ISSA Proceedings 2002 – Rationality And Judgment
Philosophical/epistemic theories of rationality differ over the role of judgment in rational argumentation. According to what Harold I. Brown (1988) calls the ‘classical model’ of rationality, rational justification is a matter of conformity with explicit rules or principles. On this view, a given belief, action or decision is rational only in so far as it is rendered so by a relevant rule or principle. These rules or principles must themselves be justified by appeal to other rules or principles. According to the classical model, judgment plays no role in the determination of rationality; whether a belief or action is rational is a matter, not of judgment, but of its relation to the appropriate rules(i). Critics of the classical model, e.g. Brown and Trudy Govier (1999), argue that the model is subject to insuperable difficulties. They propose, instead, that rationality be understood in terms of judgment rather than (or in addition to) rules. Govier criticizes some of my previous work on the subject for being overly committed to the classical model, and for equivocating on ‘judgment.’
In this paper, I consider Brown’s and Govier’s criticisms of the classical model and their defense of what I will call the ‘judgment model’ of rationality, as well as Govier’s critique of my earlier discussions. While my own commitment to the classical model is (I think and hope) somewhat more nuanced than Govier alleges, and so avoids at least some of her criticisms, the main burden of my paper will be not so much to defend my view from those criticisms, but rather, first, to articulate what I think are two deep problems for the view that Brown and Govier advocate: its inability to distinguish between rational and irrational judgment, and its inability to avoid recourse to rules. This inability, I will argue, renders the view inadequate as an account of rationality, critical thinking, or argument appraisal. Second, and more positively, I hope to show that, properly understood, an adequate account of rationality will centrally involve both rules and judgment.
The Classical Model and its Problems
According to Brown, it is basic to any acceptable account of rationality that “rational beliefs must be based on reasons” (p. 38; see also p. 183). That is, whatever else a given theory says about the character and constitution of rationality, it must at least include a provision according to which reasons are fundamental. The point is perfectly general, and runs far beyond belief: for a belief, but also an action, a hope, a fear, a decision, a vote, or anything else to be rational, it must in some sense be ‘based on reasons.’ It is endorsed by most contemporary theorists of rationality, including myself(ii), and I will presume it in what follows. Granting that rationality involves reasons, what else is needed for a belief, action, decision or whatever to be rational?
On Brown’s articulation of the classical model, there are three further necessary conditions on rationality: for X (a belief, action, decision, etc.) to be rational, it must be universal, in that a given body of evidence will render X rational (or not) universally, i.e. for any arbitrary individual agent; necessary, in that X must follow necessarily from the relevant body of reasons/evidence; and, finally, must be a matter of rules, in that X must both conform to and be based on appropriate rules (pp. 5-19). While all three conditions are worth extended discussion, I will focus on the last in what follows.
As Brown puts it, according to the classical model “the rationality of any conclusion is determined by whether it conforms to the appropriate rules.” (17) Various sorts of rules can be employed here, according to the model: logical rules, inductive rules, methodological rules, evidential rules, etc. So, for example, according to the classical model it is rational to conclude or believe that
r: Amsterdam is in Europe
on the basis of
p: Amsterdam is in The Netherlands
q: The Netherlands is in Europe
along with the rule which can be stated as
R: If a is in b, and b is in c, then a is in c.
Similarly, on the classical model it is rational to believe/conclude that
q’: Amsterdam is a beautiful city
on the basis of
p’: Amsterdam has many beautiful canals, buildings, and parks
and the rule which can be stated as
R’: If a city has many beautiful features, it is (probably) a beautiful city.
As a final example, according to the classical model it is rational to believe/conclude that
r’’: Amsterdam is a more permissive city than any U.S. city
on the basis of the reasons/premises
p’’: Amsterdam has a large and thriving ‘red light’ district, and many cafes in which marijuana and hashish can be openly smoked
q’’: In U.S. cities, prostitution and drug use are both illegal and actively discouraged by law enforcement officials
and the rule which can be stated as
R’’: If a city permits activities that other cities do not, the former city is, ceteris paribus, more permissive than the others.
