Relationship between the strength of Czech accent and the duration of vowels before obstruents

This bachelor thesis aims to inspect the presence or absence of a feature called pre- fortis shortening in English spoken by Czechs. The term denotes the shortening of a vowel preceding a voiceless obstruent. This feature is known to appear in various languages like Russian, French, Italian; indeed, Matthew Chen even suggests it is language-universal. In English the feature is very prominent and because it affects speech perception, it is even considered a primary indicator of the voicing of the following obstruent. A study included in this thesis examines the extent to which 12 female speakers of Czech English, sorted into 3 categories according to their proficiency in pronunciation, mark the distinction between words like "bet" and "bed" by vowel shortening. The study does not exploit minimal pairs like these, but vowel-obstruent sequences taken from long read passages of BBC news. The the fortis/lenis character of the final obstruent contained in these sequences is a discerning parameter which always separates the data into two groups, the vowel durations of which are then compared. Statistical analysis of the data showed that contrary to the expectations, speakers with native-like pronunciation were not the ones who displayed the most massive usage of pre-fortis shortening. Possible reasons for..