Historical Linguistics

Problems and Perspectives

Charles Jones

• Publisher: Routledge
• Year: 2014
• eBook ISBN: 9781317898993
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

The contributors to this volume cover the international range of scholarship in the field of Historical Linguistics, as well as some of its major themes. The work and ideas they discuss are relevant not only to other aspects of Historical Linguistics but also to more general developments in linguistic theory. Along with Professor Jones' Introduction, their comments provide a major overview of Historical Linguistics that will be the reference point for its development for many years to come and form an important contribution to general theories of linguistic behaviour.

Chapter 1

Parameters of syntactic change: a notional view

John Anderson

Syntactic theory circumscribes syntactic variation and thus syntactic change: potential changes (ones that might be contemplated pre-theoretically) are, in the first place, either impossible, as infringing universals of syntax, or possible. Assessing the likelihood of possible changes, potentiality for recurrence, is apparently more complex, however. Many recurrent changes involve minimal incremental or decremental modification, such that the change involves a local increase or decrease in what is allowed within the syntax of a particular language, where locality is defined by syntactic theory. But it has been claimed that the histories of languages can also experience major restructurings. Such, for instance, would be involved in the innovation of a category ‘modal’ in the history of English, as proposed by Lightfoot (1974, 1979: ch. 3). Here I consider various changes in the syntax of English, some of them with quite widespread repercussions (and including the history of the ‘modals’), which nevertheless can be related to local, minimal modifications to the syntax of English: I suggest that these are preferred, most likely.¹ I am concerned specifically with ‘categorial’ changes, with respect to which it might further be suggested that ‘radical restructuring’ is even impossible, excluded by syntactic theory; if such changes can be restricted to the minimal modifications outlined and exemplified below, then we can put firm bounds on the notion of possible change in categorization. The establishment of this depends on the adoption of a restrictive theory of syntax such as has been developed within notional grammar.

Notional grammar provides for each language a set of notionally defined categories out of which syntactic representations are built in accordance with particular parameter settings appropriate to that language. A hierarchy of categories is established by the inherent complexity of the representations they require: some categories are more basic and require fewer (notional) features. Consider, for instance, a notional theory whereby categories are identified in terms of the notional components, or features, P(redicative) and N(ominal), presence of P being associated with categories that are situation-defining, and of N with categories that are entity-defining. In these terms, (finite) verbs and (proper) names can be simply characterized, respectively, as {P}, i.e. as (categorially) predicative, and as {N}, nominal, and functors (adpositions) as { }, categorially unspecified; while other categories, such as adjectives, are identified by combinations of the two features, and are thereby more complex, less basic.

Combinations may be simple, symmetrical, or asymmetrical. Some basic possibilities, involving asymmetrical and symmetrical combination, are illustrated in (1):

(1)

P finite verbs

P;N non-finites

P:N adjectives

N;P common nouns

N names

The feature on the left of the semi-colon preponderates over, or governs, that on the right, which is dependent; the colon indicates mutual dependency. With non-finite verbs P governs, preponderates over, N; with common, as opposed to proper, nouns, the preponderance is reversed; with adjectives, the two properties are mutually preponderant. Further, more complex categories can be provided for by second-order combinations, such as are illustrated in (2.a) and exemplified in (2b):

(2.a)	P;(P:N) participles	(P:N);P deverbal adjectives
(2b.i)	Bruce is neglecting Sheila (participle)
(2b.ii)	Bruce is neglectful of Sheila (deverbal adjective)

In (2.a) the second-order representations are composed of the same elements, P, the ‘finite verb property’, and P:N, the ‘adjectival property’, but the preponderance, or direction of dependency, of ‘finite-verb-ness’ and ‘adjectiveness’ is reversed between the two. Participles are more verb-like than deverbal adjectives: see s. 3.2 below.

Elsewhere (notably Anderson 1988, 1989a, 1990, 1991b) I have discussed the notional definitions of the features and the motivations for attributing particular combinations to particular categories; the role of the representations in providing appropriately for cross-classification (in terms of shared features or combinations); markedness (in terms of relative inherent complexity); and hierarchization – e.g. degree of ‘nouniness’ – (in terms of the relative preponderance of a particular feature or combination). And there, too, I have illustrated how other syntactic properties (syntagms, linearity) are constructed on the basis of such notional characterizations for individual words. It has also been argued that the representation of phonological segments, most analogously of the categorial gesture, displays the same properties (cf. e.g. Anderson 1985a, 1986, 1987a; Anderson and Ewen 1987). Here, I am not concerned directly with these issues, though some of what follows will be relevant to them; rather, we shall be focusing on the role of such a conception of syntax in circumscribing possible syntactic change.

Such a notional theory of categories allows for languages to vary in the number of categorial distinctions made: these are recognized on the basis of the distributional and morphological characteristics of the language concerned, and identified on notional grounds. The minimal system seems to be {P} vs. {N} vs. { } (Anderson 1990).

Languages may also vary in the categories that are contrastive, given lexical status, associated with distinct lexical classes. Thus, many languages have distinct finite and non-finite verb classes, or one class that can be both and one that is only non-finite (with distinctive finites often being called ‘auxiliaries’), while others allow all verbs to occur in both finite and non-finite positions and/or to show both finite and non-finite morphology. Not all the categories appropriate to the syntax of a language will be recognized as classes. Also, though a lexical class will contain as prototypical members items satisfying the notional requirements of the category, there are likely to be items which are distributionally appropriate but notionally less so. Thus, some nouns (‘relational nouns’), such as friend in English, are perhaps as much situation-defining as entity-defining.

Unless otherwise constrained, the theory sets no upper limit on the set of possible categories. It may be possible to establish such a limit: e.g. second-order combinations of the character of (2) may instantiate the maximal complexity to be allowed to categorial representations. (Sub-categorial representations, distinguishing sub-classes whose distribution and/or morphological differentiation is a proper sub-part of that associated with the category as a whole, e.g. gender sub-classes of the noun, are not relevant to this question.) More importantly, however, at least from the point of view of our present concerns, the theory provides a restricted alphabet out of which any categorial representation must be constructed, and the representations themselves provide an inherent measure of complexity. The complexity measure imposes a hierarchy on representations, and on language systems. Systems can be ranked in terms of the complexity of the representations they invoke.

It is a natural property of such a hierarchy of systems that the set of representations invoked by any system be continuous and minimalist, at a contrastive level. We require that the presence in a system of a complex representation presupposes the independent presence in the same system of the components out of which it is constructed (e.g. adjectives presuppose verbs and names). Inter-language variation in categories invoked is not randomly distributed. Let us refer to this property of...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Notes on the Contributors
• Editor’s Preface
• Acknowledgements
• 1 Parameters of syntactic change: a notional view
• 2 Change and metatheory at the beginning of the 1990s: the primacy of history
• 3 Typology and reconstruction
• 4 On the phonetic basis of phonological change
• 5 Internally and externally motivated change in language contact settings: doubts about dichotomy
• 6 How real(ist) are reconstructions?
• 7 Why UG needs a learning theory: triggering verb movement
• 8 On the social origins of language change
• 9 The phonetics of sound change
• 10 Nicaraguan English in history
• 11 Language change as language improvement
• 12 Bidirectional diffusion in sound change
• Index