(Cambridge studies in linguistics 82) john m anderson a notional theory of syntactic categories cambridge university press (1997)

Tiếng Anh và mức độ quan trọng đối với cuộc sống của học sinh, sinh viên Việt Nam.Khi nhắc tới tiếng Anh, người ta nghĩ ngay đó là ngôn ngữ toàn cầu: là ngôn ngữ chính thức của hơn 53 quốc gia và vùng lãnh thổ, là ngôn ngữ chính thức của EU và là ngôn ngữ thứ 3 được nhiều người sử dụng nhất chỉ sau tiếng Trung Quốc và Tây Ban Nha (các bạn cần chú ý là Trung quốc có số dân hơn 1 tỷ người). Các sự kiện quốc tế , các tổ chức toàn cầu,… cũng mặc định coi tiếng Anh là ngôn ngữ giao tiếp.

the lexical classes they define It revives the traditional idea that these are to be distinguished notionally (semantically) It allows for there to be peripheral members of a lexical class which may not obviously conform to the general definition The author proposes

a notation based on semantic features which accounts for the syntactic behaviour of classes The book also presents a case for considering this classification - again in rather traditional vein - to

be basic to determining the syntactic structure of sentences Syntactic structure is thus erected in a very restricted fashion, with- out recourse to movement or empty elements.

General Editors: s. R ANDERSON, J BRESNAN, B COMRIE,

W DRESSLER, C EWEN, R HUDDLESTON, R LASS,

D LIGHTFOOT, J LYONS, P H MATTHEWS, R POSNER,

S ROMAINE, N V SMITH, N VINCENT

A notional theory of syntactic categories

52 MICHAEL s ROCHEMONT and PETER w CULLICOVER: English focus constructions

and the theory of grammar

53 PHILIP CARR: Linguistic realities: an autonomist metatheory for the generative enterprise

54 EVE SWEETSER: From etymology to pragmatics: Metaphorical and cultural aspects

57 ALESSANDRA GIORGI and GIUSEPPE LONGOBARDI: The syntax ofnoun phrases:

configuration, parameters and empty categories

58 MONIK CHARETTE: Conditions on phonological government

59 M H KLAIMAN: Grammatical voice

60 SARAH M B FAGAN: The syntax and semantics of middle constructions: a study with special reference to German

61 ANJUM P SALEEMI: Universal Grammar and language learnability

62 STEPHEN R ANDERSON: A-Morphous Morphology

63 LESLEY STIRLING: Switch reference and discourse representation

64 HENK J VERKUYL: A theory of aspectuality: the interaction between temporal and atemporal structure

65 EVE v CLARK: The lexicon in acquisition

66 ANTHONY R WARNER: English auxiliaries: structure and history

67 P H MATTHEWS: Grammatical theory in the United States from Bloomfield to Chomsky

68 LJILJANA PROGOVAC: Negative and positive polarity: a binding approach

69 R M w DIXON: Ergativity

70 YAN HUANG: The syntax and pragmatics of anaphora

71 KNUD LAMBRECHT: Information structure and sentence form: Topic, focus, and the mental representations of discourse referents

72 LUIGI BURZIO: Principles of English stress

73 JOHN A HAWKINS: A performance theory of order and constituency

74 ALICE c HARRIS and LYLE CAMPBELL: Historical syntax in cross-linguistic perspective

75 LILIANE HAEGEMAN: The syntax of negation

76 PAUL GORRELL: Syntax and parsing

77 GUGLIELMO CINQUE: Italian syntax and Universal Grammar

78 HENRY SMITH: Restrictiveness in case theory

79 D ROBERT LADD: Intonational phonology

80 ANDREA MORO: The raising of predicates: predicative noun phrases and the theory

of clause structure

81 ROGER LASS: Historical linguistics and language change

82 JOHN M ANDERSON: A notional theory of syntactic categories

Earlier issues not listed are also available

A NOTIONAL THEORY OF SYNTACTIC CATEGORIES

JOHN M ANDERSON

Professor of English Language,

Department of English Language, University of Edinburgh

CAMBRIDGE

UNIVERSITY PRESS

40 West 20th Street, New York, NY 10011-4211, USA

10 Stamford Road, Oakleigh, Melbourne 3166, Australia

© Cambridge University Press 1997

First published 1997

A catalogue record for this book is available from the British Library Library of Congress cataloguing in publication data

Anderson, John M (John Mathieson),

1941-A notional theory of syntactic categories / John M 1941-Anderson.

p cm - (Cambridge studies in linguistics: 82)

Includes bibliographical references and index.

syntax, can either be omitted altogether, or else be better studied under another name.

(Latham 1862: 577)

'A9f]vas

Preface page x List of abbreviations xii

1 Prelude i

i i Notionalism 21.2 Analogism 61.3 Minimalism 8

2 Fundamentals of a notional theory 13

2.1 Syntactic categories and notional features 132.2 Relations between elements 292.3 Further categories: the role of feature dependencies 432.4 Markedness and category continuity 612.5 Cross-classification 642.6 Gradience and second-order categories 732.7 Secondary categories 1042.8 Non-complements 132

3 The syntax of categories 146

3.1 Verbal valencies 1493.2 The content of the functor category 1683.3 The basic syntax of predications 1743.4 The formation of ditransitives 2363.5 Variation in argument structure 2443.6 Verbals as arguments 2523.7 The structure of primary arguments 292

References 320 Index 345

There were the bits he understood They were bad

enough But the bits he didn } t understand were worse.

The initial quotation here and the chapter epigraphs are drawn from Patrick

White's novel The solid mandala, which and particularly Arthur therein have

much to say about grammar and meaning The present work was concluded byits author when the bits he didn't understand were/are still overwhelming; anddrawing a line at this point is, as is normal, relatively arbitrary This means that

a number of issues which many might regard as crucial to present-day concernwith syntax have scarcely been touched on in what follows, or have even beenignored, and those areas treated have received vastly varying degrees of atten-tion Also, the recognition here of previous work remains partial, and adapted

to the needs of the arguments put forward in particular sections But there is forsuch as the present enterprise no great virtue in comprehensiveness of biblio-graphical reference (or even a possibility thereof) Nor have I striven to estab-lish throughout a consistent temporal endpoint for reference: in a work whosewriting extends over any length of time, achieving this would be equivalent tohaving all of the Forth Bridge freshly painted at the same time It is my hope thatnevertheless - or rather, as a result, to some extent - the following spells out suf-ficiently, over a wide enough area, and in appropriate sub-areas in enough detail,the general structure of a notional theory of syntactic categories and the majorconsequences for the syntax of adopting the views that syntactic categories are

so based and that syntax itself involves the interaction of structures projected bythese categories with pragmatically based requirements involving crucially theorganisation of information - though it should be conceded that the existence ofthe latter is asserted here rather than fully motivated and articulated The dis-cussion which follows thus involves an attempt to support (what has been formost, in the context of the latter half of the twentieth century) an unfamiliarstrategy, both by detailed investigation of some subsystems and with reference

xi

to a wide variety of language types; as such, its reception will no doubt founderbetween and among the Scylla of 'true believers' (responsory motto: That's notwhat we do (anymore)'), the Charybdis of the 'bug collector' (That doesn'thappen in (say) Amharic') and the Deep Blue Sea of scholarly inertia (The lit-erature is what my friends tell me about') I am, however, and of course, solelyresponsible for failures to launch things worth receiving, despite the help ofthose whom it is my pleasure in the following paragraph to both acknowledgeand absolve.

