AUTHOR: Anderson, Stephen R. TITLE: Aspects of the Theory of Clitics SERIES: Oxford Studies in Theoretical Linguistics PUBLISHER: Oxford University Press YEAR: 2005 ISBN: 019927990X ANNOUNCED IN: http://linguistlist.org/issues/17/17-669.html
Peter M. Arkadiev, PhD, Institute of Slavic Studies, Russian Academy of Sciences, Moscow
The book under review is a result of a more than a decade's research, some of which has already been disclosed to the linguistic audience as articles (in particular Anderson 1993, 2000). Anderson's new book is probably his second most important book after the famous 'A-Morphous Morphology' (1992), and will certainly rouse no less discussion than the latter. It contains probably the most comprehensive survey of phenomena and problems associated with the notion of 'clitic' up to date, and argues for a unified and coherent formal theory of the relevant phenomena, which, in author's opinion, help to solve these problems.
The core idea of Anderson's approach to clitics can be summarized as follows: 'clitics' are prosodically deficient elements introduced into syntactic domains by rules akin to those introducing morphological material into words, and assimilated to prosodically non-deficient material by general phonological rules. Thus the component of grammar clitics are dealt with in is morphology, notably phrase-level morphology rather than word-level morphology (this position is maintained from Anderson 1992), and not syntax nor phonology, although both of these play an important role in cliticization phenomena. Anderson's approach is cast in an Optimality-Theoretic (OT) framework, at least he himself considers it to be so, because actually the theory presented in the book departs from the classical OT tradition in several important respects, some of which seem to be somewhat more substantial than Anderson's explicitly stated reluctance to use standard OT tableaux.
The scope of the book is very broad, and deals not only with almost every important kind of cliticization phenomenon and descriptive and theoretical problems it may pose, but also with more general issues, such as the role and place of the morphological component in the architecture of grammar, the tension between rule-based and constraint-based approaches to morphology, the level of complexity of clause structure and the question of whether the functional content of the clause should be represented as separate projecting heads in the syntax or rather as rich feature structure of a fairly limited number of syntactic nodes. Some problems not directly linked to clitics are also reviewed, especially the so-called Verb Second, the relation between argument structure, agreement, and 'polysynthesis', and noun incorporation. The range of languages whose data is discussed in the book is also impressive.
The Introduction briefly presents the reader to the major problems associated with the notion of 'clitic', i.e. the very definition of what a clitic is, and where in the grammar these peculiar elements should be tackled with, and outlines the main theoretical claims of the book.
Chapter 2 ('What is a Clitic?') is devoted to the rigorous definition of clitics and to the general typology of clitics. Anderson critically reviews the classical theory of Arnold Zwicky (1977), which states that clitics are linguistic elements intermediate between independent words and bound affixes and provides a three-way typology of such elements ('special clitics' vs. 'simple clitics' vs. 'bound words'), and suggests that (i) it is misleading to relate clitics to full-fledged words by phonological rules, and consequently (ii) that Zwicky's notion of 'bound word' is unnecessary. Anderson informally defines clitics as linguistic elements which (a) lack autonomous accent, and (b) are phonologically subordinated to another word via a general mechanism called 'Stray Adjunction'. Such elements may be characterized by some kind of peculiar syntactic behavior, and those which show such behavior are 'special clitics', and those which do not are 'simple clitics'. Simple clitics are exemplified by determiners and case-markers in a Wakashan language Kwakw'ala (Kwakiutl), which are adjoined to the immediately preceding word. Another case study briefly presents the well-known problems with auxiliary reduction in English. Then Anderson turns to a more formal definition of clitics, and suggests the following typology:
Phonological clitic: A linguistic element whose phonological form is deficient in that it lacks prosodic structure at the level of the (Prosodic) Word (p. 23).
Morphosyntactic clitic: A linguistic element whose position with respect to the other elements of the phrase or clause follows a distinct set of principles, separate from those of the independently motivated syntax of free elements in the language (p. 31).
Importantly, morphosyntactic clitics are usually phonological clitics as well, but that is by no means necessary (e.g., Italian 3rd person plural pronominal clitic ''loro'' is stressed but behaves exactly like the unstressed elements ''la'', ''li'' etc.).
Finally, the distinction between affixes and clitics is discussed. The properties of these two kinds of linguistic elements examined by Zwicky and Pullum (1983) are shown to follow from the assumption that they are introduced and manipulated at different levels of grammar, i.e. lexical and post-lexical phonology and morphology, respectively.
