Files
Abstract
This study is the first to investigate the effects of orthography on the phonetic production of spontaneous speech. It is also unique in that it utilizes corpus speech data. The speech utilized for this study is extracted from the Buckeye Speech Corpus and the Santa Barbara Corpus of Spoken American English. Homophone pairs with differing orthographic vowels were compared acoustically in tokens from these corpora. The results of two-tailed t-tests show that speakers may have different pronunciations in words with different orthographic vowels and the difference appears primarily in the F2. However, many of the control pairs were also found to be significantly different in their acoustic measurements. Therefore, any significant changes to the acoustics of the homophone productions could be caused by a variety of factors not necessarily including orthography. Further investigation with more homophone pairs and greater token numbers is required to clarify these findings.