The Syntax of Local Processes

Collected Essays by Joseph E. Emonds, Volume I



1 Preface: A Reader's Guide

7 Alternatives to Global Derivational Constraints (1973) Glossa 7. (109-138)

37 Arguments for Assigning Tense Meanings after Certain Syntactic Transformations Apply (1974) In Formal Semantics. E. Keenan, ed. Cambridge University Press, Cambridge. (351-372)

59 The Verbal Complex V'-V in French (1978) Linguistic Inquiry 9, Number 2. (151-175)

85 Appositive Relatives Have No Properties (1979) Linguistic Inquiry 10, Number 2. (211-243)

119 Word Order in Generative Grammar (1979) In Explorations in Linguistics: Papers in Honor of Kazuko Inoue. G. Bedell, E. Kobayashi, and M. Muraki, eds. Kenkyusha Press, Tokyo. (58-88)

151 The Prepositional Copula as (1984) Linguistic Analysis 13, Number 2. (127-144)

169 Generalized NP-@ Inversion: Hallmark of English (1986) In Twentieth Anniversary Volume. Indiana University Linguistics Club, Bloomington, IN. (Revised English version of Inversion généralisée NP-Alpha: marque distinctive de l'anglais (1980) Langages 60: Syntaxe générative et syntaxe comparée. Alain Rouveret, ed. (1-65)

235 Grammatically Deviant Prestige Constructions (1986) In A Festschrift for Sol Saporta. M. Brame, H. Contreras, F.J. Newmeyer, eds. Noit Amrofer Press, Seattle, WA. (93-129)

273 Parts of Speech in Generative Grammar (1987) Linguistic Analysis 17, Number 1-2. (3-42)

Preface: A Reader's Guide to The Syntax of Local Processes

The purpose of publishing this volume is to make available a collection of articles around a unified theme, articles that have become almost unavailable due to the premium on novelty in the publishing industry and the apparent shrinking of older library stocks.

In retrospect, many of these articles' titles don't indicate the central empirical content or the current theoretical interest of the topics treated. This is because most of these articles were written to support theoretical positions taken for granted today or to combat views that have lost currency or credibility; the titles often refer to these now past debates rather than the nature of the evidence presented. Nonetheless, not only are the paradigms presented in these studies of interest in themselves, they support analyses for a variety of constructions that are arguably still defensible. And in many places, these articles contain some first formulations of arguments and claims still widely in use. As far as any individual, article-specific theoretical issues are concerned, I let the titles and the introductory sections speak for themselves. A reader interested in the history of a particular theoretical controversy can simply use the title of an appropriate article as a first approximation to the content.

The still current interesting general theme of these articles is encapsulated in the volume's title. This focus on local syntactic processes developed as follows. As I wrote final drafts of chapters of my 1976 book, I came to realize more and more that the dichotomy of transformational operations which organized my doctoral dissertation (Root and Structure-Preserving Transformations) needed to be supplemented by a third type of transformational rule. Hence a class of "local transformations" appearing in that book's title were treated in detail in its final chapter as well as in my simultaneously appearing UCLA Working Paper on local transformations in English and French.

These transformations differed from the Root and Structure-Preserving operations which effected movement, deletion, or copying of X0 and XP over strings of variable length. For example, WH-movement is described in note 10 of Chapter 5 of Emonds (1976) as simply moving the feature WH to COMP, with universal principles or language-particular lexical properties accounting for all its other characteristics. For example, the +WH value of the COMP at the landing site is due to structure-preservation. (The definition of structure-preservation in all these works is more restrictive than the weakened concept in N. Chomsky's Barriers; a target or "landing site" node must be of the same category as a moved XP and only substitution not adjunction, is allowed).

In contrast, local rules excluded string variables and required categories mentioned in a transformation to be adjacent. While non-local movement and copying seemed to affect all members of categories and to have other properties which could in large part be characterized by considerations arguably part of universal grammar, local rules were language-particular and affected not all members of a category X0 or XP but rather small subsets or only one member (i.e. "designated elements") of an X0 category. Finally, they could apply in dependent clauses and not preserve structure.

These properties taken together constituted what I called local transformational operations, and they seemed to me, until I published a second book A Unified Theory of Syntactic Categories in late 1984, the best way to characterize language-particular syntax. That is, when Chomsky's "Principles and Parameters" model came to be widely adopted, one way to formalize "language-particular parameters" was as local transformations. In fact, this idea contributed to H. Borer's more restrictive formulation in her Parametric Syntax, where she proposed that language-particular rules were not just local transformations but in fact lexical (or "inflectional") rules which characterize what I had taken to be the designated elements which underwent local transformations.

