19.68, Review: Syntax: Aarts (2007)

[LINGUIST Network]

LINGUIST List: Vol-19-68. Tue Jan 08 2008. ISSN: 1068 - 4875.

Subject: 19.68, Review: Syntax: Aarts (2007)

Moderators: Anthony Aristar, Eastern Michigan U <aristar at linguistlist.org>
            Helen Aristar-Dry, Eastern Michigan U <hdry at linguistlist.org>
 
Reviews: Randall Eggert, U of Utah  
         <reviews at linguistlist.org> 

Homepage: http://linguistlist.org/

The LINGUIST List is funded by Eastern Michigan University, 
and donations from subscribers and publishers.

Editor for this issue: Randall Eggert <randy at linguistlist.org>
================================================================  

This LINGUIST List issue is a review of a book published by one of our
supporting publishers, commissioned by our book review editorial staff. We
welcome discussion of this book review on the list, and particularly invite
the author(s) or editor(s) of this book to join in. To start a discussion of
this book, you can use the Discussion form on the LINGUIST List website. For
the subject of the discussion, specify "Book Review" and the issue number of
this review. If you are interested in reviewing a book for LINGUIST, look for
the most recent posting with the subject "Reviews: AVAILABLE FOR REVIEW", and
follow the instructions at the top of the message. You can also contact the
book review staff directly.

===========================Directory==============================  

1)
Date: 08-Jan-2008
From: Randall Eggert < randy at linguistlist.org >
Subject: Review: Syntax: Aarts (2007)

 

	
-------------------------Message 1 ---------------------------------- 
Date: Tue, 08 Jan 2008 15:08:20
From: Randall Eggert [randy at linguistlist.org]
Subject: Review: Syntax: Aarts (2007)
E-mail this message to a friend:
http://linguistlist.org/issues/emailmessage/verification.cfm?iss=19-68.html&submissionid=165660&topicid=9&msgnumber=1  


Announced at http://linguistlist.org/issues/18/18-1510.html 

AUTHOR: Aarts, Bas
TITLE: Syntactic Gradience
SUBTITLE: The Nature of Grammatical Indeterminacy
PUBLISHER: Oxford University Press
YEAR: 2007

Oliver Schallert, Department of Linguistics, Philipps-Universität Marburg

SUMMARY
This book deals with syntactic gradience in the sense of categorical gradience,
which has a rather disputed status in modern linguistic theorizing: Whereas
scholars working in a formalist framework tend to even deny the existence of
such a phenomenon by assigning it an epiphenomenal status, within the
functionalist school, categorical fuzziness plays a prominent role not only as a
phenomenon but more importantly as a theoretical concept. An interesting
parallel to this debate can be found in the disputed status of
grammaticalization in both frameworks, which has drawn some attention to itself
recently, see e.g. Newmeyer (1998, 2001), Campbell (2001), Hawkins (2004),
Bisang et al. (2004). Against this background, the author's motivation for
writing this book ''grew out of a feeling of discomfort with not only the views
of (...) most linguists working in formal syntactic frameworks, but also with
those of eclectic linguists for whom anything goes, with a
'gradience-is-everywhere' perspective'' (p. 4). Thus, the aim of this book lies
in studying the phenomenon of syntactic gradience, conceived of as ''categorical
indeterminacy'', in close detail and on the empirical basis of English. To avoid
confusion, the author states that  gradience in the sense of  ''indeterminacy in
the area of acceptability/grammaticality judgements'', as explored at length by
Fanselow et al. (2006) lies outside the scope of this book. Among the problems
investigated are: What exactly is gradience? Are there ''diagnostics'' for
establishing the existence of gradience? Is gradience a grammatical phenomenon,
or merely a by-product of performance, as has recently been argued? 

