What do disyllabic words tell us about syllable structure, vowel quality, and stress in English?
This thesis explores syllable structure, vowel quality, and stress in English in the light of disyllabic words. Disyllabic words tell us much about these phenomena. Despite the huge literature on syllable structure, vowel quality, and stress in English, there are numerous questions in relation to these phenomena which remain untackeled, and, on the other hand, questions whose proposed answers in the literature are unsatisfactory. In my study of disyllabic words, I used the John Wells’ (2008) Longman Pronunciation Dictionary, as the source of my data. I collected all the disyllabic words in Wells’ dictionary and put them into categories based on their syllabic structure, vowel qualities (in both syllables), and the position of primary stress. Given these factors, disyllabic words have come into eighteen categories with significantly varying sizes. The largest category contains 7989 disyllabic words while the smallest category contains only two disyllabic words. The size of a category is interpreted as reflecting how acceptable a pattern (or a category) is in the phonology of English. I propose an approach to each of the above phenomena. Regarding syllable structure, I employ sonority as a principle of constructing syllables. I argue that an onset in English contains a maximum of three slots while each rhymal constituent contains a maximum of two slots. A coda contains a maximum of two consonants; the initial is necessarily [+sonorant] while the second is [-sonorant]. A second postvocalic [+sonorant] consonant forms what I call a dummy syllable whose nucleus is occupied by a null vowel, unless that consonant fits into the onset of the following syllable. On the other hand, a second postvocalic [-sonorant] consonant that is preceded by another [-sonorant] consonant is linked directly to the syllable node (rather than the coda) unless that consonant again fits into the following onset. As for vowel quality, I suggest that English has four vowel-classes: lax vowels, short tense vowels, long vowels, and schwas, with long vowels being classified into long monophthongs, diphthongs, and centring diphthongs. However, not all these vowels exist in underlying representation. I argue that only lax vowels and short tense vowels do. Diphthongs are combinations of a lax vowel and a short tense vowel or two short tense vowels. However, it is not necessary that diphthongs exist as tautosyllabic sequences in underlying representation. I argue that long monophthongs do not exist underlyingly. Nor do centering diphthongs. Short tense vowels serve as the underlying forms of long monophthongs. Centring diphthongs are merely combinations of a lax vowel and a compensatory schwa. I also argue that there are six schwas in English that contrast as to their underlying qualities. In terms of stress, I argue that in a disyllabic word both syllables may be stressed. The non-primary-stressed syllable in a disyllabic word is subordinate-stressed if it invariably surfaces with a full vowel. I argue that vowel quality (i.e. whether a vowel surfaces as full or as a schwa) is subject to special constraints that regulate the distribution of full vowels and schwas in English words and are irrelevant to stress. I then propose an approach to stress assignment (not only in disyllabic words but in all English words) where I argue that vowel quality is determined prior to and independently of stress rules. Stress rules apply after vowel quality has been determined and assign stress to syllables with full vowels and skip over those with schwas. We also propose a set of rules that are responsible for locating primary stress in words with multiple stressed syllables.