Elements of a cybernetic epistemology: sequence learning systems

Abstract

Purpose

The purpose of this paper is to analyze how sequence learning can build on pattern‐recognition systems and how it can contribute to the behavioral options of goal‐oriented systems.

Design/methodology/approach

A functional approach is used to develop the necessary cybernetic structures of a subsystem for sequence learning, that can recognize patterns, register patterns occurring repeatedly and connect these to sequences. Based on that it is analyzed how goal‐oriented systems can use information about reoccurring sequences.

Findings

A subsystem for sequence learning basically requires pattern recognition and it needs a structure for the directed connection of single standards for pattern matching to standards for sequences, given that it can learn both new patterns and new sequences. Such a subsystem for sequence learning may recognize a certain pattern and with that the end of a certain sequence. So it may deliver more than one output signal at a point in time, and therefore needs additionally a subsystem for directing attention.

Practical implications

The paper analyses the principles of an “associative” way of connecting standards for pattern matching to standards for sequences. Also it shows the cybernetic necessity of an attention directing system that has to decide how to deal with the multiple outputs of a subsystem for sequence learning, i.e. to decide to act either towards a pattern or a whole sequence.

Originality/value

The paper investigates basic mechanisms of sequence learning and its contribution to goal‐oriented behavior. Also, it lays the base for an analysis of attention directing systems and anticipatory systems.

Citation

Nechansky, H. (2012), "Elements of a cybernetic epistemology: sequence learning systems", Kybernetes, Vol. 41 No. 1/2, pp. 157-176. https://doi.org/10.1108/03684921211213007