The Sceptical Challenge

Ruth Weintraub

• Publisher: Routledge
• Year: 2012
• eBook ISBN: 9781134770878
• Topic: Philosophy
• Subtopic: Philosophy History & Theory

Do we really know the things we think we know? Are any of our beliefs reasonable? Scepticism gives a pessimistic reply to these important epistemological questions - we don't know anything; none of our beliefs are reasonable. But can such a seemingly paradoxical claim be more than an intellectual curiousity? And if it is, can it be refuted? Ruth Weintraub answers yes to both these questions.
The sceptical challenge is a formidable one, and should be confronted, not dismissed. The theoretical and practical difficulties it presents - in that the sceptical life cannot be lived, and the doctrine seems self-defeating - are in fact superficial, according to Ruth Weintraub. Her study looks at the sceptical arguments of Descartes, Hume and the ancient Greek sceptic, Sextus Empiricus. The author argues that by drawing on philosophy, rather than science, the sceptical challenge can be answered.
The Sceptical Challenge is a bold and original response to scepticism; it represents a new way of looking at the field for philosophers of epistemology.

1

KNOWLEDGE, JUSTIFICATION AND TRUTH

INTRODUCTION

Scepticism can be more or less radical, or comprehensive, depending on the set of beliefs upon which it is directed. But it can also vary with respect to the kind of doubt it casts. One sceptic questions the justifiability (or rationality) of our beliefs, and another doubts our knowledge claims. In this chapter I shall argue that we should focus on scepticism about justification and need not separately address scepticism about knowledge. The argument proceeds in the following way. I first clarify the notion of justification which is most pertinent to our concerns, and explain its epistemic significance. Then, I consider various ways of interpreting knowledge attributions, and argue that even when very modestly construed, knowledge claims are undermined by justification scepticism, and, furthermore, that we needn't be perturbed by knowledge scepticism which leaves intact justification claims. Just fication, rather than knowledge, I shall conclude, is the central epistemological notion.

THE CONCEPT OF EPISTEMIC JUSTIFICATION

Our beliefs are the product of a complex causal process. The term ‘justified’ (and its synonym ‘rational’) is used to evaluate certain aspects of this process. This epistemic appraisal is made relative to one particular desideratum: maximal truth- and minimal falsity-content in our beliefs. To use Popper's (1972) terminology, we wish to attain a maximal degree of verisimilitude, our regulative ideal being ‘the truth, the whole truth, and nothing but the truth’. With respect to this ideal, belief can be more or less successful, and in this it is akin to action. But there are important differences. Most obviously, belief is less (if at all) voluntary.

An objection to taking verisimilitude to be our cognitive objective should here be preempted. It must be conceded that attempts to define verisimilitude have not been successful. But this, far from impugning the concept, shows it to be pretheoretically meaningful. If we did not have intuitions about verisimilitude, we couldn't cite counter-examples to proposed definitions. We need not wait for an adequate analysis, just as we do not eschew the concept of causation pending its analysis. One may not even be forthcoming (for verisimilitude or for causation): why think that every natural language expression can be defined?

There are other standards by reference to which belief can be assessed, most notably the practical one. There is some overlap between the practical and the cognitive value of a belief: plausibly (pace Stitch, 1990, ch. 5), our actions are (typically) more successful when guided by true beliefs. But some beliefs which hardly affect action may be theoretically important: we are interested in understanding reality, not just in controlling it. Furthermore, belief has consequences in addition to the actions it motivates. In terms of these consequences, a belief of greater verisimilitude may be less beneficial, as the following two examples show.

First, a terminally ill person may reasonably want to have a false belief about his situation: some beliefs are just too awful to have. Second, if persuaded by Pascal's argument, one will have a reason for believing in the existence of God, even if his existence is very unlikely. In both cases, the goal of the agent in forming his beliefs isn't cognitive (increasing truth and avoiding falsity). He aims to have beliefs which are—although false—practically beneficial.

I should make a terminological point here. One could restrict the use of the word ‘justified’ to express a purely epistemic evaluation, and use the word ‘rational’ more widely, to incorporate other considerations (practical, for instance). But my concern is epistemological, and I shall not be saying much about practical rationality. For stylistic reasons, therefore (so as to avoid repeating the word ‘justified’), I use the words ‘rational’ and ‘justified’ interchangeably and narrowly: ‘justified’ (‘rational’) should be read as ‘epistemically justified’.

Justification is most paradigmatically concerned with the inferential aspect of belief-formation. Rationality, after all, has been defined as ‘the use of reason or logic in thinking out a problem’ (Collins English dictionary). But this does not mean that logical omniscience is required for rationality. Economists and philosophers who use the term ‘boundedly rational’ to characterize (human) agents who are not logically omniscient are departing from ordinary usage. We do not think ourselves irrational in failing to believe in Goldbach's conjecture even if it is a logical consequence of axioms (Peano's) which we accept. This is best explained, I believe, by the supposition that judgements of rationality are relativized to agents' cognitive capacities. It is the agent's reason, his intellectual potential for making inferences, which constitutes the standard relative to which assessments of rationality are made. If we ignore this relativization, we may be tempted to think of rationality as a skill (Reynolds, 1991). In fact, a rational person is more aptly likened to one who properly uses his (possibly mediocre) intellectual skill.

