Rationality, Judgment, and Critical Inquiry

Paul Healy

Swinburne University of Technology


[0] In an insightful recent work, Harold Brown has revealed the shortcomings of the classical rule-governed (foundationalist) concept of rationality, and has sketched the groundplan for a new model which does greater justice to the findings of recent research in the history and philosophy of science. Particularly worthwhile features of the new model are its vindication of the (generally ignored but crucial) role of judgment in inquiry, the social (intersubjective) basis of rational decision-making, and the contextual and historically conditioned nature of evidence assessment. Taking these aspects of Brown's analysis as its starting point, the present paper seeks to provide additional arguments in support of his conclusions, while refining and expanding Brown's thesis at points at which his arguments are found wanting. I begin with consideration of the case for judgment as an intrinsic component of critical inquiry.


[1] As Brown documents, the dominant conception of rationality in recent philosophy has favoured a rule-governed approach.{1} Whatever the strengths of the classical model in terms of its efforts to provide a public and objective mode of determining rationality, a major weakness has been its neglect of the role of judgment in critical inquiry.{2} Indeed, in many instances judgment has not just been neglected in discussions of this topic, but its influence has been positively derided as a significant source of subjectivism and error, if not irrationality. It is against this background that Brown has sought to reinstate the credentials of judgment as a crucial element in reasoning. Brown establishes his point by a detailed analysis of the rule-governed paradigm. This analysis reveals that, despite its strengths, the classical model breaks down at several important junctures. Reflection on these weaknesses reveals the need for a new model of rationality, one which does justice to the proper claims of judgment.{3}

[2] From the perspective of present concerns, the key weaknesses in the classical model are documented by Brown in a section entitled "Appropriate Rules" (Brown 1988: 70-78). If, as has often been maintained since the time of Aristotle, rationality concerns the appropriate means for achieving a certain end or goal, the first problem for a rule-governed notion of rationality (a problem which to date has proved intractable) is that of how to determine the appropriate goal at which to aim other than by an arbitrary decision. Rules, it seems clear, could not be invoked in a non- circular fashion to justify the choice of ends, since they themselves are justified in terms of their effectiveness in reaching the goal, once this has been (independently) decided upon. Second, unless one appeals to self-evident intuitions, there would seem to be no non-arbitrary way to establish the appropriate starting point for the (subsequent) application of rules. Third, since there are always alternative rules available for tackling any given problem, we are faced with the problem of deciding between them. How do we decide which are the appropriate rules? Either we decide in an arbitrary way or we appeal to a meta-rule. But, as Brown observes, if we try to invoke a meta-rule, "we might well find competing meta-rules, and will have to seek higher level meta- rules to choose between them," thereby giving rise to a potential infinite regress (71). Finally, there is the problem that rules can be misapplied, "so that even if we have an appropriate set of rules, we must still know how to apply those rules" (71). Once again, unless we terminate the process by making an arbitrary decision or appeal to a self-evident intuition, we find that an infinite regress threatens. Since these kinds of flaws are intrinsic to the very structure of the classical model, Brown concludes that we must "reconsider the idea that rationality is captured in rules" (78). Moreover, in his subsequent discussion of rationality and science (chapter III), he provides extensive additional evidence for this view by demonstrating that contemporary analyses (by Kuhn and others) of scientific theorizing, traditionally regarded as the very paradigm of rational activity, show that it resists analysis in rule-governed terms.

[3] In his vindication of judgment, Brown's central argument is that judgment must be appealed to to stop the regress of justification, generated by appeal to rule-following alone. It is noteworthy, however, that in making this point, Brown insists (mistakenly, I think) that judgment is always ancillary to rule- following, when algorithmic procedures are available. Thus, in the case of such well-defined procedures as balancing a checkbook or differentiating a function (two of Brown's favorite examples of algorithmic processes),{4} Brown effectively dismisses the need for judgment. Instead, judgment is held to come into its own in "situations in which we run out of rules"; and in particular, in "cases in which we are attempting to develop new rules, cases in which we have to choose between a number of competing rules, and cases in which familiar rules fail us" (139). The case of writing a computer program or constructing a proof in formal logic provides a good illustration of the first two situations mentioned; scientific revolutions in science of the third. The core point that Brown wants to get at in his analysis of such cases is, of course, that in attempting to proceed rationally there is (by definition) no rule-governed way to make the right decision, to choose between such sets of rules as are available, or to decide on the next step to take. There are, however, better and worse ways of proceeding, and the best way forward must be adjudicated on the basis of informed judgment. Brown readily acknowledges that judgments are fallible, but he insists that they are not, on that account, subjective, still less baseless or irrational (144-45). It is true that decisions made on the basis of judgment are always "open to reconsideration under appropriate circumstances," but if they are made in a reliable way they nonetheless provide a firm (though clearly not indubitable) basis for rational analysis, one which provides a means for stopping the infinite regress of rule applications.

