Publications

This study focuses on the development of lexical stress perception during the first year of life. Previous research shows that cross-linguistic differences in word stress organisation translate into differences in word stress processing from a very early age: At 9 months, Spanish-learning infants, learning a language with variable word stress, can distinguish segmentally varied stress-initial (e.g. níla, túli) and stress-final (e.g. lutá, pukí) nonsense words in a headturn preference procedure. However, French infants, who learn a language with fixed word stress, can only distinguish between initial and final stress when no segmental variability is involved (Skoruppa et al., 2009). The present study investigates the emergence of this cross-linguistic difference. We show that neither Spanish nor French 6-month-olds encode stress patterns in the presence of segmental variability (Exp. 1), while both groups succeed in the absence of segmental variability (Exp. 2). Hence, only Spanish infants, who learn a variable stress language, get better at tracking stress patterns in segmentally varied words between the ages of 6 and 9 months. In contrast, all infants seem to be able to discriminate basic stress patterns in the absence of segmental variability during the first nine months of life, regardless of the status of stress in their native language.

Before the end of the first year of life, infants begin to lose the ability to perceive distinctions between sounds that are not phonemic in their native language. It is typically assumed that this developmental change reflects the construction of language-specific phoneme categories, but how these categories are learned largely remains a mystery. Peperkamp, Le Calvez, Nadal, & Dupoux (2006) present an algorithm that can discover phonemes using the distributions of allophones as well as the phonetic properties of the allophones and their contexts. We show that a third type of information source, the occurrence of pairs of minimally-differing word forms in speech heard by the infant, is also useful for learning phonemic categories, and is in fact more reliable than purely distributional information in data containing a large number of allophones. In our model, learners build an approximation of the lexicon consisting of the high-frequency n-grams present in their speech input, allowing them to take advantage of top-down lexical information without needing to learn words. This may explain how infants have already begun to exhibit sensitivity to phonemic categories before they have a large receptive lexicon.

Using a picture pointing task, this study examines toddlers' processing of phonological alternations that trigger sound changes in connected speech. Three experiments investigate whether 2;5- to 3-year-old children take into account assimilations, processes by which phonological features of one sound spread to adjacent sounds for the purpose of word recognition (e.g., in English, ten pounds can be produced as tem pounds). English toddlers show sensitivity to native place assimilations during lexical access in Experiment 1. Likewise, French toddlers compensate for French voicing assimilations in Experiment 2. However, French toddlers do not take into account a hypothetical non-native place assimilation rule in Experiment 3, suggesting that compensation for assimilation is already language-specific.

Adult speakers of different free stress languages (e.g., English, Spanish) differ both in their sensitivity to lexical stress, and in their processing of suprasegmental and vowel quality cues to stress. In a headturn preference experiment with a familiarization phase, both 8-month-old and 12-month-old English-learning infants discriminated between initial stress and final stress among lists of Spanish-spoken disyllabic non-words that were segmentally varied (e.g. ['nila, 'tuli] vs. [lu'ta, pu'ki]). This is evidence that English-learning infants are sensitive to lexical stress patterns, instantiated primarily by suprasegmental cues, during the second half of the first year of life.

Listeners of various languages tend to perceive an illusory vowel inside consonant clusters that are illegal in their native language. Here, we test whether this phenomenon arises after phoneme categorization or rather interacts with it. We assess the perception of illegal consonant clusters in native speakers of Japanese, Brazilian Portuguese, and European Portuguese, three languages that have similar phonological properties, but that differ with respect to both segmental categories and segmental transition probabilities. We manipulate the coarticulatory information present in the consonant clusters, and use a forced choice vowel labeling task (Experiment 1) and an ABX discrimination task (Experiment 2). We find that only Japanese and Brazilian Portuguese listeners show a perceptual epenthesis effect, and, furthermore, that within these participant groups the nature of the perceived epenthetic vowel varies according to the coarticulation cues. These results are consistent with models that integrate phonotactic probabilities within perceptual categorization, and are problematic for two-step models in which the repair of illegal sequences follows that of categorization.

