Vous êtes sur la page 1sur 250

Bosse_cpi_cb_rauch dd no metallic.

qxd 4/11/2015 9:22 AM Page 1

93 BERKELEY
INSIGHTS

Bosse
IN LINGUISTICS
Applicative Arguments: A Syntactic and Semantic Investiga- AND SEMIOTICS
tion of German and English presents formal semantic and
syntactic analyses of German and English applicative argu- 93
ments. These arguments are nominal elements that are not
obligatory parts of a sentence. Both German and English

Applicative Arguments
have several types of applicative arguments, including so-
called benefactive and malefactive constructions. More
specifically, the research relies on tests to differentiate the
different types of applicative arguments based on this contri-
bution to meaning: Some applicatives contribute only not-at-
issue meaning, whereas others contribute only at-issue
meaning, and still others contribute both types of meaning.
These tests are applied to both German and English to
uniquely identify the applicative arguments in each language.
Formal analyses of the identified type of applicative argu-
ments are presented that provide an account for each type of
applicative identified for each language, explaining the
applicatives’ differences and similarities.

Solveig Bosse received her Ph.D. in linguistics from the


University of Delaware. She is currently Assistant Professor of
Theoretical Linguistics in the Department of English at East
Carolina University. Her research focuses on syntactic and
formal semantic analyses of German and English with occa-
sional other cross-linguistic comparisons.
APPLICATIVE
ARGUMENTS
A Syntactic AND Semantic
PETER LANG

Investigation OF
German AND English

WWW.PETERLANG.COM SOLVEIG BOSSE


Bosse_cpi_cb_rauch dd no metallic.qxd 4/11/2015 9:22 AM Page 1

93 BERKELEY
INSIGHTS

Bosse
IN LINGUISTICS
Applicative Arguments: A Syntactic and Semantic Investiga- AND SEMIOTICS
tion of German and English presents formal semantic and
syntactic analyses of German and English applicative argu- 93
ments. These arguments are nominal elements that are not
obligatory parts of a sentence. Both German and English

Applicative Arguments
have several types of applicative arguments, including so-
called benefactive and malefactive constructions. More
specifically, the research relies on tests to differentiate the
different types of applicative arguments based on this contri-
bution to meaning: Some applicatives contribute only not-at-
issue meaning, whereas others contribute only at-issue
meaning, and still others contribute both types of meaning.
These tests are applied to both German and English to
uniquely identify the applicative arguments in each language.
Formal analyses of the identified type of applicative argu-
ments are presented that provide an account for each type of
applicative identified for each language, explaining the
applicatives’ differences and similarities.

Solveig Bosse received her Ph.D. in linguistics from the


University of Delaware. She is currently Assistant Professor of
Theoretical Linguistics in the Department of English at East
Carolina University. Her research focuses on syntactic and
formal semantic analyses of German and English with occa-
sional other cross-linguistic comparisons.
APPLICATIVE
ARGUMENTS
A Syntactic AND Semantic
PETER LANG

Investigation OF
German AND English

WWW.PETERLANG.COM SOLVEIG BOSSE


APPLICATIVE
ARGUMENTS
B E R K E L E Y
I N S I G H T S
I N LINGUISTICS
AND SEMIOTICS

Irmengard Rauch
General Editor

Vol. 93

This book is a volume in a Peter Lang monograph series.


Every volume is peer reviewed and meets
the highest quality standards for content and production.

PETER LANG
New York  Bern  Frankfurt  Berlin
Brussels  Vienna  Oxford  Warsaw
SOLVEIG BOSSE

APPLICATIVE
ARGUMENTS

A Syntactic AND Semantic


Investigation OF
German AND English

PETER LANG
New York  Bern  Frankfurt  Berlin
Brussels  Vienna  Oxford  Warsaw
Library of Congress Cataloging-in-Publication Data
Bosse, Solveig.
Applicative arguments: a syntactic and semantic investigation
of German and English / Solveig Bosse.
pages cm. — (Berkeley insights in linguistics and semiotics; v. 93)
Originally published as author’s dissertation entitled
“The syntax and semantics of applicative arguments in German and English”, 2011.
Includes bibliographical references and index.
1. Grammar, Comparative and general—Applicative constructions.
2. German language—Syntax. 3. German language—Semantics.
4. English language—Syntax. 5. English language—Semantics. I. Title.
P291.25.B67 425—dc 3 2014010339
ISBN 978-1-4331-2726-7 (hardcover)
ISBN 978-1-4539-1354-3 (e-book)
ISSN 0893-6935

Bibliographic information published by Die Deutsche Nationalbibliothek.


Die Deutsche Nationalbibliothek lists this publication in the “Deutsche
Nationalbibliografie”; detailed bibliographic data are available
on the Internet at http://dnb.d-nb.de/.

© 2015 Peter Lang Publishing, Inc., New York


29 Broadway, 18th floor, New York, NY 10006
www.peterlang.com

All rights reserved.


Reprint or reproduction, even partially, in all forms such as microfilm,
xerography, microfiche, microcard, and offset strictly prohibited.
Acknowledgments
This book is a revised version of my Ph.D. dissertation (2011, University of
Delaware). Consequently, it would not have been possible without the help of
Dr. Benjamin Bruening. Thank you for your support throughout the years! Many
thanks go to my Ph.D. committee members Dr. Satoshi Tomioka, Dr. Gabrielle
Hermon and Professor Dr. Daniel Hole for their support and invaluable com-
ments. Thank you to everyone at the Department of Linguistics at the University
of Delaware!
Many thanks to all the different audiences that have provided feedback on
my work at conferences and workshops over the last few years. I enjoyed the
discussions and they provided me with many details discussed in this book.
A big THANK YOU to everyone who has supported me since I graduated,
especially to the members of the Department of Linguistics at The University of
Iowa and very special thanks to the Department of English at East Carolina Uni-
versity. Thank you for your continued support.
To my parents Raphaele and Arnd Bosse: Thank you for you support, en-
couraging words, and for always being there for me – even if there's an ocean
between us. DANKE!
Last but not least to all of my friends, new and old, for your support: Thank
you for being there when I needed to vent or celebrate. Time passes but the
memories will stay forever.


Contents
1 Introduction 11
1.1 Defining Applicative Arguments 13
1.2 Goals and Structure of this Book 17
2 Preliminaries 19
2.1 Introduction 19
2.2 Not-At-Issue Meaning 19
2.3 Event Semantics 23
2.4 Pylkkänen (2002, 2008) 26
3 Affected Experiencers 29
3.1 Introduction 29
3.2 Distinguishing Affected Experiencers 30
3.2.1 Benefactives (Chapter 5) 31
3.2.2 Part-Whole Applicative Arguments (Chapter 6) & Pertinence
Datives 32
3.2.3 Other German Applicatives 35
3.3 At-Issue and Not-At-Issue Meaning 36
3.3.1 Affected Experiencers 36
3.3.2 Pertinence Dative 38
3.4 Analysis 40
3.5 Consequences 44
3.5.1 Bi-Eventivity 44
3.5.2 Again-Modification 46
3.5.3 Negation 48
3.6 Pertinence Dative 49
3.6.1 Analysis 50
3.6.2 Previous Analyses 56
3.6.3 Conclusion 65
3.7 Verbal Restrictions 65
3.7.1 Valency 65
3.7.2 Semantic Licensing 67
3.8 Affected Experiencers following Potts (2005) 72
3.9 Digression: Parametric Variation of Aff 73
3.9.1 Japanese: Attachment Height Variation 73
3.9.2 Not-At-Issue Affected Experiencers 78
3.9.3 Parametric Variation 86
3.10Conclusion 87
4 Not-At-Issue Applicative Arguments 89
4.1 Introduction 89
4.2 Ethical Dative 89
4.2.1 Description 90


8   

4.2.2 Analysis 96


4.2.3 Summary 100
4.3 Subject Co-Referential Applicative Arguments 101
4.3.1 Description 101
4.3.2 Features 102
4.3.3 Form 104
4.3.4 Role of the Direct Object 106
4.3.5 Not-At-Issue Meaning 106
4.3.6 Analysis 108
4.3.7 Consequences 109
4.3.8 Previous Analyses 111
4.3.9 Summary 115
4.4 Not-At-Issue Applicatives 115
4.5 Conclusion 117
5 Benefactives 119
5.1 Introduction 119
5.2 Description 119
5.2.1 English 121
5.2.2 German 122
5.3 At-Issue Content 124
5.4 True Applicative Benefactives 125
5.4.1 Analysis 125
5.4.2 Consequences 126
5.4.3 Verbal Restrictions 129
5.4.4 Hole (2005; 201X) 131
5.4.5 Summary 135
5.5 Recipient Benefactives 135
5.5.1 Meaning 135
5.5.2 Ditransitives and Benefactives 136
5.5.3 Analysis 137
5.5.4 Consequences 138
5.5.5 Summary 140
5.6 Prepositional Alternation 140
5.6.1 Analysis 140
5.6.2 Consequences 143
5.6.3 PP-shells 144
5.7 Conclusion 147
6 Part-Whole Applicatives 148
6.1 Introduction 148
6.2 Description 150
6.2.1 Differentiating Part-Whole Applicative Arguments 150
6.2.2 Characteristics 152


  9

6.2.3 Restrictions on Part-Whole Applicative Arguments 154


6.3 At-Issue Meaning 158
6.4 Analysis 161
6.4.1 Possessor Raising 161
6.4.2 Pylkkänen (2002) 161
6.4.3 Hole (2008) 163
6.4.4 Tomioka and Sim (2007) 167
6.4.5 Analysis 170
6.5 Consequences 175
6.6 Conclusion 176
7 Dative of Inaction 178
7.1 Introduction 178
7.2 Description 179
7.3 At-Issue Meaning 182
7.4 Analysis 183
7.5 Consequences 185
7.6 Conclusion 186
8 Co-Occurrence of Applicatives 188
8.1 Introduction 188
8.2 German 188
8.2.1 Ethical Dative 190
8.2.2 Dative of Inaction 192
8.2.3 Subject Co-referential Applicative 193
8.2.4 Affected Experiencer 194
8.2.5 True Benefactive 195
8.2.6 Part-Whole Applicative 198
8.2.7 Summary 199
8.3 English 200
8.4 Conclusion 202
9 Conclusion 203
Appendix 1: Affected Experiencers 207
Appendix 2: Not-At-Issue Applicatives 213
Appendix 3: Benefactives 215
Appendix 4: Part-Whole Applicative 221
Appendix 5: Dative of Inaction 225
Glossary 227
Bibliography 231
Index 239

 

1 Introduction
This book is concerned with detailed formal semantic and syntactic analyses of
applicative arguments in German and English. These arguments are typically
defined as nominal constituents not selected by the lexical verb or a preposition
of the sentence they appear in. In other words, they seemingly increase (or re-
flect an increase of) the verb's valency. One of my goals is to provide an over-
view of different types of applicative arguments found in German and English
and to provide formal tests and characteristics that differentiate the types. In (1),
an example of each identified type is given: part-whole applicatives (1a, Ger-
man), true benefactives (1b, German), recipient benefactives (1c, English), af-
fected experiencers (1d, German and English), ethical applicatives (1e, Ger-
man), subject co-referential applicatives (1f, German and English), and Datives
of Inaction (1g, German).
1. a. Jan zerbeulte dem Auto die Tür.
Jan dented the.Dat car the.Acc door
'Jan dented the door of the car.'
b. Jan buk ihm einen Kuchen.
Jan baked him.Dat a cake
'Jan baked a cake for him.'
c. John brought him a cake.
d. Alex zerbrach Chris Bens Vase.
Alex broke Chris.Dat Ben.Gen vase
'Alex broke Ben's vase on Chris.'
e. Du sollst mir pünktlich nach Hause kommen.
you shall me.Dat on.time to house come
'You shall be home on time (and I want that to happen).'
f. John1 killed him1 a bear.
g. Der Oma zerbrach die Vase.
the.Dat grandma broke the.Nom vase
'The vase broke on grandma's watch.'
Each of these types will be described in detail to unambiguously identify and
characterize that type. This is important because the different types often seem
identical on first glance, especially in German.
Applicative arguments are not a new phenomenon. Abraham (1973) was one
of the first researchers who tried to differentiate German applicative arguments
based on their syntactic behavior. Before that, they were typically classified
based on their meaning alone. Given the framework Abraham was working in,


12   

his classifications are not adequate anymore. A re-evaluation of the types of ap-
plicative arguments in German is necessary.
English applicatives have also been studied before: the recipient benefactive
has extensively been discussed, especially in combination with the structure of
ditransitive verbs, which they resemble. The other two types have received less
attention. I provide formal analyses of all types, showing why previous analyses
need to be revised.
The main dimension that I use to differentiate and analyze applicative argu-
ments is the type of meaning that each applicative argument contributes: at-issue
(roughly "asserted") meaning or not-at-issue (roughly "implied") meaning (Kart-
tunen and Peters 1979, Potts 2005). Elements contributing not-at-issue meaning
behave syntactically different than those that contribute at-issue meaning. I use
the family of sentence tests to distinguish at-issue applicatives from not-at-issue
applicatives. The tests are discussed in detail in Chapter 2. German and English
applicative argument constructions have never been systematically analyzed
with respect to this dimension. Both languages have applicative arguments that
contribute only at-issue meaning or only not-at-issue meaning. In addition, I
show that both languages also have one applicative argument that contributes
meaning to both tiers of meaning (Bosse et al. 2012).
A few other researchers have investigated applicative arguments that carry
not-at-issue meaning in other languages and/or in German and English (Lamiroy
and Delbecque 1998, O'Connor 2007, Halevy 2007, Gutzmann 2007, Horn
2008). These studies often lack formal analyses and do not set the discussed ap-
plicative argument in relation to other applicatives in that language. To do that is
another goal of this book.
Besides the meaning contribution, applicative arguments can be divided into
low and high applicatives, with the former involving a transfer of possession and
the latter relating an individual to an event (Pylkkänen 2002). Similarly to the
analysis by Cuervo (2003) for Spanish applicative arguments, I show that the
occurrences of applicative arguments in English and German are more varied
than expected under Pylkkänen's categorization of only high and low applicative
arguments.
The remainder of the introduction is dedicated to defining what applicative
arguments are. I develop my definition of applicative arguments based on Hole
(2008). He developed a definition for German applicative arguments which I
adapt to a more general notion of applicative arguments. It will be shown
throughout the book that this definition is too stringent, and that it cannot easily
be adapted to cover only the appropriate cases in a cross-linguistic setting. Thus,
it will remain a working definition.


  13

1.1 Defining Applicative Arguments


In this book, I focus on applicative arguments in German and English with occa-
sional references to other languages. As stated above, applicative arguments are
nominal elements of a sentence that are not selected by the lexical verb or a
preposition of that sentence. They are independent constituents of the sentence.
Consequently, optional elements that appear inside other nominal constituents
are not applicative arguments.
The nominal constituents I am concerned with have been called free, non-
core or applicative because of the fact that they can be added seemingly freely
to a sentence. This in turn is often taken to mean that when these constituents are
deleted from the sentence, the sentence is still grammatical. Such a German ex-
ample is given in (2).
2. a. Der Computer ist mir abgestürzt.
the computer is me.Dat crashed
'The computer crashed on me.'
b. Der Computer ist abgestürzt.
the computer is crashed
'The computer crashed.'
Sentence (2a) is grammatical without the dative NP mir 'me' (2b), showing that
the pronoun is an applicative argument. Due to this behavior, applicative argu-
ments are often characterized as valency-increasing because they look like a
verbal argument.
Hole (2008) argues against using this simple omission test to determine
whether a constituent should be considered an applicative argument. One reason
that he brings forth is that asking whether a sentence is grammatical without a
given element disregards the change in semantics that (potentially) follows from
the deletion of an element. Consequently, Hole (2008) proposes the entailment
omission test for free datives1.
3. The Entailment Omission Test for Free Datives2
A dative that is not governed by a preposition is free in a (simple, non-
negated declarative) sentence iff the sentence without the dative does not en-
tail that there is at least one individual3 that is involved in the asserted event

1 Hole is concerned only with German in which applicative arguments are dative-marked.
2 original: Die implikative Weglassprobe für freie Dative: Ein (nicht von einer Präposi-
tion regierter) Dativ ist frei in einem (einfachen nicht-negierten Aussage-)Satz gdw. der
Satz ohne den Dative nicht impliziert, dass es mindestens ein Individuum gibt, das an
dem assertierten Sachverhalt beteiligt ist und das durch einen Dativ ausgedrückt werden
könnte.
3 Individual here is a technical term of semantic types, not necessarily an animate entity.

 
14   

and could be referred to by a dative expression. (Hole 2008, 1.9, my transla-


tion)
I adopt this definition for applicative arguments with some minor adjust-
ments. I do not restrict applicative arguments to dative-marked constituents be-
cause languages that do not employ this case for applicative arguments are also
considered here. Furthermore, I take "involvement" here to be rather vague as it
can range from having special interest in the event coming true (as with ethical
applicatives) to being affected by it (as with affected experiencers) to benefitting
from it (true benefactive applicatives) to being a(n intended) recipient of an enti-
ty being part of the event (recipient benefactives) to a whole entity one part of
which is involved in the event (part-whole applicatives). Thus, my definition of
applicative arguments can be stated as follows.
4. Applicative Argument
An NP Y of a simple, non-negated declarative sentence that is not governed
by a preposition is an applicative argument iff the sentence without Y does
not entail that there is at least one individual that is involved in the asserted
event and could be referred to by Y.
This definition covers the core of what all applicative arguments share,
namely being independent constituents that are not selected by the verb; yet, the
definition is flexible enough to allow for different types of applicative argu-
ments. Furthermore, the definition is consistent with the fact that a given type of
applicative argument will receive the same interpretation regardless of the lexi-
cal meaning of the verb. This is not true of selected arguments.
My definition of applicative arguments (4) covers all examples in (1). First,
in none of the given examples is the applicative argument introduced by a pre-
position. Furthermore, the involvement of the referent of the applicative argu-
ment is not entailed in any of these examples. In (1a), the door is understood to
be a material part of the car; in (5), where the applicative argument is not pre-
sent, there is no entailment of the door being a material part of another entity
(beyond the specificity supplied by the definite determiner). The sentence is still
grammatical.
5. Jan zerbeulte die Tür.
Jan dented the.Acc door
'Jan dented the door.'
In (1b), the applicative argument is also not a necessary constituent of the
sentence (6). It denotes a beneficiary in (1b) but the existence/ involvement of a
beneficiary is not implied in (6).
6. Jan buk einen Kuchen.
Jan baked a cake
'Jan baked a cake.'


  15

In (1c), the applicative argument denotes a(n intended) recipient. Again, this
referent is not entailed in the sentence without the applicative argument (7). Fur-
thermore, (7) shows that the applicative argument is not selected by the verb be-
cause it is not required for the sentence to be grammatical.
7. Jan brought a cake.
(1d) is also grammatical without the applicative argument (8). This shows
that the applicative argument is not an obligatory constituent. (8) also does not
entail that there is anyone who is affected by the event.
8. Alex zerbrach Bens Vase.
Alex broke Ben.Gen vase.Acc
'Alex broke Ben's vase.'
The ethical dative of (1e) is not selected by the verb; it can be omitted from
the sentence without causing ungrammaticality (9). Furthermore, the involve-
ment of the referent of the ethical dative in (1e) is not entailed in (9).
9. Du sollst pünktlich nach Hause kommen.
you shall on.time to home come
'You shall be home on time.'
In (1f), the applicative argument is not required for the sentence to be
grammatical (10). Also, the applicative argument is not entailed (despite the
presence of the co-referential subject).
10. John killed a bear.
Finally, (1g) does not require the applicative argument to be present (11).
The involvement of an individual is also not entailed in (11) showing that der
Oma in (1g) is an applicative argument.
11. Die Vase zerbrach.
the vase broke
'The vase broke.'
The definition in (4) rules out implicit and cognate arguments as being ap-
plicative arguments. Such arguments, underlined in (12), are not governed by
prepositions. They are not necessary and may be omitted from the sentence.
12. a. He ate food.
b. She danced a dance.
However, these arguments are entailed in the sentences that lack them (13) and
are therefore not applicative arguments in (12).
13. a. He ate.
b. She danced.
While the definition of applicative arguments as given in (4) covers all of
the cases discussed so far, it is only a working definition. It rules out for instance
that the underlined argument in (14) is an applicative argument because it is in-
troduced by a preposition.

 
16   

14. Alex broke Ben's vase on Chris.


However, apart from the preposition, this argument displays similar behavior to
the affected experiencer applicative argument of German (1d). Consequently, I
provide similar analyses for (14) and (1d) in Chapter 3 even though (14) does
not include an applicative argument according to the definition in (4). Thus, ra-
ther than (4) being an absolute criterion for applicative argument status, I treat it
as a working definition.
Hole (2008) shows that following his definition of free arguments (3), there
is one type of German applicative arguments that I have not mentioned so far,
namely the judgment dative (15).
15. Die Treppe ist ihm zu steil.
the stairs is him.Dat too steep
'The stairs are too steep for him.'
Hole (2008) explains that there is no entailment with the judgment dative
because the sample sentence without the dative (16) could be uttered in a con-
text in which construction workers installed stairs in a house and realize after-
wards that the stairs are steeper than called for by the blue-print. If the workers
have no personal interest in the stairs, (16) would be a felicitous sentence to say.
(15) would be infelicitous because in this context there is no entailment that
there is an individual who judges or experiences the stairs as being too steep.
16. Die Treppe ist zu steil.
the stairs is too steep
'The stairs are too steep.'
no entailment: 'There is an x for who the stairs are too steep.'
(Hole 2008, 1.10e, my translation)
I disagree with Hole on the judgment datives. I assume (similarly to Abra-
ham (1973)) that the comparative element zu 'too' can license an additional overt
argument, which denotes the entity whose standards are used for the compari-
son. If it is not realized (16), some contextual (or common) standard is being
used for the comparison. Thus, (16) in the given context means that the stairs are
too steep for this particular house (to be functional). In a different context, (16)
could be used to indicate that the stairs are too steep for movers to carry a couch
upstairs or too steep for someone to walk them down safely. Thus, zu 'too' marks
that there is some judgment made about the situation and that the cause/reason
for this judgment may be overtly expressed but does not have to be. Consequent-
ly, I take the judgment dative to be selected and will not address it further.
Lastly, Pylkkänen (2002) includes locative and instrumental applicatives in
her description and analyses of applicative arguments. She provides the follow-
ing example of an instrumental applicative from Chichewa.
17. Chichewa instrumental


  17

Mavuto a-na-umb-ir-a mpeni mtsuko.


Mavuto SP-Past-mold-Appl-Asp knife waterpot
'Mavuto molded the waterpot with a knife.'
(Pylkkänen 2002, example 14)
In German (18) and English (translation), instrumentals are expressed as PP-
adjuncts.
18. Mark formte den Wasserbehälter mit einem Messer.
Mark molded the waterpot with a knife
'Mark molded the waterpot with a knife.'
It is not possible to express the instrument with the typical form of an applica-
tive argument, i.e. in German with an additional dative-marked NP (which is not
selected by a preposition).
19. *Mark formte einem Messer den Wasserbehälter
Mark molded a knife the waterpot
'Mark molded the waterpot with a knife.'
As a result, the Chichewa-type instrumental (and locative) applicatives men-
tioned by Pylkkänen will not be addressed here because they are not found in
German or English. However, future research should investigate these and other
applicatives that meet my definition (4) for their features and meaning contribu-
tion.
In summary, my definition of applicative arguments is that given in (4). This
definition will be used to argue that the elements under discussion are indeed
applicative arguments. Yet, it remains a working definition and some preposi-
tional phrases that resemble applicative argument will also be discussed.
1.2 Goals and Structure of this Book
There are three major goals of this work. First, I show that previous analyses
disregard the type of meaning contribution (at-issue/ not-at-issue) of some of the
applicatives found in German and English. However, this distinction is im-
portant because it is reflected in certain semantic and syntactic behaviors of the
applicatives. Second, I show that each type of applicative argument can be iden-
tified unambiguously. Applicatives are not as uniform of a category as analyses
such as Pylkkänen's classification of only high and low applicatives suggest. Ra-
ther applicative arguments are varied and each type requires its own, detailed
analysis. This ties in with the third goal which is to provide formal analyses for
each identified type of German and English applicative argument. Each type re-
ceives an analysis in the framework of event semantics which reflects the mean-
ing as well as the syntactic behavior. These formal analyses of applicatives can
be used as tools to identify applicative arguments in languages other than Ger-
man and English in the future.

 
18   

This book is organized as follows: in Chapter 2, I discuss the preliminaries


that are need to understand the analyses of the types of applicative arguments
found in English and German. In section 2.2, I explain the difference between
at-issue and not-at-issue meaning and how constituents with either type of
meaning behave syntactically. In section 2.3, I give a brief introduction to Event
Semantics which is used in my formal analyses as well as several previous ana-
lyses. Lastly, I summarize Pylkkänen's (2002, 2008) classification of applicative
arguments into high and low applicatives. Throughout the book, I frequently re-
late my analyses and observations to hers.
Chapters 3-7 are dedicated to the descriptions and analyses of the applica-
tive arguments of German and English as exemplified in (1). The full semantic
derivations for each type are presented in the corresponding appendices. In
Chapter 8, I discuss how the analyses of the applicatives in each language form
a consistent and coherent system by discussing the co-occurrence possibilities of
the applicative arguments. In the conclusion (Chapter 9), my overall findings
and open questions are presented.


2 Preliminaries
2.1 Introduction
In this chapter, I discuss some background assumptions and concepts that should
be kept in mind for the following chapters in which the applicative arguments
are analyzed. The first issue concerns the diagnostic tools to differentiate at-
issue and not-at-issue meaning. This is one of the dimensions that I use to show
that the various types of applicative arguments need different analyses. Here, I
introduce the tests used to distinguish these two tiers of meaning based on the
behavior of the elements on each. This also includes a discussion of the system
for at-issue and not-at-issue meaning proposed by Potts (2005). The second is-
sue discussed in this chapter concerns the semantic framework used throughout
this book, namely Neo-Davidsonian event semantics. I give a brief overview of
this approach in section 2.3. Lastly, in section 2.4, I summarize Pylkkänen's
(2002, 2008) work on differentiating applicative arguments into low and high
applicatives. This is the background to my analyses. It will be shown throughout
the book that Pylkkänen's analysis is too coarse to handle all details of the dif-
ferent types of applicative arguments. A reader familiar with the tests for (not-)
at-issue meaning, Neo-Davidsonian event semantics and Pylkkänen's work may
skip this chapter entirely.
2.2 Not-At-Issue Meaning
One major difference from many previous analyses of applicative arguments is
that I argue that in some applicative constructions, some or all of the meaning of
the applicative is not-at-issue ("implied") meaning. This means that in addition
to the asserted, truth-conditional proposition of the verbal event, there is mean-
ing contributed to a second tier of meaning. This second tier of meaning encodes
not-at-issue meaning. This idea of having two tiers of meaning has been argued
for by Karttunen and Peters (1979) as well as Potts (2005). Potts contends that
the not-at-issue content is independent of the asserted at-issue content that is
conveyed in a sentence. He shows that not-at-issue content is different from at-
issue material and presuppositions. According to him, Conventional Implica-
tures (CIs) contribute only to the not-at-issue tier of meaning. These CIs share
the following characteristics:
20. Features of Conventional Implicatures (CIs):
1. CIs are part of the conventional meaning of words;
2. CIs are commitments, and thus give rise to entailments;
3. These commitments are made by the speaker of the utterance 'by vir-
tue of the meaning of' the words he chooses;


20   

4. CIs are logically and compositionally independent of what is 'said (in


the favored sense)', i.e., independent of the at-issue entailments.
(Potts 2005, example 2.10)
The first feature refers to the fact that a specific lexical item carries the not-at-
issue meaning. Thus, it can be pinpointed which element of a sentence is re-
sponsible for the not-at-issue contribution. The second feature is that not-at-
issue meaning contributes to the backgrounded information in the discourse in a
similar way that at-issue meaning does: it is added to the common ground. Con-
sequently, it can be used in (and interact with) subsequent contributions to the
discourse. The third feature refers to the speaker-orientation of the CI content.
This means that the contribution of a CI always projects to the highest level and
reflects an attitude or comment made by the speaker. CIs cannot be embedded in
order to attribute their contribution to someone other than the speaker. The
fourth feature is the independence of at-issue and not-at-issue content. This is
the idea of two tiers of meaning: one tier for at-issue and one tier for not-at-issue
meaning. As shown below, some interaction between the tiers is possible but
neither tier depends on the other, and the contributions of each tier can be stated
individually.
Potts (2005) argues that appositives are an example of this kind of Conven-
tional Implicature. Their meaning is not relevant for the truth-condition of the
main assertion. Rather, they add a comment by the speaker. Thus, in (21) the
appositive the king of France is a comment made by the speaker.
21. We invited Louis, the king of France.
(Bosse et al. 2012, example 4a)
The appositive does not alter the truth conditions of the sentence We invited
Louis (assuming Louis refers to the same individual).
The family of sentence test4 can be used to detect non-asserted meaning. A
simple declarative sentence is put into different syntactic environments, such as
embedding under modals, interrogation, and negation, to see whether non-
asserted meaning is still projected. This is known as the projection problem in
semantics: non-asserted meaning that is present in a simple declarative sentence
might not be projected if that very sentence is operated on syntactically. For in-
stance, cleft sentences pass their presupposition on, even if they are embedded
under epistemic modals, whereas a simple declarative sentence does not. Con-
sider the following set of sentences:
22. a. Lauri has solved the projection problem. simple declarative
b. It is Lauri who has solved the projection problem. cleft


4 Roberts et al. (2009) cite Chierchia and McConnell-Ginet (1990) for this name.


  21

c. Someone has solved the projection problem.


d. Maybe Lauri has solved the projection principle. modal
e. Maybe it's Lauri who has solved the projection principle.
(Soames 1982, examples 5, 6a, 7a, 8, 9)
The cleft sentence (22b) presupposes the statement in (22c). This even holds
true if a modal is included in the cleft (22e). This contrasts with the behavior of
the sample declarative sentence (22a) which entails (22c) but does not entail it if
a modal is included (22d). The projection problem is concerned with meaning
that projects, when it projects and how it can be predicted whether any meaning
will project.
I use the following tests to identify not-at-issue meaning, exemplified here
with appositives (which contribute not-at-issue meaning (Potts 2005)). First,
not-at-issue meaning projects past yes/no-questions. This means that regardless
of the answer to the yes/no-question, the not-at-issue material is accepted by the
speaker and the listener, and it becomes part of the common ground. If the lis-
tener does not want to accept it, (s)he needs to explicitly say so. Thus, (23) pro-
jects the meaning that Louis is the king of France regardless of whether he was
invited or not.
23. Did you invite Louis, the king of France?
(Bosse et al. 2012, example 8)
If the answer to this question is just "no", it can only mean that Louis was not
invited. It cannot mean that Louis was invited but he is not the king of France. In
order to convey this, a more elaborate answer than "no" is needed. In other
words, by just answering "no", the listener accepts the truth that Louis is the
king of France.
A second characteristic of not-at-issue meaning is that it cannot be targeted
by negation. Negation can only negate at-issue information. In (24), negation
cannot negate just the information of the appositive the king of France in the
sense that Louis was invited but he is not the king of France.
24. We didn't invite Louis, the king of France.
(Bosse et al. 2012, example 6)
Third, not-at-issue meaning is not taken into account in the truth conditions
of a conditional clause. Thus, if an appositive is part of a conditional sentence,
its truth conditions are the same as for the sentence without the appositive. As-
suming Louis refers to the same individual in (25a) and (25b), the situations in
which I will have to pay are exactly the same. There is no situation which would
make only one sentence true but not the other.
25. a. If Louis actually does visit you, I'll give you a hundred dollars.

 
22   

b. If Louis, the very haughty king of France, actually does visit you, I'll
give you a hundred dollars.
(Bosse et al. 2012, example 10)
Next, meaning on the not-at-issue tier cannot be questioned using wh-
questions. This means that (elements of) appositives cannot undergo wh-
movement or even be part of multiple wh-questions (where the wh-word stays
in-situ).
26. a. *Which country did you invite Louis, the king of?
b. *Who invited Louis, the king of which country?
(Bosse et al. 2012, examples 4c, 5)
Finally, variable binding only works if the quantifier is on the at-issue tier,
binding a variable which is on the not-at-issue tier (and not vice versa). Thus,
binding a variable in an appositive is possible (27a), but it is not possible for a
quantifier in an appositive to bind a variable on the at-issue tier (27b).
27. a. Each girl1 tried to ditch Miss Marple, her1 chaperone.
b. *Miss Marple, each girl1's chaperone, berated her1.
(Bosse et al. 2012, examples footnote 5 i, 9b)
These different environments (yes/no-question, negation, conditional, wh-
question, and quantifier binding) are the tests that I use below to diagnose
whether an applicative argument has at-issue or not-at-issue meaning. In short,
not-at-issue elements are expected to project past yes/no-questions, to not be
negatable, to not be available for wh-questioning, to be irrelevant for the truth
conditions of a conditional clause, and to be unable to bind at-issue content. At-
issue elements are expected to project past yes/no-questions, to be negatable, to
be able to be wh-questioned, to be relevant for the truth-conditions of if-clauses,
and to be able to bind not-at-issue content.
Roberts et al. (2009) argue that a variety of not-at-issue meaning can be de-
tected by these tests (e.g. presuppositions, implicatures). Consequently, not all
applicatives that contribute not-at-issue meaning are CIs. I will address the sta-
tus (in Potts's system) for each not-at-issue applicative individually.
According to Potts (2005), it is not expected that an element can contribute
to the at-issue tier and the not-at-issue tier at the same time. However, McCrea-
dy (2009) shows that elements with this kind of mixed contribution exist. He
discusses the Japanese adverb sekkaku as such an element. The predictions for
these elements with respect to the family of sentences tests are not straightfor-
ward: it is not clear which tests an element with mixed contribution should pass
(i.e. the not-at-issue meaning is still present) and which ones it should fail (i.e.
the not-at-issue meaning does not project). I will come back to this in Chapter 3.


  23

2.3 Event Semantics


As a semantic framework, I use Neo-Davidsonian event semantics (Parsons
1990). In this approach, each verb is associated with an event variable (of se-
mantic type v5). The participants of the event receive thematic roles and are as-
sociated to the event via these thematic roles. In other words, the syntactic ar-
guments are analyzed as event properties (Kratzer 1996). This approach was de-
veloped mainly for the semantic treatment of verbal modification and is there-
fore well suited for applicative arguments (which often seem to increase a verb's
valency).
According to Parsons (1990), events can hold at a given time and can culmi-
nate at a specific (point in) time. In addition, events can be divided into subev-
ents. Events and subevents can be related to each other, e.g. they can overlap or
occur sequentially. Each event can be located in time and space and can be char-
acterized by the event participants (individuals and objects) (Eckardt 2002).
The whole verb phrase (below tense and modals) is argued to denote a set of
event(ualities) which has certain other properties. Each participant is represented
as a separate conjunct in the semantic derivation. For instance, (28a) is taken to
denote a set of events that can be described as kissing. Other properties of this
event are that Mary is the one being kissed (Theme) and John is doing the kiss-
ing (Agent), as shown in (28b). This set is associated with the characteristic
function in (28c). The logical expressions in (28b) and (28c) are defined in (29).
28. a. John kisses Mary.
b. {kiss(e) & Thm(e, Mary) & Agt(e, John)}
c. λe. KISS(e) & Thm(e)(Mary) & Agt(e)(John)
29. a. kiss → λe. e is a kissing event
b. Thm → λx.λe. x is the theme of e
c. Agt → λx.λe. x is the agent of e
The verb kiss requires an individual (λx) of the semantic type e to be used
grammatically. This individual will be interpreted as the theme of the event
(30a). The interpretation of the VP (30b) is achieved by functional application of
the verb kiss and the NP Mary, as in (30c).
30. a. [[kiss]] = λx. λe. KISS(e) & Thm(e)(x)


5 The basic types are: e (individuals), t (truth-values), v (events). Sometimes s is alterna-
tively used for events. I do not consistently include the types (as subscripts) in the syn-
tactic trees and semantic derivations but only as necessary.

 
24   

b. VP<v,t>
3
V<e,vt> NP<e>
kiss 5
Mary
c. [[VP]] = [[kiss]] ([[Mary]])
= λe. KISS(e) & Thm(e)(Mary)
The NP Mary that the verb combines with is of type e and can thus become the
theme. The verb is then satisfied and does not require any other element. By
combining the verb and the NP, the VP denotes a kissing event with Mary as the
theme (30c). (The event variable λe will later be closed.)
As Kratzer (1996) argues, the external argument is not an argument of the
verb. Rather, it is introduced by a syntactic head, Voice, as defined in (31a). It is
combined with the VP by event identification (31b): event identification com-
bines two functions (each with an event variable) into a single function (with
one event variable). (31c-d) show the syntactic structure of the whole VoiceP
p
and its semantic interpretation.
31. a. [[Voice]]]] = λx. λe. Agt(e)(x)
b. Event Identification: <e,<v,t>> <v,t> → <e,<v,t>>
c. VoiceP<v,t>
3
NPe Voice'<e,vt>
5 3
John Voice<e,vt> VP<v,t>
6
kiss Mary
d. [[VoiceP]] = λe. KISS(e) & Thm(e)(Mary) & Agt(e)(John)6
The open event variable (λe) is eventually existentially bound (with scope as
narrow as possible) unless there is another source of event quantification.
Modifiers can be added to the sentence and correspond to separate conjuncts
in the semantic derivation. This is exemplified in (32).
32. a. John kissed Mary in the park.
b. λe. KISS(e) & Thm(e)(Mary) & Agt(e)(John) & Location(e)(park)
Each event can have each thematic role assigned only once. Thus, it is not
possible to have an event with, for instance, the agent role assigned twice (Par-
sons 1990).7

6 An alternative, yet equivalent, way of writing this is:
[[VoiceP]] = λe. KISS(e) & Thm(e, Mary) & Agt(e, John)
7 However, it is possible for one referent to have two (or more) thematic roles associated


  25

One verb can involve one or more events, and each event can have sub-
events. Modifiers like again can target these subevents. For instance, again can
receive a restitutive reading when it modifies the transition from a process to a
state (33).
33. (Context: John has bought brand new glasses that have never been
dirty. Then he dropped them in some mud.)
John cleaned his glasses again.
. . . his glasses were clean before.
Again can receive a repetitive reading when it modifies a state or process (34).
34. (Context: John has had glasses for years. They get dirty every day.)
John cleaned his glasses again.
. . . John has cleaned his glasses before.
Causatives (derived from intransitive verbs) involve two events. The follow-
ing example illustrates this. (Cul stands for Culmination.)
35. a. Mary flew the kite.
b. (e) [Agent(e, Mary) & Cul(e) & (e') [Flying(e') & Cul(e') &
Theme(e', kite) & CAUSE(e, e')]]
(Parsons 1990, p. 109)
The derivation states that there is an event (which cumulates). Mary is the agent
of this event, and this event is the cause for a second event, which is a flying
event. The kite is the theme of this flying event.
Modifiers can apply to any event (36). The blank in the derivation in (36b)
can either be filled with e or e'. In the latter case, the resulting meaning is that
the flying takes place behind the museum; in the former case, Mary does some-
thing behind the museum (which causes the kite to fly).
36. a. Mary flew the kite behind the museum.
b. (e) [Agent(e,Mary) & Cul(e) & (e') [Flying(e') & Cul(e') &
Theme(e', kite) & Behind(_, museum) & CAUSE(e, e')]]
(Parsons 1990, p. 118)
While Neo-Davidsonian event semantics was developed as a purely seman-
tic framework, it has consequences for my syntactic analyses. First, the thematic
roles that the applicative arguments receive are often introduced by phonologi-
cally null (functional) heads. Second, the surface structure of the sentences will
not be addressed here (unless where necessary). As a matter of fact, I abstract
away from all higher projections (including tense) for the sake of simplicity (un-
less where necessary). This means that the provided syntactic structures are in-
complete and the semantic analyses do not include time intervals and cumula-
tion points. For German, this abstraction from surface structure is also advanta-

with it in one event according to Parsons.

 
26   

geous because it allows me to disregard the order of the constituents, which de-
pends on many constraints. Vogel and Steinbach (1998) cite Lenerz (1977) (and
Cooper (1994) for the translation) for the following ordering constraints that ap-
ply to German:
37. German Constituent Order
"a. Theme/Rheme Condition: the theme tends to precede the rheme; b. Definiteness
condition: definite tends to precede indefinite; c. Law of Growing Constituents (Ge-
setz der wachsenden Glieder, following Behagel (1932)): heavier constituents tend
to follow lighter ones; d. Sentence Bracket Condition: the tendency, not to end a
sentence on a light constituent if the sentence bracket is open, i.e. if the clause does
not end with a verb; e. Subject/Agent condition: subject/agent tends to precede other
constituents." (Vogel and Steinbach 1998, p. 68)
As this quote shows, the surface order in German is influenced by many differ-
ent factors and abstracting from lets me focus solely on the applicative argu-
ments and their analyses.
2.4 Pylkkänen (2002, 2008)
Pylkkänen (2002; 2008) claims that there are two basic types of applicative ar-
guments: a high applicative which is introduced by a head attaching outside of
VP and relating an individual to an event as well as a low applicative argument
which is introduced by a head attaching below VP and relating two individuals
involved in a transfer of possession, either to or from someone. The high appli-
catives are responsible for adding an additional participant to the event whereas
low applicatives are only responsible for the transfer of possession, independent
of the event denoted by the verb. Pylkkänen gives the following examples for
each type.
38. a. High Applicative (in Chaga)
N- ã- ï- lyí- ì- á m- ká k-èlà
Foc- 1sg- Pres- eat- Appl- FV 1- wife 7-food
'He is eating food for his wife.'8
(Pylkkänen 2002, example 2a)


8 In the original source Bresnan and Moshi (1990, example 2), the translation is given as
'... for/on his wife.' In my categorization, the preposition for indicates a benefactive ap-
plicative while on is used for affected experiencers. The fact that Pylkkänen glosses
over this can be taken as a first indication that her classification is too coarse.


  27

VoiceP
3
he 3
Voice 3
wife 3
ApplBen 3
eat food
(Pylkkänen 2002, example 16a)
b. Low Applicative: 'I baked him a cake.'
VoiceP
3
I 3
Voice 3
bake 3
him 3
Appl cake
(Pylkkänen 2002, example 16b)
As shown in (39), the high applicative covers different interpretations for the
applicative argument, e.g. benefactive, malefactive. It simply adds the corre-
sponding thematic role to the semantic derivation as a conjunct.
39. High Applicative
λx. λe. APPL(e,x)
(collapsing APPLBEN, APPLINST, APPLLOC and so forth)
(Pylkkänen 2002, example 23)
In contrast, the low applicative head takes the two NPs and then the verb as
its arguments. By doing so, it can relate the two individuals involved in the
transfer of possession to each other before relating the transfer to the verbal
event. Pylkkänen distinguishes two types of low applicatives depending on the
direction of the transfer (to/from).9 Their semantic denotation is given in (40).
40. Low Applicative
Low-APPL-TO (Recipient applicative):
λx. λy. λf<e<s,t>>. λe. f(e,x) & to-the-possession(x,y)
Low-APPL-FROM (Source applicative):
λx. λy. λf<e<s,t>>. λe. f(e,x) & from-the-possession(x,y)
(Pylkkänen 2002, example 25)
As I show throughout this book, this distinction between high and low ap-
plicatives is not sufficient to account for the types of applicative arguments that
are found within a language or across languages. What Pylkkänen (2002) de-

9 Larson (2010) points out several problems with these denotations.

 
28   

fines as high applicatives is not a uniform category. The examples in (1), with
the exception of (1c), all pass Pylkkänen's tests for high applicatives. First, they
relate an individual (or entity) to an event. In (1a), the car is added to the event
of denting the door. In (1b), he is added as a beneficiary of the cake baking. In
(1d), Chris is added as a psychologically affected individual. In (1e), the speaker
includes himself/herself and his/her interest in the situation. In (1f), the applica-
tive argument indicates that this is a special or remarkable event. Finally, in
(1g), the grandma is responsible for the broken vase (although she did not ac-
tively break it). In none of these examples is there a transfer of possession. Re-
gardless, these types are as different from each other as they are from low ap-
plicatives and must consequently be distinguished.
Pylkkänen (2002) shows that high applicative arguments can also be differ-
entiated from low applicatives by the type of verbs they can occur with. High
applicatives attach outside of VP and should be allowed to appear with unerga-
tive and stative verbs. In contrast, low applicatives require the presence of an
object NP and therefore cannot occur with unergative and stative verbs. This is
not a helpful diagnostic for German because most applicatives that relate an in-
dividual to an event in German are licensed only by certain verbs (or VPs) and
thus are not as unselective as predicted by Pylkkänen.10 The occurrence with
unergative and stative verbs is not a reliable diagnostic of types for German ap-
plicative arguments, as shown throughout this book.
As for the semantics proposed by Pylkkänen (2002), I show that the high
applicative is only suitable for two types of applicatives found in German,
namely the true benefactive applicative arguments (Chapter 5) and the Dative of
Inaction (Chapter 7). All other applicatives require analyses different from Pyl-
kkänen's proposal.


10 Wegener (1983) consequently argues that many applicatives of German are not truly
free arguments but rather verb-specific arguments. I disagree that this makes the argu-
ments verb-specific. Rather, I argue that some applicative heads select only certain VPs,
for instance resultative VPs. This does not make the applicatives arguments of the verb.


3 Affected Experiencers
3.1 Introduction
In this chapter, I discuss affected experiencer applicative arguments in German
and English, as exemplified in (41).11
41. a. Alex zerbrach Chris Bens Vase.
Alex.Nom broke Chris.Dat Ben.Gen vase.Acc
'Alex broke Ben's vase on Chris.'
= Alex broke Ben's vase, and this matters to Chris.
(Bosse et al. 2012, example 1a)
b. Gerald broke the vase on me.
Affected experiencer applicative arguments denote an individual who is not di-
rectly involved in the event, but the event affects the individual. In the German
example (41a), it is Chris who is affected by Alex breaking Ben's vase while in
the English example (41b) I am affected by Gerald breaking the vase.
This English example might be surprising because the affected experiencer
applicative argument is included in a PP (and consequently does not meet my
definition of an applicative argument (4)). Due to the striking similarities be-
tween German and English affected experiencers (explained below), I analyze
them similarly despite the English affected experiencers being PPs. I argue be-
low that the preposition on is the overt applicative head for affected experiencers
in English.
Before I analyze the affected experiencer arguments in detail in section 3.4,
note that these affected experiencer applicative arguments should be classified
as high applicatives because an individual is related to an event (Pylkkänen
2002). Being high applicatives, first it is expected that they can occur with uner-
gative and stative verbs. As mentioned in Chapter 2, this is not a good diagnostic
for German because of other verbal restrictions (see also section 3.7). As shown
in (42), unergative and stative verbs are allowed in English with the affected ex-
periencer.
42. a. John cried on Mary.
b. He held on to the bag on Mary.
Second, high applicatives do not denote a possessive relation. This is also true
for affected experiencers. In fact, it is possible to have overt possessors in addi-
tion to the affected experiencer (41) (Isacenko 1965). Thus, affected experiencer
applicative arguments should receive the same analysis as other high applica-

11 This analysis of affected experiencers presented in this chapter is our analysis as given
in Bosse, Bruening and Yamada (2012).


30   

tives according to Pylkkänen (2002, 2008). As I show below, this does not do
them justice, in part because they involve not-at-issue meaning (see Chapter 2
and section 3.3), which is not included in Pylkkänen's analysis.
I distinguish the affected experiencers from other applicative arguments in
section 3.2. The German pertinence dative is addressed in section 3.2.2 an 3.6.
This applicative argument has often been argued to constitute its own type of
applicative. I show that it is identical to the affected experiencer construction but
includes an additional conversational implicature. Therefore, it does not require
a different formal analysis than the affected experiencer.
Following the discussion of the pertinence dative, I address the verbal re-
strictions on affected experiencers and pertinence datives (section 3.7). Semantic
and valency restrictions are used to characterize the verbs that license the affect-
ed experiencer construction. I take a look at affected experiencer applicative ar-
guments as elements in Potts' system of not-at-issue elements in section 3.8.
Lastly, I digress from the German and English applicative arguments in section
3.9 where I investigate some cross-linguistic parametric variation with respect to
affected experiencers. In particular, I show how the analysis of German and
English affected experiencers can be extended to Japanese, Hebrew and French.
I conclude this chapter in section 3.10.
3.2 Distinguishing Affected Experiencers
In this section, I distinguish affected experiencers (41) from applicatives dis-
cussed in other chapters. Differentiating English affected experiencers from oth-
er English applicatives is straightforward, as the former are part of a on-PP,
whereas the latter are not:
43. a. Gerald broke the vase on me. (affected experiencer)
b. John baked Mary a cake. (recipient benefactive)
c. John1 killed him1 a bear. (subject co-referential)
Furthermore, the affected experiencer construction can be distinguished
from other English on-PPs (e.g. locatives) based on the contributed meaning,
namely a psychological affectedness. This goes along with a sentience require-
ment for the affected experiencer (44).
44. Gerald broke the vase on his dead grandfather.
The non-sentient individual denoted by the argument his dead grandfather in
(44) cannot be interpreted as an affected experiencer but rather must be under-
stood as a location or as a personification of the dead grandfather, for instance
the grandfather's ghost who could be psychologically affected. Other uses of on-
PPs, for instance the locative one, do not have this requirement (45).
45. John broke the vase on the beach.


  31

The affected experiencer cannot refer to entities that are not sentient because its
semantic contribution is that of a psychological experience. Thus, English af-
fected experiencers can easily be identified because of the required preposition
on marking psychological affectedness.
At first glance, German affected experiencer applicative arguments (41a)
can be mistaken for other types of applicative arguments. Many sentences are
ambiguous between the different types (see among others Rosengren 1975, Hole
2008, Colleman 2010). However, they need to be distinguished and can be, as
shown in the following subsections.

3.2.1 Benefactives (Chapter 5)


German affected experiencer applicative arguments and benefactive applicative
arguments can be distinguished based on three characteristics (Bosse and
Bruening 2011). First, affected experiencers are psychologically affected by the
event, unlike beneficiaries, and they do not (necessarily) benefit from it. Thus,
(41a) can be used in the context of Chris being happy about the broken vase (as
he never liked it) or in the context of Chris being sad about it (as he was collect-
ing vases) (cf. Hens 1997). In contrast, true benefactive constructions require
some benefit for the referent. Example (46) can only be understood as there be-
ing some benefit for the (dead) father.
46. German true benefactive
Dennis malte das Bild seinem toten Vater.
Dennis.Nom painted the.Acc picture his.Dat dead father
'Dennis painted the picture for his dead father.'
(Bosse and Bruening 2011, example 6a)
This benefit is, for instance, a dedication or that Dennis painted a picture and put
it on his father's grave. The affected experiencer interpretation is ruled out be-
cause of the sentience requirement for the referent of affected experiencers. The
referents must be sentient as they have to be able to be psychologically affected.
Consequently, (46) cannot be an affected experiencer construction. Similarly,
(47b) is infelicitous because the referent of the affected experiencer (Paul) is
already dead and cannot be affected.
47. German Affected Experiencer (Hole 2005, example 14)
Context: Paul died first.
a. Dann starb auch seine Mutter.
then died also his mother
'Then his mother died, too.'
b. #Dann starb ihm auch seine Mutter.
then died him.Dat also his mother
'Then his mother died on him, too.'

 
32   

This shows that affected experiencer applicative arguments have to refer to sen-
tient beings. This does not hold for benefactives. In the benefactive example
(46), the father is the beneficiary even though he has already died. These exam-
ples show that the referent of the affected experiencer must be able to perceive
the event (and therefore be in a sentient state of mind) whereas the referent of
the benefactive applicative argument does not necessarily have to (be able to)
perceive the event.
The third difference is that, unlike benefactives, German affected experi-
encer applicative arguments do not alternate with a prepositional variant (see
also Chapter 5). (48a) is largely synonymous with (46), whereas no grammatical
example of a prepositional affected experiencer (48b) exists.
48. a. Dennis malte das Bild
Dennis.Nom painted the.Acc picture
für seinen toten Vater.
for his.Acc dead father
'Dennis painted the picture for his dead father.'
b. *Lisa lobte gegen/ auf/ wider
Lisa.Nom praised against/ on/ against
ihren Mann den Anzug.
her.Acc man the.Acc suit
(Bosse and Bruening 2011, example 8)
These three characteristics (meaning, sentience, PP-variant) distinguish af-
fected experiencers from benefactives in German (in addition to the benefactive
contributing only at-issue meaning whereas the affected experiencer contributes
some not-at-issue meaning (see section 3.3)).

3.2.2 Part-Whole Applicative Arguments (Chapter 6) &


Pertinence Datives
The part-whole applicative argument (49a) differs from the affected experiencer
applicative argument (49b) most clearly in its meaning; the latter denotes that
the applicative argument is psychologically affected whereas the former encodes
the whole of a part-whole relationship.
49. a. Jan zerbeulte dem Auto die Tür.
Jan.Nom dented the.Dat car the.Acc door
'Jan dented the door of the car.'
b. Alex zerbrach Chris Bens Vase.
Alex.Nom broke Chris.Dat Ben.Gen vase
'Alex broke Ben's vase on Chris.'


