LINGUIST List 12.645
Thu Mar 8 2001
Review: Senft, Systems of Nominal Classification
Editor for this issue: Terence Langendoen <terrylinguistlist.org>
What follows is another discussion note contributed to our Book Discussion
Forum. We expect these discussions to be informal and interactive; and
the author of the book discussed is cordially invited to join in.
If you are interested in leading a book discussion, look for books
announced on LINGUIST as "available for discussion." (This means that
the publisher has sent us a review copy.) Then contact Simin Karimi at
simin
linguistlist.org or Terry Langendoen at terry
linguistlist.org.
Directory
Wolfgang Schulze, Review of Senft (ed.) Nominal classification
Message 1: Review of Senft (ed.) Nominal classification
Date: Thu, 08 Mar 2001 17:18:58 +0100
From: Wolfgang Schulze <W.Schulzelrz.uni-muenchen.de>
Subject: Review of Senft (ed.) Nominal classification
Senft, Gunter (ed.) (2000). Systems of Nominal Classification.
Cambridge: Cambridge University Press. vii, 350 pages.
Claudia Gerstner-Link, University of Munich
Research on Nominal Classification has a rather long standing tradition
in Linguistics (starting with Gerlach Royen's well-known opus magnum
'Die nominalen Klassifikationssystem in den Sprachen der Welt' (1929)).
Systems of nominal classification have attracted the interest of
linguists not only because of their diverging morphosyntactic behavior
and by universal features possibly underlying them, but also because
they seem to offer a tool to relate linguistic categorization to
principles and procedures of how human beings categorize (or learn to
categorize) their worlds in a cognitive and/or social perspective.
No wonder that research groups specialized in topics of
cognitive linguistics such as the Cognitive Anthropology Research Group,
now the Department of Language and Cognition at the Max Planck Institute
for Psycholinguistics (Nijmegen) have devoted much work on this
promising subject. Gunter Senft, one of the members of this research
group had organized a workshop 'Back to Basic Issues in Nominal
Classification' in 1993 the output of which is - in parts - documented
in the latest volume of Nominal Classification, namely in Senft (2000).
This volume explicitly concentrates on the 'semantic' (or - in a
rather broad sense - cognitive) foundations of nominal classification.
It offers ten articles by renowned specialists in the topic to which is
added an introduction by the editor (pp. 1-10) and an index (languages,
subjects, names). Obviously, a single volume cannot comprehensively
cover the empirics of the phenomenon. It has to waver between in-depth
studies concentrating on single language data and a broader typological
perspective. Fortunately enough, the authors concentrate on the single
language (or contrastive) perspective and avoid a necessarily
superficial cross-linguistic approach (mass comparison). Yet, the reader
probably wants to know about the motives of the editor concerning the
choice of languages that are presented in this volume. Is one of the
motives the basic condition that the systems in question should have a
'conceptual-semantic basis rather than a formal basis' (p.1)? Or is the
choice simply determined by the special research interests of the
authors?
The following languages are covered (more or less extensively) by the
authors: Kilivala (Austronesian, Trobriand Islands) (Senft), Tariana
(North-Arawak) (Aikhenbald), Akatek (Mayan) (Zavala), Arrernte
(Pama-Nyungan) (Wilkins), Japanese (Inoue), Tongan (and other Oceanic
languages) (Broschart), Bantu languages (Demuth), Russian, Quafar (East
Cushitic), Ghodoberi (East Caucasian), Zande (Niger-Kordofanian) and
Arapesh (Papuan: Torricelli) (Corbett & Fraser). Colette Grinevald's
contribution offers data from a number of other languages (in a
typological perspective). If we bear in mind that some languages such as
Ghodoberi, Zande, or Quafar are treated in a rather superficial manner we
arrive at the impression that the choice of languages is not very
equilibrated and that it does not conform to the standards of typological
sampling. I confess that editing a volume such as Senft's book generally
necessitates some kind of compromising: Do I want to demonstrate the
universe of nominal classification as it shows up in 'the languages of the
world'? In this case, prominent language groups (or languages) should have
been comprehensively monitored as for all relevant aspects of the
phenomenon (including the diachronic perspective). Or: Are there arguments
for a 'categorial' subdivision (and choice of data) which would show us
the universe of nominal classification as such? I am left with the
impression that the editor has taken the 'third' road, namely to leave the
decision to the authors which systems (languages) are to be included in
the volume.
