This book "asserts that the origin and spread of languages must be examined primarily through the time-tested techniques of linguistic analysis, rather than those of evolutionary biology" and "defends traditional practices in historical linguistics while remaining open to new techniques, including computational methods" and "will appeal to readers interested in world history and world geography."
Date: Sun, 25 Jan 2004 10:38:33 +0200 From: Phoevos Panagiotidis Subject: Subjects, Expletives, and the EPP
Svenonius, Peter, ed. (2002) Subjects, Expletives, and the EPP, Oxford University Press, Oxford Studies in Comparative Syntax.
Phoevos Panagiotidis, Cyprus College
This book is a collection of seven original contributions (plus an introduction by the editor) targeting graduate students of linguistics as well as professional syntacticians and semanticists. The eight chapters are as follows (contributors in parentheses): Introduction (P. Svenonius), The 'Que/Qui' Alternation and the Distribution of Expletives (K. T. Taraldsen), Icelandic Expletive Constructions and the Distribution of Subject Types (O. Vangsnes), Expletives, Subjects, and Topics in Finnish (A. Holmberg & U. Nikanne), The EPP in a Topic-Prominent Language (K. ��. Kiss), The Extended Projection Principle as a Condition on the Tense Dependency (I. Roberts & A. Roussou), Parameters of Subject Inflection in Italian Dialects (M. R. Manzini & L. Savoia), Subject Positions and the Placement of Adverbials (P. Svenonius).
The Introduction (Svenonius) looks at the issues pertaining to the notion of "subject" and its multi-faceted character; it then turns to the existence of expletive subjects, popularised through the study of English 'there' and 'it' in generative grammar; it continues with a discussion of the status of the Extended Projection Principle (EPP), namely the requirement that all clauses have a subject. A presentation of the individual papers in the volume concludes the chapter.
Taraldsen's "The 'Que/Qui' Alternation and the Distribution of Expletives" compares French with Vallader (a Rhaeto-Romance variety) and suggests that the 'que/qui' alternation (see (1) below) in the former boils down to the presence versus absence of an expletive subject 'i', also manifested as 'ti' in Colloquial French in cases like the one in (2):
(1) Quel livre crois-tu que /*qui les filles vont acheter? which book think-you that the girls will buy?
Quelles filles crois-tu *que /qui vont acheter ce livre-l��? which girls think-you that will buy that book-there?
(2) Pourquoi tu dois-ti partir? why you must-TI leave?
By looking into the distribution of '(t)i', Taraldsen concludes that it is a numberless, genderless expletive, hence a "pure expletive" like English 'there'. This last postulation is examined in the face of French Stylistic Inversion, where '(t)i' does not show up attached on 'que'.
Vangsnes' chapter looks into Icelandic Transitive Expletive Constructions, such as the sentence in (3) below:
(3) ��a�� hefur (einhver k��ttur) veri�� (einhver k��ttur) �� eldh��sinu there has (some cat) been (some cat) in kitchen.the "There has been some cat in the kitchen."
The expression "some cat", called the associate, can appear either in an "intermediate" position between the auxiliary 'has' and the verb, or in a "postverbal" position, following both. Vangsnes examines in detail the type of associates that can appear in each position. He establishes that, in Icelandic Transitive Expletive Constructions, partitive and universally quantified expressions only appear in the intermediate position, identified as the specifier of the Tense Phrase (SpecTP), whereas non- quantificational bare indefinites obligatorily occupy the postverbal position. Indefinites, such as "some cat" in (3), can appear in either position, despite displaying "Diesing effects": when in the intermediate position, they receive a presupposition reading. Vangsnes goes on to account for the facts by capitalising on (a) de Hoop's (1996) notion of strong and weak Case: Tense can license strong Case, hence taking quantified expressions in its specifier; (b) a distinction between "lexical" and "agreement" features, as well as the need for functional categories to be identified: this is why the expletive '��a��', conceived as deictic, appears in the specifier of the highest Agreement Projection (AgrP).
