Rationality is Problem-solving, Critical and Social

The task of the following four chapters is to explain, on the one hand, how individuals may be the only actors, and rational only to a degree-thereby rejecting any functionalist theory of society-while still explaining how they are dependent on their institutional contexts for the exercise of rationality and, on the other hand, how their rational action in institutional contexts may be used to develop social scientific explanations. In order to carry out these tasks, we need a new theory of rationality which explains how institutions shape the rationality of individuals and how the partial rationality of individuals may be used to explain how institutions grow and change in often quite unintended ways as the consequences of their actions. This theory of rationality should take better advantage of Popper’s insights, that rationality is problem-solving, critical, and social. In order to achieve this end four problems need to be solved: (1) How do institutions steer events, even though only individuals have aims, pursue them, and thereby bring about change? (2) How are institutions and individuals related? (3) How can we use social structures to explain social events? And (4) How can we explain apparently irrational social events such as the popularity of totalitarianism in terms of the rationality of individuals? How can we, for example, explain the rise of totalitarian societies in terms of the quest to be rational? We may solve these problems with the theory that all social events are products of attempts by individuals to solve problems in institutional contexts. Problems individuals pose have structures, which are a product of the institutional context in which they are formulated. The degree of rationality found in attempts to solve problems is dependent on the quality of the institutions which shape the context in which problem-solving takes place. Institutions influence the degree of rationality of attempts to solve problems by determining how problems are formulated, which solutions are allowed, and the quality of criticism of proposed solutions. Individuals vary, of course, in their ability to make use of institutions. But this variance often tells us less about the relative degree of rationality of individual efforts than about the degree of rationality which institutions allow. The theory of rational action and the theory of the pursuit of truth may be unified, since both are problem-solving, critical and social. In both cases individuals try to solve problems in institutional contexts. Institutions may be studied as real, insofar as they set the conditions for problem-solving activity relative to specific problems set by individuals. Popper’s theory of the attraction of totalitarian societies may be

improved by deeming the attraction of such societies to be a product of a misguided attempt to preserve rationality by excessively fixing the social conditions for problem-solving, rather than a rejection of rationality in favor of the closed society. Revolutions are attempts to radically change the way institutions affect problemsolving activity and the rationality of religions consists in the frameworks they offer for problem-solving. This theory explains why social change is not merely a product of changes of belief. Individualist theories presume that changes of belief lead directly to social changes and that changing beliefs is easy. Neither presumption is true. Changes of belief do not lead directly to social change, because they are often not readily institutionalized, that is, taken as norms in the regulation of affairs. This change requires institutional processes of ordered change which are much more complex and difficult than a mere change of belief. Changing beliefs is also not easy, because such changes are products of problem-solving activity and this activity is controlled to a high degree by institutions. There are no free-floating thinkers, who do not change their beliefs by using existing institutional standards. Thus, new theories must partially meet old standards or they will not be adopted. The degree of rationality which any individual exercises is not merely a question of his integrity, intelligence and courage, but also of his institutional context. The same may be said of his autonomy, which requires a high level of rationality which can only be rendered possible by institutions. Revolutions are not, as on Weber and still too much on Popper’s view, a product of setting aside rationality, but exaggerated attempts to improve rationality by radically changing the social conditions for it. Totalitarian revolutions seek both to change and to fix social conditions for problem-solving. Religions seek to set and preserve problem-solving social structures. When Karl Popper tried to explain how social science was possible, he used his own discoveries in the theory of rationality to build on the results of, above all, Weber and Hayek. But he could not solve the fundamental problem of reconciling the rationality of individuals with social scientific explanation. His defense of the rationality principle, that is, that all social events should be explained as the results of the rational action of individuals, is ingenious. But it also leads to a theory which is so complex, that the defense he offers is quite dubious. He cannot avoid the fact that individuals are for the most part by no means so rational as this principle presumes. We may trace his difficulties back to his adoption of too much of the theory of rationality endorsed by Weber and Hayek. The overwhelming majority of contemporary social scientists today presume that individuals are rational only to the extent that they follow coherent plans to realize goals they have clearly set. This theory led such thinkers as Weber, Hayek and Popper, along with the vast majority of social scientists, to seek to satisfy the high demands of strict individualism with holistic remedies: the coherence sought in individual thought is interpreted as an ideal type, that is, the way individuals might think if they coherently portrayed ideas dominating their society. They presume that some group of individuals act in

