| Summary: | Time-code theories of temporal memory argue that memories are tagged with dedicated temporal information which assist temporal judgments (e.g., G. D. A. Brown & Chater, 2001). Though this account is studied mainly with short-term memory paradigms, some models propose such information is available across time periods. This thesis investigates whether intrinsic time information may facilitate temporal judgments at longer spans, using long-term memory (LTM) paradigms and investigations of remote memory. An alternative proposal is considered where we may rely on contextual associations to make these judgments. In LTM, judgments of recency on studied items were not more accurate for recently seen items, contrary to the time-code hypothesis. Neither do temporal ratios of the distance between items and the present relate to accuracy scores. Instead, the presence of a robust primacy effect, preserved when rehearsal is minimised, supports a reconstructive approach where the beginning of a list acts as a temporal landmark. This position is supported by experiments which establish that this landmark effect can be reproduced at other list positions, for events that follow that landmark, and that a corresponding recency effect is not evoked by greater expectation for the end of the list. Investigation of remote memory revealed that for a set of public events dating was no more accurate for individuals who had lived through them, and for those cases dating accuracy was unrelated to measures of primary memory, such as memorability or content knowledge, but did relate to where the event could be localised within a personal life period. A case series investigation of neurological patients suggests an interrelatedness between measures of order memory and forms of context memory, and presents cases where order memory is impaired despite a normal ability to estimate time durations. These studies are broadly supportive of a contextual account of temporal memory.
|
|---|
