PubMed2006
In the present paper we reflect on some critically important issues in theory construction from the point of view of a practicing scientist. The starting point is to suggest the need for a minimal base of common agreement on the role of successfully working theories. It is proposed that scientific knowledge is not composed of singular facts but rather of relational structures connecting facts. Useful theories are both receptive and productive. Theories provide models, i.e., idealised representations of reality, expressed, in their most developed phases, in a mathematically formalised language. We further focus on the notions of rationality and universality, and show that these are mutually related and actually inseparable. Universality means description of observable phenomena in terms of universally valid laws that are essentially of a rational character, i.e., stated in terms of relational invariants preserved in variant, contingent conditions. Law-like components of a theory are universal by definition, not given by circumstances, and rational by their form, not by their content. Facts, on the other hand, are irrational elements unless they can be derived from law-like relations of another theory. Relational definition of rationality is self-consistent and independent from vaguely defined notions like 'reason'. Pertinent to studies of human nature, including psychophysiology, is the problem of individuality. To reconcile the claim of universality with an adequate account of unique individuality, we advocate a 'distributed nomothesis', distinguishing first-order laws ruling in an individual 'idioversum', from the higher-order, universal laws. Idioversal laws play the role of 'facts' in construction of universal theories.