This paper examines whether adults can adapt to novel accents of their native language that contain unfamiliar context-dependent phonological alternations. In two experiments, French participants listen to short stories read in accented speech. Their knowledge of the accents is then tested in a forced-choice identification task. In Experiment 1, two groups of listeners are exposed to newly created French accents in which certain vowels harmonize or disharmonize, respectively, to the rounding of the preceding vowel. Despite the cross-linguistic predominance of vowel harmony over disharmony, the two groups adapt equally well to both accents, suggesting that this typological difference is not reflected in perceptual learning. Experiment 2 further explores the mechanism underlying this type of phonological learning. Participants are exposed to an accent in which some vowels harmonize and others disharmonize, yielding an increased featural complexity. They adapt less well to this regularity, showing that adaptation to novel accents involves feature-based inferences.

Previous studies have documented that speakers of French, a language with predictable stress, have difficulty distinguishing nonsense words that vary in stress position solely (stress "deafness"). In a sequence recall task with adult speakers of five languages with predictable stress (Standard French, Southeastern French, Finnish, Hungarian and Polish) and one language with non-predictable stress (Spanish), it was found that speakers of all languages with predictable stress except Polish exhibited a strong stress "deafness", while Spanish speakers exhibited no such "deafness". Polish speakers yielded an intermediate pattern of results: they exhibited a weak stress "deafness". These findings are discussed in light of current theoretical models of speech perception.

We probed simultaneous French-Spanish bilinguals for the perception of Spanish lexical stress using three tasks, two short-term memory encoding tasks and a speeded lexical decision. In all three tasks, the performance of the group of simultaneous bilinguals was intermediate between that of native speakers of Spanish on the one hand and French late learners of Spanish on the other hand. Using a composite stress 'deafness' index measure computed over the results of the three tasks, we found that the performance of the simultaneous bilinguals is best fitted by a bimodal distribution that corresponds to a mixture of the performance distributions of the two control groups. Correlation analyses showed that the variables explaining language dominance are linked to early language exposure. These findings are discussed in light of theories of language processing in bilinguals.

During the first year of life, infants begin to have difficulties perceiving non-native vowel and consonant contrasts, thus adapting their perception to the phonetic categories of the target language. In this paper, we examine the perception of a non-segmental feature, i.e. stress. Previous research with adults has shown that speakers of French (a language with fixed stress) have great difficulties in perceiving stress contrasts (Dupoux, Pallier, Sebastián & Mehler, 1997), whereas speakers of Spanish (a language with lexically contrastive stress) perceive these contrasts as accurately as segmental contrasts. We show that language-specific differences in the perception of stress likewise arise during the first year of life. Specifically, 9-month-old Spanish infants successfully distinguish between stress-initial and stress-final pseudo-words, while French infants of this age show no sign of discrimination. In a second experiment using multiple tokens of a single pseudo-word, French infants of the same age successfully discriminate between the two stress patterns, showing that they are able to perceive the acoustic correlates of stress. Their failure to discriminate stress patterns in the first experiment thus reflects an inability to process stress at an abstract, phonological level.

Japanese shows an asymmetry in the treatment of word-final [n] in loanwords from English and French: while it is adapted as a moraic nasal consonant in loanwords from English, it is adapted with a following epenthetic vowel in loanwords from French. We provide experimental evidence that this asymmetry originates in the way Japanese speakers perceive word-final [n] in English and French. In Experiment 1 we use a forced choice identification task and show that Japanese speakers identify French [n]-final stimuli as ending in a vowel significantly more often than English ones; the identification of this vowel is shown to correlate with the spectral energy contained in the release of the nasal consonant in the stimuli. In Experiment 2 we use an ABX discrimination task and show that Japanese speakers have difficulties distinguishing French forms ending in [n] from corresponding ones ending in [nu]. Given the primacy of perception over production, we conclude that the Japanese loanword adaptations originate in perception and consist of phonetically minimal transformations. We discuss consequences of our findings for models of loanword adaptation.