some coherent way in the pursuit of some goal and ask what kind of system results, as Weber did in his explanation of capitalism as a product of Puritanism. This will not do, since societies do not consist of groups of individuals who are rational in this way. Societies institutionalize competing ideas and individuals are often confused because they adopt aspects of competing theories in ad hoc ways. In developing this view I have borrowed heavily from Joseph Agassi’s theory of rationality. He has proposed that rationality is partial. All individuals are rational to a degree, he says, and the central problem of rationality is not how to be comprehensively rational, but rather to raise each person’s level of rationality. He has also corrected Popper’s Puritan ethic, arguing that rationality is natural and no burden. I take over all of these views here. He has also greatly improved Popper’s philosophy of science by developing the theory of problems. He argues that they are shaped by metaphysical frameworks, so that frameworks and problems are integrated into a fallibilist theory of science and not simply allowed by it. But I still see a need to develop further the view that rationality is social and in this way to integrate the theory of rationality as the pursuit of truth with the theory of rational action. If we do this, improvements can be made in the theory of the social sciences and ethics. A change in Agassi’s theory of problems is needed, when problems are viewed as determined by social contexts rather than metaphysical ones, since social contexts lack the coherence of (good) metaphysical theories. Problems often reflect struggles to meet partial, confused, even contradictory, standards, rather than the coherent standards of some metaphysical theory. The theory of explanation in the social sciences proposed here lowers standards of rationality in order to achieve more. Rational thought does not need to be coherent or systematic in order to have an interesting degree of rationality and rational action does not need to be based on such coherent plans if it is to be understood. In the place of the quest for finding coherence and system in rational thought and action, this theory places the quest to understand problems and how they are socially determined. Further, rationality is found in the way individuals use the frameworks and tools they have available to pose problems, offer solutions to them, and to critically appraise them. Problems within the social sciences, then, need to be reformulated so as not to require coherent planning by individuals and/or coherent social systems. Some high standards of the social sciences have to be lowered. But new problems open up about, on the one hand, the degree of rationality institutions allow and, on the other hand, the degree of rationality of individuals. The methodological advantage this perspective offers is that it is realistic and offers hope of partial knowledge of social systems in the place of overambitious constructions which distort reality. How to overcome problems in Popper’s approach Although Popper rejects the view that rationality requires system, we can still find in his theory of the social sciences remnants of the traditional view of rationality

which block him from making progress. He presumes, namely, that we understand individual actions by finding the (coherent) plans which individuals form and use to pursue their ends. These plans and the actions which follow from them are then used as the basis for social scientific explanations. This seemingly trivial assumption prevents him from adequately taking into account how intellectual contexts effect the way individuals plan their actions. The assumption forces him to build all institutional factors into the beliefs of individuals. This is the only way they may be treated as posing some problem situation, given the traditional view, since only then can a coherent plan be formed by individuals as a response to it. But this means that the impact of institutions, regardless of the beliefs of individuals, cannot be explained. This remnant of the traditional view of rationality prevented Popper from developing a new theory of institutions and their relations to the rational actions of individuals. Because he treated institutional aspects of the logic of situations as the way individuals perceive institutions as they formulate their (coherent) plans. He has no theory of how institutions steer events regardless of the beliefs and intentions of individuals, even though he presumes that they do. Methodological individualism should explain why individuals act the way they do as a consequence of the logic of their situation. But it can only explain why they act the way they do on the basis of their perceptions of their situations. But the logic of a situation and the perceptions of it regularly diverge. The basis of the explanation in the situation is diluted if not lost. If we purge the remnants of traditional views of rationality from Popper’s view, we can improve it and make better use of the progress he made. There are three aspects of Popper’s theory of rationality in science which we may take advantage of. These aspects are (1) the proposal that rationality is social, (2)—generalizing on Selz’s thought psychology-the suggestion that rationality is problem-solving, and (3) the thesis that rationality is critical and does not justify. A more thorough use of these aspects of his view enables us to find a more powerful alternative to the rationality principle. It provides an improved theory of the roles institutions play and the methods of studying them, that is, an improved theory of how to explain how institutions steer events. Popper’s theory of the logic of the situation does not distinguish between institutional and non-institutional or relatively arbitrary factors, since they are all mere factors in an individual’s deliberations as to how he can achieve his ends. This is a serious hindrance to the formulation of social scientific theories, whose central task is to explain the role of institutions in steering social processes. It fails to distinguish those aspects of social situations which may be deemed typical and understandable-mainly institutions-from those which are not. It led Popper to the view that the aim of social science must be to construct models of social situations, which are as complete as possible, rather than to describe institutions and their impact as accurately as possible. But even here his theory is weaker than mmmmmmmmmmmmmmmmm

