Pernilla Hallonsten Halling

Summary and conclusion

This chapter set out to consider the issues that surround the determination of adverb as a part of speech category in general. Different approaches to parts of speech were then discussed, with a particular focus on typological accounts. The approach described in greatest detail was that of prototypical categories based on discourse functions (Croft 1991, 2001, 2003), which is also the basis for my discussion of adverb as a part of speech. Based on the results of the present study, adverbs were established as constituting a cross-linguistically prototypical part of speech. The nature of such a category, and the peculiarities that follow from it, were examined in depth. In conclusion, adverbs constitute a divergent category in being less frequent, displaying less inflection, and not acting as a

³ As discussed in chapter 5, it is in fact doubtful whether even ruttu is really a simple adverb, since it could also be treated as a case form of the Noun rutt ‘haste’ cf. Wiedemann (1973).

9.5. Summary and conclusion

base for derivation. These divergences are not inexplicable, but are rather characteristic of the category, and they elucidate its nature. The small number of semantic types of adverbs are restricted by the fact that not many properties can be ascribed to events without taking into account the complexity of events in terms of participants, tense, and aspect. This is also why property words used as adverbs often shift meaning towards particular characteristics of events. Even with all its peculiarities, adverb is attested as a prototypical part of speech, in its unmarked occurrences and semantic types cross- linguistically.

8. Concluding discussion

   1. Summary and evaluation

The aim of this dissertation was to carry out a typological study of adverbs, in order to examine their encoding across languages. The encoding of adverbs was also compared to that of attributive and predicative adjectives, respectively, in order to see to what extent the encoding of these three functions overlaps. A worldwide sample consisting of 60 languages was used. The analysis was carried out at three different levels: the root, the lexeme, and the construction. Beside capturing the levels of encoding as such for each language, this analysis allowed us to compare the types of encoding overlaps in an articulated manner. For the adverbial function specifically, I examined to what extent simple adverbs can be found in the sample languages, and more specifically, whether they can be found in languages that lack simple adjectives. For the languages that have simple adverbs, I looked at the semantics of these adverbs in order to find out whether it varies or is constant cross-linguistically (cf. the research questions formulated in chapter 1). In the sections that follow, the results will be summarized and evaluated, based on the adverbial function (section 10.1.1), overlaps at the root and lexeme levels (section 10.1.2), overlaps at the construction level (10.1.3), and semantic types and semantic shift (section 10.1.4). Implications of these findings will then be discussed in section 10.1.5, where two implicational universals will also be established. Prospects for future research are treated in section 10.2, and the thesis is concluded in section 10.3.

1.