I don't want to give the impression that local transformations mapping trees into trees still seem to be the best way to express language-particular syntax. The future of scientific linguistics seems rather to lie in properly formalizing the syntactic contexts and derivational insertion level for every closed class lexical item. While such distributional statements for lexical items are in a sense transformational and will doubtless reflect the locality of the various transformations in this volume, I think all language-particular syntax needs to be recast in a format explicitly designed for lexical insertion.

Each article in this volume happens to have an empirical focus on one or more local syntactic processes. Though formulated in terms of local transformations, the properties and paradigms discussed remain relevant for understanding how the individual members of functional categories can best be lexically characterized. Often, the very locality of these processes is at issue and is established by the paradigms and arguments presented. These essays thus constitute a relatively rich source of (principally English) data for supporting a range of formalized and predictive grammatical analyses. Using original pagination, I point out in the following paragraphs for each article which constructions are treated in greatest empirical detail and indicate what I think is currently most useful and relevant.

1. Alternatives to Global Derivational Constraints (1973). On pp. 109-122, the English "Double-ing" constraint discovered by Ross is shown to be a string-adjacent local surface condition on V-ing-V-ing sequences not interrupted by an NP boundary. That is, the second V-ing in excluded sequences heads an XP which is not an NP. The tests proposed for post-verbal NP status involve post-verbal particles, passivization (due to P. Rosenbaum), cleft sentences, and coordination. Thus, the article provides a range of evidence showing that the complements to temporal aspect "semi-auxiliaries" are not NPs, contra Ross's claim.

2. Arguments for Assigning Tense Meanings after Certain Syntactic Transformations Apply (1974). Section 1 of the essay reviews and extends paradigms which show that the English perfect periphrastic (have - V - en ...) construction can include past tense adverbs if and only if a syntactic past tense with past meaning is not available (as in non-finite constructions, after modals, and in counterfactuals). Section 4 shows that this co-occurrence can be parsimoniously expressed only at a derivational level subsequent to transformations, in what we today call Logical Form ("... if the information in these lexical entries is inserted in the tree after the rules producing non-finite clauses including TENSE-deletion apply", p. 365).

3. The Verbal Complex V'-V in French (1978). Sections 7 and 8 are the origin of the widely accepted "V to I Raising" of French finite verbs. It may be noted that the arguments here preclude raising in French infinitives or in English and do not extend in a natural way to finite verbs in other Romance languages. The article defends other strict correlations between syntactic structure and morphology; for example, section 3 matches paradigms with structures distinguishing English present and passive participles. Section 4 defines the difference between grammatical (closed class) and lexical (open class) verbs.

The central construct of the essay is a verbal complex V' which unites grammatical verbs with a following lexical head. In the light of criticisms of V' in K. Zagona's Verb Phrase Syntax, my recent work on Romance clitics defines a version of an "equal status" (or sisterhood) hypothesis for certain sequences of grammatical + lexical verbs, answering the criticisms in Table 1 in this article. In particular, I now view clitic placements as late lexical insertions of verbal prefixes which license ("alternatively realize") empty or doubled phrases within the same VP.

4. Appositive Relatives have no Properties (1979). The analysis systematizes and extends Ross's observations in Constraints on Variables in Syntax that appositive relatives have main clause properties. Thirteen converging paradigms are explained by deriving appositive relatives in the same way as are parentheticals in Emonds (1976, Ch. 2): Clauses set off by commas signal immediate domination by root nodes, and what follows a parenthetical has been moved rightwards over it. Ross's earlier hypothesis that only single constituents move then predicts the empirical discovery of this article that post-parenthetical strings always constitute a single phrase.

The "MCH" (Main Clause Hypothesis) deconstructs many persistent unexamined assumptions: that relative clauses must form constituents with immediately preceding antecedents, that any NP accepts an appositive relative, that rightward phrasal movements are limited to extrapositions (or don't exist), and that at some level relative clauses are always subordinate. Moreover, the analysis shows that the local licensing of a WH-pronoun by an (adjacent) antecedent, like that of other free pronouns, does not require c-command. Readability is somewhat compromised by an editorially imposed, meandering introductory defense against many possible objections (pp. 214--221, up to "We see then that...."), which perhaps can be quickly skimmed or skipped.