The scope of all ten papers is described in Senft's introduction.
The first two papers (G. Senft: 'What do we really know about nominal
classification systems', pp.11-49) and C. Grinevald: 'A morphosyntactic
typology of classifiers', pp. 50-92) can be regarded as introductory
sections on the semantics (Senft) and morphosyntax (Grinevald) of
nominal classification systems. The next five papers are devoted to
single language data from specific perspectives: A.Y. Aikhenvald:
'Unusual classifiers in Tariana', pp.93-113; R. Zavala: 'Multiple
classifier systems in Akatek (Mayan), pp.114-146, D.F. Wilkins: 'Ants,
ancestors and medicine: a semantic and pragmatic account of classifier
constructions in Arrernte (Central Australian), pp.147-216; K. Inoue:
'Visualizing ability and nominal classification: evidence of cultural
operations in the agreement rules of Japanese numeral classifiers',
pp.217-238; J. Broschart: 'Isolation of units and unification of
isolates: the gestalt-functions of classifiers', pp.239-269.The next two
papers turn again to more general issues of noun classification: K.
Demuth: 'Bantu noun class systems: loanword and acquisition evidence of
semantic productivity', pp.270-292, G.G. Corbett & N.A. Fraser: 'Gender
assignment: a typology and a model', pp.293-325. Finally, J. Lucy sums
up the discussion ('Systems of nominal classification: a concluding
discussion').
In this review, I cannot discuss all papers in detail. I will focus on
the following ones: Senft, Grinevald, Aikhenvald, Zavala, Wilkins, and
Lucy. The choice is determined by the assumption that these papers best
represent the make-up and argumentative tendencies of the volume in
question. In discussing the data from Kilivala (Austronesian), G. Senft
tries to outline a psychologically and ethnolinguistically grounded
approach to nominal classification. He interprets classification as a
global means of human beings to interact with their environment into which
linguistic classification is embedded. He enumerates (and exemplifies) the
different techniques of linguistic classification and criticizes the fact
that much too often such techniques are described as single events, but
not in terms of a complex network structure which would allow to describe
in more detail the interaction (or co-paradigmatization) of different
classification techniques in one language. In order to pursue this
question Senft elaborates the system of Kilivala, a so-called numeral
classifying language. In criticizing the traditional division between
qualifying (sortal) and quantifying (mensural) procedures, he discusses a
feature based approach which focuses on the qualitative aspect (without
telling us how qualitative features intervene with deictic and
quantitative features). Senft describes classification systems as some
kind of feature matrix the realization of which he characterizes as being
culturally determined. Most importantly, the relation between nouns and
classifiers is described as a dynamic (and variable) relation which is
determined by discourse strategies. He arrives at a complex network that
constitutes the semantics and functions of classifiers. Some classifiers
seem to be strictly localized, other more variable which allows the
speaker to pragmatically apply switching strategies. Senft concentrates on
the description of communicative effects of dynamic classification but
unfortunately refrains from looking at a cognitive motivation for this
remarkable phenomenon.