We explore whether infants can learn novel phonological alternations on the basis of distributional information. In Experiment 1, two groups of 12-month-old infants were familiarized with artificial languages whose distributional properties exhibited either stop or fricative voicing alternations. At test, infants in the two exposure groups had different preferences for novel sequences involving voiced and voiceless stops and fricatives, suggesting that each group had internalized a different familiarization alternation. In Experiment 2, 8.5-month-olds exhibited the same patterns of preference. In Experiments 3 and 4, we investigated whether infants' preferences were driven solely by preferences for sequences of high transitional probability. Although 8.5-month-olds in Experiment 3 were sensitive to the relative probabilities of sequences in the familiarization stimuli, only 12-month-olds in Experiment 4 showed evidence of having grouped alternating segments into a single functional category. Taken together, these results suggest a developmental trajectory for the acquisition of phonological alternations using distributional cues in the input.

Previous research by Dupoux et al. found that French speakers, as opposed to Spanish ones, are impaired in discrimination tasks with stimuli that vary only in the position of stress. However, what was called stress 'deafness' was only found in tasks that used high phonetic variability and memory load, not in cognitively less demanding tasks such as single token AX discrimination. This raised the possibility that instead of a perceptual problem, monolingual French speakers might simply lack a metalinguistic representation of contrastive stress, which would impair them in memory tasks. We examined a sample of 39 native speakers of French who underwent formal teaching of Spanish after age 10, and varied in degree of practice in this language. Using a sequence recall task, we observed in all our groups of late learners of Spanish the same impairment in short-term memory encoding of stress contrasts that was previously found in French monolinguals. Furthermore, using a speeded lexical decision task with word-nonword minimal pairs that differ only in the position of stress, we found that all late learners had much difficulty in the use of stress to access the lexicon. Our results show that stress 'deafness' is better interpreted as a lasting processing problem resulting from the impossibility for French speakers to encode contrastive stress in their phonological representations. This affects their memory encoding as well as their lexical access in on-line tasks. The generality of such a persistent suprasegmental 'deafness' is discussed in relation to current findings and models on the perception of non-native phonological contrasts.

In their article 'What we know about what we have never heard: Evidence from perceptual illusions', Berent, Steriade, Lennertz, and Vaknin (2007) argue that English listeners' perception of onset clusters is reflected by the markedness of these clusters. Specifically, they show that the more an onset cluster is marked, the more likely it is perceived with an illusory epenthetic vowel by English listeners. English words do not contain any of the onset cluster types used in the experiments; hence markedness effects cannot be the result of phonological learning. I argue that their conclusion is not justified for two reasons. First, all experiments test whether different types of clusters are perceived with an epenthetic schwa or not. Epenthesis, however, is not the only possible perceptual repair. Hence, we cannot rule out that those cluster types that tend not to undergo perceptual epenthesis are more often subject to some other perceptual repair, thus undermining the inference that they are perceived more faithfully. Second, although the authors do consider a phonetic explanation of the results, their arguments against such an explanation are inconclusive.

Phonological rules relate surface phonetic word forms to abstract underlying forms that are stored in the lexicon. Infants must thus acquire these rules in order to infer the abstract representation of words. We implement a statistical learning algorithm for the acquisition of one type of rule, namely allophony, which introduces context-sensitive phonetic variants of phonemes. This algorithm is based on the observation that different realizations of a single phoneme typically do not appear in the same contexts (ideally, they have complementary distributions). In particular, it measures the discrepancies in context probabilities for each pair of phonetic segments. In Experiment 1, we test the algorithm's performances on a pseudo-language and show that it is robust to statistical noise due to sampling and coding errors, and to non-systematic rule application. In Experiment 2, we show that a natural corpus of semiphonetically transcribed child-directed speech in French presents a very large number of near-complementary distributions that do not correspond to existing allophonic rules. These spurious allophonic rules can be eliminated by a linguistically motivated filtering mechanism based on a phonetic representation of segments. We discuss the role of a priori linguistic knowledge in the statistical learning of phonology.