neoclassical economics which constructs models on the assumption that specific well-functioning economic systems exist. Having quite properly rejected Weber’s functionalist theory of ideal types and unwilling to endorse Hayek’s theory that (economic) systems result from rational action, Popper had nothing more to say about institutions than that individuals act in the context of ‘the logic of a situation’. These defects can be remedied by (1) analyzing the partial rationality of individuals-individuals do not normally possess fully coherent plans for the realization of their aims, but their problem-solving activity may nevertheless be understood with the aid of institutional analyses of problems structures, that is, by showing how various aspects of problems are embedded in social contexts, and by (2) institutional analyses of problem-structures, which do not presume the existence of even partial social systems and which are, therefore, applicable to open societies. Why problems need to be seen in their institutional context The basic tool for the construction of social scientific theories on the view proposed here is not the rationality principle, according to which individuals pursue aims, in accord with beliefs, given the logic of their situation, but the social analysis of problem-structures, according to which individuals pursue goals, which are to a high degree a product of their institutional contexts, and which they try to achieve by satisfying institutional standards. The rationality principle presumes that beliefs are coherent and problems well-formulated according to some standard. The portrayal, then, of the logic of the situation in accord with the rationality principle requires that individuals have a coherent view of their situation. But problemstructures often do not meet these high standards and individuals lack sufficient social analyses to sort this out. Yet, individuals must meet some standards or there would be no rationality at all. They often do so piecemeal, with no entirely coherent view of their situation, but nevertheless with ideas of aspects of it and even how they fit together to a degree. Given the rationality principle, this would make their actions more or less irrational and inexplicable to a high degree. But this is too quick, since they must have some degree of rationality. This degree of rationality may be found in their attempts to solve problems by meeting standards, even when these standards are neither clearly spelled out nor even coherent. In order to explain my view, then, it is useful to explain both (1) why those standards which are normally presumed in methodological individualist theories are too high for analyses not only of human action but also of science, and (2) that the rejection of these high standards as a basis for descriptions of problems individuals try to solve does not lead to the denial that individuals are acting with some degree of rationality. Even when such high standards are not met, we may still see the rationality of the problems individuals pose and their attempts to solve them. We comprehend this rationality, because we can discern how they are attempting to solve problems given some established standards. We can then explain social

