Perception and acquisition of linguistic rhythm by infantsIntroductionLanguage discrimination at birthThe role of rhythm--a crosslinguistic approachThe role of rhythm--an acoustic approachSummaryAcquisition of the rhythmic properties of the native languageSummaryConclusion and future directionsAcknowledgementsReferencesPerception and acquisition of linguistic rhythm by infantsThierry Nazzia,*, Franck RamusbaLaboratoire Cognition et Deeveloppement, CNRS, Universitee Paris 5, 71, Avenue Edouard Vaillant,92774 Boulogne Billancourt Cedex, FrancebInstitute of Cognitive Neuroscience, University College London, UKAbstractIn the present paper, we address the issue of the emergence in infancy of speech segmentation procedures that werefound to be specific to rhythmic classes of languages in adulthood. These metrical procedures, which segment fluentspeech into its constitutive word sequence, are crucial for the acquisition by infants of the words of their native lan-guage. We first present a prosodic bootstrapping proposal according to which the acquisition of these metrical seg-mentation procedures would be based on an early sensitivity to rhythm (and rhythmic classes). We then review severalseries of experiments that have studied infantsÕ ability to discriminate languages between birth and 5 months, in anattempt to specify their sensitivity to rhythm and the implication of rhythm perception in the acquisition of thesesegmentation procedures. The results presented here establish infantsÕ sensitivity to rhythmic classes (from birth on-wards). They further show an evolution of infantsÕ language discriminations between birth and 5 months which, thoughnot inconsistent with our proposal, nevertheless call for more studies on the possible implication of rhythm in theacquisition of the metrical segmentation procedures.Ó 2002 Elsevier Science B.V. All rights reserved.1. IntroductionThe focus of the present paper is on the issue ofthe segmentation of fluent speech into words, andmore particularly on the development of speechsegmentation procedures during infancy. Foradults, speech segmentation involves language-specific phonological procedures (see below, andCutler, McQueen and Norris, this volume) thatallow for the retrieval of the acoustic sound pat-terns of words from fluent speech, and the con-necting of these sound patterns to the lexicalrepresentations stored in the lexicon. Hence foradults, this task could be facilitated by the lexiconitself, making it a problem of word recognition aswell as (or rather than) a problem of word seg-mentation. However, the segmentation task has tobe different for the infant who, starting with nolexicon and no language-specific phonologicalknowledge, has to discover, rather than recognize,the words in the input and learn the speech seg-mentation procedure appropriate to the languageto be learnt. In the following, we present severalstudies that have recently investigated the issue ofword segmentation in early infancy. These studieshave first established that speech segmentationemerges between 6 and 7.5 months of age in En-glish-learning American infants, hence months*Corresponding author. Tel.: +33-1-5520-5992; fax: +33-1-5520-5985.E-mail address: [email protected] (T.Nazzi).0167-6393/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.doi:10.1016/S0167-6393(02)00106-1Speech Communication 41 (2003) 233–243www.elsevier.com/locate/specombefore the onset of lexical comprehension andproduction (Jusczyk and Aslin, 1995). They havethen set out to specify the kind of information thatinfants rely on to postulate word boundaries influent speech.Some of these studies have established thatyoung American infants are sensitive to differentkinds of potential word-boundary markers that arelanguage-specific: allophonic information (the factthat the distribution of allophones within words isposition-dependent), phonotactic information (thefact that some but not all phonemic sequences arelegal at the lexical level), and prosodic/metricalinformation (the fact that lexical stress is predom-inantly word initial in English). A sensitivity toallophonic differences was found in infants asyoung as 2 months of age (Hohne and Jusczyk,1994), as attested by their ability to discriminatebetween pairs such as ‘‘nitrate’’ and ‘‘night rate’’.Infants have also been found to become sensitive tophonotactic properties of their native languagebetween 6 and 9 months of age (Friederici andWessels, 1993; Jusczyk et al., 1993b, 1994; Mattyset al., 2001). This is shown by the emergence of apreference for legal or frequent sequences of pho-nemes in their native language with respect to ille-gal or infrequent ones (e.g., the frequent andinfrequent non-word sequences ‘‘chun’’ and‘‘yush’’). Last but not least, a preference for wordswith the predominant English strong–weak stresspattern (e.g., ‘‘porter’’) over less frequent weak–strong words (e.g., ‘‘report’’) also emerges betweenthese two ages (Jusczyk et al., 1993a; Turk et al.,1995), revealing the emergence of a sensitivity tonative word stress patterns.Other studies have investigated whether, oncethey are sensitive to these language-specificmarkers, infants actually use them to infer word-like units. First, some studies showed that al-though 9-month-old infants only use allophoniccues to locate familiar words in fluent speech whenthey are guided by distributional cues, 10.5-month-old can rely on these sole allophonic cues(Jusczyk et al., 1999a). Moreover, a recent studyhas shown that the fact of providing infants withword-boundary phonotactic information helpsthem extract familiarized words from fluent speech(Mattys and Jusczyk, 2001).Finally, it was shown that 9- to 10.5-month-oldAmerican infants rely on the typical stress patternof English to group syllables into word-like units(Morgan and Saffran, 1995) and to remember fa-miliar sequences of syllables (Echols et al., 1997).The ability to retrieve familiar words from fluentspeech was also found to depend on their stresspattern. Indeed, Jusczyk et al. (1999b) found thatinfants begin segmenting strong–weak nouns (e.g.,‘‘doctor’’ and ‘‘candle’’) from fluent speech at 7.5months, but begin segmenting weak–strong nouns(e.g., ‘‘guitar’’ and ‘‘beret’’) only at 10.5 months.These authors suggested that this processing ad-vantage of strong–weak words