Holmberg and Nikanne explore subjecthood in Finnish, a language that can front subjects as well as objects. Moreover, Finnish permits expletives as well as Transitive Expletive Constructions. What Holmberg and Nikanne argue for is that in Finnish the presupposed argument, the topic (which is marked with a [-focus] feature), must be externalised and move to the specifier of a Finiteness Phrase (SpecFP), no matter whether it is a subject or an object. Finiteness is taken to subsume the category Agreement and differences between fronted subjects and fronted objects are claimed to boil down to the fact that Agreement is inherently nominative. Finally, the chapter shows that the Finnish expletive subject 'sita' also occupies the SpecFP position.
In a similar vein, E. Kiss draws on evidence primarily from Hungarian to reformulate EPP along the following principles (her (30), (31) and (40)):
(4) Statements express predication, or quantification, or both. (5) A sentence expressing predication must contain a topic. (6) Of the arguments of a predicate, one must be marked as a subject.
Fronted arguments occupy the specifier of a Topic Phrase (SpecTopP). English observes (6) in the form of the subject moving to SpecTP, which then moves to SpecTopP abiding by (5), whereas Hungarian does so by marking an argument as subject in the lexicon. Because SpecTP is irrelevant in Hungarian, the language lacks grammatical function changing operations, such as passivisation, 'tough'-movement, syntactic secondary predication and syntactic middles.
Roberts and Roussou propose a unification of the EPP with the typically Germanic Verb-Second (V2) property. In their chapter they advance a proposal whereby Complementiser and Tense constitute a dependency from which the time of the event the clause denotes is defined. In this dependency, functional heads must be identified (by phonological material, whether a phrase or a head). Roughly, if this head is Tense, then we derive a condition such as the EPP in both null and non-null subject languages; if this head is C, then V2 is also derived. As far as V2 is concerned, Roberts and Roussou argue against the relevance of topic features being checked at C and propose that, while the verb is in C to identify the C-T dependency, a full phrase is also needed in V2 environments in order to type the clause as a declarative, in the absence of any "rich" content of C.
Manzini and Savoia use evidence from Italian dialects to illustrate the need for multiple Agreement projections, as well as the need to eliminate pro and A-movement, in favour of the movement of aspectual / thematic features from the verbal domain to base-generated subjects (and other arguments) higher up in the clause, or to the heads they are attached to. Furthermore, expletive-associate and doubling constructions are unified as instances of movement of aspectual / thematic features; non-agreeing or partly agreeing associates are subsequently explored in the light of the above hypotheses. Interestingly, the chapter does not restrict itself to describing Italian dialects but moves on to hypothesise universal characteristics of subjects and expletive-associate configurations, providing evidence for a microparametric conception of syntactic variation.
The last chapter, by the editor himself, is roughly divided in two parts. In the first one, the question of the position of adverbs is discussed. The author argues against the influential model of Cinque (1999), where each adverb is in the specifier of a dedicated functional head. He argues instead that ordering restrictions between subjects and adverbs in Scandinavian and Italian can be better explained if adverbs are adjuncts with their relative orderings constrained by semantics. In the second part, Svenonius looks into subject positions in Germanic. After reviewing the different behaviour of the high position (the specifier of AgrP, SpecAgrP) and the lower one (SpecTP) in Germanic varieties, he analyses the observed variation along the following lines: the highest specifiers in V2 constructions preferentially host shift topics; the lower SpecAgrP hosts continuous topics and the lowest SpecTP practically everything else. Now, languages like Danish and English mark all their subjects as topics by default, hence they all land in SpecAgrP, in Swedish only non-focused subjects are marked as topics (hence focused Swedish subjects remain in SpecTP), German and Norwegian only mark real continued topics as such, shift topics move higher and everything else remains in SpecTP.
