press (1st ed. Philadelphia, 1938).
29
Abbreviationes et index fontium composuit C. van de Kieft, adiuvante G. S. M. M. LakeSchoonebeek. Leiden-New York-Kln: E. J. Brill, 1993 (1st ed. 1976).
OED = The Compact Edition of the Oxford English Dictionary. Complete text reproduced micrographically.
2 vols. Oxford: Clarendon Press, 1971.
OLD = Oxford Latin Dictionary. ed. P. G. W. Glare. Oxford: Clarendon Press, 1983.
REW = Romanisches etymologisches Wrterbuch von W. Meyer-Lbke. Heidelberg: Carl Winter
Universittsverlag, 1992 [1934]).
30
INTRODUO
O corpus das 500 cantigas d amigo, compostas entre 1220 e 13501, por um total de 88 poetas,
o maior corpus de poesia amorosa de voz feminina que sobreviveu da Europa medieval e antiga.
Oferece um campo ainda pouco explorado para o estudo da voz feminina, ou seja, do discurso, do
direito, da sexualidade, da mentalidade (por muito que essa voz possa ser manipulada, os aspectos arcaicos destes poemas, a nvel social, lingustico e musical, sugerem que essa voz genuna
nas suas origens).2 Alm disso, muitos estudiosos defenderam que entre as cantigas d amigo h
sequncias organizadas para execuo. Se elas existem e acredito que possa ser demonstrado
que sim estas seriam as primeiras sequncias de cantigas amorosas em qualquer lngua verncula
na histria da literatura europeia, providenciando uma oportunidade nica para estudar a evoluo de uma forma de arte lrico-dramtica a partir das suas partes constituintes, algo com que
Aristteles aparentemente apenas podia sonhar.
Assim, lemos cantigas d amigo no apenas porque as achamos belas, musicais, engenhosas,
erticas, bem delineadas, mas porque so a fonte principal para um captulo ainda por escrever
na histria da cultura europeia. Mas isso ser assunto para uma outra ocasio.3 Aqui direi, to
brevemente quanto possvel, quais as suas fontes, quando e como foram editadas at hoje e
como as tratei nesta edio.
Nenhuma cantiga d amigo pode ser datada com segurana antes de 1220 e a maior parte dos poetas em relao a quem
temos datas confiveis estavam activos entre c. 1225 e c. 1285. A morte de D. Dinis, em 1325, pe um terminus ante quem nas
suas composies, mas estas provavelmente datam de muito mais cedo. Assim, podemos dizer que praticamente todo o
corpus foi composto entre c. 1220 e c. 1300 (Oliveira 1998b).
2
Ferreira (1986; 2000: 249) salienta aspectos musicais arcaicos; alguns desses aspectos sociais e formais j tinham sido
apontados por Lang (1894: lxiii-ciii). Apesar das atribuies e do papel dos homens como compositores, executantes e
escribas, estes no podem ser donos de uma tradio potica que foi passando de mes para filhos durante sculos. Alguns
desses filhos presumivelmente foram estes autores.
3
Algumas das minhas opinies sobre a potica das cantigas d amigo foram j abordadas em Cohen 1996b. O trabalho que
tenho desenvolvido sobre este assunto, para fazer sentido, tem que esperar pela publicao desta edio.
53
O humanista italiano Angelo Colocci (1474-1549) mandou copiar estes dois manuscritos, certamente
depois de 1504, possivelmente cerca de 1525-1526 (Ferrari 1979, Gonalves 1986: 41-45). Anna Ferrari
(1979) analisou B, mas continua por realizar [2001] um estudo codicolgico de V (ver Ferrari 1993c).
C = Autori portughesi, in cod. lat. 3217, Biblioteca Apostolica Vaticana, um index que
Colocci fez de B. Chamado por essa razo La Tavola Colocciana, transcrito primeiro por Ernesto
Monaci (1875a), foi objecto de um estudo exaustivo por Elsa Gonalves (1976a). A Tavola
importante, no por fixar os textos (no possui nenhum), mas como evidncia para a atribuio e soletrao de nomes, especialmente para os nomes dos poetas cujo trabalho
desapareceu, e o nmero de poemas a eles atribudo, em partes da tradio manuscrita
para as quais o nosso nico testemunho.
H tambm uma cpia de V feita entre 1592 e 1612, que reapareceu em 1983 e foi adquirida
pela Universidade da Califrnia, Berkeley, onde est depositada na Bancroft Library. Este manuscrito, como Arthur Askins (1991) demonstrou, um descriptus parcial e imperfeito de V. Sem autoridade independente, no citado nesta edio.1
Nem V nem B contm o corpus inteiro das cantigas d amigo: em cada manuscrito faltam alguns
textos que se encontram no outro.
Sete cantigas d amigo encontram-se apenas em B:
B 47 no f.15 v (Fernan Figueira de Lemos);
B 265 no f.68v (Roi Queimado);
B 332 no f.77r (Rodrig Eanes Redondo);
B 368 no f.84r (Rodrig Eanes de Vasconcelos);
B 373 no f.85r (Pero Mafaldo);
B 383 no f.86r (Pero Mafaldo);
B 456 no f.101r (El Rey Don Affonso de Leon).
Para nove cantigas d amigo e grande parte de uma pastorela, V o nosso nico testemunho:2
V 554 no f.87v (uma pastorela de Johan Airas; B 967, f.209r, termina no v. 5 Apertandosse);
V 590, vv. 4-14 em f.94r (Roi Marti~iz d Ulveira; B 1001 f.216v termina no v. 3);
V 591 no f.94r (Roi Marti~iz d Ulveira);
V 594 no f.94v (Johan Airas);
V 595 nos ff.94v-95r (Johan Airas);
V 596 no f.95r (Johan Airas);
V 597 no f.95r (Johan Airas);
V 598 no f.95r-v (Johan Airas);
V 599 no f.95v (Johan Airas);
V 600 no f.95v (Johan Airas).
1
Contudo, pode oferecer algumas leituras de V melhores do que as que estavam diposio de Monaci, uma vez que a
tinta corrosiva usada ao copiar V danificou o manuscrito. A descrio de Askins indica que improvvel que lance mais do
que uma luz apenas superficial em problemas textuais especficos. As cpias que observei confirmam-no.
2
Dois flios foram arrancados de B entre 216v e 217r. O texto termina depois do v. 3 de B 1001 (= V 590) e recomea com
o v. 3 de B 1012 (= V 602).
54
Muitas letras e palavras, algumas frases e versos e at algumas estrofes inteiras que esto presentes num manuscrito faltam no outro. Todas essas omisses so referidas no aparato crtico.
Os sete poemas de Martin Codax so tambm transmitidos (e na mesma ordem) pelo Pergaminho Vindel, ms. 979, Pierpont Morgan Library, Nova Iorque, siglum N. Descoberto em 1913 e sendo
provavelmente um manuscrito de fins do sculo XIII ou princpios do XIV, o famoso pergaminho
um documento nico na filologia romnica no esteve acessvel aos estudiosos desde 1922
at ao seu reaparecimento, de forma algo dramtica, meio sculo depois (Ferreira 1986: 61, n. 1;
cf. Pellegrini & Marroni 1982: 25, n. 4). Foi adquirido em 1977 pela Morgan Library. Manuel Pedro
Ferreira publicou um estudo de N em 1986 e a sua anlise, recentemente actualizada, fundamental para o entendimento da msica.1
Assim, as nossas fontes para as cantigas d amigo so B e V, com a ajuda inestimvel de N no
caso de Martin Codax. O stemma codicum tem sido disputado, mas parece agora claro que tanto B
como V so cpias de um nico exemplar (ver Gonalves 1993b).
Chegar a esta concluso no foi um processo rpido, mesmo depois da edio da Tavola Colocciana
por Elsa Gonaves (1976) e da anlise codicolgica de B por Anna Ferrari (1979). Tanto estas estudiosas (Gonalves 1983: 412 e Ferrari 1991: 322) como Jean Marie DHeur (1984: 23-43) rejeitaram depois
o stemma de Tavani e substituram-no por um stemma mais simples que irmana os manuscritos B e V.2
Entretanto, a estrutura interna da tradio manuscrita, a cronologia e geografia dos poetas e o
seu contexto social e poltico foram investigados e analisados num livro importante (1994; ver
tambm 1995 [e agora 2001]) e numa srie de artigos (1987, 1988, 1989, 1993, 1995, 1997a, 1997b,
1998a, 1998b [ver agora 2001a]) por Antnio Resende de Oliveira.
1.2 Edies
Ernesto Monaci publicou uma transcrio paleogrfica de V em 1875 (Monaci 1875a).
Usando esse texto, excelente mas no sem erros, Teophilo Braga produziu uma edio crtica de
V, muito fraca, em 1878. Embora os seus erros e defeitos desafiem qualquer descrio (e diz-se terem
enraivecido Monaci), no totalmente desmerecedora (como mostra o aparato desta edio).
A transcrio de Enrico Molteni das partes de B que no esto em V (Molteni 1880) contm
apenas sete cantigas d amigo (as listadas acima como exclusivas de B). At agora no surgiu
nenhuma transcrio paleogrfica completa de B.3
1
Antes denominado PV ou R (por Pellegrini e outros, na suposio incorrecta de se tratar de um rtulo [Ferreira diz que
uma folha volante morfologicamente distinta, mas funcionalmente equivalente aos rtulos mencionados em BV]), o
justamente famoso Pergaminho Vindel ir ser aqui referido como N (por Nova Iorque, a sua actual localizao) por razes
prticas, para evitar confuses com V. No que diz respeito relao entre a tradio representada por BV e a representada
por N, ver Cunha 1986: 65-83 (argumentando contra Spaggiari 1980; ver tambm o prefcio de Cunha a Ferreira 1986). (Ver
agora Ferreira 1998b, com uma importante crtica relativamente ao seu trabalho anterior.)
2
O trabalho sobre o stemma comea com Monaci 1875a, e novamente Monaci em Molteni 1880, seguido por Michalis CA,
vol. II, pp. 210-226, 286-288, at ao trabalho de Tavani e DHeur. Tavani nunca encurtou o seu volumoso stemma, apesar dos
protestos de DHeur. Para o stemma e descrio dos manuscritos ver Gonalves 1993b, 1993c e Ferrari 1993b, 1993c. Para o
Cancioneiro da Ajuda, ver Ramos 1993, 1994.
3
J. Ruggieri 1927, a nica comparao disponvel das variantes de grande parte de B no transcritas por Molteni, foi
condotto su la copia fotografica del codice posseduta dalla Biblioteca Nazionale di Roma (1927: 460) e provavelmente
por essa razo que padece de tantos erros.
55
Elza Pacheco Machado e Jos Pedro Machado publicaram, entre 1949 e 1964, uma edio semidiplomtica, semicrtica, de B. Embora justamente criticada por graves defeitos (Cintra 1983) e ridicularizada desde o incio (por exemplo, por Sansone 1954) como bizarra e hbrida, oscilando entre
uma transcrio e uma edio crtica1, de longe em longe a nica com a leitura correcta.2 Tem sido
dito, de forma cruel, que o seu maior contributo foi a publicao, pela primeira vez, de fotografias do
manuscrito. Quaisquer que sejam os seus mritos e defeitos, no uma transcrio de B.
Portanto, a edio de V feita por Monaci uma transcrio paleogrfica; a edio dos Machados de B no nem uma transcrio nem uma edio crtica. A edio crtica de Braga agora de
interesse sobretudo para os estudiosos da histria do texto. Deve ser sublinhado que Braga no
esperou para ver as leituras de B, o que lhe teria permitido corrigir centenas de erros em V; em vez
disso apressou-se a publicar o seu livro no mesmo ano em que B foi descoberto (1878).
O nico texto de (quase) todas as cantigas de amigo que se autoproclama de edio crtica
baseada em B e V e N para Codax o de Jos Joaquim Nunes: Cantigas damigo dos trovadores galego-portugueses. Edio crtica acompanhada de introduo, comentrio, variantes, e glossrio por . 3 vols. (Coimbra:
Imprensa da Universidade, 1926-1928). Bastante citada, demasiadas vezes de pouca confiana,
eivada de centenas de erros tanto no texto (vol. II) como no aparato (Variantes, no vol. III, pp. 441-573).3 Apesar de tudo representou, falando de modo geral, um avano no seu tempo.4
Nunes no inclui uma cantiga d amigo (B 368 [=Rodrig Eanes de Vasconcelos-4], embora
Tavani diga [1986: 321] que este poema pode ser interpretado como uma cantiga d amor homossexual ou como um dilogo entre um homem e uma mulher). Omite tambm Guilhade-9 (B [748bis]
/ V 351), que pertence s cantigas d amigo daquele poeta, mesmo que no seja em si mesma uma
cantiga d amigo (dado que parte de uma sequncia organizada na qual tem um importante
papel estrutural; ver Cohen 1996a, 1996b: 27-36). Por outro lado, Nunes inclui correctamente Dinis-21 (B 572/ V 176), apesar de ser o rapaz quem comea o dilogo (cf. Cohen 1987: 104 [n. 7] e
113 [n. 36]).
1
No vol. VIII, pp. 473-474, os Machados respondem numa Autocrtica, que termina com estas palavras pouco
tranquilizadoras: certo crtico italiano achou-a [scil. edio] cmica, o que redobra a sua utilidade em funo do divertimento proporcionado (p. 474).
2
Ocasionalmente os Machados destacam-se como os nicos editores com a leitura correcta, por lerem acertadamente
com BV (e, em princpio, deveriam ler sempre de acordo com B) onde outros editores erroneamente mudam o texto, ou
porque outros falham em prestar ateno a uma leitura correcta de B (por exemplo, Solaz-3, v. 10 nen lho quis sol caber) ou
ainda porque, em ocasies mais raras, separam correctamente as letras dos manuscritos para apresentar uma sintaxe e
um sentido coerente (ver Guilhade-20, v. 21 chuf an; Solaz-1, v. 2 Agor a e v. 5 Penad a). Nota bene: Ao longo de praticamente
toda a edio refiro as cantigas d amigo pelos nmeros do manuscrito e pelo nome do poeta. Aqui, na introduo, tomei
a liberdade de usar os nmeros correspondentes aos poemas desta edio (ver infra, seco 8). Da nesta nota a indicao
Solaz-3 = B 830 / V 416 (Solaz). Espera-se que este novo sistema facilite as referncias s cantigas d amigo no futuro. Usei-o
apenas algumas vezes na edio (por exemplo, nas notas em Johan Airas-26) porque o sistema foi estabelecido depois de
a maior parte das notas ter sido escrita.
3
A reedio de 1973 (h muito fora de circulao) no incorpora as correces e adies de Nunes ao texto, aparato,
comentrio e glossrio (ver vol. I, pp. 465-471 e vol. III, pp. 707-728; outras correces e sugestes relativas ao texto dos
poemas esto espalhadas pelo comentrio [vol. III, pp. 1-437], e algumas leituras so encontradas no vol. I).
4
E ainda oferece o nico glossrio para o corpus (vol. III, pp. 577-704), embora o de Nunes seja o mais fraco dos glossrios para os quatro gneros. Inclui palavras baseadas em leituras erradas ou conjecturas pouco afortunadas (afogar,
enguear, merger, reassanhar-se [ver Charinho-2, Reimon Gonalves-1, Guilhade-18, Berdia-4]). E naturalmente no regista
palavras que foram eliminadas do texto (enartado, fogir, jogar; [ver Garcia Soares-2, Guilhade-20, Charinho-2], ou que so
o resultado de trabalho tardio no texto (avol, acoomhar-se, uviar [ver Guilhade-21, Bolseiro-8, Bolseiro-2]). Para alm
disto, muitas palavras comuns (Deus, luz, madre, mar, mundo, noite, rio, sal, sol, terra, banhar, chegar, chorar, mandar, negar,
preguntar, vencer, etc.) no constam do glossrio, aparentemente porque os estudantes portugueses no precisariam de
ajuda para entender estas palavras. (No entanto, algumas palavras que mereciam entrar no glossrio so tambm
56
Contudo, inclui erroneamente uma cantiga d amor (CCCLXXXIII = B 1151-1152 / V 754 [Johan
Zorro]) e acrescenta duas pardias de cantigas d amigo encontradas em BV entre as cantigas
d escrnio e mal dizer (CXLV e CXLVI = B 1390 / V 999 e V 1008 [Gonal Eanes do Vinhal]), embora
o seu gnero seja clarificado pela sua localizao e pela rubricas que as precedem.1
Nunes inclui as pastorelas, tal como eu fao, embora deva ser sublinhado que s uma delas
de Johan Perez d Avoin est numa seco da tradio manuscrita que contm exclusivamente as
cantigas d amigo do mesmo autor (mas cf. a posio de Airas Nunes-4 em BV).2
As edies fac-similadas de V (1973) e B (1982) tornaram as fontes manuscritas das cantigas
d amigo mais acessveis aos estudiosos. Para ser justo, as fotografias reproduzidas no fac-simile
de B publicadas pelos Machados (pouco a pouco, no fim de cada um dos oito volumes entre 1949
e 1964), embora reduzidas em tamanho, cortando assim material que estava nas margens, so
por vezes mais claras do que aquelas da edio fac-similada de 1982. E um dos muitos mritos do
estudo de N feito por Ferreira (1986) foi divulgar uma reproduo a cores.3
As edies monogrficas das composies de poetas individuais comeam com Lopez de Moura
1847 (Dinis). E devemos prestar especial ateno a E. Monaci 1875b (as cantigas d amigo de
Johan Servando). Mas as primeiras edies que podemos apelidar de modernas e filologicamente
confiveis (embora no isentas de erros) so as de Henry Roseman Lang 1894 (Dinis) e Oskar
Nobiling 1907a (Johan Garcia de Guilhade). O trabalho neste sentido retomado apenas em 1945
com um fluxo contnuo, de 1960 em diante, em grande parte resultante do trabalho de especialistas italianos, estudiosos treinados em Itlia ou estudiosos que seguiam o modelo italiano, sendo
Celso Ferreira de Cunha (1949, 1956 [ver agora Cunha 1999]) uma notvel excepo. Temos agora
cerca de metade do corpus em edies monogrficas e vrias apareceram ao longo da ltima
dcada mas mesmo assim cerca de metade dos 88 poetas a quem as cantigas d amigo so
atribudas est por trabalhar.
Nem todas as edies tm a mesma qualidade. O meu juzo sobre cada uma delas ficar claro ao
longo do texto, aparato e notas desta edio, por isso no h necessidade de entrar em pormenores
aqui. Algumas edies parecem-me excelentes, a maior parte delas revelou-se til, vrias so manifestamente medocres e umas poucas so indignas de referncia. Nos ltimos anos todas as edideixadas de fora como por exemplo gabar e referir; [ver Guilhade-12, Garcia Soares-2]). Mesmo no que consta do glossrio
encontramos muitas imprecises, anlises gramaticais incorrectas (por exemplo, o no-existente en que [por e que] introduzindo oraes concessivas [ver Pardal-2, v. 7 e Vasconcelos-3, v. 5]) e passagens mal traduzidas (chufan transitivo com
dois objectos, em vez de chufa n intransitivo em Guilhade-20 [na verdade erro de Nobiling; ver Cohen 1996b: 43-44, nota
35]). Alm disso, Nunes no seguiu Michalis (1920) em fornecer etimologias (nem o fez Lapa no seu Vocabulrio das CEM
[1965, 1970] to-pouco Mettmann no seu Glossrio s CSM [1972]). De qualquer forma, o glossrio de Nunes para as
cantigas d amigo continuar a ser usado, e a ser til, desde que manejado com precauo, at que um novo possa ser
editado. (O ideal seria um novo glossrio baseado numa concordncia de texto gerada por computador, com informao sobre frequncia na ocorrncia de palavras, partes do discurso, morfologia verbal, construes gramaticais e
etimologias.)
1
Contudo, deve ser frisado que estes casos arrastam questes importantes, entre as quais: (1) Como definir gneros na lrica
galego-portuguesa? (2) Como e por qu distinguir entre uma cantiga d amigo e uma pardia de uma cantiga d amigo?
(Tavani, por exemplo, [em Rep. e em La Poesia Lirica], usa a classificao cantiga d amigo burlesca, mas ver Cohen 1996b
passim. (A propsito, em V 1008 a fiinda, tal como foi reconstruda por Michalis, e aceite por Nunes e outros, parece-me
errada. Acredito que se deva ler: E, por Deus, que vos <tan grand> onra <fez> / perdoade a don Anrriqu esta vez repetindo assim
as palavras rimantes do refro na fiinda como tantas vezes acontece. (Ver infra 3.4 para convenes editoriais.)
2
A definio e a existncia da pastorela como um gnero na lrica galego-portuguesa tem sido objecto de controvrsia. Tavani (1986: 217-223) d uma viso geral do assunto e referncias. Ver Stegagno Picchio 1975, 1991.
3
Fotografias de N foram disponibilizadas por Vindel (1915) e amplamente reproduzidas, mas as de Ferreira (1986),
depois do restauro, so muito superiores. Ampliaes a preto e branco de cada seco do pergaminho so tambm
fornecidas.
57
Trabalho
empolgante!
es so acompanhadas de aparato, notas, anlise mtrica, introduo e glossrio (ou index verborum),
mas o texto em si por vezes tratado com pouco cuidado. Alguns editores chegaram mesmo a
recuar no progresso feito h muito tempo por Lang, Michalis, Nobiling, Nunes e outros.
Por conseguinte, talvez seja de lamentar que, em geral, apesar dos progressos (aparentes e
reais), to poucos fillogos tenham respondido ao real, se bem que subestimado, desafio de
editar as cantigas d amigo. E deixando de lado a qualidade das edies monogrficas, podemos perguntar o porqu de centenas destes poemas terem esperado tanto tempo at que algum os editasse uma segunda vez.
Afinal de contas, este corpo de lrica amorosa de voz feminina praticamente virgem: a maior
parte dele quase no foi tocada.1 Passaram 160 anos desde a redescoberta de V e 120 anos desde
que B foi realojado, no entanto apenas Nunes editou todos os 500 textos naquilo que se pode
assemelhar a uma edio crtica. Isto h 75 anos atrs. E no nos devemos esquecer que Nunes
( parte outros erros) nunca viu V, baseando-se completamente na transcrio de Monaci, e disps
de pouco tempo, antes de o seu texto ser editado em 1926, para ver e considerar as leituras de B,
que deu entrada na Biblioteca Nacional a 26 de Fevereiro de 1924 (Nunes, Cantigas damigo, vol. II,
p. 442; Cintra 1982).
Assim, no meu entender, e por muitas razes, uma nova edio das cantigas d amigo era
claramente, e tem sido por muito tempo, um desideratum. Sem uma edio creio no ser possvel
um estudo srio deste gnero. A tendncia relativamente s edies monogrficas no sculo XX
parece ser a de olhar para a edio completa de um gnero como uma coisa do passado. Mas h
boas razes para reunir estes 500 textos.
Mesmo sem outras provas, as cantigas d amigo podem ser distinguidas do restante corpus da
lrica medieval galego-portuguesa pela sua posio na tradio manuscrita (ver Oliveira 1994).
Mas tambm temos outras provas: a potica fragmentria em B (ver DHeur 1975b [e agora Tavani
1999]) e referncias nos prprios textos: cantigas damigo em Baveca-4 e Estevan Coelho-2; un cantar
novo d amigo e un cantar d amigo em Sevilha-7 (cf. cantar d amor nen d amigo em V 1032 [Loureno]).
De qualquer forma, todos os 500 textos so exemplos de um tipo de poesia que pode ser distinguido dos outros gneros principais na base de vrios critrios:
(1) Gnero do emissor: feminino, quer seja a Rapariga, a sua Me ou uma das Amigas da
Rapariga (um destes emissores fala no incio do discurso ou ento ao longo dele; existem
apenas onze excepes: Estevan Coelho-1, Afonso Sanchez-2, Solaz-1, Bonaval-1, Zorro-1, Zorro6, Meogo-5, Giinzo-8, Loureno-2, Loureno-3, Dinis-17).2
(2) Relao social entre receptor e possveis destinatrios e o emissor: se a Rapariga fala, geralmente
para a Me, Amiga(s) ou para o Rapaz (ela pode invocar Deus ou mencionar o Rei; outras
personae incluem a Outra uma Outra Rapariga [a quem, no entanto, a Rapariga nunca se dirige
directamente]) e h alguns destinatrios no humanos, muito citados: ondas, cervas e flores do
verde pino; se a Rapariga no o emissor, a ela que a Me ou as Amigas se dirigem ou, ocasionalmente, o Rapaz (ou uma voz masculina).
1
Michalis declara (CA, vol. I, p. x) ser sua inteno publicar uma edio das cantigas d amigo (ou Cancioneiro das Donas,
como ela lhe chama). Embora nunca tenha concludo o projecto, publicou muitas delas no CA: sete no vol. I (CA nos. 331,
413, 416, 428, 434, 444, 452; e muitas mais no vol. II (nas pp. 995-996 refere dezanove que so integralmente impressas
neste volume).
2
Existem tambm cerca de 30 textos com discurso citado, incluindo as palavras de um mensageiro (mandadeiro) em Treez-4 e
de informantes no identificados em Johan Airas-39, fiinda.
58
(3) A relativamente alta frequncia de aaB e formas relacionadas: embora no seja em si prprio um
critrio suficiente, uma tendncia formal que caracteriza o gnero Amigo em relao aos
outros gneros principais (Cohen 1996b: 20-23).
(4) Sintaxe: geralmente coextensiva com a estrofe, e subdividida verso por verso, muitas
vezes hemistquio por hemistquio (de tal forma que o enjambement fortemente sentido)
novamente esta uma tendncia mais marcada no gnero Amigo do que nos outros gneros;
(5) Lxico: marcado no apenas por certos arcasmos, ou tendncias arcaizantes, como tem
sido muitas vezes apontado (Diez 1863, Monaci 1873, Lang 1894, Michalis 1904, vol. II, etc.),
mas pela natureza feminina dos seus limites auto-impostos e por elementos-chave tirados de
uma tradio oral que preservou um cdigo de conduta que regula as negociaes para um
contrato nupcial ou pr-nupcial, chamado fala (Cohen 1996b: 6-7, 37-39 [notas 4 e 6]).
(6) Variedade de actos sociais (ertico-jurdicos): embora estritamente limitados, so mesmo assim mais variados do que nas cantigas d amor (por exemplo, aces discursivas ou jogadas,
como Me, deixa-me ir ver o meu namorado, Vai-te embora, Bem-vindo).
(7) Sequncias, ou conjuntos para execuo, podem ser identificadas pela anlise formal e
constituem uma espcie de teatro lrico.1
Esta edio foi feita a partir de uma nova transcrio de B e da minha verificao e comparao com V e baseada na minha prpria leitura de todos os 500 textos dos manuscritos. O meu
procedimento foi o seguinte:
Primeiro transcrevi todas as 500 cantigas d amigo a partir de fotografias de B na posse da
Biblioteca Nacional (Lisboa, Junho a Setembro de 1991) e seguidamente, feita a collatio codicum de
B e V, preparei uma edio provisria dos 500 textos, li-os directamente de V (Roma, Maio de
1992), e verifiquei a minha transcrio de B com o manuscrito aqui em Lisboa (Novembro de
1992).2 Entre 1991 e 1993 preparei o aparato crtico e acrescentei um primeiro rol de notas; pontuao e acentuao foram adicionadas em 1996; durante os ltimos anos muitas notas foram reescritas, outras acrescentadas, outras eliminadas. Em resumo, texto, aparato crtico e notas iam
mudando de acordo com o avanar do entendimento dos problemas em questo e dos mtodos
disponveis para os resolver. Novos dados compilados, mas no publicados, sobre vrios aspectos do corpus permitiram-me abordar muitos dos problemas com novas ferramentas.
2. Critrios editoriais3
Sem que a isso faa referncia:
dividi palavras que estavam juntas nos manuscritos e juntei partes de uma mesma palavra
que estavam separadas no manuscrito. Onde parece existir uma dvida legtima, e os editores
no concordam na diviso da palavra, registei no aparato opinies divergentes;
Nas CEM encontramos pardias com uma composio sequencial, presumivelmente inspiradas nas cantigas d amigo.
O exemplo mais sustentado e brilhante a sequncia de Lopo Lis contra os Zevres. Para quase toda a bibliografia sobre
sequncias, ver Weiss 1988 (que rejeita sequncias que outros afirmaram ter encontrado, mas descobre sequncias onde
ningum mais as v e, finalmente, no consegue encontrar nem produzir nenhum mtodo para as identificar). As minhas
tentativas para sugerir mtodos podem ser encontradas em Cohen 1987, 1996b.
2
Consultei N em Agosto de 1988 mas desde ento baseio-me nas fotografias em Ferreira 1986.
3
Michalis, CA, vol. I, pp. xiv-xxviii constitui ainda uma discusso til de factos e problemas bsicos.
59
desenvolvi as abreviaturas, convencionais e j aceites, dos manuscritos, a menos que a leitura seja duvidosa, nesse caso as abreviaturas (e sobrescritos) so dados no aparato;1
desenvolvi a abreviatura 9 como os ou us, dependendo da ortografia dessa mesma palavra
noutras passagens dos manuscritos. (Assim se9 olh9 desenvolvido como seus olhos.) No considerei ser necessrio registar aqueles casos onde V tem us e B tem 9 ou vice-versa;
imprimi a conjuno e (<e t) como tal sem registar as variantes grficas no aparato, quer o
manuscrito tenha e, et, ou & (= 7), e sem levar em conta se B e V tm a mesma leitura;
eliminei:
consoantes duplas no etimolgicas;
o h no etimolgico, tanto inicial como intervoclico;
o h etimolgico nas formas de aver (habere) e nas palavras ome (homo, hominis) e oje (hodie).2
distingui entre i/j e u/v;
usei j em palavras onde por vezes os manuscritos tm um g palatal como em desego, oge,
pugi. (So excepes algumas formas de trager e as palavras linhage, sage, etc. [onde leio com os
manuscritos]);
substitui y por i, quando vogal, com j em casos como deseyo (fazendo referncia no aparato);
simplifiquei as combinaes gu e qu (para g e c, respectivamente) em formas elididas como
migu, amigu, vosqu, pouqu onde o gu e o qu so convenes grficas destinadas a mostrar que g ou c
so oclusivas. (Deste modo, g final e c final so sempre oclusivas nesta edio.3 Do mesmo modo
no meio de uma palavra (por exemplo, roguar dado como rogar). Mas, cedendo s modernas
convenes ortogrficas, mantive (e em algumas ocasies restaurei) este u no interior de formas
como roguei, neguedes, guisar (fazendo referncia a estes casos no aparato).
regularizei o uso do s duplo em algumas palavras e formas comuns, registando a leitura dos
manuscritos no aparato. Assim, a leitura do texto assanhar em B 795 / V 397 (Talaveira), v. 7, e o
aparato indica asanhar BV; em B 780 / V 363-364 (Ornelas), v. 8, a leitura do texto morasse e o
aparato indica morase BV.
Os smbolos usados no aparato para representar abreviaturas e letras sobreescritas no so (nem o poderiam ser, em
muitos casos) representaes precisas do que encontramos nos cdices. Assim: q representa uma variedade de marcas (e
mesmo desenhos) nos MSS (e deve normalmente ser lido como que, mas outras vezes como qua); q = quen ou quan; ; s = ser.
p um p com a perna traada (e quase sempre = per), contudo essa perna raramente est traada; p indica um p com uma
roda volta da perna (= pro); pr (= por) indica um p seguido por um r sobreescrito, independentemente da forma do r nos
MSS, mas por vezes, em vez de um r sobreescrito encontramos uma marca que representei simplesmente por um apstrofo; po (= por) significa que o sobreescrito tem uma forma que se assemelha a um crculo. Mas os escribas de B e V
parecem ter simplificado muitas vezes (ou lido mal) algumas das abreviaturas no exemplar, como aquela que frequentemente se encontra para ar.
2
Da e no h, ome e no home, mesmo onde ambos os manuscritos tm a forma com h. B e V apresentam mais vezes estas
palavras sem o h (pelo menos nas cantigas d amigo). Os falantes modernos de algumas lnguas romnicas (especialmente se no estiverem habituados a ler textos medievais) podero estranhar as primeiras vezes que virem oje ou oj (em vez de
hoje), ome ou om (em vez de home e hom), ou (em vez de h) mas que dizer do oi espanhol, do oggi, uomo italiano, e do a,
avoir francs? No tenho qualquer lealdade lingustica nestes assuntos; a minha preocupao apresentar um sistema
grfico coerente e consistente (que facilite a pesquisa electrnica; ver nota 12 supra) para um pblico internacional. De
mais a mais, como Wittgenstein (Tractatus 3.327) observa, Wird ein Zeichen nicht gebraucht, so ist es bedeutungslos. Das ist
der Sinn der Devise Occams. (Se um sinal no necessrio ento no tem significado. Esse o significado da sentena de
Okham). Foneticamente, o sinal h nestas palavras no necessrio (e h tornou-se mudo em incio de palavra em Latim,
mais de um milnio antes).
3
Nas raras ocasies onde (por exemplo) o e final de trage elidido nos MSS, vi-me obrigado, a bem da consistncia, a
escrever traj, registando as leituras do manuscrito no aparato crtico.
60
61
usei o acento agudo para distinguir formas que de outro modo seriam iguais (o timo latino
ou a funo sintctica so dados entre parntesis a bem da clareza):
a (artigo, pronome ou preposio) / (< h a b e t)
da / d (< d a t)
de (< d e ) / d (< d e m or d e t )
e (< e t) / (< e s t)
en (< i n) / n (< i n d e)
esta (< i s t a) / est (< s t a t)
este (< i s t e) / ste (< e s t) / est (< s t e t)
la ([artigo ou pronome] < i l l a-) / l ([advrbio = al] < i l l c )
vos / vs;
usei o acento agudo para:
a ltima slaba de verbos na terceira pessoa no futuro do indicativo (por exemplo, partir,
partirn, para distingui-las das paroxtonas do mais-que-perfeito partira, partiran; mas tambm ser, far, embora no seja possvel qualquer confuso com as formas correspondentes
desses verbos [fora e fezera, respectivamente]);
certas formas de or, soer e seer (o, ostes, oa, soa, sia etc.);
a primeira pessoa do singular do perfeito activo do indicativo de doer e sair: do (< d o l u i)
em Dinis-10, v. 9 e sa (< s a l i (v) i) em Roi Fernandiz-4, v. 15;
os advrbios al, l, al e ac (nota bene: o advrbio *c no existe nas cantigas d amigo);
dei a leitura mais completa nos casos onde uma vogal elidida foi eliminada num dos manuscritos mas mantida no outro (por exemplo, madre o V : madro B). Alguns textos apresentam a forma
elidida num manuscrito e a forma completa no outro, ou a forma elidida num verso e a forma
completa noutro (por exemplo, madreo B contra madro V [B 566 / V 169]). Nestes casos respeitei a
variao quando a primeira vogal devia ser elidida, e acrecentei a vogal em falta onde a mtrica
assim o exigia;1
2
pontuei to bem quanto pude;
abstive-me em geral de pontuar no fim das estrofes (cantigas ateudas at a fiinda so por isso
marcadas principalmente pela falta de letra maiscula no incio das sucessivas estrofes e da fiinda, mas tambm,
onde necessrio, pela presena de pontuao no final, geralmente uma vrgula, no fim das estrofes);
usei aspas para o discurso citado (no caso de um discurso citado dentro de outro um
fenmeno raro: ver as pastorelas de Johan Perez d Avoin, Pedr Amigo de Sevilha, Airas Nunes usei
um segundo e distinto tipo de aspas).
Poderia ter removido todas as vogais que (presumivelmente) deveriam ser elididas. Ao manter as formas mais longas
onde estas ocorrem, permito ao leitor ver que muitas vogais elididas estariam provavelmente presentes num ponto anterior da tradio manuscrita, e tambm que os escribas eliminaram vogais elididas inconsistentemente e muitas vezes de
forma incorrecta.
2
Stegagno Picchio (1984: 661), numa espcie de retractatio, observa: A chamada pontuao interpretativa fixa a uma e a
uma s interpretao (e no importa se a de hoje) um texto por sua natureza polissmico como o texto potico. A poesia
no devia ter pontuao, ou pelo menos devia t-la com muita moderao.... Tentei fazer isto em 1984 (Cohen 1987), e
prefiro este mtodo para meu uso pessoal, mas aqui segui o exemplo de Lang, Nobiling, Michalis e da prpria Stegagno
Picchio. Um texto cuidadosamente pontuado tanto uma ajuda leitura como um guia para a interpretao. (Apercebi-me de que pontuar estes textos muitas vezes clarifica o seu significado, raramente o limita. Alm disso, quem conseguir
ler estes textos sem pontuao tambm pode retirar mentalmente a minha pontuao e deixar os significados irradiar
livremente. O ideal seria termos um texto electrnico que nos permitisse remover toda a pontuao ou escolher o nvel e
tipo de pontuao que desejssemos: cantigas pontuadas consoante o gosto pessoal podero ser a onda do futuro.)
62
3. Smbolos e convenes
3.1 Smbolos e convenes usados no texto
< > Parntesis angulares indicam letras e palavras que no se encontram nos manuscritos,
e so tambm usados aqui (embora a negrito) para todas as partes do refro no presentes em
pelo menos um manuscrito.1
. . .
Pontos espaados indicam lacunas no texto; cada ponto representa uma slaba.
[ ]
Parntesis rectos so usados no texto para marcar palavras que esto presentes nos
manuscritos mas que deviam, na minha opinio, ser eliminadas. Em tais casos, ento, subsiste
uma dvida razovel. Noutros casos (quando parece no haver tal dvida), letras e palavras foram
retiradas e relegadas para o aparato.
negrito
Cruces: usadas para marcar uma palavra ou palavras que creio estarem corrompidas ou
ininteligveis mas que deixei, no entanto, permanecer no texto. Por vezes usei cruces para indicar
uma rima inaceitvel (ver Pardal-4, Galisteu Fernandiz-3), na suposio de que pelo menos parte
do verso onde aparece a rima errada est corrupto.
()
Parntesis normais so usados no texto com a sua funo normal (como pontuao).
< > Parntesis angulares so usados no aparato para delimitar palavras ou letras acrescentadas pelos editores, independentemente de estes editores, nas suas edies, terem usado itlico (por exemplo, Lang 1894) ou parntesis rectos (por exemplo, Nunes 1926) ou mesmo nada
(por exemplo, Braga 1878).
( )
Parntesis normais so usados no aparato para marcar letras, palavras e frases que
foram canceladas nos manuscritos.3 Quando estrofes inteiras foram erradamente copiadas e depois canceladas num manuscrito so reproduzidas numa nota e o facto mencionado no aparato.
1
Isto segue a conveno usada na filologia clssica (e tambm, neste campo, por Stegagno Picchio 1968). Correia (1992)
defende que os parntesis no so necessrios para este propsito e Parkinson (1997) diz que as variaes dos manuscritos nos refres ao longo das Cantigas de Santa Maria no precisam de ser registadas no aparato. Mas nas cantigas d amigo
incertezas sobre o texto e mtrica tal como o problema da variatio no refro pesam a favor do uso dos parntesis angulares
no texto e de informao no aparato.
2
Quando o refro no igual de estrofe para estrofe, mas muito parecido, considerei tais casos como refres com variao
e coloquei-os em negrito. Os exemplos (alguns deles questionveis) so: Carpancho-6; Zorro-5; Bolseiro-3; Baveca-8;
Giinzo-4; Johan Airas-44; Johan Airas-45; Dinis-23; Dinis-33; Dinis-36.