There is a difference between irrationality and stupidity. A logical wizard who hasn't drawn a logical conclusion from the evidence is failing to use his reason, and is properly branded irrational. The label ‘irrational’, on the other hand, should not be attached to someone who hasn't drawn the conclusion because the inference is beyond his logical capacity. Ceteris paribus, a consum-mate logician is required—if he is to be justified—to display greater logical coherence than the average person. He is blamed for harbouring ‘remote’ inconsistencies, for instance.

I have argued that judgements of rationality must be relativized to agents' (logical) capacities. But this claim engenders the following difficulty. Can one fail to use one's capacity to reason? True, psychologists suggest that people reason fallaciously even when they are capable of assessing the evidence correctly (Nisbett and Ross, 1980). But it might be objected that the only evidence we have for a person's reasoning ability is the way he actually reasons. Properly understood, the claim is correct but doesn't constitute an objection. We may grant the absurdity of ascribing a capacity the exercise of which is always curtailed. But when a musician (or an athlete) is ‘off form’, the quality of his performance falls short of the standard he can attain; a standard which is displayed on other occasions. In reasoning, admittedly, some fallacies are perpetrated by the same agent regularly, but they can still be characterized as lapses from a capacity (rather than cases of incompetence), whose ascription we can substantiate by appealing to inferences of equal complexity which the agent draws successfully. Complexity can be characterized formally, independently of agents' inferential performance.

I now come to an additional feature of justification. To be rational, a belief should be properly caused. This consideration can be wielded in two kinds of case. First, an agent who believes something as a result of reading his horoscope isn't justified in his belief even if it is entailed by something else he believes: his belief is caused by something which isn't evidence for it. The second kind of case is that of a belief which is formed under pressure from some affective drive (wishful thinking, ideological bias, etc.).

THE CIRCLE OF BELIEF

The conception of justification which I have articulated contravenes the popular thought that ‘[t]here is no exit from the circle of one's beliefs’ (Lehrer, 1974, p. 188). This vague intuition I shall construe as the claim that justification supervenes on the totality of one's beliefs: two agents with identical beliefs will be in-distinguishable with respect to which of their beliefs are justified and to what degree. More formally, a property, P, supervenes on a set of properties, X, if and only if for every y, if y is P, there are properties Q1,… in X such that y has Q1…, and, necessarily, whatever has Q1,… has P.

How is it that we can exit the circle of belief? If rationality depends not only on what the agent believes, but on how his beliefs are caused, then two doxastically identical agents may be justified to different extents. For instance, an agent who has repressed from consciousness all evidence against a proposition he wishes to believe has subverted reason so perfectly that detection from ‘within’ has been rendered impossible. His beliefs are evidentially coherent, but he isn't justified in them.

Bonjour (1985, pp. 150–1) would disagree. He points out that we judge justified the person who is deluded by an evil demon or a mad scientist. And the deluded person is analogous, Bonjour thinks, to the person who has repressed evidence. The fact that ‘the mechanism which produces the beliefs is in some sense part of the person rather than an outside force’, Bonjour claims, is irrelevant. ‘[In] relation to his overt cognitive processes his unconscious is an outside force.’

There seems, in fact, to be a world of difference between the two cases. It is the agent, and not his conscious mind, that we are assessing. Epistemic rationality requires that one be motivated unitarily, one's goal being verisimilitude. And the wishful believer, the coherence of his (conscious) belief system notwithstanding, is being motivated by a competing goal. We needn't condemn him for this: truth-related reasons do not exhaust the field. But a belief which is motivated by other considerations just isn't epistemically justified, even if it is—everything considered—reasonable. Note that this is only a necessary condition. That an agent forms a belief with a view to improving verisimilitude isn't sufficient for justification. If a student lacking in self-confidence brings himself to believe he is capable of great things so as to overcome his anxiety and improve his capacity to learn, his belief is not (epistemically) justified (Smullyan, 1980, pp. 131–2). To be epistemically justified, a belief should be formed with a view to its own verisimilitude.

We can cite a second reason for thinking that justification does not supervene on belief; that irrationality, reason malfunctioning, needn't be discernible from the agent's beliefs. Perhaps below a certain threshold of logical competence the epistemic concepts ‘justified’, ‘rational’, etc., are not applicable. But above that threshold, appraisals of justification are sensitive not just to what the agent believes, but also to what he is inferentially capable of believing (see the previous section). And this means that two doxastically identical agents may be differentially justified. A fairly subtle inconsistency within a belief system may count as irrational in the case of the logical wizard while being perfectly justified in the case of a moderately intelligent person.

WHY DOES JUSTIFICATION MATTER?

Why should we care about being rational if our aim is truth (verisimilitude), for which justification is neither a necessary nor a sufficient condition? Well, justified beliefs are at least more likely (objectively) to be true than unjustified ones. This link between truth and justification has been contested. In defending induction, Goodman (1973, p. 62) has argued that we do not ‘know that [our] predictions will turn out to be correct’, nor even that they will turn out to be correct more often than random guesses. We are, nonetheless, justified in holding them, because they conform to our inductive practice.

