Bridging the gap between learning and memory

This thesis uses a fusion of recognition memory and implicit learning methods to investigate performance based on implicit learning. A series of experiments exposed participants to a study list composed of natural words that conform to a conjunctive rule-set involving the frequency and the concreteness of the words. Participants were asked either to identify words seen on the study list or to identify rule-consistent words. Across a variety of learning conditions signal-detection analyses revealed that participants used both the episodic status of the words (the episodic effect) and the structural status of the word (the structural effect) in making their decisions. Questionnaires indicated that participants could not verbalise the conjunctive rule-set. Increasing the number of repetitions of each word on the study list increased the magnitude of the episodic effect but not that of the structural effect. In addition, a classic strength-based mirror effect was found in which endorsements to words on the study list increased with repetitions but endorsements to both new rule-consistent and new rule-inconsistent words decreased. Discussion of recognition-memory models and a set of MINERVA simulations demonstrated that current recognition memory models cannot account for these results. Implicit learning theories also struggle to account for the invariance of the structural effect to repetitions. It is concluded that familiarity underlies both the structural effect and a portion of the episodic effect, but that the precursors of familiarity are different in each case with structural familiarity being insensitive to repetitions and episodic familiarity being sensitive to repetitions. Implications for recognition and implicit learning theories are discussed

Name dc:type.qualificationname

Ph.D.

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

University of Southampton

Year dc:date.issued

2011

Author dc:creator

• Neil, Gregory James

Advisor dc:contributor.advisor

• Higham, Philip A.