Put most generally, the author's concern is to show that ''some sort of
compromise between the two positions is possible'' (ibid.), i.e. between
''analogists'' (nomothetically inclined linguists) and ''anomalists'' (who view
languages as inherently flexibly entities). However, while acknowledging the
status of gradience as ''an undeniable property of grammar'' it is nonetheless
''incumbent on linguists to eliminate gradience where it comes about as a result
of sloppy description'' (ibid.). Thus, the author's stance is to argue for a weak
form of gradience which might be called ''constrained indeterminacy''. The reason
for this lies in the fact that ''the well-motivated setting-up of discrete
categories of form classes is logically prior to claiming that gradience obtains
between them'' (p. 201).

The book is divided into three parts and comprises eight chapters: The first
part, consisting of the chapters 2-4, lays out the theoretical background by
discussing gradience in a broader linguistic and philosophical background.
Chapters 2 and 3 give an overview of the historical background of concepts such
as categorization or gradience in linguistics and philosophy. The two types of
gradience pursued in this book are presented. Chapter 4 is devoted to setting
apart gradience as laid out in the present context (i.e. as categorical
indeterminacy) from related concepts of fuzziness in linguistics, i.e. serial
relationship or markedness. The second part, consisting of chapters 5-7,
assesses the two types of gradience, ''subsective gradience'' (SG) and
''intersective gradience'' (IG), which can apply to different types of grammatical
form classes: word classes, phrases, clauses, and constructions. The third part,
which solely includes chapter 8, presents a formal model of categorical
gradience as conceived of by Aarts as well as some of its applications.

Chapter 1 (pp. 1-5) sets the stage by giving a short overview on the
philosophical and linguistic dimensions of categorization and laying out the
aims of the book. The classical problem in this respect is Eubulides of Megara's
Paradox of the Sorites (or Heap), which basically deals with the question at
which point we can call a collection of grains a heap. Whereas the Aristotelian
view holds that the categories which we use to class the phenomena in the world
around us are hard and inviolable, other opinions emphasize that we have to
recognize fluidity between taxonomic constructs. Within linguistics, this debate
can be translated as follows: On the one hand there are nomothetically inclined
linguists (= formalists), who feel that languages are orderly, rule-based
systems, on the other hand we can find those who think languages are inherently
flexible entities that are not (exclusively) susceptible to an indiscriminate
application of rules (= functionalists). By studying the phenomenon of
gradience, this book tries to find some sort of compromise between the two
positions. While the author shares with formal syntacticians ''a belief that
syntax is autonomous'' (p. 2), he is ''unhappy with their unyielding views about
categorization'' (ibid.). At the same time, he notes, the views of linguists who
regard continuous phenomena in language as given are often equally
unsatisfactory. The position to be argued for, then, is the following: Even if
some alleged cases of gradience can be dispensed with since they are merely the
result of sloppy description, gradience in the weak form, i.e. as ''constrained
indeterminacy'', is an undeniable property of grammar. Generally, two types of
category fluidity need to be distinguished: (1) SG, an intra-categorical
phenomenon which allows members of a class to display the properties of that
class to varying degrees; (2) IG, an inter-categorical phenomenon which comes
about when two form classes ''converge'' on each other. Although IG is not as
widespread as often claimed (e.g. in the literature about grammaticalization),
both types are ''grammatically real'' (p. 5).

