Johnstone, Theresa; (1999) Structural versus processing accounts of implicit learning. Doctoral thesis (Ph.D), UCL (University College London).

Text Structural_versus_processing_a.pdf Download (3MB)

Abstract

Artificial grammar learning (AGL) experiments are used to investigate the possibility that separate general and specific learning systems exist that store implicit, abstract rule knowledge versus specific, episodic knowledge. Chapter 1 concludes that the most urgent priority is to settle continuing debates about whether memorisation of grammatical examples leads to rule, exemplar, or fragment knowledge. In Chapter 2, it is recommended that a biconditional grammar be used to settle the knowledge debate, as rule, exemplar, and fragment knowledge are inevitably correlated in finite- state grammar generated stimuli. Chapters 3 and 4 present evidence that memorising grammatical training examples leads to fragment, but not rule or exemplar knowledge. In contrast, active hypothesis testing is required to gain rule knowledge. Chapters 5 and 6 demonstrate that knowledge gained by memorising training examples is explicit according to objective recognition, cued-recall, and subjective confidence tests. Using a biconditional grammar, there is no support for a dichotomy between general and specific learning systems The results are best explained by one episodic- processing system that records processing of specific structural aspects of grammatical items in order to meet the demands of the training task. During classification, knowledge may appear to be implicit when participants are unaware of the source of fluent processing (i.e., when they unconsciously use fragment knowledge to classify test items as grammatical or ungrammatical). However, instructions that draw attention to the relevant knowledge (i.e., in cued-recall and recognition tests) reveal that classification knowledge is explicit.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item