Date: Wed, 22 Jun 2005 12:13:01 +0200 (CEST) From: Ferid Chekili Subject: The Oxford Handbook of Comparative Syntax
EDITOR: Cinque, Guglielmo; Kayne, Richard S. TITLE: The Oxford Handbook of Comparative Syntax SERIES: Oxford Handbooks PUBLISHER: Oxford University Press YEAR: 2005
Ferid Chekili, Department of English, Faculte des Lettres, University of Manouba, Tunisia.
This volume is an edited collection of twenty-one essays written within the framework of the Principles-and-Parameters (P&P)conception of universal grammar (UG), and representing one area of syntactic investigation known as Comparative Syntax. As explained in the 'Preface', comparative syntax, not only "attempts to deepen our understanding of the 'parameters' side of the human language faculty"... but also..."provides us with a new and highly promising tool with which to deepen our understanding of the 'principles' side, the invariant core, of the human language faculty".
Chapter 1: Some Notes on Comparative Syntax, with Special Reference to English and French, by Richard S. Kayne. The author provides an analysis of certain French and English constructions based on the "nonpronunciation" of certain elements. For Kayne, an important question _to do with the existence of overt elements in one language and their absence in another _ is to determine whether differences between languages may result from an "irreducible parametric property" of a given element or be attributed to independent factors (p.37).
He develops a conception of parametrisation whereby the syntactic differences between English and French depend on what can or cannot be pronounced. For example, despite the directness of the relation between 'Marie a mange peu de sucre' and 'Mary has eaten little sugar', he proposes that 'little' is an adjective, whereas 'peu' a noun. Further, he shows that 'little' corresponds to 'petit' and 'peu' to 'bit'. The differences between the two languages here, depend on what can or cannot be pronounced (pp.43-44).
Kayne also outlines the aims and methods of comparative syntax, including the requirement that any proper syntactic analysis of a given language must take into account other languages (cf. p.50).
Chapter 2: On the Grammatical basis of Language Development: A Case Study, by Luigi Rizzi. Rizzi addresses the issue of continuity in language development through the comparative study of early and adult grammatical systems, more specifically, the parametric approach to comparative syntax. The case study which he uses to illustrate this link between parameters and development is the analysis of subject drop in early linguistic production.
An important question addressed is the following: why should certain parameters -- such as subject drop -- differ from others -e.g. word order parameters -- in not obeying Wexler's (1998) Very Early Parameter Setting and in manifesting developmental effects? In fact, he shows that the 'early subject drop' phenomenon is a distinct phenomenon from the null subject of null subject languages.
He provides a functional explanation to early subject drop but argues for the grammatical status of the phenomenon -- what he calls "Privilege of the Spec of the Root" -- which also appears in certain adult systems. In his own words, "language development is grammatically based but performance driven".
He provides a parametric analysis of the "privilege of the root" using the approach to the left periphery in Rizzi (1997,2004). He calls the parameter 'the truncation parameter' and applies it to the development of early subject drop.
Chapter 3: Comparative Syntax and Language Disorders, by Arhonto Terzi. Terzi examines the link between comparative syntax and language disorder in terms of the contributions of each to the study of the other. Clitic placement in Greek is often used here to illustrate the objectives above.
Although, he argues, impaired varieties may not contribute much to the study of normal language (as they are different from dialects), there is a sense in which we can talk of a contribution: "if a certain phenomenon has been offered a similar account cross-linguistically for normal language, one expects that its impaired ... variant should manifest similar behaviour as well across the same languages. If not, the account offered on the basis of normal adult language may have to be reexamined".
Also, impaired language can shed light on certain aspects of development (that early language cannot make clear because of the faster pace of development it shows)and ultimately, offer insights into the syntactic structure of normal language.
Concerning the other side of the relationship, i.e. the contributions of comparative syntax to language disorder, he reports on the discovery, in English-type languages, of a "clinical marker" the Extended Optional Infinitive (EOI) which is based on the Unique Checking Constraint, and shows that it does not apply to Greek. However, there may be a counterpart of EOI in Greek (a language without infinitives), namely, 'auxiliary omission' in Standard Greek and (in view of the absence of auxiliaries in Cypriot Greek) the 'misplacement of pronominal clitics' in the latter language.
