Abstract
Many thanks to Jamie Cohen-Cole and David Teira for their commentaries. They both raise very interesting challenges for the analysis presented in How Reason Almost Lost Its Mind – although these are challenges that push in somewhat opposite directions.
Cohen-Cole begins by emphasizing the contrast that we – the six authors of the book – have drawn between ‘Enlightenment Reason’ and ‘Cold War Rationality’ (CWR henceforth) in order to argue (a) that we assert that the latter ‘displaced’ the former, but that (b) CWR was in fact not ‘the defining feature of the human sciences’ in the USA of the post-WWII era. To support (b), he claims that there was a group of thinkers in these disciplines, such as Margaret Mead and David Riesman, alongside Joseph Weizenbaum, George Miller and Noam Chomsky, among others, who rejected attempts to explain human rationality in instrumental and computational terms. This, in Cohen-Cole’s view, raises issues about the continuities or discontinuities assumed in writing a history on this whole topic.
Teira, in contrast to Cohen-Cole, claims that CWR is far from being dead. Even after the end of the Cold War, philosophers or social scientists continue to use, for instance, probability theory or rational choice theory. He then raises the question of how to explain the dominance of expected utility maximization (a theory that uses, as Teira notes with a perhaps fair amount of irony, an Enlightenment concept, namely Bernoullian utility functions): can this dominance really be explained by appeal to a priori considerations, such as Dutch Book arguments (according to which you can be made to accept monetary bets that you will necessarily lose, unless you conform to axioms of probability and rational choice)? Or is its prevalence due to some ideological background or context? These are formidable challenges. I shall address them by first removing, piece by piece, certain misunderstandings and then counter the challenges more directly where necessary.
1. Cohen-Cole overlooks the fact that we use the term ‘CWR’ not simply in order to denote a kind of thinking, or a theory of rationality. Indeed, at various times individuals have sought to articulate an ideal unified theory of rationality that would allow for the analysis of problems into simple, sequential steps, and that would provide solutions that are formal, optimizing, algorithmic and mechanizable (i.e. implementable on computers) – these being the central features we associate with CWR (Erickson et al., 2013: 3 f.). However, the very attempt to develop that theory clearly and comprehensively led to numerous thorny problems: problems such as the famous paradoxes of rationality – Allais’ paradox, the Ellsberg paradox, Newcomb’s paradox, and so on – that every student of rational choice is well aware of. It was typically those who were themselves fully trained and engaged in the modes of reasoning characteristic of CWR who identified these problems in the first instance. They therefore initiated fundamental discussions about what rationality really is. Because of this, we explained that we use the term ‘CWR’ and its cognates also to refer to the debates over what an ideal theory of rationality should look like, what its potentials and limits are (ibid.: 4). It is this historical dynamic between conceptualizations of rationality and the discussion about them, that interests us.
2. With respect to kinds of thinking, or families of similar theories, Cohen-Cole claims that we maintain that CWR ‘displaced the Enlightenment understanding of the mind, which had emphasized judgment and practical reason’. In fact, it did never really ‘displace’ it, nor do we maintain that there was a unique ‘Enlightenment understanding of the mind’, let alone one that emphasized ‘judgment and practical reason’. To begin with, the views of Enlightenment authors about what the mind is were too varied and too complex to be summarized this way. Furthermore, Enlightenment views about reason or the rational, ‘higher’ mental faculties persisted throughout the 20th-century, although – and this is our real historical claim – they no longer always occupied centre stage. Insofar as they have returned, they did so in altered forms.