[4] But clearly to rehabilitate judgment it is not sufficient merely to establish its potential role as a starting point for rational analysis. It must be established as well, against the scepticism of the traditionalists, that judgment is a reliable and trustworthy component of the reasoning process. Following leads given by Brown, the question of the reliability of judgment is the theme I take up in the next section of the paper. But before embarking on that topic, a significant divergence between Brown's views and my own on how judgment functions in critical inquiry should be noted and commented on. While Brown's arguments go a long way towards rehabilitating its significance, it is nonetheless the case that in characterizing judgment as a skill to which we have recourse when we run out of rules, Brown vastly underestimates the centrality of judgment to the reasoning process. Indeed a much stronger conclusion is warranted on the basis of the evidence that he himself provides, the conclusion, namely, that the exercise of judgment transcends and underpins rule-following at every step of the way. Recall that a core aspect of Brown's thesis is that there are three good reasons as to why rule-following by itself is never enough to determine how to proceed rationally. Specifically, without appeal to some additional factor (like judgment), we could never establish the appropriate starting or finishing points for analysis, nor could we determine just what the appropriate rules were in any given case, nor could we know that we were applying the rules correctly. But from this analysis, Brown draws only the minimal conclusion that we need to have recourse to judgment in situations where we run out of rules. In his reluctance to challenge the rule-governed paradigm head-on, Brown fails to notice that his own arguments warrant the strong conclusion that judgment always underpins the reasoning process, even when we are employing fully algorithmic procedures (as when, say, balancing a checkbook or differentiating a function).

[5] What Brown overlooks is that, so far as the logic of his analysis is concerned, the situation is really no different regardless of whether we are dealing with a familiar or an unfamiliar situation. In other words, so far as the need to invoke judgment is concerned, we are effectively in the same boat regardless of whether we are engaged in a rule-governed process like balancing a checkbook or in a non-algorithmic process like constructing a proof in formal logic. Even in the former situation we face the same kind of dilemmas identified by Brown with regard to the latter; here too we must appeal to judgment to establish that we have the appropriate rules at our disposal and how they are to be applied. That the process may be more familiar, even "automatic," in the one case than in the other in no way diminishes this fact. It would seem that only the sheer familiarity of some examples employed by Brown could have prevented him from realizing that rule-following always presupposes judgment, and does not do so only in obscure or difficult cases. Ironically, although he fails to draw this important conclusion, Browns's very definition of judgment (as "the ability to evaluate a situation, assess evidence, and come to a reasonable decision without following rules" (137)){5} implicitly acknowledges its validity.

Good Judgment

[6] The claim that judgment is intrinsic to the process of critical reasoning might seem relatively unproblematic were it not for the fact that on a widely accepted view the exercise of judgment smacks of subjectivism, if not irrationality. Indeed, for many, the formality of rule-following is needed as a restraint on the impulsiveness and subjectivism of judgment. How can acknowledgment of the pervasive role of judgment be reconciled with rationality and objectivity in critical inquiry? The defense of judgment against the charge of subjectivism necessitates appeal to a concept of good or reliable judgment. It is only informed judgment, responsibly exercised, that can claim a fundamental role in rational inquiry. Accordingly, the present section explores, albeit in a very preliminary way, some of the key features of reliable judgment. In so doing, Brown's analysis will again provide the main touchstone.

[7] As Brown convincingly argues, the key factor in establishing the reliability of judgment is expertise,{6} acquired through training and practice in a particular area, and responsibly exercised. It is expertise and "know-how" in the exercise of judgment that diminishes its arbitrariness as it enhances its reliability. The way in which reasoning skills are enhanced can be seen by reflection on the way in which good judgment is developed through training and practice in a variety of fields. The difference in performance that results from developed expertise becomes clear by reflection on how one performs before and after acquisition of the requisite skill. Before one has gained facility in a field, one must proceed cautiously by coming to know and painstakingly applying whatever rules are available. But with increased training and practice, the available rules together with a generalized "know-how" as to what is required in the situation are developed and "internalized," and become, as it were, "second nature" to the skilled practitioner. The difference between an unskilled and skilled performance may be illustrated (to take one of Brown's examples (see 170-71)) by the contrast between the experienced and the inexperienced logician. When, say, evaluating the validity of a proof in logic, the skilled logician can proceed in an immediate and insightful way, without having to consciously call to mind the valid rules of inference, as the beginner must do.