These examples appeal to several different sorts of rules. Although Brown provides illuminating discussions of these and other sorts of rules, I will pass over those discussions in order to focus directly on Brown’s main point: the classical model’s insistence that rationality is determined (in part) by conformity with relevant rules creates difficulties that it cannot satisfactorily resolve. The basic problem is that, in any given case in which we are concerned to determine the rationality of a given conclusion (or belief, action, etc.) by seeing whether it conforms to the appropriate rules – i.e., whether it can be rationally concluded from initial premises or information by appeal to such rules – we will have to answer two questions: (a) with what information/premises shall we begin?, and (b) to which rules shall we appeal? Brown argues that both questions can be answered only by appeal to further rules, thus launching an infinite regress (or a vicious circle or arbitrary decision). Let us focus for the moment on (b). How can we determine whether or not our appeals to/belief in/acceptance of the rules R, R’, and R’’ in the examples above are themselves rational? There seem to be only two possibilities open to us, on the classical model, both of which are problematic: (1) appeal to further rules, which immediately raises the specter of infinite regress; or (2) do not appeal to further rules, which seems to render our appeal to those rules arbitrary. As Brown summarizes his analysis:
…the classical model of rationality faces serious problems when it is consistently developed. The model requires that rationally acceptable claims be justified, and that the justification proceed from rationally acceptable principles in accordance with rationally acceptable rules. Each of these demands leads to an infinite regress unless we can find some self-evident principles and rules from which to begin, but these have not yet been found, and there is no reason to expect that they will be forthcoming (p. 77).
Brown’s argument for this conclusion depends on his claim that “the classical model of rationality requires a foundational epistemology” (p. 58) in order to satisfy the requirements it places on rationality, coupled with a detailed critique of several types of foundationalism. I will not discuss the case Brown makes here because, while I agree with him that the versions of foundationalism he considers are defective, I think that the classical model – or at any rate the view that rationality is fundamentally a matter of rules – need not and should not be tied to foundationalism. Below I will try to show both that any successful account of rationality will perforce need to appeal to rules/principles, and that such rules/principles do not require a foundationalist epistemology. Before addressing these matters, however, it will be helpful first to consider the alternative to the classical model that Brown and Govier favor.
An Alternative to the Classical Model: The ‘Judgment Model’ of Rationality
Central to the ‘judgment model’ of rationality, of course, is the notion of judgment. So it is imperative that we be clear on how this notion is to be understood. Happily, Brown’s and Govier’s accounts of it are sufficiently similar – in fact, Govier on the whole simply embraces Brown’s – that they can, for the moment, be treated together. Brown writes: “Judgement is the ability to evaluate a situation, assess evidence, and come to a reasonable decision without following rules” (p. 137). Govier writes: “What is judgment? It is what we exercise when we are able to make reasonable and sensible decisions without appealing to rules” (p. 129)(iii). There are three elements of these accounts, as articulated in the passages just cited, to which we must attend: for both Brown and Govier, judgment is:
1. an ability, something we exercise;
2. is not exercised by following or appealing to rules; and
3. results in reasonable decisions. All three elements will be addressed below. First, let us briefly consider some other features of Brown’s and Govier’s accounts.
For Brown (as for Govier, p. 127), the exercise of judgment is fallible (pp. 144-6). Brown emphasizes that neither this fallibility nor the fact that it is not rule-governed entail that judgment is arbitrary. A key reason Brown offers for thinking that judgment is not arbitrary is that, on his view, judgment cannot be made in ignorance: “judgment on a topic can only be made by those who have mastered the body of relevant information” (p. 146). He further emphasizes that, in addition to expertise, the exercise of judgment requires skill, that some people are more skilled at it than others, and that its exercise does not require or involve an appeal to rules (pp. 156-65). Finally, Brown’s new model of rationality differs from the classical model, and the foundationalism with which Brown thinks the latter is ineluctably intertwined, by
1. making rational agency fundamental, and rational belief derivative;
2. taking the ability to make judgments in situations in which decisions cannot be determined by rules to be a basic feature of rational agency; and
3. requiring that candidate beliefs and judgments be submitted to the consideration of the relevant community of experts for their evaluation (pp. 185-7). Thus, on Brown’s positive account, judgments – i.e. the results of the exercise of judgment – are not always rational. To be so they must be the result of the skilled exercise of judgment, reflect mastery of the relevant information, and “have been tested against the judgements of those who are also capable of exercising judgement in… critical debate” (pp. 196-7).
Govier endorses and helpfully explains and develops Brown’s account, in the context of her consideration of the place of rationality in the theory and practice of critical thinking and argument analysis. On Govier’s view, “judgment is indispensable in critical thinking.” (p. 123) She takes argument analysis and evaluation to be central to critical thinking and so to rationality (as do I). Her discussion begins with a basic observation concerning argumentation:
In the analysis and evaluation of arguments there are many points where decisions have to be made which are not generated by algorithms. That is, we have to decide what to do or say, and there are no universal rules which we can call upon to generate or justify our decision. This is where judgment comes in (p. 123).