  33

Thus, in (49a) the door is interpreted as being a part of the car and not as being
(psychologically) affected. In contrast, in (49b), Chris is psychologically af-
fected and Ben's vase is not a (material) part of him. This should make differen-
tiating the two types easy. However, a complication is the existence of the so-
called pertinence dative (see also section 3.6). The pertinence dative shares
characteristics of both the part-whole and the affected experiencer constructions.
Some examples of the pertinence dative are provided in (50).
50. a. Der Stein fällt dem Mann auf den Kopf.
the.Nom brick falls the.Dat man on the.Acc head
(Neumann 1995, example 1)
b. Mir schmerzt der Rücken.
me.Dat hurts the.Nom back
'My back hurts.'
(Abraham 1973, example 26, my translation)
c. Der Regen tropft mir auf die Schuhe.
the.Nom rain drips me.Dat on the.Acc shoes
'The rain is dripping on my shoes.'
(von Polenz 1969, example 23, my translation)
The dative constituent in each of these examples is said to encode that the de-
noted individual is the possessor of the body part/ piece of clothing mentioned in
the sentence and is also affected by the event. In other words, the pertinence da-
tive involves affectedness (like the affected experiencer) as well as a part-whole/
possessive relationship to another entity (like the part-whole applicative argu-
ment).12
Typically, the German pertinence dative (50) has been described as having
the characteristics given in (51) (according to von Polenz (1969), unless other-
wise indicated).
51. Characteristics of the Pertinence Dative
a. The applicative argument is not selected by the verb.
b. The sentences are transformable into possessive/genitive construc-
tions (though not with exactly the same meaning (Hole 2008)). Fur-
thermore, the reverse does not hold: not all possessive/ genitive con-
structions can alternatively be expressed as a pertinence dative (Isacen-
ko 1965, Wegener 1983).
c. The pertinence construction is permissible with different types of
verbs.
d. The referent of the pertinence dative must be animate.

12 Lamiroy and Delbecque (1998) argue that the interpretation oscillates between a posses-
sor interpretation and an affected interpretation.

 
34   

e. The referent of the pertinence dative must be participating/ involved


in the event. Some researchers mention a highlighting or focusing of
the referent.
f. There must be a possessive or part-whole relationship between the
applicative argument and another NP.
g. The possessed NP is typically a piece of clothing (Neumann 1995)
or an inalienable part (Bierwisch 1966).
All of these requirements are met in (50a). The verb fallen 'fall' does not select a
dative-marked argument. The sentence can alternatively be expressed as a geni-
tive construction (see below for the meaning difference between (50a) and (52)).
52. Der Stein fällt auf den Kopf des Mannes.
the.Nom brick falls on the.Acc head the.Gen man.Gen
'The brick falls onto the man's head.'
The referent of the applicative argument in (50a) is animate (and sentient) and
involved in the event (as he gets hit by the stone). There is a possessive/ part-
whole relationship between the man and the head; namely, it is the man's head.
The head is an inalienable part of the man.
However, not all of these characteristics are met by the part-whole applica-
tive argument construction (49a). Specifically, the referent of the applicative
argument dem Auto ‘the car’ is not animate. Thus, according to these features
the part-whole applicative construction and the pertinence dative can be argued
to differ in the animacy of the referent. There is no animacy requirement for the
part-whole applicative argument (Chapter 6).
The animacy requirement of the pertinence dative is tied to the interpretation
of the referent as being affected. This affectedness can be observed in the mean-
ing change between the pertinence dative (50a) and the alternating genitive con-
struction (52). The genitive sentence has been argued to lose some of the mean-
ing of the pertinence dative construction, namely a vague affectedness of the
referent by the event (von Polenz 1969, Neumann 1995, Lamiroy and Delbecque
1998, Hole 2008). Thus, in (50a) the man is affected by the event whereas in
(52) this is not (necessarily) the case. In a way, (52) is more of an objective de-
scription of the event (Lamiroy and Delbecque 1998, Neumann 1995).
Isacenko (1965) points out that the pertinence dative might be the same as
the affected experiencer. He discusses the fact that even though the possessive
interpretation is salient for pertinence datives, it is not plausible in all cases. In
particular, when another possessor is added the function of the dative seemingly
changes to that of an affected experiencer ("dativus commodi/incommode" for
him). Compare (53a) and (53b).


  35

53. a. Der Vater ist uns gestorben.


the.Nom father.Nom is us.Dat died
'Our father died.'
b. Peters Vater ist uns gestorben.
Peter.Gen father.Nom is us.Dat died
'Peter's father died on us.'
(Isacenko 1965, example 34, my translation)
The applicative argument uns 'us' is interpreted as a pertinence dative in (53a)
but as an affected experiencer in (53b) which includes the overt possessor Peters
'Peter's'. Instead of arguing for two different types of applicative arguments,
namely one with an overt possessor and one with a 'dative possessor', I argue
that the pertinence dative is a subtype of the affected experiencer applicative
argument (section 3.6).13 Here it is important that, in general, part-whole applica-
tive arguments can be distinguished from affected experiencers/ pertinence da-
tives.
Part-whole applicative arguments are possible with inanimates (49a),
whereas affected experiencers require a sentient referent (47). Animate referents
of the part-whole applicative argument tend to be interpreted as pertinence da-
tives (in the absence of another possessor), although in general, animate refer-
ents of the applicatives should be ambiguous between affected (pertinence) and
part-whole applicatives. I tease the two interpretations of animate applicative
arguments apart in section 3.3 where it is shown that affected experiencers con-
tribute some not-at-issue meaning. Part-whole applicative arguments do not con-
tribute any not-at-issue meaning (sections 3.6, 6.3).

3.2.3 Other German Applicatives


In the previous sections, I have shown that German affected experiencers can be
differentiated from benefactives and from part-whole applicative constructions.
Differentiating them from ethical datives (Chapter 4) can easily be done, as the
ethical dative must be a first person pronoun used in a sentence with directive
force. As the examples in this chapter demonstrate, this requirement does not
hold for the affected experiencer.
Affected experiencers can also easily be distinguished from subject co-
referential applicative arguments (Chapter 4) because affected experiencers do
not have to co-refer with the subject (as can be seen from the examples above).
Furthermore, the contribution of the subject co-referential applicative argument


13 Isacenko (1965) does not follow through on his discussion of this observation but rather
leaves it to future research.

 
36   

is only to the not-at-issue tier of meaning. The affected experiencer contributes


some at-issue meaning (as shown below).
Finally, the affected experiencer can be distinguished from the Dative of In-
action (Chapter 7) based on the type of meaning each contributes. Affected ex-
periencer applicatives contribute both at-issue and not-at-issue meaning whereas
DIs contribute only at-issue meaning. Furthermore, affected experiencers are not
limited to non-agentive (intransitive) verbs as DIs are. Thus, affected experi-
encers can be unambiguously identified in German.
3.3 At-Issue and Not-At-Issue Meaning
In this section, I show that affected experiencers in English and German involve
both at-issue and not-at-issue meaning. Using the tests described in Chapter 2, it
will become clear that this type of applicative argument contributes to the at-
issue as well as the not-at-issue tier, although Potts (2005) claimed that this
should not be possible. This issue is addressed in detail in section 3.8.
First, I discuss the "typical" affected experiencer, i.e. sentences like (41a). In
section 3.3.2, I address the pertinence dative (50). It is shown there that it be-
haves akin to the affected experiencer construction.

3.3.1 Affected Experiencers


Affected experiencers contribute some not-at-issue meaning. This can be seen
by negating sentences with affected experiencers. Negation cannot target the
notion of affectedness. Only the whole event can be negated, as (54) shows for
English and German.
54. a. Alex zerbrach Chris Bens Vase nicht.
Alex.Nom broke Chris.Dat Ben.Gen vase Neg
'Alex didn't break Ben's vase on Chris.'
1. Alex didn't break Ben's vase (but if he had, it would have mat-
tered to Chris.)
2. *Alex broke Ben's vase, but it didn't matter to Chris.
(Bosse et al. 2012, example 30a)
b. The car didn't break down on me.
1. The car didn't break down (but if it had, it would have mattered).
2. *The car broke down but it didn't matter to me.
The fact that the affectedness cannot be targeted by negation indicates that it is
encoded on the not-at-issue tier of meaning.
Similarly, yes/no-questions show this not-at-issue behavior of the affected-
ness. The affectedness of the affected experiencer projects past the question, as
shown in (55).


  37

55. a. Zerbrach Alex Chris Bens Vase?


broke Alex.Nom Chris.Dat Ben.Gen vase
'Did Alex break Ben's vase on Chris?' (If Alex broke it, it would
matter to Chris.)
(Bosse et al. 2012, example 31a)
b. Did the car break down on you? (If it broke down, it would matter to
you.)
These questions do not just ask whether the event took place or whether the
event mattered. Rather, the meaning that the event mattered to the individual
denoted by the affected experiencer applicative argument is already present in
the question. Furthermore, by answering this question using only "no", the an-
swering person still accepts that the individual denoted by the affected argument
would have been affected if the event had happened. A negative answer cannot
be used to indicate that the event took place but did not matter to the affected
experiencer. For this, a more elaborate answer is necessary. Thus, the yes/no-
question provides support for the affectedness being encoded on the not-at-issue
tier of meaning.
However, there is also evidence that the affected experiencer applicative ar-
gument contributes meaning to the at-issue tier. The experiencer is syntactically
accessible. It can be a wh-phrase and undergo overt movement (56). This should
not be possible if the meaning it contributed was only on the not-at-issue tier
(Chapter 2).
56. a. Wem hat Alex Bens Vase zerbrochen?
who.Dat has Alex.Nom Ben.Gen vase brokem
'On whom did Alex break Bens vase?'
(Bosse et al. 2012, example 32a)
b. Who did the car break down on?
Similarly, the affected experiencer contributes to if-clauses. This is further
evidence that there is meaning contributed to the at-issue tier.
57. a. Wenn Lisa ihrem Mann den
if Lisa.Nom her.Dat husband the.Acc
Anzug lobt, dann bekomm Jan 100Euro
suit praises then get Jan.Nom 100Euro
von ihm.
from him.Dat
'If Lisa praises the suit on her husband, then Jan will get 100 Euros
from him.'
(Bosse et al. 2012, example 37)

 
38   

b. i. If the CEO has a breakdown on us in the next two weeks, we'll all
quit.
ii. If the CEO has a breakdown in the next two weeks, we'll all quit.
Having the affected experiencer in the if-clause influences the conditions of the
clause. In (57a), Jan will only receive the money if the husband is affected by
the praising. If the affected experiencer argument were not included in the
clause, he would get the money as long as Lisa praised the suit (regardless of an
effect on the husband). Similarly, in (57b-ii) any breakdown of the CEO will
make us quit, whereas in (57b-i) the breakdown must have an effect on us to
quit. The if-clause test thus shows that the affected experiencer contributes at-
issue meaning.
Furthermore, the affected experiencer argument can be a quantificational
expression which can bind a variable that carries at-issue meaning (58).
58. a. Ich habe jedem Jungen1 seine1 Vase zerbrochen.
I.Nom have every.Dat boy his.Acc vase broken
'I broke his1 vase on every boy1.'
(Bosse et al. 2012, example 33)
b. I passed out on every one of my friends1 on his1 birthday.
As explained in Chapter 2, this binding pattern is also indicative of the element
binding the pronoun being on the at-issue tier of meaning.
In summary, these tests support the contention that the affected experiencer
construction involves both at-issue and not-at-issue meaning in both German
and English.

3.3.2 Pertinence Dative


In this section, I discuss the tiers of meaning that the pertinence dative contri-
butes to. As explained in section 3.2.2, the sentences with the pertinence dative
are ambiguous with the part-whole applicative argument (for speakers that allow
the part-whole construction). Due to this, I compare the pertinence interpretation
to the pure part-whole interpretation. Details on the part-whole applicative can
be found in Chapter 6.
I show two possible interpretations of the pertinence dative next to each oth-
er below, namely "possessor and affected" as well as a "(pure) part-whole" in-
terpretation. I use a sample sentence of the pertinence dative given by Neumann
(1995, example 1) but provide the two possible readings (59). The pure part-
whole reading is not salient without context for most native speakers but can be
made available by providing a context. For instance, example (59) can be inter-
preted as a part-whole applicative argument more readily if it is assumed that
instead of bricks, small stones are dropped onto the man which he may not no-
tice. This yields the part-whole interpretation. Alternatively, this interpretation


  39

can be brought about by assuming that it is a dead man's body onto which bricks
are dropped. In that case, the part-whole interpretation is available (while the
pertinence dative/ affected experiencer interpretation is not possible because the
referent is not sentient).
59. Der Stein fällt dem Mann auf den Kopf.
the.Nom brick falls the.Dat man on the head
1. The brick is falling on the man1's head on him1. (possessor & affect-
ed)
2. The brick is falling on the man's head. (part-whole)
Applying the not-at-issue tests (Chapter 2) to pertinence datives shows that
on the "possessor and affected" interpretation, the pertinence dative behaves like
affected experiencers. In contrast, on the (pure) part-whole interpretation, it be-
haves like a part-whole applicative argument.
The pertinence dative can be negated (60).
60. Der Stein fiel dem Mann nicht auf den Kopf.
the.Nom brick fell the.Dat man Neg on the.Acc head
1. The brick did not fall onto the (man's) head, but if it had, if would
have mattered.
*The brick fell onto the (man's) head but it didn't matter.
(possessor & affected)
2. The brick did not fall onto the man's head. (part-whole)
The negation test shows that on the "possessor and affected" interpretation the
affected meaning survives under negation. It is only the event that can be negat-
ed and not the affectedness itself. This is just like the affected meaning in a "typ-
ical" affected experiencer construction (54a). In contrast, on the pure part-whole
interpretation, negation indicates that the event did not happen. No meaning is
present that cannot be the target of negation.
Yes/no-questions also show that the pertinence dative construction behaves
just like the affected experiencer construction (55a) on the "possessor and af-
fected" interpretation, namely by contributing some not-at-issue meaning. On
the pure part-whole interpretation, it behaves like a part-whole applicative ar-
gument, contributing only at-issue meaning.
61. Fiel dem Mann der Stein auf den Kopf?
fell the.Dat man the.Nom brick on the.Acc head
1. Did the brick fall onto the (man's) head on the man?
(possessor & affected)
2. Did the brick fall onto the man's head? (part-whole)
The yes/no-question can be answered with "no". On the "possessor and affected"
interpretation, it means that the brick did not fall on the man's head but if it had,
it would have mattered to the man. It cannot mean that the brick actually hit the

 
40   

man in the head, but he was not affected. To get this meaning, a longer explana-
tion is needed. In contrast, on the part-whole interpretation the negative answer
would indicate that the event did not happen (either because the brick missed or
because the person was actually a woman or because it hit the man's shoulder).
Thus, the "possessor and affected" interpretation of the pertinence dative be-
haves just like the (typical) affected experiencer (55a) and the part-whole inter-
pretation like the part-whole applicative argument.
This is also true for if-clauses and wh-questions. In these constructions, on
both interpretations the pertinence dative contributes at-issue meaning.
62. Falls der Stein dem Mann auf
if the.Nom brick the.Dat man on
den Kopf fällt, ist es deine Schuld.
the.Acc head falls is it your fault
1. If the brick falls onto the (man's) head on him, it's your fault. (pos-
sessor & affected)
2. If the brick falls onto the man's head, it's your fault. (part-whole)
As (62) shows, the applicative argument contributes to the conditions of the if-
clause on both readings. In the "possessor and affected" case, the man must be
hit in the head by the brick and be affected by that, for it to be your fault. In the
part-whole case, the stone must actually hit the man's head. Again, on both in-
terpretations the pertinence dative behaves like the corresponding applicative
arguments.
Also, on both readings it is possible to ask a wh-question about the applica-
tive argument, showing again that at-issue meaning is contributed.
63. Wem ist der Stein auf den Kopf gefallen?
who.Dat is the.Nom brick on the head fallen
1. Whose head did the brick fall on (and affected that person)? (posses-
sor & affected)
2. Whose head did the brick fall on? (part-whole)
Thus, the pertinence dative has two interpretations (for speakers that allow
the part-whole applicative): that of the part-whole applicative and that of the af-
fected experiencer. Each interpretation behaves like the corresponding applica-
tive argument. In other words, the pertinence dative with the "possessor and af-
fected" interpretation is an affected experiencer. Its possessive interpretation is
addressed in section 3.6.
3.4 Analysis
As discussed in the previous section, affected experiencer applicative arguments
involve both at-issue and not-at-issue meaning. In that, the analysis given here
differs from all previous analyses (e.g. Abraham 1973, Neumann 1995, Pyl-


  41

kkänen 2002, Hole 200814). Before discussing the formal aspects of the analysis,
I determine the syntactic position of the affected experiencer.
In both German (64) and English (65), the subject c-commands the affected
experiencer and can consequently bind into the affected experiencer argument,
whereas the affected experiencer argument cannot bind into the subject. This
indicates that the affected experiencer attaches below the subject.
64. German
a. Alle1 haben [ihren1 Vorgesetzten]Aff den
everyone.Nom have their.Dat superiors the
Dienst quittiert.
service quit
'Everyone1 has quit the service on their1 superiors.'
b. *[Ihre1 Arbeiter] haben allen1 den
their.Nom workers have everyone.Dat the.Acc
Dienst quittiert.
service quit
'Their1workers have quit the service on everyone1.'
65. English
a. Every company1 downsized on its1workers.
b. *Its1 workers quit their jobs on every company1.
Furthermore, the affected experiencer can cause a Condition C violation
with respect to the subject (66). This again shows that the affected experiencer
should attach below the subject in the syntactic derivation.
66. *Er1 zerbrach Martins1 Bruder die Vase.
he.Nom broke Martin.Gen brother.Dat the vase
'He1 broke the vase on Martin's1 brother.'
67. *He1 broke the vase on Gerald's1 brother.
These examples indicate that in both German and English the affected experi-
encer is c-commanded by the subject.
Due to the obligatory preposition introducing the affected experiencer in
English, German and English require slightly different analyses. Both analyses
are presented in the following.
For German, I argue that the affected experiencer is introduced by the syn-
tactic head Aff(ect). This head is located outside of VP and below Voice to re-
flect that the subject c-commands the applicative argument. The denotation of
this head (68a) includes both at-issue and not-at-issue material (following the
colon). Semantically, Aff introduces an experiencing event and the experiencer.
It takes the event property denoted by its sister constituent to be the source of

14 Hole includes some not-at-issue meaning, but does not have an experiencing event.

 
42   

the affected experiencer's experience. This source information is on the not-at-


issue tier of meaning (as indicated by it following the colon in the denotation).
The affected experiencer itself is projected in the at-issue content.15
68. a. [[Aff]] = λPvt.λx.λe. P(e) & (e')(EXPER(e') & Exp(x)(e')): e''
(P(e'') → Source(e'')(e'))
b. Source → λe.λe'. e is the source of e'.
The syntactic and semantic analysis for (41a), repeated below, proceeds as
follows.16
69. a. Alex zerbrach Chris Bens Vase.
Alex.Nom broke Chris.Dat Ben.Gen vase.Acc
'Alex broke Ben's vase on Chris.'
b. VoiceP<v,t>
3
Alex Voice'
3
Voice AffP<v,t>
3
Chris Aff'
3
Aff<vt,evt> VP<v,t>
6
zerbrach Bens Vase
'broke Ben's vase'
c. [[VoiceP]] = λe. BREAK(e) & Thm(e)(Ben's vase) & Agt(e)(Alex)
& e'(EXPER(e') & Exp(e')(Chris) : e''((BREAK(e'') &
Thm(e'')(Ben's vase)) → Source(e'')(e'))
(based on Bosse et al. 2012, examples 67, 68)
The Aff head takes the VP as its complement. Semantically, this yields the VP-
event as the source of the experience of the affected experiencer. VoiceP con-
veys the not-at-issue meaning that Chris would be affected by any event of
breaking Ben's vase. The source event does not include Alex as the agent as part
of the source event. This reflects the idea that the affected experiencer c-
commands the object and is c-commanded by the subject as well as the native
speaker intuition that what matters to the affected experiencer (and is therefore
the source event) does not include the agent (70).

15 Alternatively, Aff could be defined so that it combines with the VP via event identifica-
tion. As mentioned in Bosse et al. (2012), functional application and event identification
(of Voice) can yield the same result. Having Voice combine by event identification
keeps the analysis here consistent with the ones presented in other chapters.
16 The full semantic derivation is given in Appendix 1, #1.


  43

70. a. Alex zerbrach Chris Bens Vase.


Alex.Nom broke Chris.Dat Ben.Gen vase
'Alex broke Ben's vase and…
1. … it matters to Chris because it was Ben's vase.'
2. … #it matters to Chris because Alex did it.'
(Bosse et al. 2012, example 65)
b. Gerald broke the vase on me.
1. It matters to me because it was the vase.
2. #It matters to me because Gerald did it.
Actual contexts can make the agent more or less relevant in both German and
English. For instance, (70a) could be followed with deshalb spricht Chris jetzt
nicht mehr mit Alex ("because of that Chris doesn't talk to Alex anymore"). This
makes it more salient that Chris is angry because it was Alex who broke the
vase. However, sentence (70a) still cannot convey the meaning that Chris is af-
fected by Alex's action. It has to be the breaking event that affects Chris with the
possible (contextual) addition that Alex was the agent of the event that affected
him. In other words, in the absence of context the agent does not play a role in
the source of the experiencing event, but contextual relevancy compatible with
the agent's "responsibility" for the event can make the agent more salient. Yet, if
anyone else had caused the event, it would still have mattered to Chris. Due to
this, it is important that the applicative argument is located between Voice and
VP. If the Aff head were outside of Voice, the agent would be included in the
source event (section 3.9).
English and German have patterned alike in the c-command tests (64-67)
and native speaker intuitions (70). The major difference between English and
German is that English requires the presence of the preposition on to introduce
the affected experiencer. I argue that this preposition has essentially the same
denotation as the phonologically null Aff head in German. In other words, on is a
prepositional variant of Aff. However, since the preposition and the affected ex-
periencer form one constituent (an adjoined PP), the denotation needs to be
modified in such a way that the affected experiencer argument is the first argu-
ment and the VP the second one (71).
71. [[on]] = λx.λPvt.λe. P(e) & e'(EXPER(e') & Exp(x)(e')): e''(P(e'') →
Source(e'')(e'))
Syntactically, this difference between German and English is reflected in the
adjunction of the on-PP to the VP in English. Thus, (41b), repeated below, has
the following derivation.17
72. a. Gerald broke the vase on me.

17 The full semantic derivation is given in Appendix 1, #2.

 
44   

b. VoiceP
3
Gerald Voice'
3
Voice VP1
3
VP PPAff
3 3
V NP PAff NP
broke 5 on 4
the vase me
c. [[VoiceP]] = λe. BREAK(e) & Thm(e)(the vase) & Agt(e)(Gerald) &
e'(EXPER(e') & Exp(e')(me)): e''(BREAK(e'') & Thm(e'')(the
vase) → Source(e'')(e'))
Consequently, German and English affected experiencer constructions yield
the same meaning but differ in how they achieve this. English needs an adjoined
PP with an "affective" preposition whereas German has a phonetically null Af-
fect head which selects VP.
3.5 Consequences
The analysis of affected experiencer applicative arguments as presented in sec-
tion 3.4 has several consequences. The first one is that Aff introduces a second
event (in addition to the verbal event). I present evidence for the necessity of
this in section 3.5.1. Related to this bi-eventivity is the availability of these
events for modification, such as again-modification and negation which are dis-
cussed in section 3.5.2 and 3.5.3, respectively.

3.5.1 Bi-Eventivity
One major difference between the analysis presented in section 3.4 and previous
analyses (for instance, the one by Pylkkänen (2002)) is that the affected experi-
encer applicative construction involves a second event, namely the experiencing
event in addition to the verbal event. In Pylkkänen's analysis, the applicative
head only adds a new role (but not event) to the structure (73).
73. λx. λe. APPL(e,x)
(collapsing APPLBEN, APPLINST, APPLLOC and so forth)
(Pylkkänen 2002, example 23)
In contrast, Aff contributes the experiencer role and also an experiencing event.
74. [[Aff]] = λPvt.λx.λe. P(e) & (e')(EXPER(e') & Exp(x)(e')): e'' (P(e'')
→ Source(e'')(e'))


  45

There is evidence that this second event is necessary. In Pylkkänen's analy-


sis, Jan as the agent in (75a) and as the affected experiencer in (75b) should
equally participate in the event of each sentence. In the analysis presented here,
this is not expected because the experiencing event is distinct from the verbal
event. In other words, in Pylkkänen's analysis, Jan is the agent of the killing
event in (75a) and adversatively affected by the dying event in (75b). In the pro-
posed analysis, Jan is the agent of the killing event in (75a) but the experiencer
of an experiencing event in (75b). The dying event is the source of the experien-
cing event.
75. a. Jan brachte den Mann um.
Jan.Nom brought the.Acc man Prt
'Jan killed the man.'
b. Der Mann starb Jan.
the.Nom man died Jan.Dat
'The man died on Jan.'
The bi-eventivity of (75b) can be detected by adverbial modification (cf. Chap-
ter 2). As shown in (76a), a clausal modifier can modify the experiencing event
of an affected experiencer construction without contradicting a PP which modi-
fies the main predicate. This is not possible for agents (76b) unless a second
verb is added (76c).
76. a. Als Jan in Berlin war, starb ihm
when Jan.Nom in Berlin was, died him.Dat
Alex in Bremen.
Alex.Nom in Bremen
'When Jan1 was in Berlin, Alex died in Bremen on him1.'
b. *Als Jan in Berlin war, brachte er
when Jan.Nom in Berlin was, brought he
Alex in Bremen um.
Alex.Nom in Bremen Prt
'When Jan1 was in Berlin, he1 killed Alex in Bremen.'
c. Als Jan in Berlin war, hat er
when Jan.Nom in Berlin was, has he
versucht, Alex in Bremen umzubringen.
tried Alex.Acc in Bremen Prt-to-bring
'When Jan1 was in Berlin, he1 tried to kill Alex in Bremen.'
In Pylkkänen's analysis, if the dying event happens in Berlin and Jan is ad-
versatively affected by this event, this affecting should take place in Berlin as
well. Having two different location modifiers for one event should be impossible
(cf. Parsons 1990, each event can have each participant/ thematic role only
once). In the bi-eventive analysis, the location modifiers can modify different

 
46   

events. It is possible to refine Pylkkänen's analysis to incorporate quantification


over events/ bi-eventivity. However, as it stands the analysis by Pylkkänen can-
not handle cases of bi-eventivity.

3.5.2 Again-Modification
VP adverbs provide further support for my analysis. Assuming that adverbs typ-
ically attach to nodes of type <v,t>, there are three such nodes in the derivation
for German (69), namely VoiceP, AffP and VP, as well as three in English (72),
namely VP, VP1, and VoiceP. None of these nodes include the experiencing
event without including the verbal event. This explains why the PP in the fol-
lowing example cannot modify just the experiencing event.
77. In Berlin starb der Mann Jan.
in Berlin died the.Nom man Jan.Dat
'The man died on Jan and that happened in Berlin.'
*'The man died (somewhere) and it affected Jan in Berlin.'
Again-modification further supports the proposed analysis. As von Stechow
(1996), Bale (2007) and Beck and Johnson (2004) have argued, again can detect
nodes of type <v,t>. Depending on the attachment site, the presupposition of
again (as stated in (78)) will differ.
78. [[again]] = λPvt.λe. P(e): e'[P(e') & the run time of e' preceded that of
e]
In a simple transitive clause, again can attach to either VP or VoiceP, yielding
two different readings (79).
79. I closed the door again.
a. This door was built closed, and has never been opened. One day an
earthquake jarred it open, so I had to go close it again. (VP modifica-
tion: only [door closed] held previously)
b. I've closed this door about ten times already today. Someone just left
it open again, so I had to go close it again. (VoiceP modification: [I
closed door] happened previously)
(Bosse et al. 2012, example 76)
The derivation for German affected experiencers (69) includes three nodes of
type <v,t>: VoiceP, AffP and VP. The following three readings of an affected
experiencer construction modified with again should therefore be possible:
80. 1. VP attachment: a reading where the VP event has taken place be-
fore, without an affected experiencer and not necessarily with the same
agent;
2. AffP attachment: a reading where the AffP event has taken place be-
fore with the same affected experiencer and same VP event, but not
necessarily with the same agent;


  47

3. VoiceP attachment: a reading where the whole VoiceP event has ta-
ken place before with the same VP event, the same affected experi-
encer, and the same agent.
(Bosse et al. 2012, example 77)
There should not be a reading with the same verbal event and the same agent but
a different affected experiencer. This prediction is borne out:
81. Lisa zerbrach Martin das Puzzle wieder.
Lisa.Nom broke Martin.Dat the.Acc puzzle again
'Lisa broke the puzzle on Martin again.'
1. The puzzle is broken again. (repeated VP event)
2. Someone broke the puzzle on Martin before and now Lisa broke it
on him again. (repeated AffP)
3. Lisa had broken the puzzle on Martin before and now she did it
again. (repeated VoiceP)
4. *Lisa had broken the puzzle before and now she broke it again but
for the first time on Martin (experiencer-less repeated VoiceP).18
(Bosse et al. 2012, example 78)
English should allow (and disallow) the same readings as German because the
nodes of type <v,t> in the derivation for English (72) correspond to the ones in
German (69). This prediction is also borne out.
82. My daughter broke the vase again on me.
1. My new vase arrived broken. I glued it together, but then my daugh-
ter broke it on me again. (repeated VP)
2. My dog broke this vase on me yesterday. I put it back together, but
then my daughter broke it on me again. (repeated VP1)
3. My daughter broke this vase on me yesterday, and then she broke it
on me again today. (repeated VoiceP)
4. *My daughter broke this vase yesterday, then today she broke it on
me again. (experiencer-less repeated VoiceP)
These examples show that again cannot be used to only modify the experi-
encing event: for instance, (81) cannot mean 'Martin had some experience before
and now he had another one (because Lisa broke the puzzle).' This is expected
because there is no node of type <v,t> which includes only the experiencing
event (and excludes the verbal event). Thus, the proposed analysis is supported


18 It is possible to stress Martin to seemingly make this reading available. However, in that
case, the stress causes contrastive focus which contrasts Martin with some other indi-
vidual on whom Lisa broke the puzzle before. In other words, even in this case there is a
repetition of the experiencing event.

 
48   

by adverbial modification in general (77), and again-modification in particular


(81, 82).

3.5.3 Negation
As shown in section 3.3, negation cannot target just the experiencing event in
German or English. Thus in sentence (54a), repeated below, negation cannot be
used to negate that Chris was the experiencer of an experiencing event even
though the experiencing event is semantically represented as a conjunct (see
(83b) for the non-negated denotation).
83. a. Alex zerbrach Chris Bens Vase nicht.
Alex.Nom broke Chris.Dat Ben.Gen vase Neg
'Alex didn't break Ben's vase on Chris.'
1. Alex didn't break Ben's vase (but if he had, it would have mat-
tered to Chris.)
2. *Alex broke Ben's vase, but it didn't matter to Chris.
(Bosse et al. 2012, example 30a)
b. [[VoiceP]] = λe. BREAK(e) & Thm(e)(Ben's vase) & Agt(e)(Alex)
& e'(EXPER(e') & Exp(e')(Chris) : e'' ((BREAK(e'') &
Thm(e'')(Ben's vase)) → Source(e'')(e'))
As can be seen from the denotation, the event variable of the experiencing event
is existentially quantified. Existentially quantified variables in general do not
seem to be available as a target of negation. For instance, causatives require that
one event causes another event: thus, in (84), the event of John doing something
caused the ice cream to melt.
84. John melted the ice cream.
Semantically, sentence (84) can be described with the following denotation (be-
low tense).
85. [[(84)]] = λe. DO(e) & Agt(e)(John) & e'. MELT(e') & Pat(e')(ice
cream) & CAUSE(e)(e')
The denotation states that John did something which caused a melting event
with the ice cream undergoing that melting event. The representation of the
melting event includes an existentially quantified event variable. Just like the
experiencing event in the affected experiencer construction, negation cannot tar-
get the caused event in causative sentences. Sentence (86) is ungrammatical on
the reading that John did something that caused the ice cream not to melt. This
sentence can only be used to negate the existence of the causing event.
86. John didn't melt the ice cream.
Thus, the fact that the experiencing event in the affected experiencer construc-
tion cannot be targeted by negation can be linked to the existentially quantified
event variable (for this event).


  49

This unavailability of the experiencing event for negation contrasts with the
fact that the experiencer itself can be negated, namely when that constituent is
focused.
87. Alex didn't break Ben's vase on CHRIS (but on HIS FATHER).
Thus, the experiencer is available for contrastive focus (and by extension for
negation). It is only the existentially quantified experiencing event itself which
cannot be targeted by negation.
This discussion calls into question the reliability of negation as a not-at-
issue diagnostic (Chapter 2). It seems that for bi-eventive (monoclausal) struc-
tures, negation may not be reliable because existentially quantified events can-
not be negated (regardless of the tier of meaning they contribute to). This shall
not concern us here further because yes/no-questions also detected some not-at-
issue meaning of affected experiencer applicative arguments (55).
Another issue related to negation is that the denotation of Aff states that the
referent of the affected experiencer argument experienced something if any
event of that particular kind did happen. It is not ensured that the event has to
happen in the real world for the affected experiencer to be affected (in the real
world). This problem can be solved by adding a counterfactual meaning to the
denotation of Aff (88).
88. [[Aff]]w= λPvt.λx.λe.λw. P(e)(w) & e'(EXPER(e')(w) & Exp(x)(e')(w))
: e'' (P(e'')(w) → Source(e'')(e')(w)) & w'. P(e) = 0 in w' → e'. EX-
PER(e')(w) & Exp(x)(e')(w) = 0
(w' is maximally similar to w)
This counterfactual denotation ensures that the event must happen in the real
world for it to be the source of the experiencing event. If the event does not hap-
pen in the real world, then there is no experiencing event. I assume this version
of Aff below but use the simplified (non-counterfactual) version in the deriva-
tions because this modification does not have a direct effect on other aspects of
my analysis.
3.6 Pertinence Dative
In this section, I discuss the German pertinence dative (89) as a subtype of the
affected experiencer applicative argument. I argue that pertinence datives are
affected experiencer constructions which induce a conversational implicature
concerning (inalienable) possession. I have already shown that the pertinence
dative patterns with the affected experiencer in terms of (not-)at-issue meaning
(section 3.3.2). Below, I first show how my analysis of affected experiencers
works for pertinence datives (section 3.6.1) before addressing previous analyses
(section 3.6.2).

 
50   

89. Jan brach Martin den Arm.


Jan.Nom broke Martin.Dat the.Acc arm
'Jan broke Martin's arm.'

3.6.1 Analysis
I argue that it is not necessary to include a thematic possessor role in the analy-
sis of the pertinence dative. Rather, its possessive meaning is the result of a con-
versational implicature. This is in the spirit of Shibatani (1994), who discusses
the difference in interpretation between the pertinence dative and affected expe-
riencers in terms of pragmatically constructed affectedness, which means that it
is easier to be affected if the affected entity can easily be integrated into the
event (for instance, because it is a body part of an individual participating in the
event).
The first piece of evidence for the idea that the pertinence dative is not a
separate type of applicative argument (but rather the affected experiencer con-
struction) comes from the fact that the possession relation is not always neces-
sary. With relational nouns, body parts, and clothes (that are being worn), the
possessor relation is typically present but with other nouns it is possible (or like-
ly) (90) but by no means necessary.
90. Er ruinierte mir die Wohnung.
he.Nom ruined me.Dat the apartment
'He ruined my apartment.'
(Lee-Schoenfeld 2006, example 42)
The salient interpretation of (90) is that it is my apartment but that, in fact, is not
necessary; for instance, it could just as well be my parents' or my friend's apart-
ment or some other apartment (which I care enough about to be affected when it
is ruined). Thus, a possessive relation between the applicative argument and an-
other entity is only one possible interpretation of the sentence.
Furthermore, von Polenz (1969) observes that relational nouns used with the
pertinence dative behave unexpectedly in that deleting the possessor from the
sentence can impact the grammaticality of the sentence (91). This is unexpected
because applicative arguments are not selected but rather freely added to a sen-
tence and should consequently not impact the grammaticality (Chapter 1).
91. a. Er klopfte seinem Freund auf
he.Nom tapped his.Dat friend on
die Schulter.
the.Acc shoulder
'He tapped his friend on the shoulder.'
(von Polenz 1969, example 1, my translation)


  51

b. Er klopfte seinem Freund auf den Tisch.


he.Nom tapped his.Dat friend on the.Acc table
'He tapped on the table on his friend.'
(von Polenz 1969, example 5, my translation)
c. *Er klopfte auf die Schulter.
he.Nom tapped on the.Acc shoulder
'He tapped on the shoulder.'
(von Polenz 1969, example 5, my translation)
d. Er klopfte auf den Tisch.
he.Nom tapped on the.Acc table
'He tapped on the table.'
(von Polenz 1969, example 7, my translation)
Regardless of the noun, the sentences including the applicative argument are
grammatical (91a, b). Yet, without the applicative argument only one of them is
grammatical, namely the one with the sortal noun Tisch 'table' (91d). The sen-
tence with the relational noun Schulter 'shoulder' is ungrammatical when the ap-
plicative argument is not present (91c). However, von Polenz's judgment of
(91c) is not quite accurate. The sentence is not ungrammatical per se. Rather, it
is infelicitous because it cannot be used without context. For instance, the sen-
tence is perfectly acceptable if it is used to describe what an instructor of an
anatomy class did to a corpse. The ungrammaticality of (91c) can consequently
be attributed to the fact that shoulder is a relational noun and is not dependent
on the applicative argument. The unacceptability of (91c) (for von Polenz) can
be attributed to the fact that typically a possessor to a relational noun must be
realized (Isacenko 1965). The possessor can be realized either by adding an ap-
plicative argument (91a), a genitive (92a) or a possessive pronoun (92b).
92. a. Er klopfte auf die Schulter des Freundes.
he.Nom tapped on the shoulder the.Gen friend.Gen
'He tapped on his friend's shoulder.'
b. Er klopfte auf seine Schulter.
he.Nom tapped on his shoulder
'He tapped on his shoulder.'
This shows that the absence of the dative constituent in (91c) is not what causes
the ungrammaticality. Rather, the relation that it provides with respect to the re-
lational noun shoulder does (cf. Isacenko 1965, Wegener 1983, Shibatani 1994).
The second piece of evidence that the possessive meaning of the pertinence
dative does not need to be built into its analysis comes from sentences with sor-
tal nouns which can have a possessive interpretation (93).

 
52   

93. Tim hat der Nachbarin das Auto gewaschen.


Tim.Nom has the.Dat neighbor the.Acc car washed
'Tim washed the neighbor's car.'
(Lee-Schoenfeld 2006, example 1)
The salient interpretation of this example is that the car belongs to the neighbor.
But this interpretation is not necessary (contrary to what Lee-Schoenfeld (2006)
claims). It could also be that Tim was washing his own or someone else's car,
and that affected the neighbor in some way. This becomes even more obvious
when a possessive pronoun referring to Tim is used:
94. Tim1 hat der Nachbarin2 sein1/*2 Auto gewaschen
Tim.Nom has the.Dat neighbor his.Acc car washed
'Tim has washed his car on the neighbor.'
This sentence is grammatical, showing that the car does not have to be that of
the neighbor. The pertinence dative should be ungrammatical when a possessor
is expressed explicitly if its function is to encode the possessive relationship.
This, however, is not the case, as the following examples also show.
95. a. Der Stein fällt dem Mann auf
the.Nom brick falls the.Dat man on
den Kopf des Kindes.
the.Acc head the.Gen child.Gen
'The brick falls onto the head of the child on the man.'
b. Jana tritt dem Mann auf das Kleid
Jana.Nom steps the.Dat man on the dress
seiner Frau.
his wife
'Jana steps on the dress of his wife on the man.'
These two sample sentences are both acceptable. Yet they do not have the inter-
pretation expected for the pertinence dative but rather that of an affected experi-
encer (Isacenko 1965). This can be taken as an indication that the pertinence da-
tive is an affected experiencer.
In summary, the possessor interpretation of the pertinence dative is more sa-
lient in some examples (90, 93) compared to others (91b). This saliency seems
to be related to the noun. Relational nouns support the possessor interpretation
more easily than sortal nouns, and sortal nouns that usually have individual pos-
sessors support the possessor interpretation more readily than those that do not
have individual possessors. For instance, (90) supports the possessor interpreta-
tion easily because apartments are typically owned/rented by an individual. In
contrast, sentence (96) does not support the possessor interpretation (easily) be-
cause schools are typically not individually owned.


  53

96. Er ruinierte mir die Schule.


he.Nom ruined me.Dat the school
'He ruined the school on me.'
Similarly, table in (91b) falls somewhere in between as it may be individually
owned or communal property and can consequently be understood as including
or excluding a reference to the possessor of the table.
I argue below that the possessive relation is a conversational implicature that
arises more easily with relational nouns than with sortal nouns. The possessor
reading for (89) is obligatory (unless the context specifies a different arm) but in
(93) it is only a tendency to interpret the car as belonging to the neighbor. Thus,
I agree with Wegener (1983), who argues that the semantic contribution of the
pertinence dative is the affectedness of the referent and not the possession. The
latter is, according to my analysis, a conversational implicature.
This conversational implicature of possession is calculable based on Gricean
maxims (Grice 1967). My analysis of the pertinence dative in (89) as an affected
experiencer with a conversational implicature says that if Jan breaks the arm and
Martin is affected by that, it is a reasonable assumption for the hearer that it was
Martin's arm. If it were not Martin's arm, the speaker should have indicated that,
being a good conversational partner who makes unambiguous contributions. The
conversational implicature is also non-detachable, as it is not lost by substituting
(near) synonyms. Compare (89) to (97).
97. Jan verletzte dem Kind eine (obere) Gliedmaße.
Jan hurt the.Dat child an upper limb
'Jan hurt a(n upper) limb of the child.'
This sentence conveys that the limb is that of the child just like (89) does. Final-
ly, the possessor interpretation is cancelable, i.e. it can be denied without caus-
ing a contradiction.
98. Jan verletzte dem Kind den Arm. Es
Jan.Nom hurt the.Dat child the.Acc arm it
war der Arm des Teddys.
was the.Nom arm the.Gen teddy.Gen
'Jan hurt the arm on the child. It was the arm of the teddy.'
Thus, the possessor interpretation in pertinence datives passes the tests for being
a conversational implicature.
Treating the possession interpretation of the pertinence dative as a conversa-
tional implicature explains the fact that "the possessed item may be the subject
or direct object of a clause, or part of an adverbial phrase" (Neumann 1995, p.
749). A conversational implicature can hold between any constituents. This also
means that there is no need to treat NPs inside of PPs differently, as Abraham
(1973) does. He proposes different analyses for the following two sentences:

 
54   

99. a. Mir schmerzt der Rücken.


me.Dat hurts the.Nom back
'My back hurts (on me).'
(Abraham 1973, example 26, my translation)
b. Der Regen tropft ihm in den Kragen.
the.Nom rain drips him.Dat into the.Acc collar
'Rain is dripping into his collar (on him).'
(Abraham 1973, example 45, my translation)
Even though both sentences are pertinence dative constructions, two different
analyses are needed according to Abraham because in the second example the
inalienable entity is inside of a PP whereas in the first it is not. My conversa-
tional implicature analysis can explain this straightforwardly because there is no
requirement on which constituents are needed to establish the conversational
implicature.19
This approach of possession being conversationally implicated also explains
the difference in grammaticality reported by Wegener (1983) for the following
two sentences.
100. a. Ich sah vom Fenster aus, wie mir
I.Nom saw from.the window Prt how me.Dat
das Kind vors Auto lief.
the.Nom child in.front.the car ran
*'I saw from the window how the child ran in front of my car.'
'I saw from the window how my child ran in front of the car.'
b. Ich sah vom Fenster aus, wie mir
I.Nom saw from.the window Prt how me.Dat
das Kind vor mein Auto lief.
the.Nom child in.front my car ran
'I saw from the window how the child ran in front of my car.'


19 This can also be used to differentiate part-whole applicatives from pertinence datives.
Part-whole applicatives are restricted in which NP can denote the part to the whole (de-
noted by the applicative), as explained in detail in section 6.2.3. The possessor in a per-
tinence dative can be the direct object in the presence of a goal PP. This is not the case
for the part NP in a part-whole applicative construction (see i. and ii. below).
i. *Jan legte dem Auto das Rad auf den Tisch.
Jan laid the.Dat car the.Acc tire onto the table
'Jan put the car's tire onto the table.' (part-whole)
ii.Jan legte dem Patienten das Bein auf das Bett.
Jan laid the.Dat patient the.Acc leg onto the bed
'Jan put the patient's leg onto the bed (on the patient).' (pertinence dative)


  55

(100a) cannot be interpreted as "my car" (but possibly as "my child"). This is
due to the fact that it is clear that I am not driving the car. Consequently, it is not
likely that it is my car. To convey this meaning, it needed to be overtly encoded
that it is my car (100b). Also, if the context supports that is it my car, the impli-
cation of the possessive relationship used in (100a) is possible:
101. Ich fuhr um die Kurve und sah
I.Nom drove around the corner and saw
gerade noch, wie mir das Kind
just Prt how me.Dat the.Nom child
vors Auto lief.
in.front.the car ran
'I was driving around the corner and saw how the child ran in front of
(the/my) car (on me) just at that moment.'
Thus, rather than being hard-wired into the semantic denotation of the sentence,
the possessor relationship is heavily influenced by the context.
In summary, the pertinence dative is not a separate type of applicative ar-
gument but rather the affected experiencer construction with an additional con-
versational implicature. This conversational implicature is virtually required
with relational nouns, whereas it is optional with sortal nouns ("typical affected
experiencer").
Consequently, the derivation of (59a), repeated below, proceeds as fol-
lows.20
102. a. Der Stein fällt dem Mann auf den Kopf.
the.Nom brick falls the.Dat man on the.Acc head
'The brick falls on the man's head.'
b. VoiceP
3
der Stein Voice'
3
Voice AffP
3
dem Mann Aff'
3
Aff VP
6
fällt auf den Kopf


20 The full semantic derivation is guven in Appendix 1, #3.

 
56   

c. [[VoiceP]] = λe. FALL(e) & Goal(e)(the head) & Agt(e)(the stone) &
e' (EXPER(e') & EXP(e')(him)) : e'' (FALL(e) & Goal(e)(the
head) → Source(e'')(e'))
d. Conversational Implicature: The man was affected because his head
was hit.
This approach of the possession being conversationally implicated might al-
so explain why the possessive interpretation can be found in configurations oth-
er than the affected experiencer construction (cf. Neumann (1995)). Two such
examples are given in (103).
103. a. Der Hund hebt das Bein.
the.Nom dog lifts the.Acc leg
'The dog lifts its leg.'
(Neumann 1995, example 34)
b. Die alte Frau wackelt mit dem Kopf.
the.Nom old woman is.wobbling with the.Dat head
'The old woman's head is wobbling.'
(Neumann 1995, example 37)
In these examples, a possessive relation between two entities is understood but
this relation is not overtly encoded. If the possessor of a relational noun can be
satisfied by a conversational implicature, these examples can straightforwardly
be explained.

3.6.2 Previous Analyses


There have been many attempts to analyze the pertinence dative, starting at least
as early as von Polenz (1969). The approaches taken vary widely; I briefly
summarize some of them here and address why they do not explain all aspects of
the pertinence dative.
3.6.2.1 Binding
Hole (2005; 2008; 201X) presents a binding approach to pertinence datives. He
argues that there are two basic types of applicative arguments in German, name-
ly what I discuss as part-whole applicative arguments ((Chapter 6); 'landmark'
for Hole) and experiencers/ beneficiaries. The pertinence dative is taken as a
combination of the semantics of the part-whole applicative argument and the
semantics of the affected experiencer. This seems reasonable at first because the
pertinence dative displays features of both (see section 3.2.2).
The referent of the pertinence dative must be able to perceive the event and
be the possessor of an entity (body part) involved in the event. Hole (2005) pro-
vides the following analysis. An Aff head is merged outside of VP. It provides
the thematic role of Affectee (104).


  57

104. [[Aff]]0 = λx. λe. Affectee(x)(e)


(Hole 2005, example 19a)
This head also inserts an abstractor between VP and Aff, as schematized in
(105).
105. 3 → 3
Aff0 VP Aff0 VP
3
λn VP
(Hole 2005, example 19b)
Semantically, the VP and Aff combine by Variable Identification, as shown in
(106).
106. f g → h
<e,<s,t>> <e,<s,t>> <e,<s,t>>
(Hole 2005, example 19c)
The VP has the semantic type <e ,<s,t>> because Hole assumes that every noun
has a relational counterpart and that one is used inside the VP. Consequently,
there is still an open individual argument (type e). Hole's sample (partial) syn-
tactic and semantic derivation is given in (107).
107. Sie will dem Jungen den Kopf streicheln.
she.Nom wants the.Dat boy the head stroke
'She wants to stroke the boy's head.'
Aff
λe. Affectee(the boy)(e) & stroke the boy's head(e)
3
SpecAffP Aff'
dem Jungen λx.λe.Affectee(x)(e) & stroke x's head(e)
3
Aff0 VPhigh
λx.λe.Affectee(x)(e) λx.λe.stroke X's head(e)
3
λ5 VPlow
λe. stroke g(5)'s head(e)
6
den Kopf streicheln
(Hole 2005, examples 15, 20)
The possessive relation is established between the theme and the affectee via the
abstractor inserted by Aff.
Apart from the not-at-issue meaning of affected experiencer/ pertinence da-
tive construction that I have argued for above, Hole's binding approach runs into
the following problems. It is unclear which variable is bound in examples like

 
58   

(98), repeated below. The salient reading of the first clause is that of a perti-
nence dative: the child is the possessor of the arm. The second sentence cancels
this relation in order to establish that the arm belongs to the teddy instead.
108. Jan verletzte dem Kind den Arm. Es
Jan.Nom hurt the.Dat child the.Acc arm it
war der Arm des Teddys.
was the.Nom arm the.Gen teddy.Gen
'Jan hurt the arm on the child. It was the arm of the teddy.'
Due to the canceling of the possessive relation, dem Kind should not bind into
den Arm because it is the teddy's arm and not the child's arm. Thus, there is
nothing for dem Kind to bind. Hole (p.c.) claims that the teddy must belong to
the child in this case. Consequently there is a silent pronoun bound by dem Kind.
It is indeed one possible interpretation that the teddy belongs to the child. How-
ever, this approach becomes problematic in the case when the teddy does not
belong to the child, which is also a possible interpretation of the sentence. For
instance, the child could be watching TV and adoring the teddy who is part of
the program. In that case, Jan could be the mean kid who hurt the teddy's arm on
the program (to the detriment of the child). It seems farfetched to say that the
child is the possessor of the teddy. For Hole, the dative-marked constituent
could bind some silent purpose/ beneficiary variable which in this case must be
something along the lines of "for the child to become upset/ sad". In this case,
the possessive relation would be reduced to a beneficiary. Hole's binding ac-
count is then essentially saying that every sentence has a silent purpose variable
available to be bound by an applicative. The interpretation of that silent purpose
clause is then contextually specified (possession, benefit, affectedness). In other
words, the cancelability of the possessive relation of the pertinence dative leads
in Hole's system to the pertinence dative being reduced to a beneficiary (with a
silent purpose clause).
This problem also arises in cases where there is an overt possessor that dif-
fers in its referent from the applicative argument (94).
109. Tim1 hat der Nachbarin2 sein1/*2
Tim.Nom has the.Dat neighbor.Fem his.Acc.Masc
Auto gewaschen.
car washed
'Tim has washed his car on the neighbor.'
In this example, the applicative must also bind into a silent purpose clause on
Hole's approach because the possessor of the vase is specified. The applicative
cannot bind into the possessive pronoun because that would lead to a gender
mismatch. Thus, it is necessary to assume a silent purpose clause for the applica-


  59

tive to be able to bind. This is not necessary on my analysis of the pertinence


dative.
In cases where the direct object is a pronoun, the same problem can be ob-
served.
110. Jan hat sie mir aufgekratzt.
Jan.Nom has them.Acc me.Dat scratched.open
'Jan has scratched them open on me.'
This sentence can be interpreted as them (for instance, wounds) being either part
of me or of Jan (or belonging to neither of us). For the interpretation that Jan
scratched mine open, Hole must assume that the pronoun has a silent possessor
that can be bound. This possessor cannot be overtly expressed.
111. *Jan hat meine sie mir aufgekratzt.
Jan.Nom has my them.Acc me.Dat scratched.open
'Jan has scratched them (belonging to me) open on me.'
The interpretation of (110) where they belong to Jan requires the agent to bind
the possessor of the pronoun and the dative to bind another silent purpose
phrase. For my analysis that is not a problem because me can be interpreted as
an affected experiencer (or benefactive) without requiring a possessive relation.
Finally, the third interpretation of (110), sie 'them' belonging to a third individu-
al, requires again a silent purpose clause on Hole's account because there is
nothing else for the applicative to bind. Again, on my analysis me can receive an
affected experiencer (or benefactive) interpretation. No further relation to the
other entities of the event is necessary.
In short, Hole must pose unpronounced purpose clauses in sentences where
there is no good motivation to have them. My account does not require these
additional unpronounced clauses.
To support his account, Hole argues that sloppy identity effects can be ob-
served with pertinence datives. This is taken to indicate that the dative-marked
constituent must bind. Hole (201X) provides the following example.21


21 Hole assumes that binding into (bridging) definites and prepositions including definite
determiners is possible.

 
60   

112. Dem Patienteni platzte ein Stück Gips


the patient.Dat cracked a piece cast
(vomi/ von seinemi) Arm ab, und dem Arzt
off.the off his arm off and the doctor.Dat
auch.
too
'It happened to the patienti that part of the cast on his1 arm came off, and
it happened to the doctorj that part of the cast on hisj came off, too.'
*'It happened to the patienti that part of the cast on his1 arm came off,
and it happened to the doctorj that part of the cast on hisi came off, too.'
(Hole 201X, example 11)
The sloppy reading of this sentence is available. Sentence (112) can felicitously
be used in a context where the patient and the doctor each had a cast on their
arms. Let's assume that the plaster is faulty, then (112) can describe that some
plaster came off of each of their casts. The reading that Hole claims is not avail-
able is in fact available as well (at least for some native speakers, including my-
self): let's assume that the doctor had put a cast on the patient's arm (and the
doctor does not have a cast himself). The plaster of the cast is faulty and some
came off when the patient was at home. He came back and when the doctor ex-
amined the cast, another piece came off. Sentence (112) can describe this situa-
tion. This shows that this sentence does not only have a sloppy reading; a strict
reading is possible. Consequently, binding is not required for pertinence datives.
Another problem for the binding approach is the observation by Wegener
(1983) that (100a), repeated below, cannot be interpreted as "my car". It is not
clear why the applicative cannot bind into the PP in this case on Hole's account.
113. Ich sah vom Fenster aus, wie mir
I.Nom saw from.the window Prt how me.Dat
das Kind vors Auto lief.
the.Nom child in.front.the car ran
*'I saw from the window how the child ran in front of my car.'
'I saw from the window how my child ran in front of the car.'
As discussed above, if the possessive interpretation relies on a conversational
implicature, this meaning can be explained.
Many of the issues that Hole (2005) addresses are not problematic for my
approach. For example, Hole argues that binding cannot cross CP boundaries
but it can cross relational and event nouns.


