Date: Wed, 19 Jan 2005 08:47:02 +0200
From: Seppo Kittilä <sepkit@utu.fi>
Subject: Non-nominative Subjects, vol. 1
EDITORS: Bhaskararao, Peri; Subbarao, Karumuri Venkata
TITLE: Non-nominative Subjects
SUBTITLE: Volume 1
SERIES: Typological Studies in Language 60
PUBLISHER: John Benjamins
YEAR: 2004
Seppo Kittilä, Department of Linguistics, University of Turku, Finland
OVERVIEW
The book under review is a collection papers dealing with non-nominative
subjects (NNS) from a cross-linguistic perspective. The collection
comprises 15 papers, which examine data from a wide range of languages and
from different perspectives and in different frameworks. The papers go
together rather well, even though some of the papers written in a
generative framework are not easy to follow for a linguist with a more
functional background. Despite this, a variety of themes are examined (cf.
below), and the book thus serves as a valuable starting point for future
studies on the topic.
R. Amritavalli's paper 'Experiencer datives in Kannada' deals with dative
(not only experiencer) subjects in Kannada. The paper examines the
construction both synchronically and diachronically, even though synchrony
is focussed on. Both coding (morphological) and behavioral (syntactic)
properties of the construction are examined. It is shown that on the basis
of these features, dative subjects are more like subjects than datively
marked objects, which lack the relevant properties. Semantics of the
construction type is not discussed in any detail. The paper is written in
a rather formalistic framework, but the theory is not too strongly
stressed, which makes the paper accessible to functionally oriented
linguists as well.
Harbir Kaur Arora's and Karumuri Venkata Subbarao's paper 'Syntactic
change and convergence' examines the nature of non-nominative subjects
Dakkhini and Konkani from the perspective of language change due to
contact and convergence. The paper explains the differences in NNS's in
the studied languages on the basis of different contact languages, and the
explanations mostly seem plausible, even though examples illustrating the
actual changes are not given.
Josef Bayer's paper 'non-nominative subjects in comparison' is not quite
what a general typologist expects it to be. The paper does not deal with a
variety of different constructions having non-nominative subjects, but it
focusses on quirky subjects in German and Icelandic (occasionally data
from Bengali is also examined). The paper examines NNS's in the
aforementioned languages in light of behavioral properties, such as
control and conjunction reduction. It also discusses the differences and
similarities of NNS's and (dative) objects.
Balthasar Bickel's paper 'The syntax of experiencers in the Himalayas'
discusses the formal realization of events with an experiencer from a
comparative perspective (Indo-Aryan vs. Tibeto-Burman languages). Both
coding and behavioral properties of experiencers are examined in the
paper. Coding properties refer mainly to case marking, while behavioral
properties have to do with syntactic pivots. With regard to coding
properties, two constructions types, experiencer-as-goal and experiencer-
as-possessor are distinguished (in the former case, the experiencer
surfaces in the dative, while in the latter case the experiencer takes a
genitive case). It is shown that the languages spoken in the area diverge
with regard to the morpho-syntactic nature of the construction type in
question. For example, experiencers have access to syntactic pivothood
only in Tibeto-Burman, not in Indo-Aryan.
Bernard Comrie's paper 'Oblique-case subjects in Tsez' examines
constructions with a non-canonical (i.e. non-ergative) subject in Tsez.
These comprise experiencer constructions, accidental constructions, and
potentials, so also the semantic nature of the constructions is touched
upon. The subject of these constructions in not in the dative, but either
in the lative or the possessive. After introducing the construction type,
they are examined in light of behavioral properties, such as the formation
of imperatives and definite future, reflexivization, and the obligatory
coreferential noun phrase deletion. It is shown that oblique-case subjects
differ from ergative subjects in the formation of the imperative and the
definite future, but other properties are not helpful in this respect.
Probal Dasgupta's paper 'Some non-nominative subjects in Bangla' discusses
a variety of non-nominative subjects in Bangla (Bengali). The examined
constructions comprise obligational dative constructions, experiencer
subject constructions, generic locative constructions, and reciprocity
locative constructions. These constructions are also formally
distinguishable, since the obligational dative construction takes a dative
subject, while experiencers occur in the so-called indirect case. In the
case of the locative constructions, the subject is naturally in the
locative. This paper focusses on presenting data, and as such is a
valuable contribution, even though detailed analyses of the constructions
are lacking (which the author himself also states).