Chapter 4: Object Shift, Verb Movement, and Verb reduplication, by Enoch Olade Aboh. The author argues that object shift and verb movement are obligatory in the Gbe languages, motivated by case or strong Extended Projection Principle (EPP) features and aspect features respectively. The differences between verb-object (VO) and object-verb (OV) can be accounted for in terms of the interaction between object shift and verb movement on the one hand, and differences in clause structure on the other. Differences between monoclausal VO structures and biclausal OV structures, together with the attendant movement possibilities, can account for the asymmetries. Further, Reduplication is argued to be a last resort phenomenon whereby the verb reduplicates in order to license a null expletive.
Chapter 5: Finiteness and Negation in Dravidian, by R. Amritavalli and K. A Jayaseelan. The Dravidian languages investigated here, include Kannada and Malayalam. This paper argues against the universality of 'Finiteness-is-Tense', as finiteness in Dravidian is not a property of a tense category. Instead, finiteness is argued to be constituted by the presence of a Mood Phrase which contains agreement morphology and therefore, explains the correlation of agreement with finiteness.
Comparative syntax is at play in this essay on two levels: first, the Dravidian languages are compared with European languages, and second, with one another. An example of the former: whereas in Dravidian languages, finiteness and tense information are realized in different projections -- a Mood Phrase and a Asp Phrase respectively -- in English-type languages, there is a single head Tense realizing the two classes of features: a parametric difference in terms of features. An example of the latter: starting with the observation that negative clauses in Kannada and Malayalam are superficially different, they in fact show -- on the basis of different complements of negation in different Dravidian languages -- that the Kannada negative clause is a "principled fact" of the whole language family.
Chapter 6: On Some Descriptive Generalisations in Romance, by Paola Beninca and Cecilia Poletto. Clitic wh-elements, pronominal clitics, and negative clitics in Romance are considered with respect to their theoretical relevance for syntactic theory. The aim, as stated by the authors, is to show "how cross- linguistic variation can direct our research toward a precise path and thereby narrow down the number of possible analyses of a given phenomenon".
The paper shows how the development of clitic forms for wh-items, pronouns and negation is influenced by certain factors: It is sensitive to both semantic factors (a semantically poor element tends to become a clitic) and syntactic ones (clitics are sensitive to verb movement).
The authors use Chomsky's concept of feature movement (Chomsky 1995) by proposing that clitic forms are the overt counterpart of the mechanism of feature movement at Logical Form (LF): the clitic checks the features of an element in a higher projection leaving the complex category behind. They also suggest that the development of clitics is tied to the loss of movement of the independent category.
Chapter 7. Classifiers in Four Varieties of Chinese, by Lisa L.-S.Cheng and Rint Sybesma. The use of classifiers in four varieties of Chinese is considered with the aim of accounting for the "distribution and interpretation of nominal expressions in these languages" and their variations which are handled by means of parameters. Certain form-interpretation correlations between the nominal expressions in different varieties are identified.
It is argued that a functional head (in the nominal domain) performs the deictic function. In languages with determiners, this function is performed by Determiner (D); in Chinese, by the classifier. Variation between the different varieties , in terms of definiteness, is accounted for by postulating different strctures and different mechanisms.
Chapter 8: Morphology and Word Order in "Creolization" and beyond, by Michel Degraff. Certain VP-related properties in Haitian Creole (HC) (mainly verb and object placement) are compared with some of HC's source languages (e.g. French and Gbe dialects).
The question of Creole "Exceptionalism" is also dealt with, with the conclusion that, from a Universal Grammar/ mentalist perspective, there is no difference, in terms of mental processes, between Creole and non Creole languages. Starting with the observation that the major sources of HC have certain word order and morphosyntactic properties not found in HC, the author argues that this does not necessarily lead to Creole Exceptionalism.
Using what he calls a "verb placement" parameter, he argues that HC is V- in-situ while French is V-to-I. This parameter is triggered by verbal inflectional morphology which is "rich" in certain languages but "poor" in others. He also argues that certain VP-related "discrepancies" in Germanic and Romance diachrony are as 'significant' as in the diachrony of HC. This, he says, weakens any empirical basis for Creole Exceptionalism.
Chapter 9: The Slavic Languages, by Steven Franks. A number of issues in Slavic comparative morphosyntax are examined, including case and agreement, the genitive of negation, numerals, argument structure and voice, clitics, wh-movement, negation, binding, aspect, and word order. The aim is to highlight the crosslinguistic variation and to sketch proposals for analyzing the phenomena. The belief is that Slavic morphosyntactic properties contribute to understanding the cross- linguistic diversity.
