Séminaire de psycholinguistique (Printemps 2017)
Responsable: Dr Julie Franck (email@example.com)
Les séminaires ont lieu les lundis de 12h15 à 13h45 dans la salle M3393 à Uni Mail
Programme des invités externes
6 Mars 2017
Andrew Nevins (UCL, London): When subject-verb agreement is influenced by object-verb agreement: ERP evidence from Basque
Previous cross-linguistic research has found that comprehenders are immediately sensitive to various kinds of agreement violations across languages (e.g., Sally runs/*run) Molinaro, Barber & Carreiras (2011); Nevins, Dillon, Malhotra & Phillips (2007). Many studies have examined how comprehenders compute agreement based on different features (e.g., person, number, gender) and between different constituents (e.g., subject-verb vs. object-verb agreement), but much less is known about the relationship between agreement processing and the grammatical properties of the constituents involved. For example, how does sensitivity to subject-verb (S-V) agreement violations vary as a function of verb transitivity? We focus on Basque, an SOV ergative language with both S-V and object-verb (O-V) agreement and hypothesize that S-V agreement processing (specifically, agreement with the ergative) may be affected by the presence vs. absence of O-V agreement. We found that S-V agreement violations elicited qualitatively different event-related brain potential (ERP) responses in transitive (ergative+absolutive) vs. intransitive (absolutive only) sentences. This provides evidence for distinct real-time agreement computations between ergative subjects (the grammatically-marked subjects of transitive verbs) and absolutive subjects (the grammatically-marked subjects of intransitive verbs), and leverages the typological contribution of a language in which they are morphosyntactically distinguished.
6 Fevrier 2017
Micah Murray (Centre Hospitalier Universitaire Vaudois and University of Lausanne): The multisensory scaffolding for perception across the lifespan
This talk summarizes our efforts to identify the spatio-temporal brain mechanisms and behavioural relevance of multisensory interactions in humans and the consequence such has had on our understanding of the organization of the brain, the functional selectivity of low-level cortices, and plasticity across the lifespan. Across studies we have used combinations of psychophysics, ERPs, fMRI and TMS, taking advantage of innovations in signal processing to yield greater mechanistic interpretability of the data. Several general conclusions are supported by the collective data. First, (near) primary cortices are loci of multisensory convergence and interactions. Second, these effects occur at early latencies (i.e.
Paul Matusz (University Hospital Centre (CHUV) – University of Lausanne) : Taking attention back to school: Multisensory contexts reveal multi-pronged effects of experience on selective attention
The last decades have highlighted the role of both sensory and top-down influences (e.g., memory, goals) in adult attention. This research enabled investigations into how learning and object perception occur in real-world environments, where tasks typically differ in their difficulty and information stimulates multiple senses at once but where also individuals differ in their ability to select task-relevant information (selective attention). Are the current models, derived primarily from unisensory and/or adult research, sufficient to explain the challenges of selecting information in multisensory environments and its interactions with experience? In this talk, I will first present results demonstrating how individuals of different ages allocate attention towards simple, highly familiar objects: coloured shapes (Matusz et al. 2015 Cognition). Using an audiovisual, age-appropriate version of the response-competition task, we showed that attention is most strongly allocated to peripheral distractors matching target features across both senses, despite these stimuli being task-irrelevant. With increased task demands (larger search arrays), in older children this interference was still present, but in children entering school it disappeared. Thus, developing selective-attention skills paradoxically “shields” young individuals from multi-sensory distraction. We next tested whether experience/expertise interacts with developing selective attention, using Arabic digits as education-relevant but less familiar to children objects. Participants now searched for pre-defined visual numerals, while we measured the attentional “costs” and “benefits” of distractors representing the three object format types. In response to digit verbal labels, search in adults and older children was only reliably slower (to target-incompatible sounds), whereas in school-entering kids only faster (to target-compatible sounds). Notably, school-entering kids were generally faster when distractor sounds vs. shapes accompanied search. As digit sounds are the object formats learnt faster in childhood, these results suggest an initial involuntary positive influence of object experience on selective attention – that further experience turns then into a cost. Together, these results challenge the traditional models of selective attention development that portray top-down control processes as gradually improving till they reach an “adult state”, and highlight the value of adaptations of rigorous adult-attention paradigms in clarifying the dynamic interplay between selective attention, object perception and experience in real-world, multisensory environments.