On Govier’s view, judgment is involved whenever our decisions are not guided – i.e., generated or justified – by algorithms or ‘universal’ rules. She argues persuasively that judgment is required more or less throughout the activity of argument analysis and evaluation – it is involved in argument interpretation, reconstruction, and throughout the process of evaluation (pp. 123-5). But she (like Brown, p. 138) is not a skeptic about rules; she denies neither their existence nor their applicability to real cases of argument analysis/evaluation. Her view, rather, is that rules cannot be the whole story:
I am not saying we should never use rules, or that we should seldom use rules, and I am not denying that it is often useful and appropriate to articulate possible rules and try to test their applicability. I am merely saying that abiding by universal rules is not all there is to rationality. Rules are bound to run out at some point, whatever the endeavor… (p. 125).
Govier is clear that her thesis about rules is not restricted to matters of rationality, but is rather completely general:
Rules cannot do everything. We must sometimes make decisions about what to believe or what to do when we have no rules to appeal to. This, I have submitted, certainly happens when we are interpreting and evaluating arguments, or attempting to think critically about a claim, theory, or issue. And it happens in a vast variety of other contexts as well. It is, I submit, a perfectly general feature of human life and thought. If we run out of rules sometimes when explaining how to identify, interpret, and assess arguments, we need feel no special theoretical embarrassment about this fact….The fact that what we do and think is not a product of applying a universally applicable algorithm is ubiquitous in human life; it is by no means restricted to the area of critical thinking and argument evaluation (p. 127).
If rules are not sufficient for rationality, what is? Govier’s alternative account of rationality centers, of course, on judgment. She endorses Brown’s account, according to which judgment is “the ability to evaluate a situation, assess evidence, and come to a reasonable decision without following rules” (p. 127, citing Brown, p. 137). She develops her view of judgment in ways which will be considered below. But we are now in position to consider some problematic aspects of the judgment model, to which I turn next.
3. Problems with the Judgment Model.
Earlier we identified three features common to Brown’s and Govier’s versions of the judgment model: on both their accounts, judgment is an ability which persons exercise; it results in reasonable decisions (as we have now seen, for Brown at least, only if the additional constraints of expertise, skill, and social testing/evaluation are met); and its exercise does not involve following or appealing to rules. Let us consider these elements in turn.
A: Judgment is an ability which persons exercise.
According to the judgment model, ‘judgment’ refers to an ability which people exercise, rather than the products produced by the exercise of that ability. It is important to realize that very often the term is used to refer to the latter rather than the former. For example, when we utter sentences like ‘I judge that it will be more efficient to take the southern rather than the northern route’, or ‘she judged that Jones would be a better Prime Minister than Smith’ – indeed, whenever we use, or utter sentences which presuppose, the expression ‘judge that’ – we refer to the product of the exercise of the ability, rather than the process in which the ability is exercised.
This point will perhaps be more clear if we consider some utterances or sentences which, I think, are unproblematically regarded as judgments. To all of them the expression ‘I judge that’ can be prefixed:
a: ‘Amsterdam and Miami are both culturally heterogeneous.’
b: ‘This stock has performed well in the past, so it is worth buying today’ (Suggested by Govier, p. 129).
c: ‘Physicists will unify quantum mechanics and general relativity by 2020.’
d: ‘The tyranny of evolutionary theory will be overcome in my lifetime’ (spoken by Philip Johnson).
e: ‘Chocolate ice cream tastes better than pecan butter crunch’ (spoken by my daughter).
f: ‘Chocolate stains are harder to remove from white blouses than pecan butter crunch’ (spoken by my wife or me).
g: ‘Nine year old children cannot handle the following assignment: “Write an
essay evaluating your own character, accomplishments, and goals”’ (asserted by Govier, p. 130 [not cited verbatim]).
h: ‘Deliberation, options, and the weighing of pros and cons all enter into judgment’ (asserted by Govier, p. 129 [not cited verbatim]).
i: ‘The classical model of rationality requires a foundational epistemology’ (asserted by Brown, p. 58).
