1.3.4.3 Adjectives

We now turn to the [–F, +N, +V] category: adjectives. As their categorial features suggest they share properties with both nouns and verbs, though obviously differ from both.

Adjectives tend to describe states, properties or attributes of things, though as usual, one needs to be careful with semantic definitions of syntactic categories. This category tends to be used in one of two ways; either as a modifier of a noun or as a predicate in a sentence:

This observation, however, will also require modification once we start to consider adverbs and their relationship to adjectives.

The morphology of adjectives is an interesting area, though slightly more complex than that of verbs and nouns. There are three main adjectival morphemes which we might use to identify members of the category. First, many adjectives have three distinct forms relating to the straightforward adjective (traditionally called the positive form), the situation in which two elements are compared with respect to the property expressed by the adjective (the comparative form) and the situation in which more than two elements are compared (the superlative form):

Although there are few irregular adjectival inflections for comparative and superlative (many – more – most, good – better – best, far – further – furthest being obvious examples), there are a number of adjectives which do not take part in this morphological paradigm at all. One class of adjectives that do not have comparative or superlative forms are those which cannot be used for the basis of comparison from a semantic point of view. Obviously, the notion of comparison involves properties that can be graded into more or less: the property long, for example, covers a whole range of lengths, some longer some shorter. A long piece of string could be anything between, say 1 metre and infinitely long. We can therefore compare two elements in terms of their lengths and determine that one is longer than the other. Some adjectives however, do not express properties that can form the basis of comparison: some states such as being dead or being married are absolute or ungradable, so someone cannot be more dead or more married than someone else. Clearly ungradable adjectives are not going to have comparative or superlative forms:

In the above cases there is a semantic explanation for the lacking forms. In other cases however, there are other explanations. Quite a few adjectives are morphologically complex, being derived from nouns or verbs. It seems that morphologically complex adjectives cannot bear the comparative and superlative morphemes:

There are, however, certain exceptions to this:

It seems that adjectives formed with either ‘ y’ or ‘ ly’ are able to take ‘ er’ and ‘ est’.

However, unlike the case of the ungradable adjectives, we can express comparative and superlative notions with morphologically complex adjectives using degree adverbs more and most:

These are known as the periphrastic comparative and superlative constructions as opposed to the inflectional ones. Often it is the case that adjectives participate in either one or the other of these constructions, though there are some adjectives that can appear with both:

We need not go into this any further. The main point that concerns us here is that the less productive nature of these adjectival morphemes makes them less reliable as a test for adjectival status than we have seen in the case of verbs and nouns. Obviously, if a word can appear in a comparative or superlative form, it is an adjective, but failure to do so cannot automatically lead us to a negative conclusion.

Another morpheme closely associated with adjectives is ly. This is used with a large number of adjectives to form adverbs:

There is some debate about the status of this morpheme which revolves around the central issue of our present discussion. On the one hand, ly might be taken as a derivational morpheme which is applied to a lexical item of one category to derive another lexical item of another category. This would be similar to morphemes such as er in cook er, ic as in scen(e) ic or ment in govern ment. We have not been concerned with such morphemes so far as they tend to be rather restricted, applying to certain lexical items of a given category rather than to the category as a whole. There are, for example, no forms *exister (someone who exists), *viewic (the property of resembling a nice view) or *rulement (the collective body of people who rule). As we have been concerned in using morphological observations for identifying categories, the derivational morphemes would have been only of limited use to us. The important point about derivational morphology is that it takes place in the lexicon, forming new lexical elements from others, prior to any grammatical operation. If ly is a derivational morpheme, then adverbs are a different category from the adjectives they are derived from. However, we have no feature analysis for adverbs using the [±F], [±N] and [±V] features and as we have pointed out we cannot just introduce a new category into the system without there being some fairly substantial consequences. If we introduce a new feature to try to accommodate adverbs, we predict the existence of a further seven more categories for which we have very little evidence.

However, ly is strangely productive for a derivational morpheme, applying to many adjectives, though there are exceptions:

Yet we can explain many of these absent forms. For example, while the form fastly does not exist, the form fast can be used as both an adjective and an adverb:

In many ways, then, this is like the missing plural *sheeps or the missing past tense *putted (as past tense of put, not putt, which is putted). As such fast is just an irregular adverb. In general, colour adjectives do not tend to form adverbs and the fact that this is a semantically well-defined class of adjectives indicates that there might be semantic reasons for it. This is further supported by the fact that colour adjectives that do form ly adverbs, such as blackly, do so only if they have meanings that go beyond reference to the colour: blackly means ‘in a sinister or evil way’ and greenly can mean either innocently or enviously. Admittedly, the absence of size adverbs like *bigly and *smally is problematic given the existence of hugely and minutely. But putting this small number of problematic cases to one side, we can see that the ly morpheme is a very productive one, applying to most adjectives. As pointed out above, most derivational morphemes, being lexical in nature, are not productive and apply only to selected lexical items.

