The Organization of LTS

Tripartite Division of LTS into Episodic, Semantic, and Procedural Memory

In any individual, long-term store (LTS) contains a vast amount of information, and to be effective, it must be organized. The question of organization can be approached in many ways. Psycholinguists, for example, may be concerned with how we represent our knowledge of words and their meanings. Our initial approach will be more general, and will be based on the proposal that information in LTS takes one of three basic forms.

In 1972 Tulving put forward a conceptual view of memory which proposed that memories in LTS could be classified as either episodic or semantic.

Episodic memory, a term coined by Tulving (1972), is concerned with the storage and retrieval of personally experienced events or episodes. Recall of an episodic memory Opens in new window is usually reproductive—the event is recalled as it was experienced.

Semantic memory is our store of general knowledge about the world: concepts, rules and language. According to Tulving (1972) semantic memory Opens in new window is based on one’s stable, long-term knowledge of language and how to use it.

More recently Tulving (1985), among others (e.g. Anderson, 1995), has argued for an additional form of memory, a form that has been termed procedural memory Opens in new window. This can be defined as information in LTS that cannot be inspected consciously. Riding a bicycle, for example, is a complex skill that most of us acquire; but if we try to explain how we do it, we can give only the most superficial explanation.

Similarly, native speakers of a language can usually give no account of the complicated grammatical rules that enable them to produce correct utterances. This type of memory is contrasted with episodic and semantic memory, both of which can be inspected consciously, and their contents described fully to another individual.

Although it is necessary to distinguish them conceptually, episodic Opens in new window, semantic Opens in new window and procedural memory Opens in new window represent a highly interactive system, and, at any one time, the behavior of a normal individual may be directed by information from one or more of these sources.

In addition, the contribution of these memories to determining behavior may change across time. The formation of new semantic memories may depend initially, for example, on information from episodic memory.

Consider the problem of learning computer terminology. To newcomers, the jargon Opens in new window is wholly unfamiliar and difficult to assimilate. Thus the instruction to ‘boot’ the system is quite mysterious unless you remember that on a previous occasion it meant to start up the computer. However, with repeated use, the term becomes part of your general knowledge, and can be defined without recourse to episodic memory.

These kinds of interactions also occur while learning skills. Take learning to type, for example. At first, this involves remembering the layout of the keyboard in order to place your fingers correctly. However, as practice continues, the skill becomes increasingly automatic and independent of the ability to remember the keyboard’s layout.

At this stage, typing has ceased to rely on episodic and semantic memory, and has become incorporated in procedural memory. Indeed, when skilled typists are asked to recall the layout of the keyboard, they often find it difficult, remembering the location of some letters only by trying to type them and noticing where their finger is placed (Posner, 1973).

Tulving’s account of memory is essentially introspective, with each form of memory being associated with a different kind of conscious experience. We are not consciously aware of procedural memories; their existence is inferred from the fact that an organism responds in a consistent manner to a particular stimulus. Because of this, it is a form of memory that we can assume to be present in any organism capable of learning. For this reason, it is thought to be the most primitive kind of memory.

Semantic memory Opens in new window is open to conscious examination, because we can inwardly contemplate the features of the external world. Its contents are confirmed or altered in the light of new experience. For example, someone might believe initially that all cars run on petrol, but then s/he encounters a car that runs on diesel. This new observation then modifies that person’s knowledge about cars.

Episodic memory Opens in new window is associated with an additional and qualitatively distinct conscious awareness. It has a self-referential quality, enabling us to be aware of ourselves in the past and to imagine ourselves in the future.

The truth of episodic memories is determined entirely by their subjective familiarity, rather than by observation. When we recall personal experiences, we do not have external proof that our memories are correct; the manner in which they enter our consciousness seems to assure their authenticity.

Semantic and episodic memory are assumed to be higher and more recently evolved forms of memory, and are inextricably linked with consciousness. As yet, we cannot establish whether other animals experience consciousness, which makes it difficult to consider these memory systems in relation to species other than man.

Revised multi-store model showing tripartite division of LTS
Figure X | Revised multi-store model showing tripartite division of LTS

Many psychologists have proposed conceptual distinctions similar to Tulving’s, and there is agreement that a classification of this kind makes intuitive sense; but experimental evidence to support the distinctions is only beginning to emerge. At present, major support for this view comes from studies of the human amnesic syndrome. Figure X represents the revised view of human memory; the multistore model is retained, but with the addition of a tripartite division of LTS into episodic, semantic and procedural memory.

The division of LTS into three stores has been a popular theory, but there have been arguments against the distinction between episodic and semantic memory. It has been argued that a simpler distinction between procedural and declarative memory gives a better account of memory. Here a distinction is made solely in terms of whether information is consciously accessible or not. We have already seen that procedural memory Opens in new window is not consciously accessible; declarative memory Opens in new window is the converse: any permanent memory we can consciously describe irrespective of whether it is an autobiographical recollection or a piece of general knowledge.