Principles of Memory

Aimée M. Surprenant, Ian Neath

• Publisher: Psychology Press
• Year: 2013
• Print ISBN: 9781138882928
• eBook ISBN: 9781136950636
• Edition: 1
• Topic: Psychology
• Subtopic: Cognitive Psychology & Cognition
• Index: Psychology

In over 100 years of scientific research on human memory, and nearly 50 years after the so-called cognitive revolution, we have nothing that really constitutes a widely accepted and frequently cited law of memory, and perhaps only one generally accepted principle. The purpose of this monograph is to begin to rectify this situation by proposing 7 principles of human memory that apply to all memory. These principles are qualitative statements of empirical regularities that can serve as intermediary explanations and which follow from viewing memory as a function. They apply to all types of information, to all memory systems, and to all time scales. The principles highlight important gaps in our knowledge, challenge existing organizational views of memory, and suggest important new lines of research.

This volume is intended for people in the field of memory (from advanced undergraduates to seasoned researchers), although it will be of interest to those who would like a comprehensive overview of the fundamental regularities in cognitive functioning.

1 CHAPTER

Introduction

[These] are my principles, and if you don't like them … well, I have others.

—Groucho Marx

1.1 Principles of memory

In 1976, Robert G. Crowder published a highly regarded book called Principles of Learning and Memory. Among the many reviews, critiques, and commentaries subsequently printed was one by Cohen (1985, p. 248), who noted that “Crowder's (1976) Principles of Learning and Memory lists no principles.” Bob responded in his usual inimitable style: “I was once criticized for writing a book called Principles of Learning and Memory without ever really coming out and saying what those principles were. … Looking back, I must say Cohen had a point” (Crowder, 1993, p. 146).¹

The purpose of this monograph is to propose seven principles of human memory that apply to all memory regardless of the type of information, type of processing, hypothetical system supporting memory, or time scale. Although these principles focus on the invariants and empirical regularities of memory, the reader should be forewarned that they are qualitative rather than quantitative, more like regularities in biology than principles of geometry.

1.2 Laws, Principles, and effects

If the goal of any 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?

A first attempt at delineating the laws and principles of memory might involve a search of the literature. Teigen (2002) performed such a PsycLit search of journal abstracts for “law” citations and found a total of 3,093 between 1900 and 1999. The year with the most citations was 1927 (with 62). By 1999, there were so few mentions of the laws of psychology that one “must read around 1,000 journal abstracts before encountering a single law” (p. 109). None of the most commonly cited laws concern memory, and most of the laws were first proposed more than 50 years ago. In fact, the law that is cited most often (Weber's law, with 336 citations) was proposed in 1834. Very few new laws have been put forward, and of the few that have been suggested, almost all concern psychophysical phenomena of one kind or another (e.g., Shepard, 1987; Chater & Vitányi, 2003).

The few possible candidates for laws of memory that have been proposed have been proffered with great diffidence, if not outright reluctance. For example, Tulving and Wiseman (1975) describe a relationship between recall and recognition that suggests that the two measures are uncorrelated. Many papers refer to this description as the Tulving-Wiseman law, and indeed, it is the most cited memory law between 1990 and 1999 according to Teigen's (2002) search, with 10 citations. However, Tulving himself is ambivalent, sometimes referring to it as a function (Tulving & Flexser, 1992) and sometimes a law (Nilsson, Law, & Tulving, 1988). Similarly, the ratio rule (Glenberg, Bradley, Kraus, & Renzaglia, 1983; see also Bjork & Whitten, 1974) relates the size of the recency effect observed in free recall to the amount of time an item has to be remembered until recall and the amount of time that separates the items in the list. Despite its empirical regularity (e.g., Nairne, Neath, Serra, & Byun, 1997), it remains a rule, not a law.² As a third example, Watkins (2001, p. 195) discusses the gentle law of speech ascendancy, but he characterizes it as closer in spirit to “please keep off the grass” as opposed to “thou shalt not kill.”