Chapter 3 ('The Phonology of Cliticization') is devoted to the development of an adequate theory of phonological clitics, which would account both for their appearance and positioning. Anderson's theory of phonological clitics is based on a well articulated prosodic structure, which includes the prosodic hierarchy (1) and the requirements in (2):
(2a) Full Interpretation: In order to be well-formed at Phonological Form (PF), phonetic content has to be incorporated into prosodic structure.
(2b) Layeredness: No category dominates a higher-level category.
(2c) Headedness: Every category directly dominates (at least) one element no more than one level below it on the Prosodic Hierarchy.
On these assumptions, Anderson proposes a four-way typology of phonological clitics, based on the level of the Prosodic Hierarchy the clitic is adjoined to its host at and on whether the host projects its own PWord level. Then he develops an OT-style system of violable constraints which determine the way Stray Adjunction (the general mechanism responsible for incorporating the prosodically deficient material into the prosodic structure) operates in a given language. This system consists at least of the following constraints (or constraint types):
Exhaustivity(Ci): Every element of category Ci is exhaustively composed of elements of category Ci-1.
NonRecursive(Ci): No element of category Ci directly dominates another instance of Ci.
Prosodic Faithfulness: Prosodic structure in the input should be preserved in the output.
How different rankings of these constraints determine the way cliticization works is exemplified on the data of different Italian dialects.
Further, Anderson develops an account of how prosodic and syntactic structures are aligned, and shows, inter alia, how such an important parameter as the positioning of the clitic with respect to its host may be accounted for in this way. Then he turns to a detailed discussion of English auxiliary reduction, and shows quite convincingly that this phenomenon can be described and explained in terms of phonological cliticization only, without recourse to syntactic operations assumed by most previous analyses, e.g. Bresnan 1978.
In Chapter 4 ('Special Clitics and their Grammar') Anderson outlines the range of problems connected with 'special' or 'morphosyntactic' clitics, the most important of which is their positioning within a phrase or a clause. First of all, Anderson justifies the very status of special clitics as a linguistic category, and then presents a typology of these clitics (p. 82):
A clitic is located a. within the DOMAIN of some syntactic constituent (head or phrase); b. by reference to the FIRST vs. LAST (syntactic or prosodic) daughter constituent of that domain; and c. PRECEDING or FOLLOWING this anchor point.
Anderson proceeds arguing that special clitics bear important similarities with affixes, such as their ability to express morphosyntactic or semantic features, common restrictions on their positioning in their respective domains, possibility of non-concatenative exponents of both affixes and clitics. Anderson suggests that special clitics are best analyzed as phrasal affixes, i.e. phonological elements introduced into phrases (and not into words) and expressing the morphosyntactic features of these phrases. Some concrete examples of such elements are explored in detail: the English possessive, Polynesian definiteness markers and Kwakw'ala pronominal clitics and determiners.
Chapter 5 ('Theories of Special Clitics') is a review of the most influential approaches to the 'second-position' phenomenon. The main problem is what kind of element may precede a second-position clitic. Anderson discusses and refutes both purely phonological and purely syntactic accounts of special clitic positioning, and concludes, partly following Bošković (2000, 2001) than both syntactic and phonological mechanisms are responsible for the phenomena in question.
In Chapter 6 ('Optimal Theory of Clitic Positioning') Anderson develops his own account of clitic placement, based on his assumption that special clitics are phrasal affixes. First of all, he shows that a rule-based theory of Anderson 1992 is less adequate than a constraint-based account couched in terms of 'Stratal OT' (Kiparsky 2000). Anderson starts with developing an account of 'classical' word-level affixation, and then extends it to the level of syntactic phrases. Here the principal role is played by the constraints NonInitial(e,D), which requires that the element ''e'' be non initial in a specified domain ''D'', and LeftMost(e,D), which requires that ''e'' must occupy the closest possible position to the beginning of its domain. Also important is the Integrity(C) constraint, which prevents certain type of constituents (e.g. words) from being interrupted by phonological material. Anderson discusses data from various Slavic and Romance languages and shows how different rankings of the relevant constraints account for the distribution and positioning of clitics in these languages. Next he turns to three cases of what was assumed in the literature to be 'endoclitics' (word internal clitics), i.e. pronominal clitics in Pashto, European Portuguese and Udi (North-East Caucasian), and shows that only the latter presents an uncontroversial example of the violation of Integrity(Word) constraint. Lastly, Anderson discusses a complex case of Tagalog special clitics, showing how the distinction between ''grammatical'' and ''derivational'' elements may be extended to clitics.
