Abstract

The article starts with the introduction of the Lithuanian reader into the expected utility theory which is the core of the decision theory known under the name of the rational choice theory or by the name of the thin theory of practical rationality too. This theroy conceives practical rationality as maximization of the expected utility defined as the measure of the preferences which satisfy the following conditions: (1) reflexivity; (2) completeness; (3) transitivity; (4) continuity; (5) preference increasing with probability and (6) strong independence. This theory is qualified as the formal or minimal theory of the practical rationality. This theory is unique having two ways of application: as the empirical theory and as the normative one. The main task of the article is the disentangling and discussing of the different and mainly divergent "thickenings" which this theory undergoes in its various applications. The name of "thickenings" is given to auxiliary assumptions or supplementary conditions supplementing the core axioms of the expected utility theory. The auxiliary assumptions current in the empirical applications of decision theory are those of the self-interested behavior, of the transpersonal homogeneity and of the transpersonal stability of preferences. While vindicated by the main methodological principle of the rational choice approach in the social sciences - to keep the models as simple as possible and to make them complex only as much as necessary - the assumptions abovementioned are counterintuitive if the rational choice theory is considered as normative theory. The main problem confronting a constructor of the thick (substantive) normative theory of rationality is the diversity and incoherence of the preanalytic intuitions concerning the content of the preferences deserving to be qualified as rational. The ethical theory capable to reach the reflective equilibrium with all those intuitions could pretend the title of the theory of the absolute practical rationality. As a matter of fact such theory is as yet nonexistent, the theorists of the thick rationality beeing divided into the consequentialists relying on the Humean instrumentalist concept of rationality explicated in the decision theory and the deontologists exploiting the Kantian universalist intuitions on rationality. The article discusses the idea of the "broad" concept of rationality which is intermediate between the minimalist (thin) or formal one and the maximalist (thick) or substantive ethical one, exposed in the work of Jon Elster, Amartya Sen, Albert O. Hirschman, Martin Hollis, Stefan Gosepath et all. The rational decision in this intermediate sense is named in the article (following the proposal of S. Gosepath) "relatively practically rational decision". Relatively practically rational is the choice of an agent whose preferences in addition to the standard 6 axioms of the theory of expected utility satisfy two supplementary ones: (7) preferences are realistic and (8) they are shaped in some specific way. This way can be charaterized in the negative terms or in the positive ones. Speaking in the negative terms, to be rational the preferences must not be the products of some subintentional mechanisms or to remain acceptable for the agent after she comes to know the causal story of their genesis. Speaking in the positive terms, the rationality of the preferences is warranted if they are chosen by the agent guided by the metapreferences satisfying conditions (1)-(3). The distinction between preferences and preferences proposed by Harry G. Frankfurt gives handy means to demarcate the bounds of practical rationality too: human agents cannot pretend to achieve the absolute practical rationality in their choices because they in their choices of the preferences (or ethical preferences) cannot rely on arguments having the validating power not coming up to that possessed by the arguments substantiating the choices between the empirical hypotheses. They can achieve the relative practical rationality in so far as the rational choice between preferences is a feasible one. The concept of relative pratical rationality is thick enough to assimilate one of the most powerful intuitions concerning rationality, characteristic for the modern Western culture and proclaimed as early as in 1486 by Pico della Mirandola in his famous "Oratio de hominis dignitate". In this intuition the human rationality and the human autonomy are indissolubly bounded. The concept of relative practical rationality preserves this intuition equating the rational and the autonomous choice and defining the rational choice as the consistent choice guided by the realistic preferences which are themselves chosen on the ground of the ethical preferences (metapreferences) directing the characters planning of an agent.