events as products of efforts to meet these standards even when fully meeting them is impossible, because, say, some conflict exists between several standards. In order to explain my view of problem structures, it is useful in this section to review the reasons why the theory of problems has become central for the theory of rationality and why a new theory of them is still needed. It is this development which has laid the groundwork for a new approach to rationality. The theory of problems has become central, because difficulties stemming from both the use of logic and thought psychology have led to problems in explaining the growth of knowledge. In order to overcome these difficulties one has to turn to the theory of problems. Furthermore, existing theories of rationality, even problem-oriented ones, cannot describe rational practice adequately, because they set unrealistically high standards for the formulation of problems. They cannot describe the rationality of individuals who do not meet artificially high standards, and this is most people most of the time, whether they are planning actions or pursuing truth. So, a new approach is needed for both the philosophy of science and the theory of the methods of the social sciences. The English title of Popper’s first book should have read Logic of Research, if accuracy had been the prime criterion for the translation of the German text into English. This is a fine title for the task which Popper and many other philosophers of science set themselves at the beginning of the last century. The logic of research was thought to be an integral part of propositional logic or, perhaps, an inductive addition to it. The remarkable advances made in propositional logic in the first part of the 20th century raised hopes that a new and better theory of science could be formed with their aid. One lonely voice of protest came from R.G. Collingwood, who argued that a logic of problems was needed in order to take account of the fact that the meaning of words depended on context (Collingwood, 1939). Only when context was accounted for could one ensure that the substitution of one term for another in some argument would not invalidate the argument; only then could mistaken inferences be avoided. Collingwood was not listened to, however, so the futile search for an inductive logic dominated most efforts. Quine implicitly conceded that Collingwood’s criticism was correct with his observation that the analytic-synthetic distinction was not sharp. Many thinkers then acknowledged the problem, but still tried, as Quine did, to work around it with seemingly ad hoc measures (Wettersten 2002 a). Quite independent of these developments Popper had learned from Selz that directed thought processes-as Selz called them-could not proceed according to the principles of the dominant associationist psychology (Wettersten 1985, 1987, 1988m 1990, 1992). Selz argued that it was tasks-or problems-which guided thought processes. He further suggested that any theory of science must explain how that happens in scientific research. This was a new idea, since most philosophers of science based their inductivist methodologies on associationist psychology. But what is false in psychology could not be true in methodology. Popper, then, had already broken ranks with those who did not deem problems important, when he offered a theory of the logic of research. On his view the logic

of research consists entirely of the retransmission of the falsehood of the conclusion of any valid argument to the falsehood of at least one of the premises. This logic is, in fact, crucial and the core of scientific research, but it describes only a small part of the methodology of research. Popper added a methodology in order to make up the deficit of his logic of science alone (Wettersten 1987, 1992, Jarvie 2001). And in doing so he was led back to the importance of problems in scientific research. But his view of problems remained rather primitive, being, on the one hand, merely a response to refutations and, on the other hand, something subjective. Each starting point led back to the problem of explaining the role of problems in scientific research. The analysis of the logic of science with propositional logic to solve the problem of induction led to the problem of taking context into account when one set the meaning of terms or presumed that terms were synonymous. The description of the logic of science as the retransmission of the falsity of a valid argument to at least one its premises was guided by, and led back to, the observation that thought processes were guided by problems rather than associations or inductions. Various attempts to solve the ensuing problems have been pursued. Collingwood recommended that a logic of problems be developed to take account of context but never developed such a view. Popper emphasized the role of problems in scientific research due to the background of his view in the psychology of Otto Selz, but offered no theory of them. Quine, after coming to the same conclusion as Collingwood as he noted that the distinction between analytic and synthetic statements could depend on context, recommended that problems be posed so as not to call for great changes so that problems with meaning would not arise. Mario Bunge said we should first pose problems calling for small changes and only when then that fails should we move on to more radical changes. Joseph Agassi said problems were often a function of metaphysical frameworks and integrated this view into a fallibilist philosophy of science, following Popper’s lead. Jagdish Hattiangadi unsuccessfully tried to break out of the dilemma by claiming that all problems are contradictions and that frameworks are irrelevant (Wettersten 2002). With the exception of Hattiangadi, these views share the idea that problems are formulated with the aid of theoretical frameworks. Collingwood says that they depend on absolute propositions, Quine finds that they depend on language, Bunge on systems and Agassi on metaphysical frameworks. Those seeking a logic of science seek to analyze the presuppositions of problems and their relations to problems. I agree with these thinkers that frameworks are crucial for problems and for a theory of rationality.1 But they all share, either explicitly or implicitly, the view that well-formed problems are set by a definite set of coherent assumptions. Now it certainly is useful to note that assumptions made in the formulation of problems may not be coherent. Tools which help us do that are welcome. But in

1 I have analyzed Hattiangadi’s attempt to avoid this in detail elsewhere and have shown that it will not work (Wettersten 2002).