[8] The exercise of skilled judgment is equally evident in a variety of other areas of expertise, such as correctly interpreting what one sees through a microscope, reading an x-ray, discerning the trails of electrons and other subatomic particles in experimental situations, etc. As Brown observes, the facility of expert performance, without conscious reference to rules, has been explicated by Polanyi in terms of "tacit knowledge" (see Polanyi 1958). Consider, for example, the skilled "way of seeing" that is gradually acquired by a medical student learning to read an x-ray. At first, despite his technical knowledge of medicine, the student simply does not have the ability to discern the relevant details on the x-ray, because he has not yet acquired the skills necessary to interpret the transparency in a meaningful way. But "as he goes on listening for a few weeks, looking carefully at ever new pictures of different cases, a tentative understanding will dawn on him; he will gradually forget about the ribs and begin to see the lungs. And eventually, if he perseveres intelligently, a rich panorama of significant details will be revealed to him: of physiological variations and pathological changes, of scars, of chronic infections and signs of acute disease"(Polanyi, cited in Brown 1988: 163).{7} It is at this point that the practitioner will be able to make reliable judgments in an immediate sort of way.

[9] Moreover, the fact that reliable judgment, acquired through expertise and practice, is a crucial factor in the conduct of inquiry is well borne out in the scientific domain by Thomas Kuhn's analysis of the role of "exemplars" in effective scientific problem-solving.{8} As Kuhn has shown, for science to progress, emerging scientists must acquire, not just a methodology, but also a "way of seeing." This ability, central to the conduct of science, to discern the relevant features of a problem situation and weigh their significance in the context in question cannot be reduced to a rule-determined activity, but is rather a skilled performance which requires the acquisition of a "way of seeing" as a result of exposure to a range of significant scientific problems and the types of strategies employed for their resolution.{9} Indeed, Kuhn insists that the very ability to understand the basic vocabulary of science cannot be accounted for by recourse to methodological procedures alone, but requires immediate and repeated exposure to the relevant range of scientific phenomena.{10}

[10] As the foregoing examples illustrate, good judgment is marked by the fact that, while its exercise is spontaneous and immediate, there is nothing arbitrary about its operation. Its reliability derives from the fact that it is an expert skill, acquired through extensive training and practice. Moreover, this analysis suggests that, although judgment is not (and cannot be) a rule-governed activity, its immediate exercise, when based on expertise, is not just desirable but also crucial in the conduct of critical inquiry. It is noteworthy too that skilled judgment shares much in common with the Aristotelian notion of phronesis (practical wisdom).{11} Like phronesis, it is acquired through training and practice; and the development of a given level of skill paves the way for still more skilled performance in the future. And just as practical wisdom becomes second nature to the phronimos, good judgment becomes second nature to the accomplished reasoner.

[11] A related way in which judgment establishes itself as markedly different from rule-following is with respect to its contextuality. What the expertise that is acquired as a result of training and practise enables one to do is to directly perceive the crucial features in a situation in a manner which paves the way for an appropriate response. What is acquired is the ability to directly discern the parameters of a situation in a manner analogous to that which is involved in the appreciation of a work of art.{12} Trained judgment brings with it the ability to recognize the salient features in a situation, the relevant constellation of operative factors and patterns, their harmony or disharmony, and the weight that they should have in a particular context. Such discernment is all-important for deciding on the appropriate response in the particular situation at hand. And here it is worth emphasizing that, as the foregoing analysis makes plain, there is nothing mysterious, still less irrational, about the development of this ability to see directly into what the situation requires; it is a natural development, given the extensive training leading to its acquisition (which, as we have seen, is a precondition for the reliable exercise of judgment).{13}

[12] It should be noted too that judgment is historically conditioned in its mode of operation, and that it never operates in an entirely presuppositionless way.{14} Thus, the relevant factors that judgment, to be effective, must take account of are significantly influenced by historical considerations, and by what questions and presuppositions we approach the situation with. The point at issue here is amply illustrated by Kuhn's analysis of the history of science and the onset of scientific revolutions. As Kuhn's analysis shows, in different historical circumstances very similar perceptual data are interpreted in completely different ways (to take one of Kuhn's best-known examples: what for an Aristotelian is a stone on a piece of string subject only to the dynamics of free-fall becomes for Galileo a pendulum, obeying very different principles of dynamics). Indeed, this viewpoint is at the heart of Kuhn's claims about the significance of Gestalt switches in the onset of scientific revolutions.{15}

