Learning pitch patterns in lexical identification by native English-speaking adults

The current study investigates the learning of nonnative suprasegmental patterns for word identification. Native English-speaking adults learned to use suprasegmentals (pitch patterns) to identify a vocabulary of six English pseudosyllables superimposed with three pitch patterns (18 words). Successful learning of the vocabulary necessarily entailed learning to use pitch patterns in words. Two major facets of sound-to-word learning were investigated: could native speakers of a nontone language learn the use of pitch patterns for lexical identification, and what effect did more basic auditory ability have on learning success. We found that all subjects improved to a certain degree, although large individual differences were observed. Learning success was found to be associated with the learners’ ability to perceive pitch patterns in a nonlexical context and their previous musical experience. These results suggest the importance of a phonetic–phonological–lexical continuity in adult nonnative word learning, including phonological awareness and general auditory ability. The human nervous system has a remarkable ability to learn to integrate novel complex sounds into words. This ability is evident in both infants and adults. Although numerous studies of speech and language learning in adulthood have been conducted (e.g., Jamieson & Morosan, 1989), including recent studies examining learning-related neural changes, they were either concerned with learning foreign sounds without considering their contribution to larger linguistic contexts such as words (Bradlow, Akahane-Yamada, Pisoni, & Tohkura, 1997; Golestani & Zatorre, 2004) or they focused on word learning without considering the contribution of specific phonetic features (McLaughlin, Osterhout, & Kim, 2004; Raboyeau et al., 2004). For a fuller understanding of language processing and instruction, it is essential to understand the bridge between phonemes and words, given the widely accepted notion that phonemes are the building blocks of spoken language (e.g., Chomsky & Halle, 1968). In addition, there is a mounting literature showing that infants must be able to perceive the phonetic differences between two sounds, © 2007 Cambridge University Press 0142-7164/07 $15.00 Applied Psycholinguistics 28:4 566 Wong & Perrachione: Learning pitch patterns in lexical identification and that they must learn that these phonetic differences have phonological importance before they can use two sound patterns contrastively for word identification (Stager & Werker, 1997; Swingley & Aslin, 2002; Werker, Fennell, Corcoran, & Stager, 2002). These results suggest a phonetic–phonological–lexical continuity for speech learning, such that more basic auditory abilities1 (phoneme discrimination) mediate performance on higher level auditory tasks (word learning), consistent with bottom-up models of speech perception (e.g., Norris, McQueen, & Cutler, 2000). Bottom-up processes may be especially important in learning, which likely requires additional attention to acoustic details. For example, it has been shown that masking of acoustic details by noise is more detrimental to the perception of speech by less experienced (nonnative) listeners relative to more experienced (native) listeners (Van Wijngaarden, Steeneken, & Houtgast, 2002). The present study investigates two important facets of sound-to-word learning in adults: first, can native speakers of a nontone language successfully learn the use of pitch patterns for lexical identification; second, how does the phonetic–phonological–lexical continuity manifest itself in adult second language learners—specifically, what relation does more basic auditory ability have to successful nonnative language word learning. LEARNING OF SEGMENTAL AND SUPRASEGMENTAL FEATURES A substantial amount of recent research has expanded our understanding of how linguistic exposure early in life affects the tuning of perceptual systems for speech in adulthood (see Werker & Tees, 2005, for a review). For example, Werker and Tees (1984) found that 6to 7-month-old infants raised in an English-speaking environment and native Hindi-speaking adults had no difficulty distinguishing the Hindi (non-English) retroflex/dental stop contrast, whereas native Englishspeaking adults did experience difficulty. Similarly, Polka, Colantonio, and Sundara (2002) found that English speakers improved in their perception of a /d/–/D/ contrast during language development into adulthood, whereas French speakers showed no such improvement. These results suggest the plasticity of the perceptual systems underlying language development, such that early and extensive exposure to a phonemic