Chapter 10: The Scandinavian Languages, by Anders Holmberg and Christer Platzack. After highlighting certain similarities between Scandinavian languages, the authors turn to cases of syntactic variation including word order variation within the sentence and the noun phrase.
The variations are accounted for using the proposals of the Minimalist Program (MP), in particular, the fact that structural variation results from an interaction between Universal (UG) principles and language particular lexical properties.
The authors show that there is more variation between the Scandinavian languages within the noun phrase than the sentence. Noun phrase syntax "does not group the Scandinavian languages in the same way as sentential syntax".
Chapter 11: Noun Class, Gender, and the Lexicon-Syntax-Morphology Interfaces: A Comparative Study of Niger-Congo and Romance Languages, by Alain Kihm. This essay is concerned with the grammatical function of "class". "Class" is argued to be "a particularly apt topic for comparative grammar, precisely because of its uncertain status as far as universality is concerned".
Romance class systems include gender which consists in two or three genders; Niger Congo languages have noun class or multiple gender systems.
The aim is to compare these two systems in light of certain empirical generalizations (e.g. class exponents of gender systems are suffixes, whereas those of noun class systems are suffixes or prefixes; and they are always combined (in the two systems) with number morphemes when overt.
The discussion is conducted within the framework of Distributed Morphology.
They propose that the grammatical function of noun classes is distinct from classification as such. It is, instead, noun formation. Manjaku and Romance share the category "class", i.e. the formation of nouns from roots. However, they differ in that the lexical expressions of class are meaningful roots in Manjaku but semantically empty functional items in Romance.
Chapter 12: Agreement and its Placement in Turkic Nonsubject Relative Clauses, by Jaklin Kornfilt. The author concentrates on a productive type of head-final relative clause (RC) with a nonsubject target. Such RC's are of three types. To account for the common as well as different properties of the three types, Kayne's (1994) approach is used.
The author shows that a Kaynean derivation is able to account for the microparametric variation -the three RC types -- among the Turkic languages (e.g. Turkish, Uzbek, Azeri), by proposing, first, to move the target NP to [Spec,CP], and second, move the remnant IP to Spec of the large DP. Variation between the languages is due to the nature of the moving IP: AgrP for type one, and an Agr-less, bare nominal for types two and three. In other words, the differences among the three types are attributed to 1. a different status of Agr (if it occurs): either as a syntactic head (type one) or as a clitic (type three) 2. the moving IP.
Chapter 13: Qu'est-ce-que (qu)-est-ce-que? A Case Study in Comparative Romance Interrogative Syntax, by Nicola Munaro and Jean-Yves Pollock. The authors' aim is to show that the distribution and constraints on the occurence of wh-est-ce-q questions in Romance follow from UG and the invariant left periphery of Romance questions.
They argue that there is a difference between bare wh-words and complex wh- phrases in Romance. Bare wh-words check the uninterpretable features of two left periphery positions (an existential position and a disjunction position); complex wh-phrases only check the second feature. Bare wh-words may lexicalize the existential quantifier (e.g. Bellunese), the disjunction operator (e.g. French, Italian), or both (e.g. Bellunese).
It is also argued that French 'Que' is a clitic -- a wh-clitic. This means that 'que' can reach its target position in CP only as a 'free rider' to Force Phrase. This explains the ban on 'que' extraction from subject position. The facts of Northern Italian dialects can also be explained if we posit a null clitic with similar behaviour to 'que'.
Chapter 14: Clitic Placement, Grammaticalization, and Reanalysis in Berber, by Jamal Ouhalla. He argues that in all Berber varieties, clitics are attracted (to the preverbal position) by functional categories, never by lexical categories. Apparent variation -- such as cases where clitics appear after the verb in the presence of overt T or C -- is argued to be determined by processes of language change, i.e. grammaticalization and reanalysis.
Clitic placement -following traditional accounts -- is said to be determined by factors that are partly syntactic -- movement to a functional category -- and partly prosodic -- accounting for the ban on first-position clitics.
Chapter 15: Clitic Placement in Western Iberian: A Minimalist View, by Eduardo P. Raposo and Juan Uriagereka. The idea that prosodic factors are involved in clitic placement is shared by both Ouhalla and the present authors who are mostly concerned with European Portuguese and Galician.