Behind Cohen-Cole’s misinterpretation is a more basic issue. ‘CWR’ is to be understood and used as an ideal type concept in the Weberian sense (Erickson et al., 2013: 3). So, we should not be surprised at the infrequency with which we meet perfect instantiations of the concept in reality. There are unclear cases, and overlapping assumptions will be found when one compares certain CWR authors with those who one might not view as belonging to that group. For this reason, Cohen-Cole’s reading that we claim Enlightenment Reason to have been ‘displaced’ by CWR is incorrect. Nonetheless, the concept of CWR highlights an identifiable historical trend, and helps to contrast different thinkers and their views on this approach. Teira grasps our position better when he notes that both Herbert Simon’s ‘bounded rationality’ and Charles Osgood’s GRIT strategy in conflict negotiation were products of Cold War contexts too, but showed critical distance from the pure ideal type of CWR – which one might find more in writings of formal logic or statistics or in von Neumann’s and Morgenstern’s influential theory of games and decisions. For instance, in his works on bounded rationality, Simon gave up optimizing aspirations, but nonetheless aimed at developing algorithmic rules for problem-solving. Cohen-Cole puts Osgood into the group of CWR thinkers without any reservations whatsoever, not recognizing that his theory, while having normative intentions, was based on empirical insights from social psychology, and could thus at least not satisfy the ideal of a purely formal theory. Note, furthermore, that we avoided defining CWR in terms of some specific version of rational choice theory, or perhaps rational choice theory plus probability and logic. Even without invoking the deeper paradoxes or puzzles, there existed different versions of these formal theories: just consider the fundamental debates over the proper meaning and standards of probability. The term ‘CWR’ is meant to encompass in principle all formal theories of rationality. So, I cannot agree with Teira’s claim that ‘[s]tarting with diversity, the authors define their ideal type in the most encompassing manner, but they often argue as if the canon within Cold War rationality was rational choice theory’.
Let me strengthen the point about the complexity of the ideal types of Enlightenment Reason and CWR and their overlapping assumptions and historical (dis)continuities with an example we did not use in How Reason Almost Lost Its Mind. When David Hume speaks of ‘reason’, he thinks of a capacity to justify logical/conceptual and empirical knowledge, but not a faculty that would have any practical power sui generis – as he famously says, ‘Reason is, and ought only to be the slave of the passions, and can never pretend to any other office than to serve and obey them’ (Hume, 1978[1739-40]: section II.3.ii.). Reason cannot guide or structure our passions or desires.
This is as much a view of the Enlightenment as the opposed position of Immanuel Kant’s. When Kant speaks of reason (Vernunft), he means both a speculative faculty by which we are misled into ‘transcendental illusions’ (drawing false conclusions about the existence of God, free will, or the beginning of the Universe) but also about a practical faculty that allows us to check moral judgments (and these are only some of the functions he ascribes to ‘pure’ reason; see for example Neiman, 1994). However, both Hume and Kant agree with many 20th-century adherents of CWR that reason can of course be instrumental: it can instruct us about the means for realizing our given aims. Indeed, Hume thinks that that is the only thing reason can do, while Kant denies that practical reason is merely instrumental, claiming that reason is also a source of distinctively moral principles. Neither Hume nor Kant would have thought about mechanizing reasoning (although others did so, such as Leibniz or before him the medieval philosopher Ramon Llull). Defenders of CWR, again, would not accept Hume’s view that practical reasoning is a purely passive instrument. Theories of rational choice include, after all, formal criteria of consistency and coherence that our preferences must minimally satisfy in order to count as rational – such as preference orderings having to be complete, transitive, consistent over time, and so on. Kant, again, might be seen as having overlooked the idea that we might formalize minimal coherence conditions of decision-making into a theory of rational choice, and as having ignored the idea that reason can help us to maximize utility.
Actually, one might view him as not having missed that road accidentally, as was argued by one author whom we might group among CWR thinkers, namely the philosopher of science Charles West Churchman (1913–2004). Churchman was a founding member and later president of the Institute of Management Sciences (TIMS; now INFORMS, the Institute for Operations Research and the Management Sciences) and an influential editor of the journals Philosophy of Science and Management Science. Furthermore, like a number of other American philosophers then, he worked as a consultant for the RAND Corporation that played such an important role in promoting research on rationality during the Cold War. In an article, Churchman (1970: 107) once raised the odd question of whether Kant was a decision theorist. He pointed out that Kant saw fundamental problems in the popular understanding of ‘pragmatic imperatives’ of reason, which were usually understood as rules of strategic planning with the aim ‘to maximize a set of social values’.
