In forensic speaker comparison, there may be cases in which the speech samples to be compared are not in the same language. Research in forensic phonetics has however almost exclusively been done...Show moreIn forensic speaker comparison, there may be cases in which the speech samples to be compared are not in the same language. Research in forensic phonetics has however almost exclusively been done in monolingual contexts. This study aims to give more clarity about what features can be used in cross-linguistic speaker comparison. This is done by investigating whether proficient first language (L1) Dutch speakers of second language (L2) English adapt the vowel in and to the target language. L2 sound learning models predict that Dutch learners will have difficulty learning the difference between /ɛ/ and /æ/ and will therefore only be able to approximate nativelike pronunciation. The vowels in en and and of 35 proficient, female Dutch speakers of English were analyzed, both as a discourse marker and a coordinating conjunction. Mixed-effect models showed that for both word classes, the F1 and F2 were language-dependent. This means that the use of /æ/ as a characteristic in cross-linguistic speaker comparison is not useful.Show less