3
Pensei no incluir textos que tivessem sido cancelados, pois esse tipo de informao merece um estudo parte, que
deveria ser feito em todo o corpus, no apenas nas cantigas d amigo. Mas recuei, considerando os dados suficientemente
63
(No restante aparato, isto , fora das leituras dos manuscritos, os parntesis normais tm a sua funo
normal de pontuao).
]
Apenas quando uma palavra ou palavras referidas no aparato podem no ser bvias
usei este smbolo para indicar a que palavra ou frase se est a fazer referncia. Se dada informao sobre a leitura de apenas um dos dois manuscritos, deve assumir-se que o outro l exactamente da mesma forma que o texto impresso. Se a sugesto de um editor dada no aparato sem
mais informao mas no aceite no texto, por exemplo em Camans-2, v. 12, onde o aparato
informa guarria Nunes mas no texto se l guarira, deve assumir-se que ambos os manuscritos
lem exactamente como o texto.
Itlico Seguido por ? (por exemplo, uos ? B) indica que a leitura de uma letra ou letras incerta.
Fora das leituras dos manuscritos, contudo, os itlicos so usados no aparato para nomes de editores, operaes editoriais (ver em baixo) e outras anotaes, quer em latim quer em ingls.
[....] Foram usados para lacunas no texto de N (da, apenas no aparato e notas das cantigas
de Martin Codax).
No aparato apenas no registei variantes grficas (por exemplo, y / i / z)1 e isto inclui a
variao rr / ir, ou seja, os casos onde B tem ir e V rr.2 Tambm omiti praticamente todos os
exemplos de confuso entre u e n se o erro ocorre somente num dos dois manuscritos, fazendo
referncia apenas queles casos onde pode existir uma verdadeira variante, ou onde um problema textual num texto especfico parece pedir uma informao mais completa no aparato desse
poema.3 Indiquei todos os casos onde ambos os manuscritos tm u para n ou n para u, ou seja,
onde o erro parece ter estado presente no antecedente comum de BV.4 Todos os outros erros
(igualmente triviais, mas menos comuns) em um ou em ambos os manuscritos so indicados no
aparato. Onde B e V diferem indiquei apenas a parte relevante do texto, para que o aparato no
seja um guia na diviso de palavras para alm dos limites do que citado. Tambm no indiquei
(excepto quando relevante por uma qualquer razo5) letras e palavras acrescentadas pelo escriba
acima da linha, a menos que a leitura nesse ponto esteja em questo.
teis, o que justifica a sua incluso. De qualquer forma, quem estiver interessado neste problema (que lana uma importante luz sobre a natureza dos erros dos escribas) pode localizar palavras, versos, estrofes canceladas, etc. mais facilmente
se consultar as edies fac-similadas dos manuscritos do que tentando localiz-los num aparato crtico.
1
V frequentemente tem z onde B tem (faza / faa). Nestes casos escolhi a variante em B, em parte por causa da sua correspondncia com a moderna ortografia portuguesa.
2
Cf. Tavani 1963. Este um erro frequente e ocorre em partes das cantigas d amigo copiadas pelos seis escribas de B.
Pareceu-me desnecessrio sobrecarregar o aparato com centenas de casos onde este erro ocorre. De qualquer forma, nem
sempre possvel determinar se o grafema usado em B era suposto indicar ir ou rr.
3
Novamente, incluir todos os exemplos de confuso entre u e n teria sobrecarregado o aparato com centenas de itens sem
utilidade alguma na fixao do texto. Alm disto, em muitos casos (especialmente em B, mas tambm em V) notoriamente difcil, se no impossvel, dizer se a letra em questo u ou n.
4
Isto no quer dizer necessariamente que o erro estava presente no exemplar. De qualquer forma, os escribas de B e V poderiam
ambos ter errado, mesmo que o exemplar estivesse correcto. Ou o exemplar poderia estar errado, no entanto ambos os escribas
poderiam ter cometido o erro (independentemente do porqu ou como) na mesma letra, da corrigir o erro na fonte do texto.
5
Por exemplo, em Johan Airas-3, v. 1 o aparato indica que em V be~ se encontra escrito acima da linha na altura em que foi
omitido pelo escriba. Um erro assim (na primeira tentativa) pode ser invocado como evidncia para a omisso de <ben>
em outros stios (cf. Baveca-7, v. 12).
64
Com estas excepes, onde h alguma diferena, dei a leitura de ambos os manuscritos. Quando
a leitura correcta duvidosa, dei primeiro a leitura do manuscrito mais prxima da leitura adoptada no texto. De resto, a leitura de B precede a de V. Nos casos onde o meu texto concorda com
a emenda, suplemento ou conjectura de um editor, tal vem em primeiro lugar, seguido pela leitura dos manuscritos, seguida por sugestes de outros editores.
Para as operaes crticas mais bsicas, e para alguma outra informao, usei, onde necessrio, palavras e frases latinas. Em questes de mtrica, lineao, letras canceladas e corrigidas,
etc., usei ingls. Deste modo, o latim e o ingls, as lnguas cientficas internacionais da Europa
medieval e moderna, respectivamente, aparecem lado a lado no aparato. Para as palavras latinas
empregues, a forma completa usada para a primeira pessoa (por exemplo, distinxi), enquanto
que uma forma abreviada usada para a terceira pessoa, quer seja singular ou plural (por exemplo, distinx. = distinxit ou distinxerunt).
addidi
correxi
def. = defendit, -erunt
delere voluit
deleto
delevi
distinxi
fort. (= fortasse)
fort. recte
fort. x delendum
fort. x scribendum
iam Fulano
in. ap.
interpunxi
malim
om. = omittitur
possis
possis x vel y
probante
recepto
scripsi
seclusi
servato
spatio relicto
supplevi
transposui
4. Forma estrfica
No raras vezes, as mudanas de linha em B e V no reflectem a forma estrfica do poema.
Quando estes erros na diviso de linhas so triviais (muitas das vezes idnticos em B e em V) no
indiquei tais irregularidades. Quando, no entanto, a forma estrfica est em causa, forneo infor-
65
5. Mtrica
Existia uma opinio antiga que defende que as cantigas d amigo no tm mtrica regular.
Interrogamo-nos sobre em que se basear tal opinio. No em alguma edio fivel do corpus,
uma vez que no existia nenhuma. Nem em nenhum estudo, que de qualquer forma pressupe
uma edio. Aqueles que defendem essa opinio antiquada no encontraro suporte aqui. Pelo
contrrio, tendo examinado o problema em todas as suas vertentes, concordo com Nobiling
(1907a: 10), indo contra a posio representada, por exemplo, por Cunha 1985: 63-66.2
Como regra geral, estes poemas so isossilbicos, tal como Nobiling tinha referido. Cada verso
deve ter o mesmo nmero de slabas que o verso correspondente na mesma posio em todas as
outras estrofes (e isto inclui equivalncias do tipo 7=8). Este o princpio da external responsion
(expresso de Paul Maas). Excepes genunas a este princpio so raras.
Temos, por exemplo, dois poemas com primeiros versos hipermtricos (Carpancho-1, Airas
Paez-1) ou versos hipermtricos em I.1 e II.1 (Torneol-1, Casal-2). E h alguns poemas nos quais a
Para o problema da disposio e forma, ver Parkinson 1987; sobre versos longos copiados em colunas curtas, Parkinson
1997: 64 [ver agora Parkinson 2000b, 2000c]. Ferreira (publicao prxima) lida com evidncia musical para a colometria
em algumas das Cantigas de Santa Maria.
2
Ironicamente, os excelentes estudos de Celso Cunha sobre hiatos e elises (em Cunha 1982) teriam sido impossveis se
no houvesse regularidade mtrica. Por isso Cunha um estranho defensor de mtricas incertas: o seu prprio trabalho
fala contra ele.
66
irregularidade regular e pode reflectir o uso de duas diferentes melodias (Zorro-6, Zorro-9,
Dinis-16 [ver Ferreira 2000]).
Para alm destes casos isolados, h duas categorias principais de excepes: (1) poemas com
uma mistura de rimas graves e agudas e external responsion do tipo 7= 7; e (2) versos
hipermtricos imediatamente antes do refro na ltima estrofe. Ambas as categorias exigem alguma explicao.
Para entendermos o primeiro ponto temos de considerar misturas de rimas agudas e graves.
Se no existe mistura de rimas agudas e graves no corpo da estrofe nem no refro, ento h
quatro tipos bsicos de combinaes, que ocorrem da seguinte forma:
corpo da estrofe
refro
n. de textos
aguda
grave
grave
aguda
aguda
grave
aguda
grave
194
89
41
30
354 textos1
Mas em 95 textos rimas agudas e graves esto misturadas no corpo da estrofe.2 H uma regra
para tais misturas na potica fragmentria no incio de B. Declara esta:
pero que poder meter na cobra das u~as e das outras, se quiser, atanto que per qual guisa as meter en u~a cobra,
per tal guisa as meta nas outras (f.5r).
Por outras palavras, de estrofe para estrofe, uma rima aguda tem de corresponder a uma rima
aguda e uma grave a uma grave. Claro que esta uma regra prescritiva, provavelmente formulada
em meados do sculo XIV. No com certeza uma descrio precisa da prtica potica no nosso
corpus de cantigas d amigo do sculo XIII, como fica claro pelos factos. Esta regra observada
em apenas 34 textos, por um total de 15 poetas:
Avoin 9
Porto Carreiro 1
Padrozelos 7
J. Coelho 6, 7
R. Fernandiz 1, 4, 5
Johan Airas 5, 25, 45
Vinhal 1
S. Sanchez 1, 4
Obviamente, este total exclui as cantigas de meestria. Exclui tambm poemas que, embora no sendo de meestria, no
excedem uma nica estrofe (e podem ser fragmentrios; de qualquer forma, no se pode considerar terem refro uma vez
que no h repetio) e aquelas que no tm refro mas dificilmente podem ser chamadas de meestria, como Bonaval-1 e
Meogo-3.
2
Em sete textos rimas agudas e graves esto misturadas no refro: Carpancho-6, Pardal-5, Ponte-1, Porto-Carreiro-1, Roi
Fernandiz-5, Servando-1 e Zorro-6. Em apenas dois poemas no corpus de 500, Pardal-5 e Roi Fernandiz-5, encontramos
uma mistura de graves e agudas tanto no corpo da estrofe como no refro.
67
68
poeta/poema
escanso
forma estrfica
localizao/palavras/rimas
Calheiros-3:
J. Coelho-1:
Guilhade-6
Guilhade-17:
Charinho-3:
Berdia-4:
Servando-2:
Servando-5:
Servando-7:
*Servando-13:
Meogo-6:
*Padrozelos-5:
Cangas-1:
Airas Paez-2:
9 / 9
10 / 10
10/ 10
10 / 10
10 / 10
7 / 7
7 / 7
15 / 15
10 / 10
7 / 7
5 / 5
7 / 7
7 / 7
15 / 15
aaB
abbaC
abbaa
abbacca
ababCC
ababcCCC
ababCC
aaB
aaB
aabBB
aaB
abbaCC
abbaC
aaB
Ia amigo / migo
IIIa vedar / perdoar
Ib falou / queixou
IIc fezesse / desse
IIa diria / poderia
Ib guisei / ousei; IIb quer / quiser
Ib vingar / assanhar
IIIa par / leixar
IIIa tolhedes / dizedes
Ia -en; IIa -ades; IIIa -al; IVa -ando
Va lavei / liei; VIa liei / asperei
I -ada / -ejo; II -i / -ongo; III -ar / -ei
IIb mal / val
IIa <coi>tado / grado
Estes so, rigorosamente falando, os textos no isossilbicos, mesmo que possam ser chamados
isomtricos se permitirmos, como estes poetas claramente permitiram, equivalncias do tipo n = n.
A segunda grande categoria de excepes ao princpio de external responsion diz respeito a
cerca de 20 cantigas d amigo nas quais o ltimo verso antes do refro final hipermtrico em BV.
Apelidei esta posio de Rv-1 (= Refram Vltimo minus 1 [Cohen 1996b: 8-9]). Certamente que alguns destes versos hipermtricos em Rv-1 esto correctos, apesar da falta de external responsion
(Johan Soarez Coelho-8, Travanca-4, Beesteiros-1, Servando-6, Lopo-3, Codax-1, Codax-6, Johan
Airas-19, Johan Airas-38, Dinis-14). Outros so ilusrios e devem ser corrigidos (Ponte-3, Bonaval8). E h casos onde vogais extra que podem ser elididas provavelmente deviam s-lo (Avoin-4,
Guilhade-13, Estevan da Guarda-1; cf. Servando-11). Todavia, outros possveis exemplos envolvem problemas textuais um tanto difceis (Ulhoa-3, Airas Paez-1, Dinis-43). De qualquer forma,
versos hipermtricos em Rv-1 ocorrem vezes suficientes para que esta seja considerada como uma
69
possvel posio privilegiada na forma musical, onde um verso hipermtrico era ocasionalmente
admitido.
Similarmente, versos hipermtricos encontrados nas fiindas so provavelmente genunos (Johan
Soarez Coelho-10, Johan Airas-39, Johan Airas-43; cf. Servando-9). Recuando no stemma da tradio manuscrita sabemos que era reservado por vezes espao especial para a msica das fiindas
(ver Ramos 1984).
No caso das excepes acima discutidas, versos que parecem aceitveis em BV, mas tm slabas a mais ou a menos, so, na minha opinio, resultantes de uma transmisso errnea, tal como
a comparao entre B e V demonstra. Devemos lembrar-nos que se tivessmos apenas B ou apenas V poderamos facilmente enganar-nos, tal como Braga e outros fizeram, defendendo como
sendo irregularidades mtricas erros que o outro manuscrito demonstra serem, de facto, erros.
Centenas de erros em BV devem ser corrigidos antes de termos um texto das cantigas d amigo
pronto a ser lido. Faltas mtricas frequentemente coincidem com erros na morfologia, sintaxe,
sentido ou estilo. Por outras palavras, uma falha mtrica geralmente uma de vrias pistas que
indicam que algo de errado se passa com o texto. Por vezes a nossa primeira pista, raramente a
nica; mas mesmo assim devemos prestar ateno.
Referncias s edies monogrficas no aparato e notas devem ser procuradas na seco 2.3 da Bibliografia. Referncias aos
textos de Braga, Nunes, Machado geralmente no so mais especificadas, uma vez que as edies contm os nmeros dos
manuscritos. Referncias edio de Nunes que no se refiram ao texto no vol. II so especificadas como, por exemplo, Nunes
(III, 394) no aparato; mas como, por exemplo, Nunes, vol. III, p. 691 nas notas. Correces ao texto feitas por Monaci (excepto no
caso de Johan Servando) so tiradas de Monaci 1875a: 431-438 onde podem ser localizadas atravs da numerao em V. Referncias a Michalis, CA, vol. II so dadas no aparato apenas com o volume e nmero de pgina por exemplo, Michalis (II, 881),
enquanto que nas notas so dadas como, por exemplo, Michalis (vol. II, p. 881). Referncias a Lapa, a menos que especificado
de outra forma, so de Lapa 1982: 140-195 (originalmente publicado em 1929) e no so mais referidas no aparato, enquanto
que nas notas dada a referncia completa. Se uma edio monogrfica citada fora do texto desse poeta, dada a citao
completa, por exemplo, Lang (1904: lxviii), enquanto que a mesma edio monogrfica simplesmente citada como, por exemplo, Lang (p. 173), no conjunto de cantigas atribudas a esse poeta. CA, CEM, CSM, e Rep. so usadas ao longo de toda a edio
(ver Bibliografia 2, ad init.). Mas para as excepes acima referidas, as citaes so dadas de acordo com Parkinson 1997: 64 ou
simplesmente Parkinson 1997, dependendo da especificidade da referncia. Umas poucas obras que so exteriores a este
campo e so citadas apenas uma vez no aparecem na Bibliografia (por exemplo, Libro de Buen Amor).
2
Para este propsito considerei a edio das cantigas d amigo de Johan Garcia de Guilhade (Cohen 1996a) como uma antecipao (no sem falhas) da presente edio. Assim, em Guilhade-22, v. 12, o aparato indica: s<a>budo scripsi (em vez de s<a>budo
Cohen) apesar de tal j ter sido publicado. O mesmo se aplica s leituras que surgiram na Histria e Antologia da Literatura Portuguesa
(Sculos XIII-XIV), n. 2, (Lisboa: Fundao Calouste Gulbenkian - Servio de Bibliotecas e Apoio Leitura, 1997), onde as 74
cantigas nas pp. 9-31 foram reproduzidas a partir de uma verso anterior, no-publicada, desta edio.
70
Ao considerar mudanas ao texto, mantive-me fiel s regras filolgicas da evidncia e fiz grande uso de paralelos das cantigas d amigo. Mas onde necessrio ou conveniente recorri a todo o
corpus da lrica medieval galego-portuguesa, bem como prosa medieval, etc. Normalmente, um
exemplo do mesmo poeta no mesmo gnero ou por outro poeta no mesmo gnero ou pelo mesmo poeta num outro gnero tero mais peso do que um exemplo de outro poeta num outro
gnero. Mas gostaria especialmente de chamar ateno para a importncia de comparar BV com
o Cancioneiro da Ajuda, sigla A. Quando nas cantigas d amor tal comparao possvel (ver quadro
em Oliveira 1994: 126-127), podemos aprender valiosas lies sobre o tipo de erros que encontramos nas cantigas d amigo.
7. Notas
7.1 Notas ao aparato crtico
As notas ao aparato tm como finalidade esclarecer decises editoriais. Dizem respeito principalmente a questes relativas paleografia, mtrica, forma estrfica, morfologia, sintaxe e lxico.
Questes de interpretao so levantadas apenas quando directamente relevantes para um problema textual.1
Ainda precisamos de uma concordncia para o corpus das cantigas d amigo, um glossrio (para substituir o de Nunes no
vol. III; ver p. 56, nota 4 supra) e (embora Nunes fornea algumas notas gramaticais no vol. I, pp. 357-405) uma gramtica.
Enquanto tal no acontece, leitores e estudantes das cantigas d amigo e da lrica galego-portuguesa em geral so aconselhados a usar a gramtica de Huber (1933 [tr. port., com reviso de Lapa e nota introdutria de L. F. Lindley Cintra,
1993]) e o glossrio de Michalis (1920 [agora no vol. I da reedio de 1990 do CA]), de Lapa (CEM), e de Mettmann (CSM,
vol. IV da 1 ed. [no vol. II da 2 ed.]).
71
maiores, de Johan Airas e D. Dinis esto posicionados depois de Fernand Esquio, o que justo,
uma vez que derivam de cancioneiros individuais (Oliveira 1994). Textos do mesmo autor, na seco
inicial de B, que podem ser juntos, so-no, enquanto que os poucos que no o podem ser, nomeadamente aqueles de Fernan Figueira de Lemos, Rodrig Eanes Redondo, Pero Mafaldo, e Alfonso
X, so includos numa pequena adenda entre Fernand Esquio e os cancioneiros individuais.
As pastorelas no devem ser consideradas cantigas d amigo em sentido estrito. Critrios formais so suficientes para as distinguir, e so evidentemente um produto importado, embora adaptado tradio local. Ainda assim, foram aqui includas por causa das suas relaes bvias com o
corpus. Por razes de convenincia, e seguindo o exemplo da pastorela de Johan d Avoin, localizada
no fim das suas cantigas d amigo em BV, as outras seis pastorelas, embora no estejam nos
manuscritos juntamente com as cantigas d amigo dos respectivos autores, seguem-se-lhes
nesta edio. Assim, as sete pastorelas surgem como Avoin-12, Airas Nunes-*4, Sevilha-*12, Johan
Airas-*46, Dinis-*53*, Dinis-*54* e Dinis-*55.
Os poemas de cada autor esto numerados de acordo com a ordem em que aparecem nos
manuscritos, e qualquer texto no-consecutivo marcado com um asterisco. Um asterisco esquerda do nmero indica que em BV o poema em questo no contguo ao texto (do mesmo
autor) que o precede nesta edio; um asterisco direita do nmero indica que em BV o poema no
contguo ao texto (do mesmo autor) que o segue nesta edio.
* * *
Editei as cantigas d amigo para as tornar acessveis a todos os leitores interessados, mas
sobretudo aos estudiosos e estudantes, uma vez que apenas uma edio crtica fivel poder
servir como base de pesquisa. O texto no modernizado; a ortografia meramente regularizada de acordo com a prtica do sculo XIII. E no imponho uma interpretao geral; forneo apenas notas filolgicas e explicativas onde me pareceram necessrias.
Editar estes textos mostrou-me, repetidamente, a sua complexidade e subtileza, tantas vezes
desprezada e subestimada. O meu esforo vai ser suficientemente recompensado se mais leitores
poderem desfrutar desta poesia, mas tambm espero que os estudiosos comecem a considerar
estas composies mais seriamente do que at aqui tm feito.
As cantigas d amigo so bem merecedoras de toda a ateno e cuidado que as geraes
futuras lhes possam conceder. Estas cantigas no so uma espcie de ornamento social, so
antes uma parte dos seres humanos e da sua histria.
72
AGRADECIMENTOS
Em primeiro lugar agradeo Biblioteca Nacional (Lisboa), Biblioteca Apostolica Vaticana e Pierpont
Morgan Library. Graas a estas bibliotecas foi-me possvel ler todo o corpus das cantigas d amigo directamente dos manuscritos. Agradeo tambm o apoio das instituies que me ajudaram durante os anos
iniciais deste projecto: Fundao Calouste Gulbenkian (1991-92); Instituto de Lngua e Cultura Portuguesa
(1992-93); Instituto Cames (1993-94); Junta Nacional de Investigao Cientfica e Tecnolgica/Praxis XXI
(1994-96).
Devo agradecer tambm a vrios estudiosos nesta rea, desde os pioneiros at aos contemporneos e
estudantes. Mas tenho de recuar no tempo.
Estudante algo rebelde, beneficiei da pacincia e sabedoria de excelentes professores na Universidade
da Califrnia, em Filologia Grega, Latina, Inglesa e Romnica: professores Keith Aldrich, Alva Walter Bennett,
Kim Ziegel, e a professora visitante de Roma, Luciana Stegagno Picchio. A eles devo o que aprendi na minha
juventude sobre exactido e honestidade filolgica. E s universidades onde ensinei e cujas bibliotecas
usei, como a Universidade de Pensilvnia e a Universidade de Chicago, e a certos colegas que tive a, expresso tambm a minha gratido.
Muitos estudiosos, alguns dos quais amigos, ajudaram-me desde que iniciei este projecto. Alguns deles
talvez estranhem o agradecimento, podero dizer mas que fiz eu?. A resposta Conversou comigo quando
pedi o seu conselho. A quem poderei agradecer? Por alguma coisa, ou por muitas coisas, ou por tanta coisa
que impossvel agradecer: aos professores Stephen Reckert, Federico Corriente, Valeria Bertolucci Pizzorusso,
Arthur Askins, Yvette K. Centeno, Harvey Sharrer, K. David Jackson, Artur Anselmo, Elsa Gonalves, Carmen
Radulet, Ivo Jos de Castro, Isabel Allegro, Anna Ferrari, Pedro Ferr, Joo Dionsio, Mariagrazia Russo, ngela
Correia, Stephen Parkinson, Manuel Pedro Ferreira, Antnio Resende de Oliveira e Gema Valln. De uma
forma ou outra o contacto com cada um deles contribuu para o meu entendimento deste material e fortaleceu a minha determinao em completar esta edio.
Devo agradecer tambm a Joana Varela, da Colquio-Letras, que me ajudou numa edio de 22 cantigas
d amigo de Johan de Guilhade (e um longo artigo) em 1996-97. A velhos amigos e famlia: Jos Blanc de
Portugal, Robert W. Wallace, Stephen Malinowski, Pamela Bury, Kevin Oderman, Peter Bing, Pilar Garca
Mouton, Rafael Aburto, Mnica Surez, Peter Steiner, Srgio de Portugal, Maria Jos Machado, Martin D.
Cohen, Todd Cohen, Wendy Phillips e Pieta Bellizia.
Michael Weiss, outrora meu aluno de latim, hoje perito em lingustica histrica, ajudou-me em alguns
assuntos tcnicos. E tambm aprendi com os meus alunos aqui em Lisboa, no campo, sempre arriscado, da
filologia medieval galego-portuguesa: Patrcia Ferreira, Marta Castelo Branco Gonalves, Manuela de Sousa
Pereira e Isabel Rodrigues. Em particular, durante uma reviso (1998-99), Isabel Rodrigues pacientemente
confrontou o meu texto com as edies fac-similadas dos manuscritos e bibliografia relevante e fez muitas
correces e sugestes valiosas.
Estou tambm extremamente grato Johns Hopkins University, onde tenho sido professor convidado
durante o ano acadmico 2002-2003. Muito obrigado, portanto, a Daniel H. Weiss, Dean da Krieger School
of Arts and Sciences, e ao Professor Stephen G. Nichols, Chefe do Departamento de Lnguas e Literaturas
Romnicas.
Finalmente, dedico este livro, aos meus filhos e a Luciana Stegagno Picchio, Professor Emeritus da
Universit di Roma, La Sapienza. Arquitecta dos estudos modernos neste campo, Luciana foi tambm a
inspirao para esta edio e aconselhou-me em inmeras ocasies. Mas tenho ainda uma outra razo.
Durante vinte anos, Luciana tem permanecido sempre, em todos os momentos, a mais verdadeira dos verdadeiros amigos.
73
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
75
76
139
140
Airas Carpancho
Chegades [vs, ai] amiga, du meu amigo
Tanto sei eu de mi parte quanto de meu coraon,
Madre velida, meu amigo vi,
A maior coita que eu no mund ei
Que me mandades, ai madre, fazer
Madre, pois vs desamor avedes
A mha coita non lhi sei guarida;
Por fazer romaria puj en meu coraon
141
142
143
144
146
147
148
149
Vaasco Gil
Irma, o meu amigo, que mi quer ben de coraon
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
Estevan Reimondo
Amigo, se ben ajades,
Anda triste <o> meu amigo,
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
197
198
Roi Queimado
O meu amigo, que me mui gran ben
O meu amig, ai amiga,
Quando meu amigo souber
Dize<n> mh ora que non verr
199
200
201
202
203
204
205
206
Estevan Coelho
Sedia la fremosa seu sirgo torcendo,
Se oj o meu amigo
207
208
Estevan Travanca
Por Deus, amiga, que preguntedes
209
77
78
210
212
213
214
215
216
217
218
219
220
221
222
Pero Viviaez
Pois nossas madres van a San Simon
Por Deus, amiga, punhad en partir
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
Estevan da Guarda
A voss amig, amiga, que prol ten
252
Pero d Ornelas
Avedes vs, amiga, guisado
253
Afonso Sanchez
Quand, amiga, meu amigo veer,
Dizia la fremosinha:
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
Fernan Froiaz
Juravades mi [vs], amigo,
Que trist anda meu amigo,
Amigo, preguntar vos ei
Por que se foi daqui meu amigo
271
272
273
274
79
80
Fernan Velho
Vedes, amigo, <o> que oj o
275
276
277
278
279
280
281
283
284
285
287
289
Pero da Ponte
Vistes, madr, o escudeiro que m ouvera levar sigo?
Vistes, madr, o que dizia
Mha madre, pois se foi daqui
Foi s o meu amigo daqui
Pois vos ides daqui, ai meu amigo,
Por Deus, amig, e que ser de mi,
Ai madr, o que me namorou
290
291
292
293
294
295
296
297
299
300
301
302
303
Johan Garcia
Donas, fezeron ir daqui
O meu amigo, que eu sempr amei,
304
305
Reimon Gonalvez
Foste<s> vos vs, meu amigo, daqui
306
Garcia Soares
Filha, do voss amigo m gran ben,
Madre, se meu amigo veesse,
307
308
309
310
Meendinho
Sea m eu na ermida de San Simhon
311
313
314
315
Airas Nunes
A Santiag en romaria ven
Bailemos ns ja todas tres, ai amigas,
Bailade oje, ai filha, que prazer vejades,
O oj eu u~ a pastor cantar (pastorela)
316
317
318
319
321
322
323
324
Gomez Garcia
Diz meu amigo que me serve ben
326
Roi Fernandiz
Conhosco me, meu amigo,
Se vos non pesar ende,
Id meu amigo daqui
Ai madre, que mui<t eu err>ei
Madre, pois amor ei migo
Ora non dev eu prear parecer
Madre, quer oj eu ir veer
327
328
329
330
331
332
333
Pae de Cana
Vedes que gran desmesura,
Amiga, o voss amigo
334
335
Sancho Sanchez
Amiga, ben sei do meu amigo
Amiga, do meu amigo
Ir vos queredes, [ai meu] amigo, <d aqui>
Que mui gran torto mi fez, amiga,
En outro dia en San Salvador
336
337
338
339
340
81
82
341
342
343
344
345
346
347
348
Pero de Berdia
Sanhudo m meu amig e non sei,
Jurava mh o meu amigo,
Deu-lo sabe, coitada vivo mais ca soa,
Assanhou s o meu amigo
Foi s o meu amigo daqui
349
350
351
352
353
Nuno Porco
Irei a lo mar vee-lo meu amigo;
354
Pero de Veer
Ai Deus, que doo que eu de mi ei,
Assanhei me vos, amigo, noutro dia,
A Santa Maria fiz ir meu amigo
Do meu amig, a que eu quero ben,
Assanhei me vos, amigo, per ba fe, con sandece
Vejo vos, filha, tan de coraon
355
356
357
358
359
360
Bernal de Bonaval
Fremosas, a Deus grado, tan bon dia comigo,
Quero vos eu, mha irmana, rogar
Ai fremosinha, se ben ajades,
Pois mi dizedes, amigo, ca mi queredes vs melhor
Se veess o meu amigo a Bonaval e me visse,
Diss a fremosa en Bonaval assi:
Rogar vos quer eu, mha madre <e> mha senhor,
Filha fremosa, vedes que vos digo:
361
362
363
364
365
366
367
368
Johan Servando
Quand eu a San Servando fui un dia daqui
Ir se quer o meu amigo,
A San Servand en oraon
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
Johan Zorro
Quen visse andar fremos~ia
Os meus olhos e o meu coraon
Per ribeira do rio
El rei de Portugale
Cabelos, los meus cabelos,
Pela ribeira do rio
Mete el rei barcas no rio forte;
Jus a lo mar e o rio
Pela ribeira do rio salido
Bailemos agora, por Deus, ai velidas,
386
387
388
389
390
391
392
393
394
395
396
397
398
Juio Bolseiro
Sen meu amigo manh eu senlheira,
Da noite d eire poderan fazer
Fui oj eu, madre, veer meu amigo
Nas barcas novas foi s o meu amigo daqui,
Vej eu, mha filha, quant meu cuidar,
Que olhos son que vergonha non an,
Mal me tragedes, ai filha, por que quer aver amigo,
Buscastes m, ai amigo, muito mal
Fez u~ a cantiga d amor
Ai madre, nunca mal senti<u>
Ai meu amigo, meu, per ba fe,
Aquestas noites tan longas que Deus fez en grave dia
399
400
401
402
403
404
405
407
408
409
410
411
83
84
412
413
414
Martin Campina
O meu amig, amiga, vej andar
Diz meu amigo que eu o mandei
415
416
Pero Meogo
O meu amig, a que preito talhei,
Por mui fremosa, que sanhuda estou
Tal vai o meu amigo, con amor que lh eu dei,
Ai cervas do monte, vin vos preguntar:
<Levou s aa alva>, levou s a velida,
Enas verdes ervas
Preguntar vos quer eu, madre,
Fostes, filha, eno bailar
Digades, filha, mha filha velida,
417
418
419
420
421
422
423
424
425
Martin de Caldas
Per quaes novas oj eu aprendi,
Madr e senhor, leixade m ir veer
Mandad ei migo qual eu desejei
Foi s un dia meu amigo daqui
Ai meu amig e lume destes meus
Nostro Senhor, e como poderei
Vedes qual preit eu querria trager,
426
427
428
429
430
431
432
Nuno Treez
Des quando vos fostes daqui, meu amigo, sen meu prazer,
San Clemeno do mar,
Non vou eu a San Clemeno orar, e fao gran razon,
Estava m en San Clemeno, u fora fazer oraon,
433
434
435
436
Pero dArmea
Sej eu fremosa con mui gran pesar
Amiga, grand engan ouv a prender
Mhas amigas, quero m eu des aqui
Amigo, mando vos migo falar
437
438
439
440
441
442
443
444
445
446
447
448
449
451
453
455
Johan Baveca
Amiga, dizen que meu amig
Por Deus, amiga, preguntar vos ei
Ai amiga, oje falou comigo
Amigo, sei que mui gran sazon
Pesa mh, amiga, por vos non mentir,
Filha, de grado queria saber
Vossa menaj, amigo, non ren,
Amig, entendo que non ouvestes
Como cuidades, amiga, fazer
Amigo, vs non queredes catar
Madr, o que sei que mi quer mui gran ben
Ora veerei, amiga, que far
Amigo, mal soubestes encobrir
457
458
459
460
461
462
463
464
465
466
467
468
469
Pero d Ambroa
Ai meu amigo, pero vs andades
470
Pae Calvo
Foi s o namorado, madr, e non o vejo;
Ai madr, o que ben queria
471
472
Martin Padrozelos
Eu louana, en quant eu viva for,
Gran sazon , meu amigo,
Amig, avia queixume
Madr, enviou volo meu amigo
Ai meu amigo, coitada
Por Deus, que vos non pes,
Amigas, sejo cuidando
Fostes vos vs, meu amigo, daqui
Id oj, ai meu amigo, led a San Salvador,
473
474
475
476
477
478
479
480
481
Lopo
Pois vs, meu amigo, morar
Polo meu mal filhou <s ora> el rei
And ora trist e fremosa
Por que se foi meu amigo
482
483
484
485
85
Filha, se gradoedes,
Por Deus vos rogo, madre, que mi digades
Disseron m agora do meu namorado
Assanhou se, madr<e>, o que mi quer gran ben
86
486
487
488
489
Galisteu Fernandiz
O voss amigo foi s oje daqui
Meu amigo sei ca se foi daqui
Por Deus, amiga, que pode seer
Dizen do meu amigo ca mi fez pesar,
490
491
492
493
Loureno
Ir vos queredes, amigo,
~a moa namorada
U
Tres moas cantavan d amor,
Assaz meu amigo trobador,
Amiga, des que meu amigo vi,
Ja gora meu amigo filharia
Amiga, quero m ora cousecer
494
495
496
497
498
499
500
Golparro
Mal fa eu velida, que ora non vou
501
Johan de Cangas
En San Momed, u sabedes
Fui eu, madr, a San Momed u me cuidei
Amigo, se mi gran ben queredes,
502
503
504
Martin de Giinzo
Como vivo coitada, madre, por meu amigo,
Se vos prouguer, madr, oj este dia
Treides, ai mha madr, en romaria
Non poss eu, madre, ir a Santa Cecilia,
Ai vertudes de Santa Cecilia,
Non mi digades, madre, mal e irei
Nunca eu vi melhor ermida nen mais santa,
A do mui bon parecer
505
506
507
508
509
510
511
512
Martin Codax
Ondas do mar de Vigo,
Mandad ei comigo
Mia irmana fremosa, treides comigo
Ai Deus, se sab ora meu amigo
Quantas sabedes amar amigo,
Eno sagrado en Vigo
Ai ondas que eu vin veer,
513
514
515
516
517
518
519
Airas Paez
Quer ir a Santa Maria [de Rea,] e, irmana, treides migo,
Por vee-lo namorado, que muit que eu non vi,
520
521
Fernan do Lago
D ir a Santa Maria do Lag ei gran sabor
522
Johan de Requeixo
Fui eu, madr, en romaria a Faro con meu amigo
A Far<o> un dia irei, madre, se vos prouguer,
Pois vs, filha, queredes mui gran ben
Atender quer eu mandado que m enviou meu amigo
Amiga, quen oje ouvesse mandado do meu amigo
523
524
525
526
527
Fernan d Esquio
O vosso amigo, assi Deus m empar,
O voss amigo trist e sen razon
Vaiamos, irmana, vaiamos dormir
Que adubastes, amigo, al en Lug u andastes,
528
529
530
531
addendun
Fernan Figueira de Lemos
Diz meu amigo que lhe faa ben B 47 f. 15v
532
533
Pero Mafaldo
Ai amiga, sempr avedes sabor B 373 f. 85r
O meu amig, amiga, que me gran ben fezia, B 383 f. 86r
534
535
Alfonso X
Ai eu coitada, como vivo en gran cuidado B 456 f. 101r
536
Cancioneiros individuais
Johan Airas de Santiago
Todalas cousas eu vejo partir
Dizen, amigo, que outra senhor
O que soa, mha filha, morrer
Par Deus, mha madr, o que mi gran ben quer
O meu amigo novas sabe ja
537
538
539
540
541
87
88
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
565
566
567
568
569
570
571
572
573
575
576
577
578
579
580
581
582
583
584
585
Dinis
Ben entendi, meu amigo,
Amiga, muit gran sazon
Que trist oj meu amigo,
Dos que ora son na oste,
Que muit ja que non vejo
Chegou m or aqui recado,
O meu amig, amiga, non quer eu
Amiga, bon grad aja Deus
Vs que vos en vossos cantares meu
Roga m oje, filha, o voss amigo
Pesar mi fez meu amigo,
Amiga, sei eu ben du~ a molher
Bon dia vi amigo,
Non chegou, madre, o meu amigo,
De que morredes, filha, a do corpo velido?
Ai flores, ai flores do verde pino,
Levantou s a velida,
Amig e meu amigo,
O voss amigo tan de coraon
Com ousar parecer ante mi
En grave dia, senhor, que vos o
Amiga, fao me maravilhada
O voss amig, amiga, vi andar
Amigo, queredes vos ir?
Dizede por Deus, amigo:
Non poss eu, meu amigo,
Por Deus, amigo, quen cuidaria
O meu amigo de mal assaz
Meu amigo, non poss eu guarecer
Que coita ouvestes, madr e senhor,
Amig e fals e desleal,
Meu amigo ven oj aqui
Quisera vosco falar de grado,
Vi vos, madre, con meu amig aqui
Gran temp , meu amigo, que non quis Deus
Valer vos ia, amigo, <meu ben>,
Pera veer meu amigo,
Chegou mh, amiga, recado
De morrerdes por mi gran dereit ,
Mha madre velida,
Coitada viv, amigo, por que vos non vejo,
O voss amig, ai amiga,
Ai fals amig e sen lealdade,
Meu amig, u eu sejo,
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
604
605
607
608
609
610
611
612
613
614
615
616
618
619
620
621
622
623
624
625
626
627
628
630
631
632
633
89
90
634
635
636
637
638
639
640
641
642
643
644
532
533
536
150
225
252
253
270
275
306
311
326
341
354
470
501
522
2 Pero MAFALDO
Pero Garcia BURGALS
Estevan REIMONDO
Estevan COELHO
Afonso Meendez de BEESTEIROS
Pero VIVIAEZ
Afonso SANCHEZ
Johan GARCIA
Garcia SOARES
Vaasco Rodriguez de CALVELO
Pae de CANA
Martin CAMPINA
Pai CALVO
Airas PAEZ
(14 poets, 28 texts)
534-535
134-135
178-179
207-208
218-219
223-224
254-255
304-305
307-308
309-310
334-335
415-416
471-472
520-521
123-125
220-222
281-284
285-289
313-315
342-344
103
396-398
502-504
119-122
187-190
214-217
203-206
209-213
199-202
226-229
271-274
316-319
321-324
345-348
433-436
437-440
490-493
528-531
136-140
276-280
336-340
349-353
523-527
264-269
297-303
355-360
180-186
191-198
290-296
327-333
426-432
494-500
513-519
111-118
126-133
141-149
256-263
104
Bernal de BONAVAL
LOPO
Martin de GIINZO
(7 poets, 56 texts)
361-368
482-489
505-512
9 Pero MEOGO
Martin PADROZELOS
417-425
473-481
10 Johan ZORRO*
386-395
151-162
441-456
13 Johan BAVECA
457-469
15 Juio BOLSEIRO
Johan Soarez COELHO
399-414
163-177
16 Johan SERVANDO*
369-385
230-251
537-585
586-644
105
108
109
110
10
15
20
10 mais om. V
13 Fiey B : Eiey V
vv. 6, 9: non prix seu castigo (com prix < *prensi) significa
No prestei ateno ao seu aviso (ou no segui o
seu conselho).
