The earliest stages of second language learning: A behavioral investigation of long-term memory and age

Diana Pili-Moss
DOI: https://doi.org/10.31234/osf.io/7p5f2
2019-07-25
Abstract:A study with 40 L1 Italian 8-9 year old children and its replication with 36 L1 Italian adults investigated the role of declarative and procedural learning ability in the early stages of language learning.The studies investigated: (1) the extent to which memory-related abilities predicted L2 learning of form-meaning mapping between syntax and thematic interpretation, word order and case marking; and (2) the nature of the acquired L2 knowledge in terms of the implicit/explicit distinction.Deploying a computer game in incidental instruction conditions, the participants were aurally trained in the artificial language BrocantoJ over three sessions. Standardized memory tasks, vocabulary learning ability, and an alternating serial reaction time task provided measures of visual/verbal declarative and procedural learning ability. Language learning was assessed via a measure of comprehension during practice and a grammaticality judgment test.Generalized mixed-effects models fitted to both experimental datasets revealed that, although adults attained higher accuracy levels and were faster learners compared to children, the two groups did not differ qualitatively in what they learned. However, by the end of the experiment, adults displayed higher explicit knowledge of syntactic and semantic regularities. During practice, declarative learning ability predicted accuracy in both groups, but procedural learning ability became an increasingly stronger predictor of L2 comprehension only in children. The role of procedural learning ability emerged again only in the child grammaticality judgment test dataset, where it was a statistically significant predictor of learning of word order. In the practice data declarative learning ability and vocabulary learning ability interacted negatively with procedural learning ability in children, whereas declarative learning ability interacted positively with procedural learning ability in adults. Moreover, the positive interaction in adults only obtained for a subset of practice stimuli, i.e. sentences where the processing of linking between morphosyntax and thematic interpretation was required. Overall, the findings support age-related differences and linguistic target differences in the way abilities related to long-term memory predict language learning.
What problem does this paper attempt to address?