Goodman's concession is fatal. Perhaps we cannot persuade the sceptic that our practice is truth-conducive, rather than a mere convention, but we believe that it furthers our cognitive aims. It would otherwise be arbitrary, and rightly denigrated by the sceptic.

That justification is truth-conducive doesn't quite explain its peculiar significance in epistemology. Rational people may do better vis-à-vis the truth, but so do people with good eyes. So what is special about justification? It may be thought that justification is important as a practical aid to belief-formation. Bonjour, thus, suggests that justification is a

means to truth, a more directly attainable mediating link between our subjective starting point and our objective goal. We cannot, in most cases at least, bring it about directly that our beliefs are true, but we can presumably bring it about directly (though perhaps only in the long run) that they are epistemically justified.

(Bonjour, 1985, p. 7, original italics)

We should presume no such thing. We do not have a privileged epistemic access to justification. It is quite easy for me, standing outdoors, to discover that your belief in impending rain is false, and rather more difficult to glean that it is irrational: caused by some wish, rather than being based on an erroneous weather report. The same is true in my own case. The causal psychological facts on which the rationality of a belief depend may be more difficult to discover than the state of affairs upon which it is directed.

Justification also depends on logical (deductive and inductive) relations between beliefs. But even if these are a priori, they can be more difficult to discover than the truth-value of the (empirical) propositions they relate. Were Galileo's opponents irrational and bigoted? Or was theirs a ‘well-defined, sophisticated and empirically successful system [and Galileo's]…an unfinished and absurd hypothesis’ (Feyerabend, 1975, p. 153)? This remains a controversial question even when the truth of the theory is no longer in doubt. The extent to which a scientific theory is justified depends on its simplicity, the variety of evidence in its favour, the plausibility of auxiliary theories jointly with which it gives rise to observable consequences, and its (perceived) fruitfulness in solving new puzzles (Kuhn, 1970). The assessment of these is highly susceptible to error.

One might suppose that our epistemic access to justification is privileged at least in the following modest sense. Isn't the belief in p justified, one may wonder, if the belief in ‘p is justified’ is? With respect to justification, it could be argued, there can be no gap between justifiably thinking oneself justified and actually being justified. This might be one instance where truth and justification go together.

In fact, they don't. If the belief in p is caused by a wish, the belief in ‘p is justified’ typically is as well (the same wish, in fact). But this is a psychological fact about wishful thinking (on which depends its efficacy). It need not always be the case that wishful thinking causes both beliefs. It is logically possible for an agent irrationally (wishfully) to believe p while rationally believing ‘p is justified’: judgements of rationality (those about oneself included), dependent as they are on subterranean psychological facts, are subject to rational error.

Justification isn't ‘a path to truth’ (Bonjour, 1985, p. 8) in the sense of being more easily recognized. So where does its significance lie? Well, from the first-person perspective justification and truth are very intimately linked. I cannot (rationally) judge false a proposition my belief in which I judge would be justified. In the case of someone else this possibility is quite straightforward: my two judgements (concerning the truth and rationality of his belief) are made in the light of something I know and he doesn't. But if it is myself I am assessing, then anything I know which undermines the truth of a proposition ipso facto undermines the rationality of my belief in it. The converse is also true: when I judge that my belief is unjustified, I ipso facto judge that my confidence is misplaced. This is not true when the judgements are made about someone else, where I may have information he doesn't.

Although justification isn't particularly easy to detect, it nonetheless has a regulative role with respect to belief modification and retention. To characterize one's belief as justified is to approve of it: to mark it as one which one ought to hold on to (or adopt) if one's aim is verisimilitude. Belief-formation goes by justification because, roughly, that is our best (cognitive) strategy.

These considerations suggest that we shouldn't be perturbed by Unger's (1975, p. 199) suggestion that ‘there is at most very little in which one will be justified or reasonable’. According to Unger (1971), justification requires a logical guarantee of truth, and that is why we are so seldom ‘j...

Table of contents

• Cover
• Half Title
• INTERNATIONAL LIBRARY OF PHILOSOPHY
• Full Title
• Copyright
• Dedication
• CONTENTS
• Acknowledgements
• Abbreviations
• Introduction
• 1 KNOWLEDGE, JUSTIFICATION AND TRUTH
• 2 THE SCEPTICAL LIFE
• 3 THE CHALLENGE
• 4 THE RESPONSE
• 5 DESCARTES' SCEPTICAL CHALLENGE
• 6 INDUCTIVE SCEPTICISM
• 7 SCEPTICISM AND THE STRUCTURE OF JUSTIFICATION
• 8 INDUCTIVE SCEPTICISM REVISITED
• 9 TRANSCENDENT SCEPTICISM AND INFERENCE TO THE BEST EXPLANATION
• 10 THE DEMON ARGUMENT REVISITED
• 11 THE DREAM ARGUMENT REVISITED
• Notes
• Bibliography
• Index