111
10
15
In V the reading is apparently agora amigo but the r is not clearly legible, and the a of amigo seems to have been cancelled.
112
10
113
114
10
10
115
10
6 mi B
The way the repetition of the refrain is handled in strophes II and III (with a break after tornarey) suggests that
it might consist of two verses (although it has been copied on one line in strophe I).
116
10
15
20
117
10
15
10 fort. de que
The text is copied in short lines in BV, but Nunes (vol. I, p. 414) and Lapa (1982 : 160) recommend long verses.
118
10
15
119
Cuidades vs, meu amigo, ca vos non quer eu mui gran ben
e a mi nunca ben venha, se eu vejo no mundo ren
que a mi tolha desejo
de vs, u vos eu non vejo
E, maca-lo vs cuidades, eno meu coraon vos ei
tan grand amor, meu amigo, que cousa no mundo non sei
que a mi tolha desejo
<de vs, u vos eu non vejo>
E nunca mi ben querades, que mi ser de morte par,
se souberdes, meu amigo, ca poss eu ren no mund achar
que a mi tolha <desejo
de vs, u vos eu non vejo
10 no] n V
120
10
10
15
121
10
15
vv. 3-4: The sense is: I look and I see and I cannot see
anything, my friend, that could please me. The
expression te~er olho seems on occasion also to mean,
have an eye for, lust after.
Cf. B 200 (Burgals), v. 14 quando a vejo, quam fremosa , / e a vejo falar, per ba fe, / teend olho, sayo de meu sen! And see also CSM
249.32 non tev ollo por al. Previous editors seem to have been blind to the eye in this verse (Nunes prints tenho e vej; Lapa
[1982: 161] defends tenho-lhe vej), but the reading seems confirmed by the references to eyes in vv. 5 and 13-14. (The
expression can also be found in medieval prose, e. g. in the Demanda do Santo Graal [ed. Irene Freire Nunes, Lisboa, Imprensa Nacional-Casa da Moeda, 1995]: ... e teve olho pelo paao u siiam aas mesas e viu antre outros u~u~ homem... [p. 131]; E o cavaleiro teve
olho ps si quando ouiu Tristam ir deps si, e esteve [p. 283].)
2
Nunes (who emends to enfinta) seems not to have understood the phrase nen en preito non . Cf. CA 210, vv. 7-10 (Seabra)
and CSM 273.36, 385.32.
122
10
v. 8: por is causal. For this sense of dizer cf. B 672 / B 274 [Avoin], v. 22 ca fa eu quanto dizedes).
123
10
gaar BV
6 que BV : ca Nunes2
6, 10, 14 q BV : ca Nunes
Nunes prints mais, Deus! que cuida mi a gar / de lhi saberem ca mi quer gram ben, pointing out (vol. III, p. 73) that ainda hoje se usa
de que com o valor de porque. A personal pronoun can be the direct object of gar (cf. CSM 16.33), but the conjunction e often
introduces a question (see the glossaries of Michalis, Mettmann). [This reading has been accepted by Vallin 1995, ad loc.]
2
Among the three cantigas d amigo of Pae Soarez we find dizia que (B 638 / V 239, vv. 1-2) and dizia ca (B 640 / V 241, vv. 13-14).
124
10
15
16 auia V
125
10
15
20
126
10
5 mizo V
7 <O> Nunes
haralhades V
8 quero] q to V
] e B : a V
127
128
2 e qui BV
3 boa B
10
15
129
10
15
20
130
3 moirer moyrer B1
131
132
The ess of the MSS could represent a prior misreading of est and in v. 4 e could have been e (an e with a horizontal stroke
above), an abbreviation for est (e. g. B 792 / V 376 [Talaveira], v. 17, where e in B corresponds to est in V while both MSS have
est at the corresponding place in the refrain in the first two strophes).
2
The two cantigas d amigo of Pero Garcia Burgals, preceded by his name, follow this text in both B and V.
133
10
7 u ir Nunes1 : oyr BV
8 des] de V
We find o (the normal spelling in medieval Spanish) instead of u (< Lat. ubi) in the MSS in B 683 / V 285 (J. S. Coelho), v. 2,
but both cases could be scribal errors rather than graphic (or phonetic) variants. Blasco prints non foy oyr a vya, which
allegedly would mean that the boy did not hear or failed to follow instructions on how to get to the tryst.
134
10
15
Along with the refrain, three rhyme words from this text veestes (4), dia (9) and fezestes (2) appear (in rhyme) in B 474
(= CEM 4, vv. 1-3), apparently a parody of our poem.
135
10
Nunes first suggestion (printed in his text, and accepted by Tavani) scans badly if at all, does not provide an exact rhyme,
and results in awkward syntax (requiring an abrupt stop at the end of the verse). The restoration of the final <e> of morte
supplies the missing syllable and the syntactic link between vv. 9 and 10. For the enjambment ai / filha cf. B 1243 / V 848
(Padrozelos), vv. 13-14 ai / madr and B 718 / V 319 (Tenoiro), vv. 3-4 meu / amigo.
136
10
15
20
6 disera BV 9 vee-lo Nunes : uelo V : Volo B 12 guarria Nunes 13 mandei Nunes : Mando B
14 pr V : Pre B del] dele BV 19 mandei Nunes : Mandoy B : mandoy V
ent B : enc V
Monaci reads moiria, but I can make out nothing between m and ia. B very frequently has ir where V has rr (and for that
reason I have not noted such variants in the apparatus), but here, if in fact the manuscript was legible when Monaci
inspected it, we have the reverse.
2
Tavani prints mando-vo-lo, but cf. v. 19, where mandey has been misread as mandoy (enton confirms the tense). The mother is
referring to an action that took place the other day (v. 1). There are no finite verbs in the present tense in this poem.
137
10
138
8 poi-l<h>o Nunes
10 Rog(u)ey B
11 vos Nunes
Cf. for example B 646 / V 247 (Torneol), v. 4 E morrer por mi, tant coitado.
Monaci reads morte, but I cannot make out an r. This aside, BV give us Por que de morte me quer gram ben de coraon but how would
we explain de morte? Tavani prints Porque morr e me quer gran bem de coraon (cf. B 651 / V 252 [Camans], v. 2), which makes sense
and cuts a syllable but maybe the wrong one since de could have come from a misreading of ele (monosyllabic here).
2
139
cnydar ? B : euydar V
Tavani (Rep. p. 287) puts this text among the frammenti non classificati and gives no rhyme sound for the last verse. One would
expect a pattern of the type abbaCC (the most common in Galego-Portuguese lyric). Were that correct, <d ir> seems a
possible supplement (assuming the sixth verse had the same syllable count as the others). Although I find no exact
parallel for triste d ir, a similar construction occurs in B 1243 / V 848 (Padrozelos), vv 1-3 que vos non pes // d ir.
140
AIRAS CARPANCHO 1
10
Minervini follows Nunes and eliminates vos ai; Pellegrini (1930: 311) would keep them in, though they seem to add two
extra syllables (all the other verses have twelve). It could be argued that the extra syllables are an intentional stylistic
device, since they occur in the first verse of the first text in the set of cantigas damigo by this poet (cf. B 641 / V 242 [Torneol],
vv. 1 and 4).
2
In none of the other eight occurences of con in the cantigas d amigo of Carpancho do we find the form cun.
3
Pellegrini, who defends the reading of B, translating assai mi rallegro di voi, per la molta ragione che ho di dirvi
(accepted by Minervini), also suggests reading per (though the abbreviation pr never represents per). But con vs is odd (and
vs is usually not written u9) and the asyndeton is harsh (hence Nunes pois). Perhaps the reading had been conu9q (i. e.
convosquo) suggested by Nunes in his rejected text (vol. II, pp. 82-83, where he had indeed made a botch of the strophic
form [his revised text is in vol. II, p. 466]). We would be more tempted by Gran ben m vosco, per muit ei que vos diga were it
not for the pattern of pauses in this poem, whereby each verse divides into 5 + 5.
4
Monaci reads ea but the letters are now barely legible in V due to corrosion.
141
AIRAS CARPANCHO 2
10
The verb should be seer as at the end of this verse () and in the refrain (for).
142
AIRAS CARPANCHO 3
10
15
8, 13 moyragura BV
Minervini prints agura in vv. 8 and 13, but maintains agora here, while Nunes opts for agora in all three places. The word
does not occur elsewhere in the cantigas d amigo of Carpancho.
143
AIRAS CARPANCHO 4
10
15
20
v. 13 qual i quissese: The expression fazer ben i is common (e. g. B 932 / V 520 [Roi Fernandiz], v. 20 Madre, faredes
me ben i) and seems to make sense, but we could read qual eu quissese here too (cf. v. 7).
vv. 8, 14 o meu: used in this way, this expression is unparalleled in the cantigas damigo as a way of referring to
a boyfriend (B 1174 / V 780 [Bolseiro], vv. 4 and 10 are not parallel since we easily understand the substantive
amigo from vv. 3 and 8 respectively). Further, o meu would be strangely otiose here as reiteration of an already
understood grammatical subject. If, then, it does not refer to the amigo mentioned in v. 2, o meu would seem
144
rather to refer to another boy, and the malicious sense of the poem would be that the girl dare not tell her
boyfriend about her other boyfriend, whose death would be a terrible injustice. Editors, however, have seen no
such meaning, though it must be stressed that they have had difficulty seeing any at all. Sense and syntax
may be intentionally riddling (compare the arguably ironic tone of the girl towards her amiga and potential
rival? in B 656 / V 257 [Carpancho]).
1
Nunes (vol. III, p. 87) says this is one of the most obscure cantigas and Lapa agrees. (Nunes takes several stabs at vv. 7 and
13 and Lapa (1982: 162) thinks a strophe is missing between the first two, a suggestion rightly rejected by Minervini [pp.
59-60], who nevertheless acknowledges reali difficolt d interpretazione [p. 58].) Indeed, vv. 7 and 13 are unintelligible
in the transmitted text: Que lh eu [lh i in v. 13] quisesse ben de coraon is senseless in context (Nunes suggests that que is
conditional; Minervini prefers optative [pp. 60-61]). My suggestion considers quelheu and quelhi as errors for qual eu and qual i
and the result is a cantiga ateuda at a fiinda whose interstrophic syntax seems to underwrite the correction.
2
Lapa (1982: 162) bravely and I think rightly restores the older valver (< u a l u e r i t). At any rate, the future subjunctive valer
would have still had an open e as opposed to a close e in the infinitive. See B 820 / V 405 (Pardal), v. 13 and note.
145
AIRAS CARPANCHO 5
10
15
146
AIRAS CARPANCHO 6
10
15
v. 13: The refrain changes slightly: se poder instead of mi quer ben (vv. 3 and 8).
1
This excision could be supported by comparison of this strophic form with those found in six other texts. See note on
B 1171 / V 777 (Bolseiro).
2
Minervini declines to include mi in the refrain, despite the evidence of v. 14; Nunes would take out madr. Both decisions
seem geared to get ten syllables out of the verse, so that its measure corresponds to that of the verses in the body of the
strophe. If that is our aim, better to cut vs.
147
AIRAS CARPANCHO 7
10
148
AIRAS CARPANCHO 8
3 me B
4 <bon> Michalis1
5 e] por
Minervini declines to accept the emendation of Michalis, but see V 1004 (= CEM 172 [Vinhal]), v. 5 que bon tempo faa.
149
VAASCO GIL 1
2 pard B
3 Irmaa B : irmaa V
Short lines in BV. Lapa (1929a: 325) sees long verses in the body of the strophe. The refrain could be split into
two short verses, yielding the form aaBB.
150
10
15
20
19 dica V
22 en de ? B : eu de V
Some scholars assign this text to Vaasco Gil, to whom the poem just before this one is attributed in B and V. In V, Colocci
has written Don Joham da voyn above this text, but in B he has written Dom Joham Davoyn in the right hand margin beside the
next cantiga (665 bis). In C, next to the number 665, we also find Dom Joham da voyn. Gonalves (1976a: 59), who would give
this poem to Gil, argues that the confusion is due to the repetition of the number 665. On stylistic grounds, the attribution
to Avoin seems certain.
151
10
15
20
Whereas come occurs in this text in vv. 3, 6 (and so 12, 18, 24, but these are not copied out in the manuscripts) and 15, como
occurs only here. The phrase come vs a min ought to be identical here and at v. 15. Elsewhere Avoin uses como in comparative
clauses, but here the clause is implicit and come governs a simple pronoun. Cf. Michalis, Glossrio, s. v. come. Note also how
the (older) Toledano manuscript of the Cantigas de Santa Maria preserves come against como in e. g. 74.32 come To : como TE
(elsewhere T against E, e.g. at 85.49 come T : como E).
152
10
9 cuydades V : cuyda B
11 ueiria B : uema V
Nunes would take out mi to remove a syllable, but we might instead consider aphaeresis of the initial vowel of enton after
the tonic final dipthong of jurou (jurou nton).
153
154
9 pardon B
11 coma V
22 quero] qre BV
10
15
20
10
15
20
155
10
15
20
156
10
15
20
11 eu del. Nunes
157
10
15
20
158
vv. 4, 6, 9-10: por causal (relativo s razes rejeitadas) nos vv. 4 (porque no podes falar comigo) e 9-10
(porque no fao tudo o que uma mulher faz pelo
homem que ama), mas final no v. 6 (para me fazeres
sofrer).
10
15
20
16 tornades V
Lapa (1929a: 324) sees the strophic form aaBB here, that is, long verses with internal rhymes in the body of
the strophe, and short verses in the refrain. The text could also be printed as aaB with internal rhymes in all
three verses.
159
10
15
20
An example of the mechanical (erroneous) recopying of the beginning of a refrain (cf. Parkinson 1987).
Though Avoin does not use this common formula elsewhere in his cantigas damigo we find it several times in his other
poems: A 157, v. 15 ben o sei eu; V 1010 (=CEM 222), v. 17 mais ben sei eu quen troba ben ou mal; V 1011 (=CEM 223), v. 5 e ben sei
eu por que aquesto faz. Among the cantigas d amigo we find the formula in: Johan Garcia de Guilhade (B 754 / V 357, v. 10; B 755
/ V 358, v. 2), Pedr Eanes Solaz (B 829 / V 415, vv. 27 and 29), Johan Airas de Santiago (B 1026 / V 616, v. 13; B 1031 / V 621,
v. 19; B 1041 / V 631, v. 19; B 1043 / V 633, v. 2; B 1047 / V 637, v. 13 [and cf. V 594, v. 14]) and Pedr Amigo de Sevilha (B 1218
/ V 823, v. 23)
2
160
10
15
20
2 quer] quei V
8 a] ia V
15 pr B : pre V
23 cado B
161
10
15
20
Monaci (1873: 32; 1875) reads uehesse o in V, but the final letter of the first word is now a blur.
162
10
15
v. 1: mui fremosa modifies the first person subject, pace DHeur (1975b: 363-364, 377 and esp. 428-429, note 18),
who suggests that it may be vocative (but cf. v. 10 of the next poem). More pertinent: Do we read fremos e
sanhuda? Or , if not, do we take sanhuda as substantival (a very pretty sanhuda [cf. Cohen 1996b]) or consider
fremosa sanhuda all together (fremosa being the substantive, sanhuda adjectival) with mui qualifying the whole
expression (very much an angry beauty)?
1
The chiastic construction ei perdud ... perdud proposed by Lang (metrically acceptable if we read el<e>; cf. B 689 / V 291,
v. 11) seems to have been overlooked by Nunes. For word division in rhyme (far more common in other genres than in the
cantigas damigo, where it is quite rare), see B 685 / V 287 (J. S. Coelho), v. 9 viss e; B 1092 / V 683 (Elvas), v. 10 lum e; B 1119 /
V 710 (Berdia), v. 12 ant o (only this last is certain).
163
10
164
5 Quis B : Que V
10 quant<o> oj eu
10
15
20
16 mh] rih V
165
10
15
166
5 moirer B : morre V
9 dizr B : diz V
10
15
ueer B : ucer V
2 ousarey B : usarey V
v. 6 : The syntax of Por en qual (Poren ql BV) coita is explained by Nunes (vol. III, p. 105) : Por [a] coita en [a] qual
mia madre mi ten.
167
10
15
7 haralhar V
14 el e
The last letter is blurred in B, but appears to be an o. Cf. B 649 / V 250 (Burgals), v. 7 and note.
See the variants in CSM 295, v. 13 (fiz E : fix F) and compare the rhymes in CSM 265, vv. 125-127 (fix/quix/dix). A similar
problem in B 1142 / V 734 (Servando), vv. 8-10.
3
Nunes prints ele, which is acceptable (cf. B 678 / V 280, v. 4 and B 689 / V 291, v. 11). But the conjunction e frequently begins
a question (cf. the glossaries of Michalis and Mettmann) and arguably introduces an indirect one here.
4
This is the only occurrence in the cantigas d amigo of salvar-se. Cf. Lang (1894: 122) and Gonalves (1991: 28-29). The reading
lhi would seem to be justified by (e.g.) CA 29 (Somesso), vv. 11-12 E non osm og eu, nen o sei / per que me lhe possa salvar. But lh i
could be defended both by the emphatic i in the refrain (though the MSS have hy in v. 5 and hi in vv. 11 and 17) and by CA
178 (J. S. Coelho), v. 13 Non me sei n d outra guisa salvar and B 71 (Calheiros), v. 10 Porque o fiz, non me poss n salvar. We might
also read lh i in vv. 4 and 14; cf. CA 167 and 168 (J. S. Coelho), vv. 10 and 21: lh i faria.
2
168
10
15
5 boa BV
10 xera B : sera V
14 nen] ueu BV
me9 B : mr9 V
169
10
15
20
170
10
15
Though aguis in v. 3 might seem irrefutable (and our poet uses the verb aguisar at CA 175.1), Nunes emendations in vv. 9
and 10 should probably be rejected (and we might even correct v. 3 to read lhi guis eu). Whereas guisar is common, aguisar is
scarcely documented in the entire corpus of cantigas damigo. I find four other apparently certain examples: (1) B 265 (Queimado), vv. 7- 8 E, por que non quij eu con el falar / quand el quisera nen se mh aguisou; (2) B 582 / V 185 (Dinis), v. 14 quand aguisastes
que per nulha ren; (3) B 1199 / V 804 (Caldas), v. 7 se mi Deus aguisar de o aver; (4) B 1269 / V 875 (Cangas), vv. 7-8 Serei vosc en San
Momede do mar / na ermida, se mho Deus aguisar. B 1148 / V 740 (Servando), v. 2 e lho Deus aguisa polo seu amor is hypermetric and
guisa should be read. Cf. B 1168 / V 774 (Bolseiro), v. 7, and note. See also B 1364 / V 972 (M. Soares), v. 8 where desaguisado
is hypermetric.
2
Formulas of this kind (from the Latin sic plus optative subjunctive) are found in Galego-Portuguese lyric with se, si, ase, asi,
asse and assi. Mettmann, s.v. asse, notes that this variant is found only in optative clauses. Cf. CSM 24.39 asse Deus vos perdon
where the Toledano has ase.
171
10
15
172
se <vs> Nunes
8 iredes] uedes V
173
10
174
10
175
soo BV
The text is copied on four short lines in BV. Long verses are recommended by DHeur (1975b: 429).
176
10
4 senor BV
8 przer B : przeri V
~to B : qnta V
9 qu
10 E B : a V
177
ESTEVAN REIMONDO 1
10
15
20
The past tense desejei (maintained by all editors) is out of place in this text (where there are no other past tenses) and the
precise parallel eu...desejo in v. 11 supports the correction. See B 1290 / V 895 (Requeixo), v. 5. (Nunes [vol. III, p. 114; cf. p. 68]
shows discomfort with the perfect, which he fails to explain convincingly.)
178
ESTEVAN REIMONDO 2
10
15
3, 9, 15 migo scripsi (cf. v. 6) : comigo BV, edd. 7, 11, 13, 17 trist<e> o Nunes (cf. v.5) : tristo BV
Meter uncertain in vv. 1, 3, 5 and 6 of each strophe. As printed here, the form is AbAbAA, scanning 8 9 8 9
8 8.
179
180
15 Peder V
17 mi~ BV
10
15
20
10
E, se eu end al soubesse,
que nunca lhi ben quisesse
181
10
15
182
10
15
The change is unnecessary since pud is a variant form of the third person perfect indicative (cf. Mettmann, s.v. poder).
The emendation fiz might seem attractive in view of vv. 1, 10 and 16, but the point here seems to be that the girls
senseless behavior caused an equally senseless reaction on the part of the boy.
2
183
10
co<n> Nunes
5 renho V
9 est<>]
v. 9 is presented on one line in V. The rest of the poem is copied on short lines in both MSS. Lapa (1982: 163)
recommends long lines, as printed here.
184
10
9 faz2] far V
10 outr olho
Short lines in BV. Lapa (1982: 164) sees long verses, at least in the body of the strophe.
v. 10: se rostro outr olho cata: Presumably, if (his) eye
gazes at another face.
Nunes (vol. III, p. 119) thinks lho unmetrical and senseless, but there is no lho if we divide correctly (cf. B 637 / V 238
[Sandin], v. 3 [and note], where the same word has fallen through the cracks).
185
10
15
7 <e> supplevi
vv. 5, 10, 15: <meus>: the first verse of the intercalated refrain counts 11 syllables. Presumably, then, vv. 5, 10
and 15 should be of the same length (cf. B 697 / V 298 [Ulhoa], vv. 10-11).
1
Cf. for example CA 243 (Talaveira), v. 14, where A correctly has fezo but B offers fez, leaving the verse a syllable short.
186
10
15
16 eu V : en B : en Nunes1
187
6 gisa B
For the strophic form (botched by Nunes, who prints an impossible aaB with no attention to meter or rhyme),
see B 627 / V 228 and B 628 / V 229, both by Calheiros.
1
Nunes maintains veer at the end of vv. 2 and 5, but prouguer (with open r) does not rhyme with veer (with close er).
The adverb i, present in both MSS in v. 8 , is missing in v. 4 (where it has already been restored by Machado) and, I believe,
here. It adds the wanting syllable and provides a rhyme (wholly lacking in Nunes) for vi. The place where the girl wants to
go is also mentioned in v. 2 (u meu amig).
2
188
10
15
20
Though en quanto can be construed with the indicative (or the present subjunctive), sense and syntax seem to require the
future subjunctive here. Closely parallel are B 749 / V 352 (Guilhade), v. 2 falade sempr u~as outras / en quant el falar comigo and
B 783 / V 367 (Afonso Sanchez), vv. 2-3 en quanto lh eu preguntar u tardou, / falade vs nas donzelas enton.
2
Elsewhere among the cantigas d amigo we find co el in the MSS at B 699 / V 300 (Ulhoa), v. 4, B 730 / V 331 (Beesteiros), v. 4
and B 1022 / V 612 (Johan Airas), v. 3. Nunes restores the final nasal in all four cases.
3
Nunes is probably right to add eu here. See note to B 674 / V 276 (Avoin), v. 9.
189
10
15
9 sedes V
10 muyt B : muyr V
v. 4: O sentido dos vv. 3-6 parece ser: Quando ficas
durante muito tempo noutro stio, sem notcias minhas, meu amigo, ests triste ento?.
It is not clear that quanto ever functions as equivalent to quando. It is found several times in the manuscripts where we
would expect quando, but apparently this is due to faulty transmission. Cf. Rodrguez 1980: 200 (note ad v. 5) and the
glossaries of Michalis, Lapa and Mettmann. Here, given the parallels with vv. 9 and 14 and the awkwardness of explaining
muito in the same verse, the case for quant (which Nunes prints in vol. II) is weak.
190
10
15
20
vi~ia
outra (outr BV [the abbreviation, as countless comparisons between BV and A will prove, represents ra not ro]) modifies
coita (v. 17) retroactively (and the phrase nulha coita non avia / ... / outra is semantically parallel to por al coitado in v.11).
191
10
15
20
25
192
10
15
20
In the MSS all verses except 7, 8, 21 and 22 contain eleven syllables, with a caesura after the seventh, so that the last four
syllables form a rhythmic group. In v. 7, which seems a syllable short, the suggested change is easily justified. In v. 21 the
extra syllable is syntactically unecessary and can be cut. Here, the caesura would have impeded the elision (graphically
represented in the manuscript transmission) of amigo and e.
2
Given the verse structure in this poem (see previous note), the triple occurrence of no coraon in vv. 4, 10 and 16, and the
ease with which a verse-initial e could disappear, the addition of the conjunction seems preferable, rhythmically, syntactically
and paleographically, to the modification suggested by Nunes. [This supplement has now been printed by Vez.]
193
10
15
20
But there is a dot in B after sempre (which might represent the lost til).
The article is not needed and is often omitted with the word morte, as noticed by Lang (1894:115-116).
194
10
15
20
25
195
I see no good reason to regard this as a real variation in the refrain (when Vinhal wants to show virtuosity he does it in
rather more spectacular ways). The MSS me could be an error for mo (= m o) or the tradition could once have had me o (as
elsewhere), from which the o dropped out.
2
The abstract noun ira needs no article and the subjunctive sena stands better unelided. The correction of senta to sena
(Nunes) is a question of historical grammar (sena < s e n t i a t).
196
10
197
10
15
The evidence of BV for the refrain, here triple what we would normally find, is two to one against desguisado. For rhythm and
for ironic force guisado should be read (against the textus receptus).
2
For the expression tornar a alguen in the cantigas d amigo cf. B 1239 / V 844 (Padrozelos), v. 17 a min vos tornade por en (where
tornar is reflexive); see also CSM 5.180-181 nunca quis / a dona tornar a el; 151.19 tornou a ela; 335.66 tornaron a ele. For the
opposition partir / tornar in rhyme cf. B 705 / V 306 (Cogominho), vv. 14-15. The corruption can be explained as follows: the
reading had been trnaua (cf. B 868-870 / V 454, where both B and V have trney in the third verse of the last strophe); the initial
t was misread as an r (cf. B 780, v. 10, where the MS offers rraz instead of traz); the superscript r was misread as ro (instead
of or); and naua was misread as uana (given the nearly identical form of u and n [cf. B 719, v. 20, where B offers toruey while V has
the correct reading, torney]). Hence trnaua became (in the source text of BV or in its exemplar) rro uana. We could read (closer
to the MSS) e que tornava (and he was on his way back), but a que tornava (to whom he was returning) makes better sense.
198
ROI QUEIMADO 1*
10
15
B 265 f. 68v
2 qer B
asanhou B 4 u~a ] nuha B
8 quand] qud B
mha guyson B
14 asanhou B
6 ssa sanharia B
10 meu] men B
199
ROI QUEIMADO *2
10
15
200
14 en outra
ROI QUEIMADO 3
10
15
20
5 barato B
10 bem B : tem V
15 catar (ne~ ) B
19 m el Nunes : mal BV
bom V
201
ROI QUEIMADO 4
10
15
202
10
15
20
21 nunca om. B
Here and in v. 8 (and thus in the corresponding verses of each refrain) Nunes emends to bring the syllable count up to
nine. Tavani (Rep.) rejects this (168: 4, 5).
203
204
8 sedes B
10
10
15
15 ma<i>s Michalis
Among the cantigas damigo we find atan gran pesar in B 737 / V 338 (Seabra), v. 14 and atan gran ben in B 559 / V 162 (Dinis),
v. 8, while in B 932 / V 520 (R. Fernandiz), v. 6 we have faredes mi atan prazer. If <a>tan in v. 5 is correct, we might also venture
mh <a>tan in v. 6.
2
The first strophe is found in both MSS among the cantigas d amor, attributed in V and C to Tenoiro, but in B (f. 89 r, before
no. 398) to Affonso Fernandez Cebolhilha (cf. Marroni 1971: 272-3). In V this folio is among those whose number has been
crossed out. In the modern pagination of V it is to be found on p. 7v.
205
10
15
20
19 te~ho B : thV
v. 20: The sense is since I entirely failed to reciprocate. This use of tornar seems to correspond to one of
the senses given by Mettmann (s.v. tornar): restituir,
devolver.
20 hypermetric1
Nunes, perhaps rightly, leaves the line as in BV (but suggests taking out por [vol. III, p. 139]), and Tavani (Rep., 160:179)
does not comment.
206
ESTEVAN COELHO 1
10
The struck lines (within which two words were cancelled before the whole strophe was crossed out) read as follows:
Par ds de cruz dona sey q dades
(q dades)
damor mui coytada q t ben cantades
cantigas
The only difference (other than the cancelled q dades) between this version and the one copied in the correct place in V is
that in the second version tan is spelled out (in B we find t).
207
ESTEVAN COELHO 2
Se oj o meu amigo
soubess, iria migo;
e<u> al rio me vou banhar<e>,
al mare
Se oj el este dia
soubesse, migo iria;
eu al rio me vou <banhare,
al mare>
Quen lhi dissess atanto:
ca ja filhei o manto;
eu al rio me vou <banhare,
al mare>
10
v. 10 ja filhei o manto: Apparently, Ive already grabbed v. 10 ja filhei o manto: Aparentemente, J agarrei (ves(put on?) my cape, suggesting that the girl is eager ti?) a minha capa, sugerindo que a rapariga est pronand ready to go.
ta e ansiosa por partir.
v. 10: Cf. Frenk (1987: 879), no. 1811: Teresica, daca mi manto, / que no puedo estar encerrada tanto; and, among the
variants cited in her critical apparatus: Toma tu manto y daca mi manto, / que no puedo estar en casa tanto.
208
ESTEVAN TRAVANCA 1
10
15
Nunes is most probably right to emend (cf. B 801 / V 385 [Sandeu]), vv. 5-6 que ser / do meu amig ou que far; CSM 6.63-64 d
me meu fillo morto / ou viv ou qual quer que seja), even though the transmitted text can be defended (as viable grammar; whereas
the question is rather a stylistic one): if he was dead, or what he will do [if he is] alive (viv o que far).
2
Given the presence of amiga in vv. 1, 4, and 7, Nunes supplement might seem attractive, but is unmetrical (vv. 1, 4, and 13
all scan 9, corresponding to 10 syllables in all other verses). Machado is on the right track (cf. v. 4 por Deus, and v. 10 por
Nostro Senhor), but the construction is more likely to be preguntade (v. 13)... /// e, <por Deus>, falade o comigo. (B 1022 / V 612
[Johan Airas], vv. 15-16 e moiro por or / novas del, e preguntade, por Deus, while closely parallel, does not help decide where to
add the exclamation). Of falade, Nunes says (vol. III, p. 142) that here o verbo falar, como est construdo, quer dizer: fazer
que alguem fale com outra pessoa but gives no parallels, nor are any to be found in the cantigas damigo or in the glossaries
of Michalis, Lapa or Mettmann. Lapa rejects Nunes explanation and suggests that falade-o comigo means simply dizei-mo, comunicai-mo.
209
ESTEVAN TRAVANCA 2
10
15
20
25
210
fiz and quix would seem to form an imperfect rhyme, so we might want to correct fiz to fix (but would have to do so twice in
the sixth verse of each strophe, against the MSS). Cf. B 683 / V 285 (J. S. Coelho), vv. 5-6 (and note) and B 1142 / V 734
(Servando), vv. 8-10. Possibly the rhyme became acceptable during the third quarter of the thirteenth century, when our
poet was active.
2
We would expect the change of subject to be marked, hence Nunes fezesse <eu>, but hiatus here (fezesse | eu) seems unlikely.
Presumably, que o <eu> fezess e, por Deus, que lhi perdoasse reports in indirect speech what would have been independent
optatives in direct speech: que o faades! e, por Deus, que me perdedes! Lapa (1982: 165) offers alternative suggestions for v. 16.
3
The rhyme-scheme of vv. 25-28, the lack of a link between them and the refrain, and the conclusive sense they provide all
suggest that they are fiindas. Cf. Tavani, Rep. (160:339) and Parkinson 1987.
211
ESTEVAN TRAVANCA 3
212
ueo BV
10
ESTEVAN TRAVANCA 4
10
My suggestion, <ben> sei, assumes another occurence of ben in a poem in which the word is used three times in the first
strophe and twice in this one (not counting the supplement). This repetition is due in part to the dobre singulars (ben in
rhyme in vv. 1 and 4; sei in vv. 7 and 10).
2
The change would regularize the meter, but see Introduction, 5.
213
10
15
11 no om. B
We find the adverb ora in vv. 9 and 14, so it is absent (in the transmitted text) only in the first strophe. Thus, since this
verse is a syllable short, the emendation seems justified on metrical, stylistic and semantic grounds.
2
The article is not needed and with it the verse scans a syllable too many. Cf. A 126 / B 241 (Fernan Garcia Esgaravunha),
v. 9, where A has De querer ben outra molher and B adds the (incriminating) preposition a: De querer ben a outra molher.
214
10
15
20
25
No previous editor has the right reading here, required by sense and meter (and present in both MSS). Nunes prints
moir<o> en (recently accepted by M. Ferreiro) but also suggests (vol. II, p. 468 and vol. III, pp. 146-147) <el> morren, thrice
ignoring the obvious. Note the future verbs in this strophe (farei, verrn, direi, outorgar mho an, direi). The construction (pois +
future subjunctive in the subordinate clause, future indicative or equivalent in the main clause) is common (for examples
with or, cf. CSM 8.8, 178.8, 198.8, 332.58, 337.6).
215
10
15
In his glossary (s.v. que) Nunes cites two examples (including this one), both conjectural, of en que introducing concessive
clauses. The MSS have e que in both texts (cf. B 821 / V 406 [Pardal], v. 7).
216
10
15
B 368 f. 84r
3 i] ay B 4 sofredes; e por n Michalis (I, 841=CA 428) : sofrer des esporen B 5 vassal<o> Michalis
(cf. vv. 11, 17) : uassal B 6 sofre Michalis : sofrer B 7 e seclusi 8 amig<o>, e Michalis : amigue B 10 mui
ben scripsi : muytem B : muit n Michalis 13-15 = A 237 (Guilhade), vv. 7-9 (but meus olhos coita in v. 8)
13 reem B 16 ffeyco B
pos B
217
10
15
218
10
15
Cf. Nunes, vol. III, pp. 148-149. In both MSS this verse stands on the next line after sigo (normally fiindas are preceded by a
blank line).
219
10
15
20
10 de om. B
16 en] eu B
The form disso does not occur elsewhere in Amigo, but see the glossaries of Michalis, Mettmann, Lapa. In A 255 / V 428 /
B 842 (Charinho), vv. 5, 11, 17 we find disso A : disse BV.
220
10
15
5 corrupt?2
8 be~ B : he~ V
Monaci (1875: 433) comments: dopo questo [v. 1] deve mancare un altro v. que forse terminava colle parole ben sey.
These words begin the second line; in V the line ends with auer while in B pod auer stands alone on the third line; in neither
MS is there any indication of a lacuna. Braga suggests (and Nunes also prints) <faa lhi quanto bem quiser>; drawing on v. 11.
If we look instead to v. 6 we might think of something like <punhen de lhi fazer ben, mais> ben sei (cf. B 1215 / V 820 [Sevilha],
v. 9 punhen ora de lhi fazer prazer).
2
vivr (with open e) does not rhyme with aver (cf. B 1258 / V 863 [Galisteu Fernandez], v. 5 and n.). Thus, unless we posit
parodic intent and a unique strophic form, we must assume a corruption. If fazer in v. 6 and faan in v. 11 can be taken as
clues to the word used in rhyme at the end of the refrain, it would have been fazer. If this is correct, fazer ben in v. 6 may be an
exact inversion of a final ben fazer in v. 5. But it is difficult to see how this end can be connected with the beginning of the
verse (se non viver migo), assuming at least that much is genuine. A verse ending, for instance, con ben fazer or a meu prazer
could either have been banalized or toned down by a jograr or meddling scribe. The phrase con ben fazer occurs twice in CEM,
while in B 925 / V 513 (Gomez Garcia), vv. 5-6 (refrain), we find por quant andou al sen meu prazer, / que ande un tempo sen meu ben
fazer. The phrase a meu prazer occurs in B 1179 / V 784 (Bolseiro), v. 25; tod a meu prazer in B 807 / V 391 (Froiaz), v. 16.
221
10
15
The verse seems a syllable short, but nen ar ei is sound (and Nunes averei wrong).
One of the few cases of double transmission among the cantigas d amigo offers us a neat parallel for the omission of ren in
just such a construction. In B 1023 / V 613 (Johan Airas), v. 21, we have que ambos ajan ren que vos gracir whereas in B 1049 / V
639 the word ren is missing in both manuscripts. Cf. also B 591 / V 194 (Dinis), v. 20 non ei ren que vos i gradecer.
2
222
PERO VIVIAEZ 1
10
15
Beltrami, printing en cos, says that no cos is not attested elsewhere. To be sure, en cos is far more common, but we find no cos
(in another sense?) in B 1619 / V 1152 (=CEM 405 [attributed to Caldeiron in B and C, to Pero Viviaez in V]), v. 14.
2
Editors have sought the missing syllable in the second half of the verse, but <i> can be justified on two grounds. First, in
every verse except this one (did we not emend) there is an accent on the fourth syllable. Second, the phrase bailar <i> in
this verse would echo the final words of the refrain (bailaremos i).
223
PERO VIVIAEZ 2
10
15
9 falem9 B : falm9 V
In both MSS there is a loop, not a cross, on the tail of the p, an abbreviation for pro, so that sempr<o>uve (Nunes) is
unnecessary and sempre <o>uve (Beltrami) worse.
224
10
15
In v. 15 there are problems with sense and meter, suggesting a corruption. If we are to understand, let alone corrrect this
verse, we must first ask whether fora is here an auxiliary from ir or a form of seer. Either would do, since sense would be
satisfied either by he would have killed himself (cf. B 563 / V 166 [Dinis], v. 14] ca sei que se fora matar) or by it would have
been just as good for him to kill himself. If the text of BV is correct at this point, ben quanto would have to mean tantamount
to. But I find no parallel for the construction seer ben quanto + infinitive. In defense of BV we could cite V 1010 (Avoin), vv. 67 ben tanto sabes tu que trobar / ben quanto sab o asno de leer where ben quanto is used in a grammatically explicit comparison, and
perhaps an elliptical form of such a comparison would be admissible here, but the additional syllable suggests that
something is amiss. If the problem lies in ben quanto we could solve it by supposing the reading to have been something like
ca, se non, fora melhor se matar (cf. melhor fora dar o rocin doado in B 1536 / V 409 =CEM 91 [Dinis], v. 19). Perhaps the intervention was
calculated to tone down the expression, found among cantigas d amor in phrases like de me matar fezera mui mellor.