The issues this volume is dealing with have received a great deal of attention and have raised considerable controversy over the last twenty years. One then has to immediately grant that putting together a volume dealing with the life of (obligatory) subjects, their position(s) and the nature and behaviour of expletives can be neither conclusive nor, certainly, exhaustive. That much becomes evident from the Introduction Svenonius has written for the volume. The Introduction is an outstandingly clear and succinctly informative review on what subjects are (not): topics, nominative phrases and/or thematically prominent arguments. Similar praise must go to the rest of the Introduction, where expletives and EPP are briefly but solidly introduced. Before moving on, it should be noted that this introductory chapter makes a more than adequate summary of the related issues, well beyond the scope of the volume, as well, and it would also make a good preparatory text to give to advanced undergraduates.
Before proceeding to assess chapters individually, I would like to raise issues pertaining to the editing of the volume. Researchers and scholars are not editors and should not be expected to substitute for them: it is not only a matter of time or workload but also, simply, of training. Nevertheless, even international publishers such as Oxford University Press, who publish this volume, assume that, at least in our field, we can also act as unassisted editors -- and save them money (although, thankfully, editors are still used for their journals). This certainly creates problems and slows jointly authored volume projects down. Some of the errors in the volume, and I am not even an editor, include: Elena Anagnostopoulou's surname misspelled (as "Anagnastopoulou") on pp. 11, 12, 44 and 58; "examing" instead of "examining" on p. 17; references to examples (5) and (6) appearing as (6) and (7) on p. 109 and Bobaljik & Thrainsson (1998) referenced as a (1997) manuscript on p.239. Publishers of joint volumes should put the hand deep in their pockets and reconsider. To the chapters, now.
Taraldsen's chapter is dealing with a very specific issue, the nature of the 'que/qui' alteration and its relation to French Stylistic Inversion. The discussion is too condensed and would have definitely benefited from some unpacking, as sometimes the text verges on obscurity. For instance, on p. 37 the functional head F is introduced, along with a functional head F' (which further down is identified with a head X: F' does not stand for an intermediate projection). Towards the end of the same page we read: "a finite clause contains two functional heads, I and F, associated with EPP features [��^����] Similarly, two distinct functional heads, I and X, carry number features." On the next page we learn that X (i.e. F') can be identified with AgrO but no indication of the identity of F is given. This and similar problems in presentation make the argument very tricky to follow.
In Vangsnes' chapter we are presented across eleven pages (pp. 43-54) with a detailed description of the two positions an associate can occupy in Icelandic Transitive Expletive Constructions, as well as with the characteristics of each position. The exposition is clear and, in parts, very useful (e.g. the table on p. 52). One reservation here would perhaps be that such a detailed descriptive presentation would be more in place in a handbook rather than a collection of primary research papers, as this volume is. The analysis builds on Case (see above) but the discussion tends to be very telegraphic in parts, such as the point on p. 56, where the relation between tense and subjective Case is (too briefly) presented, or the original distinction on p. 61 between lexical and agreement features, where the two classes are not adequately defined: this mars the clarity of the analysis on pp. 61-64. While on the issue, this very distinction itself appears to be problematic: a) It is not clear why [Case], [deixis], and [tense] are lexical features whereas [person] and [number] are agreement features (p. 61). For instance, on what grounds is [deixis] taken to be a lexical feature and [person] an agreement one? There is actually research that the two belong to the same (class of) feature: Bloomfield (1938: 225-6), Ritter (1995: 421), Panagiotidis (2002: 29) etc. consider also the well-known fact that demonstrative systems tend to "shadow" the tripartition between "speaker", "addressee" and "other"; b) there is nothing by way of explanation of why [Case], [deixis], and [person] are all significant in identifying the Complementiser field, and no reference is provided either. As a consequence, the linking of these hypotheses to the distinction between strong and weak Case as well as the relevance of topic-hood (cf. p.62) are left dangling. Leaving these points behind, a more general observation here would be that the paper is lucid and informative in its descriptive part but tends to become impenetrable and contorted in parts of its analysis.
