Alexandra Marquis

Initial morphological learning in preverbal infants

How do children learn the internal structure of inflected words? We hypothesized that bound functional morphemes begin to be encoded at the preverbal stage, driven by their frequent occurrence with highly variable roots, and that infants in turn use these morphemes to interpret other words with the same inflections. Using a preferential looking procedure, we showed that French-learning 11-month-olds encoded the frequent French functor /e/, and perceived bare roots and their inflected variants as related forms. In another experiment an added training phase presented an artificial suffix co-occurring with many pseudo-roots. Infants learned the new suffix and used it to interpret novel affixed words that never occurred during the training. These findings demonstrate that initial learning of sub-lexical functors and morphological alternations is frequency-based, without relying on word meaning.

Small-signal modeling and linear control of a dual boost power factor correction circuit

In this paper, a small-signal modeling of a single-phase AC-DC dual boost power factor correction (PFC) pre-regulator is presented. The circuit consists of two interleaved boost converters. A high rated main boost, used for current wave-shaping and energy control, operates at low frequency; whereas an auxiliary boost uses a low rated device operating at high frequency and is controlled to compensate the low frequency ripple generated by the main boost. The control scheme proposed for the dual PFC is based on a small-signal model, which is developed using the average modeling technique along with the small-signal linearization process. The linear controllers are designed in the frequency domain, and are adjusted in order to ensure high dynamic performance of the PFC. Experimental results based on a laboratory 1 kW prototype validate the proposed control approach.

Speech articulation performance of francophone children in the early school years: Norming of the Test de Dépistage Francophone de Phonologie

Abstract Good quality normative data are essential for clinical practice in speech-language pathology but are largely lacking for French-speaking children. We investigated speech production accuracy by French-speaking children attending kindergarten (maternelle) and first grade (première année). The study aimed to provide normative data for a new screening test – the Test de Dépistage Francophone de Phonologie. Sixty-one children named 30 pictures depicting words selected to be representative of the distribution of phonemes, syllable shapes and word lengths characteristic of Québec French. Percent consonants’ correct was approximately 90% and did not change significantly with age although younger children produced significantly more syllable structure errors than older children. Given that the word set reflects the segmental and prosodic characteristics of spoken Québec French, and that ceiling effects were not observed, these results further indicate that phonological development is not complete by the age of seven years in French-speaking children.

An automated version of the BAT Syntactic Comprehension task for assessing auditory L2 proficiency in healthy adults

Studies of bilingualism sometimes require healthy subjects to be assessed for proficiency at auditory sentence processing in their second language (L2). The Syntactic Comprehension task of the Bilingual Aphasia Test could satisfy this need. For ease and uniformity of application, we automated its English (. The Bilingual Aphasia Test. English version. Hillsdale, NJ: Lawrence Erlbaum Associates) and French (. The Bilingual Aphasia Test, French version. Hillsdale, NJ: Lawrence Erlbaum Associates) versions. Although the Bilingual Aphasia Test is meant to assess neurological disorders affecting language, we hypothesised that ceiling performance in L2 would be rare and L2 errors should be consistent with lack of processing automaticity. Initial data from 13 French–English and 4 English–French bilinguals confirm these expectations. Thus, the automated Syntactic Comprehension task (available online for PC and Mac platforms) is indeed suited to test bilingual English and French proficiency levels in healthy adults.

Word-meaning association in early language development

Existing word-learning experiments with preverbal infants typically use visual scenes including changing objects but constant motion, thus cueing infants to ignore the motion aspect. The present study examines whether infants naturally focus on the object or its motion if both interpretations are possible. Eleven-month-old French-learning infants were habituated to either un gope or un fime (gope, fime are nonsense words, un is a determiner), paired with a visual scene, e.g., an artificial animal jumping over a barrier. Upon habituation, infants received one baseline trial (the same audio-visual paring as in Habituation), and then two types of Test trials with scene changes: action change (the same animal doing a different motion, e.g., bouncing onto a wall), and object change (an artificial vehicle doing the same habituated motion). Test trials were presented with the same habituation speech. We expected that if infants mapped the nonsense word to the object during habituation, their looking time to the object-change scene should recover the most (indicating surprise); if they interpreted the nonsense word as the motion, looking time to the action-change scene should recover more than the object-change scene. Results showed that infants looked significantly longer at the object-change scene than the action-change and baseline scenes, suggesting that they mapped the word to the object rather than motion. Therefore, when both the object and its motion were possible interpretations, preverbal infants map a novel word as a noun. We are conducting further experiments to ascertain whether determiners contribute to infants’ noun interpretation in this task.

