This thesis is a descriptive study of the Asante dialect of Akan, an indigenous, but the most influential language spoken in Ghana. The main focus is on its phonology. This study is composed of six chapters: Chapter 1 'Introduction', Chapter 2 'Phonology and Morphology of Akan', Chapter 3 'Vowel Harmony of Akan', Chapter 4 'Tone of Akan', Chapter 5 'Rhythm of Akan' and Chapter 6 'Conclusion'.
In Chapter 2 'Phonology and Morphology of Akan', the phonological and morphological structures of Akan are thoroughly reviewed. On phonology, the vowel and consonantal systems are first described. Second, the syllable structure is examined. Third, the type of tone and its function are explained. On morphology, the morphological structures of noun, verb, copula, pronoun, adjective, adverb, conjunction, and clitic are reviewed.
Chapter 3 'Vowel Harmony of Akan' deals with vowel harmony, one of the central themes of this thesis. In this chapter, the following three issues are discussed. Firstly, the asymmetry of the vowel harmony system is dealt with. Among the ten vowels, namely i, I, e, ɛ, a, з, u, 〓, o, ɔ of Akan, the pairs i and I, e and ɛ, u and 〓, and finally o and ɔ make vowel harmonic pairs. However, a and з do not constitute a harmonic pair. But a similar distribution is observed with respect to e and o, that is з, e, and o occur essentially with i or u. This distributional restriction is interpreted as a result of historical change of these vowels. That is, unless these vowels co-occur with i or u, they change into [-Expanded] vowels.
Secondly, vowel assimilation, which takes place between word or root boundaries is discussed. Among previous studies, even between previous studies and mine, descriptive differences are seen about the trigger of vowel assimilation, i.e. whether the assimilation is triggered by i and u only, or also by e and o. I interpret these differences as a result of historical changes, in relation to the distributional restriction of з, e and o in the vowel harmonic domain, and suppose that both vowel harmony and vowel assimilation went under the same historical change, that is, the loss of influence which all [+Expanded] vowels once had on [-Expanded] vowels.
Thirdly, the proclitic, which is regarded as a prefix by all previous studies, is dealt with. There is a difference between the previous studies and mine about the change of the proclitic vowels. Although the proclitic is inevitably subject to vowel harmony in the previous studies, it behaves the same as the word or root in my observation, that is the change of proclitic vowels occur before i, u, e, and o in all the previous studies, and only before i and u in mine. This difference seems to be similar to the difference mentioned above. Therefore, I suppose that this difference on the proclitic is also the result of the same historical change, and that this change just recently occurred. For these second and third hypotheses, some acoustic correlates are shown.
After all these examinations, the conclusion is made first that, on Akan vowels, [+Expanded] is marked rather than [+Expanded] and that [-Expanded] contrasts equally. Second, the vowel harmony and vowel assimilation are basically similar phenomena, and the difference lies on the extent on which [+Expanded] vowels affect surrounding [-Expanded] vowels, according to the morphological or syntactical environment in which these vowels occur, i.e, within root, affix boundary, clitic boundary, root or word boundary.
Chapter 4 'Tone of Akan' deals with tone, the second central theme of this thesis. First, the tone of noun is discussed. It is analyzed that while suffixes have H tone, preffixes have no underlying tone, and their tone is given by rule. The tonal information of a noun root is obtained in reference to its tone melody, together with an optional accent which it may have. Then the tone change in possessive noun phrases is dealt with, and the tone change and the tone shift rule in possessive noun phrases proposed by Stewart (1986b) and Dolphyne (1986) are reexamined.
Second, the tone of verbs is discussed. By showing the surface tone pattern of each of the fifteen verb conjugations, I examine how these surface forms are derived. The conclusion is made that there are four factors that decide these conjugational forms, namely, the underlying tone of each morpheme, the proper tone of the conjugation, tone rules, and accent shift rules.
Thirdly, tones of the structure root+root or word+word are dealt with, and their behavior is analyzed. Fourthly, tone of clitics is discussed, and I argue that there are three types of clitics tonally: one is the same as root or word, the second the same as affix, and the third the same as both root/word and affix. The presence of these three types reflects the fact that clitics originated from roots/words, and are shifting toward affixes.
Chapter 5 'Rhythm of Akan' deals with the third central theme of this thesis, namely rhythm. Proverbs are focused which are abundant in Akan. There is a view that oral traditions like proverbs are usually recognized as such by the speakers of the community, and are recognized primarily because of the presence of metric structure. And this metric structure conforms to the general rhythm of the language which they have been taken down (Taylor 1931, 1985, Ben-Amos 1969). Based on this view, an acoustical approach is made to examine the metric character that makes Akan proverbs sound as such, and the general rhythm of Akan which is the base for the rhythm of these proverbs.