Holmberg and Nikanne present their case with clarity and precision. The chapter contains a brief yet informative introduction to the relevant aspects of Finnish syntax; the authors also provide us with their assumed basic clause structure for the language in the form of a tree-diagram, which is a most welcome aid. Their claim that SpecFP is the landing site for any [-Focus] argument is supported via a detailed survey of the relevant structures; if the chapter stopped here, we would be left with the conclusion that Finnish is a topic-prominent language. Nevertheless, their section 5 (pp. 86-9) truly deepens their analysis as it contrasts the interaction between the above property with the inherently nominative specification of F/Agr, hence deriving the intriguing fact that fronted objects behave as if they head a quasi-A' chain. The treatment of the Finnish expletive and that of Finnish Transitive Expletive Constructions are also enlightening. One objection must nevertheless be made before closing this short evaluation of Holmberg and Nikanne's chapter: it should have featured a conclusion rounding up the discussion and tying the various threads together, for instance the analysis of Transitive Expletive Constructions with that on "weather" subjects in Finnish.
��. Kiss puts forward a recasting of the EPP along the lines of (4), (5) and (6) above, reducing it to semantic restrictions on predication. The argumentation is coherent and the ensuing proposal is certainly interesting. Of course this is done at a cost: in eventive sentences such as (7) below (example (18c) in the chapter), we are forced to postulate an empty event argument as the filler of SpecTopP:
While there is considerable support for event arguments, especially in the more semantically-minded part of syntactic literature, the question here is whether there is any independent (cross-linguistic) evidence for it; in other words, are there any overt event arguments (cross- linguistically)? This question is important when foregrounded against the discussion of English expletives 'there' and 'it': the former is understood in this chapter to be an element that turns the verb of the clause into an existential quantifier, hence a restrictor, that becomes visible purely for Case reasons, hence it is not a "pure expletive"; the latter is seen as forming a discontinuous unit with the clause, in cases such as the following:
(8) It is believed / obvious [that ...]
No mention of the weather function of 'it' is made ('It is snowing'). In other words, the analysis works at the cost of a null eventive argument and reinterpreting the role of English expletive subjects, which are the best-studied ones. More generally, it remains for the theory ��. Kiss puts forward in this chapter to be tested against a larger number of languages, other than English and Hungarian, where it is not clear that topic-prominence entails the absence of TP, as is claimed to be the case in Hungarian (see above in the "Description" section): Finnish, as analysed by Holmberg & Nikanne, or Greek (which appears to be midway between English and Hungarian) would be some of the more accessible candidates for testing.
Roberts & Roussou present a very dense and very rich chapter that is nevertheless reader-friendly, as it contains a number of summaries and mnemonic aids such as the spelling out of the main hypotheses and generalisations as numbered items. The hypothesis itself is very intriguing: instead of the interaction between interpretable and uninterpretable features, grammatical operations are taken to be driven by the need of features and bundles thereof (i.e. functional heads) to be made visible / identified. Roberts and Roussou look into the nature of the Complementiser (field) in some detail, which is certainly something that helps lend considerable credibility to their account; moreover, their criticism of "Topic-Criterion" style analyses for V2 strikes a chord with work asserting the non-quantificational character of topic, such as Rizzi (1997), for instance. Because they analyse the fronted phrase in V2 as identifying the default declarative function of C, they can accommodate otherwise awkward cases of German V1 (p.139), like 'yes/no' questions and conditionals. Their brief excursus into Celtic remains to be expanded elsewhere.
Manzini & Savoia survey a large number of Italian dialects, which display a considerable degree of variation, in order to advance their hypothesis on the number and the nature of the four "Agreement" projections where subject clitics and full subjects attach: Determiner, Number, Nominal and Person. Building on their previous work, they elaborate on their elimination of 'pro' and its paradoxical need for syntactic licensing and identification conditions because of its PF status; this is indeed a most welcome step. On pp.167-170 they argue that in null subject languages preverbal lexical subjects are not really subjects but topics, whereas post-verbal subjects are foci. Although there is a problem with the second part of this hypothesis for null subject languages beyond Italian, I will restrict myself to pointing out that an analysis of preverbal subjects as topics in null subject languages typically encounters problems with Exceptional Case Marking, for instance in Greek where preverbal subjects can be exceptionally assigned accusative from the higher verb, something very odd if they are pure topics in A' positions:
(9) Idha ton Petro na akui musiki saw.1st the.ACC Peter.ACC SUBJUN listen.3rd music "I saw Peter listen to music."