225
10
15
d9 BV : destes Nunes1
Nunes may be right, but a spatial reference seems lacking, along with a syllable.
226
8 mhas] mays BV
14 <e> addidi
10
15
6 ueer V : uer B
7 el] ele BV
10 non] en B : p n V1
15 en (casse) B : eu V
Before n there is something like a p in V, but I hesitate to say what letter, if any, it is.
227
10
I believe that o meu amigo is either a gloss or has crept in because the verse had become short, and that des que s el foi nunca
vi prazer is a probable reconstruction of what had been there. This in turn could have arisen from the shortening of ele to el
and the loss of <ar> (cf. B 1213 / V 818 [Sevilha], vv. 2-3 ca se foi daqui / mui meu sanhud e nunca o ar vi). Reading as I would (des
que s ele foi, nunca <ar> vi prazer) we may observe that in each verse in the body of the strophe there is an accent on the 6th
syllable and the 7th syllable is an unaccented final a: fremosa in vv. 1, 6 and 10; romaria in v. 2; nunca in v. 5; and leda in v. 9.
228
10
15
But in the second strophe both have hirey on the first line of the refrain. In the third strophe B breaks after leiras while V
again puts hyrey on the same line. The body of the strophes is also copied irregularly. V presents vv. 1-2 as here, while B
breaks after hunhas. Both MSS present the second and third strophes in short lines, with breaks as follows: II nouas / sabor /
amigo / for / ; III nouas / prazer / amigo / ueer /. In the last strophe B breaks at: nouas / molher / estas / ueher while V has long lines
as here.
229
10
15
The Machados print uos (= vs) here and in v. 14, but the reading of V is clear in both places. Confusion between n and u is
notorious in the sections copied by this scribe. In this text we find ua morado (na morado in V) in v. 2 and u (n in V) in vv. 5
and 8.
2
We do not elsewhere find con sigo in the cantigas d amigo of Guilhade and so have followed the suggestion provided by
Nobiling. In his text, however, he reads pero to eliminate the extra syllable.
230
10
15
20
231
10
15
20
232
v. 14: ben come quen quer significa como qualquer outra pessoa.
10
15
233
10
15
234
10
15
tal <cousa> fits in all senses, and the euphemism is well documented, e.g. in V 1012 (=CEM 229 [J. S. Coelho]), vv. 3-6 nunca
vos eu tal cousa negarei / qual oj eu ouo pela terra dizer: / dizen que fode quanto mais foder / pode vosso mouro a vossa molher and vv. 9-10 se
vos negass eu / atal cousa qual dizen que vos faz ... [with the same reference as in v. 2]. In the Crnica Geral de Espanha de 1344 we find
tal cousa referring specifically to rape: se tal cousa a my~ aveesse (ed. Cintra, vol. II, p. 307, line 3). Elsewhere (as here, if I am
right) the expression is used to refer to surprising, shocking or shameful actions.
2
Nobiling (followed by Nunes) changes the verb. Lapa (1982: 166), taking what Nobiling (in a note) rejects, would read quen
no ferisse! (which he glosses quem lhe batesse por esse atrevimento), but V often has r for x. The reading of B, fexisse, is
morphologically untouchable and provides sense and syntax.
235
10
Nobiling prints e direy mha coyta e coyta / que tragu e que maravilha. Nunes alters this slightly: e direi mia coit e coita / que tragu e
qu maravilha (with an unlikely elision of qu). Cf. vv. 2 and 6.
2
Aside from the difficulties of the transmitted text and the paleographical plausibility of the emendation, there seems to
be a rule in Guilhades cantigas d amigo (which would be broken only here, did we not emend), namely that the poet either
(1) introduces gran in the last strophe, or (2) if he has introduced it at any earlier point, uses it again in the last strophe.
(Those who, like Reckert & Macedo [1996: 93-94], would read with the manuscripts here, could cite B 364 [Cogominho],
v. 4 ca x i coita de coita...)
236
10
15
v. 15: Reading as I do (quant for quand with Nobiling) we can extract a plausible sense from this vexed verse:
which is both a reference to the rough language she used and a lament that she should have had to say
such things to him (for the kind of tone she might mean, see cabea de can in B 777 / V 360 [Guilhade], v. 16 and
B 787 / V 371 [Guilhade], refrain and v. 8).
1
We cannot tell whether the missing verse was the second or third in the strophe (there is no space left for it in the MSS).
Nobilings suggestion (relegated to a note) would solve the problem by correcting a single letter. (If we read quand with
BV, the second half of the verse would quote something the girl said to the boy.)
2
237
10
15
V 351 f. 56v-57r
6 huha B : hir ha V 10 pola BV ; corr. Nobiling (cf. v. 16) 14 ea moelinha B : eamoelinha V : e del. Nobling
mocelinha; interpunxi 16 moa al Nobiling (cf. v. 9) : mocahi B : moa hi V 17 qutas B : quta V
On whether this text belongs among the cantigas damigo or not, see Cohen 1996b: 27-36.
vv. 14-15 en pouca sazon / de parecer todas venudas son:
I take this to be a generalization, In a brief time all girls
are superseded in beauty, which in turn is used as
foil for the specific case: in a mere instant this girl
vanquished the beauty of all she came across.
238
This poem appears in BV in six-line blocks scanning 7 7 7 7 7 7 and rhyming abcbDD. Nor has Colocci
made any annotation in B that would suggest long lines. The text could also be printed in strophes of three
verses (15 15 15) rhyming aaB (cf. Tavani, Rep. 26:13; 37:16; 244:18 [alternative schemes are given for all 26
texts listed under no. 244]). The same question occurs in B 747 / V 349, but elsewhere in his cantigas d amigo
Guilhade has no unrhymed verses (assuming we accept Monacis obvious correction visse [vistes BV] in B 746
/ V 348).
239
10
15
Monaci gives the reading of V as moire but it seems to be morre with the first r imperfectly formed. B, as usual, has ir for rr.
Nunes retains el, which could anticipate o (v. 3) or be the subject of a clause that never materializes.
3
This second que (q BV) is superfluous (though Nunes defends it) and unmetrical (que o ought not be elided).
2
240
10
15
241
10
15
20
5 uoda V : noda B
6 boa BV
20 ui r B : ui t V (ui c Monaci)
22 eu] en V, Nunes1
n (V) instead of eu (B) would yield acceptable syntax, but the pronoun is stressed here in emphatic contrast to all the
other girls: todalas donas (v. 19) ... eu (v. 22).
242
10
15
7 tomaria V
11, 17 Eiurey
243
10
15
20
The sequence of tenses here (perd<eu> ... morre) can easily be defended by comparing: A 89 / B 193 (Burgals), vv. 28-29 a
por que eu moiro e por que perdi // o sen; A 205 / B 356 (Ulhoa), vv. 6-7 atal per que ei perdudo meu sen / e por que ei mui cedo de morrer;
A 269 (anon.), vv. 1-2 Senhor fremosa, ja perdi o sen / por vs, e cuido mui ced a morrer (= Tavani, Rep. 157, 57; attributed by Oliveira
[1994: 60-63] to Pardal); and B 406 / V 17 (Afonso Sanchez), vv. 20-21 ... per que perdi o sen / e por que moir assi desemparado.
244
10
15
20
245
10
15
20
246
10
15
In vv. 5 and 11 the reading of V is not clear (Monaci reads minde), but this is a minor quibble. Another question is whether
to print viinde (Nunes) or vi~ide (Nobiling, Lapa 1982: 168).
247
10
15
Nobiling defends the transmitted text for the wrong reason. Lapa (1982: 169-170) would emend da to da<r> here (though
elsewhere in his cantigas d amigo Guilhade only once juxtaposes words ending and beginning with r [B 743 / V 345, vv.5,11,17]).
The sense would be: I went right over to her (logo lh eu foi) at the end (na cima) to explain (da<r> razon).... If this were correct,
razon would not have the same meaning as in v. 2 (the argument of a poem).
248
10
15
20
249
10
15
20
But nen sol probably ought not be broken up (see sol in Mettmanns glossary).
avol: correcting Nobilings text (a v-lo), Lapa (1982: 170-171) detects here the Provenal adjective, documented three
times in CSM, which he glosses mau, ruim. Avoleza is registered three times in CEM, four times in CSM.
2
250
10
15
20
This strophe can be punctuated (read) in a variety of ways, and Nobiling, Nunes, Machado and Domingues all offer
slightly different versions. The initial e in v. 9 is taken as a verb by Nobiling, that in v. 11 as a conjunction.
2
On my reading, there is a play between sabudo and sabern. This assumes a banal error of t for b and the loss of a
superscript a. Cf. the critical apparatus of B 780 / V 363-364 (Ornelas), v. 3 (bem V : tem B); B 709 / V 310 (Vinhal), v. 19 (soute
BV : corr. Monaci) and B 713 / V 314 (Queimado), v. 7 (sabades V : satades B).
3
As Lapa points out in the glossary of CEM, tudo is a later form of todo. Registered once there, it is not in the glossaries of
CA or CSM. So I would reject Nobilings emendation (accepted by Nunes), as well as others which might seem attractive
(e.g. es<t > tudo) but assume that tudo is correct.
251
ESTEVAN DA GUARDA 1
10
15
Since the verse seems a syllable long and the object of servirei has already been given (A Deus), it is tempting to eliminate
the pleonastic o. On the other hand, some might want to scan servirei as two syllables (Cunha 1982: 102-103, 118, 147). But
it is just possible that the verse, the last by this girl, may be intentionally hypermetric; the next, the last spoken by the
other girl (within the body of the strophe), can also be scanned as 11 syllables if we observe hiatus at the juncture of amiga, e.
On hypermetric verses at Rv-1, see Introduction, 5.
252
PERO D ORNELAS 1
10
15
253
AFONSO SANCHEZ 1
254
AFONSO SANCHEZ 2
Dizia la fremosinha:
Ai Deus val,
com estou d amor ferida,
ai Deus val
Dizia la ben talhada:
Ai Deus val,
com estou d amor coitada,
ai Deus val
Com estou d amor ferida,
ai Deus val,
non ven o que ben queria,
ai Deus val
Com estou d amor coitada,
ai Deus val,
non ven o que muito amava,
ai Deus val
10
15
5 talhada V
cuytada BV
In the first two strophes both MSS present the intercalated refrain on the same line as the verse preceding. In the last two
strophes the layout is irregular and the MSS differ. Only at its second repetition in the third strophe of B does the refrain
stands on a line by itself.
2
Cf. Parkinson 1987 on false refrains. Nunes repeats v. 3 at the end of each strophe.
255
10
15
20
Lapa (1982: 171) points out that en tal introduces final (purpose) not consecutive clauses, but defends en, considering it an
adverb. Nunes (vol. III, p. 180) recognizes that the reading of V may be correct.
2
Talaveira uses the word torto twice in B 793 / V 377 (vv. 14 and 19). Though not found in Galego-Portuguese lyric (where a
torto generally fulfills the same function), cu~ torto occurs twice in the Notcia de Torto (see Cintra 1987: 21-72). The girl insists
throughout the poem that the boys complaint is unjust.
256
10
15
20
The sense seems incomplete, and if the final vowel of queixume is elided (as in v. 13), the verse is two syllables short. (Nunes
prints Porque ouv<e> el queixume, os meus although in this poem we find ouv el in vv. 3 and 13, and ouv n in vv. 9 and 15.)
257
10
15
20
19-20 om. B
Braga and Nunes both read pudi, which is possible (cf. pude in Talaveira at CA 243.5 [where B has pudi], 245.8), but the
adverb i yields fuller sense (but I couldnt do anything else about it) and fazer i is a common expression.
258
10
15
20
19 <a>tal Nunes
259
10
15
20
17 est V : e B
Whereas both Braga and the Machados read (correctly) with the MSS, Nunes senselessly emends (fal, e ver). The anticipatory
subjunctive fale is introduced by ante... que (vv. 18-19). He will have his share of the pain that he caused her.
2
Ve<n>gar (for negar) is a nearly certain correction: we need only assume a banal error of n for u along with the loss of a til.
260
10
15
20
261
10
15
20
13 bona B 15 ne~ V : ne B
(pol) por B
See the note to that text. I believe that the poem belongs here rather than there.
262
10
15
qxara
t V :
263
10
15
One of the few places in the cantigas d amigo where both MSS have ir for rr.
The form agura occurs in B 658 / V 259 (Carpancho), vv. 8 and 13 (but not in the corresponding place in the third strophe).
In this text agora appears in the MSS only in v. 10 (in v. 16 both have amiga). Elsewhere among the cantigas d amigo of Sandeu
the word pops up once: B 798 / V 382, v. 4, where both MSS read agora.
2
264
10
7 sera V : sem ? B
9 uos B : iuos V
11 poys V : Poy B
queredes] qre V
265
10
4 en] eu V
6 comi~go B
9 bon B
In BV the text is copied on short lines according to which the rhyme scheme changes from abbaCC in strophe I to
abcbDD in strophe II, an anomaly in the corpus of 500 cantigas damigo (B 1217 / V 822 [Sevilha] is hardly parallel),
and most likely due to a corruption (though possibly to a rewriting whereby a scribe sought a rhyme-scheme
which was not there); if so, the scribe copied perduda in the wrong place. Lapas transposition (1982: 172; also in
Lapa 1929: 322, n. 2) may move the words around too much to be plausible, but Lapa is probably right to look
for a solution. The strophic form would then be either aaB (with long verses in the body of the strophe and an
internal rhyme at midverse in the refrain) or aaBB (with two short rhyming verses in the refrain).
266
10
10, 11 mi~ BV
10 mu~da V : mu~do B
267
10
15
268
13 u9 B : n9 V
10
15
9 <ar> supplevi2
This may be a scribal lapse (Lapa 1982: 172-173), an alternate spelling of the perfect, or a pluperfect in place of a perfect.
Cf. B 629 / V 230 (Calheiros), v. 10 ca ja mais non ar fui ledo meu coraon; B 684 / V 286 (J. S. Coelho), vv. 3-4 direi vos quant que
sabor / non ar ouve d al nen de mi; B 1265 / V 870 (Loureno), v. 14 e non ar quis que comigo falasse; B 1025 / V 615 (Johan Airas),
v. 3 que vos non ar falou.
2
269
10
15
20
The final verse of the first and third strophes has ten syllables, this one eight. Irregular strophic form (cf. vv.15-16, 19-20)
or scribal error? sempre could be inserted at several points.
2
Cf. A 74 / B 187 (Torneol), v. 3 pois eu non sei al ben querer.
270
FERNAN FROIAZ 1
10
15
20
Although Nunes emendation (based on the parallel expressions in vv. 9, 14 and 19) seems obvious, it is just possible that
punning is at play: in this context, m errastes would have nearly the same meaning as morastes but more punch. Cf. B 534
/ V 137 (Dinis), v. 15 pois m errastes tan en vo.
271
FERNAN FROIAZ 2
10
15
20
vv. 13-15: I take pero (v. 13) as an adverb and the boy as the subject of both verbs in v. 14 (se soubess in the protasis,
ia morrer in the apodosis).
1
Nunes would cut non, but if the girl were guarded, how could she go see the boy? (Cf. B 1246 / V 851 [Padrozelos], vv. 4-5 and 7-8.)
272
FERNAN FROIAZ 3
5 queu V
273
FERNAN FROIAZ 4
10
15
v. 15: ante is an adverb here (= beforehand), while me is construed with jurar (cf. v. 9). (In front of me would
be ante mi.)
274
FERNAN VELHO 1
10
15
Lanciani claims that we ought to read with Braga. But, as Nobiling (1908: 345) noted and Cunha (1982: 14, 66-75, 82, 168)
stresses, se was not, as a rule, subject to elision. Moreover, we have clear examples of se verdad (B 1046 / V 636 [Johan
Airas], v. 21; cf. B 752 / V 355 [Guilhade], v. 14, where verdad stands in rhyme), whereas the one instance of s verdade
alleged by Lanciani (CA 19.3 [Somesso]) is simply an editorial slip by Michalis, where we ought to read se verdad . For the
word order, cf. B 1226 / V 831 (Baveca), v. 5 se verdade for, where there can be no mistake.
275
10
15
nu~ca V : nuca B
9 ventura] uentra
The older form valvera (< u a l u e r a t ) does not occur elsewhere in the cantigas d amigo but see the glossaries of Michalis,
Mettmann, Lapa. In A 103 / B 211 (Burgals), v. 16 we find ualuera A : ualera B. See also B 659 / V 260 (Carpancho), v. 16.
276
10
15
Majorano follows Nunes, despite the censure of Lapa (1982: 174). Cf. B 728 / V 329 (Rodrig Eanes de Vasconcelos), v. 5
and note.
2
To pad the verse, Nunes (followed by Majorano) prints pesar-mi- <en>, but m ia (Lapa 1982: 174-175) is required by the
syntax (a conditional ought to respond to the subjunctive pesass in v. 7) and scans perfectly.
3
The futures sofrer, poderei, averei support my correction, and ere could have been represented by a single abbreviation (cf. B
1153, v. 13 faz = fazere and, in B 170 on f. 43v, line 6, the reading mci = mereci). The emendations of Nunes and Majorano
require the awkward (and highly improbable) avere i in v. 16.
4
Previous editors presume an elision between se and eu but se ought not be elided (Nobiling 1908: 345; Cunha 1982: 14, 66-75, 82, 168). So, since eu appears earlier in the verse, its repetition here is semantically superfluous as well as unmetrical.
277
10
15
1
The emendations proposed here would yield a syntax parallel to that in v. 3. Cf. B 975, 978 / V 562 (Rodrig Eanes
d Alvares), v. 15: Nen vistes omen tan gran coit aver / com el por mi . (Majorano suggests e <o> veredes coita<do> por en. Nunes
prints e veredes <sa gran> coita por en.)
278
10
15
11 u9 B : om. V
The end of this verse is unacceptable on two grounds: (1) rhyme: vv. 1, 4, 7, 10 and 16 are unissonans, ending in close er,
whereas this verse ends in open r; (2) meaning: the (alleged) use of entender to mean make [someone] understand is
without parallel. Still, since the general syntax seems viable (cf. B 680 / V 282 [J. S. Coelho], vv. 13-16 Des que souberdes
mandado /.../.../ enton mh averedes grado), the corruption may affect only the last four syllables and fezer veer (or something
similar) would salvage the verse and solve the problem (cf. A 87 / B 191 bis (Burgals), v. 19 E se mi-a non fezer veer).
279
10
15
20
11 eu Nunes
vv. 16-17: if the text of BV is correct, al rhymes here with al. See note on B [920 bis] / V 508 (Porto Carreiro),
vv. 10-11.
1
Nunes (followed by Majorano) tries to make up the missing syllable by printing oj<e> el, but I find no parallel in Amigo for
such a hiatus. Better to restore the final vowel of vive, which would not have been elided at a pause.
2
For the expression ne~un ome cf. CSM 281.53 espantou- / s a gente por neun ome a magestade chamar. For the error see A 229 / B 419
(Guilhade), v. 14, where A reads ome against hm in B (it is precisely such a reading that could have been mistaken for nm).
3
The form dissesse (BV) can be defended metrically and syntactically (pero takes either indicative or subjunctive) but does
not provide a perfect rhyme. Though dississe is not registered in the glossaries, we find dissiste in rhyme at CSM 6.84. In B 746
/ V 348, v. 10, Guilhade employs fexisse in rhyme (though elsewhere he uses the regular fezesse). Nunes prints pero que mi-o <a
mi> nen hun non disse.
280
10
15
20
In V, the rest of f. 65v is blank. The three cantigas d amigo attributed to Coton begin on the opposite page (f. 66r) of the
manuscript.
2
If Nunes is right, this would be a variation on the first line of the (intercalated) refrain. True, estest could easily have been
reduced to est and the initial ou would fit well (syntactically and semantically) after v. 14; but the verse must still be
emended to yield the alleged variation.
3
In V about 10 blank lines follow v. 18, suggesting that room was left for the missing strophe. In B it reads as follows:
Muy tristdads en sey eu
o pr qh poilo n
dizedemho en u9 seia greu
q estaqstou pr qo fazedes
par ds senr omha coyta emeu
Mal estou eu seo uos
281
The number of words that rhyme with greu is quite limited (eu, meu [used in v. 24], seu, deu, and some words inadmissible
here: teu, encreu, judeu [Cunha 1991, esp. 914]) and leu seems the only one that works.
5
Even if the rest of the verse is correct and that assumes that the exclamatory o (cf. Mettmann, s.v.) is not an error for ai
(cf. vv. 1, 11, 17) there seems to be a syllable missing (since the final vowel of coita is presumably elided), and meu cannot
stand alone. If I am right, <mal> meu would inversely mirror meu ben (v. 1) and also play on mal estou eu at the beginning of
v. 24 (cf. the note in Nunes vol. III, p. 207). Cf. B 723 / V 324 (Travanca), vv. 13-14 chamava m el.../ seu ben e seu mal. We find such
expressions in cantigas damor, e.g. A 179 / B 330 (J. S. Coelho), v. 14 e por que sodes meu mal e meu ben.
282
Se gradoedes, amigo,
de mi, que gran ben queredes,
falad agora comigo,
por Deus, e non mho neguedes,
amigo, por que andades
tran trist ou por que chorades?
Pois eu non sei com entenda
por que andades coitado,
se Deus me de mal defenda,
queria saber de grado,
amigo, por que andades
<tan trist ou por que chorades>?
Todos andan trebelhando,
estes con que vos soedes
trebelhar, e vs chorando,
por Deus, e que dem avedes,
amigo, por que andades
<tan trist ou por que chorades>?
10
15
283
B 827 f. 174v V 413 f. 66r (cf. B 640 / V 241 [Pae Soarez de Taveirs])
13 O edd. (ex B 640 / V 241) : E BV
15 ueio BV
284
10
15
10
15
285
vv. 3, 6, 9, 12, 15, 16: This text hinges on the refrain and in particular on the word onde whose syntactic
versatility begins in Latin. OLD, s.v. unde B. 6, reports that unde is used to introduce adverbial clauses with or
without correlative. OLD B. 6. b glosses a special use of unde as from the person from whom just the
double-duty meaning onde must bear here. For onde elsewhere in the cantigas damigo, see B 630 / V 231 (Calheiros)
v. 5 E, se m outren faz ond ei despeito; B 1156 / V 758 (Zorro), v. 3 al vai, madre, o<n>d ei suidade; B 1147 bis / V 750
(Servando), vv. 5-8 ca non ei sen vs a veer, / amigo, ond eu aja prazer; / e com ei sen vs a veer / ond eu aja nen un prazer?;
B 1037 / V 627 (Johan Airas), v. 1 Vedes, amigo, ond ei gran pesar. A closer parallel outside of Amigo in A 297
(Calvelo), v. 5 ond eu atendo ben, me ven gran mal, where ond must mean from the person from whom. In CSM
368.42 meu Fillo... ond eu fuy prennada it clearly means by whom and functions (within the relative clause) as
agent of a passive participle, as here in strophes I-III.
1
Unless we accept the word division agor a in v. 2 and penad a in v. 5, the feminine pronoun a in vv. 7 and 10 would be the
only morphological indicator of the gender of the third persona who is not, however, a lesbian beloved (as some editors
may have feared), but rather a rival, an Other Girl. See Cohen & Corriente (2002).
2
There are two possible criteria for deciding which MS to follow: vocalic and semantic. The sequence of tonic vowels in
rhyme in the first four strophes follows the pattern: a i a i. And this might lead us to follow B, which would complete the
sequence with a e. But the order of the last two strophes in V seems to present a semantic crescendo, since rogasse is
stronger than dissesse and chegasse arguably more specific than veesse (cf. vv. 3, 7, 9, 13 in B 829 / V 415 [the next cantiga]).
286
10
15
20
25
30
287
vv. 2, 4, etc.: lelia doura pode ser lido como lya ddwra
a mim a vez (= a minha vez) em rabe andaluz.
lya, um alomorfe de li a mim, para mim, foneticamente /leia/, e lelia pode ser ou um erro por leiia ou,
mais plausivelmente, uma transcrio de lya (cinco
vezes em Martin Giinzo devemos pronunciar Cecilia
como /sesia/ em rima, sendo todas as rimas perfeitas). Ed oi romance ibrico arcaico < et hodie com
vocalizao do -t- intervoclico no interior da expresso. Assim ed oi / MUDANA DE CDIGO / lya ddwra =
E hoje / MUDANA DE CDIGO / a minha vez um
verso bilingue, como ocorre em tantas kharajat.
vv. 25, 27, 29, 31: leli can be read in Andalusi Arabic as
layli, the substantive layl night with first person pronominal suffix (cf. Dutton 1964, but the noun is
collective, not the nomen unitatis, which would be
laylati). ya layli is a common exclamation, meaning
something like what kind of night I had. The first
element can be omitted for metrical reasons or to
express intense emotion.
vv. 25, 27, 29, 31: leli pode ser lido em rabe andaluz
como layli, o substantivo layl noite com sufixo pronominal da primeira pessoa (cf. Dutton 1964, mas o
nome colectivo, no o nomen unitatis, que seria
laylati). ya layli uma expresso comum que significa
qualquer coisa como que noite eu tive. O primeiro
elemento pode ser omitido por razes mtricas ou
para exprimir emoo intensa.
288
10
15
289
PERO DE PONTE 1
10
Filha, dou vos por conselho que, tanto que vos el veja,
que toda ren lhi faades que vosso pagado seja
Madre, namorada me leixou,
madre, namorada mh leixada,
madre, namorada me leixou
15
20
12 pagato V
13 me] (mha) me B
14 leixada
The body of each strophe is copied on short lines in BV and editors have printed them as though the form
were abcb, (scanning 7), whereas what we have is aa (scanning 15 each). Cf. note on B 1171 / V 777 (Bolseiro).
290
PERO DE PONTE 2
E vistes u s el partia
de mi mui sen o meu grado,
e jurando que avia
por mi penas e cuidado?
tod andava con mentira,
ca, se <non, noite nen dia,
a meos de meu mandado,
nunca s el daqui partira>
10
15
20
25
13 c V : r B
17 ya BV
18 mado B
27 uira B : iura V
291
PERO DE PONTE 3
10
15
20
292
PERO DE PONTE 4
10
15
293
PERO DE PONTE 5
10
15
8 <ca> supplevi1
muyt B : muyr V
Nunes prints Non mi tardedes com outra vegada / mi tardastes; muit<o> ei en gram medo , while Panunzio inserts <eu> before n.
A syntactic and discursive connection seems needed, and Ponte often uses ca in explanatory clauses (B 832 / V 418, v.6; B
833 / V 419, v. 8; B 834 / V 420, v. 9; B 836 / V 422, v. 10; B 837 / V 423, vv. 4 and 8), as he does in this text in v. 10.
294
PERO DE PONTE 6
10
15
20
295
PERO DE PONTE 7
10
15
20
Nunes (vol. III, p. 216) suggests (and subsequent editors print) sei <eu> but the phrase saber i conselho easily qualifies as
a formula (as in B 836 / V 422 [Ponte], vv. 4 and 10).
296
10
15
20
25
30
297
vv. 3-5: Most editors have printed this poem with a 4-verse refrain. There are genuine examples in the cantigas
damigo of 4-verse refrains with two alternately repeated rhyme-words (B 706 / V 307 [Vinhal]; B 1147 bis / V 750
[Servando]; cf. B 1139 / V 730 [Bonaval], whose refrain can be taken as four verses), but here the fourth verse
of the refrain is only partially and confusedly copied in BV and we seem rather to have one verse of false
refrain (Parkinson 1987). The disproportionate bulk of a 4-verse refrain is noted by Michalis (1911-1913:
416). Correia (1992: I.120) writes: Observando a escrita de todos os refres de tipo R10 (scil. with repeated
verses), conclu haver razes para entender as aparentes abreviaes em As froles do meu amigo [...] como erros
por duplicao. Correctly understood, the strophic form follows a pattern found in several other cantigas
whereby the 2nd verse of a 3-verse refrain is longer than the other two, but equal to the verses in the body of
the strophe. See the note on the strophic form of B 1171 / V 777 (Bolseiro).
vv. 12, 17, 21, 26: We should probably take ao as one syllable, representing the contracted form (cf. B 1235 /
V 840 [Ambroa], v. 6, s); otherwise the verses would be hypermetric.
298
.
.
.
.
.
.
.
.
10
15
299
10
15
300
10
15
20
I can find only one parallel for a break followed by ende: in B 531 / V 134 (Dinis), in the last verse of the poem, in the phrase
ende so desejador where we should probably (with Michalis 1895) correct ende to onde (cf. Nunes 1932: 163). On the other
hand, the construction proposed here is well attested, occurring for instance in B 481 (Alfonso X [=CEM 11]), v. 7 e demais
nna dei eu a vs senlheira and moreover appearing at least once in the final verse of a cantiga: B 1631 / V 1165 (Ponte) e demais
dar-lh albergu endado. The reading here, in the high reaches of the tradition, could have been uoss demays with a light
vertical stroke over the e to indicate word division, subsequently misread as a til and converted to an n. It would be illogical
to have an asyndeton here in the last verse of a cantiga ateuda ata a fiinda. The word demagis (Furthermore, moreover),
though extremely rare, is documented in early Latin in Lucilius (see OLD s.v. [which reads <sed> demagis with Marx; I would
propose <et> demagis]).
301
10
15
14 del igual V
18 morrer] mortr BV
We need <s> because has already been used in rhyme (v. 7). Moreover, s offers better syntax. Cf. B 601 / V 204 (Dinis),
vv. 15-16 cada u s / falando and B 1046 / V 636 (Johan Airas), v. 18 que enos ceos s. In CSM the form s ([< sedet] present indicative,
third singular, of seer) is often used to rhyme with (e.g. CSM 296.27 u el ss).
302
303
JOHAN GARCIA 1
10
15
20
304
JOHAN GARCIA 2
10
15
Lapa (1982: 176) notes that the girl doesnt go on about a rival girl, and that the confusion of a and o is quite frequent.
305
REIMON GONALVEZ 1
10
15
306
GARCIA SOARES 1
10
15
6 Nunca] Aunca V
11 <e>no Nunes
12 encon B : enc V
An anomaly due to faulty transmission, or variation in the refrain? In the first verse of the refrain in the next two strophes
both MSS have eu mho fiz , but here they have eu o fiz. Since the first strophe has eu mho fiz in the third verse of the refrain
(a verse not copied in the other strophes), we could suppose that in each strophe the poem had o in the first verse and mho
in the third verse of the refrain. But it seems more likely, on the evidence, that we should read mho throughout. Cf. B 710
/ V 311 (Vinhal), vv. 16 and 22.
307
GARCIA SOARES 2
10
15
20
25
308
10
15
309
10
15
20
A variant of the formula I propose occurs in B 828 / V 414 (Solaz), refrain: ond eu amor ei. Compare ond avia sabor in v. 14.
el seems an awkward (and not coincidentally unmetrical) interpolation by a performer or scribe who had not anticipated
the coming subject.
2
310
MEENDINHO 1
10
15
v. 3: On my reading, ua is used here in an archaic sense already seen in certain uses of Latin unus (OLD, s.v.
unus 7, 8).
311
Cf. B 1243 / V 848 (Padrozelos), vv. 16-18 se oj i ~ua vai / fremosa, eu serei / a u~a, ben o sei. In support of this reading we may note
that ~ua is found at the end of a single verse refrain in a cantiga of the form aaB in V 1027 (=CEM 411 [Roi Paez de Ribela])
e coz end a ~ua. That verse could be a parody of a song with a refrain ending on the word ~ua (possibly but not necessarily this
text). Compare the position (at verse-end) of two synonyms, senlheira and soo, in B 1165 / V 771 (Bolseiro), v. 1 Sen meu amigo
manh eu senlheira and B 927 / V 515 (Roi Fernandiz), v. 8 como m atende soo (where we also find the verb atender, as in our text).
Both the situation of the girl all by herself and the position of an adjective meaning alone are parodied in B 1499 / V 1109
(=CEM 213 [Guilhade]), v. 8 E cada que vs andardes senlheira and B 1383 / V 992 (=CEM 385 [Burgals]), v. 11 Por non viir a min
soa, sinlheira. See also: B 1035 / V 625 (Johan Airas), v. 27 soa; B 868, 869, 870 / V 454 (Airas Nunes), v. 3 senlheira; B 688 / V 290
(J. S. Coelho), v. 8 soo; B 332 (Redondo), v. 11 soo.
2
Tavani (1980: 84) would read e verr? ending the refrain there, but (without commenting on the unlikely syntactic break) as
an abbreviation for verr we would expect ura and not ua. Michalis (vol. II, p. 889) and Nunes print a 2-verse refrain
eu atendend o meu amigo / eu atendend o meu amigo but with scant supporting evidence: a til over the u in B, a faint mark over the
u in V. Parkinson (1987: 46) defends the hypothesis of a truncation. Correia (1993: I, 120-123) argues well and with strong
evidence against a 2-verse refrain, but then accepts Tavanis proposal without providing any arguments.
3
Lapas proposed supplement <ar> in vv. 8, 11, 13 and 16 (1982: 176-177) has not been taken seriously (cf. Montero 1991:
90-91), but there is good evidence to support it. The adverb (or perhaps we should say particle) ar is often found following
nen and just before the verb in the second (or third) negated clause in a series (in vv. 8 and 13 the verb is implied). For
examples in the cantigas damigo, see B 668 / 271 (Avoin), vv. 3-4 por quen non parece melhor de mi / nen ar val mais; B 686 / V 288
(J. S. Coelho), vv. 13-16 Ca eu nunca con nulh ome falei, / tanto me non valha Nostro Senhor, / des que naci, nen ar foi sabedor / de tal fala;
B 734 / V 335 (Barroso), vv. 13-14 Se m el quer ben, non lho quer eu nen mal / nen ar ei <ren> que lhi gradesca i; B 1200 / V 805 (Treez),
refrain que non fezeron des enton os meus olhos, se chorar non, / nen ar quis o meu coraon que fezessen, se chorar non; B 1239 / V 844
(Padrozelos), vv. 3-4 non m ar vistes / nen ar ouv eu depois migo / de nulha ren gasalhado; B 555 / V 158 (Dinis), vv. 14-15 ca o non vi
nen vio el mi / nen ar oo meu mandado. The precise collocation nen ar sei occurs in A 26 (Somesso), v. 17 and B 75 / B 1336 / V 943
(Tamalancos), v. 15. Very rare is the use of ar before a substantive, as in vv. 8 and 13. A similar construction, only possible
with an implied verb (as also in our text), is found in A 161 (J. S. Coelho), vv. 22-24 (cf. Michalis, s.v. ar), where ar precedes
a second subject (in our text it comes before a second direct object): E esta dona, poilo non souber, / non lhe poden, se torto non
ouver, / Deus nen ar as gentes culpa per.
1
312
10
15
20
The supposed correction <s>ei is influenced partly by the occurence of forms of saber in vv. 4, 8, and 16. Finazzi-Agr
acknowledges that aver in the sense of to hold, to believe is found elsewhere in medieval Galego-Portuguese, and that
example (which he cites [p. 115] but does not quote) is worth looking at (B 489 / V 72 =CEM 19 [Alfonso X], v. 10): e direi vos
ora por que o ei. Lapa glosses por que o ei with porque assim o penso (in his glossary, s.v. aver, he offers pensar, entender;
cf. OLD, s.v. habeo 11, 24, 25). Moreover ei, if correct, forms part of a word play (ei que non ) with which the poem opens.
2
The reading mhe (V) seems a simple corruption of mho which is to be taken with negar (There is no need to hide it from
me). Cf. B 1197 / V 802 (Caldas), v. 10 e non vos mester de mho negar and B 1149 bis / V 752 (Zorro), v. 9 Non vos ten prol, filha, de
mho negar.
313
10
15
vv. 12-14: avia gran sabor / con o rogar me ... / o que... basicamente significa Estava deliciada com pedir-me o
meu amigo.... (Note-se que o que [o amigo] o sujeito de rogar.)
314
10
15
20
8 ssonho V : ssol(ne~)nho B
20 sonb V
9 E eu B : cen V
315
AIRAS NUNES 1*
316
10
15
Ought we to see in como enos (BV) a correction of como to come ? Cf. B [1158 bis] / V 761 (Zorro), vv. 3 (apparatus and note)
and 9. Cf. Lang 1894: lxxviii (note 1).
317
10
15
20
318
10
15
20
25
30
35
319
1 eu] en B 2 sibeyra V : corr. Varnhagen : om. B 3 estav a<li> supplevi1 : estava <i> Nunes 4 e om. B ascuytar
or astuytar B : asenytar V 6 Solo rramo V : Sol iramo B
verd e frolido distinx. Tavani : uerde frolido BV
7-8 amigo; e Michalis (II, 880, 939)2 : amige V : amigo B 11 mha chegado BV : corr. Varnhagen 12 oyr B : oyi V
(undotted)
n B : n V 15 cantandes B3
emoyreu BV : e moir<o> eu Nunes 15-16 e pen<o> / e Nunes
(pen<o> iam Michalis)4 : e pene V : e pen e B (on the next line in both MSS) 19 fazi <u~>a guirlanda Nunes : fazia
gilanda BV : fazia guirlanda def. Tavani5 21 ent V : ent.q. B 23 nonou sar B : nonousar V : non ousar def.
Tavani6 : nom <o> ousar Braga 26 indo-s en Varnhagen : indosseu V : om. B 27 camyo BV 29 hypermetric7
30-35 on three lines in BV8
33 quen V : quem que~ B
34-35 dormir / ai Michalis (II, 939) : dormoray V :
dormoiay B
The strophic form, leaving aside for a moment the songs cited at the end of each strophe, is abbaa, and I take
the scansion to be 10 10 10 10 10. The cited songs are all xxy, scanning: I: 8 8 7; II: 9 9 6; III: 10 10 10 (or 9);
IV (where we have two of them): 7 7 2 , 5 5 4. Tavani argues that the 3rd verse of each strophe can be scanned
9, a syllable shorter than the 2nd , which scans 10. In strophes II and IV, the 3rd verse can also be scanned as
10, leaving as exceptions vv. 3 and 19, both of which seem to me to read better (grammatically and stylistically,
not to mention metrically) as here emended. The question is not easy to resolve, and the use of the dobre in
the 1st and 5th verses of each strophe (I cantar, II ben, III enton, IV pastor) could be adduced to support either
position. Note that the final rhymes of the included songs form the pattern: or (I); al (II); al (III); or, ol (IV) thus
echoing the a-rhymes of strophe I (ar) and IV (or).
1
Curiously, ali is found in three of the other six Galego-Portuguese pastorelas: B 554 / V 967 (Johan Airas), v. 15 Ali stivi eu mui
quedo; B 534 / V 137 (Dinis), v. 18 ~ua gran pea do dia / jouv ali, que non falava; B 1098 / V 689 (Sevilha), v. 22 eu non sei ren que lhi
dissess ali. On the other side (of the collocation estav a<li> senlheira) we might cite CSM 65.251 foi senpr ali senlleiro.
2
Tavani omits the conjunction at the beginning of v. 8, but its restoration is justified by the combined readings of the MSS
and fits well with the sense, syntax and rhythm of the poem.