  61

114. a. Ed hat ihrx einen Sack, in den die


Ed.Nom has her.Dat a bag in which the
Wäsche reinpasste(*s), gewaschen.
laundry fitted washed
'Ed washed a bag for her that the laundry would fit in.'
(Hole 2005, example 33)
b. Ed hat ihrx einen die Wäsche
Ed.Nom has her.Dat a the laundry
fassenden Sack gewaschen.
accommodating bag washed
'Ed washed a bag for her that the laundry would fit in.'
(Home 2005, example 36)
c. Man verweigerte ihmx die Auszahlung
one.Nom denied him.Dat the payment
des Lohns der Mutter.
of.the wages of.the mother
'They denied him the payment of his mother's wages.'
(Hole 2005, example 37)
Sentences (114a) and (114b) can be affected experiencer constructions on my
approach (or benefactives, which is not relevant here). In that case, the referent
of the affected experiencer argument must be affected by the washing of the bag.
There is no statement about whether the laundry is her laundry; it may be or it
may not be. As Hole states, these sentences can be interpreted either way. Sen-
tence (114c) is a pertinence dative, i.e. an affected experiencer with a conversa-
tional implicature (due to the use of a relational (kinship) noun). This sentence is
expected to be grammatical on my approach because the conversational implica-
ture may hold between the applicative and a NP included in a complex NP. Thus
these sentences do not pose any problems for my theory.
Finally, it is puzzling that Hole (2008) states that the part-whole construc-
tion is restricted to some dialects of German but the pertinence dative and the
affected experiencers are not. In other words, speakers that do not allow the
part-whole constructions have still independently acquired its semantics to be
used in the pertinence dative construction according to Hole.
In summary, Hole's binding approach is not well-suited to explain perti-
nence dative constructions because his approach must often be reduced to the
dative binding a silent purpose clause (or possessor) rather than an overt ele-
ment. Furthermore, the possessive interpretation must often be reduced to a be-
neficiary. In that case, it becomes unclear when each interpretation is expected.

 
62   

3.6.2.2 Possessor Raising


von Polenz (1969) suggests a transformational approach to the pertinence dative,
i.e. that the pertinence dative constructions are derived from underlying senten-
ces with possessors. Lee-Schoenfeld (2006) takes a similar approach but updat-
ed to recent theoretical developments. She aligns the German pertinence dative
with Hebrew possessor datives that Landau (1999) analyzes as possessor raising
constructions. In the analysis by Lee-Schoenfeld (2006), the pertinence dative is
generated DP-internally as a possessor of the theme/patient and moves to the
specifier of a vP-shell where it receives the male-/benefactive θ-role (which car-
ries the affectedness meaning). This is exemplified below (slightly modified
presentation of her structure).22
115. Er ruinierte mir die Wohnung.
he.Nom ruined me.Dat the.Acc apartment
'He ruined my apartment.'


22 Lee-Schoenfeld (2006) defines the non-core datives for which her analysis should hold
as those "which involve both affectedness and a possessor relation." (p. 107). This is the
hallmark of pertinence datives. It differs from other definitions in not requiring body
parts/ pieces of clothing as possessed items. Lee-Schoenfeld first explicitly excludes
ethical datives, judgment datives and "ficiaries" (benefactives, malefactives; including
my affected experiencers) though she then extends her analysis to these ficiaries and
judgment datives.


  63

vP
3
DP v'
Er 3
vP vAgent
3 ACC
DP v' ruinierte
DAT 3
mir VP vMale/Benefactive
3 <arg>
V DAT
3 tv
DP VTheme/Patient
3 <arg>
tPD D' tV
3
DPossessor NP
 6
ACC N
die Wohnung
(Lee-Schoenfeld 2006, example 42)
In Lee-Schoenfeld's analysis of (115), the applicative argument mir is generated
inside of the theme-DP where it receives the possessor θ-role. D is assumed to
be defective so that the pronoun cannot receive case there. The pronoun moves
to the specifier of vP where it receives dative case and the male-/benefactive θ-
role from v. Consequently, mir 'me' receives an interpretation of being affected
as well as the possessor of the theme.
Lee-Schoenfeld extends her analysis to sentences with overt possessors ("fi-
ciaries"). She argues that in (116), the (affected experiencer) applicative argu-
ment der Mama is generated in the specifier of vP and the possessor Omas is
generated inside the theme-DP.
116. Du hast der Mama doch hoffentlich nicht
you have the mom.Dat but hopefully not
Omas Geschirr kaputt gemacht.
grandma.Gen dishes.Acc broken made
'I sure hope you did not ruin Grandma's dishes on mom.'
(Lee-Schoenfeld 2006, example 43a)
In this case, the theme-DP requires a non-defective D which can license genitive
case for the possessor. In other words, affected experiencers and pertinence da-
tives have basically the same structure according to Lee-Schoenfeld. The differ-

 
64   

ence is whether D (of the theme) is defective. It must be for the pertinence da-
tive to surface.
On this approach, it is unclear how to allow defective D only in pertinence
datives/ possessor-raising constructions. For instance, it is unclear why raised
possessors cannot appear in ditransitive constructions (117, also see below).
117. *Jan gab Sabine der Mutter das Buch.
Jan.Nom gave Sabine.Dat the.Dat mother the.Acc book
'Jan gave the book to Sabine's mother (on Sabine).'
'Jan gave Sabine's book to the mother (on Sabine).'
It is impossible to raise a possessor either out of the direct or the indirect object
of a ditransitive structure. The possessor raising approach does not explain why
this movement is prohibited. In other words, the possessor raising approach is
not restricted enough to limit the raising to occur only in pertinence dative con-
structions.
Another issue is the cancelability of the possessor relation (98), repeated be-
low.
118. Jan verletzte dem Kind den Arm. Es
Jan.Nom hurt the.Dat child the.Acc arm it
war der Arm des Teddys.
was the.Nom arm the.Gen teddy.Gen
'Jan hurt the arm on the child. It was the arm of the teddy.'
It is unclear how this is handled in a possessor raising analysis because the first
clause would assert that the arm belongs to the child while the second one would
assert that the arm belongs to the teddy. This contradiction cannot straightfor-
wardly be resolved.
Hole (2008) points out further problems with a possessor raising analysis.
He discusses, for instance, coordination. In the following example, there is one
pertinence dative but its referent is understood to be the possessor of two entities
which participate in different events.
119. Paul hat seiner Tochter die Haare
Paul.Nom has his.Dat daughter the.Acc hairs
gewaschen und die Nägel saubergemacht.
washed and the.Acc nails cleaned
'Paul has washed his daughter's hair and cleaned her nails.'
(Hole 2008, example 13.3, my translation)
As Hole explains, possessor raising requires two traces here, one for die Haare
and one for die Nägel. However, there is only one antecedent (seine Tochter).
To resolve this problem, it must be stipulated that one antecedent is unpro-
nounced (or two identical ones are phonologically merged). My analysis does
not have this problem because the two VPs can be coordinated and Aff can at-


  65

tach outside of this constituent. The conversational implicature can then be es-
tablished between the affected experiencer and each relational noun individual-
ly.
Overall it can be concluded that the possessor raising analysis cannot ade-
quately explain the phenomenon of the German pertinence dative.

3.6.3 Conclusion
In this section, I have argued that the pertinence dative is an affected experiencer
construction with an additional conversational implicature. Previous analyses in
which the possessive meaning of the pertinence dative is explicitly included in
the analysis cannot explain why this meaning only surfaces in some cases.
3.7 Verbal Restrictions
Besides the animacy requirement for the referent of the applicative argument
(section 3.2), the verb also plays an important role in licensing the affected ex-
periencer construction. Affected experiencers (including the pertinence dative)
are only possible with some verbs in German. This speaks against the claim
made by Roberge and Troberg (2007) that only languages in which applicatives
are clitics have verbal restrictions: German allows full NP applicatives and has
verbal restrictions. The restriction is not related to the valency of the verb but
rather to the meaning of the verbs (Wegener 1983, Hole 2008, among others).
First, I discuss why valency restrictions are not good enough to describe the li-
censing conditions of affected experiencers (section 3.7.1). I then discuss the
semantic licensing conditions of the affected experiencer applicative argument
(section 3.7.2).

3.7.1 Valency
Affected experiencers can appear with transitive and intransitive verbs, but not
with ditransitives (120).
120. *Jan gab ihr Maria einen Kuss.
Jan.Nom gave her.Dat Maria.Dat a.Acc kiss
'Jan gave her a kiss on Maria.'
This is expected on my approach because Aff selects a VP of type <v,t>: in ac-
cordance with the analysis of ditransitives proposed by Bruening (2010), I as-
sume that Aff cannot attach between the projections introducing the two objects.
Lamiroy and Delbecque (1998) as well as Neumann (1995) show that the
possessed entity in a pertinence dative can be the subject of an intransitive verb
(121a), the object of a transitive verb (121b) or part of a PP with either intransi-
tive (121c) or transitive verbs (121d).

 
66   

121. a. Mir beschlägt die Brille.


me.Dat mist.up the.Nom glasses
'My glasses are misted up.'
(Lamiroy and Delbecque 1998, example 28b)
b. Sie haben ihm den linken Arm gebrochen.
the.Nom have him.Dat the.Acc left arm broken
(Lamiroy and Delbecque 1998, example 30a)
c. Die Haare hängen ihm ins Gesicht.
the.Nom hair hang him.Dat in.the.Acc face
'His hair is hanging in his face.'
(Lamiroy and Delbecque 1998, example 29b)
d. Er hat ihm etwas ins
he.Nom has him.Dat something.Acc in.the.Acc
Ohr geflüstert.
ear whispered
'He whispered something into his ear.'
(Lamiroy and Delbecque 1998, example 31a)
This is expected on my approach because Aff can attach to the VP and the con-
versational implicature of the pertinence dative can hold between any constitu-
ents. The unaccusative construction in (121a) receives the following derivation.
122. AffP
3
me Aff'
3
Aff VP
6
mist up the glasses
[[AffP]] = λe. MIST UP(e) & Thm(e)(the glasses) & e'(EXPER(e') &
Exp(me)(e')): e'' (MIST UP(e'') & Thm(e'')(the glasses) →
Source(e'')(e'))
Conversational Implicature: I was affected because they were my
glasses.
The derivation of a transitive sentence with an affected experiencer is given
above (69, 102).
However, this description is not completely adequate because not all intran-
sitive and transitive verbs allow the affected experiencer construction; some un-
grammatical examples are given in (123).
123. a. *Jan arbeitet ihr.
Jan.Nom works her.Dat
'Jan works on her.'


  67

b. *Jan läuft ihr.


Jan.Nom runs her.Dat
'Jan runs on her.'
c. *Er entdeckt ihm einen Löwen.
he.Nom discovers him.Dat a.Acc lion
'He discovers a lion n him.'
d. *Sie wirft ihm den Ball.
she.Nom throws him.Dat the.Acc ball
'She throws the ball on him.'
This pattern of only some transitive and some intransitive licensing the affected
experiencer has led to semantic analyses of verbs that allow this construction
(Isacenko 1965, Wegener 1983) which I discuss in the following subsection.

3.7.2 Semantic Licensing


Isacenko (1965) argues that affected experiencers are only possible with verbs
that physically affect the object. He discusses the following example.
124. Er wusch (behandelte, verletzte, tätowierte)
he.Nom washed treated hurt tattooed
ihm den Arm.
him.Dat the.Acc arm
'He washed (treated, hurt, tattooed) his arm (on him).'
(Isacenko 1965, example 49, my translation)
Isacenko (1965) contrasts this with verbs that do not denote physical affected-
ness. These are ungrammatical with the affected experiencer applicative argu-
ment (125).
125. a. *Er bewunderte ihr die Hände.
he.Nom admired her.Dat the.Acc hands
'He admired her hands (on her).'
(Isacenko 1965, example 50, my translation)
b. *Er betrachtete ihr das Haar.
he.Nom looked.at her.Dat the.Acc hair
'He looked at her hair.'
(Isacenko 1965, example 51, my translation)
c. *Er lobte ihr die Figur.
he.Nom praised her.Dat the.Acc figure
'He praised her body (on her).'
(Isacenko 1965, example 52, my translation)
However, Wegener (1983) argues that physical affectedness is not the deci-
sive characteristic. Verbs that describe psychological affectedness can also li-

 
68   

cense the affected experiencer/ pertinence dative; for instance, perception verbs
of non-physical contact (126).
126. Er schaut ihr in die Augen.
he.Nom looks her.Dat in the.Acc eyes
'He looks into her eyes (on her).'
(Wegener 1983, example page 151, my translation)
Wegener explains that (127) is degraded compared to (126) because looking at
the back of someone does not have a (positive or negative) effect on that person
but looking into someone's eyes does.
127. ?Er schaut ihr auf den Rücken.
he.Nom looks her.Dat on the back
'He looks at her back (on her).'
(Wegener 1983, example page 151, my translation)
However, the native speakers I consulted agree that this sentence is not degrad-
ed. Consequently, psychological affectedness cannot be the right characteristic
semantically for delimiting affected experiencers either. (Although it is possible
that these native speakers would state that looking at someone's back involves
psychological affectedness for the person being looked at.)
Wegener (1983), Lamiroy and Delbecque (1998), Hole (2008) all discuss re-
sultativity as the important characteristic of VPs that license affected experi-
encers. Looking at intransitive verbs first, their categorization by Sorace (2000)
supports the tendency towards resultative verbs licensing affected experiencers.
Change of state verbs license affected experiencers (and the new state is the re-
sult).23
128. Der Baum ist mir verrottet.
the.Nom tree is me.Dat rotten
'The tree rotted on me.'
In contrast, continuation of condition verbs only marginally allow affected expe-
riencers.24
129. ??Der Hund hat mir überlebt.
the.Nom dog has me.Dat survived
'The dog survived on me.'
Continuation of condition verbs such as in (129) do not yield a new result (but
maintain the old state) and therefore allow the affected experiencer applicative
only marginally. Context can make these sentences slightly better. For instance,
if I was severely annoyed by my neighbors' dog, and fed it poisonous dog cook-
ies, and the dog survives, I could utter (129). While the sentence would carry the

23 The example is ambiguous with the Dative of Inaction.
24 A Dative of Inaction interpretation of this example is salient.


  69

intended meaning, it would also be thought of as being a funny stylistic choice


because it is not perfectly well-formed.
Existence of state and atelic motion verbs do not allow affected experiencers
(130).
130. a. *Der Hund hat mir existiert.
the.nom dog has me.Dat existed
'The dog existed on me.'
b. *Das Paar hat mir getanzt.
the.Nom couple has me.Dat danced
'The couple danced on me.'
There is no result for these verbs and consequently the affected experiencer is
not licensed. Even providing context cannot improve the acceptability of these
sentences. However, adding a (directional) prepositional phrase to a motion verb
can license an applicative argument (cf. Rosengren (1975), Wegener (1983)).
131. a. *Er läuft mir.
he.Nom runs me.Dat
'He is running on me.'
b. Er läuft mir in den Garten.
he.Nom runs me.Dat in the.Acc backyard
'He is running into the backyard on me.'
The addition of the directional PP in (131b) can be viewed as making the VP
resultative (see also Beck and Snyder 2001); here, the result is that he is in the
backyard now. Neumann (1995) discusses the fact that locative PPs are not as
good as directional ones to license affected experiencers (although there is
speaker variation). This can be attributed to the fact that locative PPs do not
(necessarily) indicate a result. The action may still be atelic/ unbounded. Thus,
resultativity seems to be the key factor in licensing affected experiencers.
Support for resultativity as the important characteristic also comes from par-
ticle verbs. Many German verbs cannot occur with applicative arguments unless
a particle is added, which indicates some kind of result, as shown in (132).
132. a. Das Baby hat geschlafen.
the.Nom baby has slept
'The baby has slept.'
b. *Das Baby hat mir geschlafen.
the.Nom baby has me.Dat slept
'The baby has slept on/for me.'
c. Das Baby ist eingeschlafen.
the.Nom baby is Prt.slept
'The baby fell asleep.'

 
70   

d. Das Baby ist mir eingeschlafen.


the.Nom baby is me.Dat Prt.slept
'The baby fell asleep on me.'
Lüdeling (2001) has shown that some German particles are used to form resulta-
tive constructions; in combination with Wegener's (1983) claim that verbs that
denote results license the affected experiencer/ pertinence dative, the pattern ob-
served in (132) can be explained.
However, there are some problems with using only resultativity for the li-
censing of affected experiencers. First, Neumann (1995) states that "the preposi-
tion [of a PP used to license an affected experiencer] must specify a relation to
an object as extended in space (such as auf 'on', unter 'under'), not a relation to
its position (zu 'to', bei 'near'), and the statement of the spatial relation cannot be
the main purpose of the utterance" (p. 769). She exemplifies this with the fol-
lowing sentences.
133. a. Ein Blatt liegt ihm auf der Wange.
a.Nom leaf is.lying him.Dat on the.Acc cheek
'A leaf is lying on his cheek.'
b. Ein Blatt klebt ihm auf der Wange.
a.Nom leaf is.sticking him.Dat on the.Acc cheek
'A leaf is sticking on his cheek.'
(Neumann 1995, example 68)
Example (133a) is only acceptable when there is (contrastive) focus on liegt but
not if it is used as a simple description on where the leaf is located. Yet, (133b)
does not require contrastive focus to be acceptable. This yields a very murky
picture of which PPs are considered as marking a VP as resultative (and there-
fore license an applicative argument).
Second, the resultativity needed to license the affected experiencer does not
correspond to the verb being able to form a stative passive, which is typically
used as a diagnostics for a resultative verb phrase. Some verbs that license the
affected experiencer do not allow this passive (Isacenko 1965, Wegener 1983).
In (134a), the applicative argument is licensed but (134b) shows that the same
verb cannot occur in a stative passive.
134. a. Er streichelte ihr die Hand.
he.Nom stroked her.Dat the.Acc hand
'He stroked her hand.'
b. *Ihre Hand ist gestreichelt.
her hand is stroked
'Her hand is stroked.'


  71

Wegener (1983) also shows that resultative verbs that require non-
intentional experiencer subjects do not allow applicative arguments. Examples
of these verbs are finden 'find', erhalten 'receive', and verlieren lose'.25
135. *Er verlor mir meine Brille.
he.Nom lost me.Dat my glasses
'He lost my glasses on me.'
It is not expected if resultativity is the crucial characteristic that the intentionali-
ty of the agent can also have an influence on the licensing of affected experi-
encers.
However, this characterization of the verbal restrictions might also explain
why some static verbs are possible with applicative arguments while others are
not.
136. Er hielt Maria die Tasche.
he.Nom held Maria.Dat the.Acc bag
'He held the bag for/on Maria.'
137. *Der Korb enthielt mir zwei Schlangen.
the.Nom basket contained me.Dat two snakes.
'The basket contained two snakes on me.'
It could be argued that holding the bag is an intentional actions with a result (the
bag not being on the ground), whereas containing is not intentional.
As a rule of thumb, resultativity of the verb phrase is the necessary (but not
sufficient) characteristic for the licensing of affected experiencers in German.
This resultativity does not entail that a stative passive can be formed. The re-
sultative requirement is not reflected in the analysis here but I assume that this is
a selectional restriction between Aff and VP. The restriction might be related to
the fact that the VP provides the source event for the experiencing event and re-
sultative/ bounded sources make for better source events than atelic, unbounded
events. However, this restriction holds only for German and not for English (or
Albanian which has otherwise the same features for affected experiencers as
German does (Bosse et al. 2012)). This problem of which verbs/verb phrases
exactly license affected experiencer constructions needs further investigation.
Hole (2008) argues for a change of state operator which ensures that applicatives
occur only with certain verb phrases. In contrast, Beck and Snyder (2001) argue
that goal PPs (in English) carry resultative meaning. The interaction with the
other factors needs more investigation to see if either approach is on the right
track to restrict the German affected experiencer applicative argument correctly.

25 As Wegener points out, in contexts in which these verbs involve intentionality, they
allow applicatives. (135) can be used grammatically if he intentionally left my glasses
somewhere but tried to make it look like he unintentionally lost them.

 
72   

3.8 Affected Experiencers following Potts (2005)


In this section, I briefly discuss how the affected experiencer applicative argu-
ment fits into the system proposed by Potts (2005). Potts characterizes Conven-
tional Implicatures (CI) as follows (cf. Chapter 2).
138. Conventional Implicatures (CI)
1. CIs are part of the conventional meaning of words.
2. CIs are commitments, and thus give rise to entailments.
3. These commitments are made by the speaker of the utterance 'by vir-
tue of the meaning of' the words he chooses.
4. CIs are logically and compositionally independent of what is 'said (in
the favor sense)', i.e. independent of the at-issue entailments.
(Potts 2005, example 2.10)
The first statement means that not-at-issue meaning is stored with a lexical entry
(and is not derived from how the word is used). The second statement means
that CIs are added to the Common Ground of a discourse. The third one states
that CIs are always speaker-oriented. The final statement means that CIs and at-
issue meaning do not depend on each other.
While affected experiencer applicative arguments have some of these char-
acteristics, Condition 3 is not met by the affected experiencer. The affected ex-
periencer construction is not speaker-oriented. This can be seen in (139) in
which the affected experiencer is embedded in an attempt to attribute the not-at-
issue meaning to someone other than the speaker.
139. Jan glaubt fälschicherweise, dass Alex Chris
Jan.Nom believes wrongly that Alex.Nom Chris.Dat
Bens Vase zerbrochen hat, dabei ist das Chris
Ben.Gen vase broken has but is that Chris.Dat
egal.
indifferent
'Jan wrongly believes that Alex has broken Ben's vase on Chris but re-
ally Chris is indifferent about this.'
In this sentence, it is not the speaker but rather Jan who believes that Chris is
affected. In other words, affected experiencers do not project to the highest point
as CIs do. Rather they can be embedded and be attributed to someone other than
the speaker.
Potts (2005) addresses phenomena that meet all requirements of the CI ex-
cept speaker orientation briefly under the term "(merely) multi-dimensional"
elements. He shows that but is such an element and suggests that additive
modifers (e.g. too, also) and exceptive constructions (no X but Y) also fall into
this category. Thus, affected experiencers should also belong to that group.


  73

However, condition 4 is also not met by affected experiencers. At-issue meaning


is used in computing the not-at-issue meaning. As Potts (2005) puts it: "Where
we end up after removing [Condition 4] is not fully determined" (p. 217). Fur-
thermore, affected experiencers contribute both at-issue and not-at-issue mean-
ing (section 3.3). Clearly, affected experiencers are not CIs as defined by Potts.
In fact, Potts does not allow for mixed meaning contribution at all. To allow
for these elements with mixed contribution seems, however, to be a necessary
change to the system of two tiers of meaning. Bosse et al. (2012) show that in
three languages affected experiencer applicative arguments contribute both
kinds of meaning. Furthermore, McCready (2009) finds other elements (of Japa-
nese) that contribute both types of meaning.
Furthermore, Roberts et al. (2009) argue that there are different types of not-
at-issue contributions. They behave differently with respect to the projection test
(including the family of sentence tests). Roberts et al. (2009) present four catego-
ries of not-at-issue elements with different behaviors in the projection tests (and
they indicate that other categories might exist). Affected experiencers seem to fit
best into their category of locally entailed projective meaning. This category is
characterized as projecting its not-at-issue meaning always locally and being
cancelable and often informative. This description also seems to characterize
affected experiencers well but a detailed analysis in Roberts et al.'s system is
required to confirm this categorization of affected experiencers and to explain
why the not-at-issue meaning can only be detected by some tests and not others.
3.9 Digression: Parametric Variation of Aff
In this section, I provide a digression from German and English applicative ar-
guments to discuss parametric variation observed with respect to the Aff head.
Two types of variation will be addressed: first, I use Japanese to show that Aff
can attach outside of Voice in that language and discuss the differences of that to
German and English (section 3.9.1). The second variation I address shows that
there are languages in which Aff contributes only not-at-issue meaning but is
otherwise identical to the Aff head presented for German and English (section
3.9.2). In section 3.9.3, I briefly discuss possible cross-linguistic parametric var-
iation of Aff.

3.9.1 Japanese: Attachment Height Variation


Japanese has affected experiencers (140) that resemble English and German
ones (41) in some respects but also differ from them in some ways.26 Here, I

26 For instance, the affected experiencer is nominative-marked. For a detailed discussion
on how these applicatives can be identified in Japanese, see Bosse et al. (2012).

 
74   

will discuss how these similarities and differences follow from allowing for para-
metric variation in the attachment height of Aff.
140. Sachi-ga Masa-ni Aiko-no kabin-o kowas-are-ta.
Sachi-Nom Masa-by Aiko-Gen vase-Acc break-Aff-Past
'Masa broke Aiko's vase on Sachi.'
= Masa broke Aiko's vase, and ths matters to Sachi.
(Bosse et al. 2012, example 1c)
This Japanese construction, often called an “adversity passive”, patterns with
German and English in terms of at-issue and not-at-issue meaning. The affect-
ed experiencer is syntactically accessible and can be questioned (compare (141)
to (56)).
141. Dare-ga Masa-ni Aiko-no kabin-o kowas-are-ta no?
who-Nom Masa-Dat Aiko-Gen vase-Acc break-Add-Past Q
'On whom did Masa break Aiko's vase?'
(Bosse et al. 2012, example 32c)
This shows the at-issue contribution of the affected experiencer. This can also
be seen from if-clauses. Compare (142a, b) and (57).
142. a. Moshi Hanako-ga Taroo-no odor-are-tara,
if Hanako-Nom Taro-Dat dance-Aff-Cond
sekai-ga owar-u.
world-Nom end-NonPast
'If Hanako is affected by Taro's dancing, the world ends.'
b. Moshi Taroo-ga odot-tara, sekai-ga
if Taro-Nom dance-Cond world-Nom
owar-u.
end-NonPast
'If Taro dances, the world ends.'
According to (142a), Hanako must be affected by Taro’s dancing for the
world to end. In contrast, the world ends according to (142b) as long as Taro
dances (regardless of Hanako being affected or not). Thus, the affected expe-
riencer makes a difference in the truth conditions of the conditional. This is
expected for at-issue meaning.
The not-at-issue contribution can be seen from negation (143) in which the
affectedness survives (just like in German and English (54)).


  75

143. Sachi-wa Masa-ni Aiko-no kabin-o


Sachi-Top Masa-Dat Aiko-Gen vase-Acc
kowas-are-nakat-ta.
break-Aff-Neg-Past
'Masa didn't break Aiko's vase on Sachi.'
1. Masa didn't break Aiko's vase (but if he had, it would have mat-
tered).
2. *Masa broke Aiko's vase, but it didn't matter to Sachi.
(Bosse et al 2012, example 30c)
Negation cannot target the experiencing event independently. It can only be
used to express that the whole event did not take place. This is the pattern
expected for not-at-issue meaning.
The not-at-issue contribution is also reflected in the yes/no-question (144),
just like in German and English (55).
144. Sachi-wa Masa-ni Aiko-no kabin-o kowas-are-ta
Sachi-Top Masa-Dat Aiko-Gen vase-Acc break-Aff-Past
no?
Q
'Did Masa break Aiko's vase on Sachi?' (If Masa broke it, it would mat-
ter.)
(Bosse et al. 20212, example 31b)
The answer to this question cannot be a simple “no” if it is meant to negate the
experiencing event. A simple “no”-answer can only be used for indicating that
the whole event did not take place.
Finally, the affected experiencer in Japanese also needs to be sentient
(Kubo 1992, Dubinsky 1997), just like in German (47) and English (44).
145. #Iwa-ga ame-ni fur-are-ta.
rock-Nom rain-Dat fall-Aff-Past
'The rock had rain fall on it.'
(Kubo 992, example 21a)
However, Japanese differs from English and German (70) in that the
agent matters to the affected experiencer. This can be seen from the native
speaker intuition about what the source of the affecting is (146).
146. Sachi-wa Masa-ni Aiko-no kabin-o kowas-are-ta.
Sachi-Top Masa-Dat Aiko-Gen vase-Acc break-Aff-Past
'Sachi had Masa break Aiko's vase on him and …
1. … it mattered to Sachi because it was Aiko's vase.'
2. … it mattered to Sachi because Masa did it.'
(Bosse et al 2012, example 66)

 
76   

For the formal analysis, this means that the agent must be included in the source
event of the experiencing event.
C-command tests indicate that the subject is c-commanded by the affected
experiencer in Japanese. The affected experiencer c-commands the agent and
can bind it (147a). Binding from the agent into the affected experiencer argu-
ment is not possible (147b).
147. a. [Go-non-ijoo-no kodomo]1 -gaAff [sono1-ko-tachi-no
five-CL-more.than-Gen child Nom it-child-PL-Gen
hahaoya-ni odor-are-ta.
mother-Dat dance-Aff-Past
'More than five children1 has his/her1 mother dance on him/her1'
(Literally: His/her1 mother danced on more than five children1.)
b. *[Sono1-ko-tachi-no hahaoya]-gaAff
it-child-PL-Gen mother-Nom
[go-nin-ijoo-no kodomo]-ni odor-are-ta.
five-CL-more.than-Gen child-Dat dance-Aff-Past
'His/her1 mother had more than five children1 dance on her.'
(Intended: 'More than five children1 danced on his/her1 mother.')
Syntactically, this means that Aff must attach outside VoiceP in Japanese. As a
consequence, it will also automatically include the agent in the source event,
as shown in the following derivation.27


27 The full semantic derivation is given in Appendix 1, #4.


  77

148. a. Sachi-ga Masa-ni Aiko-no kabin-o kowas-are-ta.


Sachi-Nom Masa-by Aiko-Gen vase-Acc break-Aff-Past
'Masa broke Aiko's vase on Sachi.'
b. AffP<v,t>
3
Sachi Aff'
3
VoiceP<v,t> Aff<vt,evt>
3
Masa Voice'
3
VP<v,t> Voice
6
Aiko-no kabin-o kowas-
c. [[AffP]] = λe. BREAK(e) & Thm(e)(Aiko's vase) & Agt(e)(Masa)
& e'. (EXPER(e') & Exp(e')(Sachi) : e''((BREAK(e'') &
Thm(e'')(Aiko's vase) & Agt(e'')(Masa)) → Source(e'')(e')
(Bosse et al. 2012, modified examples 70, 71)
The denotation (148c) states that there is a breaking event of Aiko’s vase and
Masa is the agent of that event. This whole event is the source event of
Sachi’s experience.
When modified with again, Japanese should allow a reading which includes
the agent but excludes the affected experiencer because the AffP constituent in-
cludes VoiceP (148). These predictions are correct (149).
149. Sachi-ga Masa-ni sono-pazuru-o mata kowas-are-ta.
Sachi-Nom Masa-Dat that-puzzle-Acc again break-Aff-Past
'Sachi had Masa break the puzzle on her again.'
1. Someone broke the puzzle before and now Masa did it again. (re-
peated VP)
2. Masa broke the puzzle before and now he did it again, this time af-
fecting Sachi. (repeated VoiceP)
3. Masa had broken the puzzle on Sachi before and now he did it again.
(repeated AffP)
4. *Someone broke the puzzle on Sachi before and now Masa did it on
her again. (corresponds to German (81-2))
Thus, again-modification supports the attachment height distinction proposed
for Japaneses compared to German and English.
In summary, Japanese has an overt Affect morpheme, namely -rare. The on-
ly other difference between German and Japanese is the attachment height of
Aff (below or above Voice). This can be seen as a parameter of variation (i.e.

 
78   

a selectional restriction of Aff and which projection it requires). Allowing for


different attachment heights of Aff cross-linguistically can explain the differ-
ent native speaker intuitions ((146) vs. (70)) as well as the similarities of af-
fected experiencers across languages in terms of meaning and (not-)at-issue
contribution. As will be shown in the next subsection, there is one more pa-
rameter of variation influencing the typology of affected experiencer applica-
tive arguments, namely that cross-linguistically there are some affected experi-
encers that contribute only not-at-issue meaning.

3.9.2 Not-At-Issue Affected Experiencers


In this subsection, I discuss affected experiencer arguments that are similar in
their meaning to the ones discussed for German, English and Japanese above
but differ from those in their truth-conditional contribution, as these contribute
only not-at-issue meaning.
These not-at-issue affected experiencer applicative arguments can be found
in languages such as Hebrew (150a) and French (150b).
150. a. hem kol ha-zma mitxatnim li
they all the-time marry to-me
'They are getting married on me all the time (and it bothers me).'
(Borer and Grodzinsky 1986, examples 9a)
b. Les gosses lui ont gribouillé sur tous
the kids her.Dat have scribbled on all
les murs
the walls
‘The kids scribbled all over the walls on her.’
(Authier and Reed 1992, example 3a)
Just like the affected experiencers in English, German and Japanese dis-
cussed previously, the affected experiencer applicative arguments in Hebrew
and French denote an individual who is psychologically affected by the event
described in the sentence (Borer and Grodzinsky 1986, Authier and Reed
1992).
The main point of discussing these applicatives here is to show that cross-
linguistically applicatives with essentially the same meaning can behave differ-
ently. This behavior can be explained by parametric variation. I do not discuss
here how affected experiencer applicatives in Hebrew and French can be distin-
guished in these languages (for that, see Authier and Reed (1992), Borer and
Grodzinsky (1986), Berman (1982), Bosse et al. (2012) among others).
As expected under the definition of applicative arguments (4), these affected
experiencers are optional elements and the denoted individual is not entailed
as being involved in the event. Below, I first show that the meaning of these is


  79

contributed only on the not-at-issue tier of meaning (section 3.9.2.1). Then I


give the analysis for these applicatives and relate them to their German, English
and Japanese counterparts (section 3.9.2.2).
3.9.2.1 Not-At-Issue Meaning
The meaning contributed by the affected experiencer applicative argument in
Hebrew and French (150) is exactly like that contributed in German, English
and Japanese, namely that the denoted individual is affected by the event.
Typically this is a negative effect on the individual but it does not have to be.
In Hebrew and French, this meaning is contributed completely on the not-at-
issue tier of meaning, unlike in the languages discussed in previously. This can
be seen by applying the tests introduced in Chapter 2 . First, the notion of
affectedness survives under negation (151).
151. a. Rina lo lavsha li simla megunderet
Rina Neg put.on to.me dress fancy
'Rina did not put on a fancy dress on me.'
1. Rina did not put on a fancy dress (but if she had I would
have found it objectionable).'
(Bosse et al. 2012, example 81)
2. *Rina put on a fancy dress and it did not affect me.
b. Heureusement que les invites lui ont
fortunately that the guests 3S.Dat have
pas mangé tout ce qu'il y avait dans le
not eaten all that there was in the
frigo
fridge
‘Fortunately, the guests did not eat everything in the fridge on
him/her.’
1. Fortunately, the guests did not eat everything in the fridge on
him/her.’ (If they did, it would matter to him/her.)
(Bosse et al. 2012, Example 91)
2. *Fortunately, the guests ate everything in the fridge and it did not
matter to him/her.
Just like in German (54a), English (54b) and Japanese (143), in Hebrew
(151a) and French (151b) the notion of affectedness itself cannot be targeted
by negation. This indicates that the affected experiencer applicative argument
contributes not-at-issue meaning.
The notion of affectedness also projects past yes/no-questions in Hebrew
(152a) and French (152b), similar to the not-at-issue meaning of English
and German affected experiencers (55), as well as Japanese ones (144).

 
80   

152. a. Rina pit'om lavsha lo simla megunderet?


Rina suddenly put.on to.him dress fancy
‘Did Rina suddenly put on a fancy dress on him?’
(Bosse et al. 2012, example 82)
b. Est-ce-que les invites lui on mangé tout
is-it-that the guests 3S.Dat have eaten all
ce qu'il y avait dans le frigo?
that there was in the fridge
‘Did the guests eat everything in the fridge on him/her?’ (if they
did, it would matter to him/her)
(Bosse et al. 2012, example 90)
The listener cannot just answer “no” to the questions in (152) to convey the
meaning that the event happened but that it did not matter to the individual
denoted by the affected experiencer. For this meaning, further explanations
are required. With a “no” answer, the truth of the affected experiencer refer-
ent being affected by the event is accepted. This supports the claim that the af-
fected experiencer construction in Hebrew and French involves not-at-issue
meaning.
As expected for not-at-issue meaning (and unlike affected experiencers in
German (56a), English (56b) and Japanese (141)), the affected experiencer
cannot be questioned but is syntactically inert in Hebrew (153a) and French
(153b).
153. a *lə-mi hem mitxatnim kol ha-zman?
to-who they marry all the-time
'On who do they marry all the time?'
(Borer and Grodzinsky 1986, example 11)
b. *A qui as-tu gribouillé sur tous les murs?
to who have-you scribbled on all the walls
‘Who did you scribble all over the walls on?’
(Authier and Reed 1992, example 15b)
This syntactic inertness further indicates that the affected experiencer contri-
butes meaning to the not-at-issue tier in Hebrew and French.
Finally, the affected experiencer does not contribute to the conditions of
an if-clause in Hebrew and French. Again, this is expected for elements con-
tributing only not-at-issue meaning (Chapter 2 ) but was not observed for
German (57a), English (57b) and Japanese (142) affected experiencers.
154. a. Im hem yitxatnu li, Rina titen la'hem 100dolar.
if they will.marry to.me Rina will.give to.them 100dollar
‘If they marry (on me), Rina will give them $100.’
(Bosse et al. 2012, example 84)


  81

b. Si Elmer lui dévalise deux banques, je


if Elmer 3S.Dat rob two banks I
te donnerai $100.
you will.give $100
‘If Elmer robs two banks (on him/her), I will give you $100.’
(Bosse et al. 2012, example 92)
In (154a), Rina will have to pay if they get married regardless of whether that
affects me or not. In the same fashion, I will have to pay according to
(154b) as long as Elmer robs two banks. It does not need to affect the third
person. This behavior is associated with not-at-issue meaning. This not-at-
issue behavior cannot be attributed to the fact that the affected experiencers
are clitics in these languages. Clitics in general are possible in conditional sen-
tences and typically contribute to the truth conditions. Thus, being a clitic is
not the reason for them not contributing to the truth conditions in (154). Just
like in English and German affected experiencer if-clauses (57), the Hebrew
and French affected experiencer clitics in these sentences could contribute to the
truth conditions but these Hebrew and French elements do not (154). Thus, they
are special in their not-at-issue contribution.
Overall, all of these tests show that in French and Hebrew affected experi-
encer applicative arguments contribute only not-at-issue meaning (also see
Gutzmann (2007) for speculations about Hebrew not-at-issue clitics). In this
way, affected experiencers in Hebrew and French differ from affected experi-
encers in German, English and Japanese.
3.9.2.2 Analysis
For the analyses of German and Japanese affected experiencers, the denota-
tion for Aff in (168a) was used. It projects the affected experiencer in the asser-
tion. In Hebrew and French, the entire content of Aff must be on the not-at-issue
tier of meaning to be consistent with the findings of the previous section. Conse-
quently, the Aff head used to account for these projects the experiencer argument
and the experiencing event as not-at-issue meaning together with the source
event (168b).
155. a. Experiencer on the at-issue tier
[[Aff1]] = λPvt.λx.λe. P(e) & e'(EXPER(e') & Exp(e')(x)): e''(P(e'') →
Source(e'')(e'))
b. Experiencer on the not-at-issue tier
[[Aff2]] = λPvt.λx.λe. P(e) : e'(EXPER(e') & Exp(e')(x)) & e''(P(e'') →
Source(e'')(e'))

 
82   

The two denotations for Aff differ solely in on which tier of meaning the ex-
periencing event and the experiencer are projected (as reflected in the position
of the colon).
In section 3.9.1, it was also shown that Aff could either attach between
VP and Voice (German/ English) or above Voice (Japanese). This could be seen
from different c-command relations between the external argument and the ap-
plicative argument. This parameter of attachment height is independent of the
type of meaning (at-issue or not-at-issue) contributed. However, in Hebrew
and French, it is not possible to check c-command relations between the ex-
ternal argument and the applicative argument because the applicative argument
may not be a full NP (156).28
156. a. i. hem kol ha-zma mitxatnim li
they all the-time marry to-me
‘They are getting married on me all the time (and it bothers
me).’
(Borer and Grodzinsky 1986, examples 9a)
ii. *hem mitxatnim lə-Rani kol ha-zman
they marry to-Rani all the-time
'They are getting married all the time on Rani.'
(Borer and Grodzinsky 1986, examples 10a)
b. i. Les gosses lui ont gribouillé sur tous
the kids her.Dat have scribbled on all
les murs
the walls
'The kids scribbled all over the walls on her.'
ii. *Les gosses ont gribouillé sur tous les
the kids have scribbled on all the
murs à Marie.
walls to Marie
'The kids scribbled all over the walls on Marie.'
(Authier and Reed 1992, example 3)
Consequently, it is impossible to obtain c-command data to decide the attach-
ment height of the Aff head in these languages. Yet, we can take native speaker
intuitions of what is part of the source event as an indicator for the attachment
height (cf. German / English (70) and Japanese (146)). In Hebrew, this test
yields that Aff must attach outside of VoiceP since the external argument can
be the source of the experiencing event (157).

28
There is also no prepositional variant in which a full NP could be acceptable
(cf. Authier and Reed 1992).


  83

157. Rina lavsha li simla megunderet.


Rina put.on to.me dress fancy
'Rina put on a fancy dress on me.'
= Rina put on a facny dress and …
1. … it aggravates me that Rina did it.
2. … it aggravates me because it is a dress.
3. … it aggravates me because the dress is fancy.
4. … it aggravates me that she put on the dress.
(Bosse et al 2012, example 86)
The derivation for the Hebrew sentence (158a) procedes as follows (ab-
stracting away from the modifier all the time).29
158. a. hem kol ha-zma mitxatnim li
they all the-time marry to-me
'They are getting married on me all the time (and it bothers me).'
(Borer and Grodzinsky 1986, examples 9a)
b. AffP<v,t>
3
li Aff'
3
Aff<vt,evt> VoiceP<v,t>
3
hem Voice'
3
Voice<vt,evt> VP<v,t>
6
mitxatnim
c. [[AffP]] = λe. MARRY(e) & Agt(e)(they) : e'(EXPER(e') &
Exp(e')(me)) & e'' ((MARRY(e'') & Agt(e'')(they)) →
Source(e'')(e'))
(Bosse et al. 2012, slightly modified examples 87, 88)
The assertion in (158) is only that they are getting married. The meaning that
the event mattered to me is present only on the not-at-issue tier of meaning.
The difference to Japanese (in which Aff also attaches outside of VP) is that
the experiencing event with the experiencer is on the at-issue tier in Japanese
(148b) but not in Hebrew (158).
The situation is different for French. For sentence (150b), Authier and Reed
(1992) report that the individual denoted by the affected experiencer is aggra-
vated by the whole event. Yet, they state that the kids’ involvement in the ag-

29 The full semantic derivation is given in Appendix 1, #5.

 
84   

gravating is a pragmatic inference. We take this to indicate that the agent is


not in fact included in the source event (but can be made relevant by the con-
text, just like in German and English). For the analysis, this means that Aff
needs to attach between VP and Voice in French, just like in German and
English. This derivation is given in (159).30
159. a. Les gosses lui ont gribouillé sur tous
the kids her.Dat have scribbled on all
les murs
the walls
‘The kids scribbled all over the walls on her.’
(Authier and Reed 1992, example 3a)
b. VoiceP
3
les gosses Voice'
3
Voice AffP<v,t>
3
lui Aff'
3
Aff<vt,evt> VP
6
ont gribouillé sur tous les murs
c. [[VoiceP]] = λe. SCRIBBLE(e) & Thm(e)(the walls) & Agt(e)(the
kids) : e' (EXPER(e') & Exp(e')(her)) & e'' (P(e'') →
Source(e'')(e'))
The at-issue meaning in (159) is that there is a scribbling event with the
Theme the walls and the Agent the kids. There is also an experiencing event
of which s/he is the experiencer; this event is contributed on the not-at-issue
tier of meaning. The source of the experiencing event is the scribbling on the
wall (excluding the agent).
This is where the difference from Hebrew becomes apparent; in French
(159), the agent is not part of the source event but in Hebrew (158) it is. Fur-
thermore, the difference between French on the one hand and German and
English on the other is the experiencing event and the experiencer being on
the at-issue tier in the latter two languages (69, 72) but not in the former (159).
The difference between French and Hebrew is also reflected in again-
modification. The predictions for French are the same as for German (80),
namely that again can modify VP, AffP and VoiceP but cannot modify the

30 The full semantic derivation is in Appendix 1, #6.


  85

verbal event with the same agent but a different affected experiencer. These
predictions are borne out.31
160. Les gosses lui ont enore gribouillé tous les murs.
the kids her.Dat have again scribbled on the walls
'The kids have scribbled on the walls again on her.'
1. VoiceP: The kids had scribbled on the walls on her before and now
they did it again.
2. AffP: First, the neighbors had scribbled on the walls on her and now
the kids scribbled on the wall on her again.
3. VP: The walls were scribbled on when the family moved in, they
were renovated and now the kids scribbled on the walls again.
4. unavailable: First, the kids had scribbled on the walls on their father
and now the kids scribbled on the walls again, this time on her.
As expected, encore 'again' cannot be used to modify just the experiencing
event either.
161. Les invités lui ont mangé tout
the guests her.Dat have eaten all
ce qu'il y avait dans le frigo et les gosses
that there was in the fridge and the kids
lui ont encore gribouillé tous les murs.
her.Dat have again scribbled all the walls
'The guests ate everything on her and the kids scribbled all over the
walls on her again.'
This sentence cannot be used to indicate that she was affected twice, first by
the eating of everything and then by the scribbling on the wall. Rather, it is
only acceptable if the scribbling on the wall has happened before (affecting
her) and now it happened again.
The predictions of again-modification for Hebrew are the same as for Ja-
panese (149), repeated below. Again should be able to modify the VP, VoiceP,
and AffP.
162. Again-Modification Predictions
1. VP attachment: a reading where the VP event has taken place be-
fore, without an affected experiencer and not necessarily with the
same agent;
2. VoiceP attachment: a reading where the VoiceP event has taken
place before with the same VP event and the same agent, but without
an affected experiencer;


31 Thanks to Joelle Malou for these judgments!

 
86   

3. AffP attachment: a reading where the whole AffP event has taken
place before with the same VP event, the same agent, and the same
affected experiencer.
The reading that should not be available is where the event happens again
with the same affected experiencer and a different agent. These predictions
are essentially borne out. However, Hebrew shuv 'again' does not allow the
the VP attachment (regardless of the presence of the affected experiencer).32
163. Rina shuv lavsha li simla megunderet.
Rina again put.on to.me dress fancy
'Rina put on a fancy dress on me again.'
1. VP attachment: * Yesterday Rachel put on a fancy dress, and no-
body cared. Today, Rina put on a fancy dress on me again.
2. VoiceP attachment: Yesterday Rina put on a fancy dress, but that didn’t
bother me. Today, Rina put on a fancy dress on me again.
3. AffP attachment: Yesterday Rina put on a fancy dress and it both-
ered me. Today, Rina put on a fancy dress on me again.
4. unavailable: Yesterday, Rachel put on a fancy dress on me. Today,
Rina put on a fancy dress on me.
While the VoiceP attachment reading of shuv 'again' is not easily available, it
is not ruled out and therefore provides weak support for the high attachment of
Aff in Hebrew.
In summary, the discussion of Hebrew and French affected experiencer ap-
plicative arguments has shown that in these languages the relevant applicative
behaves just like the equivalent in German, English and Japanese, except that all
of the meaning is projected on the not-at-issue tier of meaning.

3.9.3 Parametric Variation


Affected experiencer applicative arguments are crosslinguistically expected to
take one of the forms shown in Table 1 below. The first variation is whether
the Affect head attaches between VP and Voice (low attachment) or outside
of VoiceP (high attachment). The second variation is whether the experiencer is
contributed on the at-issue tier of meaning or on the not-at-issue tier.


32 Thanks to Idan Landau for these judgments.


  87

Table 1: Typology of Affected Experiencers

high attachment low attachment


experiencer at-issue Japanese German, English
experiencer not-at-issue Hebrew French
(Bosse et al. 2012, Table 3)

The difference between German and English is that English requires a preposi-
tion as an overt Aff head. This causes a minor variation of the semantic type of
Aff.
So far, there is no affected experiencer applicative argument that contributes
only at-issue meaning. However, if that is a possible variation for affected expe-
riencers, it is expected to also have a variation in attachment height. The predic-
tions for such an affected experiencer applicative argument are very clear based
on this chapter: the affected experiencer should be syntactically accessible (as
all at-issue elements are) and should follow the prediction for again-
modification depending on the attachment height, as explained above.
3.10 Conclusion
In this chapter, I have discussed affected experiencer applicative arguments
which are characterized as denoting a sentient individual who is psychologically
affected by an event. I have shown how they can be distinguished from other
applicative arguments in English and German. In both languages, the affected
experiencer applicative argument contributes to both the at-issue tier and the
not-at-issue tier of meaning. This conclusion is reflected in the denotation of the
Aff head. It contributes meaning to both tiers. In German, this is a phonological-
ly null head; in English, it is realized as the preposition on (164).
164. a. [[Aff]] = λPvt. λx. λe. P(e) & e'(EXPER(e') & Exp(x)(e')): e"
(P(e") → Source(e'')(e'))
b. [[on]] = λx. λPvt. λe. P(e) & e'(EXPER(e') & Exp(x)(e')):
e"(P(e") → Source(e'')(e'))
In both languages, this head is attached between VP and Voice. Support for the
proposed analysis comes from again-modification and adverbials. These two
also support that the analysis includes two event variables. Cross-linguistic vari-
ations of Aff with respect to the meaning contribution and attachment height
were discussed in this chapter.
The German pertinence dative was shown to be the affected experiencer
construction with a conversational implicature. The latter contributes the posses-
sion interpretation of the pertinence dative. This interpretation is more easily

 
88   

available with relational nouns, but can also be found with sortal nouns. Apart
from this conversational implicature, pertinence datives and affected experi-
encers are identical.
In the next chapter, I discuss those applicative arguments of German and
English that contribute only not-at-issue meaning.


4 Not-At-Issue Applicative Arguments
4.1 Introduction
As shown in section 3.9.2, Hebrew and French have applicative arguments that
contribute only not-at-issue meaning. In this chapter, I discuss such applicative
arguments with only not-at-issue for German and English. Both of these lan-
guages have subject co-referential applicative arguments (166) that contribute
only not-at-issue meaning. In addition, German has a so-called ethical dative
not-at-issue applicative (165). These constructions share their not-at-issue con-
tribution and their form: they are all weak pronouns. After investigating each of
these types of applicative arguments individually, I will briefly explore the rela-
tionship of being an applicative argument with only not-at-issue meaning and
being a weak pronoun in section 4.4. This also includes a discussion of their sta-
tus in the system of CIs presented by Potts (2005).
165. Komm mir pünktlich nach Hause!
come me.Dat on.time to home
'Come home on time and I want this to happen!'
166. a. I'm gonna eat me some apple pie.
b. Ich trinke mir jetzt einen Kaffee.
I drink me.Dat now a.Acc coffee
'I'm gonna drink me some coffee now.'
(Horn 2009, example 31a)
First, I discuss German ethical datives (165) in section 4.2. Gutzmann
(2007) argues that these contribute only not-at-issue meaning. While I agree
with this, I will provide a different analysis than Gutzmann by taking their re-
stricted occurrence into account. Second, I analyze subject co-referential appli-
cative arguments of English (166a) and German (166b) in section 4.3. Horn
(2009) proposes that these applicatives contribute not-at-issue meaning. I pro-
vide a formal analysis of them, which is missing from his account. My investi-
gation of not-at-issue applicative arguments is concluded in section 4.5.
4.2 Ethical Dative
In this section, I discuss a German applicative argument that contributes only
not-at-issue meaning, namely the ethical dative ("Dativus Ethicus") (167).
167. Du sollst mir dem Papa die Schuhe putzen.
you shall me.Dat the.Dat dad the.Acc shoes clean
'You shall clean the shoes for dad and I want this to happen.'
As the translation indicates, the dative-marked pronoun denotes that the referent
of the applicative argument has a special interest in the proposition to come true.


90   

In other words, the individual referred to by the ethical dative wants the event to
take place but the individual is not otherwise involved in the actual event (see
Abraham (1973), Wegener (1989), Gutzmann (2007) and references cited there-
in).33
I describe the ethical dative including its occurrences and status as an ap-
plicative argument in section 4.2.1 before discussing the analysis of ethical da-
tives as not-at-issue elements provided by Gutzmann (2007) in section 4.2.2. In
that section, I also provide my analysis which builds on Gutzmann's and We-
gener's (1989) insights.

4.2.1 Description
In this section, I describe the characteristics of the ethical dative in detail. I limit
myself to the type that Wegener (1989) discusses as Aufforderungs-Ethicus
'order ethical dative' (167) because, as Wegener and Gutzmann (2007) state, it is
possible that the second type, the Ausrufe-Ethicus 'exclamation ethical dative'
(168), is just a variation of the former one.
168. Wie besoffen der dir war!
how drunk that.one.Nom you.Dat was
'That one was so extremely drunk (you know)!'
(Wegener 1989, example 22d, my translation)
Furthermore, the exclamation ethical dative is rarely used in contemporary
German and is not accepted by all speakers (Gutzmann 2007).
The ethical dative mir 'me' in (167) meets my definition of an applicative ar-
gument as given in (4): it is an optional element which is not selected by the lex-
ical verb or a preposition. The sentence is grammatical without it, as (169)
shows.
169. Du sollst dem Papa die Schuhe putzen.
you.Nom shall the.Dat dad the.Acc shoes clean
'You shall clean the shoes for dad.'
Also, (169) does not entail that there is another individual involved in the event.
Thus, the requirements for an applicative argument (4) are met.
The ethical dative is easily distinguishable from other applicative arguments
of German because of its form: it must be a first person singular pronoun in-
flected for dative case: mir 'me'. The following example shows that replacing the
pronoun of (169) with a full (dative) NP or other pronouns results in ungram-
maticality (cf. Wegener 1989, Gutzmann 2007).