These are all examples of judgments, as the term is often used in ordinary discourse. They are clearly not instances of the exercise of the ability the judgment model points to, but are rather examples of the outcome, or product, of that exercise. For those making them, these judgments are estimates of truth value, or of worthiness of belief. For example, it is my judgment that (a) both Amsterdam and Miami are culturally heterogeneous – I judge it to be true, and as worthy of belief – just as Brown judges it to be true that (i) the classical model of rationality requires a foundational epistemology, and Govier that (g) nine year old children cannot reasonably be thought capable of writing an essay in which they evaluate their own characters, accomplishments, and goals.
Below I’ll raise the question: what makes judgments like these rational, or reasonable(iv)? For the moment I want simply to make it clear that the term ‘judgment’ can be used to refer either to the exercise of the ability, as Brown and Govier urge, or to the product of that exercise. Now this ambiguity is not necessarily a problem for the judgment model; Brown and Govier are certainly within their rights to use the term to refer to the ability, rather than the result of its exercise. But we do need to keep aware of the shift in usage that the judgment model requires. It will be especially relevant below, when I address Govier’s suggestion that my earlier discussions use the term equivocally.
It needs also to be noticed that the classical model appears to be concerned with the rationality of the products of the exercise of judgment, rather than the rationality of the process in which that exercise occurs(v), and in that sense the judgment model might with some justice be thought to be a change of subject, rather than a new account of the subject matter of concern to the classical model.
B: Judgment results in reasonable decisions.
As we saw above in citations from both Brown and Govier, the judgment model seems committed to the view that the exercise of judgment must, apparently by definition, result in reasonable decisions. But this seems problematic.
Notice first that on the account of judgment Brown and Govier endorse – “the ability to evaluate a situation, assess evidence, and come to a reasonable decision without following rules” – judgment is a success term; the exercise of judgment automatically or necessarily results in “reasonable” decisions. All judgments, on this account, are good, i.e. normatively appropriate, judgments. Two questions immediately arise. First, are all judgments necessarily good ones? Don’t/can’t we make bad, e.g. unreasonable or irrational, judgments? Second, when our judgments are indeed good, what is it about them that renders them so? On what basis are good/rational judgments distinguished from bad/irrational ones? I address them in turn.
i. Are judgments necessarily good? Of course proponents of the judgment model can simply stipulate that judgments are always good. In that case we’ll need to introduce a new term for attempted or pseudo-judgments and decisions that appear to be judgments but are not, because they fail to meet the relevant standards and/or criteria of goodness. Let us call all such failures ‘shmudgments.’ Our problem now becomes: how, on the judgment model, are we to distinguish judgments from shmudgments?
On Brown’s view, as we have seen, there are additional constraints which must be satisfied if judgments are to be rational: they must be informed, skilled, and “submitted to the community of competent individuals for evaluation and criticism”(vi). But when these additional constraints are met, judgments will be, on his view, rational:
If the subject is one in which we have the relevant expertise, we gather information, apply whatever rules are available, weigh alternatives, and arrive at a judgement; then we discuss our judgement and the reasons for it with our peers, and re-evaluate that judgement on the basis of their recommendations and critiques. The outcome of this process is a rational decision or belief (p. 226).
As this passage makes clear, on Brown’s account of it, the rationality of judgment is guaranteed, whatever decision is made or outcome reached, so long as the exercise of judgment follows the procedure and meets the constraints that the model imposes. Whatever the content of the judgment reached – e.g., whether or not Amsterdam and Miami are judged to be culturally heterogeneous, whether or not the classical model of rationality is judged to require a foundational epistemology, whether or not nine year olds are judged capable of writing essays in which they assess their characters, accomplishments, and goals – the judgment will be, on Brown’s view, rational. Whatever the outcome of the process – whatever content, conclusion or decision is reached – its rationality is guaranteed.
The view Brown here defends, that the rationality of a judgment is guaranteed by the procedure followed, and is independent of the content of its outcome, faces two important difficulties, which I will mention but not develop in detail.
First, an adequate theory of rationality will perforce declare itself not only on matters of procedure, but also on matters of content. That is, it will recognize the legitimacy and importance of questions concerning the rational status of the conclusions, decisions, and judgments reached through the exercise of our ability to judge. Is the judgment that (a) Amsterdam and Miami are both culturally heterogeneous rational? Suppose that we judge that (¬a) these cities are not both culturally heterogeneous. Would that judgment be rational, too – whatever the evidence? The view that rationality is independent of content is not only deeply problematic; it simply ignores a central concern for any theory of rationality, namely the rationality of the content of judgments.