3
Though cantandes (B) may look like a dialectal variant, it is probably a slip.
4
Tavani declines to restore the o (of peno), whose elision is signalled in the MSS, and prints moyreu e pen / e but this leaves
the verse without a final assonance to match that of the preceding. Nunes, reaching for eleven syllables, also restores the
final vowel of moir<o>, but this is misguided, since we merely read ai storninho with aphaeresis.
5
Tavani leaves guirlanda dangling without an article (see note on meter). Note that in v. 25 it takes a definite article.
6
Nunes prints non ousar and Lapa suggests non nousar (= non no ousar), both aiming to add an object pronoun without
adding a syllable (assuming veer is bisyllabic). Tavani sees no need to add an object for veer since amigo can serve as direct
object of both infinitives. Cf B 644 / V 245 (Torneol), vv. 2 and 10.
7
The verse seems a syllable long and one might be tempted to cut ben, but in the cantigas damigo the verse before the final
refrain (Rv-1) is hypermetric with more than accidental frequency (see Introduction, 5).
8
Cf. B 1155 / V 757 (Zorro), vv. 1-3.
320
10
15
20
The word is copied on the extreme right edge of the page, and no trace of an o is visible.
Unless we prononce smelho in two syllables the verse scans long. (Perhaps Se for Pero?).
321
We could defend BV as follows: in medieval Galego-Portuguese querer is frequently used as an auxiliary verb to form a
compound future tense; thus, when the girl asked the boy if he would come back, he said querrei: either I shall or I mean
to. But the correction has a strong basis in a conversational convention that dates from early Latin (e. g. dixisti? / dixi) and
continues in force today in Portuguese. Thus the reported girl / boy dialog at parting would have been: verredes? / verrei.
2
The MSS ssuya (sua) is a variant form (from solebat) which the representation soa would obscure.
3
Nunes ei n is basically sound (cf. for example B 1016 / V 606 [Johan Airas], v. 8 and B 1134 / 725 [Veer], v. 2), but here we
need ei end in front of eu as in B 1147 bis / V 750 (Servando), v. 2 e ei end eu mui gran pesar.
322
10
7-9 om. B1
7 ramo] rrayao V
The order of strophes III-IV is uncertain (see note below) and the meter varies significantly. In the text
transmitted by BV the scansion (excluding the refrain) would appear to be as follows for the four strophes:
7 8; 7 8; 8 7; 8 7. We can even this by taking meu in vv. 1 and 4 and eu in vv. 8 and 11 as bisyllabic.
1
In B, at the bottom of the right hand margin of f. 197v there is a conspicuous +, apparently indicating that something (a
strophe) is missing. Editors assume assume that the pattern of tonic vowels in final position (in the body of the strophes)
should be a / i / a / i and B follows this order through three strophes, but V may accurately reflect an anomaly. There are
similar problems in B 644 / V 245 (Torneol), B 662 / V 263-264 (Carpancho), B 1237 / V 842 (Pae Calvo), and in two texts
(N 3 and N 6) of Martin Codax. The phrase dona virgo seems an appropriate final touch to the poem, and the appearance of
the pair dona virgo / dona d algo in that order elsewhere (cf. B 1155 / V 757 [Zorro]) does not preclude an inversion here.
323
10
15
20
324
in the last strophe of B 1039 / V 629 (Johan Airas), vv. 14-15 the last cantiga with dobre in his set al rhymes
with al, not as part of the dobre, but in addition to it: apparently one special effect on top of another. Further,
in B 786 / V 370 (Guilhade), vv. 22-23, son rhymes with son in the fiinda of the 21st of 22 cantigas.
1
In Nunes the refrain appears as follows: de viver migo, qual eu ouv i / de viver vosco, des que vos vi. For the rhyme-scheme
abbaCDDC, cf. Tavani, Rep. 177:1-3, but none of those texts has a refrain. The refrain on two lines in the MSS has internal
rhymes but no end rhymes.
2
Note that later in the same verse the same error alleged here (s for l) occurs in reverse: gse V : gle B. For the divine subject,
cf. B 1021 / B 611 (Johan Airas), vv. 4 (e por en guise mho Nostro Senhor) and 16 (e por en guise mho Deus, se quiser)
325
GOMEZ GARCIA 1
10
15
4 farey V : fazei B
vv. 7-8: avi~ir se (con alguen) means to make up, that is,
get back together with someone (after a break-up
or quarrel). Well documented in Galego-Portuguese
lyric and medieval prose, it is found only here in the
cantigas damigo (cf. B 787 / V 371 [Guilhade], where we
find a rare transitive usage).
By restoring and not eliding the final e of ande (anda V) and reading ~uu we would get a twelve syllable verse. The strophic
form would then conform to Tavani, Rep. 160:4 (one example, CA 169 [J. S. Coelho]). There are other examples of texts in
which the last verse of the refrain has two syllables more than those in the rest of the strophe (see, for instance, the cantiga
d amigo by Loureno with the same rhyme-scheme [abbaCC], B 1260 / V 865 [Tavani, Rep. 160:313; 408; 431]). Printing (with
Michalis) ande un tempo, but eliding the e, the verse scans ten syllables. The longer verse length might be defended as
strengthening the thrust of the utterance: it is here that the girl threatens what Michalis calls retaliation. This would,
however, be the only case of bisyllabic ~uu in the cantigas d amigo.
1
326
ROI FERNANDIZ 1
10
15
13 por] pr B : pre V
16 Pero B : po V
v. 3: amigar (BV), found only here in medieval GalegoPortuguese lyric, ought to mean to make up or get
back together (after a quarrel or break-up), and not
merely to be friendly (in the euphemistic sense of
carry on an amorous relationship; cf. Nunes, s.v.:
tomar por amiga ou namorada). But even if it meant
the latter, in this context it would imply the former.
Cf. vv. 9-10 and 13-14.
v. 16: The understood grammatical subject of veesse is pesar (which may be understood either from its use in
v. 15 [fiz pesar] or from its use in the the refrain [eu nunca faria / pesar] or from both, and recurs twice in the
refrain [where lo = pesar]).
327
ROI FERNANDIZ 2
2 Madrirey B : ma dirrey V
7 ffilhe B : ssilhe V
v. 5: A similar expression, apparently the passive of meter en falha (which Nunes [s.v. falha] glosses as faltar,
enganar) occurs at V 1017 (=CEM 234 [J. S. Coelho]), vv. 1-2 Luzia Snchez, jazedes en gran falha / comigo, que non
fodo mais nemigalha / d u~a vez; where jazer en falha means to be disillusioned (or disappointed).
328
ROI FERNANDIZ 3
10
15
20
25
329
ROI FERNANDIZ 4
10
15
Bragas que muyt ey, following V, is senseless and unmetrical; but Nunes version of vv. 1-3 is even worse, disregarding the
MSS and yet still making no sense. Everything the girl says in this poem admits to a grave error. Were my suggestion, based
on the combined evidence of B and V, essentially correct, vv. 1-3 could be rendered: O mother, what a mistake I made,
cause I didnt go see my friend where he could talk with me.... For the construction que muit, see B 699 / V 300, v. 2 (Ulhoa)
fui eu i muit errada; B 1236 / V 841 (Calvo), v. 6 nembrar se devia de que muito m erra; CEM 24 (Alfonso X), v. 5 Pois el agora tan muito
erra; CSM 8.43 entendeu que muit errara; CSM 272.16 tan muit avia errado. For the position in rhyme, see (among the cantigas
damigo) B 1118 / V 709 (Berdia), vv. 3-4 ca toda ren que m el a mi mandou / fazer, fij eu e nunca lh <i> errei (the form also occurs in
rhyme in Amor, CEM and CSM). In addition to the supplement printed here, we might consider several others which
assume that errei is correct: que mui<to err>ei (with hiatus); que mui<to lh err>ei (the demonstrative pronoun would refer to the
boy); que mui<t eu lh err>ei (which sounds rather too heavy); and que mui<t i err>ei (the adverb i would anticipate the mistake
described in vv. 2-3). And finally, we might also weigh que mui<t errad> ei. Cf. CSM 281.80 Amigo, eu fui errado... (at the
beginning of cited discourse).
2
Variously read as u el falasse comigo or vel falasse comigo these letters are just a version of v. 3 copied out of place, yet have
editors have transformed them into a third line of the refrain (cf. Parkinson 1987 on false refrains).
3
Misread by all previous editors and moreover attributed to the mother instead of the girl.
330
ROI FERNANDIZ 5
10
15
4 ma B : om. V where three lines are left blank 6 vee-lo Monaci : ueole BV
ueer V : iueer B 17 seu mandado V : om. B
331
ROI FERNANDIZ 6
De seui . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
<que non queirades, amigo, partir>
10
15
20
Cf . B 1013 / V 603 [Johan Airas], v. 15; B 571 / V 175 [Dinis], vv. 5, 11, 17.
nunca ja occurs often enough without mais to warrant suspecting the latter here (for example: B 732 / V 333 [Barroso], v. 7
nunca lhi ja creerei nulha ren; B 1228 / V 833 [Baveca], v. 6 nunca vos ja de ren ei a creer ; B 750 / V 353 [Guilhade], v. 13 mais nunca
ja crea molher; B 777 / V 360 [Guilhade], v. 6 mais pero nunca vos ja ben querrei). Pointedly, the phrase is used in a parody of the
renuntiatio amoris in B 1629 / V 1163 (Ponte), vv. 22-23 E nunca j comerei / com enton con el comi.
2
332
ROI FERNANDIZ 7
10
15
20
333
PAE DE CANA 1
334
10
15
PAE DE CANA 2
10
15
11 mui gm me V : om. B
335
SANCHO SANCHEZ 1
10
15
336
SANCHO SANCHEZ 2
Eu era maravilhada
por que tan muito tardava,
pero sempr esto cuidava,
se eu del seja vingada,
mais jur a Deus que quisera
<or ante que mort era>
10
15
20
11 q q sa V : om. B
15 deseia V : deseya B
17 q qsa V : om. B
337
SANCHO SANCHEZ 3
10
15
338
SANCHO SANCHEZ 4
10
15
The verb leixar is documented in the required sense both with and without a reflexive pronoun (cf. CSM, s. v.).
339
SANCHO SANCHEZ 5
10
15
340
10
15
20
341
10
15
20
342
10
4-14 om. B2
12 entendi V : entendi
Nunes also suggests (vol. II, p. 470) ~u (u~a ), to which Lapa (1982: 182) rightly objects. Tavani (Rep.) accepts Nunes Un .
Meanwhile Lapas Muit occurs in the incipit of five other cantigas damigo (though not in verse-initial position).
2
At this point begins a lacuna in B, where two pages (remains of which are visible) were torn out of the codex. V is therefore
the only witness for the rest of this text and all of the following one (as well as for seven cantigas d amigo of Johan Airas,
V 594-600).
1
343
10
V 591 f. 94r
2 quen Monaci : queu V
1
o] e V
Nunes suggestion makes sense (Braga keeps vos) but breaks the parallelism. que me veja would correspond to se me vir in
v. 9, where the boy is the subject of the verb and the girl the object.
344
10
15
Radulet (ad loc.) finds no parallels for the construction with de que (which she emends to de<s> que though I find no
examples among the cantigas damigo of des que introducing a causal clause); but both parallel and related uses of de que can
be found: B 1223 / V 828 (Baveca) vv. 8-10 Amiga, vs non fezestes razon / de que perdestes voss amig assi / quando vos el amava mais
ca si; B 796 / V 380 (Sandeu), vv. 7-8 Ben m con este mandado que ei / de meu amig, e non o negarei,/ de que se ven; and, in the same
text, vv. 13-15 Muit and eu leda no meu coraon / con meu amig, e fao gran razon,/ de que se ven [...] . Cf. also B 1152 / V 745
(Servando), vv. 3-4 eu por ben tenho de que lh aqui vin / polo veer; and, in the same text, vv. 15-16 e tenh eu de que o vi ja / que lh gran
ben. (Cf. B 632 bis / V 234 [Calheiros], v. 10, where we could read e pesa me de [do BV] que tarda.)
345
10
15
Despite the objections of Radulet, this may be one place (there is another in B 1223 / V 826 [Baveca], v. 16) where B and
V have dropped different but adjacent words. Pois Deus quis que m inda non morreu makes sense and scans. And the slightly
humourous ethical dative (once called the My dog died on me dative) would be in keeping with this girls tone.
346
10
15
Radulet registers the superscript in B (above the e in madre) as c but it could be e (the final vowel of madre would be elided
anyway). Braga (reading madr e), Nunes, Machado and Radulet all take meu ben e meu lume as addressed to the mother, but
nowhere else in the cantigas d amigo does a girl use this formula when speaking to her mother. On my reading, the phrase
forms part of the girls defense of the boy. We could read lume instead of lum e at the end of v. 10 and get approximately the
same meaning, but lum e places a special emphasis on the initial eu of v. 11.
347
3 por V : Par B
perfaz Nunes
348
PERO DE BERDIA 1
10
15
20
Errar is commonly construed with i. Cf. B 699 / V 300 (Ulhoa), v. 2 fui eu i muit errada and B 1550 (Sevilha / Talaveira), fiinda
ca errastes vos i.
2
Cf. B 483 / V 66 (Alfonso X), vv. 12-13 ben leu / querrn. In A 112 / B 225, v. 7, we find: leu A : lheu B.
349
PERO DE BERDIA 2
10
Se m el desejasse tanto
como dizia, logo ant o
tempo que disse verria,
mais sei que mi non queria
tan gran ben como dizia
15
20
We find the same combination of pronouns with the same verb in B 750 / V 353 (Guilhade), v. 2 e eu non volho creo ben . Given
the rarity of such usage, the close similarity of the two phrases suggests that they are both versions of an archaic formula.
Note that the second person plural pronoun in volh is the only reference (in any of the five cantigas d amigo of Berdia) to a
putative addressee (or addressees).
2
In the cantigas damigo the exclamation ai is invariably used with vocatives, so Bragas punctuation (followed by Nunes) is
improbable. On Monacis reading, the girl shifts from the vocative exclamation Ai fals to a rhetorical question (introduced
by e) with third person verbs.
350
PERO DE BERDIA 3
ala B : ela V
6 casse V : om. B
7 mi~ B : mi V
Lang (1894: cxviii), Nunes (vol I, p. 414 and vol. III, p. 305) and Lapa (1982: 183) see long lines here.
351
PERO DE BERDIA 4
10
15
20
25
30
352
grisado V
20 non o] nouo V
PERO DE BERDIA 5
353
NUNO PORCO 1
10
The only morphological indication of the gender of the speaker is the word namorado, copied in full only here. Since
confusion between a and o is common, editors beginning with Monaci have understandably corrected what they took to be
a simple error. (Compare B 1236 / V 841 [Pai Calvo], v. 2 and note). In doing so they may have overlooked or declined to
acknowledge the possibility that this text deliberately breaks the rules of the genre: if namorado is sound, we have a parody
of a cantiga d amigo (and whoever prefers to read namorado with the manuscripts can adduce as corroborating evidence both the
logic of the discourse [a limp response to the boys sanha] and the location of the text in BV [just at the end of a section with
poems of various genres and immediately before the Amigo section of what Oliveira 1994 calls the cancioneiro de jograis galegos]).
2
My suggestion supplies a syllable in the first half of the verse, where sense and syntax seem to stumble (Nunes) or simply
fizz out (Lapa 1982: 183) in the readings so far proposed. Compare B 1274 / V 880 (Giinzo), v. 4 e, se s assanha, non faz i torto.
The verb assanhar in this sense (to get angry) is always reflexive (except once in oratio obliqua with introductory verb and
infinitive: poilo vi assanhar / escontra mi [B 789 / V 373, vv. 9-10] where the reflexive pronoun is omitted [cf. Cohen 1996b:
9-11 and 40, note 7]). The form xi is suggested by the manuscripts ssi and supported by the position of the text in BV. See,
for example, B 1118 / V 709 (Berdia), v. 2 xi m assanhou and B 1120 bis / V 712 (Berdia), v. 27 xi m assanha.
3
Monaci and Lapa rightly see here the adverbial phrase a torto (wrongly), while Nunes, putting a semicolon after assanhou,
reads tort (but aver torto means to have been wronged and is construed with de plus the offending person).
354
PERO DE VEER 1
355
PERO DE VEER 2
possis sab<e>
3 por] p B
4 boa BV
In strophes of this form (aabb or aaBB) the verses in the body of the strophe should be of equal measure. Hence
verse 2 should scan 11 to match v. 1, but it is difficult to judge where a syllable has been lost. Neither is it clear
whether the text as it stands is complete or fragmentary. Given these uncertainties, maybe we ought merely
report that the text in the MSS scans 11 10 7 11, and that we believe a syllable has been lost in v. 2.
356
PERO DE VEER 3
2 arendi V
5 amigo (a) a B
7 el]
vv. 1, 2, 5, 6: The meter of verses in the body of the strophe is uncertain. Unless we scan Maria (v. 1) as two
syllables, the first verse scans a syllable longer than vv. 2, 5 and 6. However, v. 5 could scan 11 syllables (with
hiatus between amigo | a) and in v. 6 a vowel could have dropped out: aquel<e>. Against this (possible) metrical
norm of 11 the only exception would then be v. 2, where we might venture either atendi <eu> or <el> pos).
357
PERO DE VEER 4
10
15
358
8 gsar B :
PERO DE VEER 5
2 moher B
Lapa (1982: 183) wants long verses in the body of the strophe, short ones in the refrain (assuming we could
call it that in a text consisting of a single strophe).
359
PERO DE VEER 6
10
360
uos V :
BERNAL DE BONAVAL 1
10
Cf. B 1140 / V 731 (Bonaval), v. 7 Pois eu migo seu mandado non ei, and note that in a similar phrase in B 1279 / V 885 (Codax)
v. 4, B and V read e, but N correctly has ei.
361
BERNAL DE BONAVAL 2
10
15
362
desanganar B
8 uoss V : noss ? B
9 m end] meu d BV
11
BERNAL DE BONAVAL 3
10
11 uo leu V : u9 leu B
363
BERNAL DE BONAVAL 4
10
7 ouzo V
In both B and V the text is copied in strophes of 6 lines each; in neither, however, would the space available
have sufficed to copy the text in long lines. Cf. the remarks of Indini (p. 158), who follows the lay-out of the
MSS despite the resultant metrical irregularities. Tavani (Rep.) again offers three possibilities (26:1, 37:2 and
244:2,4,6 [followed by Indini]). Lapa (1929a: 325) takes the form to be aaBB.
1
Though unconvinced by Nunes emendation (accepted by Lapa and Indini), I do feel the need (not shared by Braga and
Machado) for the infinitive ir (cf. v. 11). My suggestion would make the refrain 16 syllables and the overall form would be
aaB 16 16 16.
364
BERNAL DE BONAVAL 5
10
9 que2] q V : n B
On the strophic form, see Lapa 1929a: 305-306; Tavani, Rep., p. 17; Indini, pp. 80, 162.
365
BERNAL DE BONAVAL 6
366
10
BERNAL DE BONAVAL 7
madre <e> ego : madre BV : madr e edd. 2, 5 BV begin new line after
One could lay the poem out according to the rhyme pattern ababC, scanning 12 9 3 4 6 (with breaks after
mal in vv. 2 and 5 and after Bonaval in vv. 3 and 6), but I think it is more significant that the form can be taken
as a simple aaB strophe with internal rhymes. See Indini, ad loc.
367
BERNAL DE BONAVAL 8
368
5 falades B
7 guardedes B
JOHAN SERVANDO 1
10
15
20
25
369
JOHAN SERVANDO 2
10
15
6 usara B
Michalis (vol. II, p. 882) and Lapa (1982: 184) favor long lines (at least for the body of the strophe) and
Nunes (vol. III, p. 317) repents having printed short lines. But in the absence of direct manuscript evidence
and of any indication by Colocci, it may be prudent not to insist. The form can also be taken as aaB or aaBB.
370
JOHAN SERVANDO 3
10
2 n (foy) B
3 e<u> Monaci : e BV
5, 11 non9 V : nou9 B
end] and B
8 o] illeg. V
371
JOHAN SERVANDO 4
10
372
JOHAN SERVANDO 5
10
4 eu] en V
7 n V : no B
11 N B : no V
Short lines in BV; Monaci prints long verses. See the note on the other version (following B 1147 bis / V 750).
1
These two versions, both attributed to Johan Servando, of what is basically the same poem, provide an opportunity to
see how a text could be altered, whether by the poet himself, by a performer, or perhaps through oral or written transmission.
Michalis (vol. II, p. 882) and Lapa (1982: 184) would print long lines.
373
JOHAN SERVANDO 6
10
15
In B Colocci has written 14 syll. in the left hand margin beside the first strophe just beneath the number
(1147) and 14 syllab at the bottom of f. 244va. This, though arguably a slip up on Coloccis part, draws
attention to a metrical problem: whereas vv. 1-2, and 6-7 have 12 syllables, v. 12 must scan 13 syllables and
v. 11 can be so scanned (if we observe hiatus between grado and oraon). If we add a paragogic e to vv. 11 and
12 we would then get up to Coloccis 14. The question remains open.
374
JOHAN SERVANDO 7
3 Hylo B : dylo V
4 el Nunes
The verse is a syllable long and aguisa is suspect. There are three other attestations of non-reflexive finite forms of aguisar
in the cantigas damigo, the normal verb being guisar (cf. note on B 686 / V 288 [J. S. Coelho], v. 3). (There is also one case of
the reflexive aguisar se in B 256, v. 8 [Queimado]). In B 1168 / V 774, v. 7 (Bolseiro) we find desaguisado in a hypermetric verse
where we should read desguisado. Elsewhere Johan Servando uses guisaria in B 1152 / V 745, v. 6 (refrain).
2
In V these lines read:
O meu amigo qmi uos tolhedes
po magora por el mal dizedes
dylo quereu
375
JOHAN SERVANDO 8
10
376
10 hypermetric
uos
JOHAN SERVANDO 9
10
For the relative positions of o and eu in subordinate clauses, cf. (among the cantigas d amigo of Servando) B 1152 / V 745,
v. 7 que o eu vi and B [1145 bis] / V 748, v. 2 des que o eu vi.
2
Cf. v. 4. de is not elided before the pronouns o, os, a, as (Nobiling 1907b: 349; Cunha 1982: 14, 25, 81).
377
JOHAN SERVANDO 10
10
15
20
378
JOHAN SERVANDO 11
3 incomplete?
4 do Nunes : de BV
deseio V : deseyo B
In BV, vv. 1-2 are copied in long lines, vv. 4-5 in short ones. Cf. Lang (1894 : cxviii), Michalis (vol. II, p. 883),
Nunes (vol. I, p. 414, vol. III, p. 324) and Lapa (1982 : 184). In v. 5 we scan 13 eliding both verria a and Servando
e (15 if we observe hiatus twice).
379
JOHAN SERVANDO 12
10
15
20
380
JOHAN SERVANDO 13
10
15
20
381
JOHAN SERVANDO 14
Disseron mi ca se queria ir
o meu amigo, por que me ferir
quiso mha madr, e, se m ante non vir,
achar s end el mal, se eu poder,
se ora for sen meu grad u ir quer,
achar s ende mal, se eu poder
Torto mi fez que m agora mentiu;
a veer m ouv e pero non me viu,
e, por que m el de mandado saiu,
achar s end el mal, <se eu poder,
se ora for sen meu grad u ir quer,
achar s ende mal, se eu poder>
El me rogou que lhi quisesse ben,
e rogo a Deus que lhi dia por en
coitas d amor, e, pois s el vai daquen,
achar s <end el mal, se eu poder,
se ora for sen meu grad u ir quer,
achar s ende mal, se eu poder>
A San Servando foi en oraon
eu, que o viss, e non foi el enton
e por atanto, se Deus mi perdon,
achar s en<d el mal, se eu poder,
se ora for sen meu grad u ir quer,
achar s ende mal, se eu poder>
10
15
20
Braga (followed by Nunes) emends to make the third verse of the refrain (copied in full in the MSS only here) identical to
the first, but the verse scans and makes sense as is.
382
10
15
20
383
10
15
20
384
4 in B, after cpanha where there is a line break, Colocci has written de donas fora on the same line1
To the left of the incipit and at the bottom of f. 246ra Colocci has written 16. syll. In B the first strophe is
copied in two long lines and a refrain, as here (in V, after the first line, the scribe gives up this procedure, for
want of space), while the remaining strophes are copied in short lines.
1
Colocci apparently meant to indicate that the verse does not end at the line-break.
385
JOHAN ZORRO 1
10
15
The correction is paleographically simple and rectifies sense, syntax and meter, yet has been overlooked (Nunes prints
am<or> ; Cunha, ama <e>). The error m for ui (which is common, as Cunha himself observes elsewhere [1986: 78]) occurs
among the cantigas d amigo of Zorro in B 1154 / V 756, v. 6 (e~uiou B : ~emou V). Cf. B 640 / V 241 (Pae Soarez), v. 16, where V
correctly has auia while B has ama; in the version of the same text attributed to Coton (B 827 / V 413) both MSS have auya.
The antecedent of que is amor (amor ... / que avia sempre sigo); cf. B 930 / V 518 (Roi Fernandiz), v. 1 amor ei migo.
386
JOHAN ZORRO 2
10
387
JOHAN ZORRO 3*
Vi remar o navio,
i vai o meu amigo
e sabor ei <da ribeira>
Vi remar o barco,
i vai o meu amado
e sabor ei <da ribeira>
10
15
388
6 e] & B : om. V
7 Vy B : V(A)y V
before 10 (Hi
JOHAN ZORRO *4
El rei de Portugale
barcas mandou lavrare,
e l ir nas barcas sigo,
mha filha, o voss amigo
El rei portugueese
barcas mandou fazere,
e l ir nas barcas sigo,
<mha filha, o voss amigo>
Barcas mandou lavrare
e no mar as deitare,
e l ir <nas barcas sigo,
mha filha, o voss amigo>
Barcas mandou fazere
e no mar as metere,
e l ir <nas barcas sigo,
mha filha, o voss amigo>
10
15
ir (Varnhagen; cf. Nunes, vol. III, p. 721) is the first of three interrelated corrections required by sense (the other two are
sigo for migo in this v. and o voss amigo for e voss amigo in v. 4). The transmitted text says My daughter and your (boy) friend
will go on the boats with me. But who would be talking to whom? The conventions of the cantiga d amigo lead us to assume
that the mother is telling her daughter: and, my daughter, on those boats your boyfriend will be going with him [the king].
This poem is preceded in BV by the cantiga d amor beginning En Lisboa sobrelo mare / barcas novas mandei lavrare, / ai mha senhor
velida. The use of a male speaker there may have thrown off the scribe of BVs source-text here.
2
sigo (Braga; cf. Nunes, vol. III, p. 721; migo in vol. II) would refer to the king, subject of the first clause. For other examples
cf. B 1249 / V 854 (Lopo), vv. 3 (ca [scil. el rei] levou sigo o meu coraon) and 5 (se o el rei sigo non levasse); and B 704 / V 305
(Cogominho), v. 10 (dos que el rei levou sigo). Elsewhere migo must be corrected to sigo in B 747 / V 349 (Guilhade), vv. 3-4 amigas,
ten meu amigo / amiga na terra sigo.
389
JOHAN ZORRO 5
All editors transpose the MSS que lhis farei, madre for the sake of rhyme, but then scrupulously leave the verse a syllable
shorter than v. 4. Elsewhere in the cantigas damigo of Zorro we find the exclamation ai in B 1148 bis / V 751, v. 5 (ai amor) and
B 1149 bis / V 752, v. 3 (ai filha). The exclamation ai madre is common (we find it, for instance, in eight incipits) and seems
particularly appropriate here both to the meter and to the girls mood.
2
The transposition produces a strophe of the form aaBB (Tavani, Rep. 37:57). For those who would maintain the text of BV
as is, there is at least one cantiga d amigo with the rhyme scheme aaBC (B 739 / V 341 [Baian]). Cf. Tavani, Rep. 47:3 (the other
three examples he gives are problematic).
390
JOHAN ZORRO 6
10
3, 8 damor V : Namor B
4, 9 nas]
The verses in the body of the strophe apparently scan 7 8. On the basis of vv. 30-31 of Airas Nunes pastorela
(B 868, 869, 870 / V 454) Pela ribeira do rio / cantando ia la virgo we might be tempted to emend here and read la
virgo in v. 2 and la dalgo in v. 7 (with hiatus cantando | ia). Compare B 1157 / V 759 vv. 11 (a virgo) and 15 (a d algo).
But the difference in scansion, since it is regular, may reflect the use of two different melodies in the body of
the strophe, one for I.1 and II.1 and another for I.2 and II.2 (see M. P. Ferreira 2000).
391
JOHAN ZORRO 7
392
JOHAN ZORRO 8
10
15
3, 7, 9, 13 aima B
vv. 11, 15: meter points to a problem, though sense is satisfactory. We can even the scansion by reading
levar<e> in both verses. Michalis wants levar<des> both metri causa and to clear the girl herself from any role in
transporting the virgo (whoever she is, wherever bound). That scruple aside, one might be tempted to emend
boldly to por levar a <dona> virgo and por levar a <dona> d algo, respectively (cf. B 1155 / V 757, vv. 2 and 7 [and
note ad loc.]).
1
Michalis reads Jus a lo mar e o rio. Nunes prints Jus a lo mar o rio but retracts (vol. III, p. 336; cf. Bell 1920: 62). Tavani (Rep.)
accepts , Cunha prints e. Our perception of syntactic links between odd and even (varying and unvarying) verses and our
assignment of verses to speakers both weigh in this decision. I believe there is only one speaker: the girl.
2
Monaci prints namorado and considers this a cantiga d amor: il cod. ha namorada, ma come si concilia coi vv. 11 e 15?
(1873: 31).
393
JOHAN ZORRO 9
4 amigo que Nunes, Cunha : amig o que Michalis (II, 894), Lapa
fezesse V : tenesse B
5 rio]
The refrain is copied on four short lines in BV. Michalis (vol. II, p. 894, note 3) favors long verses and Lapa
(1982: 185) agrees. The verses in the body of the strophe scan 10 9 in both strophes, which might point to
two different melodies (cf. M. P. Ferreira 2000). Were we determined to even the meter, we could print Per for
Pela in v. 1 (citing B 1150 bis / V 753, vv. 1 and 6 Per ribeira do rio). Lapa (1929a: 314), stretching the other way,
prints con <o> meu amigo.
vv. 2, 6: trebelhei means played (here in an amorous
sense).
394
JOHAN ZORRO 10
10
Cunha defends como (p. 60), but both MSS have come in v. 9, and come is more common in brief comparisons, especially
those not governing a clause. In the imitation of this cantiga by Airas Nunes (B 879 / V 462) both MSS have como enos in v. 9.
Rather than citing that text to decide here, I would be tempted to use this one to emend there.
395
10
15
The disruption of the dobre pattern in vv. 7-8 in BV may be due to an earlier scribe who thought to better the word order,
while failing to notice the technique he was disrupting.
2
Macchi rejects the reading of Nunes (who assigns prazeria to the male speaker), pointing out that there are no examples in
medieval Galego-Portuguese texts of expressions like prazer... prazeria. But Macchis own solution (praz er) is highly improbable,
since ar and er are normally found only before verbs. On the other hand, Roi Marti~iz do Casal uses the exclamation ai three
times in his other two cantigas d amigo (B 1162 / V 765, vv. 1 and 4; B 1163 / V 766, v. 1). In this text we find senhor in v. 4 and
mha senhor in v. 16, so ai senhor here would vary the vocative once again.
396
10
The emendation (adopted by Macchi), based on the parallelism with v. 2, seems required: andar / lidar con Mouros makes
sense whereas morar / lidar does not.
397
10
15
20
Macchi follows B and prints se migo morar, which would mean if he stays with me. This would make sense (though the
verse would be hypermetric), but the absence of a subject (the friend) makes the reading less attractive. (An attempt by a
scribe [who understood the language] to make sense of a difficult reading?)
398
JUIO BOLSEIRO 1
10
15
20
399
JUIO BOLSEIRO 2
10
15
Meter, syntax and sense all point to a corruption. The verse is hypermetric (though Reali signals a synalepha between
e and ante, the conjunction e was not normally subject to fusion of any kind with a subsequent vowel, whether tonic or
atonic [Cunha 1982: 32-47, 81, 99-128, 168]). Further, lh enviasse makes no sense. While Nunes (vol. III, p. 343) imagines that
lh enviasse a dizer ren is synonymous with comeou a falar (v. 11), Reali translates the verse e prima che gli mandassi a dire
qualcosa (p. 25), but produces no parallel (nor is there any) for enviar dizer in this sense. What the corruption hides must
have been the first syllable of a two-syllable first conjugation verb meaning could, was able to, had time to and the
best candidate (given the readings of BV) is uviar, suggested to me by Stephen Parkinson (1998, oral communication), who
points out that the uviass is metrically two syllables, the i in uviar being a semi-vowel. This would be the only occurrence of
uviare (< M. Lat. obviare < ob + *viare < via, ~ae, f. ) in the cantigas damigo. The word does occur, however, in CSM, where we find
the common imperative uviade in the sense of Help! More to the point, in CSM 65.138 ante que huviasse comer nen bocado we
see precisely the syntax and meaning required here. (See also the examples in Lorenzo, s. v. huuyar.)
2
In her critical apparatus (p. 25) Reali wrongly implies that the conjunction is present in V.
400
JUIO BOLSEIRO 3
10
15
Reali resists emending mel to me (or m), but the verse transmitted by the manuscripts is hypermetric. (It seems unlikely
that asperando should be scanned as three syllables, although Cunha [1982: 102-103, 118] suggests that pretonic e could be
syncopated before r in forms of the future, and [p. 147] that other pretonic vowels before an r may have suffered the same
fate.)
2
The verse is metrically and syntactically sound, and the variation in the refrain should be maintained. Nunes proposed
change takes out one word to add another with no gain (el in v. 16 renders the pronoun superfluous in v. 17) and arguably
some loss (eu).
401
JUIO BOLSEIRO 4
Reali prints veir here (ueir is in B, not V, as her apparatus states), but viir in v. 5. But the form veir is not documented
elsewhere in the cantigas d amigo, and among those of Bolseiro viir occurs seven times: B 1167 / V 773, v. 15; B 1169 / V 775,
vv. 2, 12; B 1170 / V 776, vv. 5, 11, 17, 23. Moreover, scribes not infrequently confuse e and i.
2
Reali maintains desaguisado. The verse as transmitted is hypermetric and the more normal form would resolve the problem
(Reali would have us swallow the a in pronunciation anyway). In B 1631 / V 1165 (Ponte), v. 3, desaguisado again occurs in a
hypermetric verse where we should probably read desguisado (against both Lapa and Panunzio). Similarly, B 1364 / V 972
(M. Soares), v. 8. Cf. CA 206, v. 4, where the corresponding version in B 357 has the same unmetrical error, and see note on
B 686 / V 288 (J. S. Coelho), v. 9.
402
JUIO BOLSEIRO 5
10
15
15 uen B : uou V
403
JUIO BOLSEIRO 6
10
15
20
devia<des> is personal here (its subject is vs) and governs an infinitive directly. Cf. CA 31.23 (Pae Soarez) mais mi-o deviades
vs gradecer and 131.14 (Queimado) non me deviades leixar morrer. (Nunes Ca vs ben vos devia <a> nembrar solves neither the
metrical problem [devia a would have been elided; not to mention the unlikelihood of scanning C a] nor the syntactic one.
Reali leaves the line untouched.)
2
These lines read in V as follows:
Par ds falsso mal ssemi gradezeu
quando uos ouurades de morrer
se eu n fosse qu9 fui ueer
mays poys u9 outr ia demi~uenzeu
(uenzeu) s
como(s)u sastes uijr antus me9
404
JUIO BOLSEIRO 7
10
15
20
7
12
15
8
9
11
7
12
15
8
9
11
5
8
7
6
6
8
7
12
15
8
9
11
5
8
7
6
6
8
405
And the seventh is a parody of a cantiga damigo: B 1383 / V 992 (=CEM 385) composed by Pero Garcia Burgals,
scanning: 10 10 5 10 5. See B 831 / V 417 (Ponte) for a similar example (where, however, the middle verse of
the refrain exceeds the outer two by a single syllable and the rhyme scheme is different): madre, namorada me
leixou, / madre, namorada mh leixada, / madre, namorada me leixou; and compare the parody in B 755 / V 358
(Guilhade), vv. 20-22 ja afou // Ja afad pan de voda, / ja afou.
406
JUIO BOLSEIRO 8
10
15
11 no] n B
13 dizestes B
15 fals, acoomhar
Nunes, prints: vedes falsa coor ar xi vos and explains vedes...o que ento vos direi: cr falsa vos toma outra vez, isto , c
tornais com as vossas falsidades (vol. III, p. 349). Reali (no. 10) makes no progress. Lapa [1982: 185-186] is on the right
track, but doesnt see the verb acoomhar. Michalis [1920, s. v. acoomiar], noting the etymology (ad+calumniare; but in Classical
Latin the verb is deponent and means to bring false accusations [OED]), glosses acoimar, levar ou por coima ou multa;
castigar, punir and Lapa (CEM), defines it as dar castigo, proibir, recusar. (Cf. B 172 [M. Soares], v. 8 [CA 398, CEM 283,
Bertolucci no. 29]). The correct reading is due to Joo Dionsio (1991, oral communication).
407
JUIO BOLSEIRO 9
10
15
To the left of the first strophe, and at the bottom of the column (after v. 10) Colocci has written 16 syll and
16 syllab respectively; it seems to make no sense here, however, to print the poem in long verses, since that
would obscure the obvious rhyme-scheme.
vv. 5, 11, 17 lirias: The sense appears to be melodic
embellishments or flourishes. We could render:
He made some lilies in the melody.
vv. 5, 11, 17: lirias is found only in this text, though thrice. Latin lirium (< Greek leirion), lily, gives GalegoPortuguese liro, lilio in CSM and lirias is presumably a variant form.
vv. 7-8 muito ben xe soube buscar / por mi ali: He knew
how to seek out muito ben for himself (xe) on my behalf
(por mi) there (ali).
vv. 7-8 muito ben xe soube buscar / por mi ali: Ele soube
como procurar muito ben para si prprio (xe) da minha
parte (por mi) ali.
vv. 7-8 muito ben xe soube buscar / por mi ali: We should take muito ben as the direct object of buscar (which is always
transitive in the cantigas damigo; cf. B 1172 / V 778 (Bolseiro), vv. 1-2 and B 1183 / V 788 [Campina], v. 16. And since
buscar cannot be both transitive and reflexive, the MSS se seems an error for the old ethical dative xe or xi.
408
JUIO BOLSEIRO 10
10
The h in the manuscripts probably indicates that the i of ouvi is to be pronounced as a y rather than fully elided.
There is no metrical problem (the final vowel amigo would be elided before u), but the decision is more complicated than
might appear. One question is whether the o (in lho) in v. 8 refers to sabor or amigo. If the antecedent is amigo, then amigo u
seems preferable (And if a woman had to have pleasure with a boy since God gave him to her); if it is sabor, then we might
want to read amig ou , (... had occasion to feel pleasure with a boy, or God granted it to her), but amigo u would still make
sense (... since God granted it). Other factors to take into consideration are the use of u in conjunction with Deus in such
clauses as u quis Deus que ll a alma do corpo saysse (CEM 114.73) and u me vus Deus fez prime<i>ro veer (CA 86.9), as well as the
semantic parallel with phrases such as e filhe o ben quando lho Deus der (B 1020 / V 610 [Johan Airas], v. 20).