Manzini and Savoia's postulation of four positions available for subject clitics is convincingly presented, although the exact characterisation of Nominal and Person is certainly going to benefit from further research.
The final chapter by Svenonius himself does an excellent job refuting Cinque's account of the positioning of adverbs. He is right in showing that, given that subjects can surface between almost any two higher adverbs, the number of optional subject positions is multiplied beyond necessity (twenty in Norwegian, fifteen in Italian: p.208); the following Norwegian example (his (15)) partly illustrates this state of affairs:
(10) at tydeligvis (Per) ikke (Per) lenger (Per) bestandig (Per) vinner that evidently Per not Per anymore Per always Per wins "that evidently Per not anymore always wins."
He defends the adjunction analysis, coupling it with semantic restrictions on the co-occurrence of certain adverbs. Nevertheless, this is not the topic of the volume in question. That is why he goes on to survey Germanic subjects (see the "Description" section above). However, when it is time, in section 5, to bring these two threads together, the result is nowhere as clear as the discussion on the adjunction site of adverbs. The fact that he invokes checking of topic features as one of the mechanisms at play and that he in effect collapses a Topic projection with Agreement (only to return to a unitary head Inflection on p.233) certainly does not improve matters. This chapter does make it clear that the proposals in Bobaljik & Jonas (1996) and Bobaljik & Thrainsson (1998) regarding the number of inflectional heads in Germanic varieties are maybe too coarse-grained and do not properly account for the interaction between subjects and adverbs in a clear way. At the same time, Svenonius' counter-proposal is not always clear and, in the case of topic-hood, probably not fully on the right track.
Finishing the volume I was left with the feeling that the contributions, although interesting in their majority, were not always one hundred percent in concord with the subject of "Subjects, Expletives, and the EPP". In other words, the scope of the title (as well as the Introduction) turned out to be much wider than that of the chapters: their number is too small, some of them deal with details not necessarily contributing to our understanding of the broader picture, the comparative-typological side of the individual chapters tends to remain underdeveloped. In other words, the end result gives the impression of an uneven collection of contributions, as far as their presentation, scope and rigour are concerned.
I wish to warmly thank Kleanthes K. Grohmann and Stavroula Tsiplakou for discussing this review with me. The usual disclaimers apply.
Bloomfield, Leonard 1938. 'Language'. London: Allen & Unwin.
Bobaljik, Jonathan & Jonas, Diane. 1996 Subject positions and the roles of TP. 'Linguistic Inquiry' 27: 195-236
Bobaljik, Jonathan & Thrainsson, Hoskuldur 1998. Two heads aren't always better than one. 'Syntax' 1: 37-71
Cinque, Guglielmo 1999. 'Adverbs and functional heads'. Oxford: OUP
Hoop, Helen de 1996. 'Case Configuration and Noun Phrase Interpretation'. New York: Garland
Panagiotidis, Phoevos 2002. 'Pronouns, Clitics and Empty Nouns'. Amsterdam: John Benjamins.
Ritter, Elizabeth 1995. On the syntactic category of pronouns and agreement. 'Natural Language and Linguistic Theory' 13: 405-443
Rizzi, Luigi 1997. The fine structure of the left periphery. In: Haegeman, Liliane (ed.) 'Elements of Grammar'. Dordrecht: Kluwer. 281-337
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Phoevos Panagiotidis is an Assistant Professor of Linguistics in Cyprus College, Cyprus. He is the author of the monograph titled "Pronouns, Clitics and Empty Nouns" (2002: Benjamins). He has also published articles in international journals (Lingua, NLLT, Linguistic Inquiry) and jointly authored volumes on pronouns, properties of Determiner Phrases and the status of arguments in null subject languages. Besides the above, his research interests include the nature of grammatical categories, language acquisition and breakdown, as well as the structure of English, Greek and the languages of the Balkan Sprachbund.