Kant, Churchman claimed, had noticed numerous limits in developing a genuine theory of rational choice, mostly due to the uncertainties involved in making ‘scientific’ statements of the form ‘Action X maximizes the utility of individual Y’. Kant’s criticism of teleological and hedonistic ethical views, as Churchman noticed correctly, contains the point that only a myopic simplification of situations made the idea of happiness-maximization plausible. In real life, the consequences of our actions are often too complex and uncertain. Similarly, too many examples in decision theory are overly simplified, and cannot be translated into the complexities of real decision-making. Churchman pointed out that this is so, in part, because the uncertainties of such judgments cannot be expressed in the classical language of Bayesian probabilities, and that a fully rational choice always requires ‘whole system judgments’. Furthermore, he claimed that decision theorists should take seriously Kant’s idea of categorical imperatives, and incorporate ethical values. Churchman illustrated his discussion of Kant through reference to typical problems dealt with in Cold War rational choice theories, such as management decisions or the Vietnam war (Churchman, 1970: 115.)
Such a reconstruction may hit the historian as absurdly anachronistic. Of course, Kant did not consider whether the relevant uncertainties could be expressed probabilistically or not. Still, the basic point about his rejection of the popular understanding of ‘pragmatic’ rules of reason is perfectly correct: Kant did indeed deny that utility maximization (be it of an individual or a group) is a sensible aim for pragmatic reasoning, and his view was based on a subtle recognition of complexities and uncertainties and the frequently shifting rules of human social action (cf. Sturm, 2009: ch. 8).
So, the devil is in the details. While Cohen-Cole makes a fair point that there are problems of how to write a history like ours, it cannot be really used to challenge us. In principle, we provided the epistemic and semantic qualifications that a concept such as CWR requires. Once they are grasped, the concept can be used as analytic tool for comparisons between authors and for the reconstruction of the historical development of the debate over what rationality is. Unless, of course, someone comes up with a better ideal type.
3. Cohen-Cole’s main objection concerns the claim, in his words, that CWR was ‘the defining feature of the human sciences’ (emphasis in the original) in the USA during the Cold War. Attacking this assumption, he argues there was another group which may be called ‘C(old War Intellectuals)…populated by three kinds of people: public intellectuals, computer and cognitive scientists, and educational policymakers’. They were ‘concerned with cultural analysis and understanding the whole character of persons or cultures. Mechanical rationality was a distant concern for…them’. Actually, Cohen-Cole not only means they were not interested in mechanizing rationality or reasoning: rather, Joseph Weizenbaum ‘concluded that designing social life and policy based on mechanical, instrumental reason not only depended on a misunderstanding of both humans and machines, but was deeply immoral as well’. Being disinterested in an intellectual or scientific trend is one thing; rejecting it is quite another. In any case, Cohen-Cole claims that these Cold War intellectuals ‘denied that thin, algorithmic rationality either characterized human nature or stood as its normative referent’. But, again, this objection fails. It fails not because it would be untrue that Mead or Weizenbaum or Chomsky were opposed, for moral reasons, to an amoral conception of rationality. Rather, it fails because Cohen-Cole finds it problematic to describe, as we do, the emergence of CWR as a mode of thinking and a group of widespread debates. We simply do not claim that CWR would have been ‘the defining feature of the human sciences’ in Cold War America. We just want to say that there is an elephant in the room that demands investigation. 1 And because we view ‘Enlightenment Reason’ and ‘CWR’ as ideal types, because we allow for intermediate cases and overlapping assumptions, there need be no sharp opposition. To point to Churchman again: he was trained in the typical modes of CWR but urged decision theorists also to recognize the importance of an ethical notion of rationality such as Kant’s (Churchman, 1970: 114 f.).