Alice Davison's paper 'Non-nominative subjects in Hindi-Urdu: VP structure
and case parameters' discusses the subject properties related to ergative
and dative subjects in Hindi-Urdu. It is shown that the two subjects
behave differently with regard to some of the subject properties, such as
reflexive binding. The mere case is not a sufficient subject criterion of
subjecthood in Hindi-Urdu. The paper is very loyal to the generative
tradition, and it is thus most valuable to scholars working within this
kind of framework
Nicholas Evans' paper 'Experiencer objects in Iwaidjan languages
(Australia)' differs from all other papers of the volume in that it, as
the title suggests, discusses experiencer objects, since the languages
under consideration do not have experiencer subjects. Similar meanings are
expressed by constructions like 'shame did me' in Iwaidjan languages. The
meaning is thus more or less the same, but its formal manifestation is
radically different. The stimulus is always the subject in these
constructions, while the experiencer is coded as an object, which,
according to Evans, reflects a high tolerance for active constructions
with inanimate subjects.
Susann Fischer's paper 'the diachronic relationship between quirky
subjects and stylistic fronting' examines the grammaticalization of quirky
subjects in Germanic languages, primarily in Icelandic. It is argued that
stylistic fronting and the emergence of quirky subjects are intimately
related. Evidence is found in the history of Germanic languages (both are
missing in Modern Swedish, but not in Old Swedish, while Old Icelandic did
not have quirky subjects or stylistic fronting, but Modern Icelandic has
both). In addition to the development of quirky subjects, their
subjecthood is examined in light of typical subject properties.
Peter Hook's and Omkar N. Koul's paper 'Case as agreement' examines non-
nominative subjects in Kashmiri, Poguli, and Gujarati. The main argument
of the paper is that the ergative affix in these languages agrees in
tense, aspect, and mood with the verb. Consequently, it is argued that the
variation between nominative and ergative, which is typical of almost all
Indo-Aryan languages with morphological ergativity, is a semantically
empty, automatic agreement. Therefore, it does not have any meaning. The
claim is perhaps somewhat bold, but the paper puts forward a strong
argument.
K. A. Jayaselaan's paper 'The possessor-experiencer dative in Malayalam'
discusses dative constructions in Malayalam (the author explicitly avoids
the term 'dative subject', since in his opinion the dative is not the
subject in these constructions). It is shown that the dative is used with
possession, experience (both mental and physical), 'know' verbs, and also
certain modals (such as 'may' and 'can'). Mental and physical experience
are formally distinct, since mental experience allows a variation between
dative and nominative, while this is not possible with physical
experience. In addition to distinguishing the dative constructions
semantically, the paper also discusses the subject properties related to
the construction type in question. Most of these properties (such as PRO)
are useless for Malayalam, since they do not distinguish between subjects
and objects.
B. Lakshmi Bai's paper 'Acquisition of dative subject in Tamil' adopts a
completely different approach to NNS's. It examines the acquisition of
these constructions by one monolingual Tamil child and two bilingual
(Tamil and Telugu) children. Subject properties of datives are not
examined in this paper. Despite this, the paper provides us with very
interesting insights into the nature of dative subjects. All the studied
children overextend the dative subject in the beginning, which means that
the acquisition of the construction as such poses no problems, but the
problems lie in the correct use of the dative. The dative is even extended
to wrong contexts, since it is used to mark an animate direct object
(DOM), and also (stative) locative, even though these are not a part of
the Tamil grammar.
Howard Lasnik's paper 'The position of accusative subject in the
accusative-infinitive construction' discusses the nature of the accusative
subject in accusative-infinitive constructions, such as 'Jack believed
Joan to be famous' in English (these are also labelled as 'Exceptional
Case Marking' constructions). It is shown that the accusative subject
shares many properties with an object.