They argue that the complex pattern of Western Iberian (WI) clitic placement (as opposed to the more homogeneous pattern in Central/Eastern Iberian variants) can be accounted for by a morphophonological property of the functional category F located in the periphery of the clause.
The position of Determiner clitics is discussed and accounted for in terms of a Phonetic Form approach, together with a distinction between two operations, namely, fusion and clitic placement. These clitics must "fuse" with an appropriate phonological host; clitic placement is understood as a "complex and costly route whereby a determiner clitic reaches its fusion host if it cannot find it in its immediate adjacency".
Finally, a number of questions relevant to the MP are addressed by the authors.
Chapter 16: Comparative Athapaskan Syntax: Arguments and Projections, by Keren Rice and Leslie Saxon. The authors investigate basic clause structure as revealed by a study of pronouns and word order.
There is evidence, they claim, for the existence of three subject positions in these languages (around forty languages spread over North America). There are also two positions for direct objects: as the complement to the verb, and as the specifier of AgrOP.
In sum, they argue for the functional projections AgrO, NumP, and AgrS, basing their arguments on evidence from the properties of idioms and incorporation, word order... .
Chapter 17: Number Agreement Variation in Catalan Dialects, by Gemma Rigau. The author's concern is primarily with Central Catalan and Ribagorcan Catalan. The aim is to show how two main varieties of Catalan can be distinguished as a result of an interaction between certain parametric properties and UG principles.
Tense (T)is argued to be the locus of the parametric variation in number agreement in existential constructions. In Central Catalan, the phi- features of T are determined separately, whereas in Ribagorcan Catalan, they are detremined simultaneously. Furthermore, the fact that in Central Catalan, T and the object D agree even when D does not erase the EPP feature in T, constitutes evidence for 'Agree' as an independent operation satisfying the EPP.
Chapter 18: Classifiers and DP Structure in Southeast Asia, by Andrew Simpson. The essay's focus is on how to account for the cross-linguistic variation with respect to the ordering of constituents in Determiner Phrases (DP's). The aim is to see whether there are any principles governing the internal structure of DP's in these languages (primarily, Thai, Khmer, Burmese, Malay, Vietnamese...) or whether the attested patterns are arbitrary.
He adduces some evidence suggesting that numerals and classifiers project two distinct positions each associated with a distinct semantic function.
Whether NP-internal NP-movement is obligatory or not, is argued to account for the word order variation. Such movement is argued to be the result of the reanalysis of an earlier adverbial form.
In general, variation is explained in terms of elements being in different stages of historical development and reanalysis. The classifier is said to contribute to understanding this variation.
Finally, it is suggested that DP's in these languages, may share the same underlying structure.
Chapter 19: The Celtic Languages, by Maggie Tallerman. A number of issues of theoretical and empirical significance are surveyed with respect to word order, the position of the subject (whether in VP, or IP, or -- as a result of the Split-Infl hypothesis -- neither VP nor IP), V2 Syntax, copular constructions, Agreement.
She concludes that the syntax of the Celtic languages (Welsh, Breton, Irish, Scottish Gaelic, as well as two extinct languages, namely, Cornish and Manx) differs significantly. They share some properties with other VSO languages. However, their analysis can be integrated with that of "more familiar" SVO languages and hence, they may be analyzed with existing syntactic tools.
Chapter 20: Preverbal Elements in Korean and Japanese, by John Whitman. The author argues that negation, in V-final languages, may be realized either as a specifier or a head: under Kayne's (1994) antisymmetric account, the SONegV pattern has neg in Spec (like French 'pas'), whereas the SOVNeg pattern has neg in head position; Korean SOVNeg results if the negator is a free morpheme, blocking raising to the lexical verb; the Japanese SOVNeg pattern results if the negator is a bound morpheme, hosting the raised verb.
The preverbal SONegV pattern is not accounted for under a (symmetric) base- generated left-branching analysis as the position of the preverbal element (before the verb) and its scope are unexpected. This points, he argues, to a view of V-Infl order as a derived order.
Chapter 21: Continental West-Germanic Languages, by Jan-Wouter Zwart. This is largely a survey of the literature. The languages of the family include High German, Dutch, Frisian, Luxembourgeois, Africaans, ... with Dutch used to illustrate these languages' syntax.
The author shows that, although the clause-final position of the verb may suggest that these languages are SOV, certain considerations point to these languages being head-initial instead. They will have undergone a number of displacement processes such as object movement and verb movement, resulting in SOV.