We investigate the influence of orthography on loanword adaptations by means of an experiment in which late French-English bilinguals produce on-line adaptations of English non-words. In half of the experiment, the stimuli are presented orally only, whereas in the other half, the oral stimuli are accompanied by their written representation. The adaptations of eight English vowels are shown to be different according to whether the input is oral or mixed (i.e. oral + written). In particular, the adaptations based on the mixed input more often reflect the way French speakers are used to read English graphemes. These results thus confirm the sensitivity of loanword adaptations to the presence versus absence of a written representation. We conclude that in order to control for orthography, loanword adaptations are best studied in an experimental framework.

We tested the effect of local lexical ambiguities while manipulating the type of prosodic boundary at which the ambiguity occurred, using French sentences and participants. We observed delayed lexical access when a local lexical ambiguity occurred within a phonological phrase (consistent with previous research; e.g. '[un chat grincheux]', containing the potential competitor word 'chagrin', was processed more slowly than '[un chat drogué]' that contains no potential competitor). In contrast, when the lexical competitor straddled a phonological phrase boundary, there was no delay in lexical recognition (e.g., '[son grand chat] [grimpait. . .]', potential competitor 'chagrin' was not delayed relative to the non-ambiguous control). These results were observed with two different on-line tasks, word-monitoring and phoneme monitoring. They suggest that lexical access occurs within the domain of phonological phrases. We discuss the implications of these results for models of lexical access.

Some, but not all, languages with fixed stress exhibit lexical exceptions to their stress rule. Under the assumption that lexical exceptions can occur in a language only if its native speakers can perceive stress contrasts, I argue that the presence of these exceptions depends on the age at which infants discover the stress rule of their language. If the stress regularity is easy to infer from the surface speech stream, then it will be acquired very early, and stress will not be encoded in the phonological representation of words in the mental lexicon; as a consequence, stress contrasts are not well perceived by adult speakers and lexical exceptions are excluded. If, by contrast, the regularity is difficult to infer, then it will be acquired relatively late, after the format of the phonological encoding of words has been fixed. That is, stress will be redundantly encoded in the mental lexicon, and lexical exceptions can thus be perceived and stored by adult speakers. A typological survey concerning the occurrence of exceptions in languages with fixed stress supports this proposal. A comparison with a metrical approach to exceptional stress is made, leading to a proposal about the division of labor between psycholinguistics and theoretical phonology.

Infants' phonological acquisition during the first 18 months of life has been studied within experimental psychology for some 30 years. Current research themes include statistical learning mechanisms, early lexical development, and models of phonetic category perception. So far, linguistic theories have hardly been taken into account. These theories are based upon the assumption that there is a common core of innate phonological knowledge across speakers of all human languages, and they contain detailed proposals concerning phonological representations and the derivations by which abstract underlying forms are mapped onto concrete surface forms. It remains to be investigated experimentally if there is innate phonological knowledge and how the language-specific phonological grammar is acquired. In the present article, the contributions to this special issue are introduced, and an attempt is made to bridge the gap between phonological theory and experimental psychology. In particular, some recent experimental work is considered in the light of phonological theories and new research avenues are sketched. What might be innate, what needs to be acquired, and how this acquisition might take place are questions that are addressed with respect to several aspects of phonological knowledge, specifically segmental representations, phonotactics, phonological processes, and the architecture of the phonological grammar.

This paper surveys various analyses of secondary stress assignment in Italian and argues that they suffer from both descriptive and explanatory shortcomings. An alternative account will be developed, in which stress is not assigned by rule but is selected from a set of conceivable candidates by a system of constraints on representations. The stress patterns that are considered are those of words when pronounced in isolation. It will be shown that monomorphemic and derived words differ only in their stress patterns in that derived words are subject to the requirement that stress from the embedded word be preserved. This requirement, though, is overridden by constraints on metrical and prosodic wellformedness. Optional reranking of these requirements in the relevance hierarchy accounts for the variability of secondary stress which is found mainly in the southern varieties of Standard Italian. Finally, it will be argued that compounds generally consist of more than one stress domain, whereas adverbs ending in -mente show hybrid behavior; they cannot be classified either with derived words or with compounds.

