The quick and the cheap: using puzzles as experimental tools. New Trends in Experimental Psycholinguistics (original) (raw)

Evaluation of linguistic abilities and mastery in children under school age can be a daunting task. Although children appear to have strong constraints on their linguistic behavior, they can often be unreliable participants in formal testing of experimental situations. When results are poor, it is often unclear whether the child understands what is asked of her, or whether the task is too difficult, or just badly designed. I report on a method to elicit structures of varying complexity, that children master at early ages, and can tap into language specific structures in an ecological way. Researchers and clinicians have occasionally used puzzles to elicit specific structures from children. These are especially useful for the evaluation of phonological and syllabic inventories (Bulle, xxxx), and even more complex structures such as verbs and their complements with case marking in German or Japanese (see Eisenbeiss, 2011). Our experiment was designed to elicit agreement in the French noun phrase and the more complex structures of the determiner phrase (DP). French speaking children seem to master agreement at very young ages. However, their productive use of adjectives is difficult to ascertain form the corpus (Royle & Valois, 2010). Our experimental technique was to develop puzzle boards of growing complexity, with puzzle pieces having minimal differences along one or two dimensions at a time (color or size). The participants are asked to name the piece they want to insert on the puzzle, and have to produce the appropriate adjectives in order to do this. This task is quite easy to use with children as young as three years old, and has proven useful in distinguishing children with SLI from normally developing children (Royle et al, 2010), as well as identifying specific difficulties in the acquisition of French L2 by Spanish-speakers (Bergeron et al, 2011) and in tracking normal development of the DP structure. These tasks are easily adaptable to other languages with similar linguistic features (here gender agreement). A Spanish version has been with Spanish-speaking children in New York (Royle et al, in preparation). The data from NYC is quite interesting as children evidence code-switching in these controlled settings, whereas Spanish-speaking children from Montreal do not (Bergeron et al, 2011). Thus the puzzles seem to tap into sociolinguistic styles in addition to morpho-syntactic structures. We will present a review of our findings for these different groups of children. Finally, in order to understand the cognitive processes underlying agreement, we have adapted our design to event related potential (ERP) methodology, presenting children with images of colorful objects concurrently with concordant or discordant auditory stimuli, while recording their brain activity (Royle et al, 2009-20012). We have established that typical ERPs related to agreement errors (LAN, P600) can be elicited without asking for grammaticality judgments form participants (Gascon et al, 2011) and that is will be useful for evaluating the emergence of agreement in young children. We will report on ongoing child data collection on this task.