Academic journal article Canadian Journal of Experimental Psychology

A Computational Account of the Production Effect: Still Playing Twenty Questions with Nature

Academic journal article Canadian Journal of Experimental Psychology

A Computational Account of the Production Effect: Still Playing Twenty Questions with Nature

Article excerpt

(ProQuest: ... denotes formulae omitted.)

Psychology often uses a divide-and-conquer strategy to understand how people learn and remember. For example, the standard multiple systems view assumes different systems for explicit and implicit memory with a further subdivision into episodic, semantic, procedural, priming, classical conditioning, and nonassociative learning (Squire, 1994, 2004). Although the strategy offers the methodological convenience of allowing researchers to address one function of memory at a time, the divide-and-conquer strategy risks leaving us with a fractured perspective: a view of the trees for lack of the forest.

An alternative view is that memory is a single system capable of producing complex and even perplexing patterns when faced with different test scenarios and materials. For example, Jamieson and Mewhort (2009a, 2009b, 2010, 2011; see also Jamieson, Holmes, & Mewhort, 2010), Higham, Vokey, and Pritchard (2000), Kinder and Shanks (2001, 2003), Nosofsky and Zaki (1998), and Benjamin (2010) have all argued that implicit and explicit learning can be understood using a single set of principles and mechanisms to handle phenomena traditionally distinguished as being implicit versus explicit. The number of memory systems needed is a fundamental issue. If we need to develop a different account for each phenomenon, scientific psychology is a futile discipline. Of course, we are not the first to note the problem.

Surprenant and Neath (2009) recently called for a critical reappraisal of theories of memory. To frame their argument, they asked "If the goal of science is to identify invariants and regularities within a particular domain (Russell, 1931; Simon, 1990), one might ask, what are the laws and principles of human memory?" (p. 2). They then pointed out that whereas 100 years of psychological research has produced an ample database of empirical effects and demonstrations, the field has failed to develop a unified explanation of those effects and demonstrations. Based on the failure, they argued that psychology should orient away from growing the already overwhelming database and focus instead on developing a coherent theoretical framework that identifies and articulates key principles and laws of human behavior. But Surprenant and Neath's criticism follows from a more classic example.

In his cri de coeur titled "You Cannot Play 20 Questions With Nature and Win," Newell (1973) pointed out that psychology had become seduced into playing an empirical game- one that he likened to playing the parlor game of 20 questions. Researchers pose a binary question such as, "Is the memorial benefit of production due to distinctiveness or strength" and then resolve the opposition by experimental analysis. Having solved that one bit's worth of uncertainty, the true state of nature becomes more certain. At first blush, this strategy is entirely rational. But, Newell argued that the strategy does not work and forecasted that in 30 years (i.e., 2003), our discipline would amount to a database of experimental demonstrations without a coherent explanation. In place of the 20 questions approach, Newell proposed that psychology should shift from its goal of empirical demonstration to refocus its effort on the goal of developing a coherent and general explanation of behavior.

Newell's (1973) advice has been largely ignored: Psychology has continued to play 20 questions with nature. Unfortunately, as Newell warned, psychology in 2016 has grown its database of demonstrations but has not tried to develop a unified account of human behavior. How, then, should we go about developing the kind of unified account of memory and cognition that Newell envisaged?

Consider the topic of this special issue: the production effect. People remember words that they have read aloud better than words that they have not. Although a memorial advantage for produced over unproduced words has been known for some time (e. …

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


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.