Academic journal article Memory & Cognition

Time versus Items in Judgment of Recency

Academic journal article Memory & Cognition

Time versus Items in Judgment of Recency

Article excerpt

In the numerical judgment of recency (JOR) task, subjects judge how many items have intervened since the test item was previously presented. Two experiments were conducted to determine whether the basis of JOR is the age of the memory (time) or the number of intervening items. Subjects went through a long list that was made up of alternating fast blocks and slow blocks, but the block structure was disguised by probabilistic selection of a short or long intertrial interval. In both experiments, JOR was found to be a simple function of time, with no added contribution from the number of items.

It is essential to everything remembered that it be something which is past; and we cannot conceive a thing to be past, without conceiving some duration, more or less, between it and the present.

(Thomas Reid, 1785/2002, pp. 258-259)

In the numerical judgment of recency (JOR) task, experimental subjects are asked to judge how many items have intervened since the test item was previously presented. The number of items is, of course, highly correlated with time. There are pragmatic reasons for asking subjects to judge recency in terms of items. Time is experienced as continuous and is divided only by convention into arbitrary measures such as seconds and minutes. By contrast, the number of discrete items is a concept that everyone immediately grasps. Regardless of the way the question is asked, a numerical JOR might ultimately be based on an intuitive sense of how much time has passed since the test item was presented before. To transform a sense of the memory's age into a numerical JOR, one would just need an idea of the overall presentation rate.

Several theories predict that JOR should be primarily determined by time. These include a model in which JOR is based on decaying trace strength (Hinrichs, 1970) and one in which it is based on the degree to which the memory trace has consolidated (Wickelgren, 1972, 1974). Others have proposed that each memory incorporates a "time tag" (Yntema & Trask, 1963) or "temporal address" (Gallistel, 1990). One version of the latter hypothesis holds that time of encoding is recorded by reference to the states of coupled oscillators of different frequencies (G. D. A. Brown, Preece, & Hulme, 2000). To formulate a JOR, one would compare the present oscillator states with those that the test item retrieves from memory and use an estimate of overall presentation rate to transform time into items.

Other theories predict that JOR should be determined by the number of intervening items. In a 1682 lecture, Robert Hooke (1705/1969) proposed that memory traces were anatomically arranged in a chain, according to serial order, and that recency could be inferred from a physical distance in the brain (see Hintzman, 2003b). Similar to Hooke's model, but limited to JOR over very short lags, is the idea that the subject counts or scans back to the test item, in a short-term buffer store (Atkinson & Shiffrin, 1968; Hacker, 1980; Murdock, 1972).

A different kind of hypothesis holds that JOR is based on changes in cognitive context. The test item retrieves contextual information from memory, and the retrieved context is compared with the current context-the greater the difference, the greater the judged lag (Hintzman, 2002). If one adopts the assumption that changes in cognitive context are driven primarily by task-related processing of stimuli (see Block, 1990; Howard & Kahana, 2002), then this view predicts that JOR will be determined mainly by the number of intervening items.

The present experiments were done to determine whether JOR is more closely related to items or to time. Subjects went through a long, continuous list of stimuli, making old-new recognition judgments to each and a numerical JOR to each item they judged to be old (see Hintzman, 2001, 2002, 2003a). The list was organized into alternating fast blocks and slow blocks, but the block structure was disguised by making selection of the intertrial interval (ITI) probabilistic. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.