I am a functionalist1 with a biologist's "bias and with "conscious awareness of other disciplines. My approach to memory research is guided by the following tenets: (a) the function of the brain is to create and retain "internal representations of the world that could guide behaviour; (b) the function of "learning is to permit the adaptation of internal representations to a changing world ("memory is the retention of these adaptations over time); (c) learning and memory require neural "plasticity for their actualization; and (d) learning and memory are "system properties, made possible by the concerted operation of multiple "levels of the system.
The aforementioned tenets yield two important consequences for memory research. First, the comprehensive investigation of the processes and mechanisms of biological learning and memory requires a multilevel approach. Second, in the analysis of learning and memory, two levels of functional organization are particularly critical. One is the behavioural level. It does not make sense to address the function of the system without addressing its input-output relationships. The other is the level in which the specific content (semantics) of internal representations emerges in the brain. Identification of the behavioural level is self-evident. Identification of the level that encodes internal representations is not. It is currently believed that the level critical for encoding the semantics of internal representations in the brain is the circuit level, or the cellular-and-circuit level. More reduced levels implement plasticity, but in the absence of the circuit "context, do not suffice to endow the representation with its semantics. It is essential, therefore, that research programmes on memory never lose sight of the circuit and the behavioural levels. This is not easy. The circuit level is often excessively complex, the behavioural level amazingly tricky. Furthermore, the remarkable success of molecular neurobiology is enticing. I thus believe in a focused, restrained "reductionistic approach to memory research (Dudai 1992). I hope that this is aptly reflected in the entries throughout this book.
Each entry opens with a definition, or a set of definitions. What a definition is, is extremely difficult to define. A liberal list contains no less than 18 different species of definitions, and multiple candidate definitions of the definitions in each species (Robinson 1954). Whenever possible, I tried to adhere to one of the following meanings of definition: (a) the minimal set of attributes that uniquely describes an item or a concept; and (b) the formulation of a thing in terms of a more elementary level of organization or theory. These meanings are not mutually exclusive, and reflect, respectively, an attempt to adhere to "Ockham's razor, and the basic reductionistic approach, which has been restrained above. It is evident, however, at the outset, that each of these types of definitions requires quite a lot of "a priori knowledge about the item to be defined.2 In the case of many items and concepts in the field of memory research, the relevant knowledge is yet unavailable. I had, therefore, to use an additional type of definition: explanation of the meaning of the term as it is to be used (stipulated definition). And as terms in memory research are occasionally used in more than one way, I provided multiple definitions when appropriate. The difficulties and uncertainties involved in definitions bring to mind the view that attempts to define entities at the cutting edge of knowledge could cause more harm than good: 'For when we define, we seem in danger of circumscribing nature within the bounds of our own notions' (Burke 1757). There is, however, the opposite view, that the risk is well worth taking. Socrates leads Meno to admit that definitions are always a must for a fertile, constructive dialogue (Plato, Meno 79d; "culture). In this debate, while being aware of Burke's caveat, I am much in favour of Meno's conviction.
Each entry ends with a short string of "associations. Bodies of knowledge in general are associative systems. I tried to "reinforce this notion by proposing selected associations. The reader is invited to form additional ones. Associations are not only aids to understanding, they are also proven "mnemonic devices: the richer the associative network, the higher the probability that the item will be stored ("metaphor) and "retrieved.
'Functionalism in its broadest sense is any view that analyses something in terms of how it functions (Lacey 1996).There are several versions of functionalism, one of which is 'functional analysis' (Cummins 1975). This is the research strategy that relies on the decomposition of a *system into its component parts while attempting to explain the working of the system in terms of the capacities of the parts and the way they are integrated with each other (Block 1980). Still, the structure of the parts and of the integrative system matters solely as much as it implements or shapes the function. Functional analysis is the sense of functionalism implied here.
2On this difficulty, which is also called 'the problem of the criterion', see 'criterion.
Was this article helpful?