Insensitivity to verb conjugation patterns in French children with SLI

ABSTRACT Specific language impairment (SLI) is characterised by persistent difficulties that affect language abilities in otherwise normally developing children (Leonard, 2014). It remains challenging to identify young children affected by SLI in French. We tested oral production of the passé composé tense in 19 children in kindergarten and first grade with SLI aged from 5;6 to 7;4 years. All children were schooled in a French environment, but with different linguistic backgrounds. We used an Android application, Jeu de verbes (Marquis et al., 2012), with six verbs in each of four past participle categories (ending in -é, -i, -u, and other irregulars). We compared their results and error types to those of control children (from Marquis, 2012–2014) matched for gender, age, languages spoken at home, and parental education. Results show that children with SLI do not master the passé composé in the same way as typical French children do, at later ages than previously shown in the literature. This task shows potential for oral language screening in French-speaking children in kindergarten and first grade, independently of language background.

The Factors Contributing to Teacher Predictions of Spelling Ability, and the Accuracy of Their Assessments

In this study, teachers of kindergarten and Grade 1 French-speaking students indicated the likelihood their students would develop later writing difficulties. Results showed that language measures, language background, the education levels of parents, and home literacy practices predicted whether children would be identified as at-risk. Moreover children’s oral language skills accounted for even more of the variance in teacher ratings than other variables. Spelling performance assessed 1-year later from a subset of children indicated that the teacher predictions were accurate. Thus, teachers appear to be an effective source for predicting children’s future literacy performance.

The Beginning of Morphological Learning: Evidence from Verb Morpheme Processing in Preverbal Infants

The goals of our chapter are to study infants’ acquisition of verb morphology and develop a model of early morphological learning. We approach this acquisition by studying French-learning infants’ initial segmentation of verb forms and their interpretation of verb morphological alternations. Our empirical findings demonstrate that infants begin to parse verbs into decomposed stems and suffixes by 11 months of age, and they have rudimentary knowledge of regular verb paradigms by 14 months of age. We show that this learning is based on distributional analyses of the input without the need for semantics.

The recognition of frequent suffixes and the parsing of verb roots in 11-month-old infants

Infants begin analysing distributional patterns of speech by 8-9 months of age (e.g., Jusczyk et al, 1994). Here we inquire whether infants recognise frequent verb suffixes and use them to parse verb roots. We measured French-learning 11-month-olds’ detection of verb roots using a looking procedure. In Experiment 1, infants heard isolated tokens of an infrequent verb root, /bif/ or /tar/, and were then tested with sentences containing both roots in conjugated forms, /bife/, /tare/. Results showed a listening preference (longer looking time) while infants heard sentences containing the targets, suggesting that infants parsed both roots. The parsing of /bif/ was particularly complex, involving not only resyllabification, but also the [i]-[I] allophonic alternation in [bIf][bife]. Experiment 2 examined if the parsing was indeed due to infants’ recognition of /e/ morpheme. Infants heard the same target roots but then heard sentences containing the roots conjugated with a nonsense morpheme /u/ (i.e., /bifu/, /taru/). We expected that infants would consider /bifu/, /taru/ as monomorphemic forms and would not extract the unrelated /bif/, /tar/ from the /u/ context. Indeed, infants who heard the target /bif/ did not prefer /bifu/ sentences, confirming our expectation. Infants exposed to the target /tar/, however, preferred /taru/ sentences. This response may reflect infants’ recognition of /ta/ (homophonic to the frequent function word “ta”) in /taru/, rather than the parsing of /tar/. We are conducting additional experiments to explore this question. Taken together, our data provide the first evidence supporting prelinguistic infants’ ability to use suffixes to parse verbal roots.

Frequency factor in the segmentation of function words in French‐learning infants

Previous work showed that 8‐month‐old French‐learning infants segment both highly and moderately frequent function words from continuous speech [Shi and Gauthier, J. Acoust. Soc. Am. 117, 2426 (2005)]. In the present study we examine if high‐frequency function words are segmented by even younger infants. Participants were 6‐month‐old Quebec‐French‐learning infants. As in the previous study, infants in the high‐frequency condition were familiarized with a target functor (des or la) and were then tested with phrases containing the target versus those containing the nontarget. Thus, for infants familiarized to la, la‐phrases were the target phrases and des‐phrases were the nontarget phrases during Test. The reverse was the case when des was the familiarization target. The lower‐frequency condition included a target functor mes versus ta (prosodically analogous to the functors in the high‐frequency condition). The results show that 6‐month‐olds listened significantly longer to phrases containing the familiari...