Applied Psycholinguistics 24 (2003), 45–57Printed in the United States of AmericaDOI: 10.1017.S0142716403000031Auditory recognition of idioms bynative and nonnative speakers ofEnglish: It takes one to know oneDIANA VANLANCKER–SIDTISNew York UniversityADDRESS FOR CORRESPONDENCEDiana Vanlancker–Sidtis, Department of Speech–Language Pathology and Audiology, New YorkUniversity, 719 Broadway, Suite 200, New York, NY 10003. E-mail: [email protected] abilities of second language speakers to discriminate the prosodic contrasts between idiomaticand literal meanings of ambiguous sentences were investigated using utterances previously shownto be reliably identified by acoustic cues. Four listener groups of varying proficiency, native speak-ers of American English, native speakers of non-American English, fluent nonnative speakers ofEnglish, and advanced students of English as a second language (ESL), judged whether single andpaired, tape-recorded, literal and idiomatic utterances were spoken with intended idiomatic or literalmeanings. Both native speaker groups performed significantly better than fluent nonnatives, whileESL students performed at chance. These results lend support to the hypothesis that abilities todiscriminate subtle prosodic contrasts are learned later than other components of speech and lan-guage.Idioms, proverbs, speech formulas, and other familiar conventional expressions,although neglected in earlier modern linguistic theory (exceptions are Chafe,1968; Fraser, 1970; Heringer, 1976; Katz, 1973; Makkai, 1972; Weinreich,1969), form a large portion of everyday communication (Coulmas, 1981, 1994;Cowie, 1998; Peters, 1983; Van Lancker, 1973, 1975, 1987; Wray & Perkins,2000). Familiar conventional expressions include a broad range of more or less“memorized” constructions (Bolinger, 1976; Lyons, 1968) such as slang, quota-tions, sayings, irreversible binomials (“salt and pepper”; Cooper & Ross, 1975;Malkiel, 1959), expletives, song lyrics, indirect speech acts, and professionaljargon (Van Lancker, 1999). Fillmore (1979) stated that nativelike fluency in alanguage is dependent largely on knowledge and proper use of formulaic expres-sions; and Pawley and Syder (1980) argued that “sounding like a native speaker”involves appropriate use of these conventionalized phrasal units. Schegloff(1998) noted that competent and routine use of formulaic language allows fordisplays of paralinguistic information. For second language (L2) learners, learn- 2003 Cambridge University Press 0142-7164/03 $12.00Applied Psycholinguistics 24:1 46Vanlancker–Sidtis: Auditory recognition of idiomsing formulaic language is a complex, even arduous, process requiring specialtreatment (Bygate, 1988; Howarth, 1998; Weinert, 1995; Wray, 1999a, 1999b).Mastery of an L2 presents a unique difficulty in part because idiomatic ex-pressions are made up of stereotyped forms associated with conventionalizedmeanings, allowing only narrow ranges of variability in usage. Words in theidiom are often not used with their usual meanings (e.g., “She has him eatingout of her hand”). The expression occurs on a narrow range of possible intona-tional contours (Bolinger, 1986, p. 280, 1989). Novel sentences appear in arange of grammatical structures and lexical choices; in contrast, appropriate useof idioms requires relative exactness in expression and usage. Lindfield, Wing-field, and Goodglass (1999) suggested that prosodic shapes are stored with pho-nological material in the mental lexicon. A production error heard in the sponta-neous speech of speakers of English as a second language (ESL), “I wouldn’twant to be in his shoes like that,” illustrates the integral role of prosodic fea-tures. For the native English speaker, not only would the expression end withshoes, but it would have the sentence accent on his. From a study revealing theinability of Japanese speakers to discriminate the /r/ versus /l/ distinction inEnglish at the cortical level (Buchwald, Van Lancker, Erwin, Guthrie, & Schwa-fel, 1994), it can be assumed that errors in production are correlated with insuffi-cient perceptual ability. It is this inference that leads to this study of perceptionof acoustic cues distinguishing literal from idiomatic expressions.The inability to produce accentless speech in a L2 learned after childhoodhas been widely noted (e.g., Lenneberg, 1967; Major, 1987; Neufeld, 1980;Ryan & Sebastian, 1980; Scovel, 1969, 1988). Much of the observed accentresides in prosodic aspects. Even after phonology, syntax, and semantics areapparently mastered, L2 speakers often retain prominent prosodic features—those of pitch, loudness, timing, and voice quality—intrinsic to their nativelanguage (Van Els & de Bot, 1987), and details of prosody inherent in theL2 are particularly difficult to teach the L2 speaker (e.g., Gilbert, 1984, 1993;Young–Scholten, 1993). The ability to hear prosodic differences has been lessstudied. Prosody may be the least “perfectly” learned aspect of adult L2 acquisi-tion. The present study examined the abilities of highly practiced and experi-enced L2 speakers on a task requiring judgments of subtle prosodic contrasts.Studies have suggested that idiomatic (or formulaic) language and literal lan-guage are learned and used according to fundamentally different psycholinguis-tic processes (Peters, 1977; Swinney & Cutler, 1979; Vihman, 1982; Wray &Perkins, 2000). The claim that they are represented in different cerebral hemi-spheres is elaborated elsewhere (Van Lancker, 1975, 1990; Van Lancker &Kempler, 1987). Lounsbury (1963) noted that use of two constructions, one adhoc and the other “familiar and employed as a whole unit,” constitutes differentbehavioral events and that “their psychological statuses in the structure of actualspeaking behavior may be quite different” (p. 561). During speech perceptionand registration, units of different sizes (phonological units, syllables, mor-phemes, words, and phrases) are computed, or “chunked” (Cole & Scott, 1974;Marslen–Wilson & Welsh, 1978; Simon, 1974). Idioms are recognized and re-membered as integral units, in contrast to novel phrases, and prosodic cues areutilized generally in the structural analysis in speech perception (Brown, 1997;Applied Psycholinguistics 24:1 47Vanlancker–Sidtis: Auditory recognition of idiomsButtet, 1988; Buttet, Wingfield, & Sandoval, 1980; Cutler, Dahan, & Donselaar,1997; Geers, 1978; Price,