International Encyclopedia of Systems and Cybernetics

A registered trace of some former stimulus and its results, in a living or an artificial system…

or, the capacity to store perceived information for later retrieval, generally in a modified form.

According to K. KRIPPENDORFF, memories are observational constructs (1986, p.48).

However M. BUNGE writes that "… memory is the activation of neural connections", while "learning is the formation of neural connections" (1979, p.138).

It seems not yet very clear if these connections are merely a potential to be evoked or if they exist permanently.

The first definition, inspired by R.L. ACKOFF and F.E. EMERY (1972, p.76), "… does not require that the memory-response be a conscious one".

The same authors define the intensity of a memory as "… the intensity of the response that defines it"; the durability of the memory "as the length of time over which it persists", and the correctness of the memory as "the efficiency of the memory response for the objective for which it is intended" (p.77).

"Memory' has been a very controversial concept among systemists, cyberneticians and computer scientists.

H.von FOERSTER writes: "Let me demonstrate the notion that is computing, rather than signaling and storing these signals, which is at the core of cognitive processes by contemplating for a moment "memory". For it is precisely the misconception of this higher mental function to be a "data storage system" which blocks the vision for the kind of systems we shall need to meet the challenge of the future.

"If memory were a data storage system it is easy to show that in order to account for what we know, each of our brains should be the size of a sphere about one mile in diameter packed with nerve cells. However, when of this size, the operation to recognize, for instance, the presence of a lion in its field of vision takes this brain about ten years. This may be helpful for the lion but, alas, not for the bearer of this brain" (1982, p.216).

In synthesis, what our brain does is not going searching for one specific datum in a colossal heap of stored data. In von FOERSTER's terms: "… (it) computes the answer to a question, rather than store all answers together with all possible questions in order to respond with the answer… when it can find the question" (p.175).

According to von FOERSTER, we should view memory as a process property of the brain: "… as if it were a computer whose internal organization changes as a result of its interaction with an environment that possesses some order. The changes of the internal organization of the computer take place in such a way that some constraints in the environment which are responsible for its orderliness are mapped into the computer's structure. This homomorphism "environment-system" reveals itself as "memory" and permits the system to function as an inductive inference computer" (p.114).

Moreover, the main practical feature of cognitive processing in relation to memory is that "cognitive living organisms… succeed to eliminate quicky all temporal aspects in a sequence of events or, in other words,… "time" is abandoned as early as possible in the chain of cognitive processes" (1982, p.140) (i.e. the only way to save oneself from the "lion's jaws!").

von FOERSTER wrote this in 1965, when networks and connection machines had not yet entered into the main currents of cybernetics and systems science. Thus the use of the term "computer" should not be taken as meaning any identification of brains with classical sequential computers which process stored data according to strictly deterministic or even stochastic programs.