33 English does not have an applicative argument with a similar meaning. Consequently,
"and I want this to happen" is used in the translation.


  91

170. *Du sollst der Mutter/ ihr/ ihm/


you.Nom shall the.Dat mother/ her.Dat/ him.Dat
ihnen dem Papa die Schuhe putzen.
them.Dat the.Dat dad the.Acc shoes clean
'You shall clean the shoes for dad, and the mother/ she/ he/ they want(s)
this to happen.'
None of the other German applicatives have a similar requirement (see affected
experiencers (Chapter 3), subject co-referential applicative (section 4.3), (true)
benefactives (Chapter 5), part-whole (Chapter 6), Dative of Inaction (Chapter
7)). As I show in section 4.2.1.2, the meaning of the ethical dative is contributed
as not-at-issue meaning. In this, the ethical dative also differs from other Ger-
man applicative arguments which contribute (some) at-issue meaning (all except
the subject co-referential applicative argument). Furthermore, the ethical dative
is highly restricted in the type of sentences it can occur in (section 4.2.1.1). No
other German applicative argument has similar restrictions in requiring a sen-
tence to have a certain force.
4.2.1.1 Distribution
The occurrence of the ethical dative is limited to directive sentences (Wegener
1989), such as declaratives with deontic force (167) and imperatives (171a). As
(171b-d) show the ethical dative is ungrammatical in simple declaratives (with-
out deontic force) and in questions.34
171. a. Paulinchen, geh mir nicht ans Feuerzeug!
Paulinchen go me.Dat Neg at.the.Acc lighter
'Paulinchen, do not touch the lighter (and I want that)!'
(Wegener 1989, example 1, my translation)
b. Paulinchen geht mir immer ans
Paulinchen.Nom goes me.Dat always at.the.Acc
Feuerzeug.
lighter
*'Paulinchen always touches the lighter and I want that.'
'Paulinchen always touches the lighter on me.'
c. Wann bist du (*mir) gestern nach Hause gekommen?
when are you.Nom me.Dat yesterday to home come
'When did you come home yesterday (*and I wanted this to hap-
pen)?'
(Wegener 1989, example 10b, my translation)


34 (171c) and (171d) are grammatical as e.g. benefactives.

 
92   

d. Bist du (*mir) pünktlich nach Hause


are you.Nom me.Dat on.time to home
gekommen?
come
'Did you come home on time (*and I wanted this to happen)?'
(Wegener 1989, example 10b, my translation)
In this extreme restriction to certain sentence types, the ethical dative resembles
a discourse particle more than an applicative argument (cf. Wegener 1989, Car-
dinaletti and Starke 1994). For instance, the particle denn can only occur in
questions (Wegener 1989):
172. a. Wo bleibt er denn so lange?
where stays he.Nom Prt so long
'What in the world is taking him so long?'
b. Er war (*denn) so lange im Garten.
he.Nom was Prt so long in.the backyard
'He spent a long time in the backyard.'
(Wegener 1989, example 10a, my translation)
Since the ethical dative can only be the first person pronoun (170), naturally
it cannot appear in a coordinated structure because there is no other element it
could coordinate with (Gutzmann 2007).
173. a. *Du sollst deiner Schwester und mir
you.Nom shall your.Dat sister and me.Dat
dem Papa die Schuhe putzen.
the.Dat dad the.Acc shoes clean
'You shall clean the shoes for dad, and your sister and I want this to
happen.'
b. *Du sollst ihr und mir dem
you.Nom shall her.Dat and me.Dat the.Dat
Papa die Schuhe putzen.
dad the.Acc shoes clean
'You shall clean the shoes for dad, and she and I want this to hap-
pen.'
Given that the ethical dative has to be a pronoun, it is also not surprising that
it follows the word order rules for pronouns. In German, pronouns tend to occur
in the left middle field, and this is also true for the ethical dative (Gutzmann
2007).35


35 Gutzmann (2007) (contrary to (Wegener 1989)) accepts the ethical dative in (depend-
ent) verb-final sentences. I agree with his judgment.


  93

174. (Ich möchte,) dass du (mir) dem


I.Nom want that you.Nom me.Dat the.Dat
Papa (*mir) keinen Ärger machst!
dad me.Dat no.Acc trouble make
'(I want) that you don't cause any trouble on/for dad, and I want this to
happen!'
(Gutzmann 2007, based on examples 3b/ footnote 2, my translation)
In this, the ethical dative contrasts with discourse particles like denn which are
less fixed in their position.
175. Was macht er (denn) seiner Frau
what makes he.Nom Prt his.Dat wife
(denn) schon wieder so lange im Garten?
Prt yet again so long in.the backyard
'What in the world is he doing in the backyard for so long again on his
wife?'
So far, we have seen that the ethical dative behaves like a pronoun in terms
of syntactic integration. However, it is not available for other syntactic trans-
formations that pronouns can usually participate in. For instance, the ethical da-
tive cannot be negated or contrasted (Wegener 1989, Gutzmann 2007).36
176. *Du sollst nicht mir (sondern Papa)
you.Nom shall Neg me.Dat but dad.Dat
pünktlich nach Hause kommen!
on.time to home come
'You shall come home on time and not me but your dad wants this to
happen!'
(Gutzmann 2007, example 2e, my translation)
The ethical dative cannot be the head of a relative clause (177) or be modi-
fied by an appositive (178) (Wegener 1989, Gutzmann 2007).
177. *Du sollst mir, die dich mehrfach
you.Nom shall me.Dat who you repeatedly
erinnert hat, dem Papa die Schuhe
reminded has the.Dat dad the.Acc shoes
putzen.
clean
'You shall clean the shoes for dad, and I who has reminded you several
times want this to happen.'


36 This sample sentence is acceptable on a benefactive interpretation "You shall be home
on time, not for me but for dad."

 
94   

178. *Du sollst mir, deiner Mutter, dem


you.Nom shall me.Dat your.Dat mother the.Dat
Papa die Schuhe putzen.
dad the.Acc shoes clean
' You shall clean the shoes for dad and I, your mother, want this to hap-
pen.'
In all of this, the ethical dative contrasts with other pronouns, e.g. when they are
used as a true benefactive applicatives (179).
179. a. Du sollst nicht mir (sondern Papa)
you.Nom shall Neg me.Dat but dad.Dat
ein Bild malen.
a picture paint
'You shall paint a picture not for me (but for dad).'
b. Du sollst mir, die dich mehrfach
you.Nom shall me.Dat who you.Acc repeatedly
gebeten hat, ein Bild malen.
asked has a picture paint
' You shall paint a picture for me who has repeatedly asked you to
do so.'
c. Du sollst mir, deiner Mutter, ein
you.Nom shall me.Dat your.Dat mother a
Bild malen.
picture paint
'You shall paint a picture for me, your mother.'
Another characteristic of the ethical dative is that it cannot be bound (180).37
180. *Nur ich will, dass mir meine Frau
Only I want that me.Dat my.Nom wife
pünktlich nach Hause kommt.
on.time to home comes
'Only I want that my wife comes home on time and I want that to hap-
pen.'
Lastly, the ethical dative must appear inside the clause that states the event
in which the speaker is interested (181).


37 The sample sentence is grammatical on a benefactive reading "Only I want that my wife
comes home on time for me".


  95

181. *Ich will mir, dass jeder von euch pünktlich


I.Nom want me.Dat that each of you on.time
nach Hause kommt.
to home comes
'I want each of you to come home on time and I want that to happen.'
This discussion has shown that in terms of form and position the ethical da-
tive behaves like any other (first person singular dative) pronoun but in terms of
syntactic accessibility the ethical dative is not available (unlike other pronouns).
Furthermore, the ethical dative is only acceptable in sentences with directive
force. In this, as well as its invariant form, it resembles a modal/ discourse parti-
cle. Wegener (1989) has consequently argued that the ethical dative is a modal
particle while Gutzmann (2007) argued that it is a (not-at-issue) pronoun. Before
discussing these analyses (in section 4.2.2.1), I show in the next subsection that
the ethical dative contributes only not-at-issue meaning.
4.2.1.2 Not-at-issue Meaning
By applying the tests for not-at-issue meaning (Chapter 2), it can be shown that
the ethical dative contributes only not-at-issue meaning. First, the ethical dative
does not contribute to the conditions of an if-clause (182).38
182. a. Komm mir pünktlich nach Hause, und
come me.Dat on.time to home and
du bekommst 100Euro!
you.Nom get 100Euros
'Come home on time (and I want this to happen) and you will get
100 Euros!'
(Bosse and Bruening 2011, example 37)
b. Komm pünktlich nach Hause, und du
come on.time to home and you.Nom
bekommst 100Euro!
get 100Euro
'Come home on time and you will get 100 Euros!'
The ethical dative does not change the conditions of the clause: if you are home
on time, you will get the money according to both (182a) and (182b). There is
no situation in which I have to pay the 100 Euros according to (182a) but not
according to (182b) or vice versa. This indicates that the meaning contributed by
the ethical dative is on the not-at-issue tier of meaning.


38 Since ethical datives can only occur in directive sentences, they cannot appear in what
would technically be an if-clause. Consequently, a semantic approximation to an if-
clause is used here.

 
96   

Further support that the ethical dative contributes only not-at-issue meaning
comes from negated sentences. In the negated sentence (183), the meaning of
the special interest by the speaker survives.
183. Du sollst mir jetzt nicht einschlafen.
you.Nom shall me.Dat now Neg fall.asleep
'You shall not fall asleep now, and that is what I want.'
*'You shall fall asleep now, and that is not what I want.'
As expected for not-at-issue meaning, the meaning contributed by the ethical
dative cannot be negated by itself.
This projection of not-at-issue meaning is also expected in yes/no-questions.
However, they cannot be used as a test for ethical datives because the ethical
dative requires that the sentence be a directive sentence. Thus, it cannot occur in
yes/no-questions (171d) or in wh-questions (171c). Regardless, the results of
negation (182) and the semantic equivalent to an if-clause (181) indicate that the
ethical dative contributes not-at-issue meaning. This is also Gutzmann's (2007)
conclusion (see section 4.2.2).

4.2.2 Analysis
In this section, I briefly explain the analyses of the ethical dative by Wegener
(1989) and by Gutzmann (2007) before presenting my analysis which combines
the strong points of both of these analyses.
4.2.2.1 Previous Analyses
The analysis by Wegener (1989) focuses on the ethical dative as being limited to
certain sentence types and its invariant form. Because of this, she argues that the
ethical dative is a (modal) particle rather than a (typical) pronoun. She claims
that the primary function of the ethical dative is illocutive but that it does not
change the illocutionary force of the sentence it is added to.39 Rather, the ethical
dative is used to make the speaker's interest in the event explicit.
In contrast, Gutzmann (2007) focuses on the ethical dative being a pronoun.
He argues that the German ethical dative is a not-at-issue pronoun. He shows
that ethical datives have all the features of CIs that Potts discusses (20). First,
the ethical dative is detachable. By replacing the ethical dative, its meaning is
not preserved (184). (184b) cannot be interpreted as an ethical dative (but only
as a benefactive).


39 Also see Abraham (1973) who claims that the (exclamation) ethical dative indicates a
"dialogue situation" (p. 12).


  97

184. a. Komm mir ja pünktlich nach Hause!


come me.Dat Prt on.time to home
'Come home on time and I want that to happen!'
b. (#)Komm für mich pünktlich nach Hause!
come for me on.time to home
*'Come home on time and I want that to happen!'
(Gutzmann 2007, examples 12c/d, my translation)
Second, the ethical dative leads to entailments and cannot be cancelled.
185. #Du sollst mir pünktlich nach Hause kommen,
you.Nom shall me.Dat on.time to home come
auch wenn es mir persönlich egal ist.
even if it me.Dat personally indifferent is
'You shall come home on time and I want that to happen even though,
personally, I don't care about it.'
(Gutzmann 2007, example 14, my translation)
The second clause of (185) cannot cancel the meaning of the ethical dative but
rather makes the sentence unacceptable and contradictory.
Third, the ethical dative is speaker-oriented. This can be seen from its form
as well as from the fact that it cannot be embedded in order to contribute its
meaning to someone other than the speaker (186).
186. a. Mutter: "Sei mir pünktlich zu Hause!"
mother be me.Dat on.time to home
Mother: 'Be home on time and I want that to happen!'
b. *Die Mutter sagte, dass ich mir pünktlich
the.Nom mother said that I me.Dat on.time
zu Hause sein soll.
to home be shall
'The mother said that I should be on time and I wanted that to hap-
pen.'
c. *Die Mutter sagte, dass Peter mir pünktlich
the.Nom mother said that Peter me.Dat on.time
zu Hause sein soll.
to home be shall
'The mother said that Peter should be home on time and I wanted
that to happen.'
(Gutzmann 2007, example 19, my translation)
Finally, the contribution of the ethical dative is independent of the at-issue
meaning of the sentence. It does not impact the truth conditions of a sentence
(see (183)) but it just reflects the speaker’s attitude towards an event.

 
98   

Consequently, Gutzmann provides the denotation (187) for the ethical dative
(in Potts's (2005) system; DE stands for 'dativus ethicus'). The ethical dative is a
pronoun that carries the meaning of the speaker's interest in the proposition.
187. meDE→ λp. meDE (p) =def λp. want(speaker)(p): <ta,tc>
(Gutzmann 2007, example 31, my translation)
Gutzmann discusses how the not-at-issue contribution of the ethical dative pro-
noun can explain its syntactic behavior. The semantic type of this element dif-
fers from the type that Potts proposed for CIs. This allows Gutzmann to explain
why the ethical dative cannot head a relative clause or be modified (177, 178)
but Potts's CI elements can.
4.2.2.2 Analysis
The problem with the analyses by Wegener (1989) and Gutzmann (2005) is that
each neglects the main focus of the other. Wegener makes it seem like it is acci-
dental that the ethical dative has the form of a pronoun. Gutzmann requires
German to have an "ethical pronoun" ("meDE") in addition to regular pronouns.
The ethical pronoun can "accidentally" only occur in directive sentences. In my
analysis below, I combine the two approaches which allows me to explain why
neither the form of the ethical dative nor the sentences it occurs in are acci-
dental.
I rely on the approach to imperatives and subjunctives proposed by Han
(1998). In his analysis, subjunctives have an [IRREALIS] feature. This feature is
used to encode a modality of unrealized interpretation. Combined with the pro-
position p of a sentence (irrealis(p)), it denotes a set of possible worlds in which
p is satisfied. "The speaker is agnostic to whether the real world is included in
this set" (Han 1998, p. 151).
In Han's approach, imperatives have an additional [DIRECTIVE] feature. This
feature reflects the directive mood (i.e. that the addressee is ordered/ asked to do
something by the speaker). Since the addressee is asked to do something in the
future, the directive feature also carries a future orientation explaining why im-
peratives cannot be used for past events (cf. Portner 2007). In imperatives, the
[IRREALIS] feature and the [DIRECTIVE] feature co-occur (directive (irrealis (p)))
which leads to the interpretation that the speaker orders the addressee to bring
about the (at the moment) unrealized situation described by the proposition p.
In my analysis, the ethical dative is introduced by an applicative head AppED
which attaches between the two projections hosting the two features [IRREALIS]
and [DIRECTIVE]. In other words, I propose the following (partial) structure for
an ethical dative construction.


  99

188.
3
AppEDP
DIRECTIVE
3
ethical dative AppED'
3
AppED 3
IRREALIS IP
Semantically, AppED takes a set of possible worlds in which the proposition is
realized as its argument and carries the not-at-issue meaning that the referent of
the ethical dative wants one world from this set to become the real world (189).
(In other words, the denotation can be thought of as a not-at-issue version of
WANT.40)
189. [[AppED]]= λP. λx. P : x wants one world from the set P to come true
The semantic denotation of (165), repeated below, with the assumed syntactic
structure of (188) is shown in (190).41
190. Komm mir pünktlich nach Hause!
come me.Dat on.time to home
'Come home on time and I want this to happen!'
[[AppEDP]] = w. e in w. COME(e) & Goal(home)(e) & On time(e) :
I want one world from the set (w. e in w. COME(e) &
Goal(home)(e) & On time(e)) to come true
AppEDP carries the at-issue meaning that there is a set of possible worlds that
have a ‘coming home on time’ event. This is identical to the meaning of the pro-
jection hosting [IRREALIS]. AppEDP also carries the not-at-issue meaning that the
speaker wants one of these possible worlds to become realized. Once the denota-
tion of AppEDP is combined with the directive force, the order that the speaker
wants the addressee to be home one time (and the not-at-issue meaning that the
speaker wants this) is derived for this sample sentence. Thus, [DIRECTIVE] acts
like a speech act operator. The ethical dative emphasizes the speaker's interest in
an as-yet unrealized situation being realized.
The form and reference of the ethical dative follow from the selectional re-
quirement that only the projection hosting [DIRECTIVE] can select AppED. Conse-
quently, the ethical dative has to denote the person who has an interest in the
proposition coming true as well as the person who has the authority to use di-
rective force. This is the speaker and, consequently, the ethical dative must be a
first person pronoun. Unlike in Gutzmann's approach (187), the reference to the
speaker is not directly encoded in the denotation of AppED in my analysis.

40 Thanks to Masahiro Yamada for pointing this out!
41 The full semantic derivation is given in Appendix 2, #1.

 
100  

Given my analysis, it also follows that the ethical dative is a pronoun and
not a particle with an invariant form. Like Gutzmann argued, its invariant form
follows from its semantics. However, in my analysis it is a regular (weak) pro-
noun and not a special not-at-issue pronoun (as Gutzmann proposed (187)). Its
restriction to certain sentence types is due to the fact that it requires the [IRREAL-
IS] and the [DIRECTIVE] features which are not present in all sentences but AppED
can only select projections with [IRREALIS] (it requires a set of possible worlds)
and can only be selected by projections hosting [DIRECTIVE]. As a consequence,
ethical datives can only appear in sentences with directive force.42
The required presence of [DIRECTIVE] explains why the ethical dative is un-
acceptable in the past tense.
191. *Ich will/ wollte, dass du mir
I.Nom want/ wanted that you.Nom me.Dat
pünktlich nach Hause kamst.
on.time to home came
'I want/ wanted that you came home on me and I want that to happen.'
Yet, the ethical dative is compatible with future tense (192). This is expected
because neither the [DIRECTIVE] feature nor the [IRREALIS] feature rule out this
tense.
192. Du wirst mir zur Schule gehen!
you.Nom will me.Dat to.the school go
'You will go to school and I want that to happen!'
I assume that the presence of the [DIRECTIVE] feature is also responsible for
the second person subject of the imperatives and the declarative sentences with
directive force. Though the addressee must be physically present when the ethi-
cal dative is used felicitously (Wegener 1989), I refrain from encoding it syntac-
tically, as, for instance, Zanuttini (2008) did because it does not seem to be nec-
essary for the addressee to be syntactically encoded in an analysis of the ethical
dative.

4.2.3 Summary
The ethical dative is a (first person pronoun) pronoun. It is introduced by an ap-
plicative head, AppED, which contributes an identity function for the proposition
on the at-issue tier of meaning and a not-at-issue version of WANT. The re-
strictions on the form and occurrence of the ethical dative stem from the fact that


42 This approach loses the explanation for why the ethical dative cannot be modified by an
appositive (178) or head a relative clause (177). This might be related to its position
outside of IP.


  101

the features [IRREALIS] and [DIRECTIVE] are required for the ethical dative to be
used grammatically.
4.3 Subject Co-Referential Applicative Arguments
Another type of not-at-issue applicative argument discussed in this chapter is the
subject co-referential applicative43 of English (193a)44 and of (dialects of) Ger-
man (193b).
193. a. John1 is gonna kill him1 a bear.
b. Ich trinke mir jetzt einen Kaffee.
I.Nom drink me.Dat now a.Acc coffee
'I'm gonna drink me some coffee now.'
(Horn 2009, example 31a, my translation)
The use of the subject co-referential pronoun in these sentences indicates a spe-
cial involvement of the subject in the event, e.g. in (193b) that I was going to
leisurely drink a cup of coffee.
The subject co-referential applicative argument has frequently been dis-
cussed for English (see Horn 2008; 2009, Webelhuth and Dannenberg 2006,
Christian 1991, Conroy 2007, among others) but, to my knowledge, Horn (2009)
is the only one who mentions it for German.
Below, I describe the subject co-referential applicative including its status as
an applicative argument and its contribution of only not-at-issue meaning (sec-
tions 4.3.1-4.3.5). I provide a semantic and syntactic analysis in section 4.3.6
and discuss its consequences (section 4.3.7) before presenting previous analyses
of this construction in section 4.3.8.

4.3.1 Description
The subject co-referential applicative argument meets my definition of an ap-
plicative argument (4). It is optional in the sentences in (193), as (194) shows.
194. a. John is gonna kill a bear.
b. Ich trinke jetzt einen Kaffee.
I drink now a.Acc coffee
'I'm gonna drink some coffee now.'
The special involvement of the individual denoted by the applicative argument
in (193) is not entailed by these sentences. The subject co-referential argument


43 The name is taken from Halevy (2007) who discusses a similar type for (biblical) He-
brew.
44 This construction seems to be prevalent in informal registers (Horn 2008) as well as in
Southern and Appalachian American English (Horn 2008, Webelhuth and Dannenberg
2006).

 
102  

is not governed by a preposition. Thus, all requirements for an applicative argu-


ment are met.
The meaning contribution of this applicative argument can be characterized
as the subject having a remarkable involvement in the situation. The exact
meaning, however, is difficult to pin down. Horn (2009) describes it as adding
the meaning of "subject intentionality/ success/ benefit/ satisfaction" (p. 12).
Webelhuth and Dannenberg (2006) state that the subject co-referential pronoun
underscores the subject's role in the event. In short, this applicative argument
carries some (positive) involvement of the agent/subject in the event (see also
Hutchinson and Armstrong 2013).
The German subject co-referential argument differs from other applicatives
of German, namely the affected experiencer (Chapter 3), the Dative of Inaction
(Chapter 7), the part-whole applicative (Chapter 6), and the benefactive (Chapter
5) because it contributes only not-at-issue meaning and displays the behavior
associated with not-at-issue elements (section 4.3.5). The subject co-referential
applicative argument can be distinguished from the ethical dative (section 4.2)
because the former must co-refer with the subject (section 4.3.2) whereas the
latter must refer to the speaker and appear in directive clauses.
The English subject co-referential applicative argument is the only English
applicative that carries only not-at-issue meaning (section 4.3.5) and can there-
fore easily be distinguished from other English applicatives because of the syn-
tactic behavior of not-at-issue elements. Furthermore, no other English applica-
tive argument must co-refer with the subject (cf. section 4.3.2).

4.3.2 Features
As the name indicates, the subject co-referential applicative argument must be
co-referential with the subject. If this is not the case, the subject co-referential
applicative is not possible (195).
195. a. i. Mary1 would love her1 some flowers.
ii. *Mary1 would love her2 some flowers.
iii. *Mary1 would love Sue2 some flowers.
(Webelhuth and Dannenberg 2006, example 31a, b, d)
b. i. Ich trinke mir einen Kaffee.
I.Nom drink me.Dat a.Acc coffee
'I'm gonna drink me some coffee.'
ii. *Ich trinke Dennis einen Kaffee.
I.Nom drink Dennis.Dat a.Acc coffee
'I'm gonna drink some coffee for Dennis's pleasure.'


  103

There is a preference for the subject co-referential applicative to be a first or


second person pronoun (Christian 1991) but all persons are possible (Webelhuth
and Dannenberg 2006, Horn 2008).
196. a. i. I1 got me1 some candy.
ii. You1 got you1 some candy.
iii. He1 got him1 some candy.
iv. We1 got us1 some candy.
v. They1 got them1 some candy.
(adapted from Webelhuth and Dannenberg 2006, example 10)
b. i. Ich trinke mir jetzt einen Kaffee.
I.Nom drink me.Dat now a.Acc coffee
'I'm gonna drink me some coffee now.'
(Horn 2009, example 31a, my translation)
ii. Du trinkst dir jetzt einen Kaffee.
you.Nom drink you.Dat now a.Acc coffee
'You're gonna drink you some coffee now.'
iii. Er trinkst sich/ *ihm jetzt einen Kaffee.
he drink self/ him.Dat now a.Acc coffee
'He's gonna drink him some coffee now.'
iv. Wir trinken uns jetzt einen Kaffee.
we.Nom drink us.Dat now a.Acc coffee
'We're gonna drink us some coffee now.'
v. Ihr trinkt euch jetzt einen Kaffee.
you.Nom drink you.Dat now a.Acc coffee
'You're gonna drink you some coffee now.'
vi. Sie trinken sich/ *ihnen jetzt einen Kaffee.
they drink self/ the.Dat now a.Acc coffee
'They're gonna drink them some coffee now.'
Also, animate reference for the applicative is preferred over inanimate refer-
ence but the latter is possible in English (197) (Christian 1991, Horn 2009).
197. That house1 needs it1 a new roof.
(Horn 2009, example 35c, originally from Sroda and Mishoe (1995))
In German, inanimate reference for the subject co-referential applicative is ruled
out (198d, e) because German only allows this applicative argument to occur
with verbs of consumption (198a) (and marginally with verbs of creation45
(198b)).


45 Often, these are difficult to judge because of their ambiguity with other applicative ar-
guments, esp. benefactives.

 
104  

198. a. Ich esse mir ein Stück Kuchen.


I eat me a slice cake
'I'm gonna eat me a slice of cake.'
b. Ich baue mir ein Auto.
I build me a car
?'I'm gonna build me a car.'
'I'm building a car for me.'
c. *Ich fahre mir nach Dortmund.
I drive me to Dortmund.
'I'm gonna drive me to Dortmund.'
d. *Das Haus braucht sich ein neues Dach.
the house needs self a new roof
'The house needs it a new roof.'
e. ??The Roboter isst sich einen Apfel.
the robot eats self an apple
'The robot is eating it an apple.'
There is no prepositional variant for the subject co-referential applicative ar-
gument (cf. Webelhuth and Dannenberg 2006). This shows that the subject co-
referential applicative is not a variant of the benefactive applicative (which al-
ternates with a prepositional variant; Chapter 5).
199. a. #John is gonna kill a bear for himself.
b. Ich trinke jetzt einen Kaffee für mich (selbst).
I drink now a.Acc coffee for me self
'I'm gonna drink a coffee for myself.'
*'I'm gonna drink me a coffee.'
The examples in (199) do not have the same meaning of extraordinary involve-
ment or intense participation of (193). Rather, the prepositional paraphrase just
indicates that there was benefit for the subject but that benefit could be different
from indulging in the action described by the sentence.

4.3.3 Form
The subject co-referential applicative must be a weak pronoun for its meaning to
surface (Webelhuth and Dannenberg 2006, Horn 2009). Other elements, espe-
cially the reflexive pronoun in English, may be grammatical in its verb-adjacent
position but do not carry the meaning of highlighting the subject's involvement
in the event. This can be seen from the following examples.
200. a. I'm gonna buy me some flowers. (subject co-referential)
b. I'm gonna buy myself some flowers. (recipient benefactive)
Whereas the subject co-referential applicative in (200a) indicates that the subject
buys flowers for his/her pleasure (but not necessarily to keep them), the recipi-


  105

ent benefactive (indicated by the reflexive) indicates that the subject bought the
flowers with the intention of passing them on to the individual denoted by the
applicative argument, here him-/herself. In other words, the difference between
(200a) and (200b) is that in (200a) the subject is intensely involved in the act of
buying flowers, and in (200b) the subject bought the flowers with the intention
of keeping them. Although the subject co-referential and the recipient benefac-
tive can occur in the same position, the subject co-referential one must be a
weak pronoun46 co-referring with the subject while the recipient benefactive
cannot be a weak pronoun co-referring with the subject (but it might be a reflex-
ive (200b)). In other words, subject co-referential applicatives seem to be blatant
violations of binding principle B (Webelhuth and Dannenberg 2006, Conroy
2007, Haddad 2011).
Furthermore, the subject co-referential applicative argument and the reflex-
ive cannot always occur in the same position (201) (Conroy 2007, Horn 2009).
201. a. He needs {*himself/ him} just a little more sense.
(Horn 2009, example 20b)
b. *I hurt me.
I hurt myself.
(Conroy 2007, example 2)
c. I'm gonna write me a letter to the president.
*I'm gonna write myself a letter to the president.
(Conroy 2007, example 12)
Consequently, the subject co-referential applicative argument cannot be seen as
an alternative to the reflexive (Conroy 2007).
Conroy (2007) argues that the English subject co-referential applicative is a
SE-anaphor. She shows that under ellipsis only a sloppy interpretation is availa-
ble (202) (see also Horn (2009)).
202. I got me a shotgun and so did Billy Bob.
sloppy identity reading only: we both got one
(Horn 2009, Example 26a)
This is only expected if the subject co-referential applicative is a bound variable.
Furthermore, like other SE-anaphors, the subject co-referential applicative ar-
gument requires a de re-interpretation: the identity of its referent and the referent
of the antecedent must be exactly the same; imagine an architect seeing a statue
of himself in an empty square. He decides to build an opera house behind it. In
that case, uttering I'm building me an opera house is not felicitous with me refer-
ring to the statue because he is going to build it behind the statue and not his
own body. Only I'm building myself an opera house is felicitous in this case.

46 See also section 4.4.

 
106  

Thus, I agree with Conroy's proposal that English subject co-referential applica-
tives are SE-anaphors and not (Principle B-violating) pronouns.
Further support for this contention comes from German in which the SE-
anaphor is used as the subject co-referential applicative. This can clearly be seen
in the third person (196b-iii, b-vi) where sich must be used. Just like in English
(201a), the self-anaphor (203) and the pronoun (196b-iii, b-vi) cannot be used as
a subject co-referential applicative argument in German.
203. a. *Ich trank mir selber/ selbst einen Kaffee.
I drank me.Dat self/ self a.Acc coffee
'I drank me a coffee.'
b. *Er trank sich selber/ selbst einen Kaffee.
he drank self.Dat self self a.Acc coffee
'He drank him a coffee.'
In summary, the subject co-referential applicative is a SE-anaphor. Like all
SE-anaphors, it co-refers with the subject (Heim and Kratzer 1998). The subject
co-referential applicative is used to indicate a special (positive) involvement of
the referent in the event. The subject co-referential applicative argument is not
an alternative for the reflexive pronoun.

4.3.4 Role of the Direct Object


The subject co-referential applicative is often used with transitive verbs (Chris-
tian (1991); see below). It has been observed that the direct object in such con-
structions is typically quantified. In fact, Webelhuth and Dannenberg (2006)
claim that generics in the role of the direct object are ungrammatical but Horn
(2008) shows that this is only a tendency and not a requirement.
204. a. *Mary1 would love her1 flowers.
(Webelhuth and Dannenberg 2006, example 33c)
b. I love me chocolate and I love me milk.
(Horn 2008, example 21d)
Hutchinson and Armstrong (2013), Storoshenko (2012), McLachlan (2011)
include this observation about the quantification in their analyses of the subject
co-referential applicative argument. In contrast to these researchers, I assume the
widest observed distribution of the subject co-referential applicative argument,
and will consequently not further investigate the role (and/or type) of the direct
object occurring with the subject co-referential applicative argument.

4.3.5 Not-At-Issue Meaning


The subject co-referential applicative contributes only not-at-issue meaning
which can be seen when applying the tests discussed in Chapter 2. This meaning
contribution has also been discussed by Horn (2008; 2009).


  107

The subject co-referential pronoun cannot be questioned in a wh-question


(205).
205. a. *Who did John1 kill a bear? - Him1.
b. *Wem hat er einen Kaffee getrunken? - Sich.
who.Dat has he a.Acc coffee drunk - self
'Who did he1 drink a coffee? - Him1.'
This is the first indication that the contribution of the subject co-referential ap-
plicative argument is on the not-at-issue tier of meaning.
Further support comes from negation. The meaning of special involvement
cannot be negated by itself (206). Only the whole event can be negated.47
206. a. John1 didn't kill him1 a bear.
1. John didn't kill a bear (but if he had, this would have been re-
markable for John).
2. *John killed a bear but this was not remarkable for John.
b. Ich hab mir keinen Kaffee getrunken.
I have me.Dat no coffee drunk
'I didn't drink me any coffee.'
1. I didn't drink me any coffee.
2. *I did drink some coffee but I didn't enjoy it.
Negation not being able to target the meaning of the applicative argument indi-
cates that it contributes not-at-issue meaning.
Similarly, the meaning contributed by the applicative argument projects past
yes/no-questions (207).
207. a. Did John1 kill him1 a bear?
b. Hast du dir einen Kaffee getrunken?
have you.Nom you.Dat a.Acc coffee drunk
'Did you drink you some coffee?'
If the answer to the question (207a) is only "no", this cannot mean that John did
kill a bear but that he did not view this as an enjoyable event. For this meaning,
a more elaborate answer than "no" is needed. Similarly, a negative answer to
(207b) means that you didn't drink any coffee. It cannot mean that you did drink
coffee but were not enjoying it/yourself. Answering "no" to these questions
means that the special involvement is accepted. This shows that the meaning of


47 German sentential negation nicht is not possible in this case due to the strong preference
for indefinite direct objects. Even with biclausal negation, the special involvement can
only be negated if that is explicitly stated. Otherwise the whole event is negated:
i. Es stimmt nicht, dass ich mir noch einen Kaffee getrunken hab.
it be.true Neg that I me.Dat still a.Acc coffee drunk have
'It is not true that I drank me some coffee.'

 
108  

a positive experience projects past the question and is on the not-at-issue tier of
meaning.
This can also be seen in if-clauses in which the subject co-referential pro-
noun does not make any difference for the truth conditions.
208. a. i. If John1 kills him1 a bear, I'll buy him some beer.
ii. If John kills a bear, I'll buy him some beer.
b. i. Wenn du dir (hier) einen Kaffee
if you.Nom you.Dat here a.Acc coffee
trinkst, dann bring ich dir ein Stück Kuchen.
drink then bring I you a slice cake
'If you drink you some coffee (here), I'll bring you a slice of cake.'
ii. Wenn du (hier) einen Kaffee trinkst,
if you here a.Acc coffee drink
dann bring ich dir ein Stück Kuchen.
then bring I you a slice cake
'If you drink a coffee (here), I'll bring you a slice of cake.'
Both sentences in (208a) require John to kill a bear for me to buy him beer. The
subject co-referential pronoun in (208a-i) does not add another condition to this.
The same is true for (208b). The addition of the subject co-referential applica-
tive argument does not alter the conditions of the if-clause. This is the expected
behavior for elements that contribute only not-at-issue meaning.
In summary, all of the tests discussed in Chapter 2 show that the meaning
contribution of the subject co-referential applicative argument is to the not-at-
issue tier of meaning. This is also the conclusion drawn by Horn (2008; 2009)
who argues that they are conversational implicatures (defined slightly differently
from Potts (2005)).

4.3.6 Analysis
I provide my formal semantic analysis of the subject co-referential applicative in
this section. Previous approaches to these applicatives are discussed in section
4.3.8. I propose that the subject co-referential applicative argument is introduced
by an applicative head, AppSCR, that attaches between Voice and VP. Semanti-
cally, the AppSCR head passes up the verbal event (as at-issue content) and com-
bines it with Voice later on. Voice serves as the third semantic argument of
AppSCR. The applicative head also introduces on the not-at-issue tier that the
event was intense/ extraordinary48 for a participant of the event. This denotation
of AppSCR is given in (209b). Since the subject co-referential applicative is an
SE-anaphor, it is generated as a variable with an index (a1) (Heim and Kratzer

48 This is identical to "remarkable" in Bosse (2013b).


  109

1998). The variable shares the index of the subject. In order for the subject to
bind the variable, it needs to move to a higher projection than the index binder.
Thus, the derivation of (193), repeated below, proceeds as follows:49
209. a. John1 killed him1 a bear.
b. [[AppSCR]] = λPv,t. λx. λQe,vt. λy. λe. P(e) & Q(e)(y): Intense(e)(x) &
x =y
Intense(e)(x) → The event e is intense for the individual x iff x is ex-
traordinarily involved in e.
c. VoiceP
3
John1 Voice2'
3
1 Voice1'
3
t1 Voice'
3
Voice AppSCRP
3
a1 AppSCR'
3
AppSCR VP
6
kill a bear
d. [[VoiceP]] = λe. KILL(e) & Thm(e)(a bear) & Agt(e)(John) : In-
tense(e)(John) & John = John
Thus, VoiceP denotes a killing event of a bear of which John is the agent. It also
carries the not-at-issue meaning that John is extraordinarily involved in this
event.

4.3.7 Consequences
Due to the requirement "x=y" included in the denotation (209b), it is ensured
that the individual in the specifier of VoiceP will be the one being extraordinari-
ly involved in the event. This reflects the co-reference requirement between the
subject and the subject co-referential applicative argument. It also explains why
the subject co-referential applicative is grammatical with unergative verbs (210)
(contrary to Christian (1991) who claims that they can only occur with transitive


49 The full semantic derivation is given in Appendix 2, #2.

 
110  

verbs).50 The subject of the unergative verb can bind the anaphor (and share its
index).
210. I'm gonna sit me down.
My analysis can also explain some of the semantic verbal restrictions ob-
served with the subject co-referential applicative argument. In English, there are
preferences for the subject co-referential applicative arguments to appear with
certain verbs over others. Horn (2009) characterizes these preferred verbs as
"down-home verbs", e.g. buy, build, shoot, get. He argues that other (transitive)
verbs may produce a register clash. In addition, the subject co-referential applic-
ative is more readily accepted with intentional actions (211) and verbs indicating
(future) possession (Horn 2008).
211. a. She caught her a catfish.
b. #She caught her a cold/ case of the clap.
(Horn 2008, example 26)
According to Horn, sentences like (211b) are marked because they do not in-
volve an intentional agent. Alternatively, they might be ruled out for pragmatic
reasons as it is difficult to see how the special (positive) involvement in these
events is justified.
In order to limit the subject co-referential applicative to certain verbs of
English and verbs of consumption/ creation in German, I assume a selectional
restriction between AppSCR and VP. Since AppSCR attaches outside of VP, it can
enforce its selectional restrictions.
There are several unresolved issues with this analysis. One problem is that
this analysis does not explain the ungrammaticality of passives (see also We-
belhuth and Dannenberg (2006), Christian (1991)).51
212. a. *A bear1 was killed it1 (by John).
b. *A bear was killed him1 by John1.
c. *Ein Kaffee wurde ihm1 von Jan1
a coffee was him.Dat by Jan
getrunken.
drunk
'Some coffee was drunk him by Jan.'
For my analysis, it is unclear what prevents the underlying object from moving
into the subject position to bind the anaphor (212a). This is a more general prob-

50 This point is not relevant for German because it only allows verbs of consumption and
creation with the subject co-referential applicative (198).
51 Sentences in which the anaphor is promoted to the subject are correctly ruled out by the
analysis because the SE-anaphor must be bound by the subject. It cannot be used as a
subject (and bind itself) (see Isacenko (1965) for German passives of constituents co-
referring with the subject (of the corresponding active sentence)).


  111

lem, namely that a derived subject cannot bind the subject co-referential applica-
tive, as unaccusative verbs do not license the subject co-referential applicative.52
213. *The captain1 sank him1 down under the sea.
This problem is similar to the chain condition discussed by Rizzi (1986) which
bans reflexive clitics with derived subjects in Italian.
214. *NP1 … [si1 … e1 …]
However, this observation still does not explain the ungrammaticality of (212b).
This problem of derived subjects not being able to bind the subject co-referential
applicative is left for future research.
Another open question for my analysis of the subject co-referential applica-
tive is that it predicts the following two readings with again (cf. section 3.5.2):
215. VP: the event happens again but for the first time with special involve-
ment
VoiceP: the event happens again with special involvement of the same
individual
The first predicted reading is not available.
216. a. Yesterday, there was a window open in here. Today, they're all
closed. I'm hot, so #I'm going to open me a window again.
b. Yesterday, there was a window open in here. Today, they're all
closed. I'm hot, so I'm going to open a window again.
This could indicate an attachment of AppSCR lower than VP. This problem re-
quires further research because a lower attachment requires a different mecha-
nism to establish the co-reference between the applicative and the external ar-
gument (see the next section).

4.3.8 Previous Analyses


In this section53, I briefly discuss two other analyses aimed at explaining the
English subject co-referential applicative argument, namely an anti-locality ac-
count proposed by Haddad (2011) and a low applicative construction suggested
by Hutchinson and Armstrong (2013). I discuss these two in turn.
4.3.8.1 Haddad (2011)
In his anti-locality account of the subject co-referential applicative argument,
Haddad (2011) focuses on why the subject co-referential applicative is not sub-

52 Experiencer subjects are acceptable:
i. I smell me some cookies.
These might have a structure different from other unaccusatives or might not be unaccu-
sative at all since they can also be passivized (unlike other unaccusatives):
ii. The cookies can be smelled everywhere in the house.
53 This section is based on Bosse (2013b).

 
112  

ject to binding condition B (and thus does not surface as a reflexive pronoun).
Haddad assumes that the subject co-referential applicative is a head (and not a
phrase). Therefore, he treats it as a clitic (rather than a SE-anaphor). This clitic
is merged outside of vP (which is the projection of the subject, just as my
VoiceP). This is schematized in (0).
217. John sold him a dozen toothbrushes.
IP
3
John I'
3
I ApplP
3
him vP
3 3
sold 'em John v'
6
sold a dozen toothbrushes
The structure in (0) includes two movement steps: (a) the verb head-moves to
the applicative clitic which cliticizes onto the verb, and (b) the subject moves
from vP to a higher position (in IP).
This analysis explains that the subject co-referential applicative surfaces as a
weak pronoun by having it attach to the head-moved verb. Furthermore, this ex-
plains why the applicative must occur in a verb-adjacent position (in English).
However, there is one major flaw with this analysis, namely the fact that the ap-
plicative is required to be a head. As Haddad himself points out, this is highly
questionable because the subject co-referential applicative argument can be
camouflaged with the colloquial phrase X's ass (218), see also Horn (2008).
218. I want my ass some quesadillas.
(Horn 2008, example 23b)
In this example, the phrase my ass functions as the subject co-referential pro-
noun and camouflages the pronoun me. Collins et al. (2008) argue that the ca-
mouflage X's ass should be analyzed as a phrase (rather than a head). Thus, by
assuming that the subject co-referential applicative is a head, Haddad cannot
account for camouflage constructions such as (218) whereas they can be ex-
plained by my analysis (209) which has a phrasal position available for the ap-
plicative argument as well as its camouflage.
There are a few other issues with this approach. First, Haddad did not spell
out a formal semantic derivation for his analysis. It is unclear how it proceeds
with the applicative being a head only. This is even more crucial as his analysis
predicts that both vP and VP should be of type <v,t> and therefore be available


  113

for again-modification. Attaching to either of these nodes, again should make


available the following reading: the verbal event happens again but for the first
time with the special involvement of the referent of the subject co-referential
applicative argument. This is due to the applicative being attached outside of vP.
It is this reading that is not available (216). Second, the applicative is attached
outside of the subject. Therefore, the applicative should not be able to impose
restrictions on the verb. However, as shown for instance in (211), the subject co-
referential applicative is not licensed to appear with all verbs.
In summary, Haddad's anti-locality approach falls short of accounting for all
of the features of the subject co-referential applicative argument.
4.3.8.2 Hutchinson and Armstrong (2013)
Hutchinson and Armstrong (2013) suggest that the subject co-referential appli-
cative is a low applicative (in the classification of Pylkkänen (2002)). That
means that the applicative is included in the VP and combines with the direct
object first.54 Their proposed phonologically null applicative head used to intro-
duce the applicative argument has the following semantic denotation. ("SAT"
stands for "satisfaction".)
219. [[ApplSAT]] = λx. λy. λf<e<s,t>>. λe. f(e,x) & Theme(e,x) : (e') [Satisfac-
tion(e') & Experiencer(e',y) & (e'')[f(e') → Source(e'',e')]]
(Hutchinson and Armstrong 2013, example 37)
This head first semantically combines with the direct object and then the appli-
cative argument (second semantic argument). The at-issue meaning is an identi-
ty function of the verb, which is the third semantic argument. In addition, there
is not-at-issue meaning, namely that the at-issue event is the source of a satisfac-
tive experience for an individual.
Hutchinson and Armstrong (2013) provide the following example for the
syntactic and semantic derivation of the subject co-referential applicative argu-
ment.


54 Hutchinson and Armstrong (2013) exclude sentences in which the subject co-referential
argument occurs with intransitive verbs from their analysis. These occurrences are at-
tributed to a "complex structure" (footnote 5).

 
114  

220. VoiceP
3
DP Voice'
John 3
Voice VP
3
V ApplP
bake 3
DP Appl'
him 3
ApplSAT DP
a cake
(Hutchinson and Armstrong 2013, example 38)
[[VoiceP]] = λe. Agent(e, John) & BAKE(e) & Theme(e, a cake):
(e')[Satisfaction(e') & Experiencer(e', him) & (e'')[BAKE(e'') &
Theme(e'', a cake) Source(e'',e')]]
(Hutchinson and Armstrong 2013, example 39)
This analysis has one advantage over my analysis (209), namely that it can
account for the again-modification facts (see the discussion of example (216)).
However, this analysis also has problems. First, Hutchinson and Armstrong ex-
plicitly state that this low applicative is syntactically identical to the English
double object construction. However, the two constructions do not behave alike
syntactically in all respects. For instance, the double object construction can be
passivized while the subject co-referential applicative may not be passivized
(221).
221. a. Mark was given a cake (by John).
b. *He1 was shot a bear (by John1).
Hutchinson and Armstrong cannot explain this difference between the double
object construction and the applicative argument. The analysis might be sal-
vaged by arguing for different phase boundaries in the two constructions resul-
ting in different movement possibilities. However, that would lose the proposed
syntactic similarity between the two.
Another problem with Hutchinson and Armstrong's analysis is the proposed
semantic denotation of the phonologically null head. This head takes the direct
object and the applicative argument as its first two semantic arguments. It relates
these two via an event and not directly, as would be expected for a low applica-
tive. The semantic denotation of the head is, in fact, that of a high applicative
denotation because it relates an individual to an event. In this, the analysis by
Hutchinson and Armstrong is similar to my analysis. In addition, the proposed
semantic denotation of ApplSAT is bi-eventive as both the verbal event and "Sat-


  115

isfaction(e')" are included in the denotation. However, this bi-eventive analysis


is not motivated. The two events cannot be independently identified.
In short, while the low applicative analysis proposed by Hutchinson and
Armstrong (2013) can explain the again-modification facts better than my ana-
lysis, their analysis of the subject co-referential applicative argument as a low
bi-eventive applicative is not well motivated. Furthermore, semantically speak-
ing it is a high applicative analysis just like my proposed analysis.

4.3.9 Summary
I have shown that the subject co-referential applicative argument is a SE-
anaphor which contributes only not-at-issue meaning. The SE-anaphor can be
observed directly in German. Further support for this comes from the sloppy in-
terpretation under ellipsis. I argue that the anaphor is introduced by an App
head, AppSCR, which attaches outside VP and introduces the applicative argu-
ment, the SE-anaphor. Some syntactic behavior of the subject co-referential ap-
plicative follows from the fact that it is a SE-anaphor on the not-at-issue tier of
meaning. However, not all syntactic behaviors of the subject co-referential con-
struction could be explained. Certain preferences for the occurrence of the sub-
ject co-referential applicative seem to be related to its pragmatic licensing condi-
tions (register, possession/consumption contexts) rather than to the syntactic
structure.
4.4 Not-At-Issue Applicatives
The types of not-at-issue applicative arguments discussed in this chapter (Ger-
man ethical datives, subject co-referential pronouns in English and German)
share a property that goes beyond their contribution of not-at-issue meaning,
namely their form: all of them are weak/clitic pronouns. Even more, the not-at-
issue affected experiencers of Hebrew and French (section 3.9.2) also share this
characteristic.55
All these not-at-issue applicatives display features that Cardinaletti and
Starke (1999) have identified for weak pronouns: first, weak pronouns cannot be
stressed. This is also true for all the discussed not-at-issue applicatives discussed
here (Borer and Grodzinsky 1986, Wegener 1989, Christian 1991, Gutzmann
2007).
222. a. *li, hem mitxatnim kol ha-zman
to.me they marry all the-time
'They marry all the time on me.'
(Borer and Grodzinsky 1986, examples 10b)

55 Thanks to Benjamin Bruening for pointing this out!

 
116  

b. *Du sollst [mir]F dem Papa die Schuhe putzen.


you shall me.Dat the dad the shoes clean
‘You shall clean the shoes for dad, and I want this to happen.’
c. *John killed [him]F a bear.
Also, the Hebrew affected experiencer (222a) and the English subject co-
referential applicative (194a) must occur verb-adjacent (cf. Webelhuth and Dan-
nenberg 2006), as is expected for weak pronouns in these languages.
223. a. *John1 killed yesterday him1 a bear.
b. *John1 killed a bear him1.
For German, it was shown that the ethical dative follows the rules for pronoun
placement as well (174).
Cardinaletti and Starke (1999) also argue that weak pronouns cannot be co-
ordinated. This is also true for the not-at-issue applicatives (although there might
be other explanations besides the weak pronoun for this for the ethical dative
(173a) and the subject co-referential applicative because they are restricted to
the speaker and the subject, respectively).
224. a. *hem mitxatnim li ve-lá kol ha-zman.
they marry to.me and to.her all the-time
‘They marry all the time on her and me.’
(Borer and Grodzinsky 1986, example 10c)
b. *You and I killed you and me a bear.
While it seems tempting to explain this pattern by linking the not-at-issue
meaning to the weak pronoun directly, Potts (2005) shows that adjectives and
nouns can also contribute not-at-issue content (cf. Chapter 2). Consequently, it is
not the case that not-at-issue content can only be carried by weak pronouns. It is
also not the case that weak pronouns have to carry not-at-issue meaning, as the
following example shows.
225. Hans hat es gestern gekauft.
Hans.Nom has it.Acc yesterday bought
‘Hans bought it yesterday.’
(Cardinaletti and Starke 1996, example 6)
Here, the weak pronoun es 'it' carries at-issue meaning. Thus, the relation be-
tween being a weak pronoun and being a not-at-issue applicative argument must
be more complex.
One possibility is that weak pronouns can be used if they refer to an entity
already prominent in the discourse (Cardinaletti and Starke 1994). This point
can easily be made for the subject co-referential applicative and the ethical da-
tive: the former has the referent as the subject of the same clause; the latter re-
quires directive force and refers to the speaker. Thus, these referents are promi-
nent in the discourse. However, it is not clear why the affected experiencer (in


  117

French and Hebrew) could not be new to the discourse. Thus, this approach can-
not straightforwardly explain the relation between form and function either.
However, the three types of not-at-issue applicative arguments are not uni-
fied in one dimension, namely in their status in the system proposed by Potts
(2005). Gutzmann (2007) shows that ethical datives meet all the requirements of
Conventional Implicatures (as presented by Potts (2005)). The other two not-at-
issue applicatives are merely multi-dimensional, meaning that they are not
speaker-oriented but meet other requirements that Potts (2005) identified for CIs
(20). Both can be embedded without projecting to the highest level.
226. a. Juliette pense de facon erronée que les
Juliette believes wrongfully that the
enfants leurs ont gribouills tous les murs.
children them.Dat have scribbled all the walls
‘Juliette wrongfully believes that the children scribbled all over the
walls on them.’
b. Jan glaubt fälschlicherweise, dass du dir
Jan believes wrongfully that you you.Dat
noch einen Kaffee getrunken hast.
still a coffee drunk have.
‘Jan wrongfully believes that you drunk you another cup of coffee.’
In neither of these examples is the contribution by the not-at-issue applicative
attributed to the speaker. Roberts et al. (2009) do not address a similar category
of not-at-issue meaning. The discussion here can be taken as an indication that
one is needed.
This shows that the relation between being a not-at-issue applicative and be-
ing a weak pronoun is not straightforward. Some applicative have a referent sa-
lient in the discourse, others do not. Some are speaker-oriented, others are not.
While the discourse saliency of subject co-referential applicatives and the ethical
dative might explain their form, the not-at-issue affected experiencers remain
puzzling. From the data analyzed here, however, it seems to be a strong cross-
linguistic generalization that applicative arguments that contribute only not-at-
issue meaning take the form of a weak pronoun. Yet, further research is required
to see if this correlation really holds.
4.5 Conclusion
In this chapter, I have provided analyses of two types of applicative arguments.
They contribute only not-at-issue meaning and are similar in form and behavior
but differ in their meaning. The German ethical dative is an applicative that is
the not-at-issue variant of WANT. In combination with the features [IRREALIS]
and [DIRECTIVE], its restricted occurrence can be explained. The subject co-

 
118  

referential applicative indicates the subject's special involvement in the event. It


is a SE-anaphor bound by a (non-derived) subject. It is introduced by the head
AppSCR which attaches to VP. All of these not-at-issue applicatives are
weak/clitic pronouns which follow the word order rules of their language but are
otherwise syntactically inert (unlike typical pronouns).
In the remaining chapters, I look at at-issue applicatives. These have very
different characteristics. They are (typically) less restricted than not-at-issue ap-
plicatives in their occurrence as well as in their form. The notable exception is
the Dative of Inaction (Chapter 7) which is also highly restricted in where it can
occur. I investigate the co-occurrence possibilities of at-issue and not-at-issue
applicatives in Chapter 8.