Grinevald's article presents a morphosyntactic typology of
classifiers as a first step towards an overall grammatical typology of
classifiers. There is no doubt that a classifier typology is needed as
a heuristic tool to better understand the morphosyntactic and
semantic-pragmatic intricacies of many complex classifier systems. Such
a typology should be based on prototypes of classifier systems each for
numeral classifiers, noun classifiers, genitive (possessive)
classifiers, and verbal classifiers. These are the main classifier
systems Grinevald is concerned with. In approaching the
characteristics of classifiers she first locates classifiers in the
middle of a continuum between lexical (measure and class terms) and
grammatical (gender and noun classes) systems of nominal classification;
then she outlines the major morphosyntactic properties of each
classifier type. A decisive argument for the existence of different
types of classifiers is the fact that languages may combine several
classifier systems: Jakaltek is one of the best examples of a multiple
classifier system. An important issue is the question whether a
particular type of classifier correlates with special semantic
features. As a working hypothesis Grinevald formulates the following
correspondences: numeral classifiers are preferably associated with
physical properties, noun classifiers with properties of material and
essence, and genitive classifiers with functional properties. On the
other hand, there is a functional similarity between all classifiers as
they are a device of marking individuation of nouns. However, the
nature of this process of individuation is far from being clear in its
details and therefore urgently calls for further research.
Particularly, it has to be thoroughly investigated how co-existing
functional differences between the classifier types match their
common property of individuation. Let us take the case of Jakaltek as
an example: If noun classifiers are operators of quality, one faces the
question how the assumption that Jakaltek noun classifiers function as
determiners is in line with their semantics of
qualification, since it can be argued that determination is at least
equally related to quantification than to qualification. Thus, the
future task is to develop in detail a functional typology of classifiers
based on Grinevald's convincing insights and suggestions that certainly
have a desirable and positive impact on any further studies of
classifier systems.
Aikhenvald's contribution deals with Tariana (North-Arawak). The
author tries to explain the so-called 'unusual' system of nominal
classification of Tariana in reference to the general classificatory
procedures in this language. By 'unusual' is meant that 'class' is
linked to deictic elements such as demonstrative pronouns and articles.
This 'unusual' procedure is only one of the multiple (six fold)
classificatory techniques of Tariana which are clearly elaborated in
this paper (among them attribute noun concordance, numeral classifiers,
classification of possessives, class markers on the verb with certain
types of diathesis and subordination). The different strategies are
finally summed up and evaluated in an impressing resume. The motivations
of classifying strategies in Tariana are comprehensively monitored by
the author; however, the argumentation lacks a more general answer to
the question of what is the basic classificatory strategy (if there is
one). Perhaps we could claim that the most prominent function of
classification in Tariana is the grading of reference. Tariana allows a
pragmatically motivated type of classification with deictic structures
(optionally, context depended, fluid) which is opposed to a more
semantic procedure of subcategorization (esp. noun classes and noun
classifiers). The technique of article based classification seems to be
operational between these two poles (deictic < article < noun class).
The objectives of Wilkins's contribution are to revise the type of
noun classifiers - itself a subtype of classifying constructions. He
opposes the data from Arrernte to those of Yidiny and Jakaltek and
convincingly elaborates the different functional properties in the three
languages. Arrernte has generic nouns which classify nouns according to
their inherent semantics, functions, and usage. The corresponding
constructions are called generic-specific constructions. The choice of
the appropriate classifier depends on the linguistic and communicative
(situative) context, in other words, it is fluid. In Jakaltek, on the
contrary, such a classification is strongly linked to determination and
individuation. Yidiny, too, lacks this aspect of determination; however,
in this language there seems to be no difference between
generic-specific constructions and noun phrases consisting only of
specific nouns. In Arrernte, isolated specific nouns are embedded
into a discourse pattern different from that of their use within
generic-specific constructions. They also call for syntactic and
semantic-pragmatic features that stand in complementary distribution
to those of generic-specific constructions. With respect to the
Arrernte data it seems crucial to describe the semantics of
generic nouns. In order to tackle this task, Wilkins makes use
of Wierzbicka's conceptual analysis and Fillmore's Frame Semantics.
He shows that a specific noun receives its interpretational frame
by means of a generic noun, or: the semantics of a generic-specific
construction results from the semantics of the two components and their
configurational structure. The technique of classification shows up
as a configuration without having true classifiers. In other words:
it is an emergent phenomenon.