SEMESTRE D'AUTOMNE 2016
5 Septembre 2016
Arne Lervag (University of Oslo): Development of Reading Comprehension and General Language Skills: Can General Language Skills be Enhanced by Training?
This talk consists of two parts: 1) the reciprocal development of reading comprehension and general language skills in primary school and 2) the training of general language skills in children with poor language skills in kindergarten.
In the first part we show that language is a consistent predictor of the growth of both early and later reading comprehension skills. In addition, language and word decoding explains almost all individual differences (approximate 95%) in the early years of primary school. Later on it still explains a lot of variance but not all (approximate 70% in Grade six). Further, even if general Language skills have an extreme relative stability between second and sixth grade, both beginning and early growth of reading comprehension are able to explain additional variance beyond the strong autoregressor – suggesting a reciprocal development.
In the second part we rapport the results of two randomized control trials (RCT’s) where we aimed to train general language skills in 1) minority speaking children and 2) monolingual children with poor language skills. The results show that it is possible to train language skills in both groups and to get moderate to large effects on several distal measures (vocabulary, grammar narrative retelling). However, even if the results suggest that it is possible to train several important language components it is difficult to reach a general language factor and thereby improve the language skills all over.
10 Octobre 2016
Lisa Morano (University Nijmegen): Looking for exemplar effects: testing the comprehension and memory representations of r'duced words in Dutch learners of French and French native speakers
L1 identity priming experiments (e.g. Tulving and Schacter, 1990) have repeatedly shown that listeners recognize words faster when they occur for the second time (as "targets") than when they occurred for the first time (as "primes") especially if the two tokens share fine acoustic characteristics such as the speaker's voice (i.e. both the prime and the target are uttered by the same person; McLennan and Luce, 2005). In an earlier paper, the authors reported the same findings with allophonic variation: e.g. better was recognized faster when its intervocalic consonant was consistently produced either as a /t/ or as a flap for both prime and target (McLennan, Luce, and Charles-Luce, 2003). These specificity (or exemplar) effects suggest that participants stored the occurrences they heard with at least a certain degree of acoustic detail, in the form of exemplars.
If most researchers now assume that the mental lexical is hybrid in nature, containing both abstract representations and exemplars, exemplar research, however, has almost never been applied to second language (L2) learners. We could find only one published study reporting exemplar effects for L2 learners manipulating indexical variation (i.e. speaker's voice; Trofimovich, 2005). We wanted to replicate this finding by manipulating acoustical (or pronunciation) variation instead of indexical (or speaker) variation and using a lexical decision task instead of an immediate repetition task.
In our first study, we tested Dutch learners of French on French schwa deletion. For example, if participants accurately recognized la f'nêtre, were they subsequently faster at recognizing la f'nêtre again than la fenêtre, and conversely? In our second study we investigated high vowel devoicing. In casual French, in a word like la cité, the /i/ can be devoiced as the voicing fails to be reestablished in time after the /s/. The question in this case was thus, if participants accurately recognized a devoiced token, were they then faster at reacting upon hearing again a devoiced token of the word than a voiced token, and vice versa?
Our results show no exemplar effects for our two experiments on French schwa alternation, and mixed results for high vowel devoicing. In one experiment on devoicing, we obtained significant exemplar effects; while in another experiment, we obtained a significant mismatch effect (e.g. the voiced token la cité, [la.si.te], was answered to faster when it was primed by a devoiced token, [la.si̥.te], than a voiced one, [la.si.te]), depending on whether the prime and the target were the same recording (experiment one) or two different recordings (experiment two).