Chapter 7 ('Verb Second as Alignment') discusses the possibility to extend the analysis of second position clitics to the well known phenomenon of verb-second (V2). Anderson presents a detailed analysis of data from Icelandic, Kashmiri, Breton, and Surmiran Rumantsch, and shows that V2 and second position clitics can indeed be handled by similar constraints, i.e. NonInitinal(V,D) and LeftMost(V,D), but nevertheless this similarity is not as profound as Jacob Wackernagel claimed in his famous paper (1892). The crucial difference between clitics and (finite) verbs lies in the mechanism by which they appear in the second position: the former are introduced there by post-lexical phonological rules, whereas the latter arrive there by means of syntactic displacement. Moreover, Anderson shows that V2 phenomena in Germanic, Kashmiri and Breton, on the one hand, and those in Rumantsch, on the other, result from fairly different processes and cannot be handled by identical constraints. Thus, the V2 languages themselves do not form a homogenous class.
In Chapter 8 ('Pronominal Clitics') Anderson discusses pronominal clitics with respect to their relation to agreement and argument structure, paying the greatest attention to different restrictions on cooccurrence of clitics and coreferential phrases (the so called 'clitic doubling'). He first presents a three-way typology of agreement systems, i.e. (i) where morphological agreement simply registers certain properties of an argument, and does not supplant the overt expression of the latter, e.g. subject agreement in French; (ii) where overt agreement on the verb and full argument phrases are in complementary distribution, e.g. in Carib language Pemon (Venezuela), or where noun phrases are adjuncts rather than arguments (cf. Baker's (1995) analysis of Mohawk); (iii) where agreement and argument noun phrases are compatible, but the latter may be freely omitted, e.g. in Georgian. In order to handle such cases, Anderson proposes to regard the case of simple 'registration' of an argument on the verb as involving copying of some relevant morphosyntactic features from the argument onto the Morphosyntactic Representation of the verb. The case of Mohawk and other similar phenomena discussed by Baker, on Anderson's view involve not only copying of features but also coindexation of argument positions with the verb's Morphosyntactic Representation. Under these assumptions the fact that in the so called 'polysynthetic' languages (under Baker's definition of the term) full noun phrases cannot appear in argument positions is accounted for in rather simple Binding-theoretical terms, without recourse to Baker's stipulation that agreement markers absorb structural Case. Finally, agreement in pro-drop languages such as Georgian or Italian is described as involving obligatory copying of features coupled with optional coindexing, while that of Pemon as consisting in obligatory coindexing but optional copying. Turning to pronominal clitics, among which he discusses not only well known object clitics, but also more 'exotic' subject clitics (data again comes from Rumantsch and some North Italian dialects), Anderson argues that they present an instance of phrase-level agreement, which is subject to the same mechanisms and constraints as genuinely morphological agreement.
Chapter 9 ('Clause Structure and the Grammar of Incorporation') at first glance seems to be somewhat out of place in a book devoted to clitics. However, as Anderson notes, ''the choice between syntactic and lexical analyses of this construction and the framework developed in this book for the description of special clitics both bear on a set of central issues in the theory of grammar: the nature and scope of 'Head Movement' and the role of abstract functional categories in clause structure''. Anderson critically reviews the syntactic movement approach to noun incorporation (Sadock 1986, Baker 1988) and shows that (i) this approach falls short of several important phenomena, such as idiosyncratic semantic and phonological properties of noun incorporation, and that (ii) lexical approach can neatly account for those phenomena which used to lie in the core of the syntactic approach, i.e. the fact that the incorporated nominal usually fills a particular semantic role of the verb. Anderson uses the ''Restrict'' operation of Chung and Ladusaw (2004), which unifies the semantics of the noun with that of the verb without necessarily saturating the relevant argument position. This mechanisms accounts for the 'doubling' of incorporated nouns and for the semantico-pragmatic restrictions on such 'doubling'. Also, Anderson presents a discussion of the relations between incorporation and agreement, which are crucial for Baker, and shows that the syntactic movement analysis is not necessary to explain the fact that agreement with the incorporated noun is dispreferred in such languages as Mohawk. Similarly, he discusses the denominal verb formation in West Greenlandic, analyzed as an instance of syntactic rather than lexical operation by Sadock, and argues for a lexical account of this phenomenon. The conclusion Anderson reaches is rather far-fetching, i.e. that since the incorporation structures can be dealt with in the lexicon, the very motivation for the mechanism of Head Movement becomes shaky, as well as for the contemporary view of clause structure as involving lots of abstract functional projections.
Anderson's book is undoubtedly one of the most important and valuable contributions to the study of clitics. His core argument — that cliticization is a special type of morphological process subject to specific syntactic and phonological constraints — is indeed well motivated and thoroughly articulated. Perhaps the greatest advantage of this book is that Anderson achieves a very good balance between a data-oriented empirical approach to phenomena he discusses, covering almost every important aspect of cliticization and providing detailed analysis of data from a huge array of languages, and a rigidly articulated theoretical account of these data. Such convergence between theory and descriptive typology is not too often found in linguists' works, and this makes Anderson's book valuable not only as a study of particular phenomena, but also as an example of how linguistics should be done.