Cohen (1985) proposed four laws, but referred to them as “potential” rather than established laws. For Cohen, “a memory law should be a statement about some empirical relationship in memory, which has transsituational generality and which looks like it will be around for some time to come” (Cohen, 1985, p. 252). Briefly, these laws are as follows: (a) the better something is learned, the greater the likelihood it will be remembered; (b) the longer something has to be retained in memory, the less the likelihood that it will be remembered; (c) the likelihood of remembering something depends on the nature of the memory test, later reformulated as “the closer the match between encoding and retrieval conditions (the greater the overlap in episodic contexts) the better the memory performance)”; and (d) individuals differ reliably in their memory capacities. Nyberg (1995) refers to five laws of memory, adding the Tulving-Wiseman law to Cohen's four laws.³

We performed a search comparable to Teigen's (2002) for the term “principle.”⁴ The most commonly cited principle was the encoding specificity principle (Tulving & Thomson, 1973) with 33 citations, but the next most frequently cited principle had only 2 citations (8 different principles were each cited twice). Between 1990 and 1999, only 22 abstracts contained named principles.

In contrast, a search for “effect” shows a different pattern. We found multiple citations to the acoustic confusion effect, the bizarre imagery effect, the changing distractor effect, the concreteness effect, the concurrent articulation effect, the cue depreciation effect, the fan effect, the for-got-it-all-along effect, the generation effect, the Hebb effect, the irrelevant speech effect, the isolation (von Restorff) effect, the lag recency effect, the levels of processing effect, the list length effect, the list strength effect, the mirror effect, the modality effect, the part-set cuing effect, the picture superiority effect, the prefix effect, the primacy effect, the Ranschburg effect, the recency (and long-term recency) effect, the revelation effect, the serial position effect, the spacing effect, the stimulus suffix effect (and the context-dependent and preterminal suffix effects), the word frequency effect, the word length effect, and many, many others.

Thus, in over 100 years of scientific research on memory, and nearly 50 years after the so-called cognitive revolution, we have nothing that really constitutes a widely accepted and frequently cited law of memory, and perhaps only one generally accepted principle.⁵ However, there are a plethora of effects, many of which have extensive literatures and hundreds of published empirical demonstrations.

One reason for the lack of general laws and principles of memory might be that none exists. Tulving (1985a, p. 385), for example, has argued that “no profound generalizations can be made about memory as a whole,” because memory comprises many different systems and each system operates according to different principles. One can make “general statements about particular kinds of memory,” but one cannot make statements that would apply to all types of memory. One reason we have reservations about the multiple systems approach to the study of human memory is this predisposition to dismiss the attempt to formulate general principles (see Chapter 2 for an overview of the systems view). We consider our ability to formulate seven generalizations about memory as a whole as further evidence against such a view.⁶

Roediger (2008) also argues that no general principles of memory exist, but his reasoning and arguments are quite different. He reintroduces Jenkins' (1979) tetrahedral model of memory, which views all memory experiments as comprising four factors: encoding conditions, retrieval conditions, subject variables, and events (materials and tasks). Using the tetrahedral model as a starting point, Roediger convincingly demonstrates that all of these variables can affect memory performance in different ways and that such complexity does not easily lend itself to a description using general principles. Because of the complexity of the interactions among these variables, Roediger suggests that “the most fundamental principle of learning and memory, perhaps its only sort of general law, is that in making any generalization about memory one must add that ‘it depends’” (p. 247). We fully agree with his arguments, up to a point: It is absolutely necessary to specify Jenkins' (1979) factors when applying and interpreting the principles. Where we differ is that we think it possible to produce general principles of memory that take into account these factors. For example, our encoding-retrieval principle explicitly takes into account the conditions at encoding and those at retrieval, several principles acknowledge the importance of materials and tasks, and although we do not have a principle specifically for subject variables, our specificity principle is at least a start.