5 Benefactives
5.1 Introduction
In this chapter, I discuss the semantic and syntactic structure of benefactive ap-
plicative constructions in German and English as exemplified in (227).
227. a. John baked him a cake.
b. Jan buk ihm einen Kuchen.
Jan.Nom baked him.Dat a.Acc cake
'Jan baked a cake for him.'
As the name "benefactive" indicates, these applicative arguments denote who
benefits from the described event. As Kittilä and Zuniga (2010) put it: "The
beneficiary is a participant that is advantageously affected by an event without
being its obligatory participant (either agent or primary target, i.e. patient)" (p.
2). This describes the meaning of the benefactive applicatives well.
In section 5.2, I distinguish benefactives from other applicatives, discuss dif-
ferent types of benefactives and show which benefactives are applicative argu-
ments according to my definition (4). It is shown that in English, only those
benefactives that receive a recipient or intended possessor interpretation are ap-
plicative arguments. In German, benefactive applicatives can receive either a
plain, deputative, or possessor interpretation. In section 5.3, I show that benefac-
tive applicative arguments contribute only at-issue meaning. Following that, I
provide the analyses for the observed types of benefactive applicative argu-
ments. The analysis for German true benefactives is based on Pylkkänen (2002)
(section 5.4) and that of English recipient benefactives on Bruening (2010) (sec-
tion 5.5). In section 5.6, I address the prepositional paraphrase of the benefactive
applicative arguments before concluding this chapter in section 5.7.
5.2 Description
Before describing benefactives in detail, I need to distinguish them from other
applicatives. As I show in detail in section 5.3, benefactives contribute only at-
issue meaning. In this, German benefactives differ from affected experiencers
(Chapter 3), subject co-referential applicatives and ethical datives (Chapter 4).
Also, unlike affected experiencers, benefactives do not require the referent to be
sentient (228) (cf. Bosse and Bruening (2011)).56
228. a. The old lady baked her dead dog some cookies.


56 Yet, the beneficiary is often sentient.


120  

b. Dennis malte seinem toten Vater das Bild.


Dennis.Nom painted his.Dat dead father the.Acc picture
'Dennis painted the picture for his dead father.'
(Bosse and Bruening 2011, example 6a)
These examples show that it is not necessary for the beneficiary to be sentient.
To distinguish benefactive applicatives from part-whole applicative argu-
ments (Chapter 6), the meaning can be used. As stated above, benefactives mark
the benefit of an individual or entity. In contrast, part-whole applicatives denote
that two entities are in a (material) part-whole relation (Chapter 6). There is no
such requirement for benefactives. Finally, benefactives can be distinguished
from the Dative of Inaction (Chapter 7) because of the verbal restrictions that
each of them faces. The benefactive applicative can only marginally occur with
intransitive verbs (section 5.4.3). In contrast, the Dative of Inaction requires in-
transitive verbs (Chapter 7).
English benefactives can be differentiated from other English applicatives
due to the fact that they are the only ones contributing only at-issue meaning and
show the syntactic behavior associated with this meaning contribution (section
5.3).
Across previous literature, it is not always clear what the term "benefactive
construction" refers to (Kittilä 2005). Van Valin and LaPolla (1997) have identi-
fied three different types of benefactive constructions:
229. a. recipient benefactives such as (227a) which require the applicative
argument to denote the (intended) possessor/ recipient;
b. plain benefactives such as Rita sang for the students (Van Valin
and LaPolla 1997, example 7.91c), in which there is some general
benefit for the applicative argument; and
c. deputative benefactives as in Pat stood in line for Kim (Van Valin
and LaPolla 1997, example p. 384), which involves an action done
by the agent instead of (and on the part of) the beneficiary.
The recipient benefactive corresponds to the low applicative discussed by
Pylkkänen (2002) because it requires the transfer of possession. Plain and depu-
tative benefactives are high applicatives in Pylkkänen's approach because they
relate an individual to an event (see Chapter 2). I discuss these three types of
benefactives and their status of being applicative arguments for English and
German individually in sections 5.2.1 and 5.2.2, respectively.57


57 Smith (2010) discusses a different possibility of categorizing benefactives, namely
whether the benefit stems from the agent's action or from the event in general. I leave it
to future research to investigate if this calls for a further (or different) distinction of
benefactive applicatives in German and English.


  121

5.2.1 English
In English, the three types of benefactives are maximally distinct. First, the
recipient benefactive can be distinguished from the other two types based on its
form. The recipient benefactive (230b) can be surface-identical to the ditransi-
tive structure (230a). In contrast, plain and deputative benefactives require the
preposition for (230c, 230d).
230. a. S/he gave me the book. ditransitive
b. S/he baked me a cake. recipient benefactive
c. *S/he went me to the market. plain/deputative
d. S/he went to the market for me. plain/deputative
(Kittilä 2005, example 5b, d-f)
The recipient benefactive can also alternatively be expressed as a for-PP (and
therefore look identical to plain and deputative benefactives).
231. S/he baked a cake for me. recipient benefactive (PP)
(Kittilä 2005, example 5c)
As Kittilä (2005) points out, the preposition for is how recipient benefactives
can be distinguished from ditransitives (230a) which take the preposition to in
their prepositional variant (232a).
232. a. S/he gave the book to me. ditransitive
(Kittilä 2005, example 5a)
b. *S/he gave the book for me. ditransitive
In general, the plain and the deputative benefactives are surface-identical in
English. However, the deputative reading can be distinguished from the plain
benefactive by fronting the for-PP. The deputative reading does not survive in
this case (233).
233. For me, she went to the market. plain, *deputative
It is unclear what causes this, and it seems to point to the necessity of distin-
guishing the plain and the deputative benefactive in English (see also Van Valin
and LaPolla (1997)). However, I will ignore this difference between deputative
and plain benefactives from here on because it has no obvious connection to ap-
plicative arguments.
In English, only the recipient benefactive meets my definition of applicative
arguments (4), repeated below, because it does not require a preposition (230c).
234. Applicative Argument
An NP Y of a simple, non-negated declarative sentence that is not gov-
erned by a preposition is an applicative argument iff the sentence with-
out Y does not entail that there is at least one individual that is involved
in the asserted event and could be referred to by Y.

 
122  

The omission test shows that the recipient benefactives of English arguments do
not have to be realized for the sentences to be grammatical (235).
235. John baked a cake.
This sentence does not entail the involvement of another individual. Conse-
quently, the pronoun in (227a) is an applicative argument according to my defi-
nition. This is not the case for English plain and deputative benefactives which
require the presence of for (230c). Thus, they are not applicatives according to
my definition.58 However, the definition for applicative arguments given in
Chapter 1 is only a working definition. It was shown in Chapter 3 that English
affected experiencer applicatives require the presence of the preposition on. Re-
gardless, they were analyzed in analogy to German affected experiencer applica-
tive arguments. Similarly, I provide a brief discussion of the benefactive for-PPs
(and the German equivalent) in section 5.6 because these PPs resemble benefac-
tive applicative arguments.

5.2.2 German
In German, all three types of benefactives (plain, deputative, recipient) can be
expressed as dative-marked constituents (Colleman 2010).
236. a. Jan brachte mir einen Kuchen. recipient
Jan.Nom brought me.Dat a.Acc cake
'Jan brought me a cake.'
b. Jan malte mir ein Bild. plain
Jan.Nom painted me.Dat a.Acc picture
'Jan painted a picture for me.'
c. Jan schnitt mir die Hecke. deputative
Jan. Nom cut me.Dat the hedge
'Jan cut the hedge for me.'
Many sentences are consequently ambiguous between the three types and are
disambiguated by the context (cf. Kittilä (2005), Kittilä and Zuniga (2010)). For
instance, (236b) can alternatively be used as a recipient benefactive (if I re-
ceived the picture) or as a deputative benefactive (if I was supposed to paint a
picture).

58 Colleman (2010) discusses that some dialects of English allow benefactives without
"for", citing examples like:
i. Sam promised to move/crush his lover a mountain.
(Colleman 2010, example 10b, citing Green (1974))
ii. All you have to do to gain my confidence is rob me a couple of banks.
(Colleman 2010, example 10bc, citing Oehrle (1976))
I will not discuss these separately but assume that they can be analyzed like the (equiva-
lent) German benefactives.


  123

All three types of benefactives can alternatively be expressed using a PP


headed by für (cf. Wegener (1983)).
237. a. Jan brachte einen Kuchen für mich.
Jan.Nom brought a.Acc cake for me
'Jan brought me a cake.'
b. Jan malte ein Bild für mich.
Jan.Nom painted a.Acc picture for me
'Jan painted a picture for me.'
c. Jan schnitt die Hecke für mich.
Jan.Nom cut the.Acc hedge for me
'Jan cut the hedge for me.'
Thus, the three types cannot be distinguished based on their form alone.
Besides the context, the verb itself plays an important role in deciding which
function the benefactive applicative carries. Verbs of transfer, communication
and preparation can only project a recipient benefactive applicative argument in
German (cf. Colleman (2010)).
238. Jan brachte mir einen Kuchen.
Jan.Nom brought me.Dat a.Acc cake
'Jan brought me a cake.' recipient
*Jan brought a cake instead of me.
*Jan brought a cake (to someone) for my benefit.
These types of verbs59 seem to have a predisposition for a recipient (or pur-
pose60) to be specified.61 With other verbs, German benefactives are ambiguous
between the three different types of benefactives. For instance, (227b) can have
all three interpretations.
Since some German verbs, e.g. bringen 'bring', entail a recipient, they do not
meet my definition of an applicative argument. I briefly comment on these Ger-
man verbs below but I analyze here only German benefactive constructions with
verbs that do not entail a recipient, e.g. öffnen 'open', schneiden 'cut', malen
'paint'. I call the structure of these verbs with a benefactive applicative argument
a true benefactive construction.


59 Some further examples are sagen 'say', erzählen 'tell', and schicken 'send'.
60 See Schmidtke-Bode (2009) for the details on the relation of benefactives and purpose.
He claims that "benefactive NP-arguments can substitute for an entire purpose clause"
(p. 1).
61 Alternatively, this can be viewed as an instance of the probable/expected participation
of the beneficiary (Kittilä and Zuniga 2010). With these verbs, it is expected that the da-
tive marks a recipient and no other (benefactive) function.

 
124  

These true benefactives of German constitute applicative arguments accor-


ding to my definition (234). Sentence (239a) is grammatical without the applica-
tive argument (239b).
239. a. Jan öffnete der Frau die Tür.
Jan.Nom opened The.Dat woman the door
'Jan opened the door for the woman.'
b. Jan öffnete die Tür.
Jan.Nom opened the door
'Jan opened the door.'
The dative-marked constituent is not governed by prepositions, and (239b) does
not entail the involvement of a beneficiary. Thus, German true benefactives are
applicative arguments according to my definition.
5.3 At-Issue Content
Both English recipient benefactives and German true benefactive arguments in-
volve only at-issue meaning. This can be seen when applying the family of sen-
tence test introduced in Chapter 2. First, it is possible to use benefactives in wh-
questions:
240. a. Who did John bake a cake? -Him.62
b. Wem öffnete Jan die Tür? -Der Frau.
who.Dat opened Jan.Nom the door the woman
'For who did Jan open the door? -The woman.'
This shows that the benefactive arguments are not syntactically inert. This is one
characteristic of at-issue meaning.
Second, nothing projects past yes/no-questions involving recipient benefac-
tives or true benefactives.
241. a. Did John bake him a cake?
b. Öffnete Jan der Frau die Tür?
opened Jan.Nom the woman the door
'Did Jan open the door for the woman?'
The yes/no-questions in (241) ask simply if the event took place. (241a) can be
answered with "no" indicating that John did not bake a cake with the intention of
giving it to him. It is possible that John baked it for himself or with no one in
mind. (241b) can also be answered with "no" indicating that Jan did not open the
door for the woman (regardless of whether Jan opened the door at all). No fur-
ther explanation is needed. Thus, by answering "no" to these questions, the lis-
tener does not accept the truth of anything. This indicates that benefactive ap-
plicatives do not contribute any not-at-issue meaning.

62 This is not grammatical for all speakers.


  125

In negated sentences (242), no benefactive meaning projects past negation.


242. a. John didn't bake him a cake.
b. Jan öffnete der Frau die Tür nicht.
Jan opened the.Dat woman the door Neg
'Jan didn't open the door for the woman.'
(242a) can mean that John did bake a cake but it was not intended for him. The
same pattern of negation holds for (242b); the sentence can mean that Jan
opened the door but did not do it for the benefit of the woman. Thus, no not-at-
issue meaning is projected. This further supports that benefactives contribute
only to the at-issue tier of meaning.
Finally, true and recipient benefactive arguments matter in a conditional
sentence (243); here, when the money will have to be paid.
243. a. If John bakes him a cake, I'll give you $100.
b. Wenn Jan der Frau die Tür öffnet,
if Jan.Nom the woman the door opens
dann gebe ich dir 100Euro.
then give I you 100Euro
'If Jan opens the door for the woman, then I'll give you 100Euros.'
In (243a), there must be at least the intent of the cake being baked for him so
that I have to pay. Similarly in (243b), the woman must benefit from the opening
of the door. If this is not the case, I do not have to pay. This is the expected pat-
tern for at-issue meaning.
In this section, I have shown that English recipient benefactives and German
true benefactive applicative arguments only contribute at-issue content. The tests
did not detect any not-at-issue meaning. In the following sections, these two
constructions are analyzed in turn.
5.4 True Applicative Benefactives
In this section, I discuss true benefactive applicative arguments of German. Pyl-
kkänen (2002) classified these as high applicatives. I show that her analysis is in
fact suited as an analysis of these applicatives (section 5.4.1). Following that, I
discuss some consequences (section 5.4.2) and restrictions (section 5.4.3). I also
address the analysis of benefactives by Hole (2005; 201X) in section 5.4.4.

5.4.1 Analysis
As discussed in Chapter 2, Pylkkänen (2002) proposes an analysis of high ap-
plicatives with an applicative head that attaches between VP and Voice and adds
the argument as a participant to the event. I use a Benefactive (Ben) head (in-
stead of Pylkkänen's general Applicative head). It assigns the thematic role of
Beneficiary. This is taken to cover the plain, deputative and recipient interpreta-

 
126  

tion by marking a vague notion of benefit. Thus, the derivation for (244a) pro-
ceeds as follows.63
244. a. Jan öffnete der Frau die Tür.
Jan opened the woman the door
'Jan opened the door for the woman.'
b. VoiceP<v,t>
3
Jan Voice'<e,vt>
3
Voice<e,vt> BenP<v,t>
3
NP Ben'<e,vt>
der Frau 3
Ben<e,vt> VP<v,t>
3
öffnete NP
die Tür
c. [[VoiceP]] = λe. OPEN(e) & Thm(e)(the door) & Ben(e)(the woman)
& Agt(e)(Jan)
VoiceP denotes a set of opening events with the agent Jan, the theme the door
and the benefaciary the woman. This analysis involves only at-issue content to
capture that nothing projects past negation and yes/no-questions, and that every
element counts for the truth conditions of conditional sentences (section 5.3).
The precise interpretation of "beneficiary" is then provided by the context. For
this sentence, the plain benefactive interpretation is most salient without context;
the deputative one can be made available, while the recipient one is blocked by
world knowledge (opening a door does not lead to a change of possession of the
door).

5.4.2 Consequences
True benefactive applicative arguments do not interfere with other syntactic pro-
cesses, but rather participate in them as expected. For instance, true benefactives
can occur in passivized sentences.
245. Die Tür wurde der Frau (von Jan)
the.Nom door became the.Dat woman by Jan.Dat
geöffnet.
opened
'The door was opened for the woman (by Jan).'

63 The full semantic derivation is given in Appendix 3, #1.


  127

This behavior is expected given the analysis. The main differences between the
active and the passive is at the level of the Voice projection (and the resulting
projection of the agent). This is not expected to interfere with the projection of
the benefactive applicative below Voice.
It is also possible to passivize the benefactive sentence so that the benefi-
ciary is the subject. In this case, the verb bekommen (or colloquially kriegen)
'get, receive' is required instead of werden 'become'.
246. Die Frau bekam/kriegte (von Jan) die
the.Nom woman got by Jan the.Acc
Tür geöffnet.
door opened
'The woman was opened the door by Jan.'
This possibility is remarkable in two ways. First, this distinguishes the be-
nefactive applicative from other applicatives. Affected experiencer applicative
arguments cannot be passivized in this way (247).
247. Chris bekam/kriegte Bens Vase von Alex
Chris got Ben.Gen vase by Alex
zerbrochen.
broken
*'Chris1 was broken Ben's vase on him1 by Alex.'
'Chris1 got Ben's vase broken by Alex for him1.'
The same is true for the Dative of Inaction (248, cf. Chapter 7).
248. Die Oma bekam die Vase zerbrochen.
the.Nom grandma got the.Acc vase broken
*'The vase was broken on grandma's watch.'
'The vase was broken for grandma.'
Neither the subject co-referential applicative argument (249, Chapter 4) nor the
ethical dative (205, Chapter 4) can be passivized.64
249. *Du bekamst einen Kaffee getrunken.
you.Nom got a.Acc coffee drank
'You were drunk you a cup of coffee.'
250. *Ich bekomme von dir pünktlich zur Schule
I.Nom get by you on.time to.the.Acc school
gegangen.
gone
'I was gone to school on time by you.'
The part-whole applicative (Chapter 6) can only marginally be passivized (251).

64 There is a general constraint on the passivization of subject co-referential constituents
that rule these out (Isacenko 1965, Bierwisch 1966).

 
128  

251. ?Das Auto bekam von Jan die Tür zerbeult.


the.Nom car got by Jan the door dented
'The car got the door dented by Jan.'
However, this example can be seen as a case of a benefactive applicative rather
than the part-whole applicative because a material part-whole relation between
the door and the car is not required anymore.
252. ?Das Auto bekam von Jan die Bustür zerbeult.
the.Nom car got by Jan the bus.door dented
'The car got the door of the bus dented by Jan.'
Thus, the generalization is that if passivization of an applicative argument is
possible, only the benefactive reading is available.
Second, the benefactive passive shows (again) that the true benefactive is
not selected by the verb because lexical datives cannot be passivized in German
(253).
253. a. Jan half dem Mann.
Jan.Nom helped the.Dat man
'Jan helped the man.'
b. *Der Mann wurde geholfen.
the.Nom man was helped
'The man was helped.'
(Steinbach 1998, example 38b, my translation)
c. *Der Mann bekam geholfen.
the.Nom man got helped
'The man was helped.'
In other words, the possibility of forming a passive with bekommen (or kriegen)
is a distinct feature of the benefactive.65 (It is not at all clear if sentences like
(247) should really be analyzed as a passive (Vogel and Steinbach 1998). They
might represent a phenomenon different from passives, e.g. a complex predicate
construction (cf. Abraham 2006). Regardless, this behavior sets the benefactive
apart from other applicatives.)
Further support for my analysis of true benefactives comes from again-
modification (as explained in Chapters 2 and 3). There are three nodes of type
<v,t> in (244b) to which again should be able to attach, namely VoiceP, BenP
and VP. Thus, the expected readings for again-modification are:
254. a. repeated VoiceP: the same event with the same agent and beneficiary
has happened before
b. repeated BenP: the same event with the same beneficiary and a dif-
ferent agent has happened before

65 Examples such as (238) may also be passivized with bekommen/ kriegen.


  129

c. repeated VP: the same event with different beneficiary and agent has
happened before
This prediction is borne out.
255. Jan öffnete der Frau die Tür wieder.
Jan.Nom opened the.Dat woman the door again
'Jan opened the door for the woman again.'
(255) can have all of the expected interpretations (although the repeated VoiceP
is preferred and more salient with wieder 'again' in the sentence-initial position).
Also, as expected, the sentence cannot mean that the woman had benefitted from
a different event before and now benefitted from the opening of the door.
Thus, this analysis of true benefactives as high applicatives finds support in
the again-modification and passivization facts.

5.4.3 Verbal Restrictions


There are some restrictions on the true benefactive applicative that are not re-
flected in the analysis as it stands. True benefactive applicatives cannot appear
with most intransitive verbs (256a,b) and only marginally with ditransitive verbs
in German (256c).
256. a. *Der Mann rannte (der) Maria.
the.Nom man ran the.Dat Maria
'The man ran for Maria.'
b. *Der Mann trank (der) Maria.
the.Nom man drank the.Dat Maria
'The man drank for Maria.'
c. ??Der Mann gab ihr (der) Maria
the.Nom man gave her.Dat the.Dat Maria
ein Buch
a.Acc book
'The man gave her a book for Maria.'
'The man gave Maria a book for her.'
There are intransitive verbs that are marginally acceptable with true benefac-
tive applicatives, namely those that can be characterized as controlled, unaffect-
ing events according to Sorace (2000).66
257. a. ?Das Paar hat mir getanzt.
the.Nom couple has me.Dat danced
'The couple danced for me.'
(controlled, motional, atelic)

66 Other interpretations of the applicative (e.g. as an affected experiencer) may be possible
in these examples.

 
130  

b. ?Das Paar ist mir in den Raum getanzt.


the.Nom couple is me.Dat into the.Acc room danced
'The couple danced into the room for me.'
(controlled, motional, telic)
c. ??Das Kind hat mir (so schön) gespielt.
the.Nom child has me.Dat so nicely played
'The child played for me nicely.'
(controlled, nonmotional, unaffecting)
Yet, other intransitives are not allowed (258).
258. a. Der Baum ist mir verrottet.
the.Nom tree is me.Dat rotten
*'The tree rotted for me.'
'The tree rotted on my watch.'
(change of state)
b. Der Hund hat mir überlebt.
the.Nom dog has me.Dat survived
*'The dog survived for me.'
'The dog survived on my watch.'
(continuation of condition)
c. Der Hund hat mir existiert.
the.Nom dog has me.Dat existed
'The dog existed for me.'
(existence of state)
d. Der Hund ist mir ertrunken.
the.Nom dog is me.Dat drowned
*'The dog drowned for me.'
'The dog drowned on my watch.'
(inchoative)
e. *Der König hat mir triumphiert.
the.Nom king has me.Dat triumphed
'The king triumphed for me.'
(controlled, affecting, process)
f. *Das Kind hat mir gehustet.
the.Nom child has me.Dat coughed
'The child coughed for me.'
(acceptable as controlled process only)
None of these examples can be made acceptable as a benefactive construction by
providing more context. Thus, this is a syntactic and not a pragmatic restriction.
Also, not all transitive verbs are acceptable with true benefactive applicative
arguments. Colleman (2010) cites Hens (1995) for the observation that only


  131

highly transitive verbs (according to the transitivity scale by Hopper and


Thompson (1980)) allow the applicative benefactive.67 This explains, for in-
stance, the grammaticality contrast in (259).
259. a. Otto beobachtet die Kollegen für seinen Chef.
Otto observes the colleagues for his boss
'Otto observes the colleagues for his boss.'
b. *Otto beobachtet seinem Chef die Kollegen.
Otto observes his.Dat boss the colleagues
'Otto observes the colleagues for his boss.'
(Colleman 2010, example 14, citing Wegener (1991))
Beobachten 'observe' is low on the transitivity scale because it is atelic and non-
punctual with low agency and the object is not affected; therefore, it does not
license the true benefactive applicative. It does allow the prepositional benefac-
tive paraphrase. This is also true for other verbs that do not allow true benefac-
tive applicatives. Compare (256) and (260).
260. a. Der Mann rannte für Maria.
the man ran for Maria
'The man ran for Maria.'
b. Der Mann trank für Maria.
the man drank for Maria
'The man drank for Maria.'
c. Der Mann gab (der) Maria ein Buch für sie.
the man gave the.Dat Maria a book for her
'The man gave Maria a book for her.'
The availability of the prepositional paraphrase shows that the described verbal
restriction is syntactic and not semantic. However, the analysis of true benefac-
tives presented here does not explain this restriction. Rather, it is expected that
the Ben head can select any VP including intransitives. This problem of exactly
which VPs are acceptable with true benefactive applicatives and how to incorpo-
rate that in the analysis is left to future research.

5.4.4 Hole (2005; 201X)


Hole (2005; 201X) also provides an analysis of German benefactive applicative
arguments. He argues that benefactive can be characterized as "(intended) af-
fectedness plus purposivity" (Hole 2005, p. 23). This aligns the benefactives
with the (affected) experiencer, and Hole provides a similar binding approach
(see section 3.6.2.1). Yet, according to Hole, the benefactive does not require a

67 See Zuniga (2007) citing Shibatani (1996) for the acceptability of different predicates
across languages.

 
132  

possessor but a purpose to be bound by the applicative. He explains this with the
following example.
261. Popeye hat Olive Oylx einen Steinx sauber gewischt.
Popeye has Olive Oyl.Dat a.Acc stone clean wiped
'Popeye wiped off a stone for Olive Oyl.'
(Hole 2005, example 43b)
Since it is not immediately obvious from this example what the benefit or pur-
pose of wiping off a stone is, the listener is forced to accommodate the scenario
by inventing a purpose; for instance, Popeye wiped the stone clean so Olive Oyl
could sit down. The applicative argument Olive Oyl then binds a variable in the
constituent einen Stein which represents the purpose. Hole further supports this
view of benefactive applicatives binding a purpose with the following example
which includes an overt purpose phrase.
262. Olive Oyl komponiert ihrem Babyx ein
Olive Oyl composes her.Dat baby a.Acc
Lied zum Vorsingenx bei Omas Geburtstag.
song to.the perform at granny's birthday.Dat
'Olive Oyl is composing a song for her baby to perfom at granny's
birthday.'
(Hole 2005, example 44)
The salient reading of this sentence is that the baby is supposed to perform the
song at the birthday party, regardless of the world knowledge that babies cannot
sing. Alternatively, the sentence can be interpreted as the song to be performed
by someone else and be, for instance, dedicated to the baby. From this, Hole
concludes that benefactive applicatives always bind a variable of a purpose
phrase because on either interpretation the applicative ihrem Baby binds a varia-
ble, either the overt purpose "perform" (262) or an unpronounced, accommoda-
ted purpose of the song (261).
I agree with Hole that the benefit in (261) is not immediately obvious (with-
out context). However, I attribute that to the general notion of benefit that is in-
cluded in the Ben head. The benefit can be anything, including the deputative
interpretation that Olive Oyl was supposed to do it but Popeye did it for her. I
contend that there is no purpose to be bound (but without context the exact na-
ture of the benefit cannot be determined).
In contrast to Hole, I argue that the salient reading of (262) (in which the
baby is supposed to perform) does not include an applicative argument. It does
not meet my definition of an applicative argument (4), which is based on Hole
(2008). The dative constituent in (262) is not governed by a preposition, and the
sentence is grammatical without the dative constituent (263).


  133

263. Olive Oyl komponiert ein Lied zum Vorsingen


Olive Oyl composes a.Acc song to.the perform
bei Omas Geburtstag.
at granny.Gen birthday
'Olive Oyl is composing a song to be performed at granny's birthday.'
However, (263) entails that someone is supposed to perform the song. This per-
son can be expressed as a dative constituent in German, as the salient reading of
(262) shows. Thus, on the salient reading, the dative-marked constituent of (262)
does not meet my definition of applicative arguments because the existence of a
"purpose fulfiller" is entailed. The entailment can be seen from the fact that even
if it is not explicitly specified who is supposed to perform the song, the exist-
ence of a performer is assumed. This is exactly what is happening on the alterna-
tive reading of (262). A performer is still assumed to exist, even though it is not
stated who that is. On the alternative reading of (262), the dative constituent is
indeed a benefactive applicative argument (with the intended performer(s) not
being expressed and the benefit not being exactly specified). The presence of the
purpose phrase (in (262) and (263)) can license the presence of a dative consti-
tuent which denotes who is to fulfill the stated purpose.68
As with the pertinence dative (section 3.6.2.1), Hole (201X) observes a
sloppy reading requirement with benefactives. He presents the following exam-
ple.
264. J.R. mixte Sue-Elleni einen Drink (zuri/ zu
J.R. fixed Sue-Ellen.Dat a.Acc drink for.the for
ihreri) Entspannung, und seiner Mutter auch.
her relaxation and his.Dat mother too
'J.R. fixed Sue-Ellen a drink so that Sue-Ellen could relax, and J.R.
fixed his mother a drink so that his mother could relax.'
*'J.R. fixed Sue-Ellen a drink so that Sue-Ellen could relax, and J.R.
fixed his mother a drink so that Sue-Ellen could relax.'
Hole states that the strict reading is not available, indicating that the benefactive
binds the purpose.69 It is true that, if the sentence is understood as Sue-Ellen re-
laxing, it has to be the mother who relaxes as well. She cannot benefit from the
drink for Sue-Ellen's relaxation. However, other interpretations of the sentence
are possible. For instance, with zur, the sentence can mean that J.R. fixed the
drinks so that he himself can relax.70 In that case, Sue Ellen and his mother are

68 This optional dative constituent can seen as being licensed in a similar fashion as the
dative constituent of a judgment dative (Chapter 1). It can either be overtly stated or be
covert (and entailed).
69 Hole assumes binding into (bridging) definites.
70 Note the grammaticality of the following example.

 
134  

beneficiaries (most saliently interpreted as recipients). In other words, we see


the same pattern as above: a "purpose fulfiller" is entailed and either explicitly
stated (Hole's sloppy identity reading) or implicit and a beneficiary is included
in the sentence. For the sloppy identity requirement of (264), I assume the iden-
tity of the elided constituent; in particular, I assume that the dative of the second
conjunct has the same role as the one in the first conjunct, namely "purpose ful-
filler".71
In addition, Hole's approach for benefactives runs into the same problem as
his binding approach for pertinence datives (section 3.6.2.1), namely that unpro-
nounced elements must be posited freely. Example (265) is identical to (264)
except a purpose is overtly stated in the second conjunct.
265. J.R. mixte Sue-Ellen einen Drink zu ihrer Entspannung
J.R. fixed Sue-Ellen.Dat a.Acc drink for her relaxation
und seiner Mutter auch zu Sue-Ellens
and his.Dat mother too for Sue-Ellen.Gen
Entspannung.
relaxaction.
'J.R. fixed Sue-Ellen a drink so that she could relax, and he fixed one
for his mother too so that Sue-Ellen could relax.'
In this example, Hole must posit an unpronounced purpose element inside Sue-
Ellens Entspannung 'Sue-Ellen's relaxation' which can be bound by seiner Mut-
ter 'his mother'. It is unclear what this element would be. On my approach,
seiner Mutter 'his mother' just needs to denote the beneficiary of the event of
J.R. fixing a drink so that Sue-Ellen can relax. No binding is necessary.
In summary, it is not necessary to assume that benefactive applicative argu-
ments bind a purpose phrase (either overt or unpronounced). My simple analysis
(based on Pylkkänen (2002)) can explain the phenomenon of benefactive appli-
cative arguments well.


i. J.R. mixte Sue-Ellen einen Drink zu seiner (eigenen)
J.R fixed Sue-Ellen.Dat a.Acc drink for his own
Entspannung, und seiner Mutter auch.
relaxation and his.Dat mother too
'J.R. fixed Sue-Ellen a drink so that he could relax, and J.R. fixed his mother a drink
so that he could relax.'
71 The same pattern holds for a similar sentence in English:
i. J.R. fixed a drink for Sue Ellen so that she could relax, and he did for his mother, too.
This sentence only has the sloppy interpretation; it cannot mean that he fixed the drink
for his mother so that Sue Ellen can relax. If Hole's analysis can be extended to English,
he must assume that Sue Ellen binds into the purpose clause.


  135

5.4.5 Summary
I have shown that true benefactive applicative arguments contribute only at-
issue meaning. They are introduced by a Ben(efactive) head which attaches be-
tween VP and Voice and provides the thematic role Beneficiary. This is the for-
mal structure proposed for high applicatives by Pylkkänen (2002). Many beha-
viors of this structure predicted by my analysis can be observed, such as again-
modification and the passive formation. However, it does not fully reflect the
verbal restrictions that true benefactive applicative arguments face.
5.5 Recipient Benefactives
In this section, I discuss English recipient benefactive applicative arguments
(228a), repeated below. These are low applicatives for Pylkkänen (2002). Reci-
pient benefactives and ditransitives (267) share their surface structure and be-
have alike in many respects.
266. Recipient Benefactive:
John baked him a cake.
267. Ditransitive:
John gave Mary a book.
I argue that this similarity is due to the fact that ditransitives and recipient bene-
factives have the same structure. The difference is that ditransitive verbs obliga-
torily project this structure while recipient benefactives are optionally added to a
transitive verb, which results in the ditransitive structure. Before presenting my
analysis, I discuss the meaning of the recipient benefactive and some of its ver-
bal constraints (section 5.5.1). I also show how ditransitives and recipient bene-
factives resemble each other (section 5.5.2) before presenting my analysis (sec-
tion 5.5.3) and its consequences (section 5.5.4).

5.5.1 Meaning
The recipient benefactive is similar in its interpretation to that of an indirect ob-
ject of a ditransitive verb; both involve the (intended) transfer of possession
(Pylkkänen 2002). In general, a recipient does not always have to benefit (e.g.
John gave Mary poisonous cookies) and a beneficiary does not always receive
anything (cf. the definition in section 5.1 where receiving is not listed as a fea-
ture of benefactives). Yet, a typical recipient benefactive involves the recipient
receiving something and therefore being assumed to benefit. This type of bene-
factive can consequently most often be observed with verbs of transfer, verbs of
communication, and verbs of prevention (Colleman 2010). This generalization
rules out, for instance, sentence (268).

 
136  

268. *John opened Mary the door.


(Colleman 2010, example 4a, citing Shibatani (1996))
Mary does not receive anything from John's action of opening the door. Conse-
quently, the sentence is not acceptable (as a recipient benefactive).
The recipient benefactive can also be observed with some verbs of creation
or preparation and verbs of obtaining. However, it is not necessary that a verb
belongs to any of these verb classes to license a recipient benefactive (Colleman
2010). This can be seen from the example (269).
269. a. *I cleared him the floor.
b. I cleared him a place to sleep on the floor.
(Colleman 2010, example 5, citing Langacker (1991))
This example shows that the recipient benefactive does not have to receive a
physical entity. Furthermore, the recipient does not have to be the permanent
possessor. Temporary possession (269), immaterial possession (270b) or even
just intended possession (271) is sufficient to license the recipient benefactive
(Colleman 2010).
270. a. Can you call me a taxi please?
b. Frederick sang Gertrude a love song.
(Colleman 2010, example 6)
271. I baked you a cake but then I dropped it on the floor.
In summary, the recipient benefactive requires the denoted entity to receive
something (denoted by the direct object). Yet, the exact specification of "bene-
fit" and "receiving" are hard to pin down, as they cover (among other things)
intended and immaterial possession.

5.5.2 Ditransitives and Benefactives


As discussed above, the recipient benefactive argument and the indirect object
of a ditransitive resemble each other semantically as both require the (intended)
transfer of possession. There is also syntactic resemblance. The indirect object
asymmetrically c-commands the direct object in ditransitives (among others
Barss and Lasnik (1986), Larson (1988)). The same is true in recipient benefac-
tive constructions; the recipient benefactive applicative c-commands the direct
object. The recipient benefactive can bind a variable inside a direct object but
the direct object cannot bind a variable inside the recipient benefactive (272b).
272. Variable Binding
a. i. I gave every girl1 her1 favorite kind of cookie.
ii. *I gave its1 #1 fan every kind of pie1.
b. i. I baked every girl1 her1 favorite kind of cookie.
ii. *I baked its1 #1 fan every kind of pie1.


  137

Due to the similar behavior and their similar meaning, I argue in the following
section that recipient benefactives and ditransitives have the same structure.

5.5.3 Analysis
I use the analysis of ditransitives by Bruening (2010) as my analysis of recipient
benefactives. It relies on an unpronounced head that I call Rec(ipient). Its deno-
tation is given in (273). It provides only at-issue meaning.
273. [[Rec]] = λx.λy.λe. HAVE(e) & Thm(e)(x) & Recipient(e)(y)
The syntactic structure and the semantic derivation for the recipient benefac-
tive (274a) are given in (274b) and (274c) and are explained in the following.72
274. a. Mary baked him a cake.
b. VoiceP
3
Mary Voice'
3
Voice vP1
3 2
v Voice him1 vP2
3 2
Rec v a cake v'
2 2
V Rec v RecP
2 2
Rec v t1 Rec'
2 2
V Rec Rec VP
2 6
V Rec bake a cake
c. [[VoiceP]] = λe. BAKE(e) & Thm(e)(a cake) & Agt(e)(Mary) &
(e') [HAVE(e') & Thm(e')(a cake) & Recipient(e')(him) &
CAUSE(e')(e)]
The Recipient head attaches outside of the VP. In this way, the recipient bene-
factive is not different from the true benefactive construction. However, the re-
cipient benefactive construction differs from the true benefactive in other ways.
Syntactically, there are three main movement processes going on: (a) the lexical
verb undergoes head movement to the Rec head, v, and Voice; (b) the direct ob-
ject moves from its base position to the specifier of v; (c) the applicative argu-


72 The full semantic derivation is given in Appendix 3, #2.

 
138  

ment moves from its base position as the specifier of Rec to an adjoined vP posi-
tion. This structure reflects the c-command relations (272).
Semantically, the challenge is to encode the transfer of possession. The Rec
head carries this meaning but it requires functional composition73 for computing
the semantic derivation. The verb is not interpreted in its base position. This al-
lows the Rec head to have semantic access to the direct object, which is its first
semantic argument (273). The second argument is the recipient benefactive ar-
gument in the specifier position of Rec. Next, v is merged and, by head move-
ment, the lexical verb (together with Rec) moves there. This is where the lexical
verb is interpreted. The moved direct object is then interpreted (again) in the
specifier of vP position. The trace of the recipient benefactive is interpreted
next, before Voice introduces the agent. This finally yields the correct denota-
tion of the sentence (274c) which includes the possessive relationship between
the direct and the indirect objects. It states that there is a cake baking event of
which Mary is the agent. This event causes him to have (and be the recipient of)
a cake.

5.5.4 Consequences
This analysis reflects the fact that the recipient benefactive applicative argument
is not an argument of the verb and is consequently optional (235). Furthermore,
it explains the restricted occurrence of the recipient benefactive. Due to the fact
that recipient benefactives require the denoted entity to be the intended posses-
sor of the direct object, they cannot appear with unergative verbs because a di-
rect object must be present (cf. McGinnis 2001).
275. *John ran him.
However, my analysis does not explain why recipient benefactives cannot
occur with unaccusative verbs. Example (276) cannot mean that the ice melted
into his possession.
276. *The ice melted him.
It is not clear what rules out that Rec attaches outside an unaccusative VP. It
should be possible for the object to move through vP to VoiceP and to become
the subject. Rec (head-)moves along with V. It is possible that it is a semantic
clash that rules this out, namely that the referent of the direct object is the agent
who causes him-/her-/itself to come into possession of someone else in this con-
figuration (cf. Bruening 2011).


73 Functional Composition allows for the semantic computation to look ahead in the struc-
ture to ensure that the derivation will not crash due to type mismatches later on in the
computation. It allows certain operations to be postponed in order to salvage the deriva-
tion.


  139

Finally, the recipient benefactive cannot appear with ditransitives (277).


This is expected if ditransitives and recipient benefactives have the same struc-
ture, and the Rec head must select VP. There is only one VP. Thus, only one of
the two (ditransitive, recipient benefactive) can be generated.
277. *John gave me Mary a book.
In other words, the ditransitive structure is not recursive (see also Chapter 8).
As for again-modification (see Chapters 2 and 3), there should be three
readings available, namely the repeating of VoiceP, RecP, and vP (cf. Bruening
(2010)). This prediction is borne out (278).
278. Mary baked him a cake again.
a. repeated RecP: He had a cake before.
b. repeated vP: Someone had baked him a cake before.
c. repeated VoiceP: Mary had baked him a cake before.
I claim that ditransitives such as give obligatorily project the proposed struc-
ture whereas other (English) verbs optionally project this structure, namely as
the applicative recipient benefactive structure.
German verbs that allow only the recipient benefactive (238) also project
this structure but allow the specifier of Rec to be phonologically null (existen-
tially quantified). There is, however, one difference between the ditransitive and
these recipient benefactive in German, namely the ability to form passives. Only
the recipient benefactive can be passivized with bekommen/ kriegen 'get, receive'
(279).
279. a. Maria bekam (von Jan) eine Vase gebracht.
Maria.Nom got by Jan a.Acc vase brought
'Maria was brought a vase (by Jan).'
b. *Maria bekam (von Jan) eine Vase gegeben.
Maria.Nom got by Jan a.Acc vase given
'Maria was given a vase (by Jan).'
In this respect, the recipient benefactive applicative behaves just like the true
benefactive applicative argument (section 5.4.2). This difference between the
recipient benefactive applicative and the ditransitive constructions might point
to them requiring different analyses. If they have the same structure, they are
expected to behave alike with respect to passivization. Yet this difference can
also be taken as support that the bekommen-passive of German is not a passive
of a benefactive structure but rather a different structure altogether (see among
others Vogel and Steinbach 1998). The addition of a particle to geben 'give' can
make the passive available.
280. Maria bekam (von Jan) eine Vase herausgegeben.
Maria.Nom got by Jan a.Acc vase out.given
'Maria was given out a vase (by Jan).'

 
140  

Particles can change a verb's argument structure. In this case, it changes the
ditransitive geben 'give' to a recipient benefactive structure in which the recipi-
ent is only optionally specified (281).
281. a. Jan gab *(Maria) eine Vase.
Jan.Nom gave Maria.Dat a.Acc vase
'Jan gave Maria a vase.'
b. Jan gab (Maria) eine Vase heraus.
Jan.Nom gave Maria.Dat a.Acc vase out
'Jan gave out a vase to Maria.'
Thus, the fact that only the true ditransitive geben 'give' cannot be passivized
remains puzzling.

5.5.5 Summary
Recipient benefactive applicatives are structurally different from true benefac-
tives but akin to ditransitives. They are introduced by a Rec(ipient) head which
selects VP. Functional composition ensures the interpretation of the direct object
as coming into possession of the entity referred to by the recipient benefactive.
Due to the fact that recipient benefactive applicatives and ditransitives are given
the same structure, their similarities are expected.
5.6 Prepositional Alternation
In this section, I address the prepositional alternative to the applicative benefac-
tives. I argue that for and its German equivalent für are vague benefactive pre-
positions which can be used for plain, deputative, and recipient benefactives (cf.
Parsons (1990) on the ambiguity of prepositions in general). I disregard other
uses of these prepositions such as introducing a purpose (Schmidtke-Bode
2009). The analysis provided here is similar to the cross-linguistic variation of
the applicative affected experiencer in German and its prepositional variant in
English (Chapter 3). While discussing the similarities of the applicative and the
prepositional benefactives, I also show that a PP-shell analysis of the true bene-
factive dative is not sufficient (section 5.6.3). This supports my contention that
the benefactive applicative arguments and their prepositional variants can re-
ceive similar, though not identical, analyses.

5.6.1 Analysis
The prepositional variant of the benefactive differs from its corresponding ap-
plicative in two major ways. First, the presence of the preposition is obligatory
(and consequently it forms a prepositional phrase in which the beneficiary is in-
cluded). Second, the PP is an adjunct of the sentence. In other respects, the
prepositional variant is exactly like the corresponding applicative, e.g. in the


  141

contribution of only at-issue meaning as there is no meaning that projects past


negation (282).74
282. a. Jan öffnete die Tür nicht für die Frau.
Jan.Nom opened the.Acc door Neg for the.Acc woman
'Jan didn't open the door for the woman.'
b. Mary didn't bake a cake for him.
In both cases, it is possible that the event did happen but not with a benefit for
the referent included in the benefactive PP. This is the same pattern observed for
the applicatives (242) and indicates that there is no not-at-issue meaning.
Similarly, no meaning survives in yes/no-questions (283).
283. a. Öffnete Jan die Tür für die Frau?
opened Jan.Nom the.Acc door for the woman
'Did Jan open the door for the woman?'
b. Did Mary bake a cake for him?
Both questions can be answered with "no" and do not require further elaboration
to indicate that the event did happen but did not have that person as a benefi-
ciary. No truth of anything is accepted if the questions are answered with "no".
Again, this is typical behavior for at-issue meaning, and the corresponding ap-
plicatives behave alike (241).
The prepositional benefactives can be used for wh-questions (284).
284. a. Für wen öffnete Jan die Tür?
for who.Acc opened Jan the.Acc door
'For who did Jan open the door?'
b. For whom did Mary bake a cake?
This again shows that the prepositional variants behave like the corresponding
applicatives (240) and contribute only at-issue meaning.
Finally, the prepositional benefactives contribute to the conditions of an if-
clause:
285. a. Wenn Jan die Tür für die Frau
if Jan.Nom the.Acc door for the woman
öffnet, geb ich dir 100Euro.
opens give I you.Dat 100Euros
'If Jan opens the door for the woman, I will give you 100Euros.'
b. If Mary bakes a cake for him, I will give you $100.
In both sentences, the prepositional benefactives contribute to the conditions: the
money needs to be paid if the woman benefits/ if he is the (intended) recipient of


74 I use German examples for true benefactives and English ones for recipient benefac-
tives, but the facts hold for both languages.

 
142  

the cake. Thus, in terms of meaning contribution the applicative (243) and the
PP behave alike.
I propose that the preposition for/ für takes the beneficiary as its first seman-
tic argument. It assigns the thematic role of Ben(eficiary) and adjoins to the VP:
286. [[for/für]] = λx.λP<v,t>.λe. Ben(e)(x) & P(e)
The denotation of the preposition is akin to the German true benefactive in that
it can be interpreted as deputative, plain or recipient depending on the context
(244, see below for some restrictions on the interpretation).75 A sample deriva-
tion for German is given in (287).76
287. a. Jan öffnete die Tür für die Frau.
Jan.Nom opened the.Acc door for the.Acc woman
b. VoiceP<v,t>
3
Jan Voice'
3
Voice VP<v,t>
3
PP VP1
3 6
P NP öffnete die Tür
für die Frau
c. [[VoiceP]] = λe. Ben(e)(the woman) & OPEN(e) & Thm(e)(the door)
& Agt(e)(Jan)
The benefactive PP is adjoined to VP. The benefactive preposition takes the
beneficiary as its first argument and the VP as its second argument. The result-
ing derivation is that there was an opening event of which Jan is the agent, the
door is the theme, and the woman is the beneficiary.
The English prepositional benefactive works in the same way.
288. a. Mary baked a cake for John.


75 English for can also mark plain and deputative benefactives (see section 5.2.1). Thus,
for and für are alike in their underspecification for a particular type of benefactive.
76 The full semantic derivation is given in Appendix 3, #3.


  143

b. VoiceP<v,t>
3
Mary Voice'
3
Voice VP<v,t>
3
PP VP1
3 6
P NP bake a cake
for John
c. [[VoiceP]] = λe. Ben(e)(John) & BAKE(e) & Thm(e)(a cake) &
Agt(e)(Mary)
There is a baking event with the agent Mary, the theme a cake and the benefi-
ciary John. The context will provide the correct interpretation of whether it is a
plain benefactive, a deputative benefactive, or a recipient benefactive.

5.6.2 Consequences
This analysis of prepositional benefactives has some welcome consequences.
First, it explains why the German verbs that entail a possessor cannot have a für
-PP denoting a(n intended) recipient (289).
289. Jan brachte Maria einen Kuchen nach
Jan.Nom brought Maria.Dat a.Acc cake to
Dortmund für seine Mutter.
Dortmund for his mother.
'Jan brought Maria a cake for his mother's benefit to Dortmund.'
*'Jan brought his mother a cake for Maria's benefit to Dortmund.'
The für-PP is an adjunct. In this case, it adjoins outside RecP (which has to be
present for German verbs that entail a recipient). The unique theta role assign-
ment requirement will rule out the recipient interpretation of the prepositional
phrase because there is already a recipient present, namely the applicative. Fur-
thermore, the applicative in this case cannot receive an interpretation other than
that of a recipient because of the verb's predisposition for a recipient. The PP für
seine Mutter must be interpreted as a plain or a deputative benefactive because
the dative constituent provides a recipient. Consequently, the recipient benefac-
tive is ruled out. In contrast, any für-PP can attach outside the VP of other verbs,
allowing all interpretations (290).
290. Jan buk für Maria einen Kuchen.
Jan.Nom baked for Maria a.Acc cake
'Jan baked a cake for Maria.' (plain, recipient, deputative)

 
144  

If an applicative is present with verbs not predisposed for a recipient inter-


pretation, the für-phrase can still only take in the theta role not taken by the ap-
plicative argument, but in this case the role of the applicative is not fixed (291).
291. Jan buk Maria einen Kuchen für seine Mutter.
Jan.Nom baked Maria.Dat a.Acc cake for his mother
'Jan baked Maria a cake for his mother.'
'Jan baked his mother a cake for Maria.'
Example (291) is ambiguous between the different possible meanings of the ap-
plicative and the benefactive PP, and any interpretation is possible.
The same is true in English. When the RecP is present, the possessive inter-
pretation of for is ruled out (292).77
292. Mary baked him a cake for his mother.
In this example, the mother cannot be interpreted as the recipient because that
role is assigned to him.
My analysis of benefactive prepositions also explains why they are not re-
stricted in their verbal selection (see section 5.4.3). The PP can unselectively
adjoin to any VP.
This analysis does not readily extend to prepositional (recipient) benefac-
tives included in an NP, as exemplified in (293).
293. Den Kuchen für Maria habe ich gestern gebacken.
the.Acc cake for Maria have I yesterday baked
'I have baked [the cake for Maria] yesterday.'
This is due to the preposition's semantic denotation which requires an individual
argument first and then a predicate. For the constituent-internal use, it should
take two individual arguments. Future research will show if the two uses can
(and should) receive the same analysis.

5.6.3 PP-shells
An alternative approach to analyzing the prepositional and true benefactives
would be to argue that the true benefactive has, in fact, the same structure as the
prepositional one. That is, both are PPs but that the preposition of the true bene-
factive is unpronounced.78
In fact, Rezac (2008) argues that all NPs have a PP-shell, i.e. an unpro-
nounced preposition heading a PP, if they are assigned theta-related case. This is
schematized in (294).


77 The full syntactic and semantic derivations are given in Appendix 3, #4.
78 This section is a short version of Bosse (2013a). There, I characterize applicative argu-
ments as DPs but describe them as NPs here. Nothing hinges on the classification.


  145

294. PP
2
P NP

Following Woolford (2006), theta-related case is defined as non-structural
case that is preserved in A-movement. That the true benefactive receives theta-
related case can be seen from example (295) in which the dative case of the true
benefactive is preserved in the passivized sentence.
295. a. Susi malte ihrer Mutter ein Bild.
Susi.Nom painted her.Dat mother a.Acc picture
'Susi painted a picture for her mother.'
b. *Ein Bild wurde ihrer Mutter (von
a.Nom picture was her.Dat mother by
Susi) gemalt.
Susi painted
'A picture was painted (by Susi) for her mother.'
(Bosse 2013a, examples 2, 5)
Assuming Rezac (2008) is correct, it is expected for the true benefactive to be
included in a PP-shell. It should follow then that both the true benefactive and
the prepositional benefactive are PPs, differing only in whether the preposition
is pronounced.
The simplest hypothesis is that these two PPs (the true benefactive and the
prepositional benefactive) are syntactically alike: they attach in the same posi-
tion and are available for the same syntactic transformations. Unfortunately, this
hypothesis cannot be maintained because the two types of benefactives behave
differently with respect to extraposition. While it is possible to extrapose a pre-
positional benefactive (296b), a true benefactive cannot be extraposed (296a).
296. a. *Susi hat Bilder gemalt ihrer Mutter.
Susi had pictures painted her.Dat mother
b. Susi hat Bilder gemalt für ihre Mutter.
Susi had pictures painted for her mother
'Susi had painted pictures for her mother.'
(Bosse 2013a, examples 12, 13)
In short, (296) shows that the true benefactive and the prepositional benefactive
do not behave alike syntactically. This points to them requiring different ana-
lyses.
Since it cannot be maintained that the true benefactive and the prepositional
benefactive are syntactically alike, there are two possible alternative hypotheses:
first, both types of benefactives could be PPs but they are generated in different
syntactic positions resulting in different extraposition possibilities, or, second,

 
146  

they are not both PPs and the extraposition facts follow from their different syn-
tactic status. Before I explore these hypotheses below, I present an analysis of
extraposition proposed by Kasai (2008) which will help to decide between these
two hypotheses.
Kasai (2008) argues that extraposition (in English79) involves the following
steps: first, the constituent to be extraposed is included in a focus projection
(FocP) headed by a phonologically null Foc(us) head. This phrase moves to the
edge of vP (due to an uninterpretable Foc-feature). After moving, the FocP pro-
jects again, resulting in vP being the specifer of the newly-projected FocP. The
relevant tree is given in (297).
297. FocP
3
FocP1 vP
2 2
Foc DP DP v'
3
v VP
2
t1 V'
2
V PP
(Kasai 2008, example 15c)
The result of projecting FocP is apparent rightward movement because the spe-
cifier (vP) will be linearized as preceding the complement.
If we assume that true benefactives and prepositional benefactives are PPs
but they attach at different points in the derivation, it seems reasonable to pro-
pose that the true benefactive attaches below the subject while the prepositional
attaches above the subject, given the again-modification facts of true benefac-
tives (254, 255). This hypothesis is not supported by the extraposition facts ei-
ther. Both types of benefactives are generated in the same phase, namely vP, and
should be able move to the edge of vP for extraposition (following the analysis
proposed by Kasai (2008)). Yet, as (297) shows, the two structures are not
equally available for extraposition. Thus, the hypothesis that the true benefactive
and the prepositional benefactive are both structurally PPs cannot be maintained.
As a result, the hypothesis that receives the most support is the one stating
that the true benefactive is an applicative argument and by that a nominal pro-
jection included in an applicative projection with an unpronounced head (as
proposed in section 5.4.1), and the prepositional benefactive is an adjoined PP

79 I assume that this analysis of English extraposition extends to German.


  147

(without an applicative projection) (as proposed in section 5.6.1). This hypothe-


sis can explain the extraposition facts (296) because the applicative argument
(true benefactive) behaves like other nominal projections in extraposition, and
the prepositional benefactive behaves like other PPs in extraposition: German
PPs can easily be extraposed whereas NPs can only rarely be extraposed (and
are highly marked when they are extraposed) (Scherpenisse 1985).
In short, a unified analysis for the true benefactive and the prepositional var-
iant cannot be maintained. The true benefactive does not have a PP-shell (unlike
what is expected under the proposal by Rezac (2008)) and does not syntactically
behave like a PP but rather is a nominal projection, an applicative argument,
which needs to be treated differently than the prepositional benefactive.
5.7 Conclusion
In this chapter, I have provided analyses for the true and the recipient benefac-
tive applicative arguments. The former are analyzed as Pylkkänen's (2002) high
applicatives, whereas the latter are analyzed akin to ditransitives (Bruening
2010). Both of these applicative arguments contribute only at-issue meaning.
Furthermore, I have discussed the prepositional variant of the applicatives. I
have argued that the benefactive preposition (English for/ German für) is am-
biguous between the different benefactive functions (plain, recipient, deputa-
tive). The prepositional benefactive is less restricted in its distribution than the
applicative benefactive. While it is not clear what limits the applicative, the wide
distribution of the prepositional benefactive is explained. I have also shown that
it is not feasible to explain the prepositional variant and the applicative benefac-
tive with identical analyses.
In the next chapter, I will investigate another type of at-issue applicative ar-
gument, the part-whole applicative. It will be shown that it differs from the be-
nefactive in many respects, despite contributing to the same tier of meaning.

