Definiteness

This final chapter will examine a number of more general theoretical implications arising from the present study. The first of these has to do with the theory of definiteness as a whole.

The analysis of definiteness I have put forward is a componential one. That is, it argues that definiteness is not a primitive, unitary concept, but is itself analysable into components. In both languages one such component has to do with quantity: quantity (±all) is thus seen as a primary term, and definiteness as a secondary one. Another component (locatability) has to do partly with what has loosely been termed reference (but see below, 9.2). A further component which is relevant to definiteness in English is extensivity, which roughly means the degree of abstraction or generality of a concept (see 2.3).

These components are of rather different kinds, partly pragmatic and partly semantic (see 9.4 below). In Hawkins' (1978) location theory, which has been a major source of inspiration for the present study, definiteness is also seen componentially, in terms of locatability and inclusiveness. Hawkins sees locatability as a pragmatic feature and inclusiveness as semantic. However, I have shown (2.2.2) that inclusiveness is also largely a matter of pragmatics. The only component that one might argue to be purely semantic is extensivity, which is not included in Hawkins' analysis at all.

The realization of definiteness

The general structure and purpose of this study is Popperian. The starting-point is neither a corpus of data nor a particular theoretical position, but a problem, or rather a set of problems. My basic strategy is first to describe the problems, then consider the solutions that have been proposed to date. I then try to show in what respects these solutions are inadequate, how they neglect or misinterpret certain kinds of data and finally suggest my own analysis. In Popper's terms I thus proceed from Problem to Hypothesis to Test to New Hypothesis (see Popper 1972). It then remains to be seen what new problems and further, more refined, hypotheses will follow from where this study leaves off.

The general issue is: what is definiteness? This is first of all a semantic, conceptual question: what does the term ‘definiteness’ mean? Consequently, and more specifically, we may ask: what does it mean to say that something is definite or indefinite?

A second level of discussion then has to do with recognition: given a definition of definiteness, definite and indefinite, how do we actually recognize that something is definite or not? This is a matter of realization, in two senses of the word.

Early studies

The general category of definiteness first appears in Finnish grammar under the name of spesies ‘species’, a term introduced by Noreen (1904). (See Hirvonen 1980 and Vähämäki 1980 for general surveys.) Noreen had distinguished three categories of spesies for Swedish, which he called definite, indefinite and general. The Finnish Language Commission adopted the term spesies in their influential 1915 report, but defined the category as having only two members: definite and indefinite. Nouns whose referents were known or previously mentioned were said to have definite spesies, and nouns with referents which were unknown or not previously mentioned had indefinite spesies (Kielioppikomitean mietintö 6, 1915: 38). ‘Known’ was later redefined to include ‘known by virtue of the situation’.

From the start it was clear that ‘in Finnish there is no one way of expressing the category of spesies which could be compared e.g. to the articles of many Indo-European languages’ (Ahlman 1928: 134; my translation). Research therefore focused on discovering the range of grammatical devices which could be used to express this opposition of known vs unknown referent. The research strategy was basically what Catford (1965) would later call commutation: a given syntactic feature of a Finnish sentence was varied, resulting in two different English, German or Swedish translations, one with a definite and one with an indefinite article.

Divisibility and quantity

I start by reconsidering the basic notions of divisibility and quantity in Finnish.

At first sight the divisibility distinction in Finnish appears straightforward, separating singular count nouns from plural count and mass nouns. Yet certain definiteness readings remain problematic. Because divisibility is a semantic concept many nouns can be used in both divisible and non-divisible senses, as in English: for instance kakku ‘cake’. In its divisible usage the word will allow the expression of partitive quantity, but in its non-divisible use it will not. In this latter usage, a partitive would have some other cause, not partial quantity. Recall examples such as the following:

Auto is non-divisible, and the partitive signifies irresultative action; the definiteness of the car is determined by the context.

The verb is resultative and juusto is divisible: it must therefore be read as some cheese, indefinite. But now consider (3):

Kakku can be divisible or non-divisible, so that three readings are possible (out of context, of course):

1. (a) Non-divisible, partitive due to irresultative verb: ‘We were eating a/the cake.’

2. (b) Divisible, partitive marking partial quantity; verb resultative: ‘We ate (some) cake.’

3. […]

Towards a tertium comparationis

I have so far been discussing English and Finnish separately, partly in terms of the respective research traditions in the two languages (chapters 2 and 6). My own conclusions have also been stated independently for each language (chapters 4 and 7). It will now be instructive to set the two descriptions side by side. A unified analysis of definiteness in both languages would make explicit any correspondences between the two, and also raise hypotheses about the theory of definiteness in general (a topic to be taken up in chapter 9).

But before we proceed to a contrastive analysis a general caveat is in order. Throughout this study – and even more so in the present chapter – I make one assumption which should now be made explicit: I have implicitly ‘reified’ definiteness. That is, I have been taking definiteness as some kind of independent ‘phenomenon’ (albeit a composite one), which manifests itself in various ways in different languages. The implication has been that definiteness is a potential semantic universal. At the level of analysis at which I am operating I hold this assumption to be valid and helpful. But beneath this abstraction there lies an issue that marks practically any linguistic argument: the relation between the technical linguistic terms (such as ‘definite’) and the empirical reality they purport to describe.

Oppositions

At this point it will be useful to step back for a moment and review the major points that have so far emerged. They can be summarized as follows.