Another reason for the lack of laws and principles may be that most of the regularities in memory are not unchanging enough to be given such a lofty status as law or principle. Given the complication that the human cognitive system must be dynamic in order to adapt to a changing world (Simon, 1990), laws and principles might not be possible. Some evidence consistent with this view might be that where laws and principles are proposed, they tend to be psychophysical in nature (e.g., Laming, 1985; Shepard, 2004), but as one moves from psychophysical tasks to ones that allow greater flexibility in how people might process information, there seem fewer and fewer laws and principles.

A fourth reason for the lack of principles might have to do with the Zeitgeist in which experiments are designed and articles or books are published. For example, Teigen (2002, p. 116) noted that “the emergence of the standard experiment in psychological research, based on Fisherian principles for null hypothesis testing, coincides with a steep reduction in law citations.” Designs that allow easy analysis with t-tests and ANOVAs tend to emphasize a search for significant effects rather than a search for functional relationships. In addition, the designs focus on the construction of binary oppositions, the results of which do not lead to general theories of cognition (Newell, 1973).

A fifth reason for the lack of principles might be that memory researchers, similar to investigators in many other areas of psychology, often become fixated on a particular procedure (see, e.g., the discussion of this problem by Craik, 1985). Rather than trying to discover how memory works, they focus on explaining a particular task. The concentration on just one task (e.g., immediate serial recall) yields enormous amounts of information and understanding of the task, but may not offer any real connections to memory in particular or scientific psychology in general.

We disagree with views that suggest that “the most fruitful way to extend our understanding of human memory is not to search for broader generalizations and ‘principles’” (Baddeley, 1978, p. 150). Rather, we answer Murdock's (1977, p. 332) question, “Have we not reached the point (at least in some areas) at which specifc experiments might yield general principles?” with an emphatic “Yes!”

We note, in particular, two advantages of searching for principles. First, it forces researchers and theorists to step back and look at the accomplishments of the field as a whole rather than concentrating on single theories, specific effects, or particular tasks. Second, as Greene (2007, p. 54) notes, “some of the most important contributions made by general principles have been those times when they spurred investigations into how they fail.”

We take the position here that general memory principles can be found and that we can take disparate-looking tasks and results and “put them all together” (Newell, 1973, p. 291). We also believe that there exist invariants in human behavior, at least at an abstract and qualitative level (Simon, 1990). We think it is time to assess the accumulated knowledge on memory and how it works and organize it in a fashion that allows us to “tighten the inferential web that ties our experimental studies together” (Newell, 1973, p. 306).

1.3 What Is a Principle?

As used in this monograph, a principle differs from a law in that it is more tentative, closer to a working hypothesis than to an incontrovertible statement. Nonetheless, our principles share much in common with most definitions of psychological laws (e.g., Cohen, 1985, p. 252; Teigen, 2002, p. 105). They must be empirically true and must be universal. In addition, we also require that principles have at least one (and hopefully more than one) of the following three properties: (a) they should state an important empirical regularity, (b) they should be able to serve as an intermediate explanation, and (c) they should offer useful information on how memory works.

One important function for a principle is to state an empirical regularity. As Toth and Hunt (1999, p. 254) put it, “One might imagine an outside observer … [inquiring] into the goal of our theoretical debates: ‘What are the fundamental empirical regularities of memory?’” To answer this question, one would (hopefully) want to...

Table of contents

• Cover
• Half Title
• ESSAYS IN COGNITIVE PSYCHOLOGY
• Full Title
• Copyright
• CONTENTS
• ACKNOWLEDGMENTS
• Chapter 1 Introduction
• Chapter 2 Systems or Process?
• Chapter 3 The Cue-Driven Principle
• Chapter 4 The Encoding-Retrieval Principle
• Chapter 5 The Cue Overload Principle
• Chapter 6 The Reconstruction Principle
• Chapter 7 The Impurity Principle
• Chapter 8 The Relative Distinctiveness Principle
• Chapter 9 The Specificity Principle
• Chapter 10 Evaluation, Limitations, and Implications
• References
• Author Index
• Subject Index