2
409
JUIO BOLSEIRO 11
10
15
14 e amigo Nunes :
In each strophe, the 4th and 5th syllables of the first verse rhyme with the 10th and 11th syllables of the fourth
verse: amigo (1) / migo (4); vejo (7) / desejo (10); gasalhado (13) / pagado (16). Hence migo, desejo and pagado (last word
in the fourth verse of each strophe) are not, as has been said, palavras perdudas. They do indeed rhyme, but in
a way subtle enough to have eluded the vigilance of Nunes (vol. III, p. 350), Tavani (Rep.) and others, though
not of Lapa (1982: 186).
1
The girl stresses throughout that the addressee is her amigo (cf. vv. 1-2, 8, 10) and the third strophe would be the only one
lacking the key word, were it not read here. Reali objects that la parola, di senso oscuro, stata liberamente interpretata
and prints cuiergo. But there is no such word (though cui ergo sounds like Latin and could, for that reason, have crept into the
text in place of an illegible or only partially legible eamigo).
410
JUIO BOLSEIRO 12
10
Cf. vol I, p. 390, where Nunes suggests quisera-meu outra vegada. See also B 657 / V 258 (Carpancho), v. 7.
411
JUIO BOLSEIRO 13
10
412
11 Pesami B : pesami V
JUIO BOLSEIRO 14
10
15
20
25
The initial p in both manuscripts is not a crossed p but a p with a loop on the tail (=pro).
413
JUIO BOLSEIRO 15
Although Reali maintains the reading of the manuscripts, she translates hanno perduto. We may have a pluperfect for
a perfect, a variant (phonetic or graphic or both) of the perfect, or a scribal error. Though any of these is possible, it should
be remembered that the use of the pluperfect for the perfect was already common in early Latin.
414
MARTIN CAMPINA 1
10
15
9 era] eu V
415
MARTIN CAMPINA 2
10
15
20
3 di(z)xi V
8 uir B : iur V
416
lhe] le B
v. 10: Que nunca me vingue dele nem de outra pessoa. aver direito de (alguen) significa obter vingana
(de alguma pessoa).
PERO MEOGO 1
10
<eu> supplevi1
For the collocation, cf. B 639 / V 240 (Pae Soarez), v. 7 Sofrer lh ei eu de me chamar senhor and B 730 / V 331 (Beesteiros), v. 13
Vee-lo ei eu, per ba fe. Outside the cantigas damigo we find an example at the end of a poem (and in rhyme) in B 1522 / V 395
(=CEM 158 [Gil Perez Conde]), vv. 16-17 E, se non, ficar-m ei eu / na mia pousada.
417
PERO MEOGO 2
10
10 el<e> supplevi
mas BV
The collocation mais ar occurs three times in the cantigas d amor and three times in CEM. It is used immediately following
a negated clause in B 498 / V 81 (Dinis), vv. 9-11 ca sei mui bem que nunca poss achar / nenhu~a cousa ond aja sabor / se nom da morte;
mais ar ei pavor / de mh a nom querer Deus tam cedo dar (text from Lang 1894). Cf. B 1388 / V 997 (Queimado), vv. 8-10 Qual desden
lhi vs fostes <i> fazer / nunca outr om a seu amigo fez; / mais ar far vos ele, outra vez...
418
PERO MEOGO 3
10
8, 10 fez<o> Lapa
In both MSS v. 9 is written on a single line, while vv. 1-8, and 10 are copied on short lines. In B, next to each
pair of shorter verses (except those which make up our v. 3), Colocci has drawn a connecting line (indicating
that they are to be taken together) and written 16 in the right hand margin (16 syll next to our vv. 9 and 10).
Lapa (1982: 187) sees long lines.
419
PERO MEOGO 4
mte : mu~ te ?
Nunes follows Braga, making cervas agree with velidas. See next note.
Mndez Ferrn and Azevedo Filho defend the reading of BV in v. 6 que faria uelidas. The girl, after telling the cervos do monte
that she has come to ask them a question (vv. 1-2, 4-5), would then suddenly turn and address her girlfriends with the
vocative velidas (vv. 3 and 6). The subject of faria would be the boy. Rather than submit to such contortions, it makes more
sense to read velida in vv. 3 and 6 and farei for faria in v. 6 (with Braga, Lapa [1982: 187], Nunes).
2
420
PERO MEOGO 5
10
15
Nunes supplement here, as well as that in v. 4, have been widely accepted by editors. But in no other verse in the
strophes of this text is there a mere inversion (from strophe to strophe) of the two halves of a verse. Bell suggests Leda dos
amores for v. 1 and Dos amores leda for v. 4. Assuming (with Lang 1894: 134 and the opinio communis) that this text underlies
Dinis Levantou s a velida (B 569 / V 172), we can then take the second verse of that poem levantou s <aa> alva as a reflection
of the missing phrase here.
421
PERO MEOGO 6
10
15
20
The imperfect provides an assonant rhyme (as in strophes II and III), but in view of the rhymes in strophes V-VI, the change
to asperara can be defended.
422
PERO MEOGO 7
10
423
PERO MEOGO 8
10
E rompestes i o brial
que fezestes ao meu pesar,
poilo cervo i ven;
<esta fonte seguide a ben,
poilo cervo i ven>
15
E rompestes i o vestir
que fezestes a pesar de min,
poilo cervo i ven;
<esta fonte seguide a ben,
poilo cervo i ven>
20
namorado, attested twice here in BV, is surely an error for cervo, read in both MSS in all following strophes. No doubt cervo
has been displaced in the first strophe by a gloss (note that namorado is not found elsewhere in Meogo). Nunes prints the
change, but other editors follow BV, and Azevedo-Filho unconvincingly defends the gloss. Cf. B 1272 / V 878 and B 1275
/ V 881 (Giinzo) and the note on B 1171 / V 777 (Bolseiro). (The possibility of variation in a refrain is not in question; what
editors who defend the gloss forget is the fundamental requirement of external responsion between strophes.)
2
This is the only occurrence of loir in Galego-Portuguese, which is apparently the only Romance language that preserves a
genuine descendant of Latin ludere (REW 5153a).
424
PERO MEOGO 9
10
15
9 amores
Reckert & Macedo (1996: 122-123) take the refrain as an aside, within parentheses.
425
MARTIN DE CALDAS 1
beni B
Braga and Nunes print lhi, but this seems odd given the presence of lhe in the next verse (though it should be noted that
elsewhere among the cantigas d amigo of Caldas we find lhi once [B 1194 / V 799, v. 9], lhe not at all). The girl refers in v. 2 to
an encounter with the boy tomorrow, so the force of i could be as much temporal (cf. CEM 14, v. 11 [Alfonso X]) as spatial.
426
MARTIN DE CALDAS 2
10
15
Nunes supplies <gran> before mal, but the loss of an initial e, especially before eu, seems more likely, and the conjunction
would provide a missing connection to the syntax (the third verse in each of the other strophes begins with e).
427
MARTIN DE CALDAS 3
10
15
20
10 <o> Nunes;
Braga and Nunes print senhora, se but senhora is extremely rare (and late) in Galego-Portuguse lyric (the formula madr e
senhor is used by Caldas at B 1194 / V 799, v. 1; cf. v. 16 below) and asse (with manuscript variants including ase) is found in
CSM in optative clauses such as asse Deus vos perdon, asse Deus m ampar and asse Deus me valla (cf. Mettmanns glossary, s.v.
asse). The space in B between senhor and a is very slight and may be unintended.
2
The reading of V is not entirely clear, and could be fabr.
428
MARTIN DE CALDAS 4
10
429
MARTIN DE CALDAS 5
10
15
7 quen] que V : q B
13 boa B
In the upper right hand column of f. 126r the scribe has copied and cancelled part of a hybrid strophe beginning with
v. 1 of the next poem, then skipping back to the end of v. 1 of this one (stes meus) and continuing down to v. 5. It reads:
Nostro senhor e como po stes me9
olhos e coyta do meu coraon
por que tardastes a muy gram saon
n mho neguedes seu9 ualha de9
ca eu querendauerdade saber
In the part of this hybrid relevant to the present text, the only differences between this version and the one the scribe had
just copied are in v. 3, where the first version has grsazon (exactly as in B).
430
MARTIN DE CALDAS 6
10
15
10 s e<l>
Cf. Nunes, vol. III, p. 363: Se bem os interpreto, stes versos querem dizer: irei com le, porque, sempre ho de falar
desta morte, que ser desventurosa. O emprgo de fr, em vez de ser, explicar-se-ia por se tratar de um futuro hipottico.
431
MARTIN DE CALDAS 7
10
15
432
14 meu V : men ? B
NUNO TREEZ 1
Des quando vos fostes daqui, meu amigo, sen meu prazer,
ouv eu tan gran coita des i qual vos ora quero dizer:
que non fezeron des enton os meus olhos, se chorar non,
nen ar quis o meu coraon que fezessen, se chorar non
E des que m eu sen vs achei, sol non me soub i conselhar
e mui triste por en fiquei e con coita grand e pesar,
que non fezeron des enton <os meus olhos, se chorar non,
nen ar quis o meu coraon que fezessen, se chorar non>
E fui eu fazer oraon a San Clemen e non vos vi
e ben des aquela sazon, meu amigo, ave~o m assi
que non fezeron des enton <os meus olhos, se chorar non,
nen ar quis o meu coraon que fezessen, se chorar non>
10
5 csselhar V : sselhar B
Short lines in BV. Taken by editors, perhaps rightly, as ababCCCC, scanning 8 throughout, but compare
B 1202 / V 807 and B 1203 / 808 (Treez). As printed here, there are internal (mid-verse) rhymes in all verses.
433
NUNO TREEZ 2
mal V
4 cleme~co B : demeno V
8 craedor V
9 dormirey V : om. B
The variant is striking, but this is the only time (in either manuscript) in the cantigas d amigo of Treez where the name
appears with initial cr instead of cl (cf. v. 4 below; B 1202 / V 807, vv. 1, 4; B 1203 / V 808, vv. 1, 4, 7).
434
NUNO TREEZ 3
10
15
435
NUNO TREEZ 4
10
15
Nunes supposes that by putting vv. 16-18 as the fourth strophe we would have a case of cobras ternas and both the
transposition and the classification are accepted by Tavani (Rep.), but in all of Galego-Portuguese lyric the only true case of
cobras ternas is Calheiros-1. There the verses (outside the refrain) in strophes I-III rhyme in -igo, while those in strophes
IV-VI rhyme in -ia. What we would have here, if we transposed, would be a different phenomenon (something like a triple
version of cobras alternantes): three rhyme sounds (-on, -ar, -er), one per strophe, in a series established in the first three
strophes and then repeated in the last three.
436
PERO D ARMEA 1
10
15
20
437
PERO D ARMEA 2
438
6 possis outr a
10
15
PERO D ARMEA 3
10
15
3 <en>o Nunes
el<e> Lapa1 4 mi~ B 7 boa B 10, 14 mhas] Mays B : mays V
que] fort. ca scribendum 14 amigus corrected to amigas B
10 amigas]
Lapa (1982: 188) is right to reject Nunes <en>o as unnecessary. See, for example, B 1586 / V 1118 (=CEM 44 [Alfonso X]),
v. 14 que o dia que s a lide juntou.
439
PERO D ARMEA 4
10
15
440
13 <eu> supplevi
17 estes B
15 vos
10
15
15 cofonda V
Cf. B 1213 / V 818 (Sevilha), v. 20 nunca pode per mi saber al and B 1239 / V 844 (Padrozelos), vv. 10-11 nen vs non podedes saber
/ nunca per outren... Here the preposition por introduces a substantival verb-phrase in which alguen is the subject of entender
and no the direct object (elsewhere Sevilha uses entender with a direct object at B 1216 / V 821, v. 27 volo entendern). Longer
examples (by the same poet) of such causal phrases introduced by por are to be found in B 1659 / V 1193 (=CEM 311), vv. 15-17 E polo bispo aver sabor / grande de Conca non aver / non lho queremos ns caber (caber BV : saber Lapa, Marroni) and B 1660 / V 1194
(=CEM 312), vv. 10-12 Si quer a gente a gran mal vo-lo ten, / por irdes tal molher gran ben querer, / que nunca vistes riir nen falar.
2
Marroni offers a plausible reading, but builds on what seems to me already an error in the text of BV. In the cantigas d amigo
there are numerous places where one or both manuscripts have n and the correct reading is unquestionably no: B 563
/ V 166 (Dinis), v. 20 (no B : n V); B 634 / V 236 (Sandin), v. 10 (no B : n V); B 795 / V 379 (Talaveira), v. 8 (no B : n V); B 1172
/ V 778 (Bolseiro), v. 11 (no V : n B); B 1176 / V 782 (Bolseiro), v. 11 (no B : n V) in v. 7 the scribe of V catches himself and
crosses out the til; B 1249 / V 854 (Lopo), v. 4 (no V : n B); B 1268 / V 874 (Cangas), v. 10 (no V : n B).
441
10
15
442
10
15
20
6 nunca vi scripsi2 : n mi V : N ui B
7 uos BV
11 suppl.
The strophic form (incorrectly reported in Rep.) is AbbaCDD (cf. B 695 / V 296 [Ulhoa]; Rep. 176:3), scanning (if
we accept nunca vi for nmi [V] in the second verse of the refrain) 7 throughout.
1
In both MSS the initial verse of the refrain is missing in the first strophe (arguably due to the similarity between the
beginnings of the two verses presumably nu~ca ui [v. 5] and n ui [v. 6]), copied partially in the second and fully in the third.
Yet Nunes and Marroni, though registering the evidence in apparatu, have failed to take heed (Nunes prints a two line
refrain beginning with Non o vi, des que fui nada and Marroni follows). Non o vi nen quen o visse / nunca vi is an emphatic inversion
of expressions like e eu vy quena viu (CSM 339.21; cf. 253.34). So CSM 195.160-163 Non foi quen oysse / nunca neno visse / est and,
in the cantigas d amigo, B 383 (Mafaldo), v. 6 Nunca vistes, amiga, quen tal amigo visse.
2
In such phrases (since I was born) nunca is the norm (not non). Cf. B 1098 / V 825 (Sevilha), v. 3 pois fui nado, nunca vi tan bela.
In the text before this one (B 1209 / V 814, v. 5) we also find nunca vi.
443
10
15
20
444
10
15
13 el] ei V
bo B : bona V
15 pecado] pado V
seu] eu V
For arguments in favor of Talaveiras authorship, see Marroni, pp. 192-194. This text probably does not belong here, but
somehow at some point higher up the stemma intruded (or was thrust) into the set of cantigas damigo of Pedr Amigo de
Sevilha (which would thus consist of ten poems, not counting the pastorela).
445
10
15
20
vv. 13-14 The subject of queira is the outro of v. 4, while the whole clause que mho non queira is the object of me
partir in v. 14 (livrar-se is one of the glosses given by Mettmann for partirse). (Marroni misreads both syntax
and sense here.)
446
10
15
15 mo B
vv. 1-2 stand on one long line in B, but this may be a matter of layout in the exemplar or its source rather than
a clue to strophic form. If instead of ababCC we take the form to be aaB or aaBB the dobre (amigo vv. 1, 3; feito
vv. 7, 9; dito vv. 13, 15) merely becomes internal (cf. B 747 / V 349 [Guilhade]). Perhaps the complexity of
strophic forms in Sevilhas other cantigas damigo may be adduced as evidence for seeing ababCC here.
1
Monaci reads mi~ in V, but the mark over the i may be an extended dot rather than a til.
447
10
15
20
25
448
10
15
20
25
449
v. 21 pois por seu for: Cf. B 547 / V 150 (Dinis [pastorela]), v. 7 por vosso vou eu, v. 18 por voss andarei and v. 20 por vosso vou.
vv. 23-24: It seems we must take creent as active,
agreeing with el: He must make the song for her
(lh-o) as if he really believed in it (ben / creent with the
enjambment providing extra emphasis) and yet you
must not get jealous.
1
In B, in the right hand margin beside this verse the scribe has written morre~ c pesar (the text has moyre~ c pesar) with a dot
over the first r (so that morr is indistinguishable from moir).
450
10
15
20
25
451
The rhyme-scheme in strophes I and III is abbcac, whereas in strophes II and IV we find aaabab. Such alternation
is without parallel in the cantigas damigo (but cf. B 798 / V 382 [Sandeu] and note).
1
In the cantigas damigo the verb is nearly always or and not ouvir, hence o unless we are to understand a geolinguistic or
sociolinguistic distinction here. We should not forget, also, that in oi, oui and ouui we may have three different ways of trying
to represent the same sound (English w).
452
10
15
20
453
454
10
15
20
25
30
35
455
456
JOHAN BAVECA 1
10
15
20
11 moire B
457
JOHAN BAVECA - 2
10
15
20
In vv. 6-7 the phrase e eu por lhi fazer / ben, though maintained by all editors, is anomalous. I find no parallel for putting any
member of a por phrase (causal or final [as here; cf. Zilli ad loc.]) before the preposition itself. There is a close parallel,
however, for the proposed construction in B 1031 / V 621 (Johan Airas), vv. 6, 12, 18 por lh eu non dar o meu (which is causal:
since I dont give him my ben). I have merely corrected muito eu (B) to muito e and assumed that lhi (BV) is an error for lheu
(cf. B 754 / V 357 [Guilhade], v. 17, where BV have lhi for lheu correctly copied in both MSS in vv. 5 and 11; and B 659 / V 260
[Carpancho], vv. 7 [lheu BV] and 13 [lhi BV]).
2
Each MS omits one of the words. Nunes and Zilli print Deus vos but cf. B 1234 / V 839 (Baveca), vv. 5 (que lhi Deus quiser dar)
and 12 (pois lho Deus d).
458
JOHAN BAVECA 3
10
15
20
Zillis emendation morrer is unnecessary (despite the parallel in v. 11). In mort era the imperfect has modal force. Cf. CSM
235.45-47 (e tan mal adoeceu / que quantos fisicos eran, cada hu~u ben creeu / que sen duvida mort era; / mas ben o per guareceu / a Virgen
Santa Maria). Braga and Machado read correctly here (Nunes morto ser, se lhi non valedes is indefensible).
459
JOHAN BAVECA 4
10
15
460
JOHAN BAVECA 5
10
15
20
The first letter in V is illegible, and the reading of both MSS may be u9.
461
JOHAN BAVECA 6
462
16 ] a V : al B
10
15
JOHAN BAVECA 7
10
15
10 n V
vv. 11-13: Por outras palavras: Mesmo que no houvesse acordo (preito) entre ns, como de facto h, era
tua obrigao, pelo ben de que sempre tanto precisaste, fazer tudo o que eu te ordenasse e, no entanto,...
This verse, two syllables short in BV, has resisted correction. Editors have understood it as a relative clause modifying
preito (v. 11), but vos foi ... mester cannot be construed with per qual (that is, we find no constructions of the type mi mester *per/
por). Rather, the verse forms part of the coming independent clause, as if we had (in reconstructed prose) por aquel ben que vos
foi sempre <mui> mester deviades fazer por mi que quer. The logic of por qual <ben> inversely echoes that of v. 4 nen por ben. Though
such incorporation of the antecedent in a relative clause (by means of qual [e. g. qual preito in v. 11]) is not common in the
cantigas d amigo, I find four other examples involving ben: B 626 / V 227 (Calheiros), v. 13 Fiei m eu tant en qual ben m el queria;
B 1152 / V 745 (Servando), vv. 5-6 se soubess eu qual ben el querria / aver de mi and vv. 19-20 e assanhar m ei un dia, / se m el non diz
qual ben de min querria; and B 559 / V 162 (Dinis), vv. 9-10 averian d entender por en / qual ben mi quer. A telling parallel for the
conjectures proposed is A 75 / B 188 (Torneol), v. 16 que me d seu ben, que m mui mester.
2
The phrase mui mester is more common outside the cantigas damigo, but see B 801 / V 385 (Sandeu), vv. 13-14 E a min era mui
mester / u~a morte que ei d aver and B 1199 / V 804 (Caldas), v. 7 E atal preito m era mui mester. Baveca uses a variant in B 1231 / V
836, v. 22 mais en tal feito muit mester al. Compare A 262 / B 439 / V 51 (Fernan Velho), v. 19 E se el sabe que m mui mester where
mui is present in BV but missing in A.
3
The verse is hypermetric. Nunes (vol. III, p. 376) says faz-se a costumada contraco entre mi e a unlikely when mi is
tonic. The preposition a (though common enough with dever) is not necessary. Elsewhere among the cantigas d amigo of
Baveca dever occurs in B 1223 / V 828, v. 18 and B 1232 / V 837, v. 6. In both texts it governs fazer as here, without a
preposition.
463
JOHAN BAVECA 8
10
15
5, 10, 15 uirey ? V1
vv. 5, 10, 15: It makes more sense to take these verses as constituting a refrain cum variatione than to call the
text a cantiga de meestria with palavra rima at the end of each strophe (ben), as does Tavani. Colocci, arguably
depending on the evidence at his disposal, has written tornel vario (above the text, to the right) in B and has
marked the refrain (of strophes I and II, though not of III) in his usual manner. Braga, Nunes and Zilli (see
p. 116) have rightly followed Colocci.
1
B reads Jurey (5, 10) and iurey (14) but in V, at all three places, either the i of iurey has been placed to the right of the letter,
or the scribe has written uirey (read by Monaci).
464
JOHAN BAVECA 9
10
15
20
465
JOHAN BAVECA 10
10
15
20
466
JOHAN BAVECA 11
10
15
11 pardon B
13 boa B
16 me] meu B
poder om. V1
In V the verse extends to the extreme right hand margin and the last word visible is seu. If poder is there, it is hidden in the
binding.
467
JOHAN BAVECA 12
10
15
20
The verse has caused needless problems. Bragas pois is presumably a causal conjunction, but the clause it allegedly
introduces is then left hanging (cf. Nunes awkward explanation ad loc.). Zilli prints e, por aquestas guardas, tantas son (and, as
to these guards, theyre so many), but this reading, though faithful to the MSS, yields awkward syntax and sense. The
apparent difficulty disappears if we merely divide aquest as. The sense is: and because of this the guards are so many
that... (aquest refers here to the boys eagerness to fulfill his hearts desire even though everyone could tell what was
happening [vv. 7-9]). In his cantigas d amor Baveca uses e por aquesto several times (see Zillis glossary, s. v. aquesto).
468
JOHAN BAVECA 13
10
15
We could defend e (B) as exclamatory, and the devil needs no article, but given the frequent confusion between e and o
(found later in this very line in the variant leue [V]), Nunes correction is probably right.
469
PERO D AMBROA 1
10
15
vv.13-14: O sentido : Continua a ser to brigo quanto queiras (no que dizes s outras a meu respeito).
vv. 13-14: In andad ora de camanho preito / vs vos quiserdes andar we have sophisticated syntax, with the antecedent
preito incorporated in the relative clause. Curiously, such constructions are more common in cantigas de maestria
and cantigas damor generally than in cantigas damigo, but see the notes on B 1228 / 833 (Baveca), v. 12, and
B 659 7 V 260 (Carpancho), vv. 6-7, 12-13. Though I find no parallel in Galego-Portuguese lyric for the phrase
andar de preito, cf. Queimado-2, v. 15 e ande comigo a guerra. The phrase draws on a meaning of the word preito that
Lapa (CEM) glosses as disputa, briga. This in turn derives from a sense Mettmann glosses as pleito,
controvrsia jurdica, demanda judicial. Cf. Niermeyer, s. v. placitus 26 litige, procs, action / claim, plea,
lawsuit. Among the citations found there, in pace teneat eam et sine placito is particularly telling.
1
All editors follow Michalis. Though peiurare is a common form of the Latin verb, it seems unlikely that we have a genuine
spoken form here.
470
PAE CALVO 1
10
15
e seclusi3
Lapa (1982: 189-190) advises printing the poem in long lines, as in the first strophe in both MSS and in the
second strophe and the first verse of the third strophe in B.
1
This presumes that the contrast between first and third person subjects would have been emphatic: el viv n coitado, moir
eu con desejo.
2
Though the reading of B is not clear, one is tempted to read en and not eu on purely paleographic grounds. But there is
another consideration: eu would be modified by coitado (unless this were taken as an adverb, which no editor has suggested)
and we would have a boy speaking to his mother about his boyfriend (odd for a cantiga d amigo, but cf. B 1127 / V 719 [Nuno
Porco], v. 3 and note). Lapa (1982: 190) would read e viv eu coitad e thereby eliding the offending o on coytado (BV). On our
reading, the o ought not to be elided, given the midverse rhyme. The adverb n would refer to the reason why the boy is sad,
namely, that there has been a rupture in the relationship (v. 1).
3
The verse is apparently hypermetric with the conjunction, and though we could accept a light syllable here (before the initial
accent of the second hemistich), asyndeton, strengthened by chiasmus (and with change of subject), seems more likely.
4
Where there are two witnesses (A and BV), we find numerous examples of the variation eu / -e and eu /-o for example: CA
89 (Burgals) se soubess eu B : se soubesse A; CA 205 (Ulhoa) atend eu B : atendo A. Similar cases, not however involving verbs,
occur at CA 99 (Burgals) quand eu B : quando A; CA 293 (Calvelo) sempr eu levei A : sempre levei BV.
471
PAE CALVO 2
10
7 E] Ee V
Editors have inexplicably followed the MSS, which invert the order of strophes I and II. But the initial strophe (second in
BV) is marked by several elements typical of incipits, including exclamation (ai), vocative (madre), euphemistic relative
clause (o que... [instead of amigo, but to identify the boy]) and generic markers of temporal and spatial deixis (ora daqui).
Moreover, there are six incipits (in the corpus of 500 cantigas d amigo) which begin with Ai madr and two of them begin as
here Ai madr, o que: B 837 / V 423 (Ponte) Ai madr, o que me namorou; and B 681 / V 283 (J. S. Coelho) Ai madr, o que eu quero
ben. On the other hand, perjurado is not found elsewhere in the incipit of any cantiga in the corpus of medieval GalegoPortuguese Lyric.
472
MARTIN PADROZELOS 1
10
15
me BV
473
MARTIN PADROZELOS 2
10
15
20
15 amiga]
Reading se achardes (as does Nunes, though without variants or explanation), the girl is telling the boy to return if he finds
anyone who contradicts her. Though perhaps less likely, sen achardes can be defended (since you wont find).
474
MARTIN PADROZELOS 3
10
15
20
quei]
475
MARTIN PADROZELOS 4
10
15
20
476
MARTIN PADROZELOS 5
10
15
15 algu~ a B : alguna V
477
MARTIN PADROZELOS 6
13-18 om. B1
17 fremosas V
478
10
15
MARTIN PADROZELOS 7
E ando maravilhada
por que tanto tarda, se
viv e sabe, per ba fe,
ca viv oj eu tan coitada
que me non pode guarir ren
<de morte, se cedo non ven>
10
boa B
479
MARTIN PADROZELOS 8
10
15
480
5 mi (nc) ? B
MARTIN PADROZELOS 9
481
LOPO 1
482
u9 B : n9 ? V
5 de V
LOPO 2
10
15
The verse seems to be missing two syllables. Tavanis agora makes sense (and cf. B 1254 / V 859, v. 1), but ora occurs in
strophes II and III and, in the cantigas d amigo of Lopo, appears twice at the beginning of a poem (B 1250 / V 855, v. 1 and B
1255 / V 860, v. 2). See also B 1149 bis / V 752 (Zorro), vv. 7-8 Que coit ouv ora el rei de me levar / quanto ben avia nen ei d aver?
Further: filhou <s ora> because the verb is odd here without the reflexive pronoun (though filhar is flexible in this regard;
cf. the glossaries of Mettmann, Lapa). Lapa (CEM) glosses filhar-se as meter-se, entregar-se, aplicar-se.
2
The verse scans a syllable short. Nunes suggests oj<e> assuming that the final vowel is not elided before a pause at midverse (cf. the rhythm of v. 16). Also possible is mund<o> ei.
483
LOPO 3
10
15
This is not the only place the MSS have malhi for mal lhi (cf. B 1047 / V 637 [Johan Airas], v. 6).
The verse scans a syllable long. See Introduction, 5.
484
LOPO 4
10
5 Ne hunha B : ne hunha V
9 li V
10 non] n a BV
The repetition of a word in rhyme, not to mention of a whole verse, is suspect (but cf. B 1256 / V 861 [Galisteu Fernandez],
vv. 3, 15 and note). If this verse is corrupt in whole or in part, it might have ended sen mho dizer (cf. B 1250 / V 855 [Lopo],
v. 9 non mho dizia and B 925 / V 513 [G. Garcia], vv. 13-14 Por que se foi, e o ante no vi / sen mho dizer, a cas del rei morar) or sen lho eu
dizer (cf. B 847 / V 433 [Reimon Gonalvez], v. 2 sen volo eu mandar).
485
LOPO 5
Filha, se gradoedes,
dizede que avedes
Non mi dan amores vagar
Filha, se ben ajades,
dized e non menades
Non mi <dan amores vagar>
10
486
8 hides B : lhides V
10 luter B : louter V
11 cuydd V : cuydad B
LOPO 6
4 mi Nunes
<vs> supplevi
7 re~ B : te~ V
9 leu B : len V
487
LOPO 7
488
LOPO 8
10
9 Assanhou
All verses in the body of the strophe (1-2, 5-6, 9-10) scan 12 syllables (pace Nunes [vol. III, p. 398] and Tavani
[Rep. 37:29]). This count is reached by restoring the final vowel of madr<e> in v. 1 and by restoring an e in
Assanhou se, e in v. 9 and then observing hiatus between pronoun and conjunction: assanhou se | e. Though not
regular, this hiatus can be supported by close parallels (for instance in B 1203 / V 808, [Treez], v. 11 ar comeou
m<e> | a falar, where meter also requires that we observe hiatus after the combination finite verb + pronoun). Cf.
B 1165 / V 771 (Bolseiro), v. 22 mais mostra mi | as noites d avento (mi as Nunes: mhas BV). In v. 5 we twice observe
obligatory hiatus: Sabe | o San Leuter a que | o eu roguei; and in v. 6, once: que | o.
1
On the restoration of the final e of madr<e>, see (at a few folios distance in the MSS) B 1273 / V 879 (Giinzo), v. 4: madr<e>
aver Nunes: madraver BV where the correction brings that verse into line with eleven others.
489
GALISTEU FERNANDIZ 1
10
15
20
P_ dordes B : perdoardes V
13 d9 B : da V
16 este B
19 Veiasse B
vv. 2-3, 14-15: The irregular repetition of two rhyme-words in the same position in the strophe is without
parallel in the cantigas d amigo. The exact repetition of v. 3 in v. 15 is also odd, but compare B 1251 / V 856
(Lopo), where v. 3 = v. 8.
1
Cf. B 675 / V 277 (Avoin), vv. 20-21 mais por esto ca por outra razon, / por vos veer and A 85 / B 189 (Burgals), vv. 9-11 nen mi-a
mostrou se por aquesto non: / por aver eu eno meu coraon / mui grave coita ja...
490
GALISTEU FERNANDIZ 2
10
15
20
3 dormuj V
5 boa B
6 por] poz V
16 en] eu B
17 bo V : boa B
The MSS e makes little sense here (Nunes [vol. III, p. 400] shows discomfort with it). The girl doesnt need to find out if
she herself is sad. Rather she wants to know if he is pining for her just as (com) she is for him. Presumably an abbreviation
was misread here.
491
GALISTEU FERNANDIZ 3
10
15
20
25
492
GALISTEU FERNANDIZ 4
10
vv. 3-4: Que me queria fazer tanto mal quanto gostaria que eu lhe fizesse a ele, ou seja, nenhum.
B reads mora el + amigas while V offers mora amigas. Thus B has a + precisely where a word is missing in V. Moreover, it is this
word, el, that seems to be wanting in v. 6.
2
Though both MSS have the imperfect here, in vv. 7 and 11 (at the corresponding place in the other strophes) they have the
conditional (possibly because of confusion with the conditional in v. 4). The imperfect seems preferable, both semantically
(he allegedly did her harm in the past, and queria refers to his intentions then) and stylistically (marking a contrast with the
conditional in the next verse).
3
Nunes fails to register the reading of the MSS in his apparatus (vol. III, p. 563).
493
LOURENO 1
10
15
20
494
LOURENO 2
~
Ua moa namorada
dizia un cantar d amor,
e diss ela: Nostro Senhor,
oj eu foss aventurada
que oss o meu amigo
com eu este cantar digo
A moa ben parecia
e en sa voz manseli~a
cantou e diss a men~i a:
Prouguess a Santa Maria
que oss o meu amigo
<com eu este cantar digo>
Cantava mui de coraon
e mui fremosa estava,
e disse, quando cantava:
Pe eu a Deus por pedion
que oss o meu amigo
<com eu este cantar digo>
10
15
hun V : huu B
9 Cantou B : tton V
495
LOURENO 3
10
15
20
496
5 Dizedamigas B : dizedemi(r)gas V
LOURENO 4
10
15
Tavani reports that the o is in B, but the letter in question seems to be a small s with a vertical bar to the right: s|sey.
497
LOURENO 5
10
8 por1] pr B : pre V
10 nos] nus V : u9 B
Editors have seen no need to emend, but the syntax is parallel to that of vv. 1-2: subordinate clauses in v. 4 (since I first
saw him and spoke to him), independent clauses in v. 5.
498
LOURENO 6
10
15
20
<o> sen
In the second of two negated clauses we often find nen ar. Among examples before forms of aver we find a parallel in B 88
(Sandin), v. 9 que o non sabe, nen ar poder / de o saber.
499
LOURENO 7
10
15
B often has ir where V (correctly) has rr and this variation is not registered in the apparatus (see Introduction, pp. 42 and 64),
but here the scribe of B seems to have misread rr as ir and replaced the supposed i with y.
500
GOLPARRO 1
10
el B : e V : possis el<e>
5 short lines BV
8 om. BV1
Meter and strophic form uncertain. In Nunes version the strophe has three verses, and the refrain four, the
first two of which consist of our v. 4, split in half. Tavani (Rep. 44:1) accepts this reconstruction, but whereas
v. 4 is syntactically linked to v. 3, it can hardly stand alone after his v. 10. In B, Colocci has marked the refrain.
The transmitted text appears to have the rhyme scheme aabbB and to scan:
11 11 10 10 10
11 11 11 10
If the verse presumed to be missing had ten syllables, the only irregularity (assuming one measure in the first
two verses of each strophe and another in the third and fourth) would be v. 9, and among the cantigas d amigo
the last verse before the refrain in the last strophe (Rv-1) can scan long (Introduction, 5). Nunes version,
though it provides an isosyllabic cantiga (thanks to the emendation of v. 3 and the alleged four verse refrain,
scanning 5 5 5 5), yields a strophic form unparalleled in Galego-Portuguese lyric. The strophic form as printed
here corresponds to no. 42 in Rep.
1
There is no sign of a missing verse in the MSS, but the repetition of ca at the beginning of successive verses (our vv. 7 and 9)
would be awkward, and v. 9 appears prelude to a refrain; so the third verse of the second strophe seems to have dropped out.
501
JOHAN DE CANGAS 1
10
15
20
502
6 uistes B : iustes V
7 p B : pre V
JOHAN DE CANGAS 2
10
15
16 dix] diz B
<aqu>esta supplevi2
Piel (1989: 274) calls this the survival of um elemento formativo muito arcaico, privativo dos falares galaico-portugueses,
uma autntica relquia do latim popular do noroeste da Pennsula Ibrica, namely the Latin tristedo, -edine(m), of which this
appears to be the only example in medieval literature.
2
The verse scans a syllable short. The phrase occurs elsewhere at verse end: B 429 / V 41 (Estevan Faian), v. 10 e servio vos por
aquesta razon; B 513 / V 96 (Dinis), v. 26 e por aquesta razon. See A 114 / B 230, v. 18 where we find aqueste A : este B.
503
JOHAN DE CANGAS 3
504
MARTIN DE GIINZO 1
10
4 pr V : pre B
amado] amigo B
8 digo] falo B1
10-12 om. B2
Colocci has written 14 syll. to the left of vv. 1-2 and at the bottom of f. 286rb.
1
2
The scribe of B has skipped down to the last word of the next strophe (and so of the poem).
Here V reads:
Camenuyou mdado qsse uay no fossado
eu a sc ceilia de cora o falo
e pr el uyuo
505
MARTIN DE GIINZO 2
10
15
<por> supplevi1
5 ca] sa V
Cf. B 739 / V 341 (Baian), vv. 1-3 Ir quer oj eu fremosa de coraon, / por fazer romaria e oraon / a Santa Maria das Leiras.
506
9 eno]
MARTIN DE GIINZO 3
10
15
20
15 E B : om. V
The verse is a syllable short. The adverb i would refer both back to u chaman Santa Cecilia (v.11) and forward to i in the
second verse of the refrain. The sense would be: since she has brought him (o que ben queria) to me there. Elsewhere in the
cantigas d amigo the verb aduzer is construed with a direct and an indirect object (to bring something to someone), but
cf. CSM 123.42 poi-la candea adusseron y. Nunes prints pois madusso que <eu> ben queria which could be right.
507
MARTIN DE GIINZO 4
10
15
V reads:
Morreru9 ey c aquesta perfia
esseme leixassedes hir guarria
con meu amigo
508
MARTIN DE GIINZO 5
10
The correction has been made for the sake of consistency within this text, but, as in vv. 2, 5 and 10, the verse scans 9.
509
MARTIN DE GIINZO 6
10
15
20
510
17 q V : Q B
MARTIN DE GIINZO 7
5 coytado B : cuydado V
6 periurado B : peruirado V
after 6 m~
The rhyme scheme aaaBBB is not found without refrain in any cantiga damigo. As printed here the text scans
12 12 12 7 8 6. If we want 7 7 7 in the refrain we can cut mi in v. 5 (or take que as one syllable, against the
norm) and read o <meu> perjurado in v. 6.
1
That is, M<artin> Codaz: I dont find this one with musical notation (pontada). This marginal note, probably incorporated
into the text of the exemplar of BV, apparently refers to the sixth cantiga of Martin Codax, which is copied in V on f. 140r and
found without musical notation in the Pergaminho Vindel. Gonalves (1989: 632) thinks that this poem may have immediately
preceded that one in a livrinho of cantigas de romaria.
511
MARTIN DE GIINZO 8
10
8 (uagar) lezer V
9 muyr V
11 <e>
vv. 2, 5, 7, 10 aduffe: pandeiro bimembrafone quadrangular [M. P. F. ]. The word, found only here in GalegoPortuguese lyric, is Semitic in origin. Basically a frame-drum, a small hand-held percussion instrument, the
aduffe is known since ancient times. The New Grove Dictionary of Music and Musicians cites Hebrew tof (hatof with
the article, from the root tpp to strike [HALOT, vol. IV, pp. 1771-1772, 1779]), though the precise kind of
instrument meant by tof in Biblical Hebrew is likely to remain unclear (hand-drum, tambourine, according
to HALOT [vol. IV, p. 1772]). While the mention of the aduffe in this cantiga of Giinzo does not necessarily imply
any kind of dance [M. P. F], maybe we ought not rule out the possibility in view of the somewhat enigmatic
sense of the song (on the dancing girls in the miniatures of the Cancioneiro da Ajuda, see the fresh observations
of Manuel Pedro Ferreira in Cantus Coronatus [forthcoming]; three miniatures [those marking the beginning of
the sets of cantigas attributed to Pero Garcia Burgals, Roi Paez de Ribela and Afonso Lopez de Baian] show a
girl with tambourine in hand).
vv. 8, 11: the implied subject of davan may be the same as the implied indirect object of mandou in vv. 2 and 5
(she ordered them to play the aduffe, so they didnt give the instrument any rest); but then again it may be amores
(from the refrain), in which case lhi refers to the girl: love would give her no rest. See B 1252 / V 857 (Lopo),
refrain non mi dan amores vagar and B 644 / V 245 (Torneol), v. 17 e non mi dan seus amores vagar.