Chapter 2 (pp. 9-33) deals with the problem of categorization in linguistics and
some of its philosophical underpinnings. In its most general sense,
categorization can be seen as a process of systematization of acquired
knowledge. ''To a considerable degree categorization is the imposition of a
meta-reality on the world which involves a good deal of idealization'' (p. 10).
As far as contemporary linguistics is concerned, no grammatical framework can do
without categories - in the sense of, e.g. one set or other of word classes and
relational categories - however they may be conceived. From antiquity onwards,
categorization in the sense of setting up an adequate system of parts of speech,
has been a central concern of grammar writing. Basically, the notion ''category''
(Gr. _kate:goria_) derives from Aristotle and originally meant nothing more than
'statement'; according to Aristotle, a particular entity can be defined by
listing a number of necessary and sufficient conditions that apply to it (Gr.
_symbebe:kóta_). Associated with this view is the all-or-none principle or the
Law of the Excluded Middle, i.e. that something must be either inside or outside
a category. This is what can be called the classical, scholastic or Aristotelian
theory of categorization, which has been of eminent influence to linguistics.
Within the realm of classical philosophy (linguistics as a part of it)
Aristotle's position falls within the realm of the analogists, which stresses
the regularities in language and its propensity for order and systematicity. The
other position is that of the anomalists, which points to the ''messiness'' of
language. It was also the Greek grammarians who coined the notion of ''part of
speech'' (_méros lógou_). Through the intermediation of the Greek and Roman
grammarians (e.g. Thrax, Dyscolus, Donatus, Priscian) as well as Renaissance
scholars like Petrus Ramus, the problem of categorization was taken up by the
emerging language sciences, where it soon took on different shapes, like e.g.
the 18th century grammarians' uncertainties of how to delineate word classes or
the 19th century Junggrammarians' conception of sound laws as being
exceptionless without exception. In the 20th century, the debate between
analogists and anomalists was rekindled: On the one hand, there are various
Generative grammars (including unification-based modes such as Head-driven
Phrase Structure Grammar or Generalized Phrase Structure Grammar) and its
predecessor, American structuralism, advocating a rather strict or ''all-or-none''
view of category membership, while on the other hand, varying degrees of
categorical fuzziness are admitted by a rather divergent anomalist party. The
latter group, divergent as it is, comprises e.g. Generative Semantics, Cognitive
Linguistics, functional-typological and discourse typological linguistics.

Chapter 3 (pp. 34-79) deals with the history of concepts such as ''vagueness'',
''indeterminacy'', and ''imprecision'' in the philosophical literature, ranging from
Aristotle and the stoics to the nineteenth and twentieth century (Frege,
Wittgenstein, Popper) and even to contemporary philosophical thought. This is
followed by a broad discussion of linguistic (mainly syntactic) approaches to
gradience, ranging from American and European Structuralism
(post-Bloomfieldians, Prague School) to the positions of Transformational
Grammar and Generative Semantics, Descriptive grammar, functional-typological
and discourse-typological linguistics, Optimality Theory, and Probability
Theory. The chapter is concluded by a preliminary definition of the two types of
gradience, i.e. SG and IG.

Chapter 4 (pp. 80-94) deals with setting apart gradience in the sense of
categorical indeterminacy from other phenomena or concepts associated with them.
The most important ones are serial relationship, syntactic mixing (mergers),
multiple analysis and reanalysis, gradience and prototype theory, gradience and
markedness theory. Serial relationship, a term coined by Quirk (1965), refers to
constructional frames in which grammatical categories (e.g. verbs) may appear.
Its relationship to gradience, however, is at best superficial, despite the fact
that in the mentioned paper, Quirk talks about ''overlapping gradience'' with
respect to the different constructions a lexical element can enter. 

Syntactic mixing (mergers) can be seen as the process of amalgamation of
phrases, constructions, and clauses, e.g. in (1a), which seems to be a mixture
of (1b) and (1c).

(1) a. It's not the actual story, or even the people, that attract me to write
about something.
b. It's not the actual story, or even the people, that attract me.
c. It's not the actual story, or even the people, that persuade me to write
about something.

Structures like (1) are referred to by Aarts as ''syntactic mergers'' (p. 85),
which are defined as ''more or less spontaneous mixing of two different
constructions'' (p. 85f.) and set apart from blends and fusions, which are said
to involve conventionalized mixings. The differences between these types are
discussed in more detail in chapter 7 under the heading ''constructional
gradience''. Suffice it to say that mergers are part of a ''cline of syntactic
integration, with fusions positioned at the highly integrated end of the
gradient (...), and anakolutha at the least integrated end (...). Blends and
mergers are positioned roughly half-way along the gradient'' (p. 192) . Although
they clearly bear some resemblances to constructional gradience, by having ''the
potential to become 'recognized constructions' of the language'' (p. 192),
mergers are regarded by Aarts as ''a distinct phenomenon from gradience'' (ibid.). 