Second, this independence of content makes a mystery of Brown’s insistence that rationality be a matter of reasons. Why must rational decisions be based on reasons, if not because the rationality of those decisions depend on their content, and basing them on reasons helps to establish the rational status of that content? These problems suggest that the judgment model is incapable of distinguishing rational from irrational judgments – or, to put it more generally, of speaking to the epistemic status of judgments reached by way of the process of judgment Brown articulates – and thus fails to accomplish a task basic for a theory of rationality. Following the procedure and satisfying the constraints Brown recommends, however salutary, cannot be sufficient for rationality, since the rational status of the content of judgments reached by it is left open.
Interestingly, Govier also (albeit indirectly) raises questions about the adequacy of Brown’s ‘automatically rational’ view. She does so by pointing out that judgments admit of normative appraisal, and, far from being automatically rational, can be made well or badly (In the following passages the emphases are added):
A rational person, he [i.e. Brown] proposes, is one who can exercise good sense and good judgment in difficult cases… He or she can… decide and act sensibly in cases where there are no rules (p. 128).
We will fall back on judgment when there are no rules to guide us – when we have to devise or amend rules, choose between rules, handle an unusual case, or decide whether other things are equal. To do this, and do it well, is to be rational (p. 129).
Good judgment… requires sensitivity, good sense… a sense of what is realistic… [and] a sense of proportion, of what is more or less significant…. A person with good judgment will be able to recognize what is relevant, what rules and principles bear on a case, what if anything is exceptional about that case, what the consequences or implications of various decisions are likely to be, and so on and so forth (p. 130).
… judgment can… be reasonable or unreasonable (p. 132).
… people often need good judgment, but have bad judgment (p. 133).
We should try to improve our judgment in whatever ways we can (p. 135).
Govier is, in my view, clearly correct that judgments – whether exercises of ability (processes) or products – admit of normative evaluation in this way. This poses a problem for Brown, since, as we have seen, on his view judgments which satisfy his (content-less) constraints are automatically rational. But I leave this as an in-house dispute for Brown and Govier to resolve as they see fit. Supposing that Govier is correct that judgments admit of normative evaluation, how can/do we evaluate them? Here we come to the second question posed above.
ii. When our judgments are indeed good, what is it about them that renders them so? On what basis are good/rational judgments distinguished from bad/irrational ones?
The answers to these questions seem clear: It is the satisfying of relevant standards or criteria that renders judgments good; we distinguish the good ones from the bad by seeing whether or not candidate judgments meet them. But now a new question, of considerable moment for proponents of the judgment model, arises: Can this be done without appeal to rules or principles? If not, Brown and Govier seem thrown back to the classical view they wish to reject. If so, how? Presumably, the answer will be: by using one’s judgment. But now the problems which plague the classical model, i.e. circularity, regress, or arbitrariness, seem to plague the judgment model as well. These problems might be avoided by simply insisting that we can use our judgment to tell whether a candidate judgment is good or bad, reasonable or unreasonable, a judgment or a shmudgment. But this way of avoiding the problems seems clearly enough either to settle the philosophical issue by stipulative definition, or to beg the question against all those, like Govier, who deny – quite reasonably, it seems to me – that all judgments are good, or rational.
So, the normative evaluation of judgments requires appeal to criteria. Can we so appeal without invoking rules or principles? Here we come to a decisive issue for the judgment model.
C: The exercise of judgment does not involve following or appealing to rules.
Brown and Govier are clearly right that when we judge well, and use our judgment appropriately, we needn’t appeal explicitly to rules or principles. Such explicit appeal cannot be required, on pain of the regress Brown exposes: If I had to appeal explicitly to a rule in order to exercise my judgment (e.g. concerning one of the examples a.-i. above), and I had to appeal explicitly to another rule in order to exercise my judgment concerning, e.g., the appropriateness or proper application of the first rule, I would never be able to exercise my judgment completely or successfully at all. So the judgment model’s insistence that judgment is a skill, that can be exercised well or badly – and which, like most skills, is not properly exercised by explicit attention to rules guiding its proper exercise – is a perfectly correct claim concerning one important sense of ‘judgment.’ For Govier (as for Brown), it is inevitable that the exercise of judgment does not involve following or appealing to rules, because rules “are bound to run out at some point” (p. 125), and when they do, “there is no alternative to judgment” (p. 135). And she is surely right that the exercise of judgment does not require following or appealing to rules, since, as we’ve seen, each attempt to so follow or appeal would itself require the exercise of judgment (concerning, e.g., to which rule to appeal or how to follow it).
