Pernilla Hallonsten Halling

Properties deadjectival
nouns
unmarked adjectives
predicate adjectives, copulas

Actions action nominals, participles, relative clauses complements,
infinitives, gerunds
unmarked verbs

In English, the markedness pattern can be exemplified with the property word red, which is unmarked when used in its prototypical function as a modifier in, e.g., the red rose. In a non-prototypical function, such as reference, the same property requires overt structural coding in the form of a derivational morpheme, e.g. the red-ness of the rose. There are also instances of zero structural coding in English object words used for modification, such as kitchen table or apple basket, which show that the object words kitchen and apple are encoded by the same amount of structural coding (i.e. zero) when used in their prototypical function of reference and the non-prototypical function of modification (cf. Croft 2001: 99). This follows the implicational universal of markedness, since the marked member (the object word kitchen or apple used as a modifier) is encoded by the same number of morphemes as the unmarked member (the property word red used as a modi- fier). When red is used in predication, English requires the copula be as in, e.g., The rose is red : here the copula is the structural coding, and constitutes the markedness. The un- markedness patterns importantly highlight prototypical nouns, verbs, and adjectives, and not absolute categories. This means that while language-particular category boundaries differ, these markedness patterns hold cross-linguistically.
The three semantic classes matching the three propositional act functions result in the traditional three major parts of speech labels, as prototypical categories. For property modification, one might nevertheless raise the question of why this points so uniformly towards prototypical adjectives. In the quote above, Croft specifies modification as modi- fication of referents. On the other hand, Croft (2003: 184-185) describes modification as “a secondary propositional act which can aid to establish reference (restrictive modification) or assert a secondary predication (nonrestrictive modification)”. Modification of referents must then correspond to aiding in establishing reference, i.e. restrictive modification.

But there is, apparently, another type of modification, called nonrestrictive modification, consisting of ‘asserting a secondary predication’. Moreover, modification as a whole is a secondary propositional act function. Although it is not quite clear what implications this characteristic of being secondary has (see further discussion in chapter 3), modification is necessarily secondary to reference and predication. Based on this assumption alone, it seems that there must be two types of modification. Croft (2001: 94) also briefly com- ments that the conceptual space for parts of speech can be elaborated: it “only represents modification of a referent; modification of a predicate (adverbial modification) would also have to be represented”. Modification, as such, does then not only pertain to referents, but also to predicates. In an entry on adverbs in Encyclopedia of Language and Linguistics, Haser & Kortmann (2006) elaborate on Croft’s comment on modification as follows:

Prototypical adverbs, much like prototypical adjectives, could then be de- fined as items that provide ‘modification by a property’, the difference being that prototypical adjectives modify referents and prototypical adverbs modify predicates. (Haser & Kortmann 2006: 68)
The fact that adverbs must modify predicates if adjectives modify referents is logical and follows immediately from Croft’s approach. Haser & Kortmann (2006) nonetheless draw further conclusions, arguing that this is a reflection of how closely related adjectives and adverbs are. They illustrate this relation with the example of how adverbs are commonly formed from adjectives, as in the case of the English -ly, and that many languages do not formally separate adjectives from adverbs (2006: 68). Not denying that there is a close connection between adjectives and adverbs, both belonging to the function of property modification, it seems questionable that such a connection should have to be unidirec- tional. Both Croft (2001) and Haser & Kortmann (2006) nonetheless seem to assume that although there are two types of property modification, modification of referents, which is manifested as prototypical adjectives, is in some sense primary. It is difficult to tell whether this focus on modification of referents follows from a tradition that treats nouns, verbs, and adjectives as the three major parts of speech, or from a lack of data on mod- ifiers of predicates. Such a lack could be due either to the absence of function-specific encoding for modifiers of predicates cross-linguistically, or to lack of available data.
A problem that arises at this point is the use of the terms reference and referent, especially in connection to modification in Croft’s approach. For instance, Croft talks about “modification of a referent”, and also “modification of a predicate” (2001: 94). This is somewhat puzzling, since it is an entire expression, including a modifier, that does the referring. In the example the red rose, it is obviously not the referent, i.e. the ac- tual rose, that is being modified, but the referring expression as such. To clarify this, I will use the terms modifier within referring expression, or alternatively, modification of a referring expression for adjectival modification. Similarly, modifier within predicating expression or modification of predicating expression will be used for adverbial modifica- tion. Since these terms are quite long, I may sometimes use the shorter modification in reference/predication.
Modification by a property accurately captures the similarities of adjectives and ad- verbs, whereas the ability of modification to act within either referring or predicating expressions illustrates their different functions. Treating prototypical adverbs as property

modifiers within predicating expressions provides a clear delimitation of what is meant by adverb, as investigated in the present study. Modification on other levels such as of the whole sentence, of an adjective, or of another adverb is not within the scope of this study. Also, only property words qualify as prototypical for this type of adverb – semantic types such as time, space, etc. are not included. In conclusion, although the role of ad- verbs as property modifiers within predicating expressions compared to other prototypical parts of speech remains somewhat unclear, Croft’s approach provides potential for a clear identification of prototypical adverbs of this type. If property modification is found both within referring and predicating expressions, and property modification is where we find unmarked prototypical modifiers, then there must either be both unmarked adjectives and adverbs, or an explanation to why the former should be predominant.

5.