Most of the writing of the book was carried out during the academic years1991-2 and 1992-3 while the author held a British Academy Readership, andwould not otherwise have been carried out: this does not overstate the debt Thefirst semester of 1991-2 was spent as an associate of the Department ofTheoretical and Applied Linguistics in the School of English at AristotleUniversity, Thessaloniki: I am grateful to Professors Efstathiadis andKakouriotis and their colleagues for their hospitality I have also had the bene-fit of presenting aspects of the research reported on here to audiences at theUniversities of Goteborg, Helsinki, Manchester and Umea And a number ofother people have contributed to the preparation of this work in diverse ways,from providing the bottle of NAOY22A to talk over to offering criticisms tochew over, to helping me avoid some garden paths I should especially like tothank the following: Nick Kontos, Scott McGlashan, Peter Matthews, NigelVincent, and anon (who also writes some pretty good stuff of his own, particu-larly the early music); and, most particularly, Roger Bohm, whose influencewill be seen to be pervasive And who could by-pass Fran Colman, longest-suffering in the process of the book's getting written? Not

John Anderson

Edinburgh

Abbreviations and conventions

NONFUT(ure)N(eu)TR(al)PART(iciple)PASS(ive)PERF(ect)PL(ural)POSS(essed)POT(ential)PRES(ent)PREV(erb)REFL(exive)REL(ativiser)S(u)B(or)D(inator)S(u)BJ(ect)S(hort)F(orm)S(in)G(ular)SUBJ(unctive)T(e)NS(e)V(erbal)N(oun)

These abbreviations are mainly used in glosses, where, as labels for phosyntactic categories, they appear in SMALL CAPS. In the text lexemes/wordsalso appear in small caps, the names of inflexional categories are capitalised ini-tially, and cited word and sentence forms are italicised Important terms appear

mor-in bold on their (re)mor-introduction, as do text occurrences, not withmor-in braces, of the

semantic features (P, loc etc.) which identify syntactic categories.

xni

I Prelude

'In the beginning was what word?' Arthur asked

This book is concerned with word classes and their categorisation, where word

is taken to be the basic unit of the syntax and their classification is determined

by this syntax, i.e how they combine to form sentences It is specifically cerned with the 'substance' of class labels (traditionally, 'noun', 'verb', etc.)and with what role this 'substance' plays in the syntax The chapters which suc-ceed this one are concerned to lay out and motivate a particular approach to the'substance' of word categorisation, one which I have dubbed 'notional', in so far

con-as it is conceived of con-as an extension of those 'traditional' grammars that saw this'substance' as ontologically based A more immediate antecedent is some sug-gestive work of Lyons (e.g 1966; 1977: ch 11; 1989); and the approach seems

to me a natural extension of my earlier work on 'case grammar' (e.g Anderson1971a; 1977; 1980b; 1984b; 1986b), as I have indicated elsewhere (Anderson1989b; 1992b) The present brief chapter is intended, as described by its title, as

a 'prelude' to the main discussion, in providing in brief some context for theassumptions and intentions that inform the latter It does not attempt to provide

a historiography of word-class studies, or even of notionalism: these warrantsubstantial treatment on their own account

What follows takes off from the combined proposition that there are nospecifically linguistic semantic categories, or level of representation, distinct

from the syntactic and that syntactic categories are themselves

grammaticali-sations of cognitive - or notional - constructs; what follows is intended to

explicate and to provide support for (particularly) this latter proposition But itshould already be clear that such a viewpoint also imposes certain researchstrategies It means, for instance, that it is predicted to be ultimately unprofitable

to pursue a theory of syntactic categories autonomously, in isolation from theirsemantics; a cross-linguistic syntactic, distributionally established category that

is not notionally defined is merely of interest as a potential counter-example to

the more retrictive theory of notional identification of word classes It alsomeans, for instance, that, with respect to interpretation, representations con-structed out of notionally defined syntactic categories serve as the input togeneral (not language-specific) deductive systems What I am concerned withhere, however, is the development of a system of notionally based syntacticcategories sufficient, in principle, to subtend the expression of a range of syn-tactic generalisations, i.e generalisations concerned with systematic aspects ofthe distribution of words in sentences I return later in this chapter to the assump-tions I adopt concerning the syntactic status of such categories, and, indeed,about the nature of syntactic structure and its relation to other aspects of thegrammar; let us at this point dwell a little on the notional basis for the categories.

I.I Notionalism

The content of a syntactic category, such as 'noun', is notional: the

cross-linguistic classes associated with the category are identified as such by the currence, as members of these classes, of linguistic items that share certainconceptual properties, as discussed in §2.1 below Crucially, in the case ofnouns, these items denote (what are perceived as) discrete physical objects Suchcategories are grammaticalised; as part of a cultural system, the class associatedwith particular languages may contain 'eccentric' members - indeed, memberswhose apparent conceptual equivalents belong to a different category in anotherlanguage Thus, whereas in English it is arguably the case that the distribution

re-of MAY and MUST warrants recognition that they belong to a class - or at leastsubclass - distinct from that associated with (other) verbs, their equivalents inmany other languages share their distribution with verbs And, as we shall bediscussing in §2.3.1, there are, for example, languages where the equivalents ofmost or all of those items that we might classify as adjectives in English areinstead (untypical) verbs and/or nouns Thus, whereas, rather uncontroversially,RED in English is distributionally not a noun, but occurs distinctively as an adjec-

tive (with nominal uses - That's a nice red - being derivative), in Ossetic, for

instance, the 'corresponding' term is, apparently, syntactically like any (other)noun (Abaev 1964: §54) Any notional characterisation of syntactic categoriesshould not only identify the central membership of categories but also provide

a basis for understanding such variation

The existence of notionally non-central members frustrates simple-mindedapplication of blanket notional definitions of categories, despite their currencyuntil quite recently, particularly in pedagogical grammars, in attempts to differ-

entiate those major syntactic distinctions traditionally embraced by the label'the parts of speech' Such definitions either fail to encompass all the exponents

of a particular category, or they are so vague and generalised as to be distinctive of that category; we are all familiar with Jespersen's (e.g 1924: 59)demolitions in this respect of the definitions of Sonnenschein and co (definitionssuch as 'Nouns name Pronouns identify without naming.') See too Lyons 1977:

non-§11.1 Jespersen further shows (1924: 60) that morphologically based ('formal')definitions, though again of ancient ancestry, are also clearly inadequate 'as thesole test' for distinguishing 'parts of speech' in a generally applicable manner.And he concludes: 'In my opinion everything should be kept in view, form,function, and meaning ' (ibid) But, in my view, these aspects he distinguishescontribute in different ways to the 'view': classes are to be distinguished on amorphosyntactic ('form' and 'function') basis, but their cross-linguistic identi-fication is based on 'meaning', the notional character of central members.Moreover, despite varying degrees of grammaticalisation - or desemantici-sation - much of the detailed syntax, as well as the nature of the morphologicaldistinctions associated with particular categories, reflects notional properties.This is quite apart from those relationships in the syntax - notably topicalisationand the like, but including many other variations in linearisation (e.g Bolinger1952) - which expound the organisation of information in an ongoing fashion.What I have in mind rather are the kinds of phenomena we shall be exploring inchapter 3 below, such that, for instance, the capacity for what I shall term'ectopic placement' - roughly equivalent to what has usually been interpreted

as the result of 'movement' - is associated with the semantic role of the elementconcerned, in particular whether it is inherently subcategorised-for or not, sub-categorisation being, of course, on a semantic basis Likewise, as we shall again

be looking at, the basic word order patterns in a language reflect the (semantic)valency status of elements (as, semantically, head, complement, circumstan-tial/adjunct, specifier) Consider here, as a small-scale example of the notionalsensitivity of syntax, one aspect of the syntactic behaviour of 'endocentricadjuncts' in English

As has often been observed, there are interesting restrictions in English onpreverbal placement of endocentric adjuncts like that in (1.1):

(1.1) a Mabel very quietly deteriorated/improved

b Bert very quietly ate the last pie

c Felicity very quietly moved away

such that the sentences in (1.2) are not nearly as happy as those in (1.1):

(1.2) a * Alfred very quietly arrived/danced

b *Bert very quietly ate

c *Felicity very quietly moved

even though variants with the adjunct in postverbal position are perfectly able:

accept-(1.3) a Alfred arrived/danced very quietly

b Bert ate very quietly

c Felicity moved very quietly

The verbs in (1.2a) differ from those in (1.1a) in being 'actions' rather than'processes' 'Actions' rather than 'processes' do not allow a preverbal adjunct

