Abstract

While it has been established that people can learn the regularities in letter or symbol sequences without intending to and without being aware of what they have learned (known as ‘implicit learning’), there is a debate whether this kind of learning contributes to second language acquisition. The debate seems particularly contentious with regards to learning the meanings of words, because unlike learning sequences of letters etc. this involves forming mappings between different domains of information. Can L2 learners infer anything about the use of new words without being specifically instructed? If so, is this knowledge always explicit? The resolution will crucially depend upon accurate measurement of implicit knowledge.

I address these issues in the context of learning the semantic collocational patterns of novel verbs. In the presented experiments novel verbs regularly took either an abstract or a concrete noun as a complement. The results suggest that people indeed unconsciously extract such semantic generalisations about verb usage.

I will present two series of experiments. In the first series, familiarity judgements were used as a measure of learning and subjective report as a measure of awareness. The results suggest that participants were likely to mistakenly think that they had seen word pairs which did not occur in the training part, but did follow the semantic regularity. This held true when participants reported no awareness of any regularity in the stimuli. In the second series of experiments reaction-times were used as an indication of learning, and subjective report as a measure of awareness. The reaction-times to the stimuli violating the initial regularity showed a significant slow-down, even when participants claim not to have realised any regularity. The findings confirm that learners’ behaviour can be directly influenced by implicit semantic knowledge of collocational patterns. The benefit of this paradigm is that it ensures automaticity both at encoding and at retrieval of the task-relevant information, therefore minimising the interference with explicit control.

I conclude that people can pick up usage patterns of new verbs without intending and without being aware. I will discuss these results in the light of the theories of SLA and implicit learning.