Zavala's article deals with Akatek (Mayan) as an example of a
multiple classifier system which contains four different classifier
paradigms. A noun phrase may exhibit up to three different classifiers,
but is only the numeral classifiers that are obligatory. If we have a
structure consisting of a numeral plus numeral classifier plus plural
classifier plus noun classifier plus noun each of the three classifiers
expresses a particular aspect concerning the individuation of the noun:
the exact quantity, plurality, and referentiality. Noun classifiers may
occur as the only element that combines with a noun or together with
other modifiers or classificatory devices. However, it remains unclear
how word order is determined in complex noun phrases: does the
numeral suffixed by a numeral classifier precede or follow the noun
classifier? Apparently both ordering principles are possible. But note
that a middle position of numerals in the NP generally
can only be found, if the preceding elements are
articles or demonstratives, that is, if the whole phrase is definite.
Thus, example (39)c. would especially support a determiner like reading
of noun classifiers as it happens to be the case in Jakaltek. By
contrast, the question remains what exactly is the function of the noun
classifier in a structure where it follows the numeral expression (cf.
examples (1), (15), (21), (22)). Unfortunately, Zavala does not discuss
this interesting puzzle. The main impression one gets from the
Akatek classifier system is that it is a system in transition that
should be further studied in order to reveal the stages of development
of classifiers systems into determiner systems.
Lucy's discussion of the state of the art of analyzing classifier
systems is very elucidatory in that he clearly indicates the still weak
points in the overall conception of this topic. Is it experience or
linguistic form that is classified? Another essential issue is the
"actual logic uniting all forms of nominal classification, namely their
contribution to adequate noun phrase reference." (p.329) This
referential function of any kind of nominal classification should be
much more focused on in future research. In determining what the
classifiers contribute to reference in a particular language and
construction one must not become biased by translation projections. This
can be prevented by reflecting the three different functional components
of a noun phrase given by Silverstein: the intrinsic type of the
referent, its extrinsic individuation status, and its discourse
presupposability. From this vantage Lucy concludes "that classifiers of
different types form a gradient from those dealing with relatively
intrinsic qualities of a referent, to those dealing with relatively
extrinsic qualities of a referent, to those dealing with relatively
transient pragmatic contingencies of the speech event." (p.339) Thus,
the (morpho)syntactic position of a classifier in a noun phrase is
crucial to understand its exact function. Although all articles of the
book present interesting results or give rise to substantial questions
within the field it is Grinevald's view that comes closest to Lucy's in
arguing for a more elaborated functional, i.e. semantic and cognitive,
perspective in debating the topic of nominal classification.
To conclude: The present volume is a welcome contribution to the study
of nominal classification both in a typological and a semantic
perspective. It owes much to the highly informed authors and the specific
phenomena of the languages presented in the ten papers. Once the reader
has worked through the huge amount of material and analyses (s)he will
gain a deeper insight into the morphosyntactic and semantic strategies of
establishing reference. It will be upon future researchers to draw their
conclusions from the analyses and interpretations proposed in the
different paper's and to test them against their own data. In this sense,
Senft's volume figures as an important step towards what is urgently in
want but not yet in sight: a universal and typologically validated theory
of the cognitive motivation and pragmatic (communicative) power of
linguistic classification.
Dr. Claudia Gerstner-Link is a Research Fellow at the Institute for
General and Typological Linguistics of the University of Munich. After
extensive fieldwork in Papua New Guinea, she is currently preparing a
major book on Kilmeri, a hitherto undescribed Papuan language. Her
research interests are Papuan linguistics, Formal and Functional Typology,
and Cognitive Linguistics.
Editors' note: This review was originally requested of Prof. Dr. Wolfgang
Schulze, Institut f�r Allgemeine und Indogermanische Sprachwissenschaft
Universit�t M�nchen. With our permission, he asked Dr. Gerstner-Link to
prepare the review.