The intricacies and exceptions of high vowel deletion in Old English have been the subject of much debate in recent historical phonology. Traditional philological handbooks such as Campbell (1959) describe the process within the assumptions of the Neogrammarian tradition. As such, high vowel deletion has been described as a phonological process that removes historically high and synchronically unstressed vowels after a heavy syllable, or two light syllables. However, the descriptions in these handbooks also reveal that exceptions are common, and as per the Neogrammarian tradition, these are usually assumed to be the result of analogy. In contrast, recent studies have sought to account for the exceptions in a way that lends more explanatory power (e.g. Stratal Optimality accounts including Bermúdez-Otero 2005). Such accounts have shown that there is more to the exceptions than analogy, and that phonological rules, as their synchronic activity declines, can become entangled with other morphological and phonological conditioning, due to the high levels of surface opacity that causes them to become unlearnable. Many of the accounts of high vowel deletion have focused on the West Saxon of Alfred (Early WS) and Ælfric (late WS), and recent descriptions of high vowel deletion have largely focused upon the noun declensions (e.g. Bermúdez-Otero in prep) and the weak verb preterites (Minkova 2012). In this study, I focus in particular upon the behaviour of high vowel deletion in the strong and weak verbs; including the past participles and both the present and preterites. The selected data represent the Early West Saxon dialect and also the Late Northumbrian dialect found in the Lindisfarne Gospel gloss. Discussion of the process as found in nouns and adjectives will also be incorporated. The study has two larger aims: 1. To provide an analysis of syncope for newly collected data sets from Early West Saxon and Lindisfarne verbs, and 2. To contribute to the debate surrounding how to account for morphophonological interaction within inflexional paradigms. The data reveal evidence to show that high vowel deletion is indeed suffering from the demise of its original phonological conditions in the verbs. It is not argued however that full lexicalization has yet taken place throughout the verbs. Instead, the data present a range of degrees of morphologization, within which the original phonological conditions have become supplemented by additional morphological conditions. Additional phonological conditioning is also in evidence. The Lindisfarne strong past participles, it is argued, represent a morphological category within which weight-based syncope is synchronically blocked. The wider question of how and why morphological and phonological conditions come to be added to existing phonological processes is addressed, and I argue that such phenomena result from unsustainable levels of opacity in the grammar (Anderson 1989), and that a theoretical framework that allows for the interaction of phonology and morphology within the grammar is necessary. The Optimality Theoretic analyses proposed in this study have the benefit of accounting for instances of phonologization through constraint interaction. It is also argued that the ways in which morphological category determines a) the way in which a phonological condition applies, and b) whether it applies at all, is best analysed using cophonological analyses (Anttila 2002a etc.).

This dissertation is an analysis of the segmental phonology of the Shangani language as spoken in the South Eastern parts of Zimbabwe. It starts by presenting the language situation in Zimbabwe and comparing the language’s status in Zimbabwe with that of its sister varieties in South Africa where it is referred to as Tsonga and in Mozambique where it is referred to as XiChangana or Changana. The dissertation is based on data collected from the speakers of Shangani using a variety of research techniques. The dissertation identifies and characterizes the language’s distinctive phonemes using the minimal pair and set tests. It presents the language’s consonants, which include aspirated, breathy-voiced, pre-nasalized, labialized and palatalized consonants. It shows that in Shangani, voiceless consonants cannot be pre-nasalized and that there is an incompatibility between that labio-velar glide /w/ and most labial consonants excpt /m/. The phonemes are analysed using Chomsky and Halle’s (1968) distinctive feature theory. The study uses Clements and Keyser’s (1983) CV phonology of the syllable structure to analyse the language’s syllable structure. The language’s canonical syllable structure is CV. It is also shown that consonant clusters are gaining their way into the language through borrowing from English, Afrikaans and other languages that have consonant clusters in their inventories. Onsetless Vs are marginally attested word-initially. In agent nouns, VV sequences are in most cases retained. These sequences are not analysed as diphthongs since they occupy different V slots on the syllable tier. The second vowel in the sequence is the onsetless syllable. Affricates, NCs, Cws and Cjs are presented as unitary segments that occupy a single C slot of the CV tier. Phonological processes that are attested in the language are also presented. Secondary articulation, vowel deletion, feature spreading, vowel coalescence and nasalization are shown to be the most common phonological process in the language. Since Shangani has the CV syllable typology, most of the phonological processes are there to resolve hiatus that would have been induced by suffixation of vowel commencing stems or suffixes to vowel final prefixes or stems. The notion of domains is shown to be a diagnostic tool for identifying a process in a hiatus situation. The study shows that vowel deletion is the least preferred strategy when secondary articulation, feature spreading, vowel coalescence have been blocked by some constraints like syllable structure processes or the language’s phonotactics / African Languages / D. Litt. et Phil. (African Languages)

This dissertation is an analysis of the segmental phonology of the Shangani language as spoken in the South Eastern parts of Zimbabwe. It starts by presenting the language situation in Zimbabwe and comparing the language’s status in Zimbabwe with that of its sister varieties in South Africa where it is referred to as Tsonga and in Mozambique where it is referred to as XiChangana or Changana. The dissertation is based on data collected from the speakers of Shangani using a variety of research techniques. The dissertation identifies and characterizes the language’s distinctive phonemes using the minimal pair and set tests. It presents the language’s consonants, which include aspirated, breathy-voiced, pre-nasalized, labialized and palatalized consonants. It shows that in Shangani, voiceless consonants cannot be pre-nasalized and that there is an incompatibility between that labio-velar glide /w/ and most labial consonants excpt /m/. The phonemes are analysed using Chomsky and Halle’s (1968) distinctive feature theory. The study uses Clements and Keyser’s (1983) CV phonology of the syllable structure to analyse the language’s syllable structure. The language’s canonical syllable structure is CV. It is also shown that consonant clusters are gaining their way into the language through borrowing from English, Afrikaans and other languages that have consonant clusters in their inventories. Onsetless Vs are marginally attested word-initially. In agent nouns, VV sequences are in most cases retained. These sequences are not analysed as diphthongs since they occupy different V slots on the syllable tier. The second vowel in the sequence is the onsetless syllable. Affricates, NCs, Cws and Cjs are presented as unitary segments that occupy a single C slot of the CV tier. Phonological processes that are attested in the language are also presented. Secondary articulation, vowel deletion, feature spreading, vowel coalescence and nasalization are shown to be the most common phonological process in the language. Since Shangani has the CV syllable typology, most of the phonological processes are there to resolve hiatus that would have been induced by suffixation of vowel commencing stems or suffixes to vowel final prefixes or stems. The notion of domains is shown to be a diagnostic tool for identifying a process in a hiatus situation. The study shows that vowel deletion is the least preferred strategy when secondary articulation, feature spreading, vowel coalescence have been blocked by some constraints like syllable structure processes or the language’s phonotactics / African Languages / D. Litt. et Phil. (African Languages)