[13] In contrast to the dominant rule-governed conception of rationality, the main aim of this section has been to establish the reliability (and non-arbitrariness) of expert judgment, and its centrality to critical inquiry. But it must be noted that the defense of this claim does not involve the presupposition that judgment is infallible. Far from it. The present approach readily acknowledges that experience teaches that even the most expert judgments are sometimes mistaken.{16} This being so, something additional is required before our best informed judgments are taken to be rationally and objectively established. Having vindicated the role of judgment in critical reasoning, consideration of what constitutes its rationality and objectivity represents the second (crucial) step in the development of a new (non-algorithmic) concept of rationality.


[14] The features of judgment that have been thus far discussed (expertise, contextuality, etc.) are sufficient to ensure its reliability, but hardly by themselves sufficient to ensure the rationality and objectivity of judgment. However reliable, however expert judgment might be, error is always possible. As Brown notes, the ability to exercise expert judgment does not ensure that such expertise will always be reliably exercised.{17} And where it is (as we might typically expect it to be), this does not preclude the possibility of other well-justified interpretations, based on alternative presuppositions. Accordingly, to qualify as rational and objective, the new model must incorporate some checks on the reliability of judgment. Moreover, for the kinds of reason that scientific inquiry requires public verifiability of data and interpretations as a condition for its objectivity and rationality, it would seem that the new model must also embody an intersubjective dimension. In sketching a new model of rationality on the basis of his critique of the old, Brown seeks to do justice to these requirements. As heretofore, reflection on Brown's proposals will provide the guiding thread for our analysis, even if these proposals themselves should prove to be in need of critical refinement.

[15] The model which Brown proposes comprises three essential steps, each of which warrants comment. On Brown's view, "the first step is to take the notion of rational agent as fundamental, and such notions as 'rational belief' as derivative in the sense that a rational belief will be one that is arrived at by a rational agent" (185). The core idea that Brown seeks to preserve here is that "a belief's rationality is connected with the way we arrive at that belief; a belief that we arrive at on the basis of an adequate body of appropriate evidence is a rational belief, while a belief that we arrive at without evidence, or against the evidence, or on the basis of irrelevant evidence, is not rational" (185). As Brown observes, the traditional model attempted to do justice to this intuition by emphasizing the logical relations between evidence and belief. But the new model, by contrast, "will take the agent to be basic, and the way in which an agent deals with evidence in arriving at a belief to be determinative of the rationality of that belief" (185). In making the rational agent fundamental, Brown is, of course, seeking to transcend rule- governed rationality. But even so, no adequate explanation is given as to why in the context of a three-stage process, it is the ratiocination of the individual agent which should be characterized as fundamental. This important question of the locus of rationality is one to which we will return below.

[16] Brown's second step returns to a theme which, unlike the first, has been well prepared for on the basis of his analysis: the fact that, though long overlooked, judgment is intrinsic to the rationality of beliefs. As will be apparent from the earlier comments, I am largely at one with Brown on this issue, the only point of significant disagreement between us being that concerning the relationship between rule-following and judgment. While Brown emphasizes the use of judgment in situations in which we run out of rules, I maintain that judgment has a more ubiquitous function: it always underpins rule-following and renders it possible. Brown's third step introduces the important notion of intersubjective verification. As noted above, the rationale for this step is that human judgment, even when it is expertly formed, is "notoriously fallible," giving rise to the need for independent checks. The challenge here is to find an adequate alternative to the traditional recourse to rule-governed procedures.

[17] Brown develops the proposed alternative by invoking the philosophies of Wittgenstein and Kuhn, who, he points out, insist on the social character of the verification process. On this basis, the key contention he wishes to defend is that, "for a belief based on judgment to be a rational one, it must be submitted to the community of those who share the relevant expertise for evaluation against their own judgments" (187). That is, to be termed rational, an individual judgment must be validated by submission for critical appraisal to the relevant expert community. But, unfortunately, except for broad allusions to the philosophies of Wittgenstein and Kuhn, and to Popper's critical rationalism, Brown is not very explicit about what is involved in this type of intersubjective validation. As is pointed out below, more can, and should, be said about this crucial process. But before engaging that issue, some comments should be made about other problems with Brown's proposed new model of rationality, particularly regarding its first step.