Italian has both a series of derivational prefixes and a series of derivational suffixes. Word-final vowels typically do not surface before vowel-initial suffixes. By contrast, if a vowel-final prefix attaches to a vowel-initial word, generally both vowels surface. This article analyzes the asymmetry between Italian prefixation and suffixation within Optimality Theory. In particular, it is argued that suffixation is stem-based, and that a single ranking of prosodic constraints governs both prefixation and suffixation.

Recent work suggests that infants can learn novel phonotactic regularities on the basis of brief exposure, and generalize them to novel sounds. An open question is how infants arrive at encoding class-wide patterns, without which no generalization could be observed. One possibility is that they track individual co-occurrence patterns and - based on these - posit a generalization. Alternatively, they might listen for general patterns without encoding sound-specific patterns. To adjudicate between these explanations, we tested French 6-month-olds infants. After familiarization to a phonotactic pattern on word-initial onsets, infants did not distinguish between novel items with the familiarized onsets and those with onsets that generalized the familiarization pattern, providing evidence in favor of the direct encoding of class-wide patterns. In a control experiment, infants did distinguish between novel items with onsets that generalized the familiarization pattern and those with onsets that violated it, showing that the null effect in the first experiment was not due to a lack of learning. Thus, infants are able to quickly form abstract generalizations without encoding sound-specific patterns.

Seventeen French-English bilinguals read aloud a set of English sentences and performed an ABX discrimination task that assessed their perception of the English /I/-/i/ contrast. Global nativelikeness in production correlated with pronunciation accuracy for the vowels /I/ and /i/, and both production measures correlated with self-estimated pronunciation skills. However, performance on the perception task did not correlate with either global nativelikeness or /I,i/ pronunciation accuracy. These results are discussed in light of theories about the relation between perception and production in L2 phonological processing.

Models of the acquisition of word segmentation are typically evaluated using phonemically transcribed corpora. Accordingly, they implicitly assume that children know how to undo phonetic variation when they learn to extract words from speech. Moreover, whereas models of language acquisition should perform similarly across languages, evaluation is often limited to English samples. Using child-directed corpora of English, French and Japanese, we evaluate the performance of state-of-the-art statistical models given inputs where phonetic variation has not been reduced. To do so, we measure segmentation robustness across different levels of segmental variation, simulating systematic allophonic variation or errors in phoneme recognition. We show that these models do not resist an increase in such variations and do not generalize to typologically different languages. From the perspective of early language acquisition, the results strengthen the hypothesis according to which phonological knowledge is acquired in large part before the construction of a lexicon.

Using an artificial language-learning paradigm with a production task, we examine whether the phonetic distance between alternating sounds has an effect on the learning of phonological alternations by adults, and if so, if the effect is discrete or gradient. To this end, we compare the learning of alternations that involve one, two, or three feature changes. We find that alternations that involve one feature change are learned faster and better than those involving two or three feature changes. No difference is observed between the latter two. These results are discussed in light of theories on the nature and origin of phonetic naturalness in phonology.

This study analyzes the effects of sociolinguistic factors on the adaptations and non-adaptations of Spanish borrowings into Nahuatl. Fieldwork was carried out in Tagcotepec, a Nahuatl community of 500 inhabitants in Puebla, Mexico. Data were collected using a stratified sample of 71 speakers, representative of the sociolinguistic traits of the community, in terms of age, sex, level of bilingualism, and location within the village. The compiled database consisted of 671 loanwords for a total of 4631 tokens. The mean percentage of adaptation was 17%, and a multivariate analysis of the data reveals that adaptation is more likely if the speaker lives in the remote part of the village, is in the older age group, is a man, talks to friends and relatives, and if the loanword has a high frequency of use.

We present a computational evaluation of a hypothesis according to which distributional information is sufficient to acquire allophonic rules (and hence phonemes) in a bottom-up fashion. The hypothesis was tested using a measure based on information theory that compares distributions. The test was conducted on several artificial language corpora and on two natural corpora containing transcriptions of speech directed to infants from two typologically distant languages (French and Japanese). The measure was complemented with three filters, one concerning the statistical reliability due to sample size and two concerning the following universal properties of allophonic rules: constituents of an allophonic rule should be phonetically similar, and allophonic rules should be assimilatory in nature.