This difference of findings between study one and two could be due to the nature of the linguistic phenomena under investigation. French schwa alternation is considered to be a categorical phenomenon and it has been shown that French native speakers probably store two abstract representations for each word, one with and one without schwa (Bürki and Frauenfelder, 2012). High vowel devoicing, on the other hand, is a continuous phenomenon and, contrary to schwa alternation, native speakers are not aware of doing it and cannot produce it consciously. In our third study, we thus wanted to test this hypothesis by testing French native speakers for exemplar effects using both schwa alteration and high vowel devoicing.
21 Novembre 2016
Theo Marinis (University of Reading): How do sequential bilingual children process sentences in real-time: asymmetries between on-line comprehension vs. production
In this talk I will present a series of studies investigating sentence processing in sequential bilingual (L2) children with Turkish as the first language and English, Dutch or Greek as a second language along with groups of monolingual (L1) typically developing children. The children took part in on-line production and comprehension tasks investigating phenomena within the verbal (subject-verb agreement) and nominal domain (articles, pronouns, gender). The results demonstrate that although L2 children make a lot of production errors, they are capable of using morpho-syntactic cues for predictive processing. Differences between on-line comprehension and production will be discussed in relation to models of production and comprehension.
5 Décembre 2016
Juliana Gerard (Ultser University): Interference effects in the acquisition of adjunct control
Previous research on 4-6 year-olds’ interpretations of adjunct control, as in (1), has consistently reported non-adultlike behavior for adjunct PRO (Goodluck 1981, Hsu et al. 1985, McDaniel et al. 1991, Goodluck & Behne 1992, Cairns et al. 1994, Broihier & Wexler 1995, Goodluck 2001, Adler 2006).
(1) John bumped Mary after PRO tripping on the sidewalk.
A number of different tasks have been used with the aim of identifying a grammatical source of children’s errors. In this talk, I will consider other factors that may have influenced children's behavior, which will link the results observed for explicit measures in children with those observed with implicit measures in adults. In particular, for sentences like (1), we find that the same kind of manipulation that is observed with adults to modulate effects of similarity-based interference is also found to influence behavior in 4-5 year old children: error rates for sentences with adjunct control go up when the similarity increases between an antecedent (John) and a linearly intervening noun phrase (Mary). This suggests that difficulties with adjunct control are to be explained (at least in part) by the sentence processing mechanisms that underlie similarity-based interference in adults. Further questions arise, then, about how these mechanisms develop, and about the properties of the memory architecture that give rise to these effects, to be investigated in future research.
12 Décembre 2016
Maria Teresa Guasti (University of Milano Bicocca): Predictive Timing in Developmental Dyslexia
In the present work, we argue that the mechanism at the basis of sensory temporal prediction may be compromised in children and adults with Developmental Dyslexia (DD). A breakdown of this mechanism may be the cause of reading impairments in individuals with DD as well as of subtle language (e.g., Cantiani et al., 2013, 2015; Rispens et al., 2006) and motor (Nicolson & Fawcett, 1990) problems frequently attested at least in some dyslexic individuals. In Study 1, eighteen adults with DD (mean age 22;3) were tested along with twenty controls in a task requiring entrainment to a given rhythm in a warning-imperative paradigm. DD participants displayed greater synchronization error and were more variable than controls in high predictable stimuli. No group difference was found in the control condition, in which the uncertain timing occurrence of the beat did not permit the extraction of regularities. The findings of Study 1 were replicated in Study 2, which looked at predictive timing abilities in fourteen children with DD (mean age 9;75). These results suggest that individuals with DD are unable to exploit temporal regularities to efficiently anticipate next sensory events. This impairment is likely to affect both the reading/language and the motor domain.
16 Janvier 2017
Klaus Oberauer (University of Zurich): Interference in working memory: Computational mechanisms and experimental evidence
I will introduce general principles of modelling maintenance of information in working memory, and explain how interference arises naturally in such a model, using a computational model of serial recall – the SOB-CS model – as an example. The model predicts that similarity between representations in working memory is helpful to performance in some conditions, and harmful in other conditions. I will present experimental evidence supporting these predictions.