[18] As already noted, in characterizing the individual agent as fundamental for rationality, Brown is seeking to break with the rule-governed model which has been the focus of his critical concern. But is his conclusion about the centrality of the agent for rationality justified? So far as I can see, Brown's claims in this regard are not in fact justified since he provides no good reasons why out of a three-step process, one step -- that focusing on the agent -- should be taken as fundamental for rationality. Unlike step two, which deals with the role of judgment in objective reasoning, the case for the foundational character of the individual agent is not explicitly argued in Brown's analysis. True, Brown justifiably seeks to break with an outmoded model which views rationality merely as an abstract relation between propositions, and to advance the view that rationality involves an active process of forming and testing beliefs in accordance with the best available evidence. But on Brown's own telling three steps are needed for this process. What remains unclear then is the justification for making one of these steps rather than the three- stage process itself the locus of rationality. Undue emphasis on the first step can lead to the mistaken impression that the second and third steps are simply derivative, if not redundant. But in fact the balance of Brown's own analysis argues against this view. For Brown subsequently argues that if it conflicts with the requirement of intersubjective verification the individual's judgment cannot be taken as rational after all. Thus, Brown explicitly maintains that "Robinson Crusoe alone on his island could exercise judgment, but he would not be able to achieve rationality" (187).

[19] On Brown's analysis, the reason why Robinson Crusoe, alone, could not achieve rationality is because de facto he does not satisfy requirement three, that of the intersubjective validation of beliefs. Accordingly, even on Brown's telling, there is little ground for maintaining that the individual agent is the locus of rationality. This point remains valid, I believe, notwithstanding the fact that Brown fails to notice that while rationality certainly requires the intersubjective validation of judgments and beliefs, this must surely be an in principle requirement. It can hardly be denied that the individual scientist (or anyone else for that matter), any less than Robinson Crusoe, functions rationally when, although working alone, she forms her beliefs on the basis of the best evidence available. Since subjectivism and error are indeed possible even in the case of the most expert judgment, we rightly require that the rationality and objectivity of the individual's beliefs are publicly validated through a process of intersubjective critical assessment. But it seems equally clear that we have no grounds for denying rationality to the individual's beliefs, provided they are formed on the basis of the best available evidence, even in the absence of actual intersubjective checks. We are, thus, committed only to an in principle requirement of intersubjective testing. This, I believe, constitutes an important modification of Brown's proposal. My overall point here, however, is that given problems of the sort just discussed, and the explicit formulation of a three-stage model, it is difficult to see why Brown has failed to draw the fully defensible conclusion that rationality is the result of the process (comprising all three steps) of forming and testing beliefs in accordance with the best available evidence. So far as I can see, this would do justice to all the elements in his analysis, while not conflicting with anything of significance. Moreover, on several occasions, Brown's way of explicating what is at issue makes clear (if implicit) reference to the centrality of the process itself rather than of any single aspect of it.{18}

[20] A further significant problem with Brown's analysis is that, while his third step is clearly crucial to the whole process, Brown does not provide a specific account of what intersubjective validation involves. As noted above, he relies mainly on allusions to aspects of the philosophies of Wittgenstein, Kuhn, and Popper to vindicate the social (as distinct from the individual) character of rational belief formation. Later on, in a section entitled "Rational Disagreement" (207-224), he illustrates the interplay of areas of agreement and disagreement involved in two classic scientific disputes, the one between Galileo and the Aristotelians, the other between Einstein and Bohr. But in neither case does he give an account of the process by which intersubjective validation works. While the present paper cannot undertake to fill in the picture in a detailed way, a promising line of development (neglected by Brown) can be presented for future consideration.

[21] With regard to the problem of intersubjective validation, the most promising current approach is, I believe, the model advocated by Jurgen Habermas. Unfortunately, even in Habermas's writings the process is only sketched in a fairly embryonic way, leaving much need for further analysis.{19} But the core idea is that rationality is achieved through a rigorous process of argumentative assessment of beliefs, relative to the evidence (support) available for them. To the extent that Habermas develops this concept, he does so with reference to the model of argumentation developed by Stephen Toulmin (1958). The argumentative model of evidence assessment outlined by Toulmin, and endorsed by Habermas, operates essentially on the principle that, when a claim is advanced successive levels of evidential support (in the form of data, warrant, and backing, respectively) must be given on demand, if the claim is to be rationally justified. Moreover, a rebuttal phase is incorporated to ensure that, in addition to being supported by positive evidence of the appropriate kind, claims must be defended against challenges from critics, if they are to establish their credentials as "rationally motivated." These procedures are held to constitute a forum for the "severe testing" (to use a Popperian phrase) of the rational grounding of beliefs or claims.

