Non-adjacence dependency learning/processing research: methodological issues

Dr. Cristina Name - (NEALP/UFJF, CNPq)

Non-adjacent dependencies (NAD) in sentence surface concern underlying abstract grammar rules. Thus, identifying its co-ocurrence patterns is crucial for language acquisition and sentence processing. Several studies have provided evidence of NAD learning/processing by infants and adults. Nonetheless, when we compare the way infant and adult learning abilities are tested, we may observe some methodological differences. For example, in a familiarization-novelty paradigm, infants were exposed to strings during a brief training phase. After this, they were presented to new stimuli containing NAD patterns, half of them congruent to the grammar underlying training strings. Infant sensitivity to (un)grammaticality of tested strings is captured by comparing mean listening times to grammatical versus ungrammatical trials, an on-line measure of an unconscious behavior (Gómez 2002, Gómez & Maye 2005, van Heuten & Shi 2010, Laguardia et al. 2015, Name & Shi 2015 among others). On the other hand, adult NAD learning/processing research has been conducted using off-line techniques, such as forced-choice task and grammaticality judgment task (e.g., van den Bos et al. 2012, Grama et al. 2013, Siegelman et al. 2018). Also, participants were conscious about the general goal of the experiments, since they were informed that they would be trained in an unfamiliar language and after would be tested on their knowledge about the language. In this talk, I will review some experiments, focusing on its methodological issues, and present an experiment in progress which aims at investigating on-line processing of NAD pattern by adults.