Anoop Mahajan's paper 'On the origin of non-nominative subjects' aims at
examining the contexts, which require a non-nominative subject. It is
suggested that non-nominative subjects in Hindi arise only in non-
accusative contexts (the author claims that this applies to other
languages as well). This means that non-nominative subjects are limited to
constructions, which lack an accusative-case assigner. Even though the
author says that the non-accusativity is responsible for all non-
nominative subjects, he adds that the phonological shape of the non-
nominative marker is determined semantically. As for the subject status of
NNS's, the author states that they are not subjects in the usual sense.
This is based on the examination of properties such as binding, control,
and pronominal obviation and deletion.
Makoto Minegishi's paper 'Southeast-Asian language: A case for the
caseless' examines NNS's in two languages without a morphological case.
The languages in question are Thai and Khmer. Despite the very interesting
goal, it is not clear to me how the question is addressed in the paper.
The author proposes two models, which he labels as cephalopod model and
centipede model, the former should explain the role marking in case-
marking languages, while the latter is applicable to caseless languages.
Non-nominative subjects are discussed only in passing based on the
expression volitionality in the examined languages (different serial verb
constructions are used here).
CRITICAL EVALUATION
This book is a rather controversial book. It has clear merits, but, on the
other hand, it also suffers from some weaknesses. Even the title of the
book is 'ambiguous'. After reading the two first chapters of the book I
realized that the title can be read in two different ways depending on
which part of the title is stressed. If the emphasis lies on 'non-
nominative', one would expect a cross-linguistic analysis of different
cases where the marking of the subject deviates from the expected
nominative (cf. the title of the Aikhenvald, Dixon & Onishi volume 'Non-
canonical marking of subjects and objects'). On the other hand, if one
stresses the word 'subjects', one rather expects an examination of
(subject-like) arguments that occur in the subject position, but do not
bear nominative marking. In this case, the goal is to show that these
arguments are indeed subjects, even though not canonical ones. As a
general typologist interested in a variety of transitivity phenomena
including agency (see e.g. Kittilä: forthcoming), I was thus somewhat
disappointed by the contents of the volume, since the latter reading of
the title is more accurate as regards the contents of this book, even
though some semantic aspects of the constructions are also touched upon.
Perhaps the biggest merit of the book is the presentation of data from a
wide range of languages. The data is valuable to anyone interested in the
examined phenomenon regardless of whether his/her approach is functional
or generative. In addition, also the development and acquisition of NNS's
is studied, which is a clear contrast to other studies dealing with
similar topics. Some of the papers written in a generative framework are
perhaps not accessible to a functionalist, but the data is nevertheless
valuable and interesting (it is always properly glossed). The book also
serves as a good starting point for future studies on the topic, since it
provides its readers with a good overview of constructions involving non-
nominative subjects (the focus in on experiencers, but some other
constructions are touched upon as well). It is interesting to see how the
same kinds of event enter non-nominative subject constructions in
structurally and genetically diverse languages. The theme
s discussed in the present volume belong to the periphery of syntactic
research (see, however, Verma & Verma 1990 and Aikhenvald, Dixon & Onishi
2001 for similar studies), but since most languages of the world have some
kind of non-nominative constructions, I hope that scholars working on
individual languages would find the time to read this book, or at least
parts of it. For example, Finnish linguists have spilled gallons of ink
discussing the partitive subject of Finnish, but unfortunately the
phenomenon has not been discussed from a cross-linguistic perspective.
Another merit of the book is that it shows that the subject is a
multilayered notion (see also e.g. Keenan 1976) with a number of
different definitions. Not the mere nominative case marking suffices for
defining the subject, since in light of a number of syntactic features
(such as PRO, reflexive binding, and pivothood) also non-nominatively
marked arguments in the subject position resemble canonical nominative
subjects rather than direct objects, for example. So, a more holistic
approach is recommendable, if one wishes to study the notion of subject in
detail. On the other hand, it seems that the notion is not clear to all of
the contributors themselves, since the notion has a variety of uses in the
book depending on the author (no unified definition is suggested anywhere
in the book, an introductory chapter would have been helpful). Moreover,
the authors usually stress syntactic (behavioral) or morphological
(coding) properties of subjects without taking the other option into
account. In other words, even though two possible ways (morphological and
syntactic) of defining the subject are in principle recognized, it is not
clear which of these approaches is the most suitable one (syntactic
features (i.e. behavioral properties of Keenan 1976) are usually
emphasized, while non-nominative case marking is regarded as a 'quirk').