[22] To ensure the rigorous testing of beliefs, it has been argued that the model is best construed in dialogical (or dialectical) terms, whereby criticism of a position can be forthcoming from a multiplicity of perspectives. The model should also be construed as a dynamic process whereby modifications are made to initial positions, if these are seen to be needed on the basis of the aforementioned critical debate. Frequently, the way will be prepared for modification of initial positions, leading to a resolution of the issue in dispute, only when in the course of the debate initially hidden presuppositions and value judgments have been uncovered, and themselves submitted to argumentative scrutiny.{20}

[23] But while it provides a productive avenue for development, the Habermas model is not without its shortcomings. Its most telling problem is, I believe, that it contains certain unwarranted assumptions. Chief among these are Habermas's presuppositions that consensus is the anticipated outcome of the argumentative process, and that an ideal speech situation must be anticipated as the forum for argumentation. Neither presupposition is justified. As can be briefly indicated, the presupposition of ongoing rational disagreement amongst diverse expert groups, always operating in less-than-ideal conditions, is both closer to reality and more likely to foster the advancement of critical inquiry than the presuppositions advocated by Habermas. An important consideration here is that the history of science (a mode of inquiry that is generally regarded as the paradigm of rationality) seems to provide telling evidence against the anticipation of consensus as a presupposition for rational inquiry. The frequent disputes among experts, not to mention the incidence of scientific revolutions, suggest that the existence of ongoing rational disagreement is a far more tenable presupposition than consensus.

[24] Moreover, there are good grounds for supposing that rational disagreement is an intrinsic (and a by no means accidental) feature of rational inquiry. For one thing, since theories are underdetermined by observational data, it follows that more than one conclusion is rationally justified on the basis of a given body of evidence. Furthermore, with regard to the process of theory choice, Kuhn has convincingly argued that even when the relevant constellation of decision criteria has been identified for a given domain (accuracy, consistency, scope, simplicity, and fruitfulness emerge as especially important in science), the choice of one theory over another still cannot be uniquely determined. Instead, questions such as those concerning the relative weighting of the various selection criteria immediately come into play, giving rise to ongoing debate (rational disagreement) among various contending, expert groups.{21} Accordingly, disagreement (not consensus) must be taken as the underlying presupposition of inquiry. Moreover, if Popper is correct, consensus amongst experts is far from being (even) a desideratum for science, since rational disagreement and mutual criticism are held to be the life blood of scientific progress and the growth of knowledge.

[25] Although providing a promising approach for the development of an argumentative model of rationality, Habermas's model nonetheless also embodies a second highly problematic presupposition, that of the ideal speech situation. As a condition for inquiry, the point of this presupposition is to provide for the possibility that all relevant contributors to a given debate should be allowed to particate without reference to the constraints of time or place (Plato and Aristotle, therefore, being as much potential contributors to contemporary debates as ourselves). This presupposition is clearly a counterfactual ideal which could not be fulfilled in practice. For Habermas, its primary justification lies in the (conceptual) safeguard it provides against beliefs being formed on the basis of coercion or bias, rather than through rational debate. Only by conducting the debate in the most open forum imaginable, wherein all qualified participants could potentially have input (and wherein the widest range of relevant evidence could be canvassed), can we be sure that the outcome is "rationally motivated," and not distorted by ideological considerations. Thus, the motive behind this requirement is a worthwhile and appropriate one.

[26] Nonetheless, as a prerequisite for rational inquiry, this requirement is misplaced. In the first instance, the presupposition is inappropriate because it completely transcends the realizable conditions for human knowing. My point is not, of course, simply that Plato and Aristotle could not be actual (in person) participants in the contemporary debate (notwithstanding the fact that they are clearly significant shapers of this debate). My point is rather that it is no part of what we understand by human knowing or rational inquiry that a debate be conducted in such counterfactual circumstances. Instead we must, I think, acknowledge that we are situated knowers, and that such rationality and objectivity as we are capable of must be achieved within finite and constrained epistemological conditions. Presupposing the possibility of a set of epistemological conditions which could not possibly obtain for human knowers does little, I believe, to assist us in coming to understand what is involved in achieving rationality in the actual circumstances of human knowing.