But the normative evaluation of the exercise of judgment (and of judgments) is central to Govier’s view, and it requires appeal to a range of criteria, several of which (sensitivity, a sense of what is realistic, a sense of proportion and of what is more or less significant, etc.) she articulates in the passage cited above. If judgment admits of normative evaluation, then its proper exercise, i.e. good judgment, is such because it satisfies relevant criteria (The parallel point applies to the evaluation of judgments (products). In general, judgment’s being exercised well (or badly) is a matter of its satisfying (failing to satisfy) relevant criteria. Moreover, the normative evaluation of either a given exercise of judgment, or the product of that exercise (i.e. the belief, decision or action to which judgment gives rise), as good/bad, inevitably requires appeal to such criteria(vii). If so, can rules be avoided in its proper exercise?
The classical model is wrong insofar as it bans judgment from rationality.
As Brown and Govier insist, judgment is inevitable. Even to determine that a given rule is applicable to a given case, judgment is required. But, as Brown insists, any adequate account of rationality will render it as dependent upon reasons. Whether or not beliefs or judgments are rational is a matter of their meeting relevant criteria. The goodness of reasons, like the goodness of judgment, is also a matter of satisfying relevant criteria (here of epistemic goodness). Thus judgments, even if made without criteria in mind, are, if rational, supportable by reasons which themselves satisfy criteria of epistemic goodness. And conformity with or appeal to reasons, like conformity with or appeal to relevant criteria, requires consistency: if p is a good reason for q in circumstances C, it will be so in all relevantly similar circumstances. (Scheffler, 1989, p. 76) Consider Govier’s own example of a (bad) judgment, i.e. that concerning (g) nine year old children being assigned an essay in which they are asked to evaluate their own characters, accomplishments, and goals. If Govier’s own judgment, that the teacher’s judgment in assigning the essay was bad, is itself good, it will be consistently so – e.g., it will concern nine year old children generally, not just her daughter; will not depend on the characters, accomplishments and goals (or gender or cultural identity, etc.) of any particular child; will not be altered by changes in her (Govier’s) mood or her work schedule; etc. That is, if Govier’s judgment is good, it will be so in virtue of the quality of her reasons for it, and those reasons will in turn be good (i.e. epistemically forceful) only insofar as they can be consistently invoked and evaluated across the range of relevantly similar cases. Consistency, in turn, is a matter of conformity with rules and principles (e.g., ‘treat like cases alike’; ‘no difference in judgments without a relevant difference in reasons for them’; ‘if a is a reason for p and p is relevantly similar to q, then a is a reason for q’; etc.)(viii).
As Brown and Govier insist, I may judge rationally without appealing to rules. But that judgment’s status, as rational, depends upon its satisfying relevant criteria. We needn’t consciously follow rules to be rational, but our judgments must meet criteria in order to be, and to be certifiable as, rational. The normative evaluation of judgment hinges upon the satisfaction of criteria. Rationality is more generally a matter of satisfying criteria, at least insofar as rationality is normative, telling us what is worthy of belief/decision/action. But then judgment – or at least rational judgment – is also a matter of criteria, and so consistency, and so also of rules and principles. In short:
judgment ( normative evaluation ( criteria ( consistency ( rules
And so, the view that judgment is innocent of rules fails. While Brown and Govier are right that particular exercises of judgment do not require explicit or conscious appeals to rules(ix), they are wrong to think that good, i.e. normatively appropriate judgment is so independently of rules. The satisfaction of relevant criteria is fundamental to normative appropriateness, and that satisfaction requires consistency, which in turn requires rules. Exercising judgment might be ‘what we do when rules won’t help,’ but exercising it well is nevertheless a function of reasons, criteria, consistency, and (so) rules.
It is important to note that my claim is that the normative evaluation of judgment requires, in the end, rules. I am not claiming that all instances of rational judgment are, as Govier puts it, “resolvable by appeals to rules” (p. 132). Whether available rules will suffice to resolve particular issues will depend very much on the issues, and rules, involved. There are obviously hard cases – in philosophy, in matters of pressing social concern, and in ordinary, personal life – in which available evidence, criteria and rules are insufficient to secure rational resolution. But resolution is a red herring here. The question is not whether reason can resolve all outstanding issues, but whether rational judgments are such independently of relevant criteria (and so rules). A given issue may well be irresolvable by appeal to rules, as Govier and Brown insist, yet judgments about it be rational in virtue of their being sanctioned by relevant criteria (and associated rules, e.g. concerning the application of those criteria and the evaluation of the relevant evidence). I am content to let resolution fall where it may. It is rather the point about rationality’s dependence on criteria, and so rules, that I am concerned to make.