 
6 Part-Whole Applicatives
6.1 Introduction
In this chapter, I provide an analysis of German applicative arguments that re-
ceive a part-whole interpretation (298).
298. Jan zerbeulte dem Auto die Tür.
Jan.Nom dented the.Dat car the.Acc door
'Jan dented the door of the car.'
To be precise, the applicative argument denotes an entity of which an entity de-
noted by another NP is a material part; in this example, the door is understood to
be a part of the car. These applicative arguments are not found in English; I use
a possessive construction in the translation.80
The part-whole applicative argument has first been extensively discussed as
a separate type of German applicatives by Hole (2006; 2008). He claims that this
type is only accepted by some native speakers of German, which might explain
why they have been described as being ungrammatical by other authors (see sec-
tion 6.2.1). I discuss these applicative arguments here because they conform to
my general definition of applicative arguments (4), repeated below, and all of
my informants (most of them from Lower Saxony) accept sentences such as
(298).
299. Applicative Argument
An NP Y of a simple, non-negated declarative sentence that is not
governed by a preposition is an applicative argument iff the sentence
without Y does not entail that there is at least one individual that is in-
volved in the asserted event and could be referred to by Y.
Part-whole applicative arguments are not required for the sentence to be
grammatical, are not governed by a preposition and the denoted entity is not en-
tailed as being involved in the event. (300) shows that sentence (298) is still
grammatical without the applicative argument; the applicative argument is not
selected by the verb.
300. Jan zerbeulte die Tür.
Jan.Nom dented the.Acc door
'Jan dented the door.'
This sentence does not entail the involvement of the car (beyond the specificity
supplied by the definite determiner which may indicate that the door is that of


80 External possessor constructions of English like John kicked Jim in the stomach are not
investigated here.


  149

the car). Thus, the part-whole applicative argument is an applicative argument


according to my definition (299).
I am excluding those "part-whole applicatives" where the part and the whole
form one constituent (301).
301. a. Dem Vater sein Hut hängt an der Wand.
the.Dat father his.Nom hat hangs on the wall
'The father's hat is hanging on the wall.'
b. Dem Auto seine Tür hat Jan zerbeult.
the.Dat car its.Acc door has Jan dented
'Jan has dented the car's door.'
These examples are not a part-whole applicative as discussed here because the
dative-marked constituent is included in a larger (nominal) constituent which
can be seen from the fact that it occurs in the sentence-initial position of V2-
sentences in (301). As discussed in Chapter 1, these are not applicative argu-
ments. This construction also differs from the part-whole applicative construc-
tion because the (agreeing) possessive pronoun must follow the dative-marked
elements (Cirtila 2006). The part-whole applicative construction as discussed
here neither requires the possessive pronoun nor is it part of a larger constituent
(cf. (298)).
The part-whole applicative should be classified as a high applicative accor-
ding to Pylkkänen (2002) because no transfer of possession is involved. Also,
the part-whole applicative can appear with unergative verbs (302) which is a
distinct characteristic of high applicatives (section 2.4).
302. Dem Baum haben die Blätter geleuchtet.
the.Dat tree have the.Nom leaves glowed
'The leaves of the tree were glowing.'
Thus, according to Pylkkänen's analysis, the part-whole applicative arguments
should be treated just like benefactives and other high applicatives. Yet, this is
not a good analysis for part-whole applicatives because it fails to explain the
part-whole relation between the two entities, as I discuss in section 6.4.2.
This chapter is organized as follows: in section 6.2, I give a detailed descrip-
tion of part-whole applicative arguments which includes differentiating them
from other German applicatives and discussing the restrictions they adhere to. In
section 6.3, I present the tests discussed in Chapter 2 which show that part-
whole applicative arguments do not contribute any not-at-issue meaning. In sec-
tion 6.4, I discuss previous analyses as well as present my own analysis of part-
whole applicative arguments. Some consequences of my analysis are addressed
in section 6.5. I conclude this chapter in section 6.6.

 
150  

6.2 Description
In this section, I differentiate the part-whole applicative argument from other
German applicative arguments (section 6.2.1). In this discussion as well as the
following subsection (6.2.2), the main characteristics of the part-whole applica-
tive argument are introduced. The restrictions on the part-whole applicative ar-
gument construction are addressed in section 6.2.3.

6.2.1 Differentiating Part-Whole Applicative Arguments


The part-whole applicative argument construction requires an applicative argu-
ment referring to an entity. A material part of this entity is involved in the event
and is encoded in another NP in the same sentence. Many authors have marked
these sentences as ungrammatical or only acceptable if the entity denoted by the
applicative argument is personified or ascribed animacy (see among others von
Polenz 1969, Wegener 1983, Neumann 1995). Abraham (1973) struggles with
deciding whether part-whole applicative arguments are grammatical or not. He
accepts them in some cases but not others. However, the authors that agree that
the part-whole construction is ungrammatical often discuss the "pertinence da-
tive" (303a) in depth (see Chapter 3). Superficially, the pertinence dative con-
struction differs from the part-whole applicative argument construction only in
the animacy of the referent. For instance, Neumann (1995) claims that (303a) is
grammatical (it is an example of the pertinence dative) and (303b) is not (unless
(303b) includes a personification of the car which allows the sentence to be
treated analogous to (303a)).81
303. a. Der Stein fällt dem Mann auf den Kopf.
the.Nom brick falls the.Dat man on the head
'The brick falls on the man's head.'
(Neumann 1995, example 1)
b. *Der Stein fällt dem Auto aufs Dach.
the.Nom brick falls the.Dat car onto.the roof
'The brick falls on the roof of the car.'
(Neumann 1995, example 14)
However, the native speakers of German that I have consulted (including
myself) disagree with the judgment that (303b) is ungrammatical; it is a genuine
instance of a part-whole applicative argument. Following Hole's claim that these
applicatives are only acceptable in some dialects, it seems that Neumann (1995)


81 There are also some authors that address a "possessor" dative with animate reference
without addressing the affectedness that is expected for a pertinence dative, e.g. Müller
(1995).


  151

(as well as other authors) and their consulted native speakers speak dialects of
German that do not allow the part-whole applicative argument construction.
The similarities and differences of the pertinence dative and the part-whole
applicative argument are discussed in Chapter 3. It was concluded in that chap-
ter that the pertinence dative is an affected experiencer construction which con-
tributes some not-at-issue meaning. This is not the case for the part-whole ap-
plicative argument (see Chapter 3 and section 6.3). The part-whole applicative
argument has no animacy requirement. This distinguishes the part-whole appli-
cative argument from the affected experiencer/ pertinence dative.82 Furthermore,
the contribution of at-issue meaning also distinguishes the part-whole applica-
tive from the subject co-referential applicative and the ethical dative as both
contribute not-at-issue meaning (Chapter 4).
The part-whole applicative argument construction can be distinguished from
the benefactive (Chapter 5) by the availability of the für-paraphrase for the bene-
factive. The part-whole applicative argument cannot be paraphrased in that way
(304).
304. Jan zerbeulte die Tür für das Auto.
Jan.Nom dented the.Acc door for the car
'Jan dented the door for the car.'
*'Jan dented the door of the car.'
In this sample sentence, there is no requirement that the door is a material part of
the car. As shown below, this is an essential requirement for the part-whole con-
struction. No other preposition is available for a paraphrase either.
Finally, differentiating the part-whole applicative argument from the Dative
of Inaction (Chapter 7) is impossible without context, if the requirements for
both are met (305).
305. Der Frau ist der Arm abgefault.
the.Dat woman is the.Nom arm rotted.off
'The woman's arm rotted off.'
'The arm rotted off on the woman's watch.'
This sample sentence can be used to either describe what happened to a woman
(part-whole applicative argument) or what happened to someone else while it
was the woman's responsibility to prevent this (DI interpretation). The context
must be used to differentiate the two.83


82 To avoid ambiguity of the part-whole applicative argument and the pertinence dative, I
use sentences with inanimate referents for the part-whole applicatives.
83 This sentence can also be interpreted as an affected experiencer/ pertinence dative. In
that case, it includes not-at-issue meaning and can be differentiated accordingly.

 
152  

6.2.2 Characteristics
The part-whole construction requires that there is a material part-whole relation-
ship between two NPs. Sentences in which this relationship is not provided are
ungrammatical. For instance, (306) is ungrammatical because the two entities
Fahrrad 'bike' and Tür 'door' are not in a part-whole relation.
306. *Jan zerbeulte dem Fahrrad die Tür.
Jan.Nom dented the.Dat bike the door
'Jan dented the door of the bike.'
Also, a subset-superset relation cannot be expressed in this way.84
307. *Jan zerstörte den Zigaretten nur die
Jan.Nom destroyed the.Dat cigarettes only the.Acc
Marlboros.
Marlboros
'Jan destroyed of the cigarettes on the Marlboros.'
Similarly, abstract entities and kinship nouns cannot be used in this construction
because a material part-whole relationship is required.
308. a. *Jan zerstörte der Idee den Kern.
Jan destroyed the.Dat idea the core
'Jan destroyed the core of the idea.'
b. *Jan zerkratzte Sabrina den Vater.
Jan scratched Sabrina.Dat the father
'Jan scratched Sabrina's father.'
Due to this required part-whole relationship between two entities encoded in the
sentence, I prefer the term "part-whole applicatives" over "landmarks" (Hole
2008).
There are sentence structures which can express a very similar meaning to
the part-whole applicative argument construction (309). I discuss these here
briefly to show what the part-whole applicative argument accomplishes in con-
trast to those other structures.
309. a. Jan zerbeulte dem Auto die Tür.
Jan.Nom dented the.Dat car the.Acc door
'Jan dented the door of the car.'
b. Jan zerbeulte [die Tür des Autos].
Jan.Nom dented the.Acc door the.Gen car.Gen
'Jan dented the door of the car'


84 In this, the German part-whole construction differs, for instance, from the Korean
"zooming in" construction (Sim 2005).


  153

c. Jan zerbeulte die Autotür.


Jan.Nom dented the.Acc car.door
'Jan dented the door of the car.'
d. Jan zerbeulte [die Tür von dem Auto].
Jan.Nom dented the.Acc door of the car
'Jan dented the door of the car.'
The difference between the applicative construction (309a) and the genitive con-
struction (309b) is that the applicative construction encodes that the dented door
was a material part of the car. The genitive construction is somewhat formal and
stilted, and, if anything, emphasizes the fact that it was the door (of the car) that
was dented (and not, for instance, the hood). The difference between the two can
be characterized as a change in focus or topic (apart from the stiltedness of the
genitive). The applicative argument of (309a) is focused or topicalized (cf.
Lamiroy and Delbecque 1998). The compound version (309c) encodes the type
of door that was dented by Jan, namely that it was a door of a car (rather than
the door of, for instance, a bus). However, this door does not have to be a mate-
rial part of a car (unlike in (309a)). It could be that out of a collection of differ-
ent types of doors, Jan dented one of the type "car door". Thus, the compound
does not (necessarily) convey a part-whole relationship. The closest paraphrase
of (309a) is (309d) which uses an of-phrase (which is included in a larger NP).
This sentence still indicates the relation of the door to the car. Compared to
(309a), it is lacking the highlighting of the car. In other words, what is encoded
in the part-whole applicative argument construction that is not encoded in the
paraphrases is the material part-whole relationship between the door and the car
as well as a vague highlighting of the car. For (309a) to be acceptable, the door
must be unambiguously associated with a certain car.
The part-whole relation is not cancelable (unlike the pertinence dative
(118)).85
310. #Jan zerbeulte dem Auto die Tür. Es war
Jan.Nom dented the.Dat car the.Acc door it was
die Tür des Buses.
the.Nom door the.Gen bus.Gen
'Jan dented the door of the car. It was the door of the bus.'
The sentences in (310) cause a contradiction.
In summary, the part-whole applicative argument construction conveys the
meaning that a material part of the referent of the applicative argument is in-
volved in the event.

85 The example is acceptable in a context where dem Auto can be construed as a type of
applicative different from the part-whole applicative.

 
154  

6.2.3 Restrictions on Part-Whole Applicative Arguments


The occurrence of part-whole applicative argument constructions is restricted by
several factors. I discuss here only sentence-internal restrictions. Hole (2008)
identifies several discourse requirements of the part-whole applicative construc-
tion. I focus on the restrictions relevant for the syntactic and semantic analysis.
There are a few sentence-internal restrictions that a part-whole applicative
argument construction must adhere to in order to be well-formed. First, the part-
whole relationship between the applicative argument and a second NP is neces-
sary (306). It follows from the required material part-whole relationship that the
sentence must include at least one NP in addition to the applicative argument. It
is not acceptable for this NP to be provided by the discourse. In example (311),
the material in parentheses is unpronounced.
311. a. Was hat Jan mit einer Tür gemacht?
what has Jan.Nom with a.Dat door done
'What did Jan do to a door?'
b. *(Er hat sie) dem Auto zerbeult.
he.Nom has her.Acc the.Dat car dented
'He dented the one of the car.'
c. (Er hat) sie dem Auto zerbeult.
he.Nom has her.Acc the.Dat car dented
'He dented the one of the car.'
The NP that is required in addition to the applicative argument expresses the
part to which the referent of the applicative argument is the whole. The part NP
can have various grammatical functions in the sentence: a direct object of transi-
tive verbs, a subject of an intransitive verb, or be included in a PP (312).
312. a. Jan zerbeulte dem Auto die Tür.
Jan.Nom dented the.Dat car the.Acc door.
'Jan dented the car's door.'
(transitive)
b. Jan legte dem Auto etwas in den
Jan.Nom laid the.Dat car something.Acc in the
Kofferraum.
trunk
'Jan put something into the trunk of the car.'
(transitive +PP)
c. Dem Baum haben die Blätter geleuchtet.
the.Dat tree have the.Nom leaves glowed
'The leaves of the tree were glowing.'
(unergative, part as subject)


  155

d. Jan pinkelte dem Klo auf den Rand.


Jan peed the.Dat toilet onto the rim
'Jan peed the toilet's rim.'
(unergative, part in PP)
e. Der Stein fiel dem Zug auf die Lok.
the.Nom stone fell the.Dat train onto the locomotive
'The stone fell onto the locomotive of the train.'
(unaccusative, part in PP)
f. Das Dach fiel dem Haus ein.
the.Nom roof fell the.Dat house Prt
'The roof of the house collapsed.'
(unaccusative, part as subject)
This also shows that the part-whole applicative argument can appear with transi-
tive, unaccusative and unergative verbs (cf. Lamiroy and Delbecque 1998).
However, even though the applicative argument can be interpreted as the whole
to a subject of an unergative verb (312c), it cannot be the whole to a subject of a
transitive verb (313).
313. *Die Schaufel zerbeulte dem Bagger die
the.Nom bucket dented the.Dat backhoe the.Acc
historischen Artefakte.
historical artifacts
'The bucket of the backhoe dented the historical artifacts.'
The part NP may also not be included in an adjunct PP (cf. Lee-Schoenfeld
2006).
314. *Jan zerbeulte dem Auto die Klappe
Jan.Nom dented the.Dat car the.Acc cover
hinter der Tür.
behind the door
'Jan dented the cover behind the door of the car.'
Furthermore, in the presence of an argument (goal) PP, the applicative can-
not be in a part-whole relation with the direct object (cf. Müller 1995). This is
the configuration in (315a). It contrasts with (315b) in which the presence of a
locative (adjunct) PP does not prevent the part-whole relation from being estab-
lished between the direct object and the applicative argument.
315. a. *Jan legte dem Fahrrad das Rad auf
Jan laid the.Dat bike the.Acc tire onto
den Tisch.
the.Acc table
'Jan put the car's tire onto the table.'

 
156  

b. Jan zerkratzte dem Auto das Rad in der


Jan scratched the.Dat car the.Acc tire in the
Garage.
garage
'Jan scratched the tire of the car in the garage.'
Some part-whole applicative argument constructions are accepted by only some
speakers when a directional PP is used (see Neumann (1995) on speaker varia-
tion for possessor datives).
316. %Jan sprang dem Auto durch das Fenster.
Jan.Nom jumped the.Dat car through the window
'Jan jumped through the window of the car.'
This might indicate that adding a goal PP does not force a VP (for all speakers)
to be resultative which is another sentence-internal requirement. Non-resultative
verbs do not license the part-whole applicative argument (317).
317. *Dem Vogel sang die Stimme.
the.Dat bird sang the.Nom voice
'The voice of the bird sang.'
Consequently, perception and emotion verbs do not license the part-whole ap-
plicative argument construction (318).
318. a. *Jan sah dem Baum die Rinde.
Jan saw the.Dat tree the bark
'Jan saw the tree's bark.'
b. *Jan fühlte dem Baum die Rinde.
Jan felt the.Dat tree the bark
'Jan felt the bark of the tree.'
These sample sentences cannot be improved by context. For instance, if Jan
touched a tree with old and brittle bark, sentence (318b) would still be unac-
ceptable. It is a syntactic restriction on the type of verbs that can occur in the
part-whole applicative construction. Resultativity of the VP is an important
characteristic in licensing the part-whole applicative argument.
Finally, there is one more remarkable sentence-internal restriction on the
part-whole applicative argument construction, namely the part NP cannot
scramble over the applicative argument within the midfield. 86

86 This only holds for full NPs. Pronouns must occur in the left midfield (cf. Chapter 2):
i. Jan zerbeulte sie dem Auto.
Jan dented she.Acc the.Dat car
'Jan dented it of the car.'
ii. *Jan zerbeulte dem Auto sie.
Jan dented the.Dat car she.Acc
'Jan dented the door of the car.'


  157

319. a. Jan zerbeulte dem Auto gestern die Tür


Jan dented the.Dat car yesterday the door
in der Garage.
in the garage
'Jan dented the car's door in the garage yesterday.'
b. *Jan zerbeulte die Tür dem Auto gestern
Jan dented the.Acc door the.Dat car yesterday
in der Garage.
in the garage
'Jan dented the car's door in the garage yesterday.'
c. Dem Auto zerbeulte Jan die Tür gestern
the.Dat car dented Jan the.Acc door yesterday
in der Garage.
in the garage
'Jan dented the car's door in the garage yesterday.'
This is another distinct characteristic of the part-whole applicative argument
construction. Affected experiencers, benefactives, and Datives of Inaction can
scramble in the midfield (320).87
320. a. Jan zerbeulte Bens Vase seinem Freund.
Jan dented Ben.Gen vase his.Dat friend
'Jan dented Ben's vase on his friend.'
(Affected Experiencer)
b. Jan buk den Kuchen seiner Mutter.
Jan baked the.Acc cake his.Dat mother
'Jan baked the cake for his mother.'
(True Benefactive)
c. Gestern verwelkten die Blumen dem
yesterday withered the.Nom flowers the.Dat
Gärtner.
gardener
'The flowers withered on the gardener's watch yesterday.'
In contrast to what is claimed in Müller (1995), it is not necessary for the
part-whole applicative argument to c-command the part NP. The part NP may c-
command the applicative argument on the surface if it is contrastively focused or
questioned (321).


87 Ethical datives and subject co-referential applicative arguments are less flexible in their
position because they must be pronouns which have their unmarked position in the left
midfield.

 
158  

321. a. [Die Tür]F zerbeulte Jan dem Auto in


the.Acc door dented Jan.Nom the.Dat car in
der Garage.
the garage
'Jan dented the door of the car in the garage.'
b. Was zerbeulte Jan dem Auto in der
what dented Jan.Nom the.Dat car in the
Garage?
garage
'What part of the car did Jan dent in the garage?'
In summary, the part-whole applicative must meet the following sentence-
internal requirements: first, the part-whole relation must hold between two NPs
of the sentence but it cannot be established between the applicative argument
and the subject of a transitive verb or an NP inside of an adjunct PP. Second, the
part NP may either be an argument of the verb or be included in an (argument)
PP. Third, the part NP cannot scramble over the applicative argument within the
midfield (yet, it may precede the applicative argument in certain other cases).
Fourth, the VP must be resultative. An analysis of the part-whole applicative
arguments should take all of these features into account.
6.3 At-Issue Meaning
As far as the meaning contribution is concerned, part-whole applicatives do not
contribute to the not-at-issue tier of meaning. However, the tests from Chapter 2
do not reveal this as straightforwardly as one would expect. This is partly due to
the fact that the part-whole applicative argument behaves similarly to the geni-
tive. Furthermore, wh-questions are problematic because of an accidental gap in
the German lexicon (see below).
Part-whole applicative arguments do not project any meaning past yes/no-
questions (322).
322. Zerbeulte Jan dem Auto die Tür?
dented Jan the.Dat car the door
'Did Jan dent the door of the car?'
The answer to the yes/no-question can be "no" either if Jan did dent a door but it
was not the car's door or if Jan did not dent a door at all. There is no further ex-
planation needed. By answering "no", no truth of anything is accepted. This in-
dicates there is no not-at-issue meaning carried by the part-whole applicative
argument.
Further support for part-whole applicative arguments not contributing any
not-at-issue meaning can be found in conditionals. The part-whole applicative
argument makes a difference in a conditional sentence.


  159

323. Wenn Jan einem Auto eine Tür zerbeult,


if Jan a.Dat car a.Acc door dents
dann bekommt er 100Euro von mir.
then gets he 100Euros from me
'If Jan dents a door of a car, then he'll get 100Euros from me.'
In example (323), which includes a part-whole applicative argument, if Jan
dents the door of anything but a car, I will not have to give him the money.
However, this is not true in the conditional sentence without the applicative ar-
gument (324).
324. Wenn Jan eine Tür zerbeult, dann bekommt
if Jan a.Acc door dents then gets
er 100Euro von mir.
he 100Euros from me
'If Jan dents a door, then he'll get 100Euros from me.'
In that case, I would have to pay if Jan dents any door at all. The part-whole ap-
plicative argument therefore contributes its meaning in conditional clauses. This
is support for the applicative argument contributing no not-at-issue meaning but
only at-issue meaning.
The third test is the effect of negation in sentences with a part-whole appli-
cative argument.
325. Jan zerbeulte dem Auto die Tür nicht.
Jan.Nom dented the.Dat car the door Neg
'Jan didn't dent the door of the car.'
Negation in (325) negates the whole event, i.e. denting of the door of the car did
not happen. This sentence cannot mean that Jan dented the door but the door
was not part of the car. However, this reading of the door not being part of the
car should be available if the part-whole applicative argument contributes only
at-issue meaning (because it is expected that negation can target each semantic
conjunct). Interestingly, the same pattern of negation can also be found with the
genitive construction (326).
326. a. Jan zerbeulte die Tür des Autos.
Jan dented the.Acc door the.Gen car.Gen
'Jan dented the door of the car.'
b. Jan zerbeulte die Tür des Autos nicht.
Jan dented the.Acc door the.Gen car.Gen Neg
'Jan didn't dent the door of the car.'
In (326b), the negation also cannot negate that the door belongs to the car, just
like in the applicative argument construction (325). Another similarity between
the part-whole applicative argument and the genitive is that in both cases chang-

 
160  

ing the word order and contrastively focusing Auto makes the reading available
(327).
327. a. Jan zerbeulte nicht [dem Auto]F die
Jan dented Neg the.Dat car the.Acc
Tür (sondern dem Bus).
door but the.Dat bus
'Jan dented the door not of the car (but of the bus).'
b. Jan zerbeulte nicht die Tür [des Autos]F
Jan dented Neg the.Acc door the.Gen car.Gen
(sondern des Buses).
but the.Gen bus.Gen
'Jan dented the door not of the car but of the bus.'
It seems that the special semantic relation between the two NPs in sentences
(325) and (326) prevents negation from targeting the genitive and applicative
argument, respectively. I contend that the negation test in (325) does not provide
support for part-whole applicative argument contributing not-at-issue meaning.
Rather, it may be the material part-whole requirement that stops negation from
being able to targeting this (see below).
Wh-questioning of inanimate part-whole applicative arguments is possible.
328. Was für einem/ welchem Auto hat er
what for a.Dat which.Dat car has he.Nom
die Tür zerbeult?
the door dented
'Which (kind of) car did he dent the door on?
This shows that the part-whole applicative argument is syntactically accessi-
ble.88
The last piece of evidence for the part-whole applicative argument having
only at-issue meaning comes from binding. The applicative argument can bind
into other constituents (329).
329. Jan zerbeulte jedem Auto1 seine1 Tür.
Jan.Nom dented every.Dat car its.Acc door
'Jan dented the door of every car.'
It was shown in Chapter 2 that this pattern of binding is not possible for not-at-
issue elements.

88 Non-D-linked questions with wem 'who.Dat' are only marginally possible because this
wh-word refers more readily to a (sentient) individual. The salient reading of the wh-
question i. is that of asking about an affected experiencer applicative argument.
i. Wem zerbeulte Jan die Tür?
who.Dat dented Jan the door
'Whose door did Jan dent?' 'On who did Jan dent the door?'


  161

In short, part-whole applicative arguments do not contribute any not-at-issue


meaning but all their meaning is contributed to the at-issue tier of meaning.
6.4 Analysis
In this section, I present my analysis of part-whole applicative arguments. First,
I discuss four previous approaches, namely a possessor raising analysis from
Lee-Schoenfeld (2006), the analysis by Pylkkänen (2002) for high applicative
arguments, the analysis by Tomioka and Sim (2007) who discuss a similar Ko-
rean construction, and the analysis of "landmark" applicatives from Hole (2008).
I show that none of these analyses can account for all aspects of the part-whole
applicative argument construction. I present my analysis in subsection 6.4.5.

6.4.1 Possessor Raising


A possessor raising analysis along the lines of Landau (1999) or Lee-Schoenfeld
(2006) does not provide an adequate analysis of part-whole applicatives. I will
use the analysis by Lee-Schoenfeld (2006) as an example again. As explained in
Chapter 3, she assumes that a defective D feature enables possessor raising.
The possessor raising analysis cannot explain when the raised structure will
surface. The defective D head should be available in cases other than the part-
whole applicative structures (also see Chapter 3), for instance with abstract enti-
ties and kinship relations (308). It is not clear why these relations do not allow a
defective D head and thereby possessor raising. This might be avoided by defin-
ing the position inside the DP as a position in which a material part-whole rela-
tion is encoded (rather than a possessive relation). Yet, this does not explain
why defective D can only occur in certain DPs. For instance, the possessor rais-
ing approach must allow possessor raising out of DPs that are included in goal
(argument) PPs, which allow the part-whole relation to be established (315). In
the presence of such a PP, raising may not take place out of the direct object. It
is not clear how these constraints can be handled with a defective D. Another
problem is that the difference between unergative and transitive subjects cannot
be explained (309c, 310) by using only a defective D in the analysis.
Thus, possessor raising cannot be used to explain part-whole applicative ar-
guments because it is unclear how to restrict raising correctly.

6.4.2 Pylkkänen (2002)


According to the categorization proposed by Pylkkänen (2002), the part-whole
applicative argument should be a high applicative because no transfer of posses-
sion is involced.89 The part-whole applicative construction is not separately dis-

89 Alternatively, one could claim that it should be a low applicative with a 'continuation of

 
162  

cussed by Pylkkänen. However, like all high applicatives in Pylkkänen's ap-


proach, the applicative argument should be assigned a thematic role by an ap-
plicative head that attaches between VP and Voice (Chapter 2). Let's assume
that the thematic role assigned by the applicative head is Possessor which indi-
cates that another entity is also present in the sentence. This yields the following
derivation in Pylkkänen's system:
330. a. Jan zerbeulte dem Auto die Tür.
Jan dented the.Dat car the.Acc door
'Jan dented the door of the car.'
b. VoiceP
3
Jan Voice'
3
Voice ApplP
3
dem Auto Appl'
3
Appl VP
6
zerbeulte die Tür
c. [[VoiceP]] = λe. DENT(e) & Thm(e)(the door) & Poss(e)(the car) &
Agt(e)(Jan)
According to this derivation, (330a) denotes a denting event of which Jan is the
agent, the door is the theme and the car is a possessor. While this seems to be
the correct denotation at first glance, it actually is not completely correct. The
first problem is that it does not reflect that the door is a material part of the car.
It only states that the car possesses something but not which entity belongs to
the car. While one might be able to infer that it is the door, this inference should
be available in other sentences as well. However, as discussed above, there are
several restrictions on the grammatical functions the part NP can fulfill, most
notably it cannot be the subject of a transitive verb (310). This cannot be ac-
counted for in this approach because the inference of which entity belongs to the
possessor should be able to be established between the possessor and any other
NP (because they are all event participants).
Another problem with this derivation is that the car is now a participant of
the denting event (Poss(e)(x)). However, it is not always true that the whole is
undergoing the same event that the part is undergoing: for instance, if (331a) is

possession' meaning. See Larson (2010) for problems with Pylkkänen's denotation for
low applicatives.


  163

true, (331b) is not necessarily true. This applies especially to change of state
verbs (Neumann 1995): when the state of the part changes, the state of the whole
does not necessarily change (or not in the same way).
331. a. Jan knickt dem Baum die Blätter ab.
Jan snaps the.Dat tree the leaves off
'Jan snaps the leaves off the tree.'
b. Jan knickt den Baum ab.
Jan snaps the.Acc tree off
'Jan snaps off the tree.'
This example shows that the referent of the applicative argument is not undergo-
ing the same event as the part. The denotation in (330c) does not reflect this. All
entities are participants of the denting event. This could be avoided by defining
Poss as not participating in the event but that it is the possessor of an entity
which participates in that event. However, by doing that, the essence of Pyl-
kkänen's approach is lost, namely not all high applicatives would receive the
same analysis. In other words, Pylkkänen's original analysis of high applicatives
is not well-suited for the part-whole applicatives because it cannot ensure that
the material part-whole relation holds between the right entities. Consequently,
some of Pylkkänen's assumptions have to be given up. By doing that, her unified
approach to all high applicatives is lost.
Overall, the additional thematic role provided by Pylkkänen's analysis is not
suited to explain the part-whole applicative argument construction because it
cannot ensure that the material part-whole relation holds between the correct
entities. In order to include that in the analysis, part-whole applicatives must be
treated differently from other high applicatives.

6.4.3 Hole (2008)


Hole (2008) proposes a different approach to part-whole applicatives. Before
discussing it in detail, it is necessary to discuss the terminology used by Hole.
Hole (2008) calls the part-whole applicative arguments "landmark datives". He
chose that name because he argues that they denote entities which are used to
locate other entities. For instance in (298), repeated below, the car is the land-
mark which is used to locate the door.
332. Jan zerbeulte dem Auto die Tür.
Jan dented the.Dat car the.Acc door
'Jan dented the door of the car.'
In line with this spatial approach to these applicative arguments, Hole also uses
the terms "ground" and "figure" to talk about the applicative argument (ground)
and the part NP (figure). In (332), the ground object is the car and the figure ob-
ject is the door. These two objects must be in a spatial relation with each other

 
164  

for the applicative construction to be acceptable; to be precise, the applicative


argument denotes a ground object for the located object. The referent of the
ground must be present for the event described in the sentence to take place be-
cause it provides the ground needed to locate the event.
Hole's semantic and syntactic analysis is as follows. Syntactically, all Ger-
man applicative arguments (as defined by him) attach between CAUSE90 and a
constituent that denotes a result and/or state (333). XP is a constituent that in-
cludes a result or state and is used to capture the verbal restrictions of applica-
tive arguments and that they cannot select all VPs.
333. 3
AGENT 3
CAUSE 3
DPdative 3
LANDMARK XP
6
constituent with result/state
(adapted from Hole 2008, example 11.15)
The "LANDMARK" head is given the denotation in (334).
334. [[LANDMARK]]α = λx : [¬y. xey & yC]. λs. : [¬z. s'. s' holds
for the neighborhood region of z(s) & s'ss & zex]. s''. s holds for
the neighborhood region of x(s'') & sss'' & s'' holds at location a(i),
where C is a contextually salient subset of D
(Hole 2008, example 10.64, my translation)
As Hole explains (in his example 10.65), this denotation addresses and includes
all of the characteristics of landmark datives that he identified: first, the dis-
course requirement that the dative denotes a ground with a spatial relation to an-
other entity is encoded in [λx. ... λs. … s''. s holds for the neighborhood region
of x(s'') & sss'']. Second, the fact that the ground is not bigger or smaller than
the actual entity involved in the event is denoted by "where C is a contextually
salient subset of D" and the condition [λs.: [¬z. s'. s' holds for the neighbor-
hood region of z(s) & s'ss & zex]]. Third, the assertion of the ground is re-
flected in the fact that its lambda term is in the truth condition (preceding the
colon). Finally, the necessity for explicit discourse linking is reflected in [λs. ...
s''. ... sss'' & s'' holds at location a(i)]. This location can either be contextually
given or be supplied in the sentence.
Hole (201X) discusses three types of support for his analysis which requires
the landmark dative to bind a constituent included in XP (similar to the benefac-

90 CAUSE is a change of state operator. It introduces a causing event that brings about the
result state. CAUSE also introduces the agent projection.


  165

tive (section 5.4.4) and the pertinence dative (section 3.6)). First, he addresses
sloppy identity readings under ellipsis (335). He argues that the applicative must
bind because a strict reading is not possible.
335. Dem Patienteni platzte ein Stück Gips
the.Dat patient cracked a piece cast
(vomi/ von seinemi) Arm ab, und dem
of.the of his arm off and the.Dat
Arzt auch.
doctor too
'It happened to the patienti that part of the cast on hisi arm came off,
and it happened to the doctorj that part of hisj came off, too.'
*'It happened to the patienti that part of the cast on hisi arm came off,
and it happened to the doctorj that part of hisi came off, too.'
(Hole 201X, example 11)
As shown in Chapter 3, this is not true. The strict reading is available for this
example. For the part-whole applicative, this cannot be shown because of the
required material part-whole relation. However, in the following example, the
salient reading is that there are two doors, one belonging to a bus and one be-
longing to a Mercedes.
336. Jan zerbeulte dem Bus die Tür und dem
Jan.Nom dented the.Dat bus the door and the.Dat
Mercedes auch.
Mercedes too
'Jan dented the door of the bus and the one of the Mercedes, too.'
On Hole's approach, dem Bus should bind die Tür and dem Mercedes should
bind an unpronounced die Tür included in the elided part.91 On my approach
(which does not require binding), it should be possible for the Mercedes to be a
whole to the door which is part of the bus. In other words, if there is a Mercedes
bus and Jan dented that bus's door, (336) should be able to describe that situa-
tion, and it does. However, in that case it cannot be distinguished from the bind-
ing approach because the door of the bus and the door of the Mercedes are iden-
tical, i.e. dem Mercedes might be binding into the elided part.
The second type of support for Hole's binding analysis comes from the ac-
commodation of possessors when there is no pronounced pronoun. Hole (201X)
does not provide a landmark example (337b) but the reasoning should be the
same as for possessors (which he discusses) (337a).


91 Hole assumes indices on all elements with anaphoric use including (bridging) definites
to be available for binding.

 
166  

337. a. Paul trat Maria gegen einen Stein.


Paul.Nom kicked Maria.Dat against a.Acc stone
'Paul kicked against a stone of Maria's and it wasn't excluded that
Maria noticed this.'
(Hole 201X, example 13)
b. Paul trat einem Baum gegen einen Ast.
Paul kicked a.Dat tree against a.Acc branch
'Paul kicked a branch of a tree.'
Hole argues that despite the absence of an overt pronoun in (337a), the sentence
can be interpreted as the stone being one of Maria's stones. Although the exact
relation of Maria and the stone is not specified, Maria can bind an implicit pro-
noun in that NP. This assumption is problematic because in (337b) the relation
of the tree and the branch is specified; it has to be a material part-whole relation.
Yet, binding of an implicit pronoun in (337b) in combination with the spatial
denotation of the landmark should allow for the branch being one that broke off
the tree and is now lying next to it. However, this example is not acceptable in
that scenario. By encoding the part-whole relation directly, I can avoid this prob-
lem as well as the necessity of assuming that unpronounced pronouns are always
available if there is nothing else available to be bound by the applicative argu-
ment.
A similar problem for the binding analysis arises when the dative should
bind into an argument but it does not. This is exemplified in (338).
338. Jan zerbeulte dem Auto1 seine2 Tür.
Jan.Nom dented the.Dat car his.Acc door
'Jan dented his door of the car.'
In this sentence, the possessive pronoun can refer to, for instance, a person who
donated one door which got mounted on the car. Thus, there is a pronoun which
could be bound by the dative constituent but it is not. Hole must assume another
unpronounced pronoun that the dative argument binds. This leads to unrestricted
positing of unpronounced pronouns.
The third argument for binding that Hole (201X) provides is locality, name-
ly that binding must take place in the local tense domain. While he does not dis-
cuss landmark datives in detail for this feature, my analysis of part-whole ap-
plicatives also predicts this because the applicative head selects a VP (as shown
below). The part-whole relation is established semantically at that point between
the complement of V (type e) and the applicative argument. Consequently, it is
impossible for the part and the whole to be generated in two different tense do-
mains.
Thus, Hole's and my approach do not differ on the sloppy identity and the
locality of part-whole applicatives. However, Hole's assumption of the availabil-


  167

ity of unpronounced pronouns everywhere is undesirable and not needed in my


analysis.
Hole's approach can handle the different occurrences of the part NP (e.g. as
an object or as included in a PP if the PP is included in the XP denoting the
state/result). However, it cannot explain why subjects of unergatives can be the
part NP (309c).
Hole argues that the scrambling restrictions (319) follow from his approach
because the applicative binds into the direct object/PP. Linear precedence in the
midfield reflects the scope and therefore the part NP must not scramble within
the midfield. This, however, is too strong because - as example i. in a footnote in
section 6.2.3 has shown - it is only full NPs (denoting the part) that do not
scramble over the applicative in the midfield.
In short, Hole's approach provides a better semantic analysis than Pyl-
kkänen's approach. However, it is too strong in some cases (unergative, scram-
bling) and requires undesirable assumptions about unpronounced elements being
available for binding.

6.4.4 Tomioka and Sim (2007)


Tomioka and Sim (2007) provide an analysis for a Korean construction similar
to the German part-whole applicative. This construction is an inalienable pos-
sessor construction in which both the inalienable part and its possessor are
marked with accusative case (339).
339. Jinhwa-ka Youngmee-lul meri-lul ppop-ass-ta.
Jinhwa-Nom Youngmee-Acc hair-Acc pull.out-Past-Dec
'Jinhwa pulled out Youngmee's hair.'
(Tomioka and Sim 2007, example 34)
Tomioka and Sim argue that both accusative-marked constituents are objects but
not of the same verb. The NP referring to the part is the object of the lexical
verb. This VP is selected by a (phonologically null) verb affect which assigns a
thematic role Theme to the possessor (=whole) NP.92 The derivation of (339)
proceeds as follows.93


92 This verb is not to be confused with the Affect head of affected experiencers (Chapter 3)
but rather means "impact". It does not have to be a psychological impact.
93 Tomioka and Sim (2007) use a modified event composition compared to Brisson (1998)
to combine VP2 and affect. It is described in their example 32.

 
168  

340. a. VoiceP<s,t>
3
NP<e> Voice'<e,<s,t>>
3
VP1<s,t> Voice<e,<s,t>>
3
NP1<e> V1'
3
VP2<s,t> affect<e,<s,t>>
3
NP2<e> V2<e,<s,t>>
(Tomioka and Sim 2007, example 31)
b. [[VoiceP]] = λe. [AFFECT(e) & Agt(e)(y) & Thm(e)(Youngmee) &
e'. [e'  e & PULL OUT(e) & Thm(e)(the hair)]]
The part-whole relationship holds between the two events (affect and lexical
verb) instead of two entities, as shown by e'  e. First, this ensures that the ver-
bal event is a subevent of the affecting event. Second, this relation between the
events means that the verbal event does not have to hold for the possessor. Thus,
pulling out Youngmee's hair does not mean pulling out Youngmee (cf. (331)).
Third, the part-whole relationship between the events means that the possessive
relation is not part of the semantics but rather "it is the most natural to interpret"
(Tomioka and Sim 2007, p. 16) the relation as a possessive relation because the
events are connected. Finally, this material part-whole relation between two
events means that the analysis is mono-eventive because one is a part of the oth-
er (despite the two event variables).
Furthermore, Tomioka and Sim (2007) argue that the eventualities of the
VP1 form a set of minimal eventualities so all and only those entities that are
relevant are included. Thus, in the affecting event the applicative argument is the
only entity and the verbal event includes only the Theme entity. In other words,
the sentence is only acceptable if the affecting event includes the possessor to an
entity included in the verbal event. This ensures that no entity can be affected
that is not relevant (341).
341. *Mary-ka John-ul Bill-euy tali-lul cha-ss-ta.
Mary-Nom John-Acc Bill-Gen leg-Acc kick-Past-Decl
'Mary affected John by kicking Bill's leg.'
(Tomioka and Sim 2007, example 38a)
Also, stipulating that this minimality condition is computed locally at VP1 en-
sures that the subject cannot be the possessor (as it has not yet been introduced)
(342, cf. 313).


  169

342. Chelswu-ka Sunhee-lul tali-lul ttayli-ass-ta.


Chelswu-Nom Sunhee-Acc leg-Acc hit-Past-Decl
intended but not available: 'Chelswu hit his own leg, and that affected
Sunhee.'
(Tomioka and Sim 2007, example 40)
However, there are also differences to the German part-whole construction.
First, Tomioka and Sim (2007) state that the degree of affectedness is a decisive
characteristic in Korean, unlike the resultative restriction that holds in German
(section 6.2.2). If the referent of the applicative argument is not affected, this
construction is infelicitous in Korean (343).
343. a. Chelswu-ka Sunhee-lul shikey-lul cap-ass-ta.
Chelswu-Nom Sunhee-Acc watch-Acc grab-Past-Decl
'Chelswu grabbed Sunhee's watch.'
b. ?Chelswu-ka Sunhee-lul shikey-lul kochi-ass-ta.
Chelswu-Nom Sunhee-Acc watch-Acc repair-Past-Decl
'Chelswu repaired Sunhee's watch.'
c. ??Chelswu-ka Sunhee-lul shikey-lul po-ass-ta.
Chelswu-Nom Sunhee-Acc watch-Acc see-Past-Decl
'Chelswu saw Sunhee's watch.'
(Tomioka and Sim 2007, example 10)
While German also does not allow a verb like see to participate in the part-
whole construction (318a), the German restriction is better characterized in
terms of resultativity than affectedness.
Furthermore, the German part-whole construction is acceptable with unerga-
tive verbs. This is not the case for Korean (which has a multiple nominative
construction not addressed by Tomioka and Sim (2007) for such cases). In the
analysis as it stands, unergative structures (312c), repeated below, cannot be
treated adequately.
344. Dem Baum haben die Blätter geleuchtet.
the.Dat tree have the.Nom leaves glowed
'The leaves of the tree were glowing.'
Unergative structures would receive the meaning that any event of the kind de-
noted by the verb will have an impact on the referent of the applicative argu-
ment. Here, any glowing event affects the tree, and the leaves are the agent of
the affecting event.
Another problem is that it is not clear why a material part-whole relation is
required. Tomioka and Sim's approach asks for a "natural" relation to the lexical
event. It is unclear why this is not available for abstract entities (308a), repeated
below, or direct objects when a resultative PP is present (315a), repeated below.

 
170  

345. a. *Jan zerstörte der Idee den Kern.


Jan destroyed the.Dat idea the.Acc core
'Jan destroyed the core of the idea.'
b. *Jan legte dem Fahrrad das Rad auf den Tisch.
Jan laid the.Dat bike the tire onto the table
'Jan put the car's tire onto the table.'
In short, while the German structure and the Korean structure share some si-
milarities, the analysis by Tomioka and Sim (2007) cannot be used to explain
the German part-whole applicative construction in full.
In the next section, I present my analysis of the part-whole applicative and
show how it avoids the problems of the previous analyses.

6.4.5 Analysis
The requirements for the part-whole applicative argument construction can be
summarized as follows: syntactically, the part NP and the applicative argument
do not form one constituent. Semantically, a material part-whole relation be-
tween the two NPs must be established. I argue that the applicative argument is
introduced by a phonologically null head, AppPW, which attaches outside of VP
(346).
346. VoiceP
3
Voice'
3
Voice AppPWP
3
Applicative NP AppPW'
3
AppPW VP
3
V NP
The semantic denotation of the AppPW head encodes the material part-whole rela-
tion between the two entities (347).94
347. [[AppPW]] = λf<e,vt>. λx. λy. λe. f(e)(s) & xy for the duration of e95
This denotation for AppPW works as follows: I assume that it attaches to VP and
prevents the constituents used to form this projection from combining semanti-


94 Thanks to Benjamin Bruening for the initial idea for this denotation!
95 "xy" means that x is a material part of y. "For the duration of e" reflects the fact that
the two entities do not have to be in the part-whole relation eternally but only
when/while the event is taking place.


  171

cally. This is due to functional composition; the semantic computation can look
ahead and prevent certain operations to ensure that the derivation converges.
ApplPW takes the two constituents forming the VP as its first and second seman-
tic arguments. The verbal event is passed up by ApplPW and the complement of
V is applied to it. At the same time, this entity becomes the material part to the
referent of the third semantic argument of AppPW, namely the applicative argu-
ment.
The derivation for (298), repeated below, is given in (348).96
348. a. Jan zerbeulte dem Auto die Tür.
Jan dented the.Dat car the.Acc door
'Jan dented the door of the car.'
b. VoiceP
3
Jan Voice'
3
Voice AppPWP
3
dem Auto AppPW'
3
AppPW VP
3
zerbeulte die Tür
c. [[VoiceP]] = λe. DENT(e) & Thm(e)(the door) & Agt(e)(Jan) & the
door  the car for the duration of e
This denotation states that there was a denting event of which Jan is the agent
and the door is the theme. Furthermore, the door is a material part of the car (for
the duration of the denting event).
This analysis respects both the syntactic and the semantic requirements of
the part-whole applicative construction. First, the applicative argument and the
direct object do not form a constituent. Nevertheless, the material part-whole
relation between the two is established. Since the part-whole relation is explicit-
ly encoded in the semantics, it is required to hold between the specified entities
and cannot hold between any other event participants (such as the agent, see
(313)).
Tomioka and Sim (2007) have observed that again-modification cannot tar-
get the smallest VP (in Korean) (349).


96 The full semantic derivation is given in Appendix 4, #1.

 
172  

349. Chelswu-ka Sunhee-lul tasi tali-lul cha-ass-ta.


Chelswu-Nom Sunhee-Acc again leg-Acc kick-Past-Decl
'Chelswu kicked Sunhee's leg again.'
(Tomioka and Sim 2007, example i)
This sentence cannot mean that Chelswu kicked Sunhee's leg for (at least) the
second time but affecting him for the first time. The same pattern holds for
German (350).
350. Jan zerbeulte dem Auto die Tür wieder.
Jan.Nom dented the.Dat car the.Acc door again
'Jan dented the door of the car again.'
This sentence cannot mean that Jan dented the door for (at least) the second time
and this time the door is a material part of the car. My analysis provided above
can explain this. As discussed in the previous chapters, again attaches to nodes
of type <v,t>. There are only two of those in the derivation (348), namely
VoiceP and AppPWP. Since the VP is not computed semantically, there is no
node of only the verbal event of type <v,t>. Both of the expected readings in
(350) are available: first, sentence (350) can mean that Jan (for the first time)
dented the door of the car which had previously been dented (by someone else).
This is the reading of again attaching to AppPW. Second, the sentence can mean
that Jan previously dented the door of the car and now he did it again. This is the
reading of again attaching to VoiceP.
Thus, the provided analysis is well-suited for transitive sentence. However,
the part-whole relation can also hold between the applicative argument and the
entity that is included in an argument PP (312b), repeated below. I assume that P
is semantically empty here because it is an argument PP so that the NP (inside
the PP) receives its thematic role from the verb. Consequently, the PP denotes
an individual of semantic type e. The derivation of (312b) proceeds as follows.97
351. a. Jan legte dem Auto etwas in
Jan laid the.Dat car something.Acc in
den Kofferraum.
the.Acc trunk
'Jan put something into the trunk of the car.'


97 The full semantic derivation is given in Appendix 4, #2.


  173

b. VoiceP
3
Jan Voice'
3
Voice AppPWP
3
dem Auto AppPW'
3
AppPW VP1
3
VP PP
3 3
legte<e,evt> etwas in den Kofferraum
c. [[VoiceP]] = λe. PUT(e) & Thm(e)(something) & Goal(e)(the trunk)
& Agt(e)(Jan) & the trunk  the car for the duration of e
Due to the fact that the PP is an argument of the verb, the VP is of type <e,vt>
after the verb and the direct object have combined. VP1 is formed syntactically
by VP and PP but is not computed semantically. AppPW takes these two elements
as its arguments. The entity provided by the PP then ends up as the material part
of the applicative argument. The direct object is not available (cf. (312a)) be-
cause it satisfies the first semantic argument of the verb. There is no delay in
combining the verb and the direct object because functional composition does
not require the delay. Consequently, they combine. The reason that only argu-
ment PPs are available for this is that they fill the position of the individual re-
quired by the verb. Adjunct PPs do not have this function and consequently do
not attach between AppPW and V (also see Ramchand (2008) for the different
attachment sites of PPs with varied functions).
This analysis is also supported by again-modification. There are two nodes
of type <v,t> in (351), namely AppPW and VoiceP. Thus, the expected readings
are that the event happened before with the same part-whole relation but a dif-
ferent agent or with the same part-whole relation and the same agent. These pre-
dictions are borne out.
352. Jan legte dem Auto wieder etwas in
Jan.Nom laid the.Dat car again something into
den Kofferraum.
the.Acc trunk
'Jan put something into the trunk of the car again.'
This sentence can be used if someone put something into the car's trunk before
and now Jan put something there too, or if Jan had done it before and did it
again now. The sentence cannot be used if Jan put something into the trunk of a

 
174  

truck before and now put something into the trunk of the car. Thus, again-
modification supports my analysis of part-whole applicatives.
The analysis can also be applied to unergative verbs (353).98 However, it
does not work exactly the same way because of the semantic type of the unerga-
tive verb, which is <v,t>. To resolve the type mismatch of AppPW, V and Voice,
I assume that V headmoves through AppPW to Voice, as shown in (353b). There,
V combines with Voice by event identification. The result of this becomes the
first argument of AppPW. The specifier of Voice becomes the second semantic
argument of AppPW. The specifier of App is adjoined outside of VoiceP where it
becomes the third semantic argument of AppPW. (It cannot stay in its base posi-
tion because it cannot combine with anything there.)
353. a. Dem Baum haben die Blätter geleuchtet.
the.Dat tree have the.Nom leaves glowed
'The leaves of the tree were glowing.'
b. VoiceP2
3
dem Baum2 VoiceP1
3
die Blätter Voice'
3
Voice2 AppP
3 3
App1 Voice1 t2 App'
3 3
App V t1 VP
6
leuchten
c. [[VoiceP2]] = λe. GLOW(e) & Agt(e)(the leaves) & the leaves  the
tree for the duration of e
Thus, (353a) denotes a glowing event of the leaves which are a material part of
the tree. This is the correct denotation.
The head movement that is available here can again be attributed to func-
tional composition. It is available because the composition can diverge in this
way. This derivation is only available for unergative verbs because of their se-
mantic type. AppPW does not move in transitive sentence because this movement
is a last resort movement. It is unnecessary with transitives and therefore may
not take place.


98 The full semantic derivation is given in Appendix 4, #3.


  175

Again-modification predicts two readings for this analysis of unergative


verbs, namely that again can attach to V or VoiceP2. These are the only nodes
of type <v,t>. These predictions are not both borne out.
354. Dem Baum leuchten die Blätter wieder.
the.Dat tree glow the leaves again
'The leaves of the tree are glowing again.'
This sentence can only mean that the leaves of the tree have glowed before and
are now glowing again. Again may not modify V: (354) cannot mean that there
was some glowing event before and now the leaves of the tree are glowing
again. That V itself is not available for again-modification is not limited to the
part-whole applicative construction though. It seems to be typical for unergative
verbs. Consider (355).
355. Die Blätter leuchten wieder.
the.Nom leaves glow again
'The leaves are glowing again.'
This sentence, too, cannot mean that something else has glowed before and now
the leaves are glowing. Thus, the missing reading of (354) must be ruled out for
independent reasons and is not a shortcoming of my analysis.
6.5 Consequences
My analysis (section 6.4.5) explains why the part-whole relation can hold be-
tween the applicative argument and the direct object in transitive sentences, the
applicative argument and the entity of an argument goal PP, or the applicative
argument and the subject of an unergative. Since the part-whole relation is
hardwired in the semantics, it must hold between these entities.
Another advantage of the explicit encoding of the part-whole relation is that
the negation pattern (325) can be explained. The established relation between
the two entities cannot be targeted by negation. Only the entities can be contras-
tively focused but their relation cannot be negated separately. The part-whole
applicative construction resembles the genitive because the genitive also estab-
lishes a relation between two entities. It is this relation that negation cannot tar-
get.
Also due to the encoding of the material part-whole relation in the semantics
(but not using a possessor raising construction), it can be explained why the part
NP can have a genitive modifier.
356. Jan zerstörte [dem Auto] [[das Fenster]
Jan.Nom destroyed the.Dat car the.Acc window
[der Tür]].
the.Gen door
'Jan destroyed the window of the door of the car.'