[27] Indeed, the presupposition of an ideal speech situation is especially paradoxical in a philosophy such as that of Habermas which emphasizes the development of knowledge in historically situated human communities. Moreover, the presupposition of an ideal speech situation conflicts sharply with the actual contextuality of human knowing, and also with the (very real) possibility that rationally justified conclusions may diverge rather than converge. Clearly, this is a large topic in its own right. But the point I want to press here is that we cannot assume that what it is rational to believe remains constant across historical periods (or indeed cultures). Rather what counts as sufficient evidence to support a claim, or as a reasonable objection to a claim already advanced, or as an adequate rebuttal of a counterclaim, or indeed what counts as a bona fide problem or solution at all, are not matters which can be decided in the abstract in an atemporal fashion. Rather, as the historical evidence provided by Kuhn (in his analysis of scientific revolutions) and Brown (in his account of the factors which make for rational disagreement in different historical epochs (Brown 1988: 207-24)) demonstrates, the appropriateness of evidence and the rationality of scientific decisions are factors which are very much conditioned by the prevailing historical circumstances. Accordingly, what needs to be emphasized (and further explored) is (not consensus or an ideal speech situation but) the contextuality and ongoing revisability of rational belief formation and assessment, and the important role which rational disagreement plays in this process.


[28] In this paper, I have sought to defend and expand the main outlines of Brown's new model of rationality, while correcting some significant deficiencies in it. However, these proposals do little more than indicate a promising way forward. Much work still remains to be done in refining and fully justifying the new model.

[29] As I see it, given my endorsement of Habermas's argumentative conception of rationality as an extension of Brown's position, a particularly pressing concern is the development of a detailed model of argumentation, adequate to fully specifying the requirements for rational justification in a dialectical forum. In the relatively recent past, significant advances have been made in our understanding of the process of justification on the basis of "good reasons" (as distinct from standard deductive or inductive reasoning as such){22} but, so far as I know, no rigorous model has yet been developed to ground this approach. Progress in this area would be an undoubted advantage in vindicating the new model. Likewise, while Kuhn has done much to gain support for the view that, in the rational appraisal of theories or beliefs, different factors should receive different weightings in different (social and historical) circumstances, we have, as yet, no clear understanding of how the significant factors are to be relatively assessed and balanced, and disputes about their relative weightings resolved among expert communities. Indeed, given the centrality of the concepts of "expertise" and "expert communities" to the whole model, more needs to be known about the conditions for membership in such communities. Significant research in these (and related) areas is much needed to gain widespread acceptance for the credentials of the new model.


Brown, Harold I. (1988) Rationality. New York: Routledge.

Gadamer, Hans-Georg (1975) Truth and Method. New York: Seabury Press.

Govier, Trudy (1988) A Practical Study of Argument, 2nd Edition. Belmont, CA: Wadsworth.

Habermas, Jurgen (1970) "Towards a Theory of Communicative Competence." Inquiry 13, 360-75.

---- (1973a) "Wahrheitstheorien." In Wirklichkeit und Reflexion, H. Fahrenbach (ed.), 211-65. Pfullingen: Neske.

---- (1973b) "A Postscript to Knowledge and Human Interests." Philosophy of the Social Sciences 3, 157-89.

---- (1984) The Theory of Communicative Action, Volume I. Boston: Beacon Press.

Healy, Paul (1987) "Critical Thinking and Dialectical Argument: An Extension of Toulmin's Approach." Informal Logic 9, 1-12.

Johnson, Ralph H. and J. Anthony Blair (1983) Logical Self-Defense, 2nd Edition. Toronto: McGraw-Hill.

Kuhn, Thomas (1970) The Structure of Scientific Revolutions, 2nd Edition. Chicago: Chicago University Press.

---- (1977) "Objectivity, Value Judgment, and Theory Choice." In The Essential Tension, 320-39. Chicago: University of Chicago Press.

McCarthy, Thomas (1978) The Critical Theory of Jurgen Habermas. Cambridge: MIT Press.

Polanyi, Michael (1958) Personal Knowledge. New York: Harper.

Toulmin, Stephen (1958) The Uses of Argument. Cambridge: Cambridge University Press.

----, et al. (1979) An Introduction to Reasoning, 2nd Edition. New York: Macmillan.