- unless they are presented, as in (i.ib-c), as semantically transitive (not sarily 'syntactically transitive' - even if the latter concept is well-defined) Notall 'processes' are conducive to the preposed adjunct, however:

neces-(1.4) *Basil quietly fell/tripped

The 'processes' in (1.1a) are specifically 'change-of-intrinsic-state' (rather than,say, 'of-place' simply) And so on: other semantic variables supervene

My point is that any understanding of the syntax of the adjunct is ble from the semantics of the adjunction And this is characteristic of syntacticgeneralisations: they refer to notional classes and to semantic relations betweenthem More generally, syntactic properties are projected from notional It is thusinadequate to suggest that the regularities involved in (1.1-4), and elsewhere, aresimply non-syntactic precisely because they are semantically based; that they donot demonstrate a semantic basis for syntax because the phenomena themselvesare not 'syntactic' This is the 'selectional restrictions' strategy - on which seee.g Radford 1988: §7.9 On such a view, nothing of syntax (subcategorisation,placement of complements, adjuncts and specifiers, 'ectopic' placements)remains - as much of the content of the following chapters is intended to illus-trate in some detail

inextrica-I shall argue, too, that the association with particular word classes of particular'secondary' categories (such as Case, Tense), so that the latter are recurrentlyrealised in the 'formal' shape of the former, i.e as part of their morphologicalstructure, is notionally non-arbitrary The recurrence in association with nouns

of definiteness as a functional category - whether inflexionally or

periphrasti-cally expressed, i.e as 'traditional' secondary category or 'article' (§2.7) - followsfrom their notional characters: crudely (see further §§2.1, 3.7), a definitearticle/affix is one of those elements that enables a noun to function as (part of)

a semantic argument (like a name or a personal pronoun) rather than simply as

a predicator

Of course, there may be, as well as individual idiosyncrasies - lexicalisations

- and minor grammaticalisations of membership or relation, as well as reaching relational grammaticalisations, such as that incorporated in the basis forselecting subjects, as discussed in §3.3.1 But these remain desemanticisations;they can be properly understood only against the backgound assumption of anotional basis for syntactic concepts

far-The basis of linguistic categorisation in meaning was the fundamental insight

of that long tradition which eventuated in the 'traditional' (and often pedagogical)grammars which - despite e.g Jespersen's attempts at toughening them up -proved such soft targets for the 'structuralists' In their different ways, and on dif-ferent sides of the Atlantic, both Hjelmslev (1961 [1943]) and Harris (1946) rejectthe notional content of syntax Thus, for the former, 'the projection of the form-hierarchy on the substance-hierarchy can differ essentially from language tolanguage' (1961: 97), where, on the plane of content, the 'form-hierarchy' isroughly, in the terms being used here, the morphosyntactic categorisation - orsystem of linguistic relations, in more Hjelmslevian terms - and the 'substance-hierarchy' is a non-linguistic description of the 'substance' of meaning With likeconsequences, Harris (1946) identifies word - or rather 'morpheme' - classes onthe basis of language-particular diagnostic environments, so that the classes hesets up for English and Hidatsa are strictly non-commensurate So, both Englishand Hidatsa have a class labelled W , but this is fortuitous, non-systematic:

whereas TV in English are 'morphemes which occur before plural -s or its natives' (Harris 1946: §4.1), in Hidatsa N are 'non-clause-final suffixes' (§5.1).

alter-It seems to me that in this respect structuralist syntax was (and is) a barrenconceptual detour, and that the history of (transformational-) generative gram-mar represents a reluctant (and characteristically dissembled) but inexorableretreat from that position One facet of that retreat is part of the backgound tothe discussion of subcategorisation in §3.1.1 In this context, I should perhapsmake it clear that I do not regard the developments described as 'generativesemantics' (cf e.g the contributions to Seuren (ed.) 1974) as in themselves rep-resenting an embracing of notionalism: for most of the proponents of thatapproach, a 'natural logic' was conceived of as forming the basis of syntax;whereas I interpret a notional approach as excluding 'natural logic', as such,from syntax (and indeed from grammar as a whole) and as having it operateupon (amongst other things) interpretations of the categories and structures pro-vided by the latter To this extent, notionalism has something in common withthe Hjelmslevian position alluded to above, while, of course, still denying the(essential) arbitrariness (as posited by Hjelmslev) of the relationship betweensyntactic representation and logical semantics

The positing of such arbitrariness frustrates any possibility of a general theory

of syntactic categories: overt syntactic diagnostics, just like the morphological,are ineluctably language-particular Of course, given a sufficiently abstractconception of syntax, it is possible to make cross-language generalisations con-cerning the syntax of 'nouns', say Thus, in any language with subjects, 'noun'phrases will constitute the central constituent type with respect to capacity tofunction as such But, as we shall see, such generalisations flow from the notionalcharacter of 'nouns', and are best formulated in these terms, with brute distrib-ution itself as a consequence of such; syntax does not involve the arrangment ofautonomous, or notionally uninterpreted elements

1.2 Analogism

I also share with Hjelmslev (and other - mostly dead - linguists) a conception

of linguistic structure as fundamentally bi-planar, as well as the assumption thatthe two planes are structured in accordance with the same principles My mainconcern in the present section is to attempt to explicate the consequences of thisassumption, in particular; but let me turn first to bi-planarity and other aspects

of linguistic organisation

We can differentiate different levels of representation of linguistic structure(along the lines of Anderson 1992&: ch 1) to the extent that the representationsassigned to different levels are governed by distinct regularities Most basically,

or strongly (Anderson 1982a), we can distinguish representations based on adistinct substantive alphabet This is what distinguishes the phonological and

the syntactic planes: phonological representations are constructed out of an

alphabet of phonetically identifiable features; syntactic structures are erected onthe basis of notionally identifiable features Morphology does not introduce adistinct alphabet In terms of a modified 'word-and-paradigm' approach to mor-phology, morphological structure proper interprets the syntactic ('secondary'and 'primary') categories, such as 'Past' and 'verb', associated with a word, interms of the organisation of the phonological material associated with the iteminto morphological units Morphological structure itself involves unlabelledrelations: root, base, affix etc are identified structurally (For discussion see e.g.Anderson 1984a; 1985a; 1992b: particularly §2.3; Colman 1990; 1991: ch 2;

mor-constrains the applicability of an otherwise very generally appropriate tion concerning linguistic structure.

assump-I espouse here the following assumption, given here in the form of a directivefor linguistic representations:

Structural analogy

Minimise (more strongly, eliminate) differences between levels that do not low from a difference in alphabet or from the nature of the relationship between the levels concerned (Anderson 1992&: ch 1)

fol-I interpret this as, within the limits indicated, favouring analyses at differentlevels - most significantly, on different planes - that deploy the same structuralrelations and principles of combination It is thus in accordance with the assump-

tions that, as in Anderson (1992&), the dependency, or head-modifier, relation

should be seen as basic to both syntactic and suprasegmental phonologicalstructure, as well as to the internal structure of segments and words, and thatsegments and words should both be characterised categorially as complexes of

simplex features These recurrences on the two planes are mutually supportive;

and, conversely, plane-specific proposals concerning structure must be shown tofollow from some independent difference between the planes Anderson (1992Z?)argues that this offers a preferable research strategy to the 'anomalisf positionadvocated by Bromberger and Halle (1989), and, pursuing some earlier work(especially Anderson 1985a; 1986b), attempts to illustrate the pervasiveness ofanalogy of structure Anderson (1987c) discusses some recent manifestations

of the 'analogist' view However, such a view is of substantive interest only

if, as well as being appropriately constrained, it is pursued systematically andcomprehensively; a theory which permits sporadic borrowings of conceptualapparatus from one level to another accords to the recurrences no more statusthan coincidence

I assume in the following chapters various aspects of the structure of syntaxand of the grammar that, whatever their internal motivations, receive furtherconfirmation from analogy Thus, I shall propose specifically (§2.2) that syn-tactic structure proper is unlabelled: it is constituted by formal objects we canrepresent as graphs, directed by the dependency relation and linear precedence;and such graphs are projected from well-formed collections of categorial repre-sentations, each member of the collection being constituted by the set of notionalfeatures associated with an individual lexical item Syntax is an unlabelled pro-jection of collections of the categorial properties of lexical items In this it isanalogous to suprasegmental structure in the phonology, which is projected fromthe categories of segmental structure Units like the syllable, foot etc are defined

structurally; such 'categorial' labels (as 'syllable', 'foot' etc.) are unnecessaryand inappropriate (Anderson and Ewen 1987: chs 2 and 3), just as phrasallabels are irrelevant to the syntax.