1. (a) Standard descriptions of the articles have too many exceptions, particularly regarding the cases where the is said to be omitted before count singulars, and where a can be used before mass nouns. Proper nouns are also held to be outside the article system proper.

2. (b) Generic NPs with different articles do not have uniform readings, and the term ‘generic’ itself lacks a unitary definition.

3. (c) There is good evidence that nouns with ‘no article’ fall into two distinct types: indefinite mass and plural nouns on the one hand, and count singular nouns (and singular proper names) on the other.

4. (d) It is not a priori clear what the category ‘article’ comprises. In addition to NPs with the, a and some, there is reason to claim that there are also two distinct types of NP which are not preceded by any realized surface article: these we have called the zero and null forms.

5. (e) ‘Definiteness’ and ‘indefiniteness’ are not simple polar opposites. Indefiniteness is particularly complex, in that the distribution and meaning of a, some and zero are not adequately explained merely in terms of the features [±count] and [±singular]. In particular, zero and some are by no means always in free variation.

6. […]

Noun classes

A comprehensive description of the articles must obviously incorporate answers to the following two questions:

1. (a) Which kinds of nouns may in principle take which article(s)?

2. (b) Under what circumstances may – or must – a given noun or noun type take a given article?

For any given noun, that is, we must be able to specify first which articles are possible out of context, and then which of the possible articles is the appropriate one within a given context.

The first of these questions has to do with the classification of noun types. Before considering the distribution of the articles in context, then, a brief discussion of the relevant noun classes will be in order.

What modern standard grammars (such as Quirk et al. 1972, 1985) refer to as the count/mass distinction derives largely from the ideas of scholars such as Jespersen and Christophersen. Jespersen (1949) distinguished ‘mass-words’ from ‘unit-words’ on the grounds that mass-words take both the zero article and the, whereas unit-words (in the singular) cannot take zero; both these are distinct from proper names, which take (what he called) zero only. Earlier (in volume II), Jespersen had used the terms ‘countables’ and ‘uncountables’: i.e. the opposition was then defined semantically, having to do with whether or not a given concept could be counted.

Background

This chapter brings together a wide range of research trends on the articles. In the course of the discussion a number of questions will be raised that do not yet seem to have been satisfactorily resolved, such as the following: Is definiteness a simple binary opposition? What do ‘definite’ and ‘indefinite’ mean? How many articles are there? What does ‘no article’ mean? What is the individual meaning of each article? The underlying theme of the chapter is that these questions can only be answered adequately within a theory of the articles that will incorporate insights from several research traditions into a coherent whole, for each of the existing descriptions is in some way or other too restricted.

In very general terms, modern research into the English articles has tended to fall into one of three broad types. One approach, starting with Russell (1905), centres on the meaning of definiteness and the expression of this meaning throughout the grammar. A second approach is illustrated by studies in the generative tradition: here the concern is with the correct derivation of the articles, the rules that will generate the correct article in a given context. The third approach has been to start with the articles themselves, their distribution and meaning, and also with the question of which forms should actually count as articles.

Adjectives

At the general level, ‘adjective’ is applied to a grammatically distinct word class in a language having the following properties:

(a) It contains among its most central members the morphologically simplest words denoting properties or states; among the most frequent and salient are those relating to size, shape, colour, age, evaluation (“good”, “bad”, etc.) and the like.

(b) Its members are characteristically used either predicatively (very often as complement to the verb “be”) or attributively, as modifier within NP structure.

(c) It is the class, or one of the classes, to which the inflectional category of grade applies most characteristically in languages having this category. (Adjectives often carry such other inflections as case, gender, number, but secondarily, by agreement, rather than being the primary locus for them.)

Whereas it is generally accepted that all languages distinguish grammatically between nouns and verbs, not all languages have a distinct adjective class. In languages which do, there is a tendency for verbs to be dynamic (denoting actions, events, etc.), adjectives static; note in this connection that in English adjectives generally occur very much more readily in the non-progressive construction than in the progressive, whereas this is not so with the majority of verbs (so that while Ed moved and Ed was moving are equally natural, Ed was tall and Ed was being tall are not).

Adverbs

At the general level the adverb is definable as a grammatically distinct word class with the following properties:

(a) Its central members characteristically modify (or head phrases which modify) verbs, adjectives and other adverbs. In languages which distinguish between adjectives and adverbs the primary difference is that adjectives modify nouns (or stand in a predicative relation to noun phrases) while adverbs modify verbs; the modifiers of verbs can, to a very large degree, also modify adjectives and adverbs, so that we then extend the definition of adverb to cover modifiers of all three open classes other than nouns.

(b) Central members commonly express manner or degree; other frequent meanings (often associated with grammatically less central members) include time and place.

(c) It is commonly the case that many members, especially those belonging to the manner subclass, are morphologically derived from adjectives.

Consider now, at the language-particular level, the properties of adverbs in English.

(a) Function. Adverbs, or the phrases they head, occur in a rather wide range of functions, notably (though not exhaustively): (α) modifier in VP structure ([She spoke] clearly), (β) modifier in AdjP structure ([She's] extraordinarily [bright]); (γ) modifier in AdvP structure ([She did it] rather [well]); (δ) peripheral dependent in clause structure (Frankly, [he's a dead loss]); (ε) complement in VP structure ([They put us] ashore).