 
176  

The genitive and the applicative argument are not competing for the same posi-
tion or function. Both are licensed independently.
Another advantage of this analysis is that the referent of the part-whole ap-
plicative argument is not an event participant. Consequently, it does not undergo
the same event as the part NP (cf. (331)). At the same time, the encoded part-
whole relation guarantees that the whole entity must be available for the event to
take place.
To explain the fact that a full NP denoting the part cannot scramble within
the midfield, I suggest that the part-whole applicative argument moves to a mid-
field focus position (independently argued for by Frey (2000)). Since the high-
lighting of the whole NP is one characteristic which distinguishes this construc-
tion from similar constructions (309), this seems reasonable. The part NP is pre-
vented from scrambling over the applicative because the midfield focus position
is assumed to be the left-most position for full NPs in the midfield. A partial
structure is shown in (357). XP is assumed to be a projection that is optionally
available for pronouns to respect the left-most position requirement for pronouns
in the midfield.
357. CP
3
C'
3
C XP
3
FocusP
3
Focus'
3
Focus …
6
VoiceP
6.6 Conclusion
In this chapter, I have discussed the part-whole applicative argument construc-
tion. This applicative argument contributes only at-issue meaning. The part-
whole applicative argument can only occur with resultative VPs and is restricted
to NPs whose referents can be in a material part-whole relation. I have argued
that a phonologically null head, AppPW, introduces the applicative argument.
Furthermore, this head prevents its sister from combining semantically and takes
the elements of the sister node as its arguments. This allows the material part-


  177

whole relation to be encoded in the semantics without generating the relevant


entities as one constituent.
In the next chapter, I address another type of at-issue applicative, namely the
Dative of Inaction. As that discussion will make clear, the different types of at-
issue applicatives share few features that extend beyond their behavior in the
family of sentence tests. That discussion further shows that Pylkkänen's (2002)
identical analysis of all high applicatives is not suitable for German.

 
7 Dative of Inaction
7.1 Introduction
In this chapter, I discuss the last type of German applicative arguments, namely
the Dative of Inaction (DI):99
358. Der Oma zerbrach eine Vase.
the.Dat grandma broke a.Nom vase
'The vase broke on grandma.'
(Hens 1997, example 2)
Though Hens gives the translation of the DI with "on grandma" in (358), a better
translation would be "on grandma's watch" or "Grandma let the vase break".
This translation avoids ambiguity with the affected experiencer in the English
translation, and it reflects the meaning of the DI better because it indicates that
the grandma was in some way in charge of the vase and could have prevented its
breaking. This is the meaning of the DI in (358).100 Following Hens (1997) and
Rosengren (1975), I argue that the Dative of Inaction needs to be distinguished
from other applicative arguments because of its semantic contribution and its
limited occurrence.
The DI is an applicative argument according to my definition (4). It is not
governed by a preposition (358), and it is not selected by the verb (359).
359. Eine Vase zerbrach.
A.Nom vase broke
'A vase broke.'
In (359), the involvement of another entity is not entailed (compared to (358)).
Thus, the DI meets all requirements of an applicative argument.
In terms of the classification of applicatives by Pylkkänen (2002), the DI
should be a high applicative argument because no transfer of possession is in-
volved. In fact, no possession requirement exists for the DI. In (358), it is not
necessarily the grandma's vase that broke (Hens 1995). Yet, as described below,
the DI cannot occur with unergative or stative verbs as would be expected for
high applicatives (Chapter 2). Nevertheless, Pylkkänen's high applicative analy-
sis is essentially the correct analysis for the DI, as I show in section 7.4.
I describe the features of the DI and how it can be distinguished from other
applicative arguments of German in section 7.2. In section 7.3, I address the at-
issue meaning contribution of the DI. In section 7.4, I present my analysis of the


99 The material presented here is also discussed in Bosse (2012).
100 Consequently, I use on X's watch in the translations below whenever no translation was
given in the original source.


  179

Dative of Inaction. Some consequences of this analysis are discussed in section


7.5 before the chapter is concluded in section 7.6.
7.2 Description
Abraham (1973) observes that there is a special type of applicative dative in
German that occurs only with intransitive verbs. This is the DI. Hens (1997)
characterizes the verbs with which the DI can occur as "nonagentive resultative
verbs" (p. 199) requiring a theme or patient subject with low volitionality. Simi-
larly, Rosengren (1975) argues that only intransitive verbs that are perfective or
bounded can license this applicative argument. Transitive sentences like (41a),
repeated below, are impossible with the applicative argument interpreted as a
DI.
360. Alex zerbrach Chris Bens Vase.
Alex.Nom broke Chris.Dat Ben.Gen vase
*'Alex broke Ben's case on Chris's watch.' (Dative of Inaction)
'Alex broke Ben's vase on Chris.' (Affected Experiencer)
This restriction also rules out intransitives that are agentive (361a), non-
resultative (≈stative) (361b), or unbounded (361c).
361. a. *Das Kind hat mir gespielt.
the.Nom child has me.Dat played
'The child played on my watch.'
b. *Der Hund hat mir existiert.
the.Nom dog has me.Dat existed
'The dog existed on my watch.'
c. *Das Paar hat mir getanzt.
the.Nom couple has me.Dat danced
'The couple danced on my watch.'
The acceptability of these sentences cannot be increased by context.
If a verb denotes a non-agentive, bounded event that cannot be prevented, it
does not license the DI either (362).
362. *Der Ton verklang mir.
the.Nom tone faded.away me.Dat
'The tone (of music) faded away on my watch.'
(Rosengren 1975, example 11, my translation)
Rosengren (1975) and Hens (1997) discuss the fact that the dative referent must
be able to act in order to prevent the change. (362) is unacceptable because the
tone cannot be stopped from fading away (eventually). Consequently, the verb
verklingen 'fade away' cannot license the DI, even though the verb is bounded
and non-agentive. In contrast, (358) is acceptable because grandma could have
done something to prevent the vase from breaking. If somehow, in a specific

 
180  

context, tones can be prevented from fading away, (362) would be acceptable (if
it was my responsibility to prevent the tone from fading away and I failed).
Thus, the semantic requirement on the verb is twofold: the verb (phrase) must
denote a bounded event and the event must be preventable.
In terms of categories of intransitive verbs presented by Sorace (2000), the
boundedness requirement of the verb means that change of state verbs (363a),
inchoatives (363b) as well as verbs denoting controlled, telic motion processes
(363c) license the DI.
363. a. Der Baum ist mir verrottet.
the.Nom tree is me.Dat rotten
'The tree rotted on my watch.'
b. Der Hund ist mir ertrunken.
the.Nom dog is me.Dat drowned
'The dog drowned on my watch.'
c. Das Paar ist mir in den Raum
the.Nom couple is me.Dat in the room
getanzt.
danced
'The couple danced into the room on my watch.'
Continuation of condition verbs are very marginally acceptable.
364. ??Der Hund hat mir überlebt.
the.Nom dog has me.Dat survived
'The dog survived on my watch.'
Sentences with these verbs can be made more acceptable by context. For in-
stance, (364) could be used in a case the dog survived and I was responsible for
preventing that from happening (e.g. I should have let it drown or tried to active-
ly kill it).
Verbs that are unacceptable with the DI are controlled non-motional unaf-
fecting processes (365a), uncontrolled processes (365b, c), change of location
(365d), and controlled affecting processes (365e).
365. a. *Das Kind hat mir gearbeitet.
the.Nom child has me.Dat worked
'The child has worked on my watch.'
b. *Das Kind hat mir gehustet.
the.Nom child has me.Dat coughed
'The child coughed on my watch.'
c. *Es hat mir auf die Schuhe geregnet.
it has me.Dat onto the shoes rained
'It rained onto my shoes on my watch.'


  181

d. *Er ist mir (an)gekommen.


he.Nom is me.Dat (Prt)come
'He came/arrived on my watch.'
e. *Der König hat mir triumphiert.
the.Nom king has me.Dat triumphed
'The king triumphed on my watch.'
These sentences cannot be made acceptable by adding more context. Thus, the
verbal restriction of the DI to only some intransitives is syntactic and not prag-
matic in nature.
Looking at the restrictions from another perspective, unergative verbs (such
as (365e)) are ruled out. Of the unaccusative verbs, only those that denote
changes of state or bounded/ resultative processes are allowed. In contrast,
(365b-d) are ruled out because they are either not resultative or the DI did not
have the potential to influence the event. This supports the observations by Hens
(1995) and Rosengren (1975) that it is only a limited, well-defined group of in-
transitive verbs that licenses the DI.
Hens (1997) argues that the verbal event must have a negative consequence
that could have been prevented by the referent of the applicative argument. Due
to this, Hens calls these applicatives Dative of Inaction because the referent does
not act to prevent the undesirable event. While Rosengren (1975) discusses a
similar interpretation of these applicatives, he does not want to fully commit to a
negative consequence, mainly because of the verb gedeihen 'flourish, blossom',
which indicates a positive consequence (366).
366. Die Blumen gediehen dem Gärtner.
the.Nom flowers flourished the.Dat gardener
'The flowers flourished on the gardener's watch.'
(Rosengren 1975, example 29, my translation)
Since the flourishing of flowers, as in (366), is generally considered a positive
consequence, the responsibility of the dative referent should not be characterized
as failing to prevent an event from happening but rather neutral as having been
in charge.
With the requirement to be able to act comes the requirement that the refer-
ent of the DI must denote a sentient being (although not necessarily human, un-
like Rosengren (1975) and Hens (1997) claim) (367).
367. a. Dem Blindenhund lief sein Herrchen auf die
the.Dat guide.dog ran his master onto the
Straße.
road
'His master ran onto the street on the guide dog's watch.'

 
182  

b. *Die Blumen gediehen dem Regen.


the flowers flourished the.Dat rain
'The flowers flourished because of the rain.'
All of these characteristics can be used to distinguish the DI from other
German applicatives. No other applicative argument is limited to such a specific
group of verbs as the DI is. Also, the DI contributes only at-issue meaning (sec-
tion 7.3). This can be used to distinguish the DI construction from affected expe-
riencers (Chapter 3), ethical datives (section 4.2) and subject co-referential ap-
plicatives (section 4.3) which all contribute (some) not-at-issue meaning. To dif-
ferentiate the benefactive (Chapter 5) from the DI, the für-paraphrase of the
benefactive can be used. The DI does not have a prepositional variant.101
There does not seem to be a syntactic way to differentiate the part-whole
applicative with a sentient referent and a non-agentive verb from the DI. Context
must be used in this case (see Chapter 6 and Hens (1995, p. 211)).
In summary, the DI denotes an individual who is responsible for a certain
event but not actively bringing it about. The individual could have acted in some
way to prevent the event from occurring. This applicative construction is only
possible with resultative, non-agentive verbs and sentient referents for the ap-
plicative argument.
7.3 At-Issue Meaning
In this section, I show that the Dative of Inaction contributes only at-issue mean-
ing based on the family of sentence test described in section 2.2. There is no
meaning of the DI that projects past a yes/no-question.
368. Verwelkten dem Gärtner die Blumen?
withered the.Dat gardener the flowers
'Did the flowers wither on the gardener's watch?'
The yes/no-question in (368) can be answered with "no" either if the flowers did
not wither (but for instance the herbs did) or if they withered while the gardener
was not in charge of them. By answering "no", no truth of anything is accepted.
This indicates that there is no not-at-issue meaning carried by the DI.
Similarly, there is no meaning that projects past negation (369).
369. Die Blumen verwelkten dem Gärtner nicht.
the.Nom flowers withered the.Dat gardener Neg
'The flowers didn't wither on the gardener's watch.'


101 Abraham (1973) argues that there is a construction with lassen 'let' that the DI is related
to. However, Rosengren (1975) shows that the lassen-paraphrase is not always available
for the DI.


  183

This sentence indicates that the flowers did not wither or did not wither on the
gardener's watch (but they may have withered on someone else's watch). This,
again, indicates that the applicative argument contributes only at-issue meaning
because no not-at-issue meaning is detected.
It is also possible to form a wh-question asking about the applicative argu-
ment (370), showing that it is syntactically accessible.
370. Wem verwelkten die Blumen?
who.Dat withered the flowers?
'Who is responsible for the flowers withering?' (or 'Who let the with-
er?', 'Who was in charge of the flowers when they withered?')
This accessibility of the applicative argument for wh-questions indicates that it
does not contribute any not-at-issue meaning. This can also be seen in an if-
clause where the applicative argument contributes its meaning to the conditional
(371).
371. Wenn dem Gärtner die Blumen verwelken
if the.Dat gardener the flowers wither
dann feuer ich ihn.
then fire I him
'If the flowers wither on the gardener's watch, then I will fire him.'
According to (371), the gardener will get fired if the flowers wither while he is
in charge of them. In contrast, if the applicative argument were not present in
(371), I would fire him as soon as/if the flowers wither. In that case, the garden-
er is not responsible for keeping the flowers from withering, but his fate is tied
to that of the flowers. This shows that the DI contributes at-issue meaning.
In summary, the DI contributes only at-issue meaning. The tests for not-at-
issue meaning did not detect any such meaning. This shows clearly that the DI is
not just a variant of the affected experiencer applicative argument although these
two applicatives share other features (such as the sentience requirement).
7.4 Analysis
Rosengren (1975) does not provide a formal analysis of the Dative of Inaction
but rather focuses on a description. Hens (1995) provides an analysis of these
datives in the framework of constructional semantics. Thus, no formal previous
analysis with a similar theoretical framework can be discussed.
In the system of low and high applicatives by Pylkkänen (2002), the Dative
of Inaction should be high because it relates an individual to an event and does
not involve a transfer of possession. Before modifying Pylkkänen's analysis to
account for the DI, I briefly discuss a Japanese construction seemingly similar to
the German Dative of Inaction, namely the Japanese adversity causative (372)
which has a nominative-marked element whose referent is affected by the event.

 
184  

372. Taroo-ga sensoo-ni-yotte musuko-o sin-ase-ta.


Taro-Nom war-by son-Acc die-Cause-Past
'Taro's son was caused to die on him by the war.'
(Pylkkänen 2002, example 158a)
Harley (2008) discusses the Japanese causative morpheme -sase-, stating that it
can be used to indicate different types of causation. The one that is relevant here
is the let-causative. It indicates an indirect causing of an event; the referent of
the nominative-marked constituent let the event happen. Although the German
DI and this indirect Japanese causative seem to be comparable in their meaning,
it is undesirable to have the same analysis for both. This is due to the fact that
the Japanese causative construction has been argued to involve a causative event
in addition to the verbal event (see Pylkkänen 2002, Harley 2008, and references
cited therein). As explained in section 3.5.3, negation should not be able to tar-
get the caused event. In DIs, however, it is possible to negate what would be the
caused event (cf. (369)). Also, native speaker intuition denies that there is any
causation involved in the DI construction (Hens 1995). Consequently, I argue
that DIs are different from the adversity causative of Japanese because the for-
mer are not bieventive, while the latter are (see Bosse (2013c) for further de-
tails).
I propose that the DI is introduced by a phonologically null head
Resp(onsible). This head assigns the DI the thematic role of Responsible. It has
the following denotation and truth conditions:
373. a. [[Resp]] = λx. λe. Responsible(e)(x)
b. That the individual x is responsible for event e is true iff it is possible
that x could have caused the event not to happen (in case the event
happened). 102
Thus, my analysis is essentially the analysis of high applicatives by Pylkkänen
(2002); a phonologically empty head is used to introduce the applicative argu-
ment and to assign a thematic role to the applicative argument. There is no not-
at-issue meaning involved. The derivation for (358), repeated below, proceeds
as follows.103
374. a. Der Oma zerbrach eine Vase.
the.Dat grandma broke a.Nom vase
'The vase broke on grandma.'
(Hens 1997, example 2)


102 Thanks to Satoshi Tomioka for putting my unorganized thoughts into the appropriate
words!
103 The full semantic derivation is given in Appendix 5.


  185

b.
3
Voice[-agentive] RespP
3
der Oma Resp'
3
Resp VP
3
zerbrach eine Vase
c. [[RespP]] = λe. BREAK(e) & Thm(e)(a vase) & Responsi-
ble(e)(grandma)
Thus, the sentence denotes a breaking event of the vase for which the grandma
is responsible (in the sense that she had the potential to prevent this). Voice is
non-agentive because only non-agentive verbs license DIs. In other words, I as-
sume that only non-agentive Voice can select RespP.
7.5 Consequences
One problem with my analysis as it stands is that it does not rule out the passive
of transitive verbs to appear with the DI. Yet, this configuration is impossible
(375).
375. Der Oma wurde die Vase zerbrochen.
the.Dat grandma became the.Nom vase broken
'The vase was broken for/on grandma.'
*'The vase was broken on grandma's watch.'
The unavailability of the DI interpretation of this sentence can be explained in
one of two ways. First, it could be argued that unaccusative verbs are fundamen-
tally different from transitives. For example, it could be said that the VP in
(374b) is not the same as a V plus Theme of a transitive verb; possibly the unac-
cusative has an InchoativeP rather than a VP. Second, it could be that passives
do not have a non-agentive Voice head, but that Voice head is agentive since the
agent is semantically still present in passives. The agent is then syntactically
suppressed by some other mechanism (cf. Bruening 2011). I assume this latter
option. Thus, (375) can be ruled out as a DI construction because the passive
requires the projection of agentive Voice but agentive Voice cannot select
RespP. Only non-agentive Voice can select ResP. This also rules out the DI oc-
curring with unergative (361a) and transitive verbs (360).
This analysis is further supported by adverbial modification patterns. As ex-
plained previously (Chapter 3), again can detect nodes of type <v,t>. According

 
186  

to the analysis in (374), there are two nodes of type <v,t>, namely RespP and
VP.104 Thus, the following two readings should be possible:
376. a. VP: The event happened before with someone else or no one in
charge.
b. RespP: The event happened before with the same person in charge.
These predictions are borne out (377).
377. Dem Verkäufer ist das Eis wieder
the.Dat salesman is the ice.cream again
geschmolzen.
melted
'The ice cream melted again on the salesman's watch.'
i. The ice cream was melted before and now it melted again while the
salesman was in charge of it.
ii. The ice cream melted before on the salesman's watch and now it
melted on his watch again.
As expected, it is not possible to have a repetition only of the salesman being in
charge by using wieder 'again' (378).
378. Dem Verkäufer ist eine Blume verrottet und
the.Dat salesman is a.Nom flower rotted and
dann ist ihm wieder ein Eis
then is him.Dat again a ice.cream
geschmolzen.
melted
'A flower rotted on the salesman's watch and then ice cream melted on
his watch again.'
This sentence can only mean that the ice cream had melted before. It cannot be
used to convey that the salesman was in charge of the flower rotting and then
was again in charge of the ice cream melting (for the first time). This is expected
on my analysis because there is no <v,t> node representing only the being in
charge. Thus, again-modification supports my analysis of part-whole applica-
tives as high applicatives with non-agentive Voice.
7.6 Conclusion
In this chapter, I have discussed the Dative of Inaction. It occurs with non-
agentive, bounded intransitive verbs. It denotes that the referent of the DI was in
charge of that event (in the sense that s/he was not actively bringing it about but
it happened on that person's watch). That individual could have acted in a way

104 Non-agentive VoiceP is also of type <v,t> but it would have the same interpretation as
when again modifies RespP because Voice is vacuous.


  187

that would have stopped the event from occurring. The DI construction has only
at-issue meaning and is mono-eventive. Consequently, it is best explained with
the approach to high applicatives by Pylkkänen (2002). A causative analysis is
not suitable (or necessary). In addition to the high applicative structure, the Resp
head that introduces the DI selects VP and is selected by non-agentive Voice.
This explains the restricted occurrence of the DI.
This analysis is rather ad hoc as it stands. Further research will show whe-
ther this mono-eventive counterpart to the Japanese adversity causative is avail-
able in other languages as well. Yet, as I show in the next chapter, my analysis
of the DI fits well into the system of German applicatives because all the ex-
pected co-occurrence restrictions of applicatives are observed with this analysis
of the DI.

 
8 Co-Occurrence of Applicatives
8.1 Introduction
In this chapter, I discuss the co-occurrence possibilities of applicative argu-
ments. Vogel and Steinbach (1998) speculate that the co-occurrence of multiple
(dative-marked) applicatives in German is "restricted only semantically (i.e.
each dative needs a different interpretation)" (p. 27). It follows from this that no
applicative should be recursive because the applicative head would assign the
same thematic role twice.105 It also follows that, if Vogel and Steinbach (1998)
are right, all applicatives should be allowed to appear together because each of
them is associated with a different thematic role. I show below that this is true
for English (section 8.3) and German (section 8.2) as long as the individual se-
lectional and verbal requirements for each applicative are met.
8.2 German
In German, no applicative head is recursive (379). This supports the contention
by Vogel and Steinbach (1998) that a different interpretation for each applica-
tive argument is a necessary requirement for the co-occurrence of applicatives.
379. Recursivity of Applicatives
a. *Der Mann zerbrach dem Kind den
the.Nom man broke the.Dat child the.Dat
Eltern das Rad.
parents the.Acc bike
'The man broke the bike on the child on the parents.'
(Affected Experiencer)
b. *Der Oma zerbrach dem Mann die
the.Dat grandma broke the.Dat man the.Acc
Vase.
vase
'The grandma was responsible for preventing the man who was re-
sponsible for the vase from breaking it.'
(Dative of Inaction)


105 Parsons (1990) also disallows having the same event participant twice per event (see
Chapter 2).


  189

c. *Komm mir mir pünktlich nach Hause!


come me.Dat me.Dat on.time to home
'Come home on time and I want that to happen and I want that to
happen.'
(Ethical Dative)
d. *Ich trinke mir mir einen Kaffee.
I.Nom drink me.Dat me.Dat a.Acc coffee
'I am gonna drink me a cup of coffee while enjoying myself.'
(Subject Co-Referential Applicative)
e. #Ich mal dir dem Papa ein Bild.
I.Nom paint you.Dat the.Dat dad a.Acc picture
not available: I draw a picture for your benefit for dad's benefit.'
(True Benefactive)
f. *Ich zerbeule dem Auto jetzt der Seite
I.Nom dent the.Dat car now the.Dat side
die Tür.
the.Acc door
'I am now denting the door of the side of the car.'
(Part-whole Applicative)
However, not all different types of applicatives can co-occur despite having
different thematic roles. Typically, if the requirements on the verb licensing the
different types of applicative arguments do not overlap, the applicatives cannot
co-occur. The following table gives an overview of the possible co-
occurrences.106

Table 2: Co-occurrences of German Applicative Arguments

Aff Exp DI Ethical Sub Co-Ref True ben Part-Whole


Aff Exp * * + * * *
DI * * ? * * *
Ethical + ? * ? + +
Sub Co-Ref * * ? * * *
True ben * * + * * +
Part-whole * * + * + *


106 The table only shows the co-occurrence possibilities of two applicatives. The possibili-
ties for three applicatives are discussed in section 8.2.7. + marks grammatical combina-
tions, * marks ungrammatical combinations, ? marks marginal combinations.

 
190  

As can be seen from the table, there are only a few options for applicatives to
co-occur. I explain this pattern in the following subsections.

8.2.1 Ethical Dative


The ethical dative can co-occur with the most other types applicatives (Lamiroy
and Delbecque 1998, among other). Only the co-occurrence with the DI (380a)
and the subject co-referential applicative (380b) are limited (in addition to it not
being recursive (379c)).
380. Co-Occurrence of Ethical Datives
a. ?Verwelk mir schon dem Gärtner!
wither me.Dat already the.Dat gardener
'Wither me on the gardener's watch and I want that to happen!'
(DI + Ethical Dative)
b. ?Du sollst mir dir hier jetzt noch
you.Nom shall me.Dat you.Dat here now still
einen Kaffee trinken.
a.Acc coffee drink
'You shall drink you a cup of coffee here now and I want that to
happen.'
(Sub Co-Referential + Ethical Dative)
c. Zerbrich mir ihm jetzt die Vase!
break me.Dat him.Dat now the.Acc vase
'Break the vase on him now and I want that to happen!'
(Affected Experiencer + Ethical Dative)
d. Mal mir dem Papa ein Bild!
draw me.Dat the.Dat dad a.Acc picture
'Draw dad a picture and I want that to happen!'
(True Benefactive + Ethical Dative)
e. Reiß mir dem Baum die Blätter ab!
rip me.Dat the.Dat tree the.Acc leaves off
'Rip the tree's leaves off and I want that to happen!'
(Part-Whole + Ethical Dative)
The marginal acceptability of (380a) can be attributed to the fact that DIs require
non-agentive verbs with patient subjects (Chapter 7), whereas the directive force
of the ethical dative requires an addressee who can follow orders (Chapter 4).
Thus, the two applicatives can only marginally co-occur because of their differ-
ent requirements. In this sentence, a flower is addressed which cannot volitional-
ly follow the order. However, this sentence could be used in a humorous con-
text. The marginality of (380b) is explained by a semantic clash of the special
involvement conveyed by the subject co-referential applicative and the directive


  191

force required for the ethical dative. The addressee is ordered to enjoy some-
thing but it is not really possible to order enjoyment. Typically, phrases like
"Enjoy yourself!" and "Have fun!" (and their German equivalents) are wishes
rather than orders (and, thus, they do not have directive force). However, while
this sentence is semantically marked, it indicates what the subject wants to hap-
pen and therefore is not completely unacceptable.
The grammaticality of the ethical dative co-occurring with the affected ex-
periencer (380c), the true benefactive (380d), and the part-whole applicative
(380e) is explained because all of these applicatives attach outside of VP while
the ethical dative is higher up in the syntactic structure. Consequently, each of
them is licensed as long as its individual requirements are met. This is schema-
tized in (381). (Here, App stands for the different applicative heads of affected
experiencers, true benefactives, and part-whole applicatives.)
381. 3
[DIRECTIVE] AppEDP
3
Ethical Dative AppED'
3
AppED 3
[IRREALIS] …
3
AppP
3
Applicative App'
3
App VP
Unlike many other applicatives, the ethical dative can also occur with se-
lected datives (Wegener 1983, Vogel and Steinbach 1998) and ditransitives
(382).
382. a. Hilf mir jetzt dem Kind!
help me.Dat now the.Dat child
'Help the child now and I want that to happen!'
b. Gib mir dem Jungen das Buch!
give me.Dat the.Dat boy the.Acc book
'Give the boy the book and I want that to happen!'
This can also be attributed to the fact that the ethical dative attaches so high in
the syntactic structure that it cannot impose any restrictions on the VP (cf. We-
gener 1983).

 
192  

8.2.2 Dative of Inaction


As shown above, the Dative of Inaction is not recursive (379b) and can only
marginally occur with the ethical dative (380a). It cannot occur with any other
applicative. In (383) are examples for the ungrammaticality of the DI co-
occurring with the affected experiencer (383a) (also see Hens 1995), the true
benefactive (383b), and the part-whole applicative argument (383c). The subject
co-referential applicative construction requires verbs of consumption whereas
the DI requires non-agentive, unaccusative verbs. Consequently, a sample sen-
tence that respects both restrictions cannot be formed.
383. Co-occurrence of DI
a. *Der Oma zerbrach ihm die Vase.
the.Dat grandma broke him.Dat the vase.Acc
'On grandma's watch, the vase broke on him.' or 'On his watch, the
vase broke on grandma.'
(Affected Experiencer + DI)
b. *Der Oma zerbrach der Frau
the.Dat grandma broke the.Dat woman
die Vase.
the.Acc vase
'On grandma's watch, the vase broke for the woman.' or 'On the
woman's watch, the vase broke for grandma.'
(True Benefactive + DI)
c. *Dem Gärtner verwelkten dem Busch
the.Dat gardener withered the.Dat bush
die Blüten.
the.Nom blossoms
'The blossoms of the bush withered on the gardener's watch.'
(Part-Whole Applicative + DI)
These sample sentences (383) show that the DI applicative phrase RespP cannot
be selected by any other applicative. This is expected because it must be selected
by non-agentive Voice (and select VP), as schematized in (384).


  193

384. 3
Voice[-agentive] RespP
3
DI Resp'
3
Resp VP
3
V NP
Consequently, no applicative can interfere in this structure because Resp can
only occur between Voice and VP. The higher attaching ethical dative is marked
when co-occurring with the DI because of different verbal requirements (section
8.2.1).

8.2.3 Subject Co-referential Applicative


The subject co-referential applicative cannot co-occur with any other applicative
argument: subject co-referential (379d), ethical dative (380b), DI (section 8.2.2),
affected experiencer (385a), true benefactive (385b), and Part-Whole Applica-
tive (385c).
385. Co-occurrence of Subject Co-referential Applicative
a. *Ich trinke mir jetzt meinem Mann
I.Nom drink me.Dat now my.Dat man
einen Kaffee.
a.Acc coffee
'I'm gonna drink me a cup of coffee now on my husband.'
(Affected Experiencer + Subject Co-Referential)
b. *Ich trinke mir dir einen Kaffee.
I.Nom drink me.Dat you.Dat a.Acc coffee
'I am gonna drink me a cup of coffee for you.'
(True Benefactive + Subject Co-Referential)
c. *Ich esse mir jetzt dem Baum die Blätter.
I.Nom eat me.Dat now the.Dat tree the.Acc leaves
'I'm gonna eat me the tree's leaves now.'
(Part-Whole + Subject Co-Referential)
This pattern of ungrammaticality points to selectional restrictions on the subject
co-referential applicative, namely that it cannot be selected by any other applica-
tive. Rather it must select VP and be selected by Voice (as reflected in its deno-
tation). The syntactic derivation of the subject co-referential applicative argu-
ment is shown in (386).

 
194  

386. 3
Subject1 3
1 3
t1 Voice'
3
Voice ApplSCRP
3
a1 ApplSCR'
3
ApplSCR VP
There is no place for any other applicative argument to attach in this structure.
The higher attaching ethical dative is only marginally available to co-occur with
the subject co-referential applicative argument for semantic reasons (section
8.2.1).

8.2.4 Affected Experiencer


The affected experiencer applicative can occur with the ethical dative (380c) but
no applicative argument discussed so far: affected experiencer (379a), DI (383a)
and subject co-referential applicative (385a). Also, true benefactives and part-
whole applicatives co-occurring with affected experiencers are ungrammatical
(387).
387. Co-occurrence of Affected Experiencer
a. *Ich mal ihm meiner Mutter ein Bild.
I.Nom draw him.Dat my.Dat mother a.Acc picture
'For him, I draw a picture on my mom.' or 'For my mother, I draw a
picture on him.'
(True Benefactive + Affected Experiencer)
b. *Jan zerbeulte seinem Freund gestern
Jan.Nom dented his.Dat friend yesterday
dem Auto die Tür.
the.Dat car the.Acc door
'Jan dented the car's door on his friend yesterday.'
(Part-Whole + Affected Experiencer)
This ungrammaticality can be explain by selectional restrictions. Aff selects
VP, and AffP cannot be selected by any other applicative head. It must be se-
lected by Voice (388). I assume that this is a selectional restriction of the Aff
head. As a consequence, the affected experiencer cannot co-occur with any other
applicatives (except the higher attaching ethical dative (380c, 381)).


  195

388. 3
Voice AffP
3
Affected Experiencer Aff'
3
Aff VP
The same pattern of co-occurrence also holds for the pertinence dative. This
is expected because I propose in Chapter 3 that the pertinence dative has the
same structure as the affected experiencer.
389. a. *Die Katze zerkratzt dir Maria gestern
the.Nom cat scratched you.Dat Maria.Dat yesterday
den Finger.
the.Acc finger
'The cat scratched Maria's finger on you yesterday.' of 'The cat
scratched your finger on Maria yesterday.'
(Affected Experiencer + Pertinence Dative)
b. *Gott erschuf Eva Adam den Arm.
God.Nom created Eva.Dat Adam.Dat the.Acc arm
'God created Eve's arm for Adam.' or God created Adam's arm for
Eve.'
(True Benefactive + Pertinence Dative)
c. *Der Finger erfror dem Arzt Maria
the.Nom finger froze the.Dat doctor Maria.Dat
'Maria's finger froze on the doctor's watch.' or 'The doctor's finger
froze on Maria's watch.'
(DI + Pertinence Dative)
d. *Ich esse mir dem Mann das Bein.
I.Nom eat me.Dat the.Dat man the.Acc leg
'I'm gonna eat me the man's leg.'
(Subject Co-Referential + Pertinence Dative)
e. Tritt mir Maria auf den Mantel!
step me.Dat Maria.Dat onto the.Acc coat
'Step onto Maria's coat and I want that to happen!'
(Ethical Dative + Pertinence Dative.)

8.2.5 True Benefactive


True benefactives cannot co-occur with Datives of Inaction (383b), subject co-
referential applicatives (385b), affected experiencers (387a) or themselves

 
196  

(379e). However, they can occur with ethical datives (380d) or part-whole ap-
plicatives (390).
390. Ich zerbeule dir dem Auto die Tür.
I.Nom dent you.Dat the.Dat car the.Acc door
'I'm denting the car's door for you.'
(Part-Whole + True Benefactive)
Since the part-whole applicative must select VP (Chapter 6, 8.2.6), this indi-
cates that the true benefactive can either select VP or a part-whole applicative. It
cannot occur with the DI because the DI must be selected by non-agentive Voice
and select VP. The same is true for the subject co-referential applicative argu-
ment (except that Voice is agentive). Aff is also hypothesized to require Voice
and VP as its surrounding projections (see section 8.2.4) and therefore the true
benefactive cannot interfere. Part-whole applicatives are the only applicatives
that do not have to be selected by Voice. The AppPWP projection is therefore
free to be selected by another applicative, such as the true benefactive (391).107
391. VoiceP
3
Subject Voice'
3
Voice BenP
3
True Benefactive Ben'
3
Ben VP/AppPW
As discussed in Chapter 5, the true benefactive can alternatively be ex-
pressed as a für 'for'-PP. If this PP is used, the true benefactive can appear with
all applicatives, except the Dative of Inaction (392a) and another true benefac-
tive (392b).
392. a. *Der Oma zerbrach die Vase für Sarah.
the.Dat grandma broke the.Nom vase for Sarah
'The vase broke on grandma's watch for Sarah.'
(DI + für-PP)
b. #Für seine Mutter malt Jan seinem Vater ein Bild.
for his mother draws Jan his.Dat father a picture
'For his mother, Jan is drawing a picture for his dad.'
(True Benefactive + für-PP)


107 Not requiring the true benefactive to select VP results in losing the explanation of why
they can only appear with transitive verbs.


  197

c. Putz mir die Schuhe für ihn!


clean me.Dat the.Acc shoes for him
'Clean the shoes for him and I want that to happen!'
(Ethical Dative + für-PP)
d. Jan zerbrach Alex Bens Vase für
Jan.Nom broke Alex.Dat Ben.Gen vase for
seine Freundin.
his girlfriend
'Jan broke Ben's vase for his girlfriend on Alex.'
(Affected Experiencer + für-PP)
e. Ich trink mir noch ein Bier für dich.
I.Nom drink me.Dat Prt a.Acc beer for you
'I'm gonna drink me another beer for you.'
(Subject Co-Referential + für-PP)
f. Jan zerbeulte dem Auto die Tür
Jan.Nom dented the.Dat car the.Acc door
für seinen Freund.
for his friend
'Jan dented the car's door for his friend.'
(Part-Whole + für-PP)
The ungrammaticality of (392a) can be attributed to the fact that the DI requires
a non-agentive verb, and it is not possible to construe that the spontaneously oc-
curring event happens for the benefit of someone. Benefactive actions often
need an intentional action of the agent (Kittilä and Zuniga 2010). The marked-
ness of (392b) just shows again that the true benefactive and the für-PP assign
the same thematic role and can consequently not co-occur (cf. Chapter108 5). The
grammaticality of the other examples shows that the für-PP adjoins to any ap-
plicative above VP and does not interfere with the (high-attaching) ethical dative
(393).109


108 The sentence is acceptable if the dative and the PP receive different thematic roles.
109 This structure makes specific predictions for again-modification which I do not further
investigate here.

 
198  

393. VoiceP
3
Subject Voice'
3
Voice VP
3
App VP
3
PP VP1
3
für NP
The availability of the für-PP also demonstrates that the ungrammatical occur-
rences of the true benefactive with other applicatives are due to syntactic re-
strictions and not semantic restrictions (except possibly the DI) because the
thematic role assigned by the für-PP and the true benefactive are the same, but
the two cannot co-occur with the same applicatives.

8.2.6 Part-Whole Applicative


The part-whole applicative can co-occur with the ethical dative (380e) and the
true benefactive (390), as schematized in (394).
394. AppEDP
3
Ethical Dative …
3
Voice BenP
3
True Benefactive Ben'
3
Ben AppPWP
3
Part-Whole App AppPW'
3
AppPW VP
The order of the true benefactive and the part-whole applicative argument are
fixed due to the fact that the part-whole applicative must select VP because of its
semantic denotation: the part-whole must select VP and can be selected by the
true benefactive. Again, the ethical dative is too high to impose selectional re-
strictions on the verb or the part-whole applicative. Consequently, the two ap-
plicatives can co-occur (380).


  199

The DI cannot appear with the affected experiencer (387b), Dative of Inac-
tion (383c), subject co-referential (385c), or itself (379f).

8.2.7 Summary
Vogel and Steinbach (1998) are right that the co-occurrence possibilities of ap-
plicatives are semantically restricted. No applicative is recursive. Furthermore,
most German applicatives select VP and are consequently restricted in their co-
occurrence options since only one VP is available. The exceptions are the ethical
dative and the true benefactive. The former attaches outside of IP, selecting the
projection including [IRREALIS] and does not impose verbal restrictions (beyond
the ones carried by directive force, e.g. no stative verbs are possible). The latter
can select VP directly or attach outside of the part-whole applicative.
The affected experiencer applicative, the part-whole applicative, the Dative
of Inaction and the subject co-referential applicative must select VP. The DI,
furthermore, must be selected by non-agentive Voice. Consequently, these ap-
plicatives are restricted in their co-occurrence options.
Given these selectional requirements, it should be possible to have the fol-
lowing three applicatives co-occurring: part-whole applicative, true benefactive
and ethical dative. Yet, this combination is only marginally possible (395).
395. ?Zerbeul mir dem Boss (noch) dem
dent me.Dat the.Dat boss Prt the.Dat
Auto die Tür!
car the.Acc door
'Dent the car's door for the boss and I want that to happen!'
(Part-Whole, True Benefactive, Ethical Dative)
I contend that this marginal grammaticality of three applicatives co-occurring is
due to the availability of the für-PP for the true benefactive. The acceptability of
sentence (395) is improved if this PP is used (396).
396. Zerbeul mir für den Boss (noch) dem
dent me.Dat for the boss Prt the.Dat
Auto die Tür!
car the.Acc door
'Dent the car's door for the boss and I want that to happen!'
(Part-whole applicative, für –PP, Ethical Dative)
(396) shows that the sentence in (395) should be semantically acceptable. Thus,
there must be some other reason for its marginality. I contend that it is the se-
mantically equivalent für-PP which must be present to avoid a higher processing
load imposed the three dative-marked applicatives in (395).

 
200  

8.3 English
For English, it should also be true that no applicative is recursive because it
would lead to having one thematic role used twice for the same event. Indeed,
none of the three English applicatives can co-occur with itself (397).
397. a. *Gerald broke the vase on his son on his wife.
(Affected Experiencer)
b. *Gerald bought her Mike a car.
(Recipient Benefactive)
c. *Gerald1 bought him1 him1 a car.
(Subject Co-referential Applicative)
Thus, the contention of Vogel and Steinbach (1998) that the co-occurrence of
applicative arguments is semantically restricted also holds for English.
The affected experiencer may co-occur with the recipient benefactive (398).
398. Gerald bought her a car on his father.
This shows that the affected experiencer can adjoin regardless of the presence of
Rec. This is schematized in (399).
399. VoiceP
3
Subject Voice'
3
Voice vP
3
PP vP
3 3
on NP NP1 vP
3
NP2 v'
3
v RecP
3
t1 Rec'
3
Rec VP
The affected experiencer is marginally acceptable with the subject co-
referential applicative (400).
400. ?Gerald1 bought him1 a motorcycle on his wife.
This again supports the affected experiencer PP's adjunct status. It can adjoin to
VP without interfering with the subject co-referential applicative, as schema-
tized in (401).


  201

401. VoiceP
3
Subject1 3
1 3
t1 Voice'
3
Voice ApplP
3
a1 Appl'
3
Appl VP
3
PP VP
3
on NP
The recipient benefactive and the subject co-referential applicative cannot
co-occur (402) (Christian 1991):
402. *Gerald1 bought him1 her a car.
This is due to the fact that both of them select VP, and there is only one VP.
Thus, it is expected that they cannot co-occur.
It is possible to have an applicative co-occur with a PP that expresses the
semantic function of another applicative. For instance, even though the subject
co-referential applicative and the recipient benefactive cannot co-occur (402),
the subject co-referential applicative can occur with a for-PP expressing the re-
cipient (403).
403. a. He's gonna buy (*him) his son a pick-up.
b. He's gonna buy {him/ *himself} a pick-up for his son.
(Horn 2009, examples 22a, b)
This shows that this co-occurrence restriction is syntactic and not semantic. As
in German (section 8.2), selectional requirements of the applicatives can explain
this. The following tables sums up the co-occurrence patterns of English ap-
plicative arguments.

Table 3: Co-occurrences of English Applicative Arguments

Aff Exp Recipient Sub Co-Ref

Aff Exp * + ?

Recipient + * *

Sub Co-Ref ? * *

 
202  

8.4 Conclusion
In this chapter, I have discussed the possibilities of different applicative argu-
ments co-occurring. No applicative head is recursive. Furthermore, in German
and English some applicative heads must select VP. This limits the co-
occurrence possibilities of different applicatives. It was shown that as long as
each applicative obeys its individual selectional and verbal restrictions, different
applicatives can co-occur.
Overall this discussion has shown that the system of applicative arguments
that I have presented in this book for German and English, respectively, is con-
sistent because the different types of applicative arguments can occur together in
a way that is predicted by the analyses I have proposed for the different types of
applicative arguments.


9 Conclusion
Throughout this book, I have discussed the applicative arguments of German
and English. For both languages, I have distinguish the different types based on
their syntactic behavior, their meaning, and whether the meaning is contributed
to the not-at-issue tier or the at-issue tier of meaning. German has three types of
applicative arguments that contribute only at-issue meaning: true benefactives,
Datives of Inaction, part-whole applicative arguments. It has two types of not-at-
issue applicatives: ethical datives and subject co-referential applicatives. Finally,
it has one type of applicative argument that contributes both at-issue and not-at-
issue meaning: affected experiencers. English has one applicative for each type
of meaning contribution: the affected experiencer contributes both at-issue and
not-at-issue meaning, the recipient benefactive contributes only at-issue mean-
ing, and the subject co-referential applicative contributes only not-at-issue
meaning. The existence of the affected experiencer applicative arguments with
both at-issue and not-at-issue meaning has shown that the system of two tiers of
meaning proposed by Potts (2005) needs to be adjusted to allow for such ele-
ments.
The fact that each type of applicative argument displays unique behaviors
and restrictions supports my contention that the broad categorization of applica-
tive arguments proposed by Pylkkänen (2002) is not sufficient. She differenti-
ates only two basic types of applicative argument: high and low. As I have
shown, the ones that she bundles as high applicatives do not behave alike and
cannot all be given the same analysis. Furthermore, the discussion of the Ger-
man ethical datives has shown that there are applicative arguments that attach
outside of VoiceP. Because of that, they do not fit Pylkkänen's system at all.
All of my analyses have relied on a functional head (more often than not
phonologically null) to introduce the applicative argument. Some of these heads
were simple heads in the sense that they only assign a thematic role to the argu-
ment. Others are more complex in that they either require functional composi-
tion or are bi-eventive. These phonologically null functional heads may be con-
sidered a flaw of these analyses, as there is at least one approach to the analysis
of applicative structures that does not require phonologically null functional
heads, namely the approach taken by Schöffer (1992) who argues that the dative
in German has one of the following three functions: recipient, benefactive or
experiencer. He claims that there are no different types of datives but rather that
the perceived types are actually inferences made based on the essential functions
of the dative. There is something to be said about the dative case and why all
German applicatives are dative-marked (but not only applicatives are dative-


204  

marked). However, the differences between the types of applicatives that I have
argued for here cannot be captured by the approach by Schöffer (1992): the part-
whole applicative argument as well as the existence of both the Dative of Inac-
tion and the affected experiencer speak against his approach. The part-whole
applicative argument does not have any of the dative interpretations he considers
essential; it does not denote an experiencer in any way, nor does the entity bene-
fit or receive anything. It is a far stretch to infer the part-whole relation and dis-
regard other possibilities such as a benefactive interpretation. The latter should
be more easily available if the dative typically carries this interpretation. The
existence of both the Dative of Inaction and the affected experiencer results in a
similar problem: many sentences are ambiguous between the two interpretation
but it is unclear where the "responsibility"-interpretation of the DI comes from
in Schöffer's approach if the affect experiencer is also available. However, it is
true that the case-marking for applicatives in German needs to be discussed.
Saying that all applicatives receive dative case in German is a good observation
but it is not an explanation. Also, it does not explain why only the ethical dative
can appear with verbs that assign dative case to their complements (e.g. helfen
'help', folgen 'follow', danken 'thank', gleichen 'be like'). No other German ap-
plicative can appear with lexical datives. Further research has to explain this in-
teraction of dative-marked applicatives and verbs assigning dative case lexically.
Further support for analyzing German and English applicatives as involving
phonologically null heads could come from languages that have overt mor-
phemes marking the presence of applicative arguments such as many Bantu and
Native American language. The correspondence of the applicatives discussed
here and the ones in those languages requires further investigation. All the ap-
plicatives that I have posited for German and English have equivalent (or at least
very similar) counterparts in other languages. The one head found in English
that is not found in German is the recipient benefactive. However, the posited
structure can also be found (as a non-applicative structure) in German (as well
as in other languages as an applicative). The German applicatives not found in
English also have cross-linguistic counterparts. True benefactives have been dis-
cussed for many languages (Kittilä and Zuniga 2010). The Dative of Inaction
has a similar Japanese structure, the ethical dative has an analogy in the French
ethical dative (which has a slightly different meaning (cf. Authier and Reed
1992, Jouitteau and Rezac 2008)), the part-whole applicative can be viewed as
being similar to the Korean inalienable possessor construction (although it does
not have exactly the same requirements and meaning). Thus, no posited applica-
tive is unique when put in a cross-linguistic perspective. The cross-linguistic
perspective might help justifying the phonological null functional heads.


  205

Finally, the relationship between the availability of an applicative argument


and a prepositional phrase in languages like German and English can help to
support the phonologically null heads. German and English both have applica-
tives that cannot be expressed alternatively with a PP (e.g. subject co-referential
applicatives), applicatives that can alternatively be expressed as PPs (benefac-
tives), and PPs that cannot be expressed as applicatives, e.g. instrumentals (cf.
Chapter 1). Clearly, there is some link between applicative structures and prepo-
sitional phrases that needs further investigation.
In short, the phonologically null heads for different types of applicative ar-
guments in German and English are reasonable although further research is
needed. This further research must also address questions that have not been ful-
ly answered here. For instance, there are still many unanswered questions with
respect to the verbal restrictions in German: Why are there verbal restrictions on
essentially all applicatives in German when that is cross-linguistically not the
case? The restrictions discussed here are as follows: the affected experiencer and
part-whole applicatives require resultative VPs, with other factors such as inten-
tionality playing a secondary role. Benefactive applicatives are best with transi-
tive verbs. The subject co-referential applicative requires verbs of consumption
(or marginally of creation). The Dative of Inaction requires non-agentive,
bounded intransitive verbs. Where do these different restrictions come from? Is
there a way to clearly test these restrictions, so that the acceptability of a certain
applicative can be predicted better? It was shown that the resultativity of affect-
ed experiencers does not correlate with typical tests of resultativity (such as the
stative passive). Hole (2008) and Beck and Snyder (2001) might be on the right
track on how to operationalize this particular restriction but further research, es-
pecially with respect to the other applicatives and their restrictions, is needed.
Another question that still needs to be addressed is the definition of what
applicative arguments really are. I have used a working definition of applicative
arguments (4) throughout this book. Sometimes, prepositional variants were in-
cluded but PPs were not generally discussed. Furthermore, the definition as it is
cannot immediately be extended to languages that mark applicatives with a ver-
bal morpheme (because in that case this morpheme needs to be omitted in addi-
tion to the nominal element for the omission test). So the question remains: how
can applicatives be defined? Furthermore, the discussion here has shown that
even within a language, the behaviors of the different applicative arguments are
quite varied (in terms of meaning contribution, restrictions and syntactic behav-
ior). Thus, the term applicative argument seems to be a cover term for a variety
of non-obligatory arguments. The individual categories covered by it and how
they can be identified, need further investigation - but the first step (for German
and English) was taken here.

 

Appendix 1: Affected Experiencers
1) German Affected Experiencer (69)
a. Alex zerbrach Chris Bens Vase.
Alex.Nom broke Chris.Dat Ben.Gen vase.Acc
'Alex broke Ben's vase on Chris.'
b. VoiceP<v,t>
3
Alex Voice'
3
Voice AffP<v,t>
3
Chris Aff'
3
Aff<vt,evt> VP<v,t>
6
zerbrach Bens Vase
c. (AffP and Voice combine by Event Identification.)
[[VP]] = λe. BREAK(e) & Thm(e)(Ben's vase)
[[Aff]] = λPvt.λx.λe. P(e) & e'(EXPER(e') & Exp(e')(x)) : e'' (P(e'') →
Source(e'')(e'))
[[Aff']] = λx.λe. BREAK(e) & Thm(e)(Ben's vase) & e'(EXPER(e') &
Exp(e')(x)) : e'' (BREAK(e'') & Thm(e'')(Ben's vase) → Source(e'')(e'))
[[AffP]] = λe. BREAK(e) & Thm(e)(Ben's vase) & e'(EXPER(e') &
Exp(e')(Chris)) : e'' (BREAK(e'') & Thm(e'')(Ben's vase) →
Source(e'')(e'))
[[Voice]] = λx. λe. Agt(e)(x)
[[VoiceP]] = λe. BREAK(e) & Thm(e)(Ben's vase) & Agt(e)(Alex) &
e'(EXPER(e') & Exp(e')(Chris) : e''((BREAK(e'') & Thm(e'')(Ben's
vase)) → Source(e'')(e'))
(Bosse et al. 2012, modified examples 67, 68)


208  

2) English Affected Experiencer (72)


a. Gerald broke the vase on me.
b. VoiceP
3
Gerald Voice'
3
Voice VP1
3
VP PPAff
3 3
V NP PAff NP
broke 5 on 4
the vase me
c. (VP1 and Voice combine by Event Identification.)
[[on]] = λx.λP.λe. P(e) & e'(EXPER(e') & Exp(e')(x)) : e'' (P(e'') →
Source(e'')(e'))
[[PPAff]] = λP.λe. P(e) & e'(EXPER(e') & Exp(e')(me)) : e'' (P(e'') →
Source(e'')(e'))
[[break]] = λx. λe. BREAK(e) & Thm(e)
[[VP]] = λe. BREAK(e) & Thm(e)(the vase)
[[VP1]] = λe. BREAK(e) & Thm(e)(the vase) & e'(EXPER(e') &
Exp(me)(e')) : e'' (BREAK(e'') & Thm(e'')(the vase) → Source(e'')(e'))
[[Voice]] = λx. λe. Agt(e)(x)
[[VoiceP]] = λe. BREAK(e) & Thm(e)(the vase) & Agt(e)(Gerald) &
e'(EXPER(e') & Exp(e')(me)): e''(BREAK(e'') & Thm(e'')(the vase) →
Source(e'')(e'))


  209

3) German Pertinence Dative (102)


a. Der Stein fällt dem Mann auf den Kopf.
the.Nom brick falls the.Dat man on the.Acc head
'The brick falls on the man's head.'
b. VoiceP
3
der Stein Voice'
3
Voice AffP
3
dem Mann Aff'
3
Aff VP
6
fällt auf den Kopf
c. (AffP and Voice combine by Event Identification.)
[[VP]] = λe. FALL(e) & Goal(e)(the head)
[[Aff]] = λP.λx.λe. P(e) & e'(EXPER(e') & Exp(e')(x)) : e'' (P(e'') →
Source(e'')(e'))
[[AffP]] = λe. FALL(e) & Goal(e)(the head) & e'(EXPER(e') & Exp(e')(the
man)) : e'' (FALL(e'') & Goal(e'')(the head) → Source(e'')(e'))
[[Voice]] = λx. λe. Agt(e)(x)
[[VoiceP]] = λe. FALL(e) & Goal(e) (the head) & Agt(e)(the stone) & e'
(EXPER(e') & EXP(e')(him)): e'' (FALL(e'') & Goal(e'')(the head) →
Source(e'')(e'))
d. Conversational Implicature: The man was affected because his head was hit.