Likewise, I shall propose that the exception with respect to this absence oflabelling in syntactic structure are those feature-attachments I shall refer to as

prosodic (§2.7.5): these are projected on to the syntactic tree from an element

associated with a particular configuration and thus acquire a syntactic domain;they are manifested primarily as concord Again, such attachments are appro-priate to the characterisation of suprasegmental structure, notably of harmony'processes' (Anderson, Ewen and Staun 1985; Anderson and Ewen 1987:

§7.6; Anderson 1987a; Anderson and Durand 1988) This is one indication thatsyntactic prosodies - or autosegments, if the reader prefers - are not an arbitrarydevice

Perhaps most fundamentally, both syntactic and phonological categories arerepresented, as indicated, as complexes of simplex features, where individualfeatures may simply combine (or be absent) or (where two are present) one may

be subsidiary to the other - there is asymmetry I interpret this subsidiariness asreflecting a (category-internal) dependency relation We have, then, at least thefirst-order categorial possibilities shown in (1.5), in the notation presented morefully in §2.3:

(1.5) a {A} ('feature A appears alone in the categorial representation')

b {A,B} ('features A and B combine')

c {A;B} ('feature A governs feature B')

Much of chapter 2 in particular is devoted to showing that this provides anappropriate notation for syntactic categories with respect to the range of 'behav-iour' we can associate with such So, cross-classification and various hierarchicalrelations which categories contract are optimally characterised - i.e in such away that simplicity of expression correlates with generality of the particularphenomenon However, a powerful (external) support derives from the replication

of like motivations for the same combinatory possibilities in the representation

of phonological segments (Anderson and Ewen 1987: ch 1)

1.3 Minimalism

The basic unit of the syntax, the word, interfaces with the items listed in thelexicon; indeed, word-lexical item is the unmarked mapping, to the extent thatothers are typically interpreted as special, 'idiomatic' An important facet of a

lexical item is the valencies and values it assumes as a word; in terms of valency

it imposes requirements on accompanying items, and it satisfies the valuerequired by other items These valencies and values - subcategorisation and cat-

egorisation - are framed in terms of notional features, including semantic

rela-tions (a.k.a 'case relarela-tions', or 'theta-roles') A well-formed predication

(realised ultimately as a sentence) has all its valencies satisfied, and only onevalue - a predicator - that does not satisfy a valency Syntactic structure -dependency trees with nodes/realising-items ordered in linearity - is projectedfrom such well-formed predications, with results as shown schematically in(1.6):

The character of these relations - lexical and projective - that the word entersinto introduces an asymmetry between the planes: individual phonological seg-ments as such are not mapped onto lexical items, and suprasegmental structure

in phonology in part interprets morphological, syntactic and informational tures, as well as reflecting the categorial content of the segments comprising thelexical, morphosyntactic or informational unit concerned (see Anderson 1987a;Anderson and Ewen 1987: ch 2) However, as noted, the projected object can

struc-in both struc-instances be represented as a directed graph, specifically a tree, withoutlabels except for prosodies

I take it that projection of syntactic structure obeys some minimalist

assump-tions, apart from its restriction to the introduction of unlabelled nodes (with theexception of the prosodic), which in itself has the consequence that syntax doesnot introduce categorial information In particular, the erection of syntacticstructure is monotonic: structure may not be erased or elements be changed inplacement ('moved') This means that, among other things, a syntactic level like'D-structure' ('associated with S-structure by the rule Move-a' (Chomsky 1981:18)) cannot be motivated; it is not merely epiphenomenal but non-phenomenal

We shall discuss such restrictions below in terms of inalterability:

cat-Lexical projection condition

Every syntactic node is projected by a category associated with a lexical item.There is no 'pro', big or little

On the other hand, a single item may be associated with more than one nodeand these nodes may be dependent on different governors, leading to possible

non-projectivity, of the character of (1.7):

where the solid lines are again dependency arcs, linking head/governor (highernode) to modifier (lower), and where I have also suppressed the categorial infor-mation (for a fuller representation see (3.137) in §3.3.4) The arc linking the

lower John node with that associated with read violates projectivity in so far as

it intersects an association line where there is no node to license this A part ofthe discussion of chapter 3 is devoted to establishing the basis for an under-standing of the circumstances in which such non-projectivities are permitted,and (eventually) to suggesting a rather restrictive condition on their occurrence

in apparently diverse circumstances

The categorisations associated with particular lexical items may be of ing complexity, such that the categorisation for one item may include that appro-priate to another, and sometimes this will be reflected in the morphology: whereone item corresponds to the base of another Such categorial relationships

vary-are regulated by redundancies However, categorial redundancies may be of

several different types, of varying particularity Thus, appropriate to many guages, at least, will be a redundancy, of the type just adumbrated, relating thecategorisation for an action verb to a nominal categorisation which includes the

lan-categorisation for the verb and which is to be interpreted as fulfilling the actorrole in the action Traditionally, this latter might be described as an 'agent-

nominalisation' (one function of -er in English - Anderson 1984Z?: §3.5), and its

formation be accounted the concern of derivational morphology As permittingapparently arbitrary exceptions and possibly showing non-compositionality

('semantic obscuration'), I count such redundancies as lexical: in particular

instances of the relationship, there may be no lexical items fully satisfying eitherterm of the redundancy More regular are those redundancies which relate items

of a particular class to different detailed categorial possibilities, such as ating a verb with alternative participial (rather than deverbal adjectival - whichwould be lexical) possibilities Such redundancies, largely the concern, tradi-

associ-tionally, of inflexional morphology, I shall term morphosyntactic Some

redun-dancies are satisfied only in the syntax: notable here is the requirement I shallpropose in chapter 3 that every predication must contain a particular semanticrelation, that I shall label 'absolutive' (a.k.a 'neutral', roughly 'theme' amongthe theta-roles of another tradition); not all predicators are subcategorised for

absolutive, but its presence is necessary to satisfy this particular syntactic

redun-dancy The satisfaction of the redundancy has interesting consequences for theresolution of certain apparent problems in providing, within this minimalistframework, for complex structures involving apparent 'movement'

However, here, as with the other redundancy types, I am anticipating quent discussion which requires more intensive preparation than has been pro-vided so far In this section (and, indeed, chapter) I have baldly stated, in (forthe most part) a necessarily very provisional form, various assumptions con-cerning the nature of syntax and its relation to word categorisation which it must

subse-be the role of the following discussion to attempt to justify In the chapter whichfollows we begin that discussion with an elaboration of the general outlines of

a particular version of notional theory and a provisional account of how thecategorial representations which most directly embody this theory induce therelations of the syntax Finally in this one, let me draw some further distinctionswhich will be important for this elaboration