4. Now let me turn more closely to Teira’s review. I mentioned already that we do not fill the concept of CWR with a specific theory of rationality, let alone a specific version of the theory of rational choice. Another claim he makes – in fact, one implicit also in Cohen-Cole’s opposition between adherents of CWR and ‘Cold War intellectuals’ – is that we insist that CWR is somehow a necessarily egoistic theory: ‘I don’t think, at least, that any other of the approaches discussed in the book exhibits to the same degree the features of the type listed on p. 5: formal methods modelling self-interested individuals in conflict, with a radical simplification of the circumstances and a step-by-step impersonal approach to a solution’. We deliberately did not include self-interestedness or egoism among the necessary features of the ideal type of CWR. Of course, not few authors of the time would have assumed the rationality of egoism, and many do so still today. Teira’s claim reflects a widespread view, which one also finds in oppositions between ‘the rational’ and ‘the reasonable’ such as espoused by John Rawls: The rational…is a distinct idea from the reasonable and applies to a single, unified agent…with the powers of judgment and deliberation in seeking ends and interests peculiarly its own. The rational applies to how these ends and interests are adopted and affirmed, as well as to how they are given priority. It also applies to the choice of means, in which case it is guided by such familiar principles as: to adopt the most effective means to ends, or to select the more probable alternative, other things equal…What rational agents lack is the particular form of moral sensibility that underlies the desire to engage in fair cooperation as such, and to do so on terms that others as equals might reasonably be expected to endorse. I do not assume the reasonable is the whole of moral sensibility; but it includes the part that connects with the idea of fair social cooperation. (Rawls, 1996: 50–1, emphasis added; see already Sibley, 1953)
5. Teira’s main challenge is that we should explain the dominance of expected utility maximization: can it be explained by means of normative or a priori considerations, such as the Dutch Book argument which Teira hints at? Or is its prevalence due to some ideological background or context? I recognize that we did not address this issue in the book, and that perhaps we could and should have done so. In principle, a full answer would be complex. One would have to look more closely at various issues here. Firstly, is it true that expected utility maximization was dominant? In which communities of researchers? Perhaps some economists did embrace maximization, but not necessarily utility maximization. Secondly, granting for the moment that expected utility maximization was indeed dominant, why was this so? Several options would have to be considered here: (a) How many people actually understood and accepted considerations like the Dutch Book argument? (b) How many accepted expected utility maximization, especially in the form in which it was made popular through Leonard Savage, just because of a heuristic of the division of cognitive labour (‘if you don’t have the time or cognitive resources to think things through for yourself, accept what the leading experts say is right’)? (c) Do any political or other ideological contexts explain the actual acceptance of the relevant theory better than (a) or (b)? If so, which ones? Clearly, raising these questions invites many thorny historical problems – how to find and assess appropriate sources, how to rule out that scientists, in making self-reports about what brought them to accept a certain theory, are not biased or do not retrospectively make up justifications, and so on and so forth. However these would be dealt with, and however Teira’s question would ultimately be answered, we did not wish to indicate that political or social context explains everything; or that the Cold War was the only thing that determined the acceptance of specific theories of rationality. Answers may vary, depending on which author(s) one looks at.
While How Reason Almost Lost Its Mind is primarily a historical book, it deals with a set of distinctively philosophical notions embedded in the scientific and political contexts of a particular time and place. Teira’s question indicates that we may not able to rest here, but instead should be inspired to take the timeless aspirations of the ideal of rationality seriously. And I do. Rationality or reason is a capacity of human beings that philosophers have frequently cherished, or even viewed as the defining feature of humanity, a number of anti-rationalists notwithstanding. This is so not because of what has been taught within philosophical seminars, but because of the positive consequences that reason/rationality is said to have for ethics and the economy, for the sciences and for society at large. What we aimed to reveal through the book is that the origin of these ideas is not always isolable from that environment, not that the ideas must necessarily remain confined to that environment.
Footnotes
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author disclosed receipt of the following financial support for the research, authorship and/or publication of this article: this article has been supported by funding from the Spanish Ministry for Economy and Competitiveness (MINECO), project Naturalism and the sciences of rationality: An integrated philosophy and history (FFI2016-79923-P).