So that imagination and memory are but one thing, which for diverse considerations hath diverse names. Thomas Hobbes, Leviathan 1.2.
1. Introduction
On October 4th, 1992, a cargo plane from the Israeli airline El Al crashed into an apartment building in Amsterdam, exploded and left 43 people dead and several hundred injured and homeless. The event dominated the local news for many days. Ten months after the accident, a group of psychologists led by H.F.M Crombag distributed two questionnaires among a hundred Amsterdam residents. The first questionnaire asked residents whether they had seen the footage showing the plane crashing, and whether, based on their recollection of the video, they could estimate how much time elapsed between the plane crash and the explosion. Participants’ estimations varied, but 55% of them remembered having seen the footage. Only 18% reported not remembering the video at all. A second, modified, questionnaire was distributed to another group, asking the residents questions about specific details of the accident as captured by the video— for example, the angle at which the plane hit the building, the exact way it broke apart, etc. Although there were some disagreements among their answers, 66% of respondents reported remembering the video vividly (Crombag et al., 1996). Here is the rub: there
recording, not even computerized reconstructions of the accident. Most of the people surveyed simply misremembered.
Far from being an unusual result, evidence gathered over the last three decades of research in cognitive science clearly show that people are prone to misremembering past experiences (Breinerd & Reyna, 2005). So-called false memories are a systematic and common occurrence in our ordinary lives, and they present a challenge to thetraditional philosophical view of the function of memory. According to the traditional view, memory20 is for remembering, and remembering consists of reproducing the contents of
past experiences. When philosophers discuss the problem of false memories, they typically try to safeguard the traditional view following one of two strategies. The first strategy is to say that if a particular mental content appears to the subject to be a memory even though it does not correspond to an experienced event, then the subject is not really
20 Two terminological clarifications. (1) It is an unfortunate linguistic fact of the English language that the
word ‘memory’ is so polysemous. Consider the sentence “She has an extraordinary memory”. It could mean that she has a good memory-qua-cognitive-system—she may be able to store a lot of information, for instance—or it could mean that she has a memory-qua-mental-state whose content happens to be out of the ordinary. As much as possible I will try to disambiguate these senses, but for the most part, when I talk about memory, I refer to the cognitive system. (2) Philosophers and psychologists recognize several kinds of memory. What psychologists call ‘procedural’ or ‘non-declarative memory’, for instance, roughly corresponds to what Bergson (1908) and Russell (1921) called ‘habit memory’, and James (1890) called ‘secondary memory’. ‘Declarative’ or ‘non-procedural memory’, which psychologists operationalize as the kind of memory whose contents can be consciously declared, more or less corresponds to James’ notion of ‘primary memory’. Declarative memory, in turn, is usually divided in ‘semantic’ and ‘episodic memory’ (Tulving, 1983). Semantic memory refers to knowledge of facts and situations about the world that we need not have witnessed; when we recall semantic memories there is no need for mental imagery associated to the place and/or time in which the remembered event occurred. Some philosophers take semantic memory as tantamount to propositional memory—i.e., mental occurrences that can be expressed as ‘remembering that p’—but this is wrong. After all, there are non-semantic memories that can be expressed propositionally, just as there are semantic memories that can be expressed with a gerund. Categories based on surface grammar just don’t square with psychological classifications. Finally, episodic memory refers to memory of experienced events. It roughly corresponds to what some philosophers have called ‘recollective memory’, ‘personal memory’, ‘experiential memory’, or ‘direct memory’ (Furlong, 1948; Locke, 1971; Malcolm, 1963; Martin and Deutscher, 1966; Bernecker, 2010). Although there is some disagreement as to whether or not these terms define perfectly equivalent categories, I am going to leave that issue aside. I am going to refer to the kind of memory I’d like to discuss simply as ‘memory’, bearing in mind that the sort of mental experience I mean to talk about falls roughly within the psychologist’s definition of episodic autobiographical memory. Examples include memories about particular events in one’s childhood, this or
exercising her memory but rather her imagination. From this perspective, then, all those individuals that reported having seen footage of the plane crashing were not really remembering: they were merely imagining. After all, since the traditional view holds that one can only remember what happened, false memories—that is, memories that do not correspond to what actually happened—are simply not genuine memories. Accordingly, one can still say that memory is for reproducing past experiences, as cases of false memories are not the products of memory at all. The second strategy philosophers take to safeguard their view is to say that false memories are indeed the result of exercising memory—and when false memories occur, memory itself is malfunctioning. From this perspective, all the people that claimed to have seen the video of the accident were indeed exercising their memories, but they misremembered because their memory systems malfunctioned. As a result, the view that memory is for remembering remains unchallenged, as false memories are simply the product of a faulty memory, the actual function of which is to reproduce past experiences.
In this chapter I argue that the traditional view is mistaken. I contend that memory is not for remembering. There are three sections to this chapter. In the first section I review some critical findings from cognitive science and neuroscience suggesting that false memories are both normal and pervasive. I argue that trying to reconcile these results with the traditional view would have unwanted consequences. On one hand, I argue that if we follow the first strategy, and consider false memories to be the product of imagination, not memory, we may have to say that some bona fide cases of remembering are not produced by memory. On the other hand, I contend that if we follow the second strategy—that false memories are the product of memory malfunctioning—we may have
to accept the counterintuitive claim that it is advantageous to have a memory system that normally fails. In order to avoid these pitfalls, I claim that our basic assumptions about the function of memory need to be revised. I undertake this revision in the second section of this chapter, where I examine the way in which we tend to individuate cognitive functions. This analysis leads me to defend a strategy according to which the function of a cognitive faculty is determined by its contribution to the cognitive organism. In the third section, I apply this strategy to the case of memory, and I reject the idea that seeing memory as a cognitive system for reproducing past experiences is the best way of making sense of its function. Instead, I offer a picture of memory as an integral part of a larger system that supports not only thinking of what was the case and what potentially could be
the case, but also what could have been the case. More precisely, I claim that the function of memory is to permit the flexible recombination of perceptual components of memory traces into representations of possible past events that might or might not have occurred, in the service of constructing mental simulations of possible future events. I conclude by showing how this account can accommodate the evidence that is problematic for the traditional view, how it allows us to say that ordinary instances of false recollection are indeed produced by memory, and how it preserves the intuition that many ordinary cases of false recollection are the result of a memory system that is, in fact, functioning quite well.