5. Word Order in Generative Grammar (1979). Section 1 provides the first arguments for deriving so-called verb-initial or "VSO(X)" languages by a local transformation raising V from within NP [VP V...] to a pre-subject position. Section 2 shows how, except for "stylistic reorderings" (e.g. so-called free word order), a constrained theory of movement allows only this permutation of the base orders SVO(X) and S(X)OV. Section 3 uses Anderson and Chung's (1976) VP-topicalization in Breton to confirm its underlying SVO(X) order (Celtic characteristically exhibits VSOX). Section 4 and Appendix I then argue that several of Greenberg's (1963) word order universals strongly suggest that all VSO languages, despite other differences, result from V-raising out of a head-initial VP into COMP(lementizer).

Warning: the article defines a class of "normal" languages N as those whose VP is distinct from S. Serious confusion would result from glossing this N as "noun".

6. The Prepositional Copula as (1984). One theme distinguishing generative from traditional grammar is that members of the generative category P select a range of complement types akin to those of verbs (e.g. Emonds, 1972; Jackendoff, 1973; van Riemsdijk, 1978). (Traditional grammar's genesis in classical morphology restricts P to a set of case-assigners.) The cited works show that classes of lexical P can also select PPs, clauses (so-called "subordinate conjunctions"), or no complement at all.

Section 1 of this essay provides seven empirical tests showing that an NP or AP complement of non-comparative preposition as is a predicate attribute. These tests equally well demonstrate how any predicate nominals differ from direct objects. Section 2 shows how six further tests confirm that phrases headed by the non-comparative as are structurally PPs. A head P, like a head V, can thus be a "copula" which cannot assign case. This supports a generative perspective in which P's structural role goes well beyond assigning case.

7. Generalized NP-@ Inversion: Hallmark of English (1986). Revised English version of Inversion généralisée NP-Alpha: marque distinctive de l'anglais (1980). This essay (section 1) aims to establish that a single local transformational permutation characterizes English-particular syntax, if one abstracts away from root transformations and constructions induced by specific grammatical morphemes. Though I presently subscribe to the research program that even the particularities of "NP-@ Inversion" can only result from lexical specifications of grammatical formatives, it is not obvious that all its subcases can be so reanalyzed.

The English constructions (all lacking in French) examined and assimilated to NP-@ Inversion are: subject-auxiliary inversion (section 3.1), certain subject-verb inversions (section 3.3), the possessive NP construction (section 3.4), leftward particle movement (section 4.1), indirect object interchange (sections 4.3-4.4), possessive subjects of gerunds (section 5.1) and raising to object (section 5.3). This last section introduces and predicts some rarely noted empirical restrictions. Other sub-sections explain the theoretical bases for excluding otherwise expected outputs of the local inversion, with the structure-preserving constraint and trace theory playing central roles.

The syntactically relevant distinction between grammatical and lexical formatives that figures centrally in my work since is introduced in section 4.4; further properties dividing the lexical formatives into primary and secondary vocabulary (which relate to English indirect object movement) are outlined in section 4.5.

8. Grammatically Deviant Prestige Constructions (1986). The essay addresses issues raised by the fact that all groups of English speakers, including the college-educated, fail to internalize prescriptive pronoun usage. Relevant empirical paradigms confirming this (co-ordinate structures, ellipted VPs, predicate nominals, first person demonstratives and appositive NPs) are presented systematically, in contrast to anecdotal popular treatments.

The article's linguistic analysis shows that the closed class of English subject pronouns are distributed not by universal case grammar but according a local transformation which respects subjacency bounds and obeys conditions on left-right order and adjacency. "Morphological Transparency" (section 2.3), which claims that productive case is transmitted only by open lexical classes, explains the inapplicability of universal case assignment. Sociolinguistic issues (overcorrection, educational policy, etc.) are treated at some length.

9. Parts of Speech in Generative Grammar (1987). The essay proposes a rationale for replacing traditional parts of speech with category distinctions based on the generative methodology of simplifying deep structure distributions. Supporting paradigms are presented fairly schematically. A Table (pp. 36-37) encapsulates the final results, adding an explanation of the non-open nature of the head category P in Kantian terms.

The article argues that there is no motivation for any generative counterpart to the traditional category "adverb." It shows that adverbs can be A (often, seldom, well), SPEC(A) (how, too, very), Proper Nouns (tonight, yesterday), Determiners (here, now, then), SPEC(V) (always, ever, still), or P (aboard, in, overhead), establishing some of these points in terms of some rarely discussed syntactic properties of the category A. The article also brings together previous work on Interjections and claims their special character results from projecting to X' but not Xmax.

In retrospect, this essay (completed well before its 1987 date) gives an overview of bar notation category theory just prior to the Fukui-Speas proposal that closed class categories such as I and D head phrases containing specifiers and complements.

I would like to thank Lída Veselovská, who has made this project a reality.

Joseph E. Emonds

Department of English Language and Linguistics

University of Durham

July 1998