Multiple analysis is a case of what might also be called syntactic ambiguity,
i.e. structures that can be analyzed in more than one way. One example would be
the lexical item ''into'' in the sentence ''The board looked into the recent
complaints'', which can either be analyzed as part of a so-called ''phrasal verb''
(''look into'') or as head of the Prepositional Phrase (PP) ''into the recent
complaints''. Reanalysis is used as a concept in diachronic linguistics, most
notably as one of the main mechanisms of grammaticalization. The similarities
between these notions and gradience are only superficial, since the former apply
to different structures for a certain syntactic string without implying that
these structures shade into each other. 

Prototype theory and gradience have obvious affinities, yet at the same time
there are some important differences between the two approaches: (1) Gradience
is predominantly a grammatical phenomenon, prototype theory, by contrast, has
applications beyond linguistics; (2) whereas prototype theory crucially uses
''real-world categories'' to characterize prototypes, ''the attributes of
grammatical categories are by their very nature abstract'' (p. 88); (3) ''the
extra-linguistic context plays no role in the assignment of elements to
linguistic classes'' (p. 89). Linguists who advocate prototypes often employ
semantic criteria (in a very broad sense) in their definitions. In the model of
Aarts, by contrast, ''the emphasis is on the purely syntactic characteristics of
linguistic formatives (ibid.).'' 

Markedness, as conceived of in the Chomskyan sense, can be seen as a
''categorical asymmetry'' and refers to the distinction between default exemplars,
displaying some expected property, and special exemplars, e.g. by means of
(structural) complexity or formal marking. Whereas overt case in German is the
unmarked situation, it constitutes a marked phenomenon in English, where only
pronouns carry case. In the Principles and Parameters framework, this
distinction is associated with the terms ''core'' and ''periphery''. For Chomsky &
Lasnik (1977: 430), Universal Grammar incorporates something like a ''theory of
markedness'', with systems that fall within core grammar being ''the unmarked
case'', by virtue of being ''optimal in terms of the evaluation metric. ''As
Battistella (1996: 85) points out, for many generative Grammarians working in
the Principles and Parameters framework, ''markedness'' refers to three related
concepts: (1) ''a distinction between unmarked core and marked periphery''; (2) ''a
preference structure imputed to the parameters and parameter values of core
grammar''; (3) ''a preference structure among the rules of the periphery''. Aarts
sees this as a hint that the Generativists, while envisaging a radical split
between core and periphery, still allowed for some kind of gradience within the
core and the periphery, i.e. by allowing ''degrees of markedness''. ''In this
framework, a particular phenomenon radically belongs either to the core or to
the periphery, but it may be a marked core phenomenon, or a marked peripheral
phenomenon.'' In the present context the question arises whether markedness (in
the sense of categorical asymmetry) is the same as subsective gradience. For
Aarts, the answer is clearly ''no'': ''Markedness divides categories into two, a
core and a periphery, whereas subsective gradience recognizes a gradient amongst
elements within categories. (...) In a markedness-theoretical approach it is
often assumed - or perhaps we should say pretended - that the distinction
between 'marked' and 'unmarked' is an obvious one'' (p. 93).

Chapters 5, 6 (pp. 97-163) deal with a broad discussion of examples of
Subsective and Intersective Gradience in English. Interesting examples for the
first type are modals as a subtype of verbs rather than a word class on their
own. Adjectives also show evidence for grades of prototypicality, as evidenced
by the differences between _happy_ or _utter_, the first of which displaying
much more adjective-like criteria than the latter (e.g. intensification: ''very
happy'', ''*very utter''; predicative use: ''he is happy'', ''*he is utter''). At the
end of the chapter, it is also discussed whether gradience applies within
phrases or even within clauses. ''If we establish SG for a particular word class
by establishing that the class contains prototypical and less prototypical
exemplars, does it then follow that we also have SG for the phrases or clauses
that these elements head'' (p. 122)? While claiming there are clear cases of
higher projection SG that are the result of word-level SG (e.g. the gradient
from verbless clauses to main clauses), this question is left open. IG is in a
parallel fashion, i.e. firstly between word classes and then up to the phrasal
level. Interesting examples include gradience between determinatives and
adjectives (e.g. the use of _many_ and _such_ in some constructions), adjectives
and nouns, adjectives and adverbs, verbs and adjectives (overlaps between -ing
adjectives and gerunds), etc.