4. The Place of Judgment in the Theory of Rationality
If the points made above are correct, what should the theory of rationality say about judgment? It should say at least the following:
First, the ‘classical model’ of rationality, as articulated by Brown, is wrong in holding that rationality does not in any way involve judgment. The reason is straightforward and compelling: to tell whether a given rule/principle is applicable and correctly applied to a given belief, which applicability and correct application are essential to the determination of rationality on that model, judgment is required.
But the ‘judgment model’, according to which rationality is a matter of rule/principle-independent judgment, and according to which rules and principles play no role in the determination of the rationality of beliefs, judgments or decisions, is also mistaken. Here, too, the reason is straightforward and compelling: if judgments admit of normative evaluation, as Brown and Govier as well as other theorists of rationality agree, that evaluation depends upon criteria of normative appraisal, which in turn involves rules and/or principles.
So an adequate theory of rationality needs rules and judgment(x). Thus I do not embrace the ‘classical model’ as Brown delineates it, but rather a hybrid model that includes both judgment and rules/principles. On my view (1988, 1997), as on Brown’s(xi), (good) reasons are the key to rationality – as Scheffler puts it, “Rationality… is a matter of reasons” (Scheffler, 1989, p. 62, emphasis in original) – and this in turn requires consistency, which in turn requires rules/principles. This is the part of my view which conforms to the classical model. But I also embrace judgment.
The “unresolved tension” in my view (p. 132) which Govier detects is genuine, and problematic, only insofar as (a) the quality of reasons is a matter of their satisfying epistemic criteria, (b) such satisfaction involves consistency, and so rules, and (c) judgment has nothing to do with rules. But since (c) is false – or at least not a part of my view – the tension Govier detects is actually neither genuine nor problematic. If ‘judgment’ in (c) – and in particular, its normative evaluation – is understood to be totally independent of criteria, then the problem of distinguishing rational from irrational judgment is irresolvable. If understood rather in a way that acknowledges that the quality of judgment is a matter of the satisfaction of criteria, and that the normative evaluation of judgment as good/bad, or reasonable/unreasonable, requires at least implicit appeal to criteria, and so, via consistency, to rules, the ‘tension’ to which Govier points disappears.
Govier also suggests that I equivocate, sometimes using ‘judgment’ as she and Brown do, “to allude to a human ability to make deliberative decisions about cases not resolvable by appeals to rules” (p. 132), and other times using it to refer to the product, or outcome, of the exercise of that ability. Here I plead guilty, but think it a mistake to regard it as a crime. As argued above, ‘judgment’ is used in ordinary discourse in both senses, and an account of judgment must include both. More importantly, the theory of rationality is vitally concerned with the rational status of those products/outcomes/conclusions. To banish that concern is to render the theory epistemically inert. My ‘equivocation’ is thus correctly seen not as a mistake, but rather as an effort to address both senses of ‘judgment’ in a way that retains the theory of rationality’s central concern with the epistemic status of those products. I believe that all the passages Govier cites and discusses in which I use the words ‘judge’ or ‘judgment’ (pp. 126-133) are readily, and unambiguously, so interpreted.
We should think of judgment not as necessarily rule- or criterion-innocent, but rather as (a) the exercise of our capacity to evaluate, assess, consider, and decide, and/or (b) the results of that exercise. In either case, criteria are essential, and so, therefore, are rules. If this is correct, we should also hold that:
a. If we are to distinguish rational from irrational judgment – which is basic for an adequate account of rationality – there is no alternative to doing so by appealing to relevant criteria and/or standards.
b. Since doing so requires appeal to rules or principles, that aspect of the classical model needs to be retained by any adequate view of rationality.
c. The sting of this last point is entirely removed by recognizing that explicit appeal to rules is not required for the exercise of judgment, but rather for its normative evaluation, and by embracing – as all parties (including Govier (p. 127) and Brown (p. 144) to this discussion do – a thoroughgoing fallibilism: all our judgments, including every aspect of our theory of rationality, might be mistaken. Of course that our judgments might be mistaken does not entail that they are in fact, or are likely to be, mistaken. (Brown emphasizes this point as well, pp. 144-6; see Evnine, 2001; Siegel, 1997)
d. There is no need to fear that any view other than the Brown/Govier view must founder due to infinite regress. Fallibilism requires that every belief and judgment be regarded as possibly mistaken; any can be called into question, and reasons can be demanded which purport to render continued belief justified. Vicious regress needn’t follow, because at every point the demand for further reasons can, in principle, be met(xii). This is the way that Brown endeavors to avoid the regress; it is open to theorists who acknowledge that rules have a place in a fallibilistic theory of rationality as well.