Syntactic categories are distinguished in terms of their component tions of notional features To begin with, at least, it is useful to distinguish, in a

combina-rather traditional way, and as anticipated in the preceding, between primary and secondary features and categories Primary categories are associated with

distinct distributional potentials Secondary categories involve notional tions - such as Tense, or Number - which are reflected in the morphosyntax of

distinc-a primdistinc-ary cdistinc-ategory, such thdistinc-at the distribution of members of distinc-a seconddistinc-ary cdistinc-ate-gory is included in the distribution of the primary category We should also note,

cate-however, that not all distinctions in primary categorisation are necessarily

asso-ciated in a particular language with a lexical difference in word class, i.e with

classes with different membership Thus, as we shall discuss more fully in

§2 i 4, it has been argued that various languages lack a distinction in word classbetween 'noun' and 'verb' What this involves is not necessarily a denial that theitems concerned can be categorised as 'behaving' in different syntactic circum-stances either as 'verbs' or 'nouns' - (roughly) as predicators or arguments - butrather a claim that the items do not fall into classes determined by their 'behav-ing' only as a 'verb' or a 'noun'; all the items can be, categorially, distribution-ally, either a 'verb' or a 'noun' A categorial distinction between 'noun' and'verb' may be relevant to the syntax without being given expression in terms oflexical differentiation What follows is, then, both about (syntactic) categoriesand about (lexical) classes, and about the extent to which they coincide.Such a distinction is not drawn by what I think is the most detailed and sys-tematic attempt to provide a classification of the 'parts of speech', that made inBr0ndal 1928 And I can perhaps distinguish further the aims of the presentdiscussion in terms of others of Br0ndal's abnegations He offers a classificationbased on notional features, as does what follows However, for Br0ndal, theclasses, not merely their cross-linguistic identification, are established on purelynotional grounds, without reference to syntax, in contrast with the positionadvocated in § 1.1 above (despite minimalist assumptions) For Br0ndal, no class

is found universally (1928: 217), but a maximum of fifteen classes is allowed bythe logic of the system (pp 217-18) I shall propose a minimum set of classes

in what follows; whereas the number of classes permitted by the calculus structed here is indefinitely extensible (the upper limit being an empirical ques-tion (Hjelmslev's (1935-7) 'maximum absolu' vs a 'maximum theorique').What follows nevertheless shares with Br0ndal's enterprise the conviction thatlanguage-particular systems are constructed out of (combinations of) a small set

con-of universal features (which for Br0ndal define the classes con-of 'relata' vs.'descripta') on the basis of a general 'logic'

2 Fundamentals of a notional theory

'Tell Mrs Musto I'm concentrating on words The Word.

But also words that are just words There's so

many kinds You could make necklaces '

In this chapter I am concerned to set out the basic properties of one approach

to the development of a notional theory of syntactic categories, primary andsecondary Chapter 3 elaborates further the properties of verbal and nominalwords and constructions; and there, too, I shall attempt to show more generallyand precisely that the notional characterisations of the various categories formthe basis for the assignment of syntactic and morphological structure Here Ipropose as the content of syntactic categories a set of notional features whosedistribution through the set of categories identifies these categories cross-linguistically, and I begin to sketch the syntactic consequences of the categori-sations I shall also try to show in a preliminary way that this distribution offeatures correlates with the typical associations between primary and secondarycategories (such as verb and Tense) and the recurrent syntactic relations intowhich the primary categories enter; again these are explored more fully in thechapter which follows These recurrences provide powerful support for thehypothesis of a notionally based grammar As was intimated already in chapter

1, a theory in which the category labels or the syntactic features in terms ofwhich the labels are defined are uninterpreted can in principle provide no expla-nation of such recurrences: the recurrences, which are in principle theorems withrespect to the notional properties of categories, must rather, in such a theory, betaken themselves as axiomatic

2.1 Syntactic categories and notional features

Givon (1979: §8.6; 1984a: §1.3.4) relates the content of the noun/verb tinction to the 'time-stability scale', such that nouns denote stable phenomena,

discrucially concrete, physical, compact, and verbs denote rapid changes, events and adjectives, where distinct, are intermediate in this respect, denoting rela-tively stable physical qualities; in the terms of Anderson (1991&), nouns are'entity-specific', verbs 'event-specific' It seems to me that this highlights, andonly partially, only one aspect of differentiation, one which I shall associate,

-in the case of the noun/verb dist-inction, with the presence vs absence of thenotional feature N, or (perceived) referentiality, or rather referentiability, attri-bution of which to an element is supported not merely by relative stability but

by relative independence (cf Lyons' 'first-order entities' (1977: §11.3; 1989);also Hopper and Thompson 1984; Langacker 1987; Croft 1991)

Verbs are in addition inherently relational, they impose overt structure on (theperception of) events, either new or confirmatory structure - they are situation-defining Nouns, on the other hand, typically label elements participating in sit-uations, situations that are presented linguistically as (re-)structured primarily byverbal labelling; and nouns are thus potential undergoers of (non-constitutive)change, changes differentiated by verbs I associate this difference, in thecase of noun vs verb, with the absence vs presence of the notional feature P, orpredicativity, or rather predicability

The 'relational' character of P is reflected in one of the contrasting effects of

verbalisation vs nominalisation: verbalisation of a noun - say, house - creates

an item - the verb house - with a complex 'argument structure' compared with the noun, as realised in They housed the refugees in the barn, for example; nom-

inalisation of a verb either merely preserves such 'structure' - as with the derived

noun of Fred's abandonment of that principle - or results in a reduction in plexity - as, for instance, with the English nouns in -ee derived from verbs

com-{deport => deportee etc.) The 'stable' character of N, on the other hand, is

reflected in the capacity of the derived noun, and of the basic house, to function

as an argument

The two aspects of content associated with P and N are related in a naturalway: verbs denote time-dependent structurisations and nouns introduceautonomous (potential) referents But the two aspects have different conse-quences: so that, as observed, verbs, as having P, so relation-inducing, can show

a range of argument types, and, in lacking N, stability, are classified by type, or type of change, and are often associated with temporal deixis (whichmay be realised as the secondary category of Tense); nouns, in lacking P, aretypically unsubcategorised, and, in having N, are classifiable in terms of stableproperties, such as concreteness and animacy (possibly manifested as Gender).Further, we shall find (in §§2.3-6) that the notional characterisation of furthercategories involves the interaction of just these two features Adjectives, for

event-instance, combine N and P, and this is reflected in their syntax: their isation will occupy §§2.3.1, 2.7.4 and (parts of) 3.7

character-The word classes verb and noun are identified cross-linguistically as utional classes with distinct membership whose members include respectively

distrib-event- and entity-denoting items We can distinguish central members whose

status as one or the other is relatively transparent: thus, central nouns will denotefirst-order entities, basically 'persons, animals and other discrete physicalentities' (Lyons 1989: 161; cf again Lyons 1977: §11.3) Other members of aclass are notionally more peripheral, as with nouns which are 'relational'

(stranger, opponent) or 'abstract' (lust) or even 'actional' (battle) The extent to

which the internal structure of the classes shows properties reflecting typicality' (e.g Rosch and Lloyd 1978) is a distinct question whose resolutiondoes not prejudice (though it might enhance) the viability of the idea of centralmembers

'proto-2./.7 Notional features as basic

The distribution of N and P is the ultimate basis for the construction of the

pred-ications which are realised as sentences Nouns are optimal arguments

(rela-tees) in the predications determined by the structures induced (optimally) by

verbs, as predicators (I adopt here Lyons' terminological suggestion(i977:

434), so that 'we can say that "play" in "Caroline plays the guitar" is a two-placepredicator independently of whether we also say that "play the guitar" is a pred-icate'.) The nature of subcategorisation, or 'argument structure', and its role inthe erection of syntactic structure, is explored in some detail in chapter 3.However, as a preliminary, I shall shortly have to somewhat refine upon thecrude view of the syntactic roles of verbs and nouns just expressed in the light

of the differentiation of further categories

The notional features also correlate with discourse function (cf again Hopperand Thompson 1984) Items with N, as stable referables, are optimal topics;items with P, as structure-inducing, are the nucleus of optimal comments In atopic-free sentence, verbs may thus occur alone, as in the Greek (2.1):