Chapter 7 (pp. 164-198) is devoted to gradience on the level of whole
constructions. Of course, the twofold distinction between Subsective and
Intersective Gradience also applies to this domain, hence obtaining either
within a particular construction-type (Subsective Constructional Gradience) or
between two different syntactic constructions (Intersective Constructional
Gradience), including clauses. After a short discussion of the history of the
notion ''construction'' in linguistics and the role its plays in different models
its use for a theory of grammar is evaluated. Whereas in most formalist models,
constructions are viewed as an epiphenomenon, it plays a key role in cognitivist
as well as in constructionist frameworks. Aarts argues for a dialectic concept
of construction which solely relies on syntactic criteria, so as to bypass the
bulk of problems a semantic definition, popular in cognitivist models, is
confronted with. The only way of doing so is ''by making reference to the
distributional potential of particular constructions. The most typical exemplar
of a construction will be the one that is distributionally the most versatile''
(p. 196). Among the examples of SCG discussed are pseudoclefts in the sense of
Ross (1987) or the passive gradient;  ICG is exemplified by genitival
constructions, or by the transitivity gradient as well as the one between
complements and adjuncts. While admitting that some instances of constructional
gradience are merely a by-product of other phenomena or processes, even the more
convincing cases rely on a more precise definition of the term ''construction''.
In this respect, Aarts only gets as clear as to an approximate definition as ''a
string of elements (words, phrases, etc.) that has a distinctive patterning
which plays a role in different parts of the grammar'' (p. 170)

Chapter 8 (pp. 201-242) deals with the formalization of the concepts of SG and
IG, respectively. The central point is, counter-intuitively perhaps, that a
useful model of gradience always presupposes rigid discreetness, even if only as
an analytical point of departure: ''(...) the well-motivated setting-up of
discrete categories of form classes is logically prior to claiming that
gradience obtains between them'' (p. 201). As mentioned above, SG allows for
degrees of prototypicality within grammatical categories. Prototypicality is
measured by the morphosyntactic properties of a formative, i.e. ''the potential
and actual distributional characteristics it displays in a particular
configuration'' (p. 206). IG, on the other hand, is characterized by a certain
element partaking of the categories a and b ''by displaying characteristics of
both classes'' (p. 207). Crucially, this type of gradience refers to sets of
properties and not to the intersection of categories themselves. ''I thus exclude
the possibility of a formative in particular configuration belonging to two form
classes at the same time'' (ibid.). Of importance is also the notion of
convergence which can be measured according to this scale of syntactic properties.

Subsequently, some applications of this formal model are discussed: SG in the
adjective class, IG between verbs and nouns as exemplified by the English
gerund, IG between adjectives and prepositions in the case of _near_ and _like_
etc. Towards the end of the chapter Aarts addresses some questions concerning
the syntactic properties of the formalism he makes use of: What are the relevant
properties of a category? How are they inter-related? Does every form class have
its unique properties? Can an element belonging to a particular class converge
on at most one other word class in any one syntactic configuration? Another
important question raised by Aarts' model is what to do with examples of ''true''
hybridity, i.e. cases of IG where a certain element shows an equal amount of
properties form class a and b. Even if there are some few examples, this type of
gradience is rather the exception than the rule, the reason for this being that
''cases where the categorical scales are perfectly balanced are presumably hard
to process mentally, and hence, disfavored by language users'' (p. 233). The
remainder of the chapter is devoted to the contiguity of grammatical categories,
i.e. to discussing the question between which categories gradience can apply
(e.g. between adjectives, adverbs and determinatives, but not between verbs and
coordinators). It is important to recognize that the hypothesis that IG can only
apply between certain categories resembles the discussion of grammaticalization
clines.