512
MARTIN CODAX 11
10
For the text of Martin Codax all interested readers should consult Ferreira 1986. A full transcription is provided there in
the anexo immediately following the color reproduction of the pergaminho (which itself follows p. 73) and immediately
preceding the black and white blowups of each cantiga (which end just before p. 84). In what follows I have accepted in the
text and noted in the apparatus such changes as are justified on metrical as well as musicological grounds. But all changes
to the text made by Ferreira which are based on his interpretation of the music, and which could not be justified here (or
in any other text) on purely metrical grounds, have been noted just beneath the apparatus. They have by no means been
rejected, but it is beyond my competence to accept them.
513
MARTIN CODAX 2
Mandad ei comigo
ca ven meu amigo,
e irei, madr, a Vigo
Comig ei mandado
ca ven meu amado,
e irei, madr, a Vigo
10
15
514
MARTIN CODAX 3
10
515
MARTIN CODAX 4
10
15
516
MARTIN CODAX 5
10
The phrase nas ondas in the last verse of strophe I is missing (though its music remains) along with much of strophes II and
III. What remains of the text in this part of the parchment looks like this:
] amado.
] mar leuado.
]emos. n. o.
]g alo mar de ugo.
e [.....]molo meu amigo.
517
MARTIN CODAX 6
10
15
Cf. note on N 3 / B 1280 / V 886, vv. 7-12. If the order in N is correct in both cases, two of the six poems of Martin Codax
(excluding the last, which has only two strophes) have an anomalous structure. Cf. B 644 / V 245 (Torneol), where a similar
structure is transmitted by B and V. In his prefcio to Ferreira 1986, Celso Cunha seems to accept the inversion (p. xiii).
518
MARTIN CODAX 7
519
AIRAS PAEZ 1
4 a mui V : mays a B
Short lines in BV, but v. 2 is written on two lines with a break after grado, suggesting that the body of the
strophe may have been composed in long verses.
1
If genuine, this would be an example of a hypermetric incipit (see Introduction, 5). If wrong, the phrase could have crept
in by accident from the refrain or could have been inserted for propagandistic reasons (cf. Correia 1993: 21-22) to specify
the place being publicized.
2
We find irmana in v. 5 and three times in B 1286-1287 / V 892 (the only other extant cantiga d amigo of Airas Paez), vv. 2, 5,
8. Michalis and Nunes keep the plural, and Nunes reads irma<s> in v. 5.
3
For a preposition repeated where redundant and unmetrical (and excised by Michalis and Bertolucci), cf. CA 53 (M.
Soares), v. 27 de me matar ou [de] me guarir.
4
It is difficult to account for the indirect object in lho. Possibly we have a misreading of ca o non vi (Nunes). But lho may be
popular usage (for other problematic examples of lho, see B 700 / V 301 [Ulhoa], v. 10; B 1118 / V 709 [Berdia], v. 9).
520
AIRAS PAEZ 2
10
The cantiga that immediately follows this one in BV is ascribed in V to Fern Do Lago (written by Colocci), but in B the
attribution (written by scribe a [cf. Ferrari 1979]) has been misplaced.
521
FERNAN DO LAGO 11
10
Toriello (pp. 30-35) would give this poem to Fernand Esquio, but cf. Oliveira 1988: 728-729.
It is by no means certain that V reads d here; the second letter may be a badly formed e with an upstroke.
3
Presumably we would find m n parti (cf. B 791 / V 375 [Talaveira], v. 3 pois m n parti), which would not scan here), not mende
parti, as Nunes suggests.
4
The reading of V is not certain. The first letter looks rather like a small capital d, but V normally uses capitals only at the
beginning of strophes and fiindas. (Compare the small capital d on f. 133v, line 6.)
2
522
JOHAN DE REQUEIXO 1
10
523
JOHAN DE REQUEIXO 2
10
10 ueher V : ueher B
vv. 1-2 (in this text) are long lines in BV, the rest short.
1
The consecutio temporum requires the perfect and the confusion between y and u is common. Cf. B 693 / V 294 (Estevan
Reimondo), v.16, where we find the error in reverse.
524
JOHAN DE REQUEIXO 3
10
15
10 o Braga : e V : E B
pardon B
525
JOHAN DE REQUEIXO 4
10
15
Que el log a mi non venha non tenh eu per ren que seja
nen que muito viver possa en logar u me non veja
e por en tenh eu que venha,
<como quer que outren tenha,
non tem eu del que non venha>
20
13 eu
526
JOHAN DE REQUEIXO 5
10
6 seu] seo V
9 U Nunes : E BV
527
FERNAND ESQUIO 1*
10
Grandes
In B, following v. 12, is a centered + and the single word Grandes stands alone on the next line; below it is a second + in the
left hand margin, followed by five blank lines. Toriello thinks Grandes is the beginning of a new strophe or fiinda (that is,
judging from its form, the initial G is not the beginning of a new poem). No sign in V that anything has been left out. (In B
1297 / V 901, v. 7 we find the phrase grandes coytas.)
528
FERNAND ESQUIO *2
10
15
20
11 da
<ar>
Cf. B 1172 / V 778 (Bolseiro), vv. 13-14 se vos eu ar vir / por mi coitado como vos vi ja; B 507 / V 90 (Dinis), v. 13 E non sei quando vos
ar veerei. This assumes that du (V) is an error for hu. But if Da (B) is an error for V a (for the confusion between capital D and V
see the manuscript readings for example at B 1231 / V 836 [Baveca], v. 19 where we find Vos V : Dos B) we might read u a que
el vira ir vir tornar. That the future subjunctive of veer (videre) is vir not veer does not seem to have bothered Toriello, who
translates the MSS veer as though it were vir [quando vedr, p. 99]).
529
FERNAND ESQUIO 3
10
15
530
FERNAND ESQUIO 4
10
15
vv. 5, 10, 15: O sentido parece ser: foi esse amor que
levei e esse amor que trago.
The emendation of Michalis makes a perfect tense out of a present. But the only form registered in Galego-Portuguese
lyric for the first person singular (assuming it is first and not third) of the perfect indicative is aduxe (see CSM, s. v. aduzer).
531
B 47 f. 15v
4 uos B
532
5 nezes B
10
15
B 332 f. 77r
1 nus nos Nobiling (1907b: 380-382) : nos nos B : nos nus Michalis (I, 821-822=CA 416) 3 amiga, e] Aamigaiee
B : corr. Michalis 6 nus Nobiling : nos B : ns Michalis
vosco] nosco B : corr. Michalis 7 e el] Et el B : et el
Michalis : ele Lapa, fort. recte 9 sen B : corr. Nobiling 10 Nnuca B 11-12 transpos. Michalis1 11 Grassa partar
soor mais poor B : de ora sapartar soo Michalis (deleto mais poor) : con saver a partir soo Nobiling : por s aver
a partir soo Lapa2 12 doo Michalis : door B 13 Mais poor prguntav B : del. Nunes (iam delere voluit Michalis
in ap.) pero o Nobiling : pero er Michalis 15 so Michalis : soorm B certa scripsi : terta B : certa Michalis
miga Michalis : Mi gam B
Short lines in BV. See Nobiling 1907b (critique of Michalis) as well as Lapa 1982:194-195 (critique of Nunes).
The strophic form appears to be aaabb. But verse 12 is double the length of the corresponding verse in
strophes I and II and it is difficult to see how it can be halved, despite the efforts of previous editors. If sound,
the verse would constitute a unique violation in the cantigas damigo of the principle of external responsion
(see Introduction, 5).
1
In order to maintain the dobre, whereby in this text in each strophe the same word appears at midverse in the first verse
(seventh and eighth syllables) and at the end of the fifth verse: I amiga, II chorando, II chorava.
2
In Grassa partar soo ought we read apartar (BV, Michalis) or partir (Nobiling)? The occurence of partir in vv. 1 and 6 strongly
supports partir. Perhaps triste de se partir soo is not impossible.
533
PERO MAFALDO 1
10
15
B 373 f. 85r
8 lhe B1 9 <e> Michalis (I, 850=CA 434)2 10 eu secl. Michalis, fort. recte3
: bona B 16 mais epoys B : e del. Michalis
1
13 ba Michalis
Michalis (followed by Nunes) maintains lhe here, but B has lhi in vv. 3, 6, 14.
There is no metrical reason to emend (Michalis <e> requires elision of the preceding vowel) nor any syntactic one; and ca
me seria mesura <e> bon prez (though I accept it as likely) just may miss the sense. Taking mesura as subject and bon prez as
predicate, we could render the manuscript version (ca me seria mesura bon prez) of vv. 7-9 as follows: You asked me eagerly to do
him good some time, saying compassion would be honorable. To show mesura would be bon prez according to the logic of the
boy (whose rogo the girlfriend is always [v. 2] delivering to the girl, asking que lhi faa ben). Those who would maintain both the
reading of the MSS and the sense Michalis gives the passage could merely add a comma: ca me seria mesura, bon prez.
3
No need to touch eu on metrical grounds, since the final vowel of rogo could be elided. Still, given the presence of eu in the
refrain, Michalis may be right to cut it here (where it could be an interpolation by someone for whom the hiatus rogo | e was
difficult to accept).
2
534
PERO MAFALDO 2
10
15
B 383 f. 86r
1 fazia Lang (1894: 134), Michalis (I, 864=CA 444)1 2 preit e menage Michalis : prdem enaige B 3 ora] oria B
carreira sa via Michalis : carrerya lauia B 4 possis mente, non 5 s a] s<e> a Michalis
cataionha B
6 amiga(l) B
qnan ? B 7 jurou] uirou B
1
Based on an incorrect reading in Molteni. There is no reason here not to maintain fezia, documented in the glossaries of
Mettmann and Lapa. The forms of fazer in this and the preceeding text are: faa, farei, fezesse, fezia, fez.
535
ALFONSO X1 1
B 4562
f. 101r
vv. 4, 8: Probably better to read Guarda (place name) than guarda (simple substantive).
1
On the attribution, see Pellegrini 1959: 78-93; Gonalves 1976a: 399; Oliveira 1994: 313.
This text, preserved only in B, is presented in two four line strophes which, with minor corrections (and regularized
spelling), would read as follows:
Ai eu coitada como vivo
en gran cuidado por meu amigo
que ei alongado muito me tarda
o meu amigo na guarda
2
536
10
15
This pronoun is the only reference in this text to the girls audience, presumably another girl or girls, and although BV
have u9 that u could as easily be wrong as right. It could be argued that we ought to read nos because with that pronoun the
girl would show solidarity with her addressees. But note that in another programmatic cantiga damigo (and there are few
such), B 766 / V 370 (Guilhade), a similar problem comes up in vv. 5, 9, 23, with Nobiling preferring <vs> / vos (vos with the
manuscripts in vv. 9 and 23, <de vs> his own conjectural supplement in v. 4) and Nunes <ns> / nos. There the verb veeredes
in v. 21 seems to resolve the question (pace Nunes). In our text, the best reason for reading vos may be that the pronoun
functions here as a particle (much like ancient Greek toi [enclitic: a specialized use of the dative singular of the 2nd person
pronoun, functioning as a particle, roughly equivalent to English you know]; cf. LSJ, s. v. toi).
2
For the indirect object cf. B 1149 bis / V 752 (Zorro), vv. 9-10 Non vos ten prol, filha, de mho negar; / ante volo terr de mho dizer. For
the word order, cf. B 639 / V 240 (Pae Soarez), vv. 1-2 Donas, veeredes a prol que lhi ten / de lhi saberen ca mi quer gran ben and B 1640
/ V 1174 (Ponte), v. 12 Ou grand estana, que prol lh . But the order du <lhi> gran prol ten can also be defended (B 1355 / V 963
[Lopo Lis], v. 25 pois que lhi prol non ten and B 516 / V 99 [Dinis], v. 9 o que mi pequena prol tem).
537
10
15
20
V 5941 f. 94v
2 sen] seu V : corr. Monaci
1
4 end ? V
7 E seclusi
14 <eu> Nunes2
Two pages of B (apparently torn out) leave V as the only source for this and the following six texts (but cf. V 596).
Cf. B 1026 / V 616 (Johan Airas), v. 13 mui ben sei eu. The formula ben sei eu occurs among the cantigas d amigo of Johan Airas
in B 1031 / V 621, v. 19; B 1041 / V 631, v. 19; B 1043 / V 633, v. 2; and B 1047 / V 637, v. 13.
2
538
10
15
20
539
10
15
20
1 grambe quer V 5 fal[e B : tale V 8 prol hi V : possis prol <l>hi vel prol <l>h i 11, 17 tale V 13 falhar V
14 cmig V 16 moira ssi distinxi : moyrassy V : moyrassy Braga : moira <a>ssi Rodrguez 20 sen] seu V :
corr. Monaci
1
Though two pages are missing from B (apparently torn out), traces of text remain in the upper left hand corner of the folio
on which this poem was copied. These belong to vv. 13-18 and read as follows:
Seu9 p
Cmi
Ed
E
fal
What are presumably the left halves of the initial Qs of vv. 19-20 are also visible, as well as Coloccis mark in the left hand
margin indicating the fiinda.
540
10
15
20
V1 597 f. 95r
2 fara<n> Monaci : fara V 5 mi<n> Michalis (II, 663-664) : mi V 6 <o> Michalis
Michalis, Lapa : come BV, Nunes2 14 <bon> Michalis 19 uir BV : viran Michalis
In B, though the page has been torn out, the number 1007 remains in what is left, along with the left half of the initial O
and a trace of the initial D in v. 2.
2
Michalis and Lapa (1982: 178) agree with Braga. Nunes prints come but (vol. III, p. 253) is willing to read com . Rodrguez,
who defends com on (dubious) metrical grounds, finds come acceptable. Michalis (CA) and Mettmann (CSM) give examples
of como as a causal conjunction, but Rodrguez offers none in his glossary of Johan Airas. Nor is the word registered in this
sense in Nunes glossary of the cantigas d amigo, but see B 714 / V 315 (Queimado), vv. 9-12 com eu serei / sanhuda, parecendo ben,
/ muito terr que baratou / mal... ; B 933 / V 521 (Cana), vv. 7-8 Como eu tevera guisado / de fazer quant el quisesse... Most telling is
B 1173 / V 779 (Bolseiro), v. 4 mais, como x mui trobador.
541
10
15
20
V 598 f. 95r-v
1 Amigo(me) V
542
2 <a> Braga
18 on V
10
15
V 599 f. 95v
1 u9 V 2 veja <e>sse Nunes : ueia sse V 9 grisa de zedo V 13 vos el<e> seja Lapa : uos el seia V : <a>
vs el seja Nunes : fort. vos <or> el seja1 14 uo l(h)o V 15 uos V 16 vs como vos] u9 como uos V
17-18 om. V2
vv. 2, 7, 14: fustan (< M. Lat. [pannus] fustaneus; cf. Corominas, s.v.), a kind of coarse cloth made of cotton and
flax (OED, s.v. fustian), refers here to a garment (Rodrguez, ad loc. [p. 162] suggests a skirt). In Medieval
Latin, we already find the meaning blouse ou chemise en futaine shirt made of fustian, acc. to Niermeyer
(s.v. fustaneum 2).
1
If correct, the adverb ora would stress the contrast between his anger now and how he will react when he sees the girl in her
fustan. Among the cantigas d amigo of Johan Airas, ora a scans as two syllables in B 1037 / V 627, v. 19.
2
Although there is no sign of the refrain in V, a poem with two strophes of six verses and a four verse fiinda (and with the
same rhyme scheme as the body of the strophe) seems highly improbable. Cf. B 566 / V 169 (Dinis), v. 18, and B 1231 / V 836
(Baveca), v. 22 and note.
543
10
15
20
V 600 f. 95v
2 amiga Rodrguez : amigas V1 4 nos] n9 ? V (u9 Monaci) 5 el] ci V 5, 17 ousa mentar distinxi : ousame~tar V
: ouse~mentar Lapa, fort. recte : ousa mentar Braga : ousa mentar Rodrguez2 8 el<e> Rodrguez <o> meu Nunes
10 preito : pruo V cf. v. 19
13 <j> Nunes
lh<o> eu Rodrguez : lheu V 14 ca scripsi : q V
ouv i distinx.
Rodrguez : ouui V : uvi Nunes 16 pugnera V ende <e>stamos Nunes : e~ destam9 V 20 quen o] qio V : i
delevi : quen i o Rodrguez3
vv. 19-20 o preito guisad en se chegar / era: Either these
negotiations are ready to reach a final agreement or
this relationship is ready to reach a climax.
1
vv. 19-20 o preito guisad en se chegar / era: Estas negociaes esto prontas para chegar a um acordo final
ou esta relao est prestes a atingir um clmax.
Cf. v. 6. Nunes prints amigas in v. 2 and amiga thereafter in the refrain. Rodrguez opts for the singular, persuaded by the
carcter ntimo of the girls revelations. Some data may be adduced to support his hunch: of the 45 cantigas damigo of
Johan Airas only four are addressed to amigas (V 597, B 1024 / V 614, B 1040 / V 630, B 1044 and 1048 / V 634 and 638) and in
none of these does the girl express her erotic aims as explicitly as here.
2
The form mentar printed by Braga, Nunes and Rodrguez, is not elsewhere documented in Galego-Portuguese lyric (Rodrguez,
based on this passage, has registered mentar in his glossary [cf. also REW 5505]).
3
A vexed passage. Rodrguez prints non quen i o comear, but recognizes la dureza de la sinalefa between i and o.
544
10
15
20
545
10
15
20
Lapa (1982: 179) is right to suspect a corruption. As an alternative to the reading printed here, he suggests e el<e> de vs,
but the parallel in v. 10 seems to me decisive. Rodrguez, clinging to the MSS, offers e d el e de vs non sei por que , claiming
that de indicates origen o procedencia. Nunes prints e el de vs, non sei porque .
2
Whether V has n9 or u9 (Monaci), the error is trivial (cf. B 571 / V 175 [Dinis], vv. 5, 11, 17) and esto non conven corresponds to
tenho que mal (v. 10). Rodrguez recognizes that non makes better sense, but follows (with Nunes) what he takes to be the
reading of the MSS (marred here, as he allows, by desconexin sintctica and ruptura narrativa).
3
In B 1047 / V 637, vv. 5, 11, 17 (Johan Airas) both MSS read malhi.
546
10
15
8 atendera ? V : attendera B
10 a seclusi1
fezer] fez V
Since neither mi (tonic) nor se is subject to elision, v. 10 scans long. The presumably interpolated a before mi may be the
culprit, but one might also (more boldly) transpose: mais log el verr / rogar a mi.
547
10
15
20
548
10
15
20
Cf. B 1047 / V 657 (Johan Airas), v. 19 <E> enton. Rodrguez accepts Este ben por mi <v-l>o fazedes? from Nunes.
549
10
15
20
550
19 per mj B
10
7 u9 V : uos B
8 queirey B : qirey V
eu]
551
10
15
Nunes prints <e> eu fui <mui> triste sempre chorei and is followed by Rodrguez (Machado prints trist e).
552
10
15
20
te~ B
After sse Colocci has written two letters which appear to be ra (anticipating sera in the second verse of the refrain?).
cedo was written by the scribe at the end of the first verse of the refrain (corrupt here in both MSS), then apparently struck
out and rewritten by Colocci at the beginning of the next line.
2
553
10
15
554
13 So V : Sc B
10
15
555
10
15
20
556
10
15
557
10
15
20
558
10
Nunes emendation diredes might seem attractive, but would weaken tone without improving sense. If I ever speak with
you again, you will say... would still suggest that the girl foresees no future conversations; but If I ever speak with you
again, go right ahead and say... sharpens the point. There is a close parallel in Don Dinis B 561 / V 164 where the girl mocks
the boy for his alleged lies about their relationship, bidding him boast all he wants to (vv. 4-6, 10-12, 16-18, 20). Cf. also
B 745 / V 347 (Guilhade).
559
10
15
20
2 fij n] fige~ BV
9 fazer] farer B
13 nos V : uos B
15 fiz V
vv. 15-16: de perder is to be construed with pavor (and not with ei, pace Rodrguez [pp. 211-212]). contra is used
here in a neutral sense (in relation to; cf. CSM 5.86 tenno que seja contra nos leal), already well documented
in Latin.
560
10
15
20
Pellegrini emends vv. 15 and 17, in part to alter (alleged) evidence that the girl is married. But while v. 15 scans, v. 17
would be long. Pellegrini (1959: 132, n. 10) would elide sa but se is tonic. Hence, for meter and for syntax, we cut a and read
se verdade. Cf. B 1037 / V 627, v. 10. More boldly, we might transpose: ca, se poderen a verdad osmar (which would salvage the
article without junking the meter).
2
Rodrguez dismisses Nunes proposal, but cf. B 823 / V 408 (Pardal), v. 14 que non venhades u eu for and B 1170 / V 776
(Bolseiro), v. 21 non venhades nunca u eu for. The boy will be able to see where the girl is (v. 24), but not to get there.
561
10
15
562
vem Braga : n BV
11 mi~ B : min V
13 pes B
10
15
20
This proposal (accepted by Rodrguez) should be rejected. The reading of both MSS poys eu mal quiser (maintained by
Braga, Machado, and Fernndez Pousa) is unassailable on any grounds (metrical, syntactic, semantic, sytlistic). Temporal
pois with the future subjunctive is sometimes used as a causal conjunction. In B 686 / V 288 (J. S. Coelho), v. 10, poi<lo> non
souber simply means since I dont know. Here we can render pois as once or since. Cf. CA, vol. II, p. 454, n. 4 (continued
from p. 453).
2
disser is my correction of quiser (quis BV), which would ruin the dobre by the repetition in rhyme of a word not involved in it.
It is easy to understand how quiser could have crept in here, since in all three strophes the dobre makes use of forms of this
563
verb (quer ben [vv. 2 and 5]; ben quer [vv. 8 and 11]; querer [vv. 14 and 17]) and the form quiser appears at the end of v. 7, but our
poet handles the technique perfectly in seven other cantigas damigo (Johan Airas-0, 2, 21, 22, 28, 34, 35) and the lapse here
is unlikely to be his. In favor of the correction, which produces the phrase faredes ... quant eu disser, see B 672 / V 274 (Avoin),
v. 22 ca fa eu quanto dizedes and B 793 / V 377 (Talaveira), vv. 2-3 que jurava que sempre fezesse / todo por mi quanto lheu dissesse.
Further, disser would be integrated in the sequence Dizedes (v. 7) // disser (v. 9) / digades (v. 10). Finally, it would make of dizer in
v. 19 yet another example of the repetition of a palavra rimante in the fiinda (Lang 1934) a technique used by Johan Airas in
24 cantigas damigo, including the next four but an example with a twist, since what appears in the fiinda is not the same
form but a rima derivata (disser [v. 9] > dizer [v. 19]). We find a similar display of combined techniques in Sevilha-3, where, in
addition to the dobre, a word which ends a verse of the refrain is repeated with rima derivata in the fiinda (visse > veedes). For our
poets use (in other cantigas damigo) of rima derivata in fiindas, see Johan Airas-30 (querria < quisera [I.2]; verria < venha [II.4]), 38
(podr < poder [II.3]) and 42 (desamor < amor [I.1]). In Johan Airas-3 we find the rhyme word dizer [refrain] repeated in the same
form in the fiinda.
3
Note the variety of constructions in this text: que eu quera mal a quen (v. 2); pois eu mal quiser a quen (v. 5); daver quen mi quer ben
mal a querer (v. 14); pois quen mi quer ben ei mal a querer (v. 17). In the former two we find querer a quen, in the latter two querer +
quen. Cf. A 165 / B 317 bis (J. S. Coelho), vv. 22-23 E tenho que fao dereit e sen / en querer mal quen vus quer mal e ben.
4
In B, v. 11 is copied on two lines (with a break after mal); on the second of these, the scribe copied the beginning of v. 12.
Colocci rewrote it on the next line (and presumably struck it on the one above). In V, where v. 11 is copied on one line, the
single letter e stands on the next.
5
Nunes (followed by Rodrguez) reads pregunta tal; Fernndez Pousa, pregunt atal. Cf. the glossaries of Lapa and Mettmann
(s.v. atal).
6
Sense in itself suggests a corruption (fazer mal quen is ungrammatical) and in accordance with the rules of the dobre this
verse, like v. 14, ought to end with querer. (The error may be due to the use of fazer in v. 15.)
564
10
15
20
The verse seems to be missing a syllable. Nunes supplies it with <mia> senhor; Rodrguez with el<e>. My suggestion is
to suppose the loss of an e, but as an initial conjunction (within the words ascribed to the boy) with exclamatory force.
Cf. V 599 (Johan Airas), v. 16 e guisade.
2
Nunes emendation toda seems obvious. Cf. B 630 / V 231 (Calheiros), v. 10 ca sobr el deit eu toda mha sanha. Rodrguez clings
to todo, claiming it functions as an adverb here; to judge by his glossary, this would be the only such use in the 81 cantigas
of Johan Airas. Those preferring Rodrguez text could cite B 832 / V 418 (Ponte), v. 13 tod andava con mentira. But if we keep
todo, we should probably then accept tod9 (B), taking the s as an error for x and reading todo x .
565
10
15
20
The restoration of the fiindas seems fairly clear except on one point: in v. 21 where B (our only source for the line) reads con
sen (maintained by Nunes and Fernndez Pousa), Rodrguez prints a tempting do seu (citing vv. 8 and 11) which I would
nevertheless resist. The speaker has referred to the addressees common sense in vv. 7-9 (you wouldnt be so senseless as
to think you can take the gifts and then turn him down) and has asked for her opinion (vosso sen) in v. 16. She now says:
Either you take the gifts and give him what he wants or you dont, being sensible, take anything and then you owe him
nothing. For the expression con sen, cf. B 1233 / V 838 (Baveca), v. 20 non sab andar en tal preito con sen and B 1377 / V 985
(=CEM 379 [Burgals]), v. 11 con mal sen. Compare con sen cristo in B 1649 / V 1183 (=CEM 365 [Ponte]), v. 9.
566
10
15
20
7 qu<is>er scripsi2 : quer BV, Braga, Nunes, Rodrguez : quer<e> Lapa <o> seu
9 mi~ B 10 dorma V 15 perdud<o> Nunes : perdude BV 16 dormital B
Nunes emendation (adopted by Fernndez Pousa and Rodrguez), overlooks the perfectly viable syntax of the transmitted
text: My friend, driven by love, since he now wants to live with me for a while if he can, had better not sleep while hes with
me... The same syntax and the same words (meu amigo [v. 1] ... non dormha [v. 4]) are repeated in vv. 9-10. (Nunes and
Rodrguez must put a semi-colon after sazon [v. 3], breaking the syntactic and rhythmic flow of the strophe.)
2
Cf. B 6 (Airas Moniz d Asme), vv. 17-19 E quen ben quiser trastornar / per todo o mundo, e ferir / mui festinho xi a pod achar; B 229
(Fernan Garcia Esgaravunha), vv. 15-16 E quen ben quiser preguntar / por gran coita, min pregunt n; and B [366 bis] (Cogominho),
vv. 7-9 Quen ben quiser meu coraon / saber, por que ensandeci, / pregunte me. Each of the parallels cited occurs, like the proposed
correction, at the beginning of a strophe in a generalizing relative clause of characteristic (whoever would like..). Cf. also
B 789 / V 373 (Talaveira), vv. 19-20.
567
10
15
20
A toss-up. In favor of lho eu vou we may note that lho eu is found at B 1044 / B 637, vv. 5 and 11.
568
10
15
20
25
10 con
27 mi]
569
10
15
Dizedor de nemiga
e dir o log, amiga
20
6 Null B
10 sen BV
11 el amiga V
Against this emendation (accepted by Rodrguez), it could be pointed out that paz does not occur elsewhere in the work
of Johan Airas, nor in the entire corpus of cantigas d amigo. But that in itself is hardly a sufficient argument for rejecting it,
and the expression aver paz is documented in CSM (cf. Mettmanns glossary). The far more frequent aver par yields little
sense here: aver par nen mesura would make an improbably harsh zeugma.
570
10
15
20
Cf. B 1028 / V 618, vv. 15-16. The change here seems authorized by cuid (BV) and meter is not affected. Pellegrini made
both suggestions to refute allegations that the girl in the poem might be married, but the paleographical, metrical and
syntactic justifications for the change are cogent without passing judgment on the civil status of the girl.
571
10
15
20
The verbal play I suggest is present here may be compared with V 1016 (=CEM 233 [J. S. Coelho]), vv. 1-3: Maria do Grave,
grav de saber / por que vos chaman Maria do Grave, / ca vs non sodes grave de foder / e pero sodes de foder mui grave...
2
This emendation, accepted by Rodrguez, is probably right. While the reading of the codices might appear to involve a
simple pun (the boy wants ben, but the girl doesnt think ben can save him from death), there seems rather to be an
uncharacteristic logical contradiction (cf. v. 9), arguably the result of a banal scribal error.
572
10
15
20
vv. 15-20: que mh assaca tal mentira = que diz uma mentira assim contra mim; quen me mal quer assacar =
quem quer dizer mal de mim.
573
vv. 13-15, 20: Among the cantigas damigo the verb assacar occurs only in this text, though twice. Elsewhere in
Galego-Portuguese lyric we find it in V 1199 (=CEM 317 [Sevilha]), v. 7, where it is probably intransitive, and
in CSM 401.69 ome que assaca, que peor que can, where it is clearly so. The editors gloss the word here variously,
to cover both the transitive use in v. 15 and the intransitive use in v. 20: Nunes: imputar, atribuir; Machado:
imputar, incriminar de; Rodrguez: achacar, imputar falsamente, inventar; Mettmann: accusar,
incriminar, imputar. Lapa (s.v. assacar) glosses CEM 317, v. 7 queres-te-lhi tu mal assacar as queres acus-lo
injustamente. If, rather than being a Galego-Portuguese formation, assacar derives from Gothic *at-sakan,
similar in meaning to the documented and-sakan speak against, the etymological sense would fit well in
both verses. The notion of accusation is already implicit in the Gothic verb andsakan and explicit in several
of its related forms derived from the same root: gasakan scold, gesahts reproach *sakjo quarrel, sakuls
quarrelsome (see Lehmann, s. v. sakan).
574
10
15
6 mi~ BV
Whereas the verses in the body of the strophe scan 8 syllables each, those in the refrain, as printed here, scan
8 9 7.
575
10
15
20
The only difference here is bendas doas B : mais doas V, where neither MSS seems right. But if we emend either for sense
alone we still get a hypermetric verse, so from Braga e partide-mas doas que vos el dar (based on V) to Nunes e partide ben das
que vos dar down to Rodrguez, editors have tended to shrink the verse, though sometimes without reaching ten syllables.
If we read with V, mh as for mais is an easy correction. If with B, ben could be kept in view of vv. 9 and 15. Cutting ben (B) and
el (BV) evens the scansion without altering the meaning of the phrase.
576
10
15
20
577
10
15
20
Nunes prints t (adopted by Rodrguez) but then decides (vol. II, p. 470) that at should stand. Finally (vol. III, p. 720) he
proposes a simple transposition: at que vos Deus m adusser.
578
10
15
20
579
<d> al addidi : al BV
6 on two lines in BV
e del. Nunes
dizede2] dizete V
The line break in v. 6 comes after mente. The rest of the text is copied in both MSS in long lines. Nunes (but see
vol. III, p. 282) and Rodrguez print this poem in short (7-syllable) verses (cf. Tavani, Rep. 219:1), which yields
a strophic form unique in the entire corpus. The form presented here, recommended by Lapa (1982: 181),
corresponds to Tavanis scheme no. 19 (aaabbb).
580
10
15
20
E meu amor (written by the scribe) stands alone on the next line, and Colocci apparently struck it out and wrote it above.
In V the same phrase also stands alone on a separate line.
2
Nunes offers: ca maior amor no mundo <non> .
581
10
15
20
582
10
15
8 Coita B : corta V
sofren B 11 a] o V 12 eu uolos
en] eu B 15 mi~ B 16 direyu9 B : direm9 V
The abbreviation, which has been displaced to the right, represents ar. In its genuine form, it consists of a horizontal
superscript curve (facing downwards) with two short thin parallel vertical strokes descending from it. We rarely find it in BV
(at least among the cantigas d amigo), but the scribe of B here (whom Ferrari 1979 calls e) has reproduced it rather exactly
in v. 18. Vs representation of it there is better placed and rather more elaborate than here, but still much like an ordinary
til. In v. 6 both MSS have cobrarey.
583
10
be~ V : de bem B
Colocci marked only the sixth (and not the fifth) verse of both strophes, thereby signalling that the refrain either in the
exemplar or in his opinion consists here of a single verse. According to this logic, v. 11 should not be identical to v. 5. This
clue is supported by the trip-up in verbal virtuosity in the transmitted text. What we need is a kind of mini mordobre of the
form x: x:: y: y but the last term is wrong. Hence, we need, but do not (fully) get: (I) pois vos queredes quitar / daqui, ... / quitade;
(II) pois vos queredes daqui / partir, ... / partide. This would explain both Coloccis marking and the otherwise irregular copying of
v. 12 (the second v. of a [non-intercalated] two verse refrain is only copied in full after the first strophe on a single occasion
among the cantigas d amigo of Johan Airas [B 1048 / V 638, last strophe]). The mixup may be due to the expectation of an
identically repeated two verse refrain and to the tricky wordplay. Moreover, partide ben o coraon here (at the end of the set?)
would echo the phrase non se pod o coraon partir in B 963 / V 550, v. 5 (arguably the beginning). Finally, notwithstanding the
variation (quitade / partide) and the evidence of Coloccis markings paradoxically a key to the emendation we should see
a two verse refrain here (otherwise the strophic form would be without parallel). There is also variation in the (one verse)
refrain of B 1050 / V 640 (v. 18 ai meu amigo).
584
10
15
20
25
The line division in V (our only source after B breaks off with Apertandosse in v. 5) often does not follow the rhyme scheme
and breaks up words (alzan / do [v. 5], as cuytar / des [v. 19]).
2
que resists elision (though nowhere else in the cantigas d amigo of Johan Airas does que occur before an initial i) and mais
could be cut.
585
DINIS 1
10
15
20
9 erates V
coytado B : cuytado V
18 e~coberto B : entoberto V
586
DINIS 2
10
15
5 Que B : O aie V
moireu B : moneu V
7 fort. muit al
9 s i Lang
praxer B
587
DINIS 3
10
15
20
588
8, 16 dereit] deyt BV
10 boa B
DINIS 4
10
15
589
DINIS 5
10
15
20
After the first verse a line has been left blank. There follows:
E. . . Mandado do meu amigo
Mandado|do meu amigo
590
DINIS 6
10
15
20
6 possedes B 7 Dixz B
19 preda V, Lang
8 ptirades B : pertirades
vv. 5, 11, 17: posse (substantive) is not found elsewhere in Galego-Portuguese lyric.
591
DINIS 7
10
15
20
592
DINIS 8
10
fort. ca mi ven
vv. 3-6: Ou podes crer... que nesse dia verei que nunca vi maior prazer ou ... que hei-de ver maior prazer
do que alguma vez vi.
The subjunctive possa, if correct, could be explained by the introductory verb creer (Huber, par. 463), but Lang seems to
think we have an elided indicative poss (presumably for a future). In that case the reading might once have been poss aaquel,
subsequently divided as possa aquel.
593
DINIS 9
10
15
20
5 qiserdes B : uiserdes V
mi~ BV
v. 8: o mentiral (only here in Amigo), meaning the liar, is the subject of fazer, while the whole infinitive phrase
is the object of the preposition por.
v. 9: cf. e esto a elles n montava muyto in Crnica Geral de Espanha de 1344, vol. IV, p. 306 (ed. Cintra).
v. 15 ao que: cf. Nunes, vol. III, p. 19 (the missing antecedent of o que must be understood as the subject of s
enfinger).
1
Cf. Ramos 1988: 621-637 (esp. 626-627, 634, 637). This would be the only occurrence of senher in the cantigas d amigo and
moreover in a text in which senhor occurs twice (vv. 4 and 16). The emendation of Michalis (accepted by Nunes and by
Cunha 1982: 50-51) was motivated by the lack of rhyme between der (v. 7) and senhor (v. 10). In vv. 7-8 Lang prints Ca demo lev
essa rem que eu der por / <tal> enfinta fazer o<u> mentir al providing a rhyme, but ignoring the line break, adding tal, and
counting que eu as a single syllable.
594
DINIS 10
10
15
14 re~ de q B : re~ cu
595
DINIS 11
10
15
20
Lang and Nunes follow Moura. The confusion between eu and en is common and the change, though not absolutely
necessary, seems correct.
596
DINIS 12
10
15
20
5 poer BV
8 sei eu] se er V
14 Delhi
597
DINIS 13
10
15
20
598
DINIS 14
E oj st o prazo saido;
por que mentiu o desmentido?
ai madre, moiro d amor
E oj st o prazo passado;
por que mentiu o perjurado?
ai madre, moiro d amor
Por que mentiu o desmentido,
pesa mi, pois per si falido;
ai madre, moiro d amor
10
15
11 pr V
599
DINIS 15
10
15
4 De] Do BV
8 ueiesta B : ueesta V
v. 3: Aparentemente significa: Est a amanhecer, vai
depressa!
liero is explained by Diez (1863: 99; cf. Lang, p. 134) as a variant of ligeiro with the loss of (secundary) intervocalic g. The
meaning of the verse is much disputed. It is not certain how this should be punctuated or understood, nor to whom it
should be assigned (to the girl or to some other voice). Lapa (1982: 156) may be right to understand Alva , vai liero! The 2nd
person singular imperative, if correct, would mean that we are dealing with a citation.
600
DINIS 16
10
15
20
601
DINIS 17
Levantou s a velida,
levantou s <aa> alva,
e vai lavar camisas
eno alto,
vai las lavar <a> alva
Levantou s a louana,
levantou s <aa> alva,
e vai lavar delgadas
eno alto,
vai las lavar <a alva>
10
15
20
25
30
602
By correcting problems of syntax (and sense) we come up with a strophe scanning 6 6 6 3 6 (not 6 4 6 3 5 as
in Tav., Rep. [119:1]).
vv. 2, 4, etc.: <aa> alva means in the dawn; <a> alva
is the white (lustrous, virgin) one (the girl).
1
Lorenzo 1977, s.v. alua, cites an example of aa alua do dia. Cf. B 365 (=CA 424 [Cogominho]), v. 9 aa noite.
603
DINIS 18
10
15
20
After Moura (Amigo, meu amigo) and Braga (Amigue, meu amigo) nearly all editors follow Lang. Cf. Cohen 1987: 112, n. 31.