In previous work using an artificial language-learning paradigm with a phrase-picture matching task, we found that adults learn unnatural, arbitrary, phonological rules to the same extent as natural, assimilatory, ones. In the present article we use the same stimuli but a more natural task, i.e. picture naming, and obtain an unambiguous effect of naturalness: natural rules are learned better than unnatural ones. We speculate that the difference between the two experiments is due to differences in task difficulty. Overall, the present results add to the growing body of evidence concerning the role of phonetic naturalness in phonological learning.

Phonologists have long held an interest in loanword adaptations, that is in the transformations that apply to words when they are borrowed into a foreign language. Within constraint-based frameworks, it has been argued that loanword adaptations are in conformity with the native phonology, and even that they provide insight into it, revealing the relative ranking of faithfulness constraints that would otherwise remain 'hidden'. Against this view, I argue that loanword adaptations are not computed by the phonological grammar of the borrowing language. First, I show that not all loanword adaptations are in accordance with the native phonology. Second, I argue that separating these problematic cases from the remaining loanword adaptations and treating them differently, by making appeal to phonetic and/or perceptual arguments, yields an ad hoc distinction between phonological and non-phonological adaptations. Finally, I propose that a principled solution lies with the hypothesis that all loanword adaptations are phonetically minimal transformations that apply during speech perception. This hypothesis is motivated independently by psycholinguistic data concerning the perception of non-native sound structures.

Standard phonological accounts of loanword adaptations state that the inputs to the adaptations are constituted by the surface forms of the words in the source language and that the adaptations are computed by the phonological grammar of the borrowing language. In processing terms, this means that in perception, the phonetic form of the source words is faithfully copied onto an abstract underlying form, and that adaptations are produced by the standard phonological processes in production. We argue that this is at odds with speech perception models and propose that loanword adaptations take place in perception and are defined as phonetically minimal transformations.

We report two experiments concerning the perception and the acquisition of phonemic versus allophonic contrasts, respectively. In the first experiment, we examine whether French adults exhibit differences in the perception of these two types of contrast. We find that a native allophonic contrast is very well discriminated when presented in isolation, but very poorly discriminated compared to a control phonemic contrast when embedded in context. In the second experiment, we use a training procedure with stimuli in context to explore whether adult listeners can rely upon distributional information in order to construct abstract phoneme categories. We find only a numerically small amount of learning overall and limited evidence for sensitivity to complementary distributions of allophonic variants during learning. These results are discussed in light of theories on statistical learning mechanisms in phonological acquisition.

As in many other Romance languages, main word stress in Neapolitan falls on one of the word's last three syllables. Although clitics are inherently stressless, a string consisting of a verb and two enclitics surfaces with two main word stresses, one on the verb and one on the first clitic. If only one clitic is attached, however, no stress falls on the clitic. In this article, it is argued that encliticization in Neapolitan gives rise to recursive prosodic word structure. In this structure, a disyllabic foot can be built if two clitics are present but not is only one clitic is present, thus accounting for the facts of enclitic stress. Furthermore, it is shown that permutation of the constraint ranking yields representations that make different predictions concerning stress assignment in encliticized forms. These predictions are shown to be borne out by two other languages spoken in Italy, i.e. Lucanian and Standard Italian.

The task of speech perception involves converting a continuous, information-rich waveform into a more abstract representation. This mapping process is heavily language-dependent - every language divides up acoustic space differently, and the mapping is distorted by context-dependent phonological rules. This is not an easy job, but it is made easier in two complementary respects. During the first year of life the human perceptual apparatus is gradually optimized to better perceive the distinctions that are crucial in the ambient language's phonological system while ignoring irrelevant variation, and phonological systems themselves are optimized from the perspective of human perception. This two-way interaction, in which perception adapts to phonology and phonology to perception, has long been of interest to phonologists, but only in recent decades have the tools necessary to explore the connection between the two become available. In this chapter we provide an overview of the debates that have arisen around each issue, as well as the research that bears on each debate We first discuss how phonology influences speech perception, both in the native language and in non-native ones, and how second language perception relates to loanword adaptations. We then address the question of how perception influences phonology, beginning with an overview of the relevant typological data, and concluding with a comparison of theoretical approaches to the data.

