Abstraction and the (Misnamed) Language Familiarity Effect

Abstract

Talkers are recognized more accurately if they are speaking the listeners' native language rather than an unfamiliar language. This "language familiarity effect" has been shown not to depend upon comprehension and must instead involve language sound patterns. We further examine the level of sound-pattern processing involved, by comparing talker recognition in foreign languages versus two varieties of English, by (a) English speakers of one variety, (b) English speakers of the other variety, and (c) non-native listeners (more familiar with one of the varieties). All listener groups performed better with native than foreign speech, but no effect of language variety appeared: Native listeners discriminated talkers equally well in each, with the native variety never outdoing the other variety, and non-native listeners discriminated talkers equally poorly in each, irrespective of the variety's familiarity. The results suggest that this talker recognition effect rests not on simple familiarity, but on an abstract level of phonological processing.

By 4.5 Months, Linguistic Experience Already Affects Infants’ Talker Processing Abilities

“…For example, language familiarity facilitates vocal emotion detection (Pell & Skorup, ; Scherer, Banse, & Wallbott, ), consonant and vowel perception is influenced by talker identity (Eisner & McQueen, ; Johnson, Strand, & Johnson, ), talker familiarity enhances word recognition in adverse listening conditions (Creel, Aslin, & Tanenhaus, ; Goldinger, ; Mullennix & Pisoni, ; Nygaard, Sommers, & Pisoni, ), and certain speech sounds carry more information about talker identity than others (Andics, McQueen, & van Turennout, ; Cutler, Andics, & Fang, ). Most importantly for this study, adults recognize talkers better in a familiar language than in an unfamiliar language, a phenomenon known as the language familiarity effect (Goggin, Thompson, Strube, & Simental, ; Johnson, Bruggeman, & Cutler, ; Levi, ; Orena, Theodore, & Polka, ; Perrachione, ).…”

Methodological considerations for interpreting the Language Familiarity Effect in talker processing

“…What unifies the voice line‐up studies is that a single voice functions as the target voice following brief exposure. The most common number of voices presented in the array is six (Goggin, Thompson, Strube, & Simental, ; Philippon, Cherryman, Bull, & Vrij, ; Schlichting & Sullivan, ; Sullivan & Kügler, ; Sullivan & Schlichting, ; Thompson, ), but other studies have used four (Goldstein, Knight, Bailis, & Conover, ; Johnson, Bruggeman, & Cutler, ) or only two (Fecher & Johnson, ). A few studies have used a single target voice, but instead of the typical line‐up during the test phase have instead used a two alternative forced‐choice task (2AFC) where on each trial listeners must indicate whether the voice is the same as the target (Köster & Schiller, ; Schiller & Köster, ).…”

“…One manipulation involves the delay between exposure and retrieval, ranging from nearly immediate (Goggin et al, ; Goldstein et al, ; Johnson et al, ; Schlichting & Sullivan, ) to a 1‐week delay (Thompson, ). A second variation across the tasks is the duration of stimuli used for exposure and testing, including a single word (Goldstein et al, ), a sentence (Fecher & Johnson, ; Goldstein et al, ; Johnson et al, ), a paragraph (Goggin et al, ; Thompson, ), 1 minute of speech (Philippon et al, ; Schlichting & Sullivan, ; Sullivan & Kügler, ; Sullivan & Schlichting, ), or 5 minutes of speech (Köster & Schiller, ; Schiller & Köster, ). Finally, in most voice line‐up studies, the target voice is present, but in some cases it is not, allowing researchers to explore the average rate of false alarms, a consideration that is critical for the legal ramifications of voice‐identification (Philippon et al, ; Thompson, ).…”

Bilingual infants excel at foreign‐language talker recognition

“…We tested infants on an unfamiliar language because, first, this leveled the playing field between monolinguals and bilinguals, ensuring that all infants were equally unfamiliar with the test language. And second, using an unfamiliar language decreased the likelihood that we would find a ceiling effect in infants’ performance (and thus increased the likelihood that we would observe a performance difference between groups), because we know from previous work that listeners generally struggle with identifying talkers in an unfamiliar language (e.g., Fecher & Johnson, ,; Johnson, Bruggeman, & Cutler, ; Levi, ; Perrachione, Del Tufo, & Gabrieli, ). If exposure to multiple languages in infancy enhances the processing of indexical speech cues, then bilinguals should outperform monolinguals in learning to identify foreign‐language talkers.…”