Interview Interview with Endel Tulving As to the second part of your question, a theory of something is an explanation of that something, and to explain something means different things to different Endel Tulving is University Professor at the University of people. Some think that memory will be explained4 Toronto. He was born in 1927 in Estonia, and came to even fear that some might claim it will be understood- Canada in 1949. He did his undergraduate work at the when the underlying synaptic mechanisms have been University of Toronto and graduate work at Harvard. He identified. There are others who think that a computer has been teaching in the Department of Psychology at program, or a mathematical model, can explain memory, Toronto since 1956, with a four-year stint at Yale in the or at least some of its phenomena. Between these ex- early 1970s. His research has been concerned with hu- tremes are other, potentially more fruitful, basic orient- Downloaded from http://mitprc.silverchair.com/jocn/article-pdf/3/1/89/1755748/jocn.1991.3.1.89.pdf by guest on 18 May 2021 man memory. His discoveries and the concepts he has ing attitudes. So, given these differences, you will get introduced to the field include subjective organization, different suggestions as to what the theory must contain input and output interference, the distinction between to cope with the known process. availability and accessibility of stored information, re- JOCN: Well, we are asking you. In particular, must a trieval cue, cue-dependent forgetting, Tulving-Colotla cognitive theory about memory that would please you measure of primary memory, recognition failure of re- be stated in a way that could be tested by brain scientists? callable words and names, the reduction method of de- ET: Sure! But an even better idea might be to demand termining trace structures, Tulving-Wiseman function of that a cognitive theory be stated in a way that the Almighty the relation between recognition and cued recall, encod- himself could pass judgment on. The point is that any ing specificity, synergistic ecphory, encodingketrieval in- interesting cognitive theory about memory (or a cogni- teraction, perceptual and ecphoric similarity relations tive theory about some interesting phenomenon of mem- in recognition, the distinction between episodic and orytthat is, behavior or experience that can be classified semantic memory, the distinction between noetic and under the rubric-is utterly beyond the pale of most autonoetic consciousness, stochastic independence be- methods and techniques of today’s brain science. Things tween perceptual priming and explicit memory mea- may change tomorrow, of course, and then we may want sures, and the quasimemory system of perceptual to include physiological or even physical reality as a priming. criterion for evaluating cognitive theories, but right now JOCN: Let us begin at the macrolevel, human memory. the insistence (you said “must it be?”) on it would bring What does it mean to have a theory of human memory? about a quick demise of cognitive theories of memory. Or, put differently, what must a theory of human memory Look, we have great difficulties making psychological contain for it to reflect the known complexity of the sense of many things we observe about memory, that is, process? just making up plausible, reasonably economical, and ET: I trust that you are talking about theories about internally consistent stories about those phenomena that human memory, rather than theories of memory as such. have caught the intellectual fancy of a particular gener- A theory of memory would be something like a theory ation of practitioners. If we had to start worrying about of light, or a theory of evolution, that tells you what it whether a favorite theory of ours is really true, that is, (light, evolution) is, and how its phenomena must be how Mother Nature planned it all, or if we began ex- what they are because of what the theory says they should pecting that people in some other branch of the science be, or how it is perfectly sensible that they are what they of memory do so, we and they would probably freeze are. There have not been any theories of memory of this in thought instantly. kind, and it is a reasonably safe bet that there never will JOCN: When people commonly think of memories, of be any. recalling past experience, they imagine the information Now, theories about memory are concerned with se- is somehow recorded in a particular site in the brain. Is lected, restricted sets of phenomena of memory; they there anything wrong with that folk notion, the same thereby escape facing the problem of complexity of notion most brain scientists believe to be true? Or jump- memory as such. In cognitive psychology, these “local” ing ahead, what are, say the five facts of memory that theories represent, as you well know, variations on the brain scientists ought to be considering as they pursue general theme of cognition as information processing. the physical dimensions of this problem? So, memory is thought of as consisting of the processes ET: Let me just answer the first part of the question. We of encoding, storing, and retrieving information, and will see later whether we get around to the second part. each of these processes consists of subprocesses. Phe- There is nothing wrong in principle with the idea that nomena of memory are explained in terms of the char- the information that is necessary for remembering some- acteristics of these processes, and their interactions. thing is recorded in a particular site in the brain. It is 0 I99 I Massachusetts Institute of Technology Journal of Cognitive Neuroscience Volume 3, Number I Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/jocn.1991.3.1.89 by guest on 25 September 2021 almost a tautology. When an event occurs that a person age and retrieval that makes the distinction so crucial as perceives and subsequently remembers, some changes you seem to be implying? must occur in the brain. That is, the brain is different ET: Cognitive psychologists “discovered” retrieval and before and after that event, or experience, or whatever. figured out how to separate it analytically and experi- We can call that before-after difference the engram, or mentally from storage in the 1960s. The predecessor of the memory trace, or the representation of the event, or memory research in experimental psychology was “ver- whatever. The engram, by definition, must be localizable bal learning,” the study of learning and retention of somewhere in the brain. This is why the idea of the verbal materials. Its pretheoretical thinking was greatly engram was scientifically respectable long before Lashley inspired by conditioning. It, too, suffered from the preoc- began his famous, and eventually inconclusive, search cupation with storage, although the concept was then for it, and why it is still so. labeled “association.” The interesting thing is that stu- JOCN: Why did you say in p?+mj%e nothing is wrong dents of verbal learning were unaware of their “storage with the idea of physical storage? bias,” for the simple reason that the distinction between ET: I said “nothing wrong in principle,” because in prac- storage and retrieval had not yet been made: You cannot Downloaded from http://mitprc.silverchair.com/jocn/article-pdf/3/1/89/1755748/jocn.1991.3.1.89.pdf by guest on 18 May 2021 tice there does exist a problem, quite apart from the be aware of something that does not exist. I should complexity of it all. The concept of engram, however it mention parenthetically that for the first six years or so is labeled, has mesmerized many brain scientists into of my post-Ph.D. life I too was one of those happy verbal acting as if there was nothing more to the problem of learners who did not lose any sleep over the storage- memory and the brain than the engram and its charac- retrieval distinction. When things changed, the field of teristics, including its location in the overall structure. verbal learning essentially died. Many of these brain scientists, like many common peo- It would not be difficult to argue that the “discovery” ple, do not seem to realize that the engram is an un- of retrieval processes permanently revolutionized the finished thought about memory, that it is at best only field of memory research in cognitive psychology. And one-half the story of memory. This being so, when they yet, as I see things, that revolution has not yet reached concentrate on the one-half and ignore the equally es- brain scientists. I have seen little evidence that retrieval sential other half, they may be doing the right thing, or processes occupy their thoughts or shape their activities. they may be doing the wrong thing. It is difficult to tell JOCN: Brain scientists ignore the larger question be- in advance. I wish someone would tell me whether they cause they do not have any idea how to study the issue are unaware of the other half, whether they are preoc- from a “neural systems” point of view. The synapse is cupied with the engram out of sheer inertia, or whether where the light is shining, which is to say it is something they have reflected deeply on the matter and deliberately that can be studied. Also, at a superficial level, it makes decided that identification of storage sites of memories sense that information storage ought to reflect structural is the number one priority, for such and such compelling and physiological changes at the synapse. Yet, no one reasons. thinks memories are stored at synapse, that is, if you JOCN By the “other half” you presumably mean re- push them on it.