(2.1) prexi 'it rains/is raining'

However, clearly, the (possibly topic-free) structure induced by a verb may alsoinclude elements characterised as N:

(2.2) prexi stin ela8a 'it rains/is raining in Greece'

And, more generally, discourse function cannot be regarded as a defining property

of N and/or P, given the independence of categorial status and discoursefunction So in (2.3):

(2.3) PADDY left

Paddy may be a comment on the topic of a 'leaving' Such is familiar from the

vast body of work on discourse functions

The subject-predicate distinction is apparently a grammaticalisation of (thetypical manifestation of) topic-comment But another factor enters here Languagestypically accord a special status to animate, particularly human, and ultimatelyegocentric arguments and their properties (cognition, volition): cf the 'animacyhierarchy' of Silverstein (1976) and others Their preferential roles in predica-

tions are grammaticalised as agent (source of the action) and experiencer

(loca-tion of the experience) On grounds of empathy (Kuno 1987), the human nounsassociated with these roles (and particularly the source of the action) are pre-ferred as topics And thus the typical subject is seen as also agentive and human.Again, we return to a more careful consideration of this in chapter 3 (and cf.already e.g Lyons 1968: §8.1) But we should further note at this point that suchconsiderations also impinge on the character of the categories defined by Nand P

Within the set of noun denotata, humans have a special status; within the set

of verbal notions, actions, particularly externally directed, or transitive actions.They are seen as epicentral to the class It is not for nothing that (2.4):

(2.4) The farmer killed the duckling

is perhaps the most familiar example in the discourse of linguistics This specialstatus of humans and (transitive) actions will be apparent in a number of places

in what follows, not merely in relation to subjecthood

N and P are simplex features (as proposed by Anderson and Jones 1974;

Sanders 1974), either of which, as noted in chapter 1, may be present or absent inthe representation of a particular category I shall also be assuming, as again antic-ipated in chapter 1, that in the representation of a particular category one featuremay predominate over another - the combination may be asymmetrical - and that,

further, this asymmetry is one manifestation of the dependency (head-modifier)

relation that may hold between linguistic elements (as illustrated further in §2.2)

We shall find that the invocation of simplex (and not n-ary-valued) features and

of dependency is crucial not just to further elaboration of the range of categoriesallowed for (§2.3), but also to the characterisation of markedness and otherhierarchical relations between categories (§§2.4-6), and of the principle whichgoverns the well-formedness of particular systems of categories (§2.4)

From this, it may be apparent, too, that I do not envisage that all languageswill necessarily share the same categories, or word classes A basic, possiblyuniversal core system is provided by the set of simple combinations of N and P

- N alone, P alone, N and P, neither - but further elaborations, involving metrical (dependency) combinations, are perhaps not necessarily universallyattested, at least as constituting a system of word classes: see §2.1.4 However,the elaboration of categories is strictly governed, and potentially upwardlybounded also, though the boundary must in our present state of knowledge beprovisional only (see further §2.6)

asym-2.1.2 Basic categories

Let us provisionally represent the category verb, universally, as {P}: its

cate-gorial specification includes P alone Nouns include N in their specification But

at this point it becomes urgent to introduce an important distinction thus far

glossed over Among the traditional 'naming' words only (proper) names

qual-ify as belonging to a category whose specification contains only N They are

never predicative, i.e predicators, unlike the (common) nouns in (2.5):

(2.5) a Fred is a doctor

b Alphonse est medecin 'A is a doctor'

c O Petros ine 8iki7oros 'P is a lawyer'

Notice in relation to this that I do not regard either of the post-copular elements

in (2.6) as predicative:

(2.6) a The man with the shakes is Fred

b The man with the shakes is the doctor

What is predicated in (2.6) is not 'Fred-ness' or 'doctorness' of 'the man withthe shakes'; rather, what is predicated is (referential) identity between two argu-ments, 'the man with the shakes' and 'Fred'/'the doctor' Unsurprisingly, (2.7)differ in interpretation from the respective sentences in (2.6):

(2.7) a Fred is the man with the shakes

b The doctor is the man with the shakes

only in assignment of unmarked discourse function

As illustrated by (2.5b and c), in many languages, such as French and Greek,

it may be that no article accompanies a predicative noun, or some predicativenouns, even if, as in French, a noun is otherwise normally accompanied by adeterminer Bolinger notes that the article may be absent in English in 'intensi-

fier' constructions, as Is he farmer enough to face drought and pests? or That

animal is only partly dog (1972: 136, 108) Moreover, occurrence of the

indef-inite article in English does not ensure that a post-copular NP is predicative.Thus, (2.8a), on one interpretation, is equative, as, of course, is (b):

(2.8) a Our teacher is a plumber from Yorkshire

b A plumber from Yorkshire is our teacher

The indefinites in (2.8) involve specific reference Contrariwise, a nominal

pred-icator may even bear a definite article: Bertram is (the) president And, as this

last example illustrates, 'relational' nouns in English are most likely to lack any

determiner when predicative; cf too e.g Hugh Walpole's He was now friend to

all the valley (Rogue Herries, part 1, 'Christmas Feast') This correlates with

their non-central status as nouns

In Classical Chinese (Graham 1967), for instance, predications of identity

are distinguished by a special copula, chi English and other languages show

a common copula in locational, noun-predicative and identifying sentences;and elsewhere differences in copula correlate with 'aspectual' differences(Anderson 1973a) Notice, however, that the question word with predica-

tors in English is what, whether verbal, adjectival or nominal, no matter how

the interrogatives involved may differ in other respects, as illustrated by(2.91):

(2.91) a What happened (to Bill)? He died/Bert killed him

b What did Bert do? He left/killed Bill/was a butcher

c What is Bert like? He is charming

d What is Bert (like)? He's a Tory

(ii) a Who(m) did he kill? He killed Bill/a plumber from Yorkshire

b Who is he? He's our teacher/a plumber from Yorkshire

the one with animate arguments, as exemplified in (ii), is who(m), and the latter

may be answered with either a definite (including names) or indefinite nominal.Names are not predicators; and, as we shall see, nouns are arguments only in sofar as they are potential names

Given the notation developed so far, we can distinguish between names andnouns as in (2.10):

(2.10) {N) {N,P}

name noun

Elements with P are possible predicators; elements with N are possible primary

arguments Secondary arguments are the {P}s which are the heads of

subordi-nate clauses; we return to these in §2.3.2 and - more extensively - in chapter 3

We can observe here already, however, that in English (for example) only

primary arguments allow the full range of (predicator-)argument labels, or'functors' (see below), to be expressed, as prepositions:

(2.11) a Mary was surprised at John/the result

b Mary was surprised *at (that) John left

However, the partially predicative character of nouns renders them less optimal

as stable labels for arguments, given that they are thus less referentially specific;they have inherent sense as set designators, they are classificatory rather thandirectly referential, while simple {N} elements differ only referentially ratherthan in sense (Lyons 1977: §7.5) The latter therefore also appear without restric-tive attributives, as being (assumed to be) sufficient in themselves for identifi-

cation of the referent of an argument: the adjective in poor John has no referential role, and that in the younger Bill involves recategorisation of Bill as a noun.

(This is, of course, not to deny that names may themselves be e.g denotationallygender-specific.)

Personal pronouns also are non-predicative: they too are simple {N}s They

are contextually determined names Proper names are taken as stable labels forprimary arguments; personal pronouns are shifting labels Compare here Lyons(1977: 214-15; also Jespersen (1922: ch 6, §7) on the personal pronouns as

'shifters' Only non-central, 'relational' nouns, such as father, enemy, are in

a sense 'shifters' Of course, the reference of a name (such as John) may be

context-dependent; but names, unlike pronouns, are not understood as ently so We might recall here that Latham describes the pronoun as 'a variablename' (1862: §646); whereas nouns ('common names') and ('proper') namesare invariable but differ in whether they 'are applied to a whole class of objects'

inher-or 'are appropriated to certain individual objects' (1862: §633) - with pronounssharing the latter property with (proper) names (§636) Like names, too, pronounsfrequently distinguish gender (even when this is not overtly associated with thenouns in a language), as well as their characteristic deixis-based oppositions

Definite determiners such as those in (2.6b) are transitive pronouns (cf.