contrast heightens sensitivity to that contrast specifically, shaping the perception of native and foreign speech sounds later in life (e.g., Flege, 1995). Various models have been postulated to account for the different patterns of results in adult foreign speech perception, including Best’s perceptual assimilation model (e.g., Best, McRoberts, & Goodell, 2001), Flege’s speech-learning model (Flege, 1995), and Kuhl’s native language magnet model (Kuhl, Williams, Lacerda, Stevens, & Lindblom, 1992). Generally speaking, these models do not focus on how nonnative phonemes can be used in words, nor do they focus on nonnative suprasegmentals (see Hallé, Chang, & Best, 2004, for an initial proposal). Despite their difficulty in perceiving some foreign sounds, it has been found that adults showed improvement in discriminating, identifying, and/or producing speech sounds that are not in their native language after shortterm (weeks) behavioral training, including three-way voice onset contrasts by English speakers (e.g., Pisoni, Aslin, Perey, & Hennessy, 1982), /θ /–/ε/ contrasts by French speakers (e.g., Jamieson & Morosan, 1989), and /r/–/l/ by Japanese Applied Psycholinguistics 28:4 567 Wong & Perrachione: Learning pitch patterns in lexical identification speakers (e.g., Bradlow et al., 1997; Lively, Logan, & Pisoni, 1993; McCandliss, Fiez, Protopapus, Conway, & McClelland, 2002). Although almost all studies concerning phonetic learning focused on the learning of consonants and vowels (segments), the learning of suprasegmentals was also investigated. Most languages of the world use lexical tones (lexically meaningful pitch patterns) and are called tone languages (Fromkin, 2000). Mandarin Chinese is a tone language that has four lexical tones: high level (Tone 1), rising (Tone 2), dipping (Tone 3), and falling (Tone 4). For example, four different words can result when the syllable /ma/ is spoken in each of the four lexical tones. Respectively, it can mean “mother,” “hemp,” “horse,” or “scold.” Similar to the segmental learning studies discussed, Wang, Spence, Jongman, and Sereno (1999) further showed that adults were able to improve in their ability to identify Mandarin lexical tones after a short period of training. In the Wang study, native English-speaking adults were trained to identify the four Mandarin lexical tones in nonlexical contexts, that is, subjects were asked to attend to and identify the trajectory of the pitch patterns without using them to contrast word meaning. With eight sessions of training, identification accuracy increased by an average of 21%. Taken together, these studies of speech learning show that despite years of reduced experiences with nonnative phonetic features, both segmental and suprasegmental, adults’ ability to identify nonnative features can generally be improved. It is worth noting that time (e.g., 10 sessions) was often the criterion for training termination, not proficiency level. Although significant improvement was often observed, it is not clear whether all subjects learned to the fullest of their potential, and factors influencing attainment and individual differences in learning were not investigated. SEGMENT-TO-WORD LEARNING In the aforementioned speech-learning studies, including both segment and suprasegmental learning, the sole focus of training the adults was for them to improve in their ability to perceive the individual speech sounds or phonetic contrasts, but not to use them in actual words or larger communicative contexts. A notable exception of these adult speech training studies (including the Wang et al. study) is one conducted by Curtin, Goad, and Pater (1998). In this study, subjects learned to use three-way Thai voicing contrasts for word identification. After a fixed training period (2 days), English and French learners improved in their ability to match the auditorily presented Thai words with the correct pictures, indicating the correct use of Thai voicing for lexical purposes. However, as far as we are aware, no published studies have been conducted that examined the learnability of suprasegmentals (lexical tones in particular) for lexical identification, let alone the factors that might contribute to learning. Although Wang et al. (1999) successfully trained native English-speaking subjects with no exposure to any tone language to identify pitch patterns embedded in various types of syllables produced by several different talkers, it remains unclear whether native English-speaking subjects can learn to use these pitch patterns for the purposes of identifying words. Anecdotal evidence suggests that even highly proficient learners of Mandarin who are native English speakers have difficulty Applied Psycholinguistics 28:4 568 Wong & Perrachione: Learning pitch patterns in lexical identification with lexical tones, despite the fact that communication is largely successful. This is likely because of the contextual and semantic redundancy occurring in most naturalistic communicative situations. High-level attainment of Mandarin proficiency, then, may not necessarily imply a high level of accuracy in producing and/or perceiving lexical tones. This anecdotal evidence is supported by a study of Mandarin tone perception by native English-speaking adults who had several years of Mandarin instruction but whose tone perception performance was still substantially lower than that of native Mandarin speakers (Gottfried & Suiter, 1997). The primary goal of the current study was to examine whether native English-speaking adults with no previous exposure to a tone language are able to learn to use pitch patterns (suprasegmentals) in a linguistic context. In an effort to examine the subjects’ individual learning ability, we used a performance-based, rather than a time-based, training termination criterion. FACTORS ASSOCIATED WITH LANGUAGE LEARNING In the language learning and development literature, it has been shown that phonological and morphological awareness are important indicators of learning success in Indo-European and non-Indo-European languages including Chinese (e.g., Berninger, 2001; McBride-Chang, Shu, Zhou, Wat, & Wagner, 2003). Such awareness is often defined as the learner’s ability to identify and/or manipulate components of linguistic units. For example, in the commonly used Phonological Awareness Test (Robertson & Salter, 1997), subjects are asked to identify rhyming words presented in pairs in the Discrimination task of the Rhyming Subtest, and they are asked to identify one phoneme by positions in a word in the Isolation Subtest. Because these tests are generally constructed to assess Indo-European languages, no tone identification subtest is included. As far as we are aware, no studies have been conducted to examine the relationship between tone awareness and learning to use lexical tones in words in either firstor second-language learning. Given the robustness of the reported contribution of phonological and morphological awareness in language learning, including studies using formal phonological awareness test battery for assessing reading acquisition (e.g., Berninger, 2001), as well as studies suggesting that phoneme (consonant) identification (in nonstandardized test) is a prerequisite for spoken word learning (e.g., Werker & Curtin 2005; Werker et al., 2002), we expect that tone awareness will likely contribute to tone language learning. Therefore, for all our learners, a tone (pitch pattern) identification test, which addresses this aspect of phonological awareness, was administered before the onset of training, and their ability to identify these tone patterns exclusive of a linguistic context was compared with their ability to learn to use them in words. In a recent study investigating the possible relationship between musical experience and lexical tone perception, Alexander, Wong, and Bradlow (2005) found that native English-speaking amateur musicians who had no previous exposure to any tone language showed increased ability identifying and discriminating Mandarin tones relative to their nonmusician counterparts. These amateur musicians had at least 6 years of formal private lessons in one instrument starting before 10 years old, whereas their nonmusicians counterparts had less than 3 years of Applied Psycholinguistics 28:4 569 Wong & Perrachione: Learning pitch patterns in lexical identification musical training. However, whether these amateur musicians also possess increased ability in learning to use pitch patterns in lexical context remains unknown; this was addressed in the current study. If they indeed possess increased ability in lexical learning, that would suggest a more basic and general auditory mechanism contributes to spoken language learning, which is consistent with the general framework of bottom-up models (e.g., Norris et al., 2000) and models suggesting detailed acoustic and vocal analyses to be prerequisites of speech perception (e.g., Belin, Fecteau, & Bédard, 2004). In the present study, native English-speaking adults who had no exposure to a tone language were trained to identify English pseudowords superimposed with three pitch patterns resembling three Mandarin lexical tones. Successful learning of the vocabulary necessarily entailed learning to use pitch patterns in words. Before training, all subjects reported their prior musical experience and participated in a separate experiment in which they were asked to identify pitch patterns similar to those superimposed in the training stimuli.