Konstanz, Univ., Diss., 2006.

Esta pesquisa analisa as vogais médias tônicas anteriores [ε] e [e] e posteriores [ɔ] e [o], em formas nominais e verbais na 1ª pessoa do singular e na 3ª pessoa do singular e do plural no presente do indicativo, especificamente, o processo de metafonia das vogais médias /e/ e /o/, as quais se assimilam em /ε/ e /ᴐ/ em posição tônica. Os objetivos gerais desta investigação são descrever e quantificar a ocorrência de metafonia e, posteriormente, analisar em quais palavras há regularidade ou não. Como objetivos específicos têm-se: i) compilar e etiquetar um corpus oral, espontâneo, sincrônico e regional, a partir de programas de rádio produzidos na cidade de Ituiutaba, Minas Gerais; ii) descrever as características do corpus a ser compilado; iii) investigar a alternância do timbre das vogais médias em posição tônica; iv) identificar ocorrências de metafonia nominal e verbal das vogais médias em posição tônica; v) descrever os casos de metafonia nominal e verbal identificados; vi) analisar as prováveis causas para a variação das vogais médias. Para a realização da análise proposta, adotamos como base teórico-metodológica modelos multi-representacionais: a Fonologia de Uso (BYBEE, 2001) e a Teoria de Exemplares (PIERREHUMBERT, 2001), aliados aos preceitos da Linguística de Corpus (BEBER SARDINHA, 2004). O corpus está constituído de 16 programas radiofônicos – oito políticos e oito religiosos – produzidos na cidade de Ituiutaba-MG, com gravações de aproximadamente 20 a 40 minutos cada programa. Constatamos, por meio dos dados gerados pelo programa WordSmith Tools®, versão 6.0 (SCOTT, 2012), que as formas analisadas apresentam pouca variação, o que mostra que a metafonia é um processo já lexicalizado para os participantes dos programas radiofônicos analisados. Concluímos que os resultados convergem com a proposta da Fonologia de Uso (BYBEE, 2001; PHILLIPS, 1984) de que as palavras menos frequentes, que não têm ambiente fonético propício a mudanças, são alteradas primeiramente. / This research analyzes the average previous stressed vowels [ε] and [e] and later [ɔ] and [o] in nominal and verbal forms in the 1st person singular and 3rd person singular and plural in the present tense, specifically the umlaut process of mid vowels /e/ and /o/, which assimilate in /ε/ and /ᴐ/ in stressed position. The general objective of this research is to describe and quantify the occurrence of umlaut and subsequently analyze in which words there is regularity or not. As specific objectives we have: i) to compile and to label an oral, spontaneous, synchronic and regional corpus, from radio programs produced in the city of Ituiutaba, Minas Gerais; ii) to describe the characteristics of the corpus to be compiled; iii) to investigate the alternating timbre of mid vowels in stressed position; iv) to identify instances of nominal and verbal umlaut of the middle vowels in stressed position; v) to describe the identified cases of nominal and verbal umlaut; vi) to analyze the probable causes for the variation of the middle vowels. To perform the proposed analysis, we have adopted as a theoretical-methodological basis multi-representational models: Phonology of Use (BYBEE, 2001) and Exemplar Theory (PIERREHUMBERT, 2001) combined with the precepts of Corpus Linguistics (BEBER SARDINHA, 2004). The corpus consisted of 16 radio programs – eight political and eight religious – from the city of Ituiutaba-MG, with recordings of about 20 to 40 minutes. We note, by means of the results generated by WordSmith Tools® software, version 6.0 (SCOTT, 2012), that the analyzed forms show little variation, which shows that the umlaut is a process already lexicalized in participants of the radio programs analyzed. We conclude that the results converge with the proposal of the Phonology of Use (BYBEE, 2001; PHILLIPS, 1984) that less frequent words that have no phonetic environment conducive to change, are changed first. / Dissertação (Mestrado)

The runic alphabet in the Viking Age (ca. 800-1100) only consisted of 16 characters. These did not cover the Rune-Swedish phoneme system. One of the runes, the so-called os-rune, was used for at least 8 different phonemes. By tradition it has been used to date Viking Age runic inscriptions in the Scandinavian countries. The present work investigates the phonetic and phonemic value of the os-rune in Rune-Swedish inscriptions on stone, altogether 1,745 instances in 961 inscriptions. On the basis of the analysis, the accepted view of the chronological value of the os-rune is assessed, and is found to be non-valid. The regional variations in the use of the os-rune are also studied, particularly in the context of individual rune-carvers, but also to see if any dialectal variation is discernable. It is found that there is a marked consistency in the use of the os-rune for either rounded or unrounded vowels. Inscriptions, where both uses occur, are concentrated to one small geographical area. There are also areas, where the use of the os-rune for rounded vowels is the rule. / <p>Digitaliserad utgåva 2016, kompletterad med en lista över viktigare rättelser</p> / De vikingatida runinskrifternas kronologi