Anderson 1976); they take nouns as complements and render them in contextname-like, 'terms', if you like Compare here, again, on determiners in general,Latham 1862: part v, chs 2-5 In this respect, too, a similar distinction to that I

am drawing between names and nouns is embodied in e.g unification

categori-al grammar (Zeevat 1988) and some of its antecedents Given such observations,

we might say that 'noun phrases' are more properly designated 'name phrases':determiners permit nouns to name Such and related observations have led toanalyses of pronouns as articles (e.g Postal 1970) or determiners as the heads

of the constructs traditionally termed 'noun phrases' (e.g Abney 1987),

renamed accordingly 'determiner phrases': for discussion, see e.g Hudson1990a: §11.2 Given that nouns, names, pronouns and determiners are all nom-inal (in terms of the notation proposed here), the drastic categorial reinter-pretations suggested by the above terms are unnecessary Perhaps we can

compromise on nominal phrase as the label for the construct.

Indefinite determiners, or 'quantifiers', such as those in Fred adores a/some

fan-dancer, combine a kind of naming function with a 'presentative', ultimately

predicative function As with the corresponding 'intransitive' pronouns - such

as French on, English someone - an entity is referred to, but not identified, by

way of asserting its existence We return to quantificational structure in §3.7.2.Here let us again note that the association of determiners and nouns is based

on their respective notional properties: common nouns, as class labels, ment the closed class of determiners, transitive pronouns, to provide a moredifferentiated set of variable names And I attribute, in turn, the properties thatnouns share with names/pronouns to the presence of the feature N, just as thepredicative role of nouns can be attributed to their also being P

comple-Names and verbs involve one-feature combinations; nouns combine twofeatures The final simple combination of N and P available is the null one, theempty set { } These are items which are neither predicator nor argument; theysimply provide an empty slot with which can be associated the label for therelation between predicator and argument With primary arguments, they aretypically realised as adpositions, as in (2.12):

(2.12) the destruction of the city by the Goths

or inflexionally (The subordinate predicatorial character of nominalisations like

destruction is discussed in §2.6 and elsewhere.) But, since the relation between

predicator and argument is often clear from the character of the predicator, as

when there is only one argument, as in Mabel danced, or from the predicator and

the 'word order' (as in (2.4) and (11)), such elements are frequently given no(other) overt expression I have put 'word order' in 'scare' quotes to indicateawareness of the fact that this considerably oversimplifies the situation; the syn-tacticisation of argument-label expression is discussed more fully in §3.1, onthe basis of the notion of dependency introduced in §2.2.1 shall term this cate-

gory, viz { }, the category of functors As we shall see, the category includes,

as well as the traditional (NP-introducing) prepositions and postpositions, other(secondary) argument-introducers ('subordinators') such as many comple-mentisers (cf e.g Anderson 1972; 1977: 105, 161, 282 n 31; Emonds 1985:especially ch 7)

2.1.3 Categories and distribution

Simple combinations of N and P provide for the range of distinctions in (2.13):(2.13) {P} {N} {N,P} { }

verb name noun functor

Let us suppose this at least constitutes a minimal categorial system, such that inparticular languages, each category displays a distinctive distribution, attribut-able to its notional character, and not (totally) included in that of another cate-gory Thus, for example, at a rather crude level, only verbs in English can befinite (i.e., for the moment, occur as the predicator of a simple clause) - names,nouns and functors require the 'support' of a copula:

(2.14) a The plumber is Fred

b Fred is a plumber

c Fred is at the door/out

(where out in (c) is an 'intransitive' functor); only names, as the quintessential

argument-type, inherent 'terms', can be the unmarked complement of an tive - 'specific' nouns (nouns made 'name-like') only can also be such; onlynouns are unmarkedly either predicator or argument; functors are neither, butlabel arguments with respect to their roles in the predication (even, I assume, in

equa-(2.14c), with be as a locative verb, in contrast e.g with the equative be of (a):

see further §3.1)

In terms of relations between the categories, we can say that the typical sitive') functor complements and is complemented by one of the other categories;the typical (non-impersonal) verb is complemented by functors; categories con-taining N typically complement functors (This again assumes that functors arenot necessarily expressed distinctively in the syntax, by adpositions.) Within theset of categories containing N, a noun, as opposed to name, may, like the verb,

('tran-be subcategorized as to functor selection {father of/to the bride, descendant

of/Ho Jonah): see further §3.7.1.

Here it could be objected that determiners (and functors - if we recognisesome of them, such as that in (2.14c), as 'intransitive'), though lacking the P fea-ture, also display subcategorisation (Roger Bohm, personal communication).But we should also observe that determiners and pronouns (and 'transitive' and'intransitive' functors) are distinguished by presence vs absence of a comple-ment rather than in terms of the character of the complement(s), as in the functor

selection of 'relational' nouns We can then distinguish between weak

sub-categorisation, manifested by determiners, which is simply a consequence of

complement structure and the fact that not all complements can be obligatory

(with respect to a particular category), and strong subcategorisption of the

differential character displayed by categories containing P We refine upon thisdistinction in the following subsection; but there we preserve the essential dis-tinction made here between the highly differentiated complementation of itemswith P compared with those without

It might also be argued that names simply display a subpart of the tion of nouns, in lacking the predicativity possibility, and are thus merely arestricted subclass of the latter; however, in suggesting that only names are quin-tessentially arguments, I am also claiming (to expand on the description above)that the occurrence of nouns as arguments is mediated by conversion toname/pronoun status, though such conversion is not always overtly signalled(by the presence of articles etc.): again, we return to this in §3.7 The basic idea

distribu-is that even vandals in (2.15):

(2.15) The car was destroyed by vandals

is a converted noun, a derived name: it is associated with a determiner that is notgiven independent expression (cf too §2.7.5)

In some languages, names/pronouns and nouns show more gross tional differences Thus, in German only the genitives of names/pronouns can

distribu-occur prenominally in the NP (Jakobs Tod 'J.'s death'), while noun genitives are restricted to postnominal position (Der Tod des Freundes 'the friend's

death') - see Anderson 1987b Interestingly, as again pointed out to me byRoger Bohm, prenominal genitives include kinship terms used as names, as in

Opas Besuch 'Granddad's visit' Let us note too, as another instance of a more

overt differentiation of names, that in many of the Polynesian languages namesare accompanied by a distinctive 'article' (cf e.g Krupa 1982: ch 4)

One might similarly object, however, with respect to this scheme of gories, that in languages where a predicative noun does not need the support of

cate-a copulcate-a, cate-as in the Russicate-an (2.16) - cf (2.14b):

(2.16) Marija rebenok 'Mary (is a) child' (Lyons 1968: §7.6.3)the distribution of verbs is included in that of nouns, and the former are thussimply a subclass of the latter But, again, nouns are not quintessential predica-tors In Russian, this is reflected in the fact that in the past tense the presence of

a copula is required Curiously, Lyons, in arguing for the 'primitiveness' ofstructures such as (2.16), contends that 'there is no convincing syntactic orsemantic reason for classifying "be" in English as a verb' (1977: 437) But itssyntactic credentials are surely impeccable, and though as a verb it may not be

notionally central, it nevertheless displays the relational, structure-inducingcharacter I have attributed to P Rather than regarding 'be' as simply a 'surface-structure element which "carries" the overt marking of some syntactic distinc-tion' (here 'tense, mood and aspect' - Lyons 1968: 322), with all the undesirablepower that the idea of a purely 'surface-structure element' introduces, I takeinstead the presence of 'be' as an indication that nouns are, even in languageswhere its distribution is restricted, not preferred predicators The kind of restric-tions on absence of copula illustrated by Russian are not uncommon (cf e.g.Ferguson 1972, on Bengali; Bhaskararao 1972, on Telugu) However, we return

in a moment to issues raised by languages where predicative nouns are neveraccompanied by 'be'