 
210  

4) Japanese Affected Experiencer (148)


a. Sachi-ga Masa-ni Aiko-no kabin-o kowas-are-ta.
Sachi-Nom Masa-by Aiko-Gen vase-Acc break-Aff-Past
'Masa broke Aiko's vase on Sachie.'
b. AffP<v,t>
3
Sachi Aff'
3
VoiceP<v,t> Aff<vt,evt>
3
Masa Voice'
3
VP<v,t> Voice
6
Aiko-no kabin-o kowas-
c. (VP and Voice combine by Event Identification.)
[[VP]] = λe. BREAK(e) & Thm(e)(Aiko's vase)
[[Voice]] = λx. λe. Agt(e)(x)
[[VoiceP]] = λe. BREAK(e) & Thm(e)(Aiko's vase) & Agt(e)(Masa)
[[Aff]] = λP.λx.λe. P(e) & e'(EXPER(e') & Exp(e')(x)) : e'' (P(e'') →
Source(e'')(e'))
[[Aff']] = λx.λe. BREAK(e) & Thm(e)(Aiko's vase) & Agt(e)(Masa) &
e'(EXPER(e') & Exp(e')(x)) : e'' (P(e'') → Source(e'')(e'))
[[AffP]] = λe. BREAK(e) & Thm(e)(Aiko's vase) & Agt(e)(Masa) & e'.
(EXPER(e') & Exp(e')(Sachi)) : e''((BREAK(e'') & Thm(e'')(Aiko's vase)
& Agt(e'')(Masa)) → Source(e'')(e'))


  211

5) Hebrew Affected Experiencer (158)


a. hem kol ha-zma mitxatnim li
they all the-time marry to-me
'They are getting married on me all the time (and it bothers me).'
(Borer and Grodzinsky 1986, examples 9a)
b. AffP<v,t>
3
li Aff'
3
Aff<vt,evt> VoiceP<v,t>
3
hem Voice'
3
Voice VP<v,t>
6
mitxatnim
c. (VP and Voice combine by Event Identification.)
[[VP]= λe. MARRY(e)
[[Voice]] = λx. λe. Agt(e)(x)
[[Voice']] = λx. λe. MARRY(e) & Agt(e)(x)
[[VoiceP]] = λe. MARRY(e) & Agt(e)(they)
[[Aff]] = λPvt. λx. λe. P(e) : e'(EXPER(e') & Exp(e')(x)) & e''(P(e'') →
Source(e'')(e'))
[[Aff']] = λe. MARRY(e) & Agt(e)(they) : e'(EXPER(e') & Exp(e')(x)) &
e'' ((MARRY(e'') & Agt(e'')(they)) → Source(e'')(e'))
[[AffP]] = λe. MARRY(e) & Agt(e)(they) : e'(EXPER(e') & Exp(e')(me)) &
e'' ((MARRY(e'') & Agt(e'')(they)) → Source(e'')(e'))
(Bosse et al. 2012, slightly modified examples 87, 88)

 
212  

6) French Affected Experiencer (159)


a. Les gosses lui ont gribouillé sur tous les murs
the kids her.Dat have scribbled on all the walls
‘The kids scribbled all over the walls on her.’
(Authier and Reed 1992, example 3a)
b. VoiceP
3
les gosses Voice'
3
Voice AffP<v,t>
3
lui Aff'
3
Aff<vt,evt> VP
6
ont gribouillé sur tous les murs
c. (AffP and Voice combine by Event Identification.)
[[VP]] = λe. SCRIBBLE(e) & Thm(e)(the walls)
[[Aff]] = λPvt. λx. λe. P(e) : e'(EXPER(e') & Exp(e')(x)) & e''(P(e'') →
Source(e'')(e'))
[[Aff']] = λe. SCRIBBLE(e) & Thm(e)(the walls) : e' (EXPER(e') &
Exp(e')(x)) & e'' (P(e'') → Source(e'')(e'))
[[AffP]] = λe. SCRIBBLE(e) & Thm(e)(the walls) : e' (EXPER(e') &
Exp(e')(her)) & e'' (P(e'') → Source(e'')(e'))
[[Voice]] = λx. λe. Agt(e)(x)
[[Voice']] = λe. SCRIBBLE(e) & Thm(e)(the walls) & Agt(e)(x) : e' (EX-
PER(e') & Exp(e')(her)) & e'' (P(e'') → Source(e'')(e'))
[[VoiceP]] = λe. SCRIBBLE(e) & Thm(e)(the walls) & Agt(e)(the kids) : e'
(EXPER(e') & Exp(e')(her)) & e'' (P(e'') → Source(e'')(e'))


Appendix 2: Not-At-Issue Applicatives
1) German Ethical Dative (190)
a. Komm mir pünktlich nach Hause!
come me.Dat on.time to home
'Come home on time and I want this to happen!'
b.
3
DIRECTIVE AppEDP
3
ethical dative AppED'
3
AppED IRREALISP
3
110
IRREALIS IP
w
c. [[IP]] = λw. e in w. COME(e) & Goal(e)(home) & On time(e)
[[IRREALISP]] = w. e in w. COME(e) & Goal(e)(home) & On time (e)
[[AppED]] = λP. λx. P : x wants one world from the set P to come true
[[AppED']] = λx. w. e in w. COME(e) & Goal(e)(home) & On time (e) : x
wants one world from the set (w. e in w. COME(e) & Goal(e)(home) &
On time (e)) to come true
[[AppEDP]] = w. e in w. COME(e) & Goal(e)(home) & On time (e) : I want
one world from the set (w. e in w. COME(e) & Goal(e)(home) & On
time (e)) to come true


110 Han (1998) does not provide formal semantic denotations for [IRREALIS] and [DI-
RECTIVE].


214  

2) Subject Co-Referential Applicative (209)


a. John1 killed him1 a bear.
b. VoiceP
3
John1 Voice2'
3
1 Voice1'
3
t1 Voice'
3
Voice AppSCRP
3
a1 AppSCR'
3
AppSCR VP
6
kill a bear
c. [[VP]] = λe. KILL(e) & Thm(e)(a bear)
[[AppSCR]] = λPv,t. λx. λQe,vt. λy. λe. P(e) & Q(e)(y): Intense(e)(x) & x = y
[[AppSCR']] = λx. λQe,vt. λy. λe. KILL(e) & Thm(e)(a bear) & Q(e)(y): In-
tense(e)(x) & x = y
[[AppSCRP]]g = λQ. λy. λe. KILL(e) & Thm(e)(a bear) & Q(e)(y): In-
tense(e)(a1) & a1 = y
[[Voice]] = λx. λe. Agt(e)(x)
[[Voice']]g = λy. λe. KILL(e) & Thm(e)(a bear) & Agt(e)(y) : Intense(e)(a1) &
a1 = y
[[VoiceP]] = λe. KILL(e) & Thm(e)(a bear) & Agt(e)(John) : Intense(e)(John)
& John = John


Appendix 3: Benefactives
1) True Benefactive (244)
a. Jan öffnete der Frau die Tür.
Jan opened the woman the door
'Jan opened the door for the woman.'
b. VoiceP<v,t>
3
Jan Voice'<e,vt>
3
Voice<e,vt> BenP<v,t>
3
NP Ben'<e,vt>
der Frau 3
Ben<e,vt> VP<v,t>
3
öffnete NP
die Tür
c. (Ben and VP, Voice and BenP combine by Event Identification.)
[[öffnete]] = λx. λe. OPEN(e) & Thm(e)(x)
[[VP]] = λe. OPEN(e) & Thm(e)(the door)
[[Ben]] = λx. λe. Ben(e)(x)
[[Ben']] = λx. λe. OPEN(e) & Thm(e)(the door) & Ben(e)(x)
[[BenP]] = λe. OPEN(e) & Thm(e)(the door) & Ben(e)(the woman)
[[Voice]] = λx. λe. Agt(e)(x)
[[VoiceP]] = λe. OPEN(e) & Thm(e)(the door) & Ben(e)(the woman) &
Agt(e)(Jan)


216  

2) Recipient Benefactive (274)


a. Mary baked him a cake.
b. VoiceP
3
Mary Voice'
3
Voice vP1
3 2
v Voice him1 vP2
3 2
Rec v a cake v'
2 2
V Rec v RecP
2 2
Rec v t1 Rec'
2 2
V Rec Rec VP
2 6
V Rec bake a cake
c. (The verb is not interpreted in its base position but Rec has access to the direct
object. VP1 and Voice combine by Event Identification.)
[[bake]] = λx. λe. BAKE(e) & Thme(e)(x)
[[Rec]] = λx.λy.λe. HAVE(e) & Thm(e)(x) & Recipient(e)(y)
[[Rec']] = λy. λe. HAVE(e) & Thm(e)(a cake) & Recipient(e)(y)
[[RecP]] = λe. HAVE(e) & Thm(e)(a cake) & Recipient(e)(1)
[[v']] = λx. λe. BAKE(e) & Thm(e)(x) & (e') [HAVE(e') & Thm(e')(a cake)
& Recipient(e')(1) & CAUSE(e')(e)]
[[vP2]] = λe. BAKE(e) & Thm(e)(a cake) & (e') [HAVE(e') & Thm(e')(a
cake) & Recipient(e')(1) & CAUSE(e')(e)]
[[vP1]] = λe. BAKE(e) & Thm(e)(a cake) & (e') [HAVE(e') & Thm(e')(a
cake) & Recipient(e')(him) & CAUSE(e')(e)]
[[Voice]] = λx. λe. Agt(e)(x)
[[Voice']] = λx. λe. BAKE(e) & Thm(e)(a cake) & Agt(e)(x) & (e')
[HAVE(e') & Thm(e')(a cake) & Recipient(e')(him) & CAUSE(e')(e)]
[[VoiceP]] = λe. BAKE(e) & Thm(e)(a cake) & Agt(e)(Mary) & (e')
[HAVE(e') & Thm(e')(a cake) & Recipient(e')(him) & CAUSE(e')(e)]


  217

3) Prepositional Benefactive (287)


a. Jan öffnete die Tür für die Frau.
Jan.Nom opened the.Acc door for the.Acc woman
'Jan opened the door for the woman.'
b. VoiceP<v,t>
3
Jan Voice'
3
Voice VP<v,t>
3
PP VP1
3 6
P NP öffnete die Tür
für die Frau
c. (VP and Voice combine by Event Identification.)
[[P]] = λx. λPvt. λe. Ben(e)(x) & P(e)
[[PP]] = λPvt. λe. Ben(e)(the woman) & P(e)
[[VP1]] = λe. OPEN(e) & Thm(e)(the door)
[[VP]] = λe. Ben(e)(the woman) & OPEN(e) & Thm(e)(the door)
[[Voice]] = λx. λe. Agt(e)(x)
[[VoiceP]] = λe. Ben(e)(the woman) & OPEN(e) & Thm(e)(the door) &
Agt(e)(Jan)

 
218  

4) English ditransitive structure with additional for-PP (292)


a. Mary baked him a cake for his mother.
b. VoiceP
3
Mary Voice'
3
Voice vP3
3 3
v Voice vP1 PP
2 3 6
Rec v him1 vP2 for his mother
2 3
V Rec NP2 v'
a cake 3
v RecP
2 2
Rec v t1 Rec'
2 3
V Rec Rec VP
2 3
V Rec V NP2
bake a cake
c. (The verb is not interpreted in its base position but Rec has access to the direct
object. vP3 and Voice combine by Event Identification.)
[[bake]] = λx. λe. BAKE(e) & Thme(e)(x)
[[Rec]] = λx.λy.λe. HAVE(e) & Thm(e)(x) & Recipient(e)(y)
[[Rec']] = λy. λe. HAVE(e) & Thm(e)(a cake) & Recipient(e)(y)
[[RecP]] = λe. HAVE(e) & Thm(e)(a cake) & Recipient(e)(1)
[[v']] = λx. λe. BAKE(e) & Thm(e)(x) & (e') [HAVE(e') & Thm(e')(a cake)
& Recipient(e')(1) & CAUSE(e')(e)]
[[vP2]] = λe. BAKE(e) & Thm(e)(a cake) & (e') [HAVE(e') & Thm(e')(a
cake) & Recipient(e')(1) & CAUSE(e')(e)]
[[vP1]] = λe. BAKE(e) & Thm(e)(a cake) & (e') [HAVE(e') & Thm(e')(a
cake) & Recipient(e')(him) & CAUSE(e')(e)]
[[PP]] = λP. λe. Ben(e)(his mother) & P(e)
[[vP3]] = λe. BAKE(e) & Thm(e)(a cake) & Ben(e)(his mother) & (e')
[HAVE(e') & Thm(e')(a cake) & Recipient(e')(him) & CAUSE(e')(e)]
[[Voice]] = λx. λe. Agt(e)(x)


  219

[[Voice']] = λx. λe. BAKE(e) & Thm(e)(a cake) & Ben(e)(his mother) &
Agt(e)(x) & (e') [HAVE(e') & Thm(e')(a cake) & Recipient(e')(him) &
CAUSE(e')(e)]
[[VoiceP]] = λe. BAKE(e) & Thm(e)(a cake) & Ben(e)(his mother) &
Agt(e)(Mary) & (e') [HAVE(e') & Thm(e')(a cake) & Recipient(e')(him)
& CAUSE(e')(e)]

 

Appendix 4: Part-Whole Applicative
1) Part-Whole Applicative (348)
a. Jan zerbeulte dem Auto die Tür.
Jan dented the.Dat car the.Acc door
'Jan dented the door of the car.'
b. VoiceP
3
Jan Voice'
3
Voice AppPWP
3
dem Auto AppPW'
3
AppPW VP
3
zerbeulte die Tür
c. (AppPWP and Voice combine by Event Identification.)
[[zerbeulte]] = λx.λe. DENT(e) & Thm(e)(x)
[[VP]] does not compute due to functional composition
[[AppPW]] = λf<e,vt>. λx. λy. λe. f(e)(x) & xy for the duration of e
[[AppPW']] = λy. λe. DENT(e) & Thm(e)(the door)& the door y for the dura-
tion of e
[[AppPWP]] = λe. DENT(e) & Thm(e)(the door)& the door  the car for the
duration of e
[[Voice]] = λx. λe. Agt(e)(x)
[[Voice']] = λx. λe. DENT(e) & Thm(e)(the door) & Agt(e)(x) & the door 
the car for the duration of e
[[VoiceP]] = λe. DENT(e) & Thm(e)(the door) & Agt(e)(Jan) & the door 
the car for the duration of e


222  

2) Part-Whole applicative with prepositions (351)


a. Jan legte dem Auto etwas in den Kofferraum.
Jan laid the.Dat car something.Acc in the trunk
'Jan put something into the trunk of the car.'
b. VoiceP
3
Jan Voice'
3
Voice AppPWP
3
dem Auto AppPW'
3
AppPW VP1
3
VP PP
3 3
legte<e,evt> etwas in den Kofferraum
c. (AppPWP and Voice combine by Event Identification.)
[[VP]] = λx. λe. PUT(e) & Thm(e)(something) & Goal(e)(x)
[[PP]] = the trunk
[[VP1]] does not compute due to funcational composition
[[AppPW]] = λf<e,vt>. λx. λy. λe. f(e)(x) & xy for the duration of e
[[AppPW']] = λy. λe. PUT(e) & Thm(e)(something) & Goal(e)(the trunk) & the
trunk  y for the duration of e
[[AppPWP]] = λe. PUT(e) & Thm(e)(something) & Goal(e)(the trunk) & the
trunk  the car for the duration of e
[[Voice]] = λx. λe. Agt(e)(x)
[[VoiceP]] = λe. PUT(e) & Thm(e)(something) & Goal(e)(the trunk) &
Agt(e)(Jan) & the trunk  the car for the duration of e


  223

3) Part-Whole Applicative with unergatives (353)


a. Dem Baum haben die Blätter geleuchtet.
the.Dat tree have the.Nom leaves glowed
'The leaves of the tree were glowing.'
b. VoiceP2
3
dem Baum2 VoiceP1
3
die Blätter Voice'
3
Voice2 AppPWP
3 3
AppPW1 Voice1 t2 AppPW'
3 3
AppPW V t1 VP
6
leuchten
c. (V head-moves through AppPW to Voice. V and Voice combine by Event
Identification.)
[[leuchten]] = λe. GLOW(e)
[[V + Voice]] = λe. GLOW(e) & Agt(e)(x)
[[AppPW]] = λf<e,vt>. λx. λy. λe. f(e)(x) & xy for the duration of e
[[Voice2]] = λx. λy. λe. GLOW(e) & Agt(e)(x) & xy for the duration of e
[[VoiceP]] = λy. λe. GLOW(e) & Agt(e)(the leaves) & the leaves y for the
duration of e
[[VoiceP2]] = λe. GLOW(e) & Agt(e)(the leaves) & the leaves  the tree for
the duration of e

 

Appendix 5: Dative of Inaction
Dative of Inaction (374)
a. Der Oma zerbrach eine Vase.
the.Dat grandma broke a.Nom vase
'The vase broke on grandma.'
(Hens 1997, example 2)
b.
3
Voice[-agentive] RespP
3
der Oma Resp'
3
Resp VP
3
zerbrach eine Vase
c. (Resp and VP combine by Event Identification.)
[[zerbrach]] = λx. λe. BREAK(e) & Thm(e)(x)
[[VP]] = λe. BREAK(e) & Thm(e)(a vase)
[[Resp]] = λx. λe. Responsible(e)(x)
[[Resp']] = λx. λe. BREAK(e) & Thm(e)(a vase) & Responsible(e)(x)
[[RespP]] = λe. BREAK(e) & Thm(e)(a vase) & Responsible(e)(grandma)



Glossary
Affected Experiencer: a type of applicative argument that is interpreted as psy-
chologically affected; the referent must be sentient
Affectedness: psychological or mental impact on a person
Animacy requirement: the referent must be animate
Applicative Argument: a (nominal) constituent that is not entailed by the sen-
tence and not required for the sentence to be grammatical
At-issue meaning: meaning that is asserted
Benefactive: a nominal constituent that denotes who is benefitting from an
event; can be an applicative argument or a selected argument
Bi-eventive: a denotation that includes two event variables (of semantic type v)
Condition B: a binding condition that requires pronouns to be free (not bound)
in their binding domain
Condition C: a binding condition that states that referential expressions, such as
proper names, cannot be bound
Conventional Implicature (CI): according to Potts (2005), elements that con-
tribute only not-at-issue meaning, are added to the common ground, are made
by the speaker, and are independent of the at-issue meaning of the sentence
Conversational Implicature: meaning that is conveyed by an utterance without
being explicitly stated; the meaning can be inferred by the conversational
partners
Dative of Inaction (DI): a type of (German) applicative argument that an event
happened while the referent of the DI should have acted to prevent the event
Deputative Benefactive: a benefactive that denotes someone did something in-
stead of another person (who was supposed to do it); contrasts with plain and
recipient benefactives
de re-interpretation: a pronoun must be interpreted as referring to a specific
(identical) entity
Ethical dative: a type of applicative argument that expresses a speaker's attitude
towards a proposition; in German, the speaker wants the proposition to come
true
Event Identification: a semantic way of combing two functions, each with an
event variable, into one function with one event variable
Family of Sentence Test: applying syntactic transformations to a declarative
sentence to see which meaning of the declarative is maintained; used here to
distinguish at-issue meaning from not-at-issue meaning


228  

Functional Composition: a semantic computation that allows the semantic


combination of two elements to be delayed if the derivation can be salvaged
due to that delay
High applicative: according to Pylkkänen (2002, 2008), a type of applicative
argument that attaches outside of VP and below Voice; it relates an individu-
al to an event
Low applicative: according to Pylkkänen (2002, 2008), a type of applicative
argument that attaches inside of VP and relates two individuals to each other;
it denotes a change of possession
Material Part-Whole Relation: a relationship between two entities where one
needs to be a material part of the other
Not-at-issue meaning: meaning that is not directly stated in an utterance but
associated with one or more words of the utterance; can be distinguished
from at-issue meaning by using the family of sentence tests
Part-Whole Applicative: a type of (German) applicative argument that denotes
a whole of which a material part is undergoing an event
Pertinence dative: a type of (German) applicative argument that denotes an in-
dividual whose inalienable part (or piece of clothing) is involved in an event;
here analyzed as an affected experiencers applicative argument with an addi-
tional conversational implicature
Plain Benefactive: a type of benefactive that denotes a general benefit; contrasts
with deputative benefactive and recipient benefactive
PP-Shell: a PP dominating a DP/NP with an empty (or unpronounced) preposi-
tion
Recipient Benefactive: a type of benefactive that denotes someone benefitting
because they receive something (or are the intended recipient of something);
contrasts with deputative and plain benefactive
SE-anaphor: a reflexive pronoun that is morphologically simplex; contrasts
with Self-anaphor in behavior and distribution
Self-anaphor: a reflexive pronoun that is morphologically complex; contrasts
with SE-anaphors in behavior and distribution
Sloppy identity: occurs in sentences with ellipsis when an elided pronoun is
interpreted as not identical to the non-elided pronoun; contrasts with strict
identity
Strict identity: occurs in sentences with ellipsis when an elided pronoun is in-
terpreted as identical to the non-elided pronoun; contrasts with sloppy identi-
ty
Strong Pronoun: a pronoun that can be stressed, can be modified and can be
referential; contrasts with weak pronoun


  229

Subject co-referential applicative: a type of (German and English) applicative


argument that is necessarily identical in reference to the subject of the sen-
tence
True Benefactive: a type of (German) applicative argument that is interpreted
as a beneficiary
Weak Pronoun: a pronoun that cannot be modified, is either co-referential or
expletive, and cannot be stressed; contrasts with strong pronoun

 

Bibliography
Abraham, W. (1973). The Ethic Dative in German. In Kiefer, F. and Ruwet, N.,
editors, Generative Grammar in Europe, pages 1–19. D. Reidel Publish-
ing Company, Dordrecht.
Abraham, W. (2006). Datives: structural vs. inherent - abstract vs. morphologi-
cal - autonomous vs. combinatory - universally vs. language-specifically
configured. In Hole, D., Meinunger, A., and Abraham, W., editors, Da-
tives and other cases, pages 3 – 46. John Benjamins Publishing Company.
Authier, J.-M. and Reed, L. (1992). Case theory, theta theory, and the distribu-
tion of French affected datives. In Bates, D., editor, The Proceedings of
the Tenth West Coast Conference on Formal Linguistics, pages 27–40,
Stanford, California. Center for the Study of Language and Information.
Bale, A. (2007). Quantifiers and Verb Phrases: An exploration of propositional
complexity. Natural Language and Linguistic Theory, 25:447–483.
Barss, A. and Lasnik, H. (1986). A note on anaphora and double objects. Lin-
guistic Inquiry, 17:347–354.
Beck, S. and Johnson, K. (2004). Double objects again. Linguistic Inquiry,
35:97–123.
Beck, S. and Snyder, W. (2001). Complex predicates and goal PPs: Evidence for
a semantic parameter. In Do, A., Dominguez, L., and Aimee, J., editors,
BUCLD 25: Proceedings of the 25th Annual Boston University Confer-
ence on Language Development, volume 1, pages 114–122. Cascadilla
Press.
Behagel, O. (1932). Deutsche Syntax, volume 4. Heidelberg.
Berman, R. (1982). Dative marking of the affectee role: Data from Modern He-
brew. Hebrew Annual Review, 6:35–59.
Bierwisch, M. (1966). Eine Hierarchie syntaktisch-semantischer Merkmale. Syn-
taktische Studien, pages 29–86.
Borer, H. and Grodzinsky, Y. (1986). Syntactic cliticization and lexical cliticiza-
tion: The case of Hebrew dative clitics. Syntax and Semantics, 19:175–
217.
Bosse, S. (2012). The German Dative of Inaction. Tampa Papers in Linguistics
3:39–51.
Bosse, S. (2013a). An argument against PP-shells for benefactive DPs in Ger-
man. Tampa Papers in Linguistics 4:1–10.
Bosse, S. (2013b). A formal semantic approach to (Appalachian) personal da-
tives. accepted Southern Journal of Linguistics.


232  

Bosse, S. (2013c). On the Absence of Causation in the German Dative of Inac-


tion. Interdisciplinary Journal of Germanic Linguistics and Semiotic
Analyses 18(2):207–232.
Bosse, S. and Bruening, B. (2011). Benefactive versus experiencer Datives. In
Washburn, M. B., editor, Proceedings of the 28th West Coast Conference
on Formal Linguistics, pages 66–77. Cascadilla Press.
Bosse, S., Bruening, B., and Yamada, M. (2012). Affected experiencers. Natural
Language and Linguistic Theory, 30(4):1185–1230.
Bresnan, J. and Moshi, L. (1990). Object asymmetries in comparative Bantu
syntax. Linguistic Inquiry, 22(2):147–186.
Brisson, C. (1998). Distributivity, Maximality, and Floating Quantifiers. PhD
thesis, Rutgers University.
Bruening, B. (2010). Ditransitive asymmetries and a theory of idiom formation.
Linguistic Inquiry, 41:519–562.
Bruening, B. (2011). Transitivity restrictions in Passamaquoddy and other lan-
guages. Presentation at WSCLA.
Cardinaletti, A. and Starke, M. (1994). The typology of structural deficiency: On
the three grammatical classes. Working Papers in Linguistics University
of Venice, 4(2):41–109.
Cardinaletti, A. and Starke, M. (1996). Deficient pronouns: A view from Ger-
manic. In Thrainsson, H., Epstein, S. D., and Peter, S., editors, Studies in
Comparative Germanic Syntax 2, pages 21–65. Kluwer Academic Pub-
lishers.
Cardinaletti, A. and Starke, M. (1999). The typology of structural deficiency: A
case study of the three classes of pronouns. In van Riemsdijk, H., editor,
Clitics in the Languages of Europe, pages 145–223. Mouton de Gruyter,
Berlin.
Chierchia, G. and McConnell-Ginet, S. (1990). Meaning and Grammar: An In-
troduction to Semantics. MIT Press.
Christian, D. (1991). The personal dative in Appalachian speech. In Trudgill, P.
and Chambers, J., editors, Dialects of English: Studies in Grammatical
Variation, pages 11–19. Longman, London.
Cirtila, B. (2006). Zum adnominalen possessiven Dative im Deutschen. Analele
Universitatii din Craiova Linguistics, 28:73–80.
Colleman, T. (2010). The benefactive semantic potential of 'caused reception'
constructions. In Zuniga, F. and Kittilä, S., editors, Benefactives and Mal-
efactives. Typological perspectives and case studies, pages 219–243. John
Benjamins Publishing Company.
Collins, C., Moody, S., and Postal, P. M. (2008). An AAE camouflage construc-
tion. Language, 84(1):29–68.


  233

Conroy, A. M. (2007). The personal dative in Appalachian English as a reflexive


pronoun. In Omaki, A., Ortega-Santos, I., Sprouse, J., and Wagers, M.,
editors, University of Maryland Working Papers in Linguistics, volume
16, pages 63–88. UMWPiL, College Park.
Cooper, K. (1994). Topics in Zurich German Syntax. PhD thesis, University of
Edinburgh.
Cuervo, M. C. (2003). Datives at Large. PhD thesis, MIT.
Dubinsky, S. (1997). Predicate union and the syntax of Japanese passives. Jour-
nal of Linguistics, 33:1–37.
Eckardt, R. (2002). Event semantics. Linguistische Berichte, 10:91–127.
Frey, W. (2000). A medial topic position for German. Linguistische Berichte,
pages 1–40.
Green, G. M. (1974). Semantics and Syntactic Regularity. Indiana University
Press, Bloomington, Indiana.
Grice, H. P. (1967). Logic and conversation. In Studies in the Way of Words,
pages 1–138. Harvard University Press.
Gutzmann, D. (2007). Eine Implikatur konventioneller Art: Der Dativus Ethicus.
Linguistische Berichte, 211:277–308.
Haddad, Y. A. (2011). The syntax of Southern American English personal da-
tives: an anti-localtity account. Canadian Journal of Linguistics 56(3):
403–412.
Halevy, R. (2007). The subject co-referential l-pronoun in Hebrew. In Bar, T.
and Cohen, E., editors, Studies in Semitic and General Linguistics in
Honor of Gideon Goldenberg, pages 299–321. Ugarit-Verlag, Munster.
Han, C.-h. (1998). The Structure and Interpretation of Imperatives: Mood and
Force in Universal Grammar. PhD thesis, University of Pennsylvania.
Harley, H. (2008). On the causative construction. In Miyagawa, S. and Saito,
M., editors, The Oxford Handbook of Japanese Linguistics, pages 20–53.
Oxford University Press.
Heim, I. and Kratzer, A. (1998). Semantics in generative grammar. Blackwell,
Malden, MA.
Hens, G. (1995). Ditransitive Constructions in German. PhD thesis, University
of California at Berkeley.
Hens, G. (1997). Constructional semantics in German: The dative of inaction.
American Journal of Germanic Linguistics and Literatures, 9:191–219.
Hole, D. (2005). Reconciling "possessor" datives and "beneficiary" datives - to-
wards a unified voice account of dative binding in German. In Maienborn,
C. and Wollenstein-Leisten, A., editors, Event Arguments in Syntax, Se-
mantics, and Discourse, pages 213–242. Niemeyer, Tubingen.

 
234  

Hole, D. (2006). Extra argumentality - affectees, landmarks, and voice. Linguis-


tics, 44(2):383–424.
Hole, D. (2008). Dativ, Bindung und Diathese. Habilitationsschrift Humboldt
Universitat Berlin.
Hole, D. (201X). German free dative and Knight Move binding.
http://www.ilg.uni-stuttgart.de/mitarbeiter/hole/BambergV4.pdf. Ac-
cessed July 21, 2010.
Hopper, P. and Thompson, S. (1980). Transitivity in grammar and discourse.
Language, 56(2):251–299.
Horn, L. (2008). "I love me some him": The landscape of non-argument datives.
In Bonami, O. and Hofherr, P., editors, Empirical issues in Syntax and
Semantics 7, pages 169–192. CSSP 7.
Horn, L. (2009). I love me some datives: Pronouns, expressive meaning, and f-
implicature. Handout from DGfS 31.
Hutchinson, C. and Armstrong, G. (2013). Grammatical diversity in North
American English. to appear.
Isacenko, A. V. (1965). Das syntaktische Verhaltnis der Beziehungen von
Korperteilen im Deutschen. Studia Grammatica, pages 7–27.
Jouitteau, M. and Rezac, M. (2008). The French ethical dative, 13 syntactic
tests. Bucharest Working Papers in Linguistics, XI:97–108.
Karttunen, L. and Peters, S. (1979). Conventional implicature. In Dinneen, D. A.
and Oh, C.-K., editors, Syntax and Semantics, volume 11, pages 1–56.
Academic Press, New York.
Kasai, H. (2008). Linearizing rightward movement. In Chang, C. and Haynie,
H., editors, Proceedings of WCCFL 26. Cascadilla Proceedings Project.
Kittilä, S. (2005). Recipient-prominence vs. beneficiary-prominence. Linguistic
Typology, 9:269–297.
Kittilä, S. and Zuniga, F. (2010). Introduction. Benefaction and malefaction
from a cross-linguistic perspective. In Zuniga, F. and Kittilä, S., editors,
Benefactives and Malefactives. Typological perspectives and case studies,
pages 1–28. John Benjamins Publishing Company.
Kratzer, A. (1996). Severing the external argument from its verb. In Rooryck, J.
and Zaring, L., editors, Phrase Structure and the Lexicon, pages 109–137.
Kluwer Academic Publishers, Dordrecht, The Netherlands.
Kubo, M. (1992). Japanese passives. In Institute of Language and Culture Stud-
ies Working Papers 23. Hokkaido University, Sapporo.
Lamiroy, B. and Delbecque, N. (1998). The possessive dative in Romance and
Germanic languages. In van Langendonck, W. and von Belle, W., editors,
The Dative, pages 29–74. John Benjamins Publishing Company.
Landau, I. (1999). Possessor raising and the structure of VP. Lingua, 107:1–37.


  235

Langacker, R. W. (1991). The foundations of Cognitive Grammar, volume 2:


Descriptive Application. Stanford University Press.
Larson, R. (1988). On the double object construction. Linguistic Inquiry,
19(3):335–392.
Larson, R. (2010). On Pylkkänen's semantics for low applicatives. Linguistic
Inquiry, 41(4):701–704.
Lee-Schoenfeld, V. (2006). German possessor datives: raised and affected.
Journal of Comparative Germanic Linguistics, 9:101–142.
Lenerz, J. (1977). Zur Abfolge nominale Satzglieder im Deutschen. Narr.
Lüdeling, A. (2001). On Particle Verbs and Similar Constructions in German.
CSLI Publications, Stanford, California.
McCready, E. (2009). Varieties of conventional implicature: Evidence from
Japanese.
http://www.semanticsarchive.net/Archive/DdmMWQyM/adverbs.pdf.
Accessed July 21, 2010.
McGinnis, M. (2001). Phases and the syntax of applicatives. In Kim, M. and
Strauss, U., editors, North East Linguistic Society, pages 333–349,
Georgetown University. GLSA.
McLachlan, L. R. (2011). I love me some Jimmy Glick: The semantic contribu-
tion of 'some' in personal dative constructions. Presentation at LSA.
Müller, G. (1995). A-bar syntax: a study in movement types, chapter Dative
Movement (Free Datives), pages 246 – 252. Mouton de Gruyter.
Neumann, D. (1995). The dative and the grammar of body parts in German.
Empirical Approaches to Language Typology, 14:745–779.
O'Connor, M. C. (2007). External possession and utterance interpretation: a
crosslinguistic exploration. Linguistics, 45(3):577–613.
Oehrle, R. (1976). The grammatical status of the English dative alternation.
PhD thesis, Massachusetts Institute of Technology.
Parsons, T. (1990). Events in the semantics of English: A study in subatomic
semantics. In Current Studies in Linguistics Series no. 19. MIT Press,
Cambridge, Massachusetts.
Portner, P. (2007). Imperatives and modals. Natural Language Semantics,
15(4):351–383.
Potts, C. (2005). The Logic of Conventional Implicatures. Oxford University
Press, Oxford.
Pylkkänen, L. (2002). Introducing Arguments. PhD thesis, Massachusetts Insti-
tute of Technology.
Pylkkänen, L. (2008). Introducing Arguments. MIT Press, Cambridge.
Ramchand, G. C. (2008). Verb Meaning and the lexicon: a first phase syntax.
Cambridge University Press.

 
236  

Rezac, M. (2008). Phi-Agree and Theta-related Case. In Harbour, D., Adger, D.,
and Bejar, S., editors, Phi Theory, pages 83–129. Oxford University Press,
Oxford.
Rizzi, L. (1986). On chain formation. In Borer, H., editor, The Grammar of
Pronominal Clitics (19). Academic Press, New York, New York.
Roberge, Y. and Troberg, M. (2007). Thematic indirect objects in French.
French Language Studies, 17:297–322.
Roberts, C., Simons, M., Beaver, D., and Tonhauser, J. (2009). Presupposition,
conventional implicature, and beyond: A unified account. web-
space.utexas.edu/dib97/RobertsetalESSLLI.pdf. Accessed May 3, 2014.
Rosengren, I. (1975). Ein freier Dativ. In Germanistische Streifzuge. Festschrift
fur Gustav Korlen, pages 209–221. Almqvist & Wiksell International.
Scherpenisse, W. (1985). The final field in German: Extraposition and frozen
position. gagl.eldoc.ub.rug.nl/FILES/root/1985-26/03/GAGL-26-1985-
03.pdf. Accessed May 4, 2014.
Schmidtke-Bode, K. (2009). The role of benefactives and related notions in the
typology of purpose clauses. In: Zuniga, F. and Seppo Kittilä (eds.): Bene-
factives and Malefactives: Typological perspectives and case studies .
pages 121–146.
Schöffer, G. (1992). Semantische Funktionen des deutschen Datives: Vorschlag
einer Alternative zur Diskussion um den homogenen/heterogenen Dativ
der deutschen Gegenwartssprache. Nodus-Publ.
Shibatani, M. (1994). Benefactive constructions: A Japanese-Korean compara-
tive perspective. In Akatsuka, N., editor, Japanese/Korean Linguistics (4),
pages 39–74. Stanford Linguistics Association, Stanford.
Shibatani, M. (1996). Applicatives and benefactives: a cognitive account. In
Shibatani, M. and Thompson, S., editors, Grammatical Constructions:
Their Form and Meaning, pages 157–194. Oxford University Press.
Sim, C.-Y. (2005). To Make a Case for All: Syntactic Structure, Semantic Inter-
pretation, and Case Morphology. PhD thesis, University of Delaware.
Smith, T. Y. (2010). Cross-linguistic categorization of benefactives by event
structure. In Zuniga, F. and Kittilä, S., editors, Benefactives and Malefac-
tives. Typological perspectives and case studies, pages 71–95. John Ben-
jamins Publishing Company.
Soames, S. (1982). How presuppositions are inherited: A solution to the projec-
tion problem. Linguistic Inquiry, 13:483–545.
Sorace, A. (2000). Gradients in auxiliary. Language, 76(4):859–890.
Sroda, M. S. and Mishoe, M. (1995). "i jus like to look me at some goats": dia-
lectal pronominals in Southern English. Handout for NWAV 24 presenta-
tion.


  237

Steinbach, M. (1998). Middles in German. PhD thesis, Humboldt-Universitat zu


Berlin.
Storoshenko, D. R. (2012). Baby I love me some chowder. Handout from 2012
Meeting of the Canadian Lingusitics Association.
Tomioka, S. and Sim, C.-Y. (2007). The event semantic root of inalienable pos-
session. Manuscript.
Van Valin, R. D. and LaPolla, R. (1997). Syntax. Structure, Meaning and Func-
tion. Cambridge University Press.
Vogel, R. and Steinbach, M. (1998). The dative - an oblique case. Linguistische
Berichte, pages 65–90.
von Polenz, P. (1969). Der Pertinenzdative und seine Satzbauplane. In Engel,
U., Grebe, P., and Rupp, H., editors, Festschrift für Hugo Moser, pages
146 – 171. Schwann.
von Stechow, A. (1996). The different readings of wieder "again": A structural
account. Journal of Semantics, 13:87–138.
Webelhuth, G. and Dannenberg, C. J. (2006). Southern American English per-
sonal datives: The theoretical significance of dialectal variation. American
Speech, 81:31–55.
Wegener, H. (1983). Der 'freie' Dativ im Deutschen und Französischen. Jahr-
buch Deutsch als Fremdsprache, 9:147–163.
Wegener, H. (1989). Eine Modalpartikel besonderer Art: Der Dativus Ethicus.
In Weydt, H., editor, Sprechen mit Partikeln, pages 56–73. Walter de
Gruyter, Berlin.
Wegener, H. (1991). Der Dativ - ein struktureller Kasus? In Fanselow, G. and
Felix, S., editors, Strukturen und Merkmale syntaktischer Kategorien,
pages 70–103. Narr, Tubingen.
Woolford, E. (2006). Lexical case, inherent case, and argument structure. Lin-
guistic Inquiry, 37(1):111–130.
Zanuttini, R. (2008). Encoding the addressee in the syntax: evidence from Eng-
lish imperative subjects. Natural Language and Linguistic Theory,
26:185–218.
Zuniga, F. (2007). Towards a typology of benefactives and malefactives.
www.uzh.ch/spw/aboutus/zuniga/downloads/Position paper 2.pdf. Ac-
cessed July 21, 2010.

 

Index
vs. pertinence datives, 30, 33–35,
Affected experiencers, 11, 15, 29–88 38–40
adverbial modification test of, 45 on-PP variant of, 30–31, 43–44,
again-modification of, 46–48, 77, 47
85–86 quantifier binding test of, 38
agent and, 75–76 vs. recipient benefactive, 30
analysis of, 40–44, 81–86, 207–12 resultative VP and, 68, 69–71
atelic motion verbs and, 69 semantic licensing of, 67–71
at-issue meaning and, 37–38 sentience requirement of, 30–32
vs. benefactives, 31–32, 119–20, subject c-command of, 41, 42, 76
127 vs. subject co-referential applica-
bi-eventivity and, 44–46 tives, 35–36
clausal modifier test of, 45 transitive verbs and, 65–67
Condition C violation of, 41 verbal restrictions on, 30, 65–71
continuation of condition verbs verbs of physical affectedness
and, 68–69 and, 67
Conventional Implicatures and, verbs of psychological affected-
72–73 ness and, 67–68
co-occurrences of, 189, 194–95, wh-question test of, 37, 74, 180
200–201 yes/no-question test of, 36–37, 75,
vs. Dative of Inaction, 36, 182 79–80
directional PP and, 69 See also Pertinence datives
vs. ethical datives, 35 Aff head, 41–44
existence of state verbs and, 69 attachment height variation in,
in French, 78–86, 212 73–78, 82–83
in Hebrew, 78–86, 116, 211 counterfactual denotation of, 49
as high applicative, 29–30 in French, 83–85, 87
if-clause test of, 37–38, 74, 80–81 in Hebrew, 83, 84–85, 87
intransitive verbs and, 65–67 in Japanese, 73–78, 87
in Japanese, 73–78, 210 parametric variation in, 73–87
location modifiers for, 45–46 on-PP variant of, 43–44
locative PPs and, 69 again-modification, 25
negation test of, 36, 48–49, 74– of affected experiencers, 46–48
75, 79 of benefactives, 128–29
not-at-issue meaning and, 36–37, in French, 84–85
48–49, 72, 73, 78–86 in Hebrew, 85–86
particle verbs and, 69–70 in Japanese, 77
vs. part-whole applicatives, 32– of part-whole applicatives, 171–
35, 151, 157 72, 175


240  

of subject co-referential applica- deputative, 120, 121, 122–24


tives, 111, 114 vs. ditransitives, 121, 129, 135,
Agent, 23–24, 43 136–37
in Japanese, 75–76 extraposition and, 146–47
Animacy requirement, 34, 35, 103–4 für-PP, 123, 131, 140–47, 196–98,
App head, 98–101 199
AppED, 98–100 as high applicative, 125–26
AppPW, 170–74 if-clause test of, 125
AppSCR, 108–9 intransitive verbs and, 129–30
Applicative arguments negation test of, 125
co-occurrence of, 188–202 vs. part-whole applicatives, 120,
definition of, 13–17, 205 127–28, 151, 157
family of sentence tests for, 12, passive and, 126–28
20–22 plain, 120, 121, 122–24
recursivity of, 188–89 for-PP, 121, 140–47, 218–19
See also specific types PP-shell analysis of, 144–47
Appositives, 20, 21–22 prepositional, 140–47
At-issue meaning, 12, 203 sloppy reading requirement for,
affected experiencers and, 37–38 133–34
Dative of Inaction and, 182–83 vs. subject co-referential applica-
ethical dative and, 97 tives, 104, 127
family of sentence tests for, 21–22 transitive verbs and, 130–31
part-whole applicative and, 158– true, 11, 14–15, 123–24, 125–35,
61 144–47
pertinence dative and, 40 variable binding and, 136–37
weak pronouns and, 116 verbal restrictions on, 129–31
verbal type and, 123–24
bekommen, 128, 139 wh-question test of, 124
Benefactives, 94, 119–47 yes/no-question test of, 124
vs. affected experiencers, 31–32, See also Recipient benefactives
119–20, 127 Ben head, 125–26, 135
as affectedness plus purposivity, Bi-eventivity, 44–46
131–34
again-modification of, 128–29, Causative(s), 25
146 adversity, 183–84
analysis of, 115–17, 125–26, 131– let, 182n, 184
34, 137–38, 144–47 C-command test
at-issue meaning and, 124–25 of affected experiencers, 41, 42,
co-occurrences of, 189, 195–98, 76
199, 201 in French, 82
vs. Dative of Inaction, 120, 127, in Hebrew, 82
182 in Japanese, 76–77


  241

of part-whole applicatives, 157– vs. part-whole applicatives, 151,


58 157
Chichewa, 17 passivization and, 185
Cleft sentence, 20–21 sentience and, 181–82
Cognate arguments, 15 telic motion verbs and, 180
Conventional Implicatures, 19–20, transitive verbs and, 179
72–73, 89, 96–98, 117 verbal restrictions on, 179–81
Conversational implicature of pos- wh-question test of, 183
session, 53–55 yes/no-question test of, 182
Co-occurrence of applicatives, 188– denn, 92, 93
202 Discourse particle, 92, 93
in English, 200–201 Ditransitives, 135, 136–37, 139, 140
in German, 188–99 Double object construction, 114
Coordination
ethical dative and, 92 Entailment omission test for free da-
pertinence dative and, 64–65 tives, 13–14
weak pronouns and, 116 Ethical dative, 11, 15, 89–101
vs. affected experiencers, 35
Dative analysis of, 96–101, 213
free, 13–14 appositive modification and, 93–
judgment, 16 94
See also Ethical dative at-issue meaning independence of,
Dative of Inaction, 11, 15, 178–87 97
vs. affected experiencers, 36, 182 binding and, 94
again-modification and, 185–86 cancelability and, 97
analysis of, 183–85, 225 contrast and, 93
at-issue meaning and, 182–83 Conventional Implicature features
vs. benefactives, 120, 127, 182 of, 89, 96–98, 117
change of state verbs and, 180 co-occurrences of, 189, 190–91,
continuation of condition verbs 199
and, 180 coordination and, 92
co-occurrences of, 189, 192–93 [DIRECTIVE] feature of, 98–101
as high applicative, 178, 183, vs. discourse particle, 92, 93
184–85 distribution of, 91–95, 100
if-clause test of, 183 event clause and, 94–95
inchoatives and, 180 exclamation, 90
intransitive verbs and, 179, 180 if-clause test of, 95
vs. Japanese adversity causative, [IRREALIS] feature of, 98–101
183–84 negation test of, 96
negation test of, 182–83 not-at-issue meaning and, 91, 95–
negative consequence and, 181 96, 117–18
as not-at-issue pronoun, 96–98

 
242  

optionality of, 90 of Dative of Inaction, 183


vs. part-whole applicatives, 150 of ethical dative, 95
pronoun negation and, 93 of part-whole applicatives, 158–
relative clause and, 93 59
speaker orientation of, 97 of pertinence datives, 40
tense and, 100 of PP-alternative benfactive, 141–
word order rules for, 92–93 42
Event semantics, 23–26 of prepositional benefactives,
Extraposition, 146–47 141–42
of subject co-referential applica-
Family of sentence tests, 12, 20–22 tives, 108
See also specific tests Implicit arguments, 15
for-PP, 121, 140–47, 201, 218–19 Instrumental applicatives, 17
Free dative, 13–14
French Japanese, 73–78, 204
affected experiencers in, 78–86, adversity causative of, 183–84
212 affected experiencers in, 73–78,
Aff head in, 83–85, 87 210
again-modification in, 84–85 Aff head in, 73–78, 87
c-command test in, 82 again-modification in, 77
für-PP, 123, 131, 140–47, 196–98, Agent in, 75–76
199 causative -sase- of, 183–84
c-command test in, 76–77
Genitive construction, 152–53 -rare morpheme of, 77–78
sekkaku adverb in, 22
Hebrew Judgment dative, 16
affected experiencers in, 78–86,
116, 211 Korean, 204
Aff head in, 83, 84–85, 87 inalienable possessor construction
again-modification in, 85–86 in, 167–70, 171–72
c-command test in, 82 kriegen, 127–28, 139
High applicatives, 12, 26–28, 203
affected experiences as, 29–30 Landmark datives, 163–67
Dative of Inaction as, 178 let-causative, 182n, 184
part-whole applicative as, 149, Locative applicatives, 17
161–63 Low applicatives, 12, 26–28, 203
verb phrase and, 27–28 subject co-referential applicatives
as, 113–15
If-clause test, 21–22 verb phrase and, 27–28
of affected experiencers, 37–38,
74, 80–81 mir, 13, 61–62, 90
of benefactives, 125 See also Ethical dative


  243

Modifiers, 24–25 analysis of, 161–76, 221–23


See also again-modification animate referents of, 34, 35
at-issue meaning and, 158–61
Negation test, 21 vs. benefactives, 120, 127–28,
of affected experiencers, 36, 48– 151, 157
49, 74–75, 79 binding analysis of, 163–67
of benefactives, 125 binding test of, 160–61
of Dative of Inaction, 182–83 cancelability and, 153
of ethical dative, 93, 96 c-command and, 157–58
of part-whole applicatives, 159– co-occurrences of, 189, 198–99
60, 175 vs. Dative of Inaction, 151, 157
of pertinence datives, 39 directional PP and, 156
of PP-alternative benefactive, emotion verbs and, 156
140–41 vs. ethical dative, 150
of prepositional benefactives, 141 vs. genitive construction, 152–53
of subject co-referential applica- genitive modifier of, 175–76
tives, 107 grammaticality of, 150–51, 152
Not-at-issue meaning, 12, 19–22, as high applicative, 149, 161–63
203 if-clause test of, 158–59
affected experiencer and, 36–37, kinship nouns and, 152
72, 73, 78–86 in Korean, 167–70, 171–72
Conventional Implicatures and, midfield scrambling and, 156–57,
19–20 167, 176
ethical dative and, 91, 95–96, negation test of, 159–60, 175
117–18 noncancelability of, 153
family of sentence tests for, 21–22 non-resultative verbs and, 156
pertinence dative and, 38–40 part NP of, 154–55
subject co-referential applicative perception verbs and, 156
and, 102, 106–8, 117–18 vs. pertinence dative, 54, 150, 151
weak pronouns and, 116–17 possessor raising analysis of, 161
sentence-internal restrictions on,
on-PP, 16, 29, 30–31 154–58
analysis of, 43–44 sloppy identity readings of, 165
spatial analysis of, 163–67
Part-whole applicatives, 11, 14, vs. subset-superset relation, 152
148–77 transitive verbs and, 155
abstract entities and, 152 unergative verbs and, 149, 169,
adjunct PP and, 155–56 174–75, 223
vs. affected experiencers, 32–35, verbal restrictions on, 156
151, 157 yes/no-question test of, 158
again-modification of, 171–72, Passivization
175 of benefactives, 126–28, 139

 
244  

of Dative of Inaction, 185 if- clause test of, 141–42


of resultative, 70 negation test of, 141
of subject co-referential applica- vs. true benefactive, 144–47
tives, 110–11 wh-question test of, 141
See also bekommen; kriegen yes/no-question test of, 141
Pertinence datives, 30, 49–65 Projection problem, 20–21
vs. affected experiencers, 30, 33– Psychological affectedness, 30–31
35, 38–40 Purpose phrase
analysis of, 50–56, 209 benefactive with, 131–34
animacy requirement of, 34
at-issue meaning and, 40 Quantifier binding test, 22
binding analysis of, 56–61 of affected experiencers, 38
coordination and, 64–65
if-clause test of, 40 -rare
negation test of, 39 in Japanese, 77–78
not-at-issue meaning and, 38–40 Rec head, 137–38
vs. part-whole applicatives, 38– Recipient benefactives, 11, 12, 15,
39, 54, 150, 151 120, 121–23, 135–40
part-whole interpretation of, 38– vs. affected experiencers, 30
39 again-modification of, 139
possessive interpretation of, 33, analysis of, 116, 137–38
34–35, 38–40, 50–56, 60 binding of, 136
possessor raising analysis of, 61– co-occurrences of, 201
65 vs. ditransitives, 135, 136–37,
relational nouns with, 50–51, 53, 139, 140
55–56, 60–61 meaning of, 135–36
silent purpose clause and, 58–59, particles and, 139–40
61 passivization of, 139
sloppy identity effects and, 59–60 unaccusative verb and, 138
sortal nouns with, 51–52 verbal restrictions on, 138–39
transformational analysis of, 61– Recursivity of applicative argu-
65 ments, 188–90
wh-question test of, 40 Reflexive pronoun, 104–5
yes/no-question test of, 39–40 Relational nouns, 50–51, 53, 55–56,
Possessive interpretation 60–61
of pertinence dative, 33, 34–35, Resp head, 184–85
38–40, 50–56 Resultativity, 68, 69–71
Possessor raising, 61–65
Preposition, 16, 29 SE-anaphor, 105–6
See also for-PP; für-PP; on-PP Sentience requirement, 30–32, 182
Prepositional benefactives, 140–47 Silent purpose variable
analysis of, 117, 142–47


  245

in pertinence dative analysis, 58– wh-question test of, 107


59, 131–33 yes/no question test of, 107–8
Sortal nouns, 51–52 Surface structure, 25–26
Subevents, 23
Subject co-referential applicatives, Thematic roles, 23–24, 25
11, 15, 101–15 Theme, 23–24
vs. affected experiencers, 30, 35–
36 Verb(s)
again-modification of, 111, 114 atelic motion, 69
analysis of, 108–15, 214 change of location, 181
animacy preference of, 103–4 change of state, 180
anti-locality account of, 111–13 continuation of condition, 68–69,
vs. benefactive applicative, 104, 180
127 controlled/uncontrolled process,
co-occurrences of, 189, 193–94, 180
201 ditransitive, 121, 129, 135, 136–
direct object and, 106 37, 140
vs. double object construction, down-home, 110
114 existence of state, 69
down-home verbs and, 110 inchoative, 180
features of, 102–4 intransitive, 65–67, 129–30, 179,
form of, 104–6 180
if-clause test of, 108 particle, 69–70
as low applicative, 113–15 of physical affectedness, 67
meaning contribution of, 102 of psychological affectedness, 67–
negation test of, 107 68
not-at-issue meaning and, 102, resultative, 71
106–8, 117–18 static, 71
optionality of, 101 telic motion, 180
vs. part-whole applicatives, 150 transitive, 65–67, 106, 179
passive and, 110–11 Verbal restrictions
pronoun preference of, 103 on affected experiencers, 30, 65–
vs. recipient benefactive, 104–5 71
vs. reflexive pronoun, 104–5 on benefactives, 129–31
as SE-anaphor, 105–6 on Dative of Inaction, 179–81
sloppy identity reading of, 105–6 on part-whole applicatives, 156
transitive verb and, 106 on recipient benefactives, 138–39
unergative verb and, 109–10 on subject co-referential applica-
verb adjacency of, 116 tives, 110–11
verbal restrictions on, 110–11 Verb phrase
verbs of consumption/creation affected experiencer, 41–44
and, 110 event semantics of, 23–26

 
246  

high vs. low applicatives and, 27–


28
pertinence dative, 55–56
resultative, 68, 69–71
Voice, 24

Weak pronouns, 115–17


at-issue meaning and, 116
not-at-issue meaning and, 116–17
Wh-question test, 22
of affected experiencers, 37, 74,
180
of benefactives, 124
of Dative of Inaction, 183
of pertinence datives, 40
of PP-alternative benefactives,
141
of prepositional benefactives, 141
of subject co-referential applica-
tives, 107

Yes/no-question test, 21
of affected experiencers, 36–37,
75, 79–80
of benefactives, 124
of Dative of Inaction, 182
of part-whole applicatives, 158
of pertinence datives, 39–40
of PP-alternative benefactive, 141
of prepositional benefactives, 141
of subject co-referential applica-
tives, 107–8


B E R K E L E Y
I N S I G H T S
IN LINGUISTICS
AND SEMIOTICS
Irmengard Rauch
General Editor

Through the publication of ground-breaking scholarly


research, this series deals with language and the multiple and
varied paradigms through which it is studied. Language as
viewed by linguists represents micrometa-approaches that
intersect with macrometa-approaches of semiotists who
understand language as an inlay to all experience. This data-
based series bridges study of the sciences with that of the
humanities.

To order other books in this series, please contact our


Customer Service Department at:

800-770-LANG (within the U.S.)


212-647-7706 (outside the U.S.)
212-647-7707 FAX
CustomerService@plang.com

Or browse online by series at:


www.peterlang.com

Vous aimerez peut-être aussi