The loss of repetition priming and automaticity over time as a function of degree of initial learning

Abstract

Two experiments were performed to investigate the buildup of repetition priming in a lexical decision task with repeated presentations and its decline over the course of 2 months. Priming was found to accumulate as a power function of presentations and to decline as a power function oftime. Accuracy measures indicated that the loss rate of priming was unaffected by the amount of initial priming. Response time measures indicated the same result when the experiments were analyzed separately; however, when the data were combined, increased initial priming was associated with greater losses in priming over time. The data were interpreted in terms of automaticity, and the power function decline in priming was taken as support for memory-based models of automaticity. Possible ways to incorporate forgetting into memory-based theories of automaticity are discussed. 611Repetition priming is the benefit accrued to the processing of a unit of information as a result of its having been encountered previously, with the benefit usually measured in terms of faster response time or greater accuracy. Duration is a fundamental aspect of the phenomenon. In the case of repetition priming in lexical decision tasks, for example, the recognition of a letter string as a word implies that the word has been seen in the past. For the concept of repetition priming to have any meaning, priming must be temporary. Understanding the transitory nature of this repetition effect is central to the understanding of repetition priming.Sloman, Hayman, Ohta, Law, and Tulving (1988) reviewed several studies on the duration of repetition priming over relatively short intervals and then conducted a series of experiments in which they measured priming shortly after study and 1 week after study. These experiments consistently indicated that the repetition effect declined quickly in the first few minutes after training, and that there was relatively little further loss between a few minutes and a week after training. Salasoo, Shiffrin, and Feustel (1985) used a different technique, extended practice, to look at priming over a longer interval. In a series of word identification experiments, they found that repetition priming could still be detected 1 year after training. They also found that repetition priming was cumulative; identification thresholds decreased continuously over 30 presentations.These two studies suggest further questions. Although Salasoo et al. (1985) showed that priming increased with Correspondence should be addressed to S. C. Grant, Life Sciences Division, Scarborough Campus, University of Toronto, 1265 Military Trail, Scarborough, ON, Canada MIC IA4 (e-mail: grant@lake.scar. utoronto.ca).increased numbers of repetitions and that the priming could be retained for a year, they did not explore the loss of priming as a functionof the number of repetitions. After examining performance at two retention intervals, Sloman et al. (1988) concluded that the loss of priming slowed after the first few minutes following training. They compared the c...

Abstractionist versus episodic theories of repetition priming and word identification

“…That is, there was no "study list." These experiments consistently found more priming for pseudowords than for words when the training session data were collapsed over number of repetitions (Grant & Logan, 1993;Logan, 1988Logan, , 1990. Also, additional repetitions continued to produce improvements in performance on both words and pseudowords before asymptoting at 8-12 presentations.…”

“…In most of Logan's (1988Logan's ( , 1990Grant & Logan, 1993) experiments, the subjects' first exposure to an item occurred in the same task that was used for subsequent presentations. Yet, in many repetition priming experiments, the subjects initially saw the items in a study list, then performed a different task with the items on their second presentation.…”

Long-term aftereffects of response inhibition: Memory retrieval, task goals, and cognitive control.

“…The ANOVAs revealed that first-presentation control RTs (750 ms; the dotted line in Figure 2, left panel) were longer than NS NS RTs , F(1, 23) (Grant & Logan, 1993). The repetition benefit on NS SI trials also suggests stimuli can be associated with interpretations even if the response is not executed (Logan, 1985).…”

“…For all trial types, memory-retrieval effects are assumed to be strongest at Lag 1 and weaker but still significant at longer lags (Grant & Logan, 1993), whereas temporary bindings should only affect performance at Lag 1. The memory-retrieval hypothesis does not make predictions about error percentages.…”

Task demands and representation in long-term repetition priming

“…Lexical decision tasks are among the most commonly used paradigms for exploring repetition priming in general and LTRP in particular (e.g., Grant & Logan, 1993;Hintzman, 1976;Kersteen-Tucker, 1991;Logan, 1988; This research was supported by a Faculty Start-Up Grant from Colgate University. Bill Bacon, Bill Badecker, Richard Braaten, Alfonso Caramazza, Howard Egeth, Anne Hillstrom, Molly Treadway Johnson, Ho-wan Kwak, Gordon Logan, 1.Toby Mordkoff, Andrew Olson, Brenda Rapp, Myra O. Smith, and Steve Yantis all provided valuable feedback on early versions of the logic, design, and interpretation of the data.…”