Nouns, {P,N}, are notionally intermediate between verbs, {P}, and names, {N};conjunction of P and N 'dilutes' their individual characteristics Syntactically,too, in a notionally based theory, they are expected to show a distribution sharedwith verbs and names, but attenuated The (even partial) presence of copulasand articles is one indication of this, as is the intermediate capacity of nouns forcomplementation by distinct functors (strong subcategorisation)

On such grounds the potential word classes of English and many other guages associated with the representations in (2.13) are categorially distinct at

lan-a lexiclan-al level, in displlan-aying distinct distributions for their members, lan-and oftenassociation with a distinct set of secondary categories In discussing, in particu-lar, the noun/name distinction I have also endeavoured to show that, though theirdistributions may overlap, one is not simply included in the other However, inmany languages this is true only if the distributions are 'interpreted', in that e.g.nouns are associated with 'determiner' categories that may not be overtlyexpressed And the non-inclusion requirement may not be necessary for theestablishment of lexical categories; indeed, it is not clear how it is to be applied

in the case of classes, such as the 'adverbs' discussed in the following section,which are categorially complex We return to these issues in §2.3, though theyalso have some relevance to the subsection which now follows

2.1.4 & minimal system?

A provisional general hypothesis might be that (2.13) represents the minimalsystem of word classes that needs to be attributed to any language Thus,although Robins (1968), for instance, attributes to the 'basic structures' ofSundanese only the word classes 'N(ouns), V(erbs), and Particles', it is clearthat pronouns (and names) are therein distinct from nouns in being non-predicative, and potentially equative And, more generally, although Schachter

claims merely that 'in most languages some grammatical distinction is made

between common nouns, which are used to refer to any member of a class of sons, etc and proper nouns, which are used to refer to specific persons, etc.

per- ' (1985: 8), such a conclusion seems to be generally applicable It is not ply the case that, as noted in §2.1.3, m some languages there is a particular'marker' of the distinction, either partially distinctive, as e.g the typical absence

sim-of an article with names in English or the normal presence sim-of a definite one in

Greek {Peter vs o Petros), or more generally, as e.g in the selection of special

'articles' in many of the Polynesian languages (see again Krupa 1982: 64-8), or

in terms of the differentiation of case/topic markers in Philippine languages(Llamzon 1979: 115); names are apparently universally distinct with respect

to their (non-)capacity for predicativity/equativity, reflecting their status asquintessential arguments or 'terms'

Say, then, that (2.13) may not be reduced in particular languages to (2.17) as

{ )

functor functor

which has only unary or less combinations; (2.17a) is 'sub-minimal' Schachteralso discusses (1985: 11-13), however, suggestions that there are languageswhich reflect a system of categories that we might, in terms of the presentnotation, represent as (2.17b), with no verb/noun distinction, or even (2.18), areduction of both of (2.17) - and thus in contradiction of what I have justconcluded about (the universality of) the names/noun distinction:

(2.18) {N} { }

verb/name/noun functor

To allow for the construction of sentences in such putative systems, whichrequires a distinction between predicator and argument to be made, particularoccurrences of the {P} class in (2.17b) (and the {N} class in (2.18)) would beprovided by redundancy with additional specification, so that arguments could

be distinguished derivatively (non-lexically) as {N,P} The choice of N rather than

P for the non-functor class in (2.18) reflects the assumed primacy of naming: see

§2.4 below

A 'sub-minimal' system has been suggested for Salish: see e.g Kuipers 1968;Kinkade 1983 The latter specifically argues for a system such as we havecharacterised in (2.18), in so far as (unlike Jacobsen 1979, on Makah) he pro-poses that names are not distinct from predicators; but the examples he offers

(1983: 29) show names as arguments in equative predications rather than tioning as predicators themselves: again names/pronouns are apparently distrib-utionally distinct from other potential classes In general, a system such as (2.18)has not been shown to be appropriate.

func-Swadesh's (1936-8) much cited description of Nootka would appear to rant a characterisation such as (2.17b), as would Bloomfield's account of Ilocano(1942), wherein are recognised as word classes only 'pronouns', 'full words'and 'particles' (Bloomfield also groups 'names' with 'full words' (1942: §6),

war-on the basis of their requiring, like 'full words', special markers when they occur

as arguments; but these seem to reflect differences in how functors are expressedrather than instantiating a fundamental difference in distribution between namesand pronouns.) Swadesh sums up the Nootka situation as follows:

Normal words [vs particles - JMA] do not fall into classes like noun, verb,

adjective, preposition, but all sorts of ideas find their expression in the samegeneral type of word, which is predicative or non-predicative according to itsparadigmatic ending' (1936-8: 78)and provides examples such as (2.19):

Let us be clear that what is at issue is the lexical recognition of categorial

dis-tinctions; whether there are classes of words limited to one or other category.The claim being advanced in relation to Nootka and the like is that predicatorand argument status is equally available to all 'full words'; we can contrast herethe common situation, illustrated rather strikingly by Australian languages (e.g.Blake 1987: 9), wherein normally a 'noun used as a verb' will be derivationallymarked There is no doubt, however, as is recognised in Sapir's description, thatthe articulation of the morphosyntax of alleged 'sub-minimal' languages, char-acterisable in terms of (2.17b), requires, as with other languages, reference to adistinction between predicator and argument, and not all arguments are names/pronouns In Ilocano, for instance, the formulation of word-order patternsdepends on such, the basic pattern being: predicator (+ argument(s)), where

arguments are riot limited to 'pronouns' but may involve members of the class

of 'full words', whose members also occur as predicators And, though Boasconcludes, with respect to Kwakiutl, that, as concerns potential P and P,Nwords, 'all stems seem to be neutral' (1911: 441), he subsequently affirms that'the classification of suffixes here given shows that a division of words intoverbs and nouns has taken place' (p 443), and the syntactic organisation of thesentence depends on the making of this distinction See further, more generally,Lyons 1977: §11.2

Discussion of this issue is bedevilled by a confusion of lexical class and tactic category Kinkade, for instance, seems to be unclear about (among otherthings) the distinction between category and word class when he concludes(1983: 32): 'as I have shown, Salish shows only predications, and there is thus

syn-no basis for claiming a distinction between syn-nominal expressions and verbalexpressions (Lyons' syntactic categories)' (cf too Thompson 1979: 699) Theevidence Kinkade presents suggests that there may be no lexical distinctionbetween noun and verb categories (though the languages concerned deservemore detailed analysis in this respect - cf van Eijk and Hess 1986); but it doesnot warrant the conclusion that the categories predicator and argument (distin-guishing, among other things, between the syntactic functions of 'full words')are irrelevant syntactically

The members of a (lexical) word class may be assigned a range of categorialpossibilities by redundancy If there are languages whose set of word classes islimited to that defined by (2.17b), their syntax will nevertheless appeal to at leastthe system of (2.13), with the categorial distinctions being made available by aredundancy such as (2.20):

(2.20) {P}^>N

i.e predicates ({P}) may also be {P,N}, making them available to marking (withdemonstratives etc.) as non-quintessential arguments and to the occupying ofargument positions

The existence and characterisation of such languages remains rather versial For instance, as concerns Nootka, Jacobsen (1979) has shown that of theset of 'full words' there are some which cannot occur as arguments unless suf-fixed, whereas others occur as such suffixed or unsuffixed; the former setincludes items which universally are (otherwise) centrally P words Now, onemight say that this simply shows that such items have a more restricted distrib-ution than other 'full words': we have a subclass rather than a class difference.(Cf again Schachter 1985: 11-13, on both Nootka and Tagalog.) Alternatively,however, in view of the correlation with the primary categorial notional features,