604
DINIS 19
10
15
2 t B : r V 4 veja correxi (ex V 116, v. 4 ueia) : uera BV : ver Moura, Lang, Nunes 8 vs2] uos B : nos V
9 mignado B 10 pode Lang : Pode~ B : poden V mui bem del Lang (ex. V 115, v. 14 muy be~ del) 14 catr B
: cati~ V 16 ente~de~ B : entender V
NOTE ON B 570 bis / V 174: This poem has also been transmitted in both B and V among the cantigas d amor of
Don Dinis. That version, corrupt in both MSS, contains some notable variants (including the inversion of the
last two strophes). Rather than put them in the apparatus I offer transcriptions:
B 523 f. 117r
O uos amigo t de cora
+ el em uos seus olhos
E tam be~ par deus amig9
Que no
+
+
Que n podel poder auer dauer
Praxer de (mha) nulha re~
Se non de u9 Vee
E qudo el ue~ hu uos fodes raz
Qr el catar qui se encobra ete~
605
V 116 f. 13v
O uos q migo t de toratom
pom el em uos seus olhos e t be~
par de9 & miga q n
ueia q nom em te~da q no podel
poder auer dauer prazer
de nulha rem se n de u9 ueer
E qudo el ue~ hu uos sodes (n)es *
qr el catar q se encobra ere~
qse~ cobre po nlhe ual p
tanos fr9 olhos ente~de~ q n podel poder
E que~ ele~ uiuer tomo el se9 olhos p
e~ uos amiga qndo arae uos ue~
se n for c muy gm meng desem
entender pode~ muy be~ del q n
podel poder au.
606
DINIS 20
10
15
5 non Moura : n9 B : u9 V
ueo BV
11 non] u9 BV
ueo BV
16 nono B : nouo V
607
DINIS 21
10
15
608
DINIS 22
10
15
20
25
9 poco B
Lang prints nom <o> pdem veer, which is hypermetrical. Nunes (whom I follow here) takes the boy as subject and the eyes
as object (of veer). Were we to read (not very plausibly) with the MSS, the eyes would be the subject, and veer would be
intransitive.
609
DINIS 23
10
15
Nunes inverts the order of these two words to match the corresponding verse in the first two strophes. Variation in a
refrain is rare (and sometimes illusory) but genuine, and here, since the fifth verse of each strophe varies, though within a
very narrow range, variation in the last verse of the final strophe is at least plausible.
610
DINIS 24
10
15
20
Lang reads Nom o queirades, criticized by Michalis (1895a: 530). Nunes prints Non querr Deus. Non queira Deus is also possible.
611
DINIS 25
10
15
612
8 Possededes B
13 creerey B : creerer V
14 fe] se V
DINIS 26
10
15
20
This emendation, defended by Michalis (1895a) and adopted by Nunes, would yield an exact rhyme, but mandardes is
probably sound. The boy responds to the girls discourse, in which the rhymes are exact, with an inexact rhyme characteristic
of many traditional parallelistic cantigas d amigo. Moreover, mandardes, on the one hand lectio difficilior (why would a performer
change an exact rhyme to an inexact one, or a scribe insert the rarer form?), is better syntax too (because it follows guarrei
and because u me mandardes is an indefinite adverbial clause).
613
DINIS 27
10
15
20
7 Disseste B
9 iurastes V : iuraste B
15 vos
v. 20: Among the senses of passar given by Mettmann (s.v.) we find transgredir (CSM 392.38 passou a postura;
411.110 non passedes de Deus seu mandamento) and ultrapassar, exceder (137.8 luxuria, que passava razon e maneira).
1
perjuras (Lang) appears to be a phantom word; the construction is rather creer per juras. Cf. (in Dinis) creede per min in B 560
/ V 163, v. 7 and B 583 / V 186, v. 10.
614
DINIS 28
10
15
20
6 Viuen B : uyuer V
pdon V : pard B
11 mi~ BV
17 mi~ B
If the reading (maintained by Lang and Nunes) is correct, this would be a modal imperfect, equivalent to a conditional,
answering to a perfect subjunctive in a contrary to fact condition. But elsewhere, where a similar modal imperfect
would appear to be documented, we sometimes find a discrepancy in the MSS between queria and querria. Cf., in the
cantigas d amigo of Dinis, B 556 / V 159, vv. 2, 6, 13 and B 563 / V 159, vv. 5, 18. In both those texts I have chosen querria on the
strength of the manuscript evidence. But note, for instance, that in the latter text, at v.5, both MSS read qria (as here), while
at v. 18, where B offers the same form, V reads querria. See also B 588 / V 19, v. 12.
615
DINIS 29
10
15
20
25
616
nen is here a non-negative coordinating conjunction. Cf. Pellegrini 1959: 110 (note ad. vv. 22-23): Sebbene le grammatiche
solitamente ne tacciano, nel portoghese antico e nello spagnolo antico (come in provenzale [...]) quando una proposizione
negativa viene coordinata a una precedente, negativa o positiva, si ha, in vece di e, un nen.... See nen in the glossaries of
Lapa (valores equivalentes, mais ou menos a ou e copulativa e) and Mettmann.
617
DINIS 30
10
15
20
618
14 Quand B
DINIS 31
10
15
Lapa (1982: 156-157) evidently does not understand Langs reading, which adheres to the manuscripts, is defended in a note
(Em dafr bezieht sich hier auf ein aus den adjectiven fals e desleal zu ergnzendes substantiv falsidade oder deslealdade
[1894: 137]), and has been printed here. Thus, though Lapa is willing to accept that en may be an adverb (as Lang intended),
he makes several proposals to correct a passage he rightly considers badly handled by Nunes (de na with two prepositions).
In its only other appearances in the cantigas damigo we find trabalhar construed with de (B 564 / V 167 [Dinis], vv. 2-3 que se
trabalha de vosco buscar / mal a voss amigo and 13-15 Ela trabalha se, gran sazon, / de lhi fazer o vosso desamor / aver...). Hence Lapa
suggests que prol de vos trabalhar-/des d a mia mercee cobrar. This would presuppose a unique example of word-break between
lines in the cantigas damigo, though cf. e.g. Dinis-24, vv. 23-24 guardar- / vos (we have many examples elsewhere in GalegoPortuguese lyric, more in CSM than in CEM or Amor). In defending trabalhar- / des here, Lapa writes: A lio dos manuscriptos
impe necessariamente o uso do verbo no modo pessoal (1982: 156). Meanwhile, he is right that we also find the construction
with en as in B 151 (=CA 61 [M. Soares]), vv. 17-18 e por esto non mh mester / de trabalhar en vos fogir. Hence two other suggestions:
trabalhar- / des en and trabalhar / en-na (preposition plus article). Finally, Lapa proposes a slight transposition in v. 2, reading que
prol vos de trabalhar / en-na mha mercee cobrar. This would perhaps be the most elegant solution if there is really a problem.
619
DINIS 32
10
15
20
620
13 eu] ea V
perca V, Lang;
DINIS 33
10
15
20
17 em eu Moura : emeu BV
v. 17: Lang, noting that en is causal, translates: Weil ich mit dem leben muss, mit welchem ich lebe (p. 137).
621
DINIS 34
10
15
20
vv. 7, 15 riio: This form scans bisyllabic (cf. Lang, p. cxxi rir) in both verses (with hiatus [at a pause] in v. 6 ledo
| e riio and with elision in v. 15 led e riio). We find the same scansion twice in Lopo Lis: B 1345 / V 952 (=CEM
259), v. 8 de que riiu a bela; and B 1346 / V 953 (=CEM 260), v. 3 riiu-s el Rei e mia esposa de mi (where Pellegrini
[1969: 22] marks both the bisyllabic pronunciation riu and the subsequent elision [mha_esposa]).
622
DINIS 35
10
15
20
13 muyto V : muyta B
623
DINIS 36
10
In a note, Lang (p. 138) suggests putting bem at the end of v. 1 (in his text the line ends with oj). In both B and V the first
two lines end with ousasse and al respectively (cf. Langs note on p. 174 ad CXII [vv. 2281-2]).
2
Cunha (1982: 49) makes a second, tempting suggestion: Mya madre, que end o poder / , vos sabe gran mal querer. A corruption
of Au9 to E u9 (B) or eu9 (V) is plausible, and the enjambment would parallel that of desamor at the corresponding place in the
first strophe. Cf. B 581 / V 184 (Dinis), v. 3 que end o poder .
3
The potential quer<r>ia would match the mood of pesa<ria> in v. 6. queria, if correct, is a modal imperfect. Cf. B 580 / V 183
(Dinis), v. 20 and note.
624
DINIS 37
10
625
DINIS 38
10
15
20
626
4 uijr B : uur V
ue~ B : ueu V
7, 16 ue~ B : ne~ V
9 er B : et V : e Lang
11 eu] en V
DINIS 39
10
15
20
3 aiades B : ayades V
4 boa BV
5 guysa B : gysa V
627
DINIS 40
Vou m a la bailia
que fazen en vila
do amor
<Vou m a la bailada
que fazen en casa
do amor>
Que fazen en vila
do que eu ben queria
do amor
10
15
20
628
cf. Niermeyer, s.v. ballivia). On this (double) reading, the girl is going to the dance, to be sure, but also returning
to the jurisdiction of love. Cf. CSM 32.46 u dem os seus ten / na ssa baylia and 115.29 de o dare / ao dem en baylia
two passages which prompt Mettmann to define bailia as bailiado, jurisdio do bailio, poder. (See also the
Machados glossrio, s. v. bailia [vol. VIII, p. 184]).
v. 20 garrida: seems to mean frivolous here. The word, derived from Latin garrio, ~ire (chatter), is wrongly
defined by Lang (verliebt) and Nunes (namorado), based on the parallelism with Langs imaginary jurada in
v. 24. Corominas, s. v. garrido, correctly observes (vol. III, p. 109) that anteriormente signific traverso, ligero
de cascos, lascivo, deshonesto, but follows Langs error regarding our passage. The word occurs in the
rubric of CSM 79 Como Santa Maria tormou a meni~a que era garrida, corda, e levou-a sigo a Parayso, while the diminutive
garridelinna appears in the text of the same poem (CSM 79.14-15): Aquesto foi feito por hu~a meninna / que chamavan
Musa, que mui fremosinna / era e aposta, mas garridelinna / e de pouco sen. Cf. CSM, s. v. garridena.
1
Lang (followed by Nunes) prints jurada, partly to cut a syllable, and Nunes takes the imperfects queria (vv. 14, 19) and amava
(vv. 17, 22) as presents (vol. III, p. 48) to fit the resulting sense.
629
DINIS 41
10
15
630
DINIS 42
10
15
20
10 co<n>fonda
Nunes gives the direct order of the phrase as de outra filhar-lhe o seu amigo (vol. III, pp. 50-51). If the infinitive is
personal, the subject should be singular, but cf. B 583 / V 186 (Dinis), vv. 2-3 and note. The plural outras could as easily be
a popular usage as the result of an error in transmission.
631
DINIS 43
10
15
14 hypermetric
v. 14: doutra sei eu que [o] ben sabia: Does this mean
another who knew well to whom... or another who
knew (it?) well, and to whom...?
v. 14: doutra sei eu que [o] ben sabia: significa outra que
bem sabia a quem... ou outra que (o?) sabia bem, e
a quem...?
Cunha (1982: 53) argues that this verse differs from vv. 4 and 9, to which it should be semantically equivalent. But his aim
is to argue (against Lapa) that que o cannot be scanned as a single syllable. Given the number of cantigas d amigo in which
the verse before the last refrain is a syllable long (see Introduction, 5), I see no problem in admitting a hypermetric verse
here, but see vv. 14, 17, 19, 22 in both B 592 / V 195 and B 568 / V 171. The question is how we interpret. Nunes says (vol. III,
p. 51) that o refers to what has just been mentioned, o gran mal e sabedoria (etc.), and this could be so.
632
DINIS 44
10
15
se~ B : seu V
15 eu Michalis
Nunes rightly prefers (vol. III, p. 710) the emendation of Michalis, U quer. This envolves the change of a single letter (q for
i), yields clear sense and syntax (wherever/whenever it is without you), and can be defended by parallel expressions in
vv. 1 (u eu sejo) and 16 (u vos non vejo). Lang, though reading with BV, belatedly notes the lack of sense or syntax in viver que sen
vs seja (Der sinn der stelle ist nicht klar [p. 174]).
633
DINIS 45
10
15
20
5 end] and B
9 lh] l B
16 rogu V
vv. 9-10 pero lh estranhei o que...: O sentido geral : Embora eu o tenha repreendido pelo que me pedia cada
vez que me via; por outras palavras, embora eu tenha recusado os seus pedidos constantes.
v. 9 estranhei: Lang (s. v.) glosses estranhar here as einem etwas vorwerfen, rendered by Nunes (s. v., vol. III,
p. 618) as lanar em rosto. Nunes own understanding (mostrar a alguem o difcil de qualquer cousa) is
way off. Mettmanns gloss, detestar, is an overtranslation of the passages in CSM. Lapas glosses (CEM, s. v.
estranhar) repreender, censurar are based on one verse of Guilhade (B 1501, v. 22) Loureno, di-lhe que sempre
trobei / por bas donas, e sempr estranhei / os que trobavan por amas mamadas where more than one meaning may be
at play. Michalis (s. v. estranhar) explains: ficar surpreendido ou admirado desagradavelmente de alg. c.,
desaprov-la, censur-la, castig-la mentalmente. Nas cantigas damor, por meio de olhares, gestos de descontentamento e palavras; mas em prosas jurdicas e nos Nobilirios tambm por meio de actos pblicos e
oficiais, como multas (calumnias) ou mesmo no corpo. See also Corominas, s. v. extrao.
634
DINIS 46
10
<ama> Moura1
5 uyu V : uyo B
10 uyo BV2
635
DINIS 47
10
15
20
In favor of verei one could point to vv. 5, 11 and 17. But in all three places the word is bisyllabic (despite the spelling of the
MSS). For this verse to scan, we need averei (B) or trisyllabic ve<e>rei (read also by Lang, though not quite the reading of V).
A scribe might plausibly change aurey to urey but the reverse seems unlikely, given the parallel constructions in the other
strophes. For the expression, cf. CSM 105.47 por averen seu solaz.
636
DINIS 48
10
15
Cunha (1982: 33) would read with Nunes, but mainly to reject the reading of Lang <e> est o meu solaz, since that would
require the elision of e alleged by Lapa. On my reading, we have no pronoun est<o>. Rather, two main verbs (placed at the
beginning of the last two verses) in each strophe: praz and ste (I omit ben vos digo v. 4 and vej i bon dia v. 9 as parenthetic). The
subject of these verbs is constituted by the pois clauses in the first three verses of the each strophe. For example, vv. 1-5
could be rendered, Since my friend says he wants to go with me, since that pleases him, it pleases me, Im telling you, its
my solaz. The presence of the verb form ste in v. 12 seems telling (the construction is: since I see that taking me away from
here is [ste] his desire [i.e. since I see he wants to take me away]), but there, too, some editors have seen est (which,
however, will not scan). All this affects our conception of the strophic form. If I am right, Rep. 33:24 (6 6 4 6 6) is wrong
twice, since it depends on reading est (vv. 5, 10, 15) and eliding qu a el (Pellegrini; Lapa 1982: 158 stretches que a el<e> to
get six syllables). I take both pois que a el praz and ste o meu solaz to be five syllable verses.
637
DINIS 49
10
15
638
DINIS 50
10
15
20
9 sabaia V
11 todera B : ro dera V
18 mi~ B
22 end] and B
639
DINIS 51
10
640
7 todo V : tedo B
pderra V
9 n sei B : om. V
10 nen <ar>
DINIS 52
10
15
641
DINIS 53 (pastorela i)
~
Ua pastor se queixava
muit estando noutro dia
e sigo medes falava
e chorava e dizia
con amor que a forava:
Par Deus, vi t en grave dia,
ai amor
10
15
20
9 come(r) V
12 dezia BV1
13 mha] raha V
17 tru~has B :
v. 3 medes: found only here in Amigo) derives from Latin -met (enclitic particle attached for emphasis to certain
pronouns [OLD]) + ipse and may be borrowed from Provenal.
vv. 11-12: And who had never, since she was born,
been used to suffering.
Lang (p. 120) defends dezia (also printed by Nunes) and the form exists, but in this text at v. 4 both MSS have dizia, and
in B 534 / V 137 (another pastorela of Dinis) both MSS have dizia at v. 23.
642
10
15
20
25
30
The word senhora is extremely rare in the cantigas d amigo (the only certain examples are in B 1299 / V 903 [Esquio], vv. 3, 8,
13; but cf. B 1195 / V 800 [Caldas], v. 9). It occurs in this text (v. 24), to be sure, but senhor comprida may, as a formula, be
protected from morphological variation (Stegagno Picchio).
643
Vi oj eu cantar d amor
en un fremoso virgeu
u~a fremosa pastor
que ao parecer seu
ja mais nunca lhi par vi,
e por en dixi lh assi:
Senhor, por vosso vou eu
Tornou sanhuda enton,
quando m est ou dizer,
e diss: Ide vos, varon;
quen vos foi aqui trager
pera m irdes destorvar
du dig aqueste cantar
que fez quen sei ben querer?
Pois que me mandades ir
dixi lh eu Senhor, ir m ei,
mais ja vos ei de servir
sempr e por voss andarei,
ca voss amor me forou
assi que por vosso vou,
cujo sempr eu ja serei
10
15
20
25
30
644
645
646
TABLE OF CONTENTS
NDICE GERAL
13
Bibliography
15
Bibliografia
Introduction
31
Acknowledgements
51
Introduo
53
Agradecimentos
73
75
91
103
Conjunto das cantigas d amigo atribudas a cada trovador (por ordem crescente)
107
647
Rip Cohen
See Correia. Cf. Parkinson on false refrains in the Cantigas de Santa Maria.
Numbering and (unless otherwise indicated) texts of the cantigas damigo are from Cohen 2003.
Angle brackets are used for letters missing in the manuscripts, bold for refrains. Punctuation has been
altered, and tils added where historical phonology would expect them and thirteenth century
manuscripts of Galician-Portuguese lyric regularly supply them. The poets names are usually given in
abbreviated form; for the full forms, see Cohen 2003, 9-14, 102-05. Translations are from Cohen
2010b; I have sometimes provided ad hoc renderings of isolated verses and phrases.
2
Rip Cohen
10
V 757 f. 120r
2 la] ia B virgo] ugo B : ugo V 3-6 on two lines, with a break after
barqs BV 3, 8 damor V : Namor B 4, 9 nas] as Michalis (II, 880)
4 rio] mo V 7 ia] ya V : om. B 9-10 suppl. edd.: om. BV
Along the side of the river
The young girl went singing
Of love:
Let the boats come along the river
Just as I like.
Along the side of the stream
The noble girl went singing
Of love:
Let the boats come along the river
Just as I like.
Editors have filled in the last two verses of the second strophe by faithfully
repeating those of the first strophe. In so doing, they assumed that the song has a
three-verse refrain, and thus produced a cantiga with a three-verse refrain. But what
we have is something else: a banal error of omission due to a scribe higher up the
stemma, long before the copying of the Italian apographs in the early sixteenth
century.3 Against all previous editors, I propose to read alto at the end of the
penultimate verse, understanding it not as part of the refrain but as a variable verse
belonging to the body of the strophe:
<Venhan-nas barcas polo alto
a sabor>.
With this reconstruction of the last two verses, the poem has an intercalated refrain
with a regular rhyme-scheme aaBaB, with assonant rhymes in the first two verses of
each strophe (I: rio/virgo; II: alto/algo). It does not switch from aaBAB in the first
Rip Cohen
strophe to aaBCB in the second.4 Furthermore, there is a dobre in vv. 1 and 4 of each
strophe: I rio; II alto. In retrospect, the dobre with the word rio in the first strophe
should have made editors wary of breaking the pattern. We should not read the
darkness so blithely.
The use of an intercalated two-verse refrain is not unusual in the cantigas damigo.
There are 26 examples (see Appendix), not counting Zorro 6, and our poet provides
two of them: Jus a lo mar e o rio, with the rhyme-scheme aBaB (Zorro 8); and
Bailemos agora, por Deus, ai velidas, with the rhyme-scheme aaaBaB (Zorro 10). In
the latter poem there is a double dobre unique is this genre.5 So the reading I propose
recommends itself prima facie, but before looking further into its merits we might first
ask how we got a vulgate that reads rio.
The original source of the error in Zorro 6 was the presumption, on the part of a
scribe, that the refrain consisted of three verses. This kind of mistake belongs to a
typology of scribal slip-ups regarding strophic form and is the reverse of a false
refrain in the sense used by Parkinson, where a scribe creates a refrain that was not
there by recopying the wrong verses in the place appropriate to a refrain.6 In our case
it was a scribes failure to keep writing that allowed the false image of a three-verse
refrain. This image has been retained and propagated in modern times by editors who
believe that if the scribe left the rest of a strophe uncopied there must have been a
refrain there. We can see from a few examples that this was not always the case.
Consider Johan Airas 33:
Vedes, amigo, ond ei gran pesar:
sei muitas donas que saben amar
seus amigos e soen lhis falar
e non lho saben, assi lhis aven;
e ns sol que o queiramos provar,
log sabud e non sei eu per quen.
Tal dona sei eu, quando quer veer
seu amigo a que sabe ben querer,
que lho non poden per ren entender
o<s> que cuidan que a guarda<n> mui ben;
e ns sol que o queiramos fazer,
<log sabud e non sei eu per quen>.
10
Here I use bold for all rhyme schemes (my current practice is to use bold only for long verses with
internal rhyme).
5
In a sequential performance (or reading) of Zorros songs, in the order in which they are found in BV,
the double dobre in Zorro 10 would have been preceded and so in effect introduced by the dobre in
Zorro 6.
6
There are half a dozen examples of this in the cantigas damigo. See, for example, Roi Fernandiz 4,
where uel falasse comigo printed as a third verse of the refrain in Nunes (and so in Brea; see Cohen
2003, 330) and Charinho 1, usually printed with a four-verse refrain (see Cohen 2003, 297-98).
Rip Cohen
15
20
Rip Cohen
shows that the rhyme scheme is a steady aaabaB and does not shift from *aaabAB to
*aaabCB (as we might have thought, if we had only B to edit).
Another similar error is to be found in Bolseiro 5:
Vej eu, mha filha, quant meu cuidar,
as barcas novas vir pelo mar
en que se foi voss amigo daqui.
Non vos pes, madre, se Deus vos empar,
irei veer se ven meu amig i.
10
15
Rip Cohen
the mistake, though similar to the one in Zorro 6, is an error of commission, not of
omission. In v. 9 the common ancestor of BV evidently had (expanding manuscript
abbreviations) Non vos pes madre se deus vos empar, which is identical to v. 4, the
corresponding line in the first strophe, even though this leaves the verse without a
rhyme. Probably, if we had no third strophe to help us catch this howler (corrected by
Bell, resuscitated by Nunes), we would be reading a poem that formally resembles the
standard version of Zorro 6: aaBAB in the first strophe, and aaBCB in the second.
In the case of Zorro 6, the error was committed by a scribe who probably glanced
quickly and thought the last three verses of both strophes were identical. Thus, like the
scribe of B in Johan Airas 33, this copyist did not bother finishing the rest of the
strophe. But here we cannot be saved by the scribe of V, since the error was in the
common source of BV. Nor can we be helped by a third strophe, as in Bolseiro 5,
because there is none. We can be rescued, as we were by Bell in Bolseiro 5, v.9,
simply by performing the two most basic operations of textual criticism: to identify
what is wrong and to correct it.
In addition to the case against rio, there is evidence that speaks in favor of alto: the
dobre, the strophic form, and the use of alto almost exclusively in the parallelistic
rhyme-pair rio/alto. These three elements are inextricably interconnected in this text,
so as we focus on each we will necessarily make reference to the others. But each also
deserves some individual attention.
Let us begin with the dobre. In Zorro 10, velidas/frolidas and loadas/granadas
figure in a double dobre in a poem of the form aaaBaB. Moreover, the fifth verse of
both strophes is a variable verse located between the two lines of the intercalated
refrain, and those variable verses (I.5, II.5) participate in one of the dobres (I.2 and I.5:
frolidas; II.2 and II.5: granadas) exactly as in our proposed version of Zorro 6,
except that here there are two dobres.
Bailemos agora, por Deus, ai velidas,
so aquestas avelaneiras frolidas
e quen for velida, como ns velidas,
se amigo amar,
so aquestas avelaneiras frolidas
verr bailar.
Bailemos agora, por Deus, ai loadas,
so aquestas avelaneiras granadas
e quen for loada, como ns loadas,
se amigo amar,
so aquestas avelaneiras granadas
verr bailar.
10
Rip Cohen
There is no example of this in the cantigas damigo. But see OsoirAnes 1 in Cohen 2009b, 18-19 and
the note ad loc. (33).
8
The only dobres unissonans are Guilhade 7, with coita in each strophe; and Johan Airas 34, which
though unissonans has a different word in each strophe: i, mi, assi (see Cohen 2009a, 131-32). For the
terminology see Lorenzo Gradn 1997.
Rip Cohen
for editors to repeat it without reflection is sloppy textual criticism. The poem should
read like this:
Pela ribeira do rio
cantando ia la dona virgo
d amor:
Venhan-nas barcas polo rio
a sabor.
10
And what can we say about the use of the word alto in this genre? It appears in six
cantigas damigo, a total of twelve times (not counting the reading proposed here).
Ten times it is a noun meaning river or stream, an archaic usage found only in the
cantigas damigo. Meendinho uses alto mar, where as an adjective alto could be taken
to mean either deep or high (vv. 10, 17); and these are the only times the word
does not appear at verse-end. Perhaps it is not coincidental that alto occurs in the last
verse (not counting the refrain) of the last of the nine cantigas of Pero Meogo: nunca
vi cervo que volvess o alto (I never saw a stag that stirred the stream). And there
too it forms part of the rhyme-pair rio/alto. In the enigmatic Levantou-s a velida
(The lovely girl arose) of D. Dinis (17) alto appears in an intercalated three-verse
refrain nearly as a line unto itself: eno alto.9 But Johan Zorro is the only poet who uses
the word in more than one text. It occurs in three of his eleven songs as the rhymeword in the first verse of the second strophe, in parallelistic alternation with rio.10
Zorro 3: Per ribeira do rio (I.1); Per ribeira do alto (II.1)
Zorro 6: Pela ribeira do rio (I.1); Pela ribeira do alto (II.1)
Zorro 8: Jus alo mar e o rio (I.1); Jus alo mar e o alto (II.1)
We find, then, that alto, used as an archaic substantive meaning river, is always
found at verse-end. It appears once in a refrain (Dinis 17) where it does not rhyme.
The other times, in Meogo 9 and three songs of Zorro, it is used as part of a rhymepair in i-o/a-o with rio. And the parallelistic alternation in rhyme of rio and alto
appears to have been a traditional element in cantigas damigo with water imagery and
9
See Cohen 2006 on the textual and interpretative problems of Dinis 17. In that text alto does not
rhyme, which is only possible in a refrain (or in the case of a palavra perduda).
10
rio appears without alto in two of Zorros other cantigas: Zorro 7, v. 1: Met el rei barcas no rio
forte; Zorro 9, v. 1: Pela ribeira do rio salido; v. 5: Pela ribeira do rio levado.
Rip Cohen
to have had deep roots in the formal, rhetorical, and symbolic conventions of the
genre. The traces of this alternation in the extant corpus may reasonably be taken as
evidence of its presence in the poetic matrix.11 Certainly Zorros use of this pair in
three of his cantigas shows that he regarded it as something more than a useful
technical resource (see Ferreira, 66-68, 74-81).
Some might argue that Zorro uses rio as a non-rhyming word at verse-end in a
refrain, as Dinis uses alto (and rio is found once in a refrain in this genre: Estevan
Coelho 2). But in Dinis 17 alto occurs in all six strophes, clearly part of an intercalated
refrain. The other occurrences of alto in the cantigas damigo of Johan Zorro compel
the conclusion that this poet regularly handled alto as one half of the traditional
rhyme-pair rio/alto and that our conjectured alto in v. 9 occurs in a variable not a
fixed verse; that is, in a verse belonging to the body of the strophe, not to the refrain.
Finally, the dobre pattern created by rio in vv. 1 and 4 of the first strophe requires the
presence of alto in the corresponding positions in the second strophe: v. 6, where we
have it; and v. 9, where it is missing in the manuscripts along with the rest of the two
final lines.
One might say that, in the end, like all questions of textual criticism, it is a matter
of judgment. Indeed. And the equivalent errors in Johan Airas 33, v. 11, where a
scribe erroneously assumes the verse to be part of the refrain, and in Bolseiro 5, v. 9,
where BV both have the wrong word at the end of the verse (empar instead of
perdon), seem sufficient evidence to weigh the scales in favor of alto. The scale with
alto plunges when we reflect that rio creates an anomalous rhyme-system for the
cantiga and that alto produces a song with a perfect aaBaB form. The case is sealed
by the dobre, paralleled in Zorros own poetic practice and required by the dobre in
the first strophe. The repeated use by Zorro of the rhyme pair rio/alto fits in with all
these considerations and corroborates the reading alto.
Although rio was a mere conjecture when it first appeared in 1878 in Bragas
edition of V, it has been sanctified by the inertia of generations of editors. But whereas
alto has compelling evidence and arguments in its favor, rio has none.12 All we need
to do, as Housman says, is to apply thought to textual criticism, the science of finding
mistakes and the art of correcting them.13 But in the cantigas damigo there is science
even in that art. The study of the manuscripts, of scribal procedures, of strophic
forms, rhyme-systems, technical virtuosity (here, the dobre), lexicon, traditionally
paired rhyme-words, the poetics of the genre, and the poets own practice all these
are methods, ways of knowing how to correct, reasonably, if not with absolute
certainty, what is wrong.
11
Here I mean both the contemporary poetic matrix an abstraction based on all we can infer from the
corpus about the poetics (form, rhetoric and pragmatics) that generated these 500 songs and the
historically prior poetic matrix in which the genre had its roots.
12
For other corrections to the text of Johan Zorro, see Cohen 2010a, 25-26, 35-38.
13
Textual criticism is a science, and, since it comprises recension and emendation, it is also an art. It is
the science of discovering error in texts and the art of removing it (Housman, 67).
Rip Cohen
10
Rip Cohen
11
Rip Cohen
12
Rip Cohen
13
2.
Critical Rights and Erotic Wrongs:
Emendation and Action in Johan Perez dAvoin 1
There are two kinds of sin in textual criticism: to fail to recognize what is wrong
(and correct it, if possible); and to change what is already right either in the
manuscript(s) or in a critical edition. Here, using among other things the pragmatics
of the genre, I would like to propose an emendation in the text of a cantiga damigo of
Johan Perez d Avoin (1) where I left intact a glistering error (Cohen 2003, 151). The
text as it appears in that edition, with the sin uncorrected, reads as follows.
Quando se foi noutro dia daqui
o meu amigo, roguei lh eu por Deus,
chorando muito destes olhos meus,
que non tardass e disse m el assi:
que nunca Deus lhi desse de mi bem
se non vesse mui ced, e non ven.
10
15
20
22 en de ? B : eu de V
Rip Cohen
14
The text of Sandeu 3 as transmitted by the manuscripts, where there appears to be an unrhyming
verse, is almost certainly corrupt, but no convincing correction has yet been found. See Cohen 2003,
266.
Rip Cohen
15
15
15
Lorenzo Gradn (2008a, 116) prints queimei in v. 14 but attributes the correction to Braga, even
though it belongs to Monaci, as reported by Cohen (2003, 226). She also rejects <e>, erroneously
believing it would produce a hypermetric verse. There is no problem with the scansion: non por mha
alma_e candeas queimei.
Rip Cohen
16
Rip Cohen
17
Eu era maravilhada
por que tan muito tardava,
pero sempr esto cuidava,
se eu del seja vingada,
Friend, today I heard
A message from my boy:
That hes alive and in love
With another lady, Im telling you,
But I swear to God I would
Rather hear that he was dead.
I was just amazed
Why he was taking so long
But I kept thinking this
So may I get back at him!
The girl had been amazed at how long her boy was taking to return (vv. 7-8), and
had suspected the cause (vv. 9-10), but now she knows why: he is indeed alive, but in
love with another girl (vv. 1-4). She swears she would rather hear that he was dead
(refrain).
One way of referring to infidelity is the verb errar. This usage was already noted
by Lang in his 1894 edition of the cancioneiro of D. Dinis, in both the glossary and
the introduction (Lang 2010, 115, 256 s.v. errar).16 Among other texts, he cites Dinis
22. There the girls ironic expressions of astonishment at the boys tarrying (strophes I
and III) are understood by her girlfriend to be accusations of infidelity, and she
promptly defends the boy against the implicit charge (strophes II and IV). The
girlfriend has no doubts about what the girl means. And when a persona in a text
interprets an expression, we should pay heed. As the poets creation, she knows the
conventional language of the genre and the connotations of words and phrases far
better than we do.
Here is a passage from Pae Gomez Charino (5, vv. 7-10) where errar appears:
Non sei, amiga, que foi ou que
ou que ser, ca sabemos que non
vos errou nunca voss amigo, e son
16
None of the examples of errar in the glossaries of Michalis or Lapa seems relevant; but the glossary
of CSM provides pellucid parallels for an erotic meaning of the verb (see below).
Rip Cohen
18
10
15
Rip Cohen
19
Rip Cohen
20
In a poem outside the genre, but closely related, a pastorela of D. Dinis (54, vv. 56 and 13-15), the verb errar appears twice.
E diss, Oimais non nada
de fiar per namorado
nunca molher namorada,
pois que mh o meu errado.
15
10
Rip Cohen
21
20
Rip Cohen
22
Rip Cohen
23
correspond; Cohen 2010d), the reading of the manuscripts can be checked against
strophic form, meter and rhyme. And any text must be checked for grammar and
sense. But we cannot really check meaning without understanding pragmatics (Cohen
2010a).
Here, in righting the wronged text, we have discovered in this song of Avoin more
than a rhyme. We have found one more instance of erotic wronging that fits into a
widespread phenomenon in the pragmatics of the genre, where the amigo often
betrays, or is thought to betray, his amiga by tarrying elsewhere with another girl.
The pragmatics of the cantigas damigo had long been neglected and only recently
has become a methodological concern (for a theoretical introduction, see Cohen 2011,
95-102). Maybe the speech-actions (Bing and Cohen, 19-21) of a nubile girl, her
mother, girlfriend and boyfriend, have not seemed important enough to warrant study,
much less to develop methods of analysis. Editors and commentators rarely provide
more than vague and often inaccurate descriptions of what happens in any given
text. Yet this corpus of 500 female-voiced love songs provides an ideal laboratory for
the study of action.
Accurate descriptions of kinds of speech-action can be conveniently conceived of
as scripts whose general form can be notated as 1 2 { , + } (Parkinson
and Cohen, 37-39; Cohen 2010a; 2011, 98-99). This notation identifies speaker and
addressee (P1 and P2) and describes background and new information (x, y + z)
leading to () a present action or emotion (A). A grammar of scripts, still under
construction, will be as useful for the textual critic as historical grammar, meter or
rhyme. The script of Avoin 1 would be:17
{
The correction merrou in Avoin 1, v. 20, is one example of the utility of pragmatics
as a check for the textual critic. To make full use of this check we need a complete
description of the actions represented in this genre (Cohen 2011, 135-37). Such a
description will not be easy to construct, but neither is a manual of historical
morphology, phonology or syntax. They are all tools that we cannot do without.
Philology makes use of, and contributes to, many fields. One of the most
fundamental, always presupposed but nearly never mentioned, is epistemology (on which
see Wittgenstein 1979). When we say we know what a text means, or how to identify
and, with knowledge and skill, correct a corrupt text, we must be able to say how we
know, or at least to explain how we think we know. And we can do that only if we have
reliable methods. A grammar of scripts is and must become accepted as one of the
basic methods needed to edit and interpret the cantigas damigo.
17
G = girl; = no identified addressee. Hence G means that the girl speaks to no addresse (or to an
unidentified one).
Rip Cohen
24
Rip Cohen
25
---. Hello, Honey: The Pragmatics of Greeting in the Cantigas dAmigo. In Stephen
Parkinson ed. Cantigas United. Papers on the Medieval Galician-Portuguese
Lyric. London: Department of Iberian and Latin American Studies, Queen
Mary, University of London. Forthcoming.
---, & Federico Corriente. Lelia doura revisited. La Cronica, 31.1 (2002): 19-40.
---, & Stephen Parkinson. The Galician-Portuguese Lyric. In Stephen Parkinson,
Cludia Pazos Alonso & T. F. Earle eds. Companion to Portuguese Literature.
Warminster: Tamesis, 2009. 25-44.
Correia, ngela. O refram nos cancioneiros Galego-Portugueses (Escrita e
tipologia). (Masters Thesis, Universidade de Lisboa [I: Dissertao; II:
Transcrio; III: Repertrio - Lista por tipos; IV: Repertrio - Lista por
autores]). 1992. 4 vols.
Cunha, Celso Ferreira da ed. O cancioneiro de Joan Zorro. Aspectos lingsticos.
Texto crtico/Glossrio. Rio de Janeiro: [Imprensa Nacional], 1949.
---. Cancioneiros dos Trovadores do Mar. In Elsa Gonalves ed. Lisbon: Imprensa
Nacional-Casa de Moeda, 1999 [contains Cunha 1949].
CSM = Walter Mettmann ed. Afonso X, O Sbio. Cantigas de Santa Maria. Coimbra:
Por Ordem da Universidade, 195972. 4 vols. (rpt. Vigo: Edins Xerais de
Galicia, 1981. 2 vols.).
Ferreira, Maria do Rosrio. guas Doces, guas Salgadas. Da funcionalidade dos
motivos aquticos nas Cantigas de Amigo. Porto: Granito, 1999.
Gonalves, Elsa. Tradio manuscrita da poesia lrica. In Giulia Lanciani &
Giuseppe Tavani, orgs. Dicionrio da Literatura Medieval Galega e
Portuguesa. Lisbon: Caminho, 1993. 627-32.
Housman, A. E. The Application of Thought to Textual Criticism. Proceedings of
the Classical Association 18 (1922): 67-84.
Lang, Henry R., ed. Das Liederbuch des Knigs Denis von Portugal. Zum ersten mal
vollstndig herausgegeben und mit Einleitung, Anmerkungen und Glossar
versehen. Halle a.S.: Max Niemeyer, 1894 (rpt. Hildesheim and New York:
Georg Olms Verlag, 1972).
---. In Lnia Mrcia Mongelli & Yara Frateschi Vieira orgs. Cancioneiro dEl Rei
Dom Denis e Estudos Dispersos. Niteri/Rio de Janeiro: Xunta da Galicia/
EdUFF, 2010.
Lapa, Manuel Rodrigues, ed. Cantigas descarnho e de mal dizer dos cancioneiros
medievais galego-portugueses. Vigo: Editorial Galaxia, 1970.
Lorenzo Gradn, Pilar. El dobre gallego-portugus o la esttica de la simetria. Vox
Romanica (1997) 56: 212-41.
---, ed. Don Afonso Lopez de Baian. Alessandria: Edizioni dellOrso, 2008.
Machado, Elza Pacheco, & Jos Pedro Machado, eds. Cancioneiro da Biblioteca
Nacional, Antigo Colocci-Brancuti. 8 vols. Lisbon: Edio da Revista de
Portugal, 1949-64.
Rip Cohen
26