Interesting and thought-provoking contributions are, however, usually not flawless, and Anderson's book is no exception. I would like to point out two important shortcomings of this book. The first one is its somewhat too broad coverage; specifically, the last chapter, devoted to noun incorporation, seems to be out of place here, not only because its topic is very loosely related to the book's main topic, but also because Anderson's discussion of noun incorporation is clearly too brief and his main conclusion perhaps too hasty. Evidently, he has more to say about noun incorporation, and indeed more has to be said than is said in the last chapter. In my opinion, the book would have only profited from the omission of the last chapter, especially if instead of dealing with a subject unnecessary or at least clearly subsidiary in a book on clitics, Anderson had provided the reader with a concluding chapter summarizing his main arguments and ideas.
The second critical point I have to make concerns the discussion of clitics itself. One cannot blame Anderson for theoretical or methodological inaccuracy, but certainly not every important conceptual problem is adequately discussed in his book, and this may seriously undermine his arguments. Anderson advocates a morphological, more precisely, a realizational approach to clitics, coupled with an OT-style system of constraints. However, he is not explicit enough about how these two components are actually going to operate and co-operate. Anderson assumes a stratal architecture of the constraint system, which separates lexical morphology and phonology dealing with affixal morphology from post-lexical ones, where clitics are handled. Even more so, he distinguishes several levels at which clitics may appear, cf. his discussion of Tagalog in Chapter 6, where 'derivational' clitics are introduced in a way similar to contentful words, while 'inflectional' clitics arise via application of realizational rules. This may look very convincing and reasonable, but how to implement such an analysis into the formal OT architecture is not so evident. So, Anderson's remark in the Introduction that the reason for not including a single OT tableau into his book is that ''most of the constraint interactions proposed here are quite straightforward'' and his invitation to the reader to construct the lacking tableaux ''as an exercise'' sound much less encouraging after the book is read than in the beginning.
Despite the aforementioned problems, I must acknowledge again that Anderson's new book is a very valuable and really enlightening contribution to the study of clitics — and indeed, to the study of morphology and its relations with other components of grammar. It should be carefully studied by everyone who consider themselves a specialist in morphology, and if it happens that this book raises more questions than it provides answers, then the greater its value is.
Anderson, Stephen R. (1992) A-Morphous Morphology. Cambridge: Cambridge University Press.
Anderson, Stephen R. (1993) Wackernagel's revenge: Clitics, morphology and the syntax of second position. In Language 69, p. 68-98.
Anderson, Stephen R. (2000) Towards an optimal account of second position phenomena. In J. Dekkers, F. van der Leeuw, and J. van de Weijer (eds.) Optimality Theory: Phonology, Syntax and Acquisition, p. 302-333. Oxford: Oxford University Press.
Baker, Mark (1995) The Polysynthesis Parameter. Oxford: Oxford University Press.
Bošković, Željko (2000) Second position cliticisation: Syntax and/or phonology? In F. Beukema, and M. van den Dikken (eds.) Clitic Phenomena in European Languages, p. 71-119. Amsterdam: John Benjamins.
Bošković, Željko (2001) On the Nature of the Syntax-Phonology Interface. Oxford: Elsevier.
Bresnan, Joan (1978) Contraction and the Transformational Cycle. Bloomington: Indiana University Linguistics Club.
Chung, Sandra and William Ladusaw (2004) Restriction and Saturation. Cambridge, MA: MIT Press.
Kiparsky, Paul (2000) Opacity and cyclicity. In The Linguistic Review 17, p. 351-367.
Sadock, Jerrold (1986) Some notes on noun incorporation. In Language 62, p. 19-31.
Wackernagel, Jacob (1892) Über ein Gesetz der indogermanischen Wortstellung. In Indogermanische Forschungen 1, p. 333-436.
Zwicky, Arnold (1977) On Clitics. Bloomington: Indiana University Linguistics Club.
Zwicky, Arnold, and Geoffrey Pullum (1983) Cliticization vs. inflection: English n't. In Language 59, p. 502-513.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Peter M. Arkadiev, PhD in linguistics (2006), is a research fellow at the Department of typology and comparative linguistics of the Institute of Slavic studies of the Russian Academy of Sciences, Moscow. His main interest are linguistic typology with focus on event and argument structure and its formal realization, and theoretical approaches to morphology. He works mainly on Lithuanian, Adyghe and Japanese.