This paper discusses the target articles by Fikkert, Vihman, and Goldrick & Larson, which address diverse aspects of the acquisition of phonology. These topics are examined using a wide range of tasks and experimental paradigms across different ages. Various levels of processing and representation are thus involved. The main point of the present paper is that such data can be coherently interpreted only within a particular information-processing model that specifies in sufficient detail the different levels of processing and representation. We first present the basic architecture of a model of speech perception and production, justifying it with psycholinguistic and neuropsychological data. We then use this model to interpret data from the target articles relative to the acquisition of phonology.

When infants acquire their native language they not only extract language-specific segmental categories and the words of their language, they also learn the underlying form of these words. This is difficult because words can have multiple phonetic realizations, according to the phonological context. In a series of artificial language-learning experiments with a phrase-picture matching task, we consider the respective contributions of word meaning and distributional information for the acquisition of underlying representations in the presence of an allophonic rule. We show that on the basis of semantic information, French adults can learn to map voiced and voiceless stops or fricatives onto the same underlying phonemes, whereas in their native language voicing is phonemic in all obstruents. They do not extend this knowledge to novel stops or fricatives, though. In the presence of distributional cues only, learning is much reduced and limited to the words subjects are trained on. We also test if phonological naturalness plays a role in this type of learning, and find that if semantic information is present, French adults can learn to map different segments onto a single underlying phoneme even if the mappings are highly unnatural. We discuss our findings in light of current statistical learning approaches to language acquisition.

Phonological rules introduce variation in word forms that listeners have to compensate for. We previously showed that compensation for phonological variation in perception is driven by language-specific mechanisms. In particular, English speakers compensate more for place assimilation than for voicing assimilation, whereas the reverse holds for French speakers. English indeed has a rule of place assimilation, whereas French has a rule of voicing assimilation. In the present study, we explore the patterns of compensation for assimilation in English learners of French and in French learners of English. We use the same design and stimuli as Darcy et al. (2009); in this design, listeners are engaged in a word detection task on sentences containing occurrences of both place assimilation and voicing assimilation. We test British English and American English learners of French as well as French learners of American English on both their native language (L1) and their second language (L2). The results show that beginners interpret their L2 in exactly the same way as their L1: they apply the native compensation pattern to both languages. Advanced learners, by contrast, succeed in compensating for the non-native assimilation rule in their L2, while keeping the native compensation pattern for L1; as little or no interference from L2 on L1 is observed for these learners, we conclude that two separate systems of compensation for phonological processes can co-exist.

Previous research has shown that native speakers of French, as opposed to those of Spanish, exhibit stress "deafness", i.e. have difficulties distinguishing stress contrasts. In French, stress is non-contrastive, while in Spanish, stress is used to make lexical distinctions. We examine three other languages with non-contrastive stress, Finnish, Hungarian and Polish. In two experiments with a short-term memory sequence repetition task, we find that speakers of Finnish and Hungarian are like French speakers (i.e. exhibit stress "deafness"), but not those of Polish. We interpret these findings in the light of an acquisition framework, that states that infants decide whether or not to keep stress in their phonological representation during the first two years of life, based on information extractable from utterance edges. In particular, we argue that Polish infants, unlike French, Finnish and Hungarian ones, cannot extract the stress regularity of their language on the basis of what they have already learned. As a consequence, they keep stress in their phonological representation, and as adults, they do not have difficulties in distinguishing stress contrasts.