The alternative to viewing ly as a derivational morpheme it to see it as an inflectional morpheme. These are morphemes like the ones we have been mainly concerned with so far. These apply to a lexical word to give back another form of the same word. So see, sees, saw, seen and seeing are all forms of the same word, not different words created from a single source as are depart, department, departmental, departmentalisation. Inflectional morphemes on the whole are a lot more productive than derivational morphemes (though we have seen a certain degree of irregularity and exceptions in most of the morphemes we have investigated) and this would seem to fit better the productive nature of ly. However, in what sense can an adjective and its related adverb be considered different forms of the same word, especially if they belong to different categories? If ly is an inflectional morpheme, it seems that we would have to consider adjectives and adverbs to be the same category. There is a certain amount of evidence in support of this view however. First, note that both adjectives and adverbs have similar distributions, if we consider their immediate environment:

We see from these examples that the same kinds of words (very, as, too, so, etc. – known as degree adverbs) are used to modify both adjectives and adverbs. Such things cannot be used to modify words of other categories:

Thus, it seems that adjectives and adverbs are closely related categories if they are not the same category. Of course, over a larger domain adjectives and adverbs do not distribute the same: adjectives tend to modify nouns and adverbs modify verbs or whole sentences:

Yet, if we consider the total set of possible positions for adjectives and adverbs, we notice that where an adverb can appear an adjective cannot and vice versa. In other words, the two are in complementary distribution, just like transitive and intransitive verbs. In the case of verbs we took their complementary distributions to be evidence that they are of the same category and, therefore, there is no reason why we should not argue the same here in relation to adjectives and adverbs.

As a further observation, adverbs, like adjectives, can appear in contexts of comparison and hence in comparative and superlative constructions:

Adverbs, however, tend not to have comparative or superlative forms:

The reason for this is probably because these adverbs are morphologically complex and as we have seen morphologically complex adjectives tend not to have such forms. This is supported by the fact that adverbs not formed with the ly morpheme can have morphological comparative and superlative forms:

Interestingly, for adjectives the derivational morpheme ly does not block the comparative and superlative morphemes, as we have seen:

This would seem to suggest that the two morphemes have different statuses and as the adjectival ly is clearly a derivational morpheme, we might use this to argue that the adverb ly is inflectional.

A final argument for seeing adjectives and adverbs as being of the same category has to do with the system of categorisation introduced in the preceding section. Above we pointed out that while verbs are able to take nominal complements, nouns are not. Adjectives are like nouns in this respect. For example, when we derive an adjective from a transitive verb, the adjective must take a prepositional complement, not a nominal one:

All adjectives are like this, even those not derived from verbs:

If we assume that this property is related to the [+N] feature, then we can account for why nouns and adjectives pattern alike in this respect, as both are [+N] categories. Note that prepositions and verbs, the [–N] categories, can have nominal complements. Adverbs behave like nouns and adjectives in not being able to have nominal complements:

We shall see a little later that the question of what complements adverbs can take (and when) is a complex issue. However, as they never take nominal complements under any circumstances it is safe to assume that they are, like adjectives and nouns, a [+N] category. As adverbs are thematic categories they are also [–F] and thus they have either of the following feature specifications:

The feature set in (118a) is that of nouns and we have no reason to believe that adverbs are a type of noun. We are therefore left with the feature set (118b), which is that of adjectives. Hence it seems we are forced to accept that adverbs and adjectives are of the same category by the system we have devised.

The difference between adjectives and adverbs is in how they are used: a [–F, +N, +V] category that is used to modify a noun is called an adjective and one that is used to modify a verb or a sentence is called an adverb. That they often have different forms is not by itself a problem, as there are certain nominal elements, for example, that have different forms depending on how they are used:

When a pronoun follows a verb, it has one form and this differs from the form it has before the verb. We will return to this in more detail in subsequent chapters, but for now all that is important to note is that these elements have different forms in different positions, but we do not therefore conclude that they belong to different categories. We will assume something similar is going on with the [–F, +N, +V] categories and that adjectival and adverbial forms are different forms of the same category determined by its use.

Though we will maintain the traditional terms for adjectives and adverbs, as there has not been a common term developed for them (Radford 1988 has suggested Adjerb or Advective, but surprisingly they did not catch on!). However, we will use the general category label A to stand for this whole category.

Finally in this section, we turn to the subcategorisation of adjectives and adverbs. We start with adjectives as these are the most straightforward. We have already seen that adjectives, like nouns, cannot take nominal complements. However, all other possibilities are open to them:

The lexical entries for these adjectives might therefore be:

The arguments of the adjectives are included as optional to allow for their non-predicative use. When an adjective is used to modify a noun, it does not typically appear with its arguments:

The subcategorisation of adverbs is a rather more tricky issue. One would have thought that if adverbs are formal variants of the relevant adjective, then they would subcategorise in the same way as these adjectives – like the present tense verb subcategorises in the same way as the past tense verb. There are some cases where this might well be true:

In this example, both the adjective (independent) and the adverb (independently) take the same prepositional complement. In other cases, however, this does not seem to work:

These observations raise a number of perplexing questions. Why, for example, do adjectives and adverbs differ in this way? And why are some cases of adverbs with complements ok? Comparing (124) with (125) we can see a difference in the functions of the adjectives and adverbs: whereas the adjectives are functioning as the predicate of the sentence, the adverb plays a modifying role, modifying the verb in these cases. It turns out that when adjectives function as modifiers, they also cannot take the complements that they usually can:

Thus, it turns out that this is not a difference which divides adjectives and adverbs, but a property that unifies them. Under what circumstance can an adverb have a complement then? If what we said above is correct, we predict that adverbs can only take a complement when they do not function as modifiers. This is indeed true in (123b) where the adverb functions as a complement of the verb. It is quite unusual to find an adverb in a non-modifying role and, therefore, it is not at all usual to find adverb with complements.