e. Finally: If the classical model is correct that the normative evaluation of judgments requires appeal to criteria, and so to rules, it must also be granted that the judgment model is also correct: as far as rationality is concerned, judgment is also required. In so far, we needn’t reject either view; the task, rather, is to integrate them in an adequate account of rationality. I hope to have taken some small steps in that direction here(xiii).
i. As well as the person’s recognition of that relation (Brown, 1988, p. 19). (All page references to Brown in the text are to his (1988).)
ii. For extended discussion, see Siegel, 1988, 1997; particularly explicit statements by Laudan and Scheffler are cited at the head of 1988, p. 32. It is endorsed by Rescher (1988) and Nozick (1993) as well, both of whom would I think be considered by Brown to be advocates of the classical model.
iii. All page references to Govier in the text are to her (1999). I should note that Brown spells ‘judgement’ with two occurrences of the letter ‘e’, while Govier spells it with one occurrence of ‘e’. I will quote them accurately, but otherwise use the shorter spelling (According to the Oxford English Dictionary both spellings are legitimate).
iv. Sometimes these terms are treated as synonymous, sometimes not. I treat them as equivalent here, since Brown and Govier seem to in their texts. For a more detailed consideration of them, see Siegel, 1997, chapter 7.
v. As we saw above, on Brown’s view, the classical model holds that “the rationality of any conclusion is determined by whether it conforms to the appropriate rules.” (17)
vi. Notice that satisfying these additional constraints is also a matter of meeting relevant criteria.
vii. Brown writes: “Our model requires that rational beliefs be based on judgement, and judgement requires assessment of evidence and arguments.” (p. 192) But surely what is required is (not ‘mere’ assessment, but) competent assessment.
viii. Consistency in application of relevant empirical generalizations, e.g. concerning the cognitive and emotional abilities of nine year old children, is also necessary.
ix. But Govier is wrong to attribute to me the contrary view, p. 126. On my view critical thinking/rationality is fundamentally a matter of evaluation in accordance with reasons. Rules and/or principles may be invoked when the reasons themselves become the object of critical scrutiny, but this is not itself inevitable.
x. And so the Brown/Govier definition of ‘judgment’ as something wholly innocent of rules forces their arguments against the classical model to rest on a false ‘judgment/rule’ dichotomy.
xi. “[R]ationality provides reasons for accepting claims, i.e., it provides grounds for considering propositions to be worthy of belief and for acting on decisions.” (1988, p. 226, emphasis Brown’s)
xii. As Brown agrees (p. 186).
xiii. An earlier version of this paper was presented at Fresno City College in April 2002; I am grateful to Robert Boyd for arranging that presentation, and to the audience on that occasion, and in particular to Boyd and Otávio Bueno, for their very helpful criticisms and suggestions. I want also to thank Trudy Govier for her close attention to my work, and Hal Brown for extensive correspondence on these matters when he kindly sent me draft chapters of what became his (1988). Many of the points I made in that correspondence are made here, but it was only when I read Govier’s (1999) that I decided to pursue these issues in print.
Brown, H. (1988). Rationality. London: Routledge.
Evnine, S. (2001). Learning from one’s mistakes: Epistemic modesty and the nature of belief. Pacific Philosophical Quarterly 82, pp. 157-177.
Govier, T. (1999). Rosebuds, judgment and critical thinking. In: T. Govier, The philosophy of argument (pp. 123-136). Newport News: Vale Press.
Nozick, R. (1993). The nature of rationality. Princeton: Princeton University Press.
Rescher, N. (1988). Rationality: A philosophical inquiry into the nature and the rationale of reason. Oxford: Clarendon Press of Oxford University Press.
Scheffler, I. (1989). Reason and teaching. Indianapolis: Hackett Publishing Company. First published in 1973 by Routledge and Kegan Paul.
Siegel, H. (1988). Educating reason: Rationality, critical thinking, and education. London: Routledge.
Siegel, H. (1997). Rationality redeemed? Further dialogues on an educational ideal. New York: Routledge.