{1} On the "classical model" of rationality, see especially chapters I and II of (Brown 1988). Return

{2} I use the term "critical inquiry" in a broad sense to refer to the process of rational belief formation and assessment in both science and everyday life, the topic which is the focus of Brown's work also. Return

{3} (Brown 1988) chapter IV contains Brown's vindication of judgment. The analysis which follows is an attempt to capture the spirit of Brown's critique rather than a point-for-point report of his views. Return

{4} Algorithms are "rules which, when applied to a problem, guarantee a solution in a finite number of steps. The rules for the various arithmetic procedures, the rules by which one finds the derivative of any function, Venn diagrams, the truth-table test for validity, and most computer programs are examples of algorithms" (Brown 1988: 20). Return

{5} Emphasis mine. See also (164-65): "The classical model of rationality takes rule-following to be a fundamental cognitive ability and attempts to capture skills in sets of rules, but this has things backwards since the ability to act in accordance with a set of rules is itself a skill." Return

{6} See, for example, (139): "Judgments must be made by individuals who are in command of an appropriate body of information that is relevant to the judgment in question." Return

{7} As Brown notes, skill acquired through training and practice is evidenced, not just in specialized matters, but also in a whole range of everyday tasks. Consider, for example, the expertise exhibited in effectively performing such tasks as: hammering a nail, riding a bicycle, speaking one's native language, discerning the meaning of what is spoken, etc. Return

{8} On the crucial role of exemplars in scientific problem solving, see especially (Kuhn 1970: 187-91). Return

{9} See (Brown 1988: 143): "Exercising these skills does not reduce to the application of an algorithm--if it did we would teach the algorithm instead of putting students through the routine of studying examples and doing problems in the expectation that this will improve their performance." Return

{10} Thus, as Kuhn puts it in one place: "If the student of Newtonian dynamics ever discovers the meaning of terms like 'force,' 'mass,' 'space,' and 'time,' he does so less from the incomplete though sometimes helpful definitions in his text than by observing and participating in the application of these concepts to problem- solution" (1970: 47). Return

{11} Nicomachean Ethics, esp. 1140a25-b29. For further comment on some Aristotelian parallels, see (Brown 1988: 150-54). Return

{12} This analogy is suggested by Hans-Georg Gadamer's analysis of the operation of understanding in terms of aesthetic appreciation in Part I, Sec. II of Truth and Method. Return

{13} In a section entitled "Judgment Naturalized," Brown shows that the operation of judgment can be adequately explained in naturalistic terms, in which the individual's education and training play an integral part (see 165-77). Return

{14} My thinking here is again influenced by Gadamer who insists that "reason exists for us only in concrete, historical terms, i.e., it is not its own master, but remains constantly dependent on the given circumstances in which it operates" (1975: 245). Return

{15} (Kuhn 1970), see especially chapter 10, "Revolutions as Changes of World View."

It is noteworthy that Brown's analysis of contextuality does not extend beyond noting two dimensions to good judgment. In addition to the body of generalized background knowledge that is at the disposal of all people trained in the field, one must also become familiar with the body of information relevant to the case at hand. "For example, in order to be in a position to make a medical decision, or design a bridge, we must master both the methods and knowledge of the applicable disciplines, and also the available information relevant to this particular situation. We should not wish to have our medical diagnoses done by individuals who have not studied medicine, but neither would we rely on the judgment of a thoroughly competent physician who had not bothered to study the facts of our particular case" (Brown 1988: 146). Return

{16} See Brown on the fallibility (but intrinsic reliability) of informed judgment (144-45). Return

{17} See (Brown 1988: 146-48). Return

{18} Thus, for example, in summarizing his position, Brown observes that "on our new model a rational belief or decision is one that an individual has arrived at through a two-step process... The belief is based on judgment--where possession of the relevant information and expertise is a necessary condition for a judgment--and his judgment has been tested against the judgments of those who are capable of exercising judgment in this case" (196; emphasis mine). Other implicit references to a process of rational belief formation are to be found on (190-95). Return

{19} The clearest recent statement of Habermas's concept of argumentative rationality is to be found in (Habermas 1984), especially pages 17-31. For earlier allusions to this notion, and for Habermas's position on consensus and the ideal speech situation, which figure prominently in the discussion below, see, e.g., (1973a) (unfortunately this essay has not been translated into English), (1973b), (1970). For critical commentary on Habermas's position, see especially (McCarthy 1978). Return

{20} These modifications are argued for in (Healy 1987). Return

{21} On this point see in particular (Kuhn 1977). It is noteworthy that while Brown explicitly acknowledges the role of rational disagreement in scientific inquiry, he seems to mistakenly attribute a consensual view to Kuhn (see, e.g., (Brown 1988: 192)). Return

{22} For the fundamentals of the "good reasons" approach to argument analysis and assessment, see, e.g., (Toulmin 1979), (Govier 1988), (Johnson and Blair 1983). Return

Paul Healy
Swinburne University of Technology

The Electronic Journal of Analytic Philosophy 1:3 (1993)
ISSN 1071-5800
ejap@phil.indiana.edu http://www.phil.indiana.edu/ejap/
Copyright 1993