<p dir="ltr">This dissertation explores clear speech which is defined as a specific speaking style people adopt when fluent communication could be compromised for various reasons. Although acoustic properties and their perceptual benefits of clear speech produced by monolingual speakers of English are well documented, there has only been a small body of past research on clear speech produced by non-native speakers despite its importance in informing second language (L2) speech fluency (Lindblom, 1990). Aiming to address the gap, I examined English clear speech produced by native Korean speakers from three different perspectives: its acoustic properties, its perceptual benefits, and crosslinguistic influence in the production of clear speech. Together, the broad question addressed in this dissertation concerns a communicative ability of native Korean speakers in their L2.</p><p dir="ltr">The first experiment investigated how L1 Korean speakers (n = 30) produce clear speech in their L2, English, compared with native English speakers (n = 20) in a laboratory setting where they read a list of English words. I analyzed acoustic parameters of clear speech that could be considered language-universal (e.g., vowel lengthening), and those which could be employed in a more language-specific manner (e.g., enhancement of the phonological voicing contrast). The results indicated that Korean speakers produced clear speech acoustically distinct from casual speech in every property. Furthermore, the directions of acoustic modifications in Korean-accented English clear speech were on par with those of native clear speech. However, the degree of several clear speech modifications was smaller in Korean speakers’ production than in native production. The specific points of divergence between the two groups suggest the influence of Korean speakers’ L1 phonology on their English clear speech.</p><p dir="ltr">The second experiment investigated the perceptual benefits of Korean-accented English clear speech, for both native (n = 64) and non-native (L1 Korean) listeners (n = 64). Four groups of talker–listener combinations were recruited to examine the intelligibility benefit provided by clear speech: native talker–native listener, native talker–non-native listener, non-native talker–native listener, and non-native talker–non-native listener. Listeners were presented with semantically anomalous stimulus sentences (e.g., <i>the wrong room sold the rain</i>), which were mixed with speech-shaped noise at 0dB signal-to-noise ratio. The findings suggested that neither talkers’ L1 nor listeners’ L1 determined the degree of the intelligibility benefit. In other words, Korean-accented English clear speech was as beneficial as native clear speech, and Korean listeners were able to take advantage of clear speech to a similar extent as native English listeners.</p><p dir="ltr">The third experiment investigated the possibility of crosslinguistic influence in Korean-accented English clear speech. English and Korean clear speech was recorded, using six English minimal pairs (e.g., <i>tab </i>vs. <i>dab</i>) and six Korean minimal (or near-minimal) triplets (e.g., <i>t</i><i>h</i><i>ant</i><i>h</i><i>anhata </i>vs. <i>tantanhata </i>vs. <i>t*ant*anhata</i>), from three groups of speakers: late Korean-English bilinguals residing in USA (n = 30), Korean monolinguals (n = 30) living in Korea, and English monolinguals (n = 20). The primary goal was to determine how English and Korean laryngeal contrasts are enhanced in clear speech, and whether non-native speakers would transfer their L1 enhancement strategies to their L2 and vice versa. The results revealed that late bilinguals employed language-specific strategies in each of the two languages. In both languages, they enhanced the acoustic parameter that correlates most strongly with laryngeal contrasts in each of the two languages: VOT and onset f0 in Korean vs. VOT in English. Furthermore, they showed a greater VOT modification in Korean clear speech compared with Korean monolinguals, but a lesser VOT modification in English clear speech compared with English monolinguals, suggesting a possibility of the influence of L2 on L1 production and of L2 on L1 production, as VOT arguably plays a more prominent role in the English phonology.</p><p dir="ltr">Taken together, the overall findings from the three studies demonstrated that late Korean-English bilinguals were successful interlocutors who were able to both produce clear speech and accommodate listeners in their L2. Moreover, they implemented either English-specific or Korean-specific clear speech strategies according to the languages they produced, which indicates their flexibility in the use of appropriate cues across the two languages.</p>

Vowel hyperarticulation and hyper affect are two characteristics of infant-directed speech (IDS), reported in several studies. Vowel hyperarticulation has been suggested to be a way in which adults subconsciously facilitate infants’ language acquisition, but it has also been argued to be the acoustic by-product of an affective speaking style. The present thesis describes the expression of vocal affect in Swedish IDS at 12 months and investigates the relationship between vocal affect and vowel hypo-/hyperarticulation (VHH) on a speaker level. Naïve listeners (n = 23) rated affect valence in filtered samples of parents’ IDS and adult-directed speech (ADS). VHH was quantified by vowelspace area (VSA) for the same parents. In line with previous studies, IDS contained significantly more positive affect than ADS, illustrating that heightened positive affect is a distinct feature of Swedish IDS. Furthermore, no relationship was found between vocal affect and degree of VHH. However, a significant negative relationship was found between parents’ difference in VSA for IDS and ADS, and the difference in vocal affect for IDS and ADS. The results show that hyper affect was not associated with hyperarticulation of vowels. In general, parents that made affective adjustments in their IDS made smaller adjustments regarding VHH. The results support previous research suggesting that vowel hyperarticulation in IDS is not simply the result of an affective speaking style. / Vokalhyperartikulation och förhöjd positiv röstkänsla är två kännetecken för barnriktat tal (BRT) som har rapporterats i flera studier. Vokalhyperartikulation har föreslagits vara ett sätt vuxna omedvetet underlättar barns språkinlärning. Andra menar att det är en akustisk sidoeffekt av den positiva röstkänslan som används i interaktion med små barn. Den aktuella uppsatsen beskriver uttrycket av röstkänsla i svenskt BRT vid 12 månader och undersöker sambandet mellan röstkänsla och vokalhypo-/hyperartikulation (VHH) på talarnivå. Röstkänslan i föräldrars BRT och vuxenriktade tal (VRT) skattades av vuxna lyssnare (n = 23) och samma föräldrars vokalarea användes som mått på VHH. I linje med tidigare studier innehöll BRT signifikant mer positiv röstkänsla än VRT, vilket illustrerar att förhöjd positiv röstkänsla är ett utmärkande drag för svenskt BRT. Vidare fanns inget samband mellan röstkänsla och grad av VHH. Däremot fanns ett signifikant negativt samband mellan föräldrarnas skillnad i vokalarea för BRT och VRT, och föräldrars skillnad i röstkänsla för BRT och VRT. Resultaten visar att förhöjd positiv röstkänsla inte var associerad med hyperartikulation av vokaler. Generellt gjorde föräldrar som anpassade röstkänsla i hög grad mindre anpassningar gällande VHH. Resultaten stödjer tidigare studier som har visat att vokalhyperartikulation i BRT inte enbart är en effekt av ett positivt uttryckssätt.