The syntactic areas which are surveyed include word order, word classes, grammatical functions, types of complementation, pronouns, displacements, noun phrase structure, negation and coordination and ellipsis.
All the papers in this volume are of a very high standard. Although most papers are not accessible to the general reader and would require some knowledge of syntactic theory (in particular of the P&P/MP type), the lengthy overviews and summaries that they contain make them fairly easy to read.
Most papers present novel data and offer new insights leading to novel analyses; some are mostly reviews of the existing literature (e.g. chapters 9 and 19).
Most papers -- except a few which are theoretically more neutral -- are written within the framework of the P&P theory and make use of major developments of the MP including the structure of the left periphery, the concept of the phase... . All papers, explicitly or implicitly, accept the idea that a comparative approach helps understand UG.
Among the areas covered in the book figure prominently morphosyntactic properties such as clitic placement, and certain concepts that are argued to be responsible for surface variation such as prosodic factors, grammaticalization, reanalysis... .
Some problematic areas include:
A huge number of speculations, conjecture, open questions, and acknowledged weaknesses of the analyses can be found almost in all the papers. This is not necessarily a weakness and can be taken to reflect the still developing nature of the field.
At least one analysis presented here was proposed elsewhere: Rizzi's explanation according to which a grammatical option may be used to alleviate a performance problem can be found in Chekili (2005) in connection with Tunisian Arabic.
Some problems of an empirical nature include, for example, the following: in chapter 1, the (very insightful) analysis of French 'peu' -which is argued to be a noun -- (pp.39-44) crucially depends on the assumption that English nominals -- unlike their French counterparts -- require a determiner. More generally, although the analysis is able to capture the distinction between English and French in terms of a single parameter, it crucially depends on the assumption that English lacks nonargument nominals of the type we find in French. However, as mentioned in Radford (1997), such nominals exist in English and can be exemplified by 'John is head of department'. In chapter 6, although the analysis is intuitively correct in accounting for the difference between French and the dialect of S. Anna, it is less so when it comes to explaining the optionality attested in Paduan (examples (37)-(38)) as it is not clear why and under what conditions the same phonological element 'No' can both be a clitic and a nonclitic head. In chapter 14, although Ouhalla is concerned primarily with Berber, he advances the possibility that clitics are attached to functional (versus lexical) categories in all languages. It is not clear how this is the case in certain languages like Arabic (e.g. Tunisian Arabic --TA) where a complementizer such as 'illi' -- unlike his examples in (11a,b) -- is clearly not a lexical category. Hence, it is not clear how such cases may be accounted for in terms of grammaticalization and reanalysis. Instead of saying that the Berber equivalent of 'that' is not a functional category, it could have been equally plausible to adopt the 'checking theory' solution according to which a (DO) clitic checks its case and phi-features against a light v (also a functional category). Although this checking theory solution may run into problems elsewhere, the data here (11), as far as I can see, do not pose any problems to this solution.
Finally, one noticeable absence in this collection of essays, is Arabic. A lot of work within the P&P/ MP framework has recently been conducted on the syntax of the Arabic dialects. This would certainly have contributed to the book and the field of comparative syntax, more generally.
Chekili, F. (2005) Introduction to Generative Syntax. Publications de la Faculté de la Manouba (in conjunction with CPU), Tunis.
Chomsky, N. (1995) The Minimalist Program. MIT Press.
Kayne, R. (1994) The Antisymmetry of Syntax. MIT Press.
Radford, A. (1997) Syntactic Theory and the Structure of English: A Minimalist Approach. Cambridge Textbooks in Linguistics. Cambridge University Press.
Rizzi, L. (1997)The Fine Structure of the Left Periphery, in L. Haegeman, (ed.) Elements of Grammar, 281-338. Kluwer.
Rizzi, L (2004) Locality and Left Periphery, in Belletti (ed.) Structures and Beyond. Oxford University Press.
Wexler, K. (1998) Very Early Parameter Setting and the Unique Checking Constraint: A new Explanation of the Optional Infinitive Stage. Lingua 106:23-79.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
The reviewer is professor of English and Linguistics at the University of Manouba, Faculte des Lettres, Tunisia. He is currently employed by Nizwa College of Education, Oman. His research interests lie in the areas of syntax, morphology, comparative syntax (English, French, Arabic), and second language acquisition.