A Destressing “Deafness” in French?

Abstract

Spanish but not French uses accent to contrast between words (e.g., tópo vs topó). Two populations of subjects were tested on the same materials to determine whether this difference has an impact on the perceptual capacities of listeners. In Experiment 1, using an ABX paradigm, we found that French subjects had significantly more difficulties than Spanish subjects to perform an ABX classification task based on accent. In Experiment 2, we found that Spanish subjects were unable to ignore irrelevant differences in accent in a phoneme-based ABX task, whereas French subjects had no difficulty at all. In Experiment 3, we replicated the basic French finding, and found that Spanish subjects benefited from redundant accent information even when phonemic information alone was sufficient to perform the task. In our final experiment, we showed that French subjects can be made to respond to the acoustic correlates of accent; therefore their difficulty in Experiment 1 seem to be located at the level of short-term memory. The implications of these findings for language-specific processing and acquisition are discussed.The native speaker's difficulties with some nonnative segmental contrasts is well documented, (e.g., Polka & Werker, 1994). Much less is known about the way in which rhythmical and structural properties of the native language affect encoding of foreign words. This study reports four experiments to assess the claim that nonsegmental properties of the native language greatly affect the way in which unfamiliar items are processed and represented.Learning a language requires, inter alia, acquiring knowledge of the language's sound pattern. Psycholinguists have mostly assumed that learning the sound pattern of a native language essentially involves learning the segments of the language. This assumption is evident in books and articles on the subject, e.g., Carroll (1960) for an early formulation of this view and Miller and Jusczyk (1990) for a later one. There is no question that segments are an essential ingredient of the sound pattern of language though many other properties of speech are also important. Foreign accents reflect segmental difficulties but also supra-segmental ones. For instance, informal observations suggest that French speakers who acThis research was supported by grants from the Human Frontier Science Program, the Human Capital program, the FrancoSpanish Exchange Program for 1993-1995, the Fyssen Foundation and DRET. We thank Jennifer Glos, Xavier Hardy and Olivier Crouzet for help in stimulus construction and data gathering. We thank Stanislas Dehaene and Brit van Ooijen for very useful comments on earlier versions of this paper. Mailing address: Emmanuel Dupoux, 54 Bd Raspail, F-75270 Paris Cedex 06, France. http://www.ehess.fr/centres/lscp/persons/dupoux quire English after puberty produce segments that are not prototypical of English, and prosodic output is usually deficient as well. It is not uncommon to hear speakers of French who fail to reduce vowels or who make the wrong vowel prominent. ...

A limit on behavioral plasticity in speech perception

“…More generally, research has shown that listeners are attuned to the properties of the speech signal which are most relevant for their language (e.g., Otake et al, 1993Otake et al, , 1996Dupoux et al, 1997). Data exist that show that this tuning actually takes place very early, in the first year of life (see Kuhl, 1994;Werker, 1994 for reviews).…”

Perception of suprasegmental structure in a non-native dialect

“…Non-native suprasegmental contrasts, like non-native segmental contrasts, can present listeners with perceptual di$culty. Thus, stress contrasts in nonsense words are di$cult for French speakers to perceive, but easy for Spanish speakers (Dupoux, Pallier, SebastiaH n-GalleH s & Mehler, 1997); Spanish is a stress language, French is not. Perception of tone in Chinese is in#u-enced for English speakers by their knowledge of pitch patterns in English (Broselow, Hurtig & Ringen, 1987).…”

“…Previous studies of the perception of suprasegmental distinctions have revealed processing di!erences between native and non-native listeners. Thus as Dupoux et al (1997) showed, Spanish and French listeners di!er in how accurately they can perceive stress contrasts in nonsense words spoken by a Dutch speaker; similarly, Dutch and Finnish listeners can use suprasegmental correlates of stress as a cue to word onset, while French listeners do not (Vroomen, Tuomainen & de Gelder, 1998). While the discrimination of tone contrasts is a!ected by word}nonword status only for native listeners and not for non-native listeners (Fox & Unkefer, 1985;Cutler & Chen, 1997), overall similarity in discriminability of tone contrasts for native and non-native listeners (Burnham, Kirkwood, Luksaneeyanawin & Pansottee, 1992;Cutler & Chen, 1997) can mask processing di!erences * for instance, greater interdependence of segmental and suprasegmental judgments for tone language speakers than for speakers of non-tone languages (Repp & Lin, 1990;Lee & Nusbaum, 1993), or dependence on native distinctions in making the non-native distinctions (Broselow et al, 1987;Nishinuma et al, 1996).…”

Stress assignment in aphasia: Word and non-word reading and non-word repetition

“…Such an interpretation is endorsed by findings that short-term memory is involved in stress assignment in non-brain damaged participants. Dupoux, Pallier, Sebastian, and Mehler (1997) reported a memory overload on performance in a study on stress 'deafness' in French non-brain-damaged participants. Spanish, but not French uses lexical stress to distinguish between minimal stress pairs of words (e.g.…”