Some of the authors take the notion as such for granted, while others
adopt a more cautious approach, or do not use the notion at all. Some
authors adopt a 'mixed' approach. For example, Fischer examines features
of subjects on the first four pages of her paper, but questions the notion
on the fifth page. In case the notion of subject is taken as given,
(usually) behavioral properties of subjects are studied without
questioning their applicability in the study of subjecthood.
On the other hand, Bickel, Comrie, Evans, and Jayseelan adopt a more
cautious approach to the notion of subject. Bickel explicitly
distinguishes case marking and syntactic features of subjects and studies
them separately, while Comrie, Evans, and Jayselaan use the notion very
cautiously. The tests typically used for subjects are either useless in
the languages they study, or they do not provide us with a comfortable
result with regard to the notion of subject (i.e. it is hard to say
whether nominatively or datively marked arguments are better subjects).
So, the subject is not a universal notion, and should not be examined as
one. Thus, even though combining different approaches and different
frameworks can as such be considered valuable, it can also lead to
confusions, which happens occasionally in the book under review. This is a
clear difference to the aforementioned Aikhenvald, Dixon & Onishi volume,
which remains outside formal theories, and which starts with an
introductory chapter. Some of the papers written in a generative framework
are very hard to follow for a linguist with a more functional background
(Bayer's, Davison's and Lasnik's papers are good examples of this). The
goal of these papers is to argue for a theory-internal notion, which makes
the paper impossible to evaluate for scholars not familiar with the
terminology and the theory used. Some of the suggested explanations are
not satisfactory, if they are evaluated against a functional background
without any detailed knowledge of the theory itself. The extreme example
of this is found in Mahajan's paper, where it is claimed that Italian is
also an ergative language, if one accepts the fact that the ergative
marker is always incorporated in the verb, and thus never surfaces. This
kind of claim may be accepted by generativists, but it does not convince
many functionalists.
Even though the book under review discusses a variety of features of non-
nominative subjects, there are properties that should be taken a closer
look at in the future. The most intriguing feature of non-nominative
subjects not studied in any detail in the present volume is their
semantics. As was noted above, the emphasis of the volume clearly lies on
morpho-syntax. The very basic semantics of non-nominative subjects is
discussed, since most of the papers list the canonical uses of the non-
nominative subjects in the examined language(s). However, a more precise
examination of NNS's would be very valuable and would probably provide us
with new insights into features like agency. A lower degree of agency (or
the lack of full control) seems like the triggering factor of the non-
nominative case marking in many of the discussed cases (see e.g. Tsez data
examined in Comrie's paper), but this is not discussed in enough detail in
the present volume. A potential basis for this kind of examination is
suggested in Kittilä (forthcoming) where one 'non-nominative subject
construction', namely so-called involuntary agent constructions (see
Haspelmath 1993: 292) are studied in light of features like the inherent
agency related to the agent, and the nature of the event itself (does it
imply agency or rather the lack of it). For example, in a number of
languages only humans can enter an involuntary agent construction, while
clauses like 'the storm accidentally broke the boat' are not acceptable.
It would be interesting to see whether these features help us in
distinguishing between different non-nominative constructions, not
formally, but semantically. Also common features shared by non-nominative
constructions may be easier to explain.
REFERENCES
Aikhenvald, Alexandra Y., Dixon, R. M. W. & Onishi, Masayuki (eds.) (2001)
Non-canonical marking of subjects and objects. Amsterdam/Philadelphia:
John Benjamins.
Haspelmath, Martin (1993) A grammar of Lezgian. Berlin/New York: Mouton de
Gruyter.
Keenan, Edward (1976) Towards a universal definition of subject. In:
Subject and topic, Charles N. Li (ed.), 303-333. New York: Academic Press.
Kittilä, Seppo (forthcoming) Remarks on involuntary agent constructions.
To appear in Word 56:3.
Verma, Mahindra K. & Mohanan, K. P. (eds.) (1990) Experiencer subjects in
South Asian languages. Stanford University: The Center for the Study of
Language and Information.
|