EVALUATION
''Syntactical Gradience'' is a very interesting and well-balanced book, which
tries to bridge the growing gap between formal and functional approaches to
linguistics and, as such, will surely attract readers from both communities: For
formalists, it offers many new directions for resolving the schizophrenic way of
dealing with vagueness, e.g. denying the existence of this phenomenon by many
Generativists, while at the same time talking of ''prepositional complementizers''
(one of the well-taken examples of Aart's book) or postulating a massive array
of functional categories, whose existence is often motivated only on
theory-internal grounds. For functionalists, and among them especially adherents
to fuzziness, it offers the important insight that the concept of fuzzy
boundaries presupposes borders, i.e. it makes no sense to talk about categorical
overlaps without taking some time in defining the necessary and sufficient
conditions for defining such entities. To put it in a nutshell: For a formalist,
the book is formally explicit enough, for a functionalist, it offers a large
body of empirical evidence as well as it highlights some problematic aspects of
their theory design.

Although one need not agree with every detail of the analyses offered in
chapters 5-7, the evidence presented in favor of categorical gradience is
undeniable. What remains unclear, however, is the status of the concept of
''constructional gradience'' discussed in chapter 7, but this might be a matter of
the reviewer's background in formal linguistics. The formal model developed in
chapter 8 is a good example of theory-building based on sound empirical
evidence. Not only are all assumptions laid out in a straightforward way, but
also it is clear to see which data would feature as evidence against this model,
i.e. it satisfies the criterion of falsifiability, which is, to my mind, the
best mark of quality. Far from pretending to have an answer to all questions,
the author gives the discussion of problematic data and observations that go
against his model more than enough space, all of which adds to the impression of
a carefully written, compact but nonetheless thoroughgoing study on the subject
of syntactic gradience.

REFERENCES
Battistella, E. L. (1996) _The Logic of Markedness_. New York/Oxford: Oxford
University Press.

Bisang, W., Himmelmann, N. & Wiemer, B., eds. (2004) _What makes
Grammaticalization? A Look from its Fringes and its Components_. Berlin/New
York: Mouton de Gruyter. 

Campbell, L. (2001) ''What's wrong with grammaticalization?'' _Language Sciences_
23: 113-161.

Chomsky, N. & Lasnik, H. (1977) ''Filters and control''. _Linguistic Inquiry_ 8/3:
425-504.

Fanselow, G., Fery, C., Schlesewsky, and Vogel, R., eds. (2006) _Gradience in
Grammar: Generative Perspectives_. Oxford: Oxford University Press.

Hawkins, J. A. (2004) _Efficiency and Complexity in Grammars_. Oxford/ New York:
Oxford University Press.

Newmeyer, F. D. (1998). _Language Form and Language Function_. Cambridge (MA):
MIT Press.

Newmeyer, F. D. (2001) ''Deconstructing grammaticalization''. Language Sciences
23, 187-229.

Quirk, R. 1965. ''Descriptive statement and serial relationship''. _Language_
41/2: 205-217.

Ross, J. R. 1987. ''Islands and syntactic prototypes''. In _Papers from the
Twenty-third Regional Meeting of the Chicago Linguistic Society_. Need, B.,
Schiller, E., and Bosch, A. (eds.). Chicago (Il.) CLS, 309-320.


ABOUT THE REVIEWER
Oliver Schallert has a M.A. in General and Historical Linguistics. Currently he
is working as scientific assistant at the department of Germanic linguistics at
the University of Marburg. Apart from that, he is at the department of
Linguistics at the University of Salzburg writing his PhD, which aims at getting
a finer-grained picture of Germanic OV and VO languages by taking a closer look
at less clear-cut cases like Yiddish, Cimbrian, and Old/Middle High German. His
research interests include the morphosyntax of verb constructions, syntactic
change, and the syntax pragmatics interface (information structure). 

 





-----------------------------------------------------------
LINGUIST List: Vol-19-68