The sound pattern of the language(s) we have heard as infants affects the way in which we perceive linguistic sounds as adults. Typically, some foreign sounds are very difficult to perceive accurately, even after extensive training. For instance, native speakers of French have troubles distinguishing foreign words that differ only in the position of main stress, French being a language in which stress is not contrastive. In this paper, we propose to explore the perception of foreign sounds cross-linguistically in order to understand the processes that govern early language acquisition. Specifically, we propose to test the hypothesis that early language acquisition begins by using only regularities that infants can observe in the surface speech stream (Bottom-Up Bootstrapping), and compare it with the hypothesis that they use all possible sources of information, including, for instance, word boundaries (Interactive Bootstrapping). We set up a research paradigm using the stress system, since it allows to test the various options at hand within a single test procedure. We distinguish four types of regular stress systems the acquisition of which requires different sources of information. We show that the two hypotheses make contrastive predictions as to the pattern of stress perception of adults in these four types of languages. We conclude that cross-linguistic research of adults speech perception, when coupled with detailed linguistic analysis, can be brought to bear on important issues of language acquisition.

Mapping word forms onto their corresponding meanings is one of the most complex tasks that young infants acquiring their native language have to perform. One of the challenges they meet in this respect is that words can surface with different phonetic forms due to the application of postlexical phonological processes; that is, surface word forms exhibit what we call phonological variation. In this article, we examine if and how infants that do not have a lexicon might undo phonological variation, i.e. deduce which phonological processes apply and infer unique underlying word forms that will constitute lexical entries. Two types of phonological rules are considered, which introduce variation in word forms and morphemes, respectively. As to the former, we propose a learning mechanism that deduces which rule applies and infers underlying phonemes and word forms, based on an examination of the distribution of either surface segments or surface word forms. As to the latter, we show that although the rules that introduce morpheme variation can be deduced in the absence of semantic knowledge, no underlying forms can be inferred. These forms, in fact, can be found only if morphological alternations are taken into account.

Recent developments in optimality theory have led to a revision of the role of faithfulness. In particular, it is argued that output candidates can be compared not only with their underlying inputs by means of input-to-output constraints, but also with the surface form of paradigmatically related words by means of output-to-output constraints. Many phonological effects of under- or overapplication formerly attributed to the cycle or to level ordering receive a straightforward one-level analysis in this type of paradigmatic phonology. In this paper, I consider several cases of under- and overapplication in the phonology of Italian, Spanish, Catalan and Dutch. I argue that analyses based on output-to-output faithfulness constraints lead to a variety of problems, which do not arise in accounts based on rule ordering and the definition of rule domains.

Prototypically, compounds consist of two independent words. Many languages, however, also have compounds that are composed of a stem followed by a word, or by two stems. This article addresses the question as to how these different morphological types of compounds are mapped onto prosodic words in Italian, a language that allows for all three morphological compound types.. Two phonological rules that have the prosodic word as domain of application, vowel raising and intervocalic s-voicing, serve as diagnostics for prosodic word status. It is shown that certain compound types exhibit hybrid behavior with regard to these rules, and that previous approaches concerning the prosodic structure of compounds that assume the Strict Layer Hypothesis are unable to account for these data. Alternatively, it is argued that the data can be accommodated if recursive prosodic words are permitted.

This article brings supporting evidence for the hypothesis advanced in previous work that clitics can and should be represented in the prosodic hierarchy without the postulation of the category clitic group. First, cases that have been argued to show the need for the clitic group are reanalyzed without reference to a category clitic group. Second, a novel argument against the clitic group is developed. That is, based on data regarding stress assignment in verb plus enclitic sequences in three languages spoken in Italy, it is argued that Romance clitics are not incorporated into prosodic constituent structure in a uniform manner. Specifically, each of the languages under consideration adopts one of the following strategies for the prosodization of clitics: adjunction to the prosodic word, incorporation into the phonological phrase, or incorporation into the prosodic word by restructuring. This interlinguistic variation is derived within Optimality Theory by the permutation of constraints on prosodic constituency and the mapping from morphosyntactic onto prosodic structure.

Prosodic Words examines phonological aspects of derivation, compounding and cliticization, with particular regard to the definition of the prosodic word. Based on data from various languages, including Italian and Spanish, the prosodization of derivational affixes, compound elements, and clitics is analyzed. Attention is paid, furthermore, to the consequences of phrasal resyllabification for the structure of the prosodic word, as well as to several morphological issues regarding derivation and compounding. The proposals are set within the constraint-based framework of optimality theory. In particular, this approach is argued to allow for a constrained account of the occurrence of marked representations in prosodic phonology.