Most philosophers accept that experimental findings can in principle bear on philosophical questions. Experimental confirmation of the theory of special relativity is relevant to the philosophy of time, as are results from experimental psychology to the philosophy of perception. The usual pattern is that experimental findings about X bear on the philosophy of X by showing something about X in real or apparent conflict with assumptions on which philosophers of X had relied. The movement calling itself ‘Experimental Philosophy’ does not fit that pattern. The experiments it promotes as bearing on the philosophy of X are not directly about X. Rather, they are about what people say about X. The people surveyed are typically not experts on X. Their reactions are supposed to bear on the philosophy of X by showing something about the everyday concept ‘X’, perhaps that philosophers had misunderstood its structure, or neglected its variation from one social group to another. Reconsidering the concept ‘X’ is supposed to lead philosophers indirectly to reconsider X itself. In the philosophy of time, such a method might involve, not physicists' experiments about time, but statistical surveys (carried out by the philosophers themselves) of ordinary people's verdicts on examples from the philosophy of time, and even (if funding permits) scans of their brains as they react.
Advocates present Experimental Philosophy as a revolution in philosophy, starting about the new millennium. But the approach is older than they suggest. During the heyday of ordinary language philosophy in the 1950s, the complaint was sometimes heard that if philosophers want to discuss the ordinary uses of words, they should go out and see how ordinary people really use those words. The Norwegian philosopher Arne Næss put an early version of the approach into practice in the 1930s with questionnaires about truth. His results had very little impact. His approach was generally treated as eccentric, although quite what entitled ordinary language philosophers to dismiss it was not altogether clear. The current philosophical climate is more favourable; an atmosphere of excitement surrounds Experimental Philosophy.
Joshua Alexander, an activist in the movement, has written a clear and accessible introduction to its work. The reader obtains a good sense of what Experimental Philosophers are doing and why. The impression is conveyed that the revolution will triumph eventually, though not without an arduous struggle. Much of the book is devoted to showing how tricky the data are to interpret and to confronting critics of Experimental Philosophy (including this reviewer).
The starting-point of Alexander's account is that ‘philosophical intuitions’, glossed as ‘what we would say or how things seem to us’ (1, 101), play a central evidential role in much contemporary philosophy. They constitute data against which theories are measured. He discusses contrasting accounts of the nature of intuitions without settling on any, but continues on the basis that we can recognize when they are being invoked. If philosophers are indeed using what we would say or how things seem to us as data, then they should be careful to find out what we really would say or how things really do seem to us, and large-scale surveys are relevant to that enterprise. The first person plural pronouns are a clue. If I am relying on what we would say or how things seem to us, what entitles me to assume without asking that what the rest of us would say is what I would say, or that how things seem to the rest of us is how they seem to me? Anyway, who are ‘we’? If we include people who differ from me in race, gender, culture, or education, isn't the assumption of uniformity in the data hopelessly shaky? If we include only those who resemble me in all such respects, what value has a philosophical theory based on so narrow a range of data? The stage is set for Experimental Philosophy. Alexander illustrates the range of roles it can play in alerting us to unexpected and theoretically suggestive complexities and variations in responses, using case studies of its application in epistemology, ethics, and the philosophy of mind. His general moral is that we are not yet in a position to draw firm conclusions: more Experimental Philosophy is needed because the data so far do not point unequivocally to a single explanation.
Alexander's case studies are detailed and readable. Unfortunately, they can mislead in philosophically crucial respects. For instance, he discusses the recent debate between contextualists and subject-sensitive invariantists in epistemology (36–48). He presents contextualism as the view that the standard for knowledge varies with how salient possibilities of error are, and subject-sensitive invariantism as the view that it varies with how much is at stake. Readers unfamiliar with the debate will not realize that the fundamental difference between contextualism and subject-sensitive invariantism has nothing to do with either salience or stakes, but concerns instead the difference between the context of the ascriber of ‘knowledge’ and the context of the subject to whom it is ascribed: according to contextualism, the truth-value of the ascription is sensitive to the ascriber's context even when the subject's context is held fixed; according to subject-sensitive invariantism, the truth-value of the ascription is sensitive not to the ascriber's context but to features of the subject's context traditionally thought to be epistemically irrelevant. Thus to hold the truth-value of ‘knowledge’-ascriptions sensitive to how much is at stake for the ascriber but not to salience is to be a contextualist rather than a subject-sensitive invariantist, while to hold it sensitive to how salient error possibilities are for the subject but not to stakes or the ascriber's context is to be a subject-sensitive invariantist rather than a contextualist. In an endnote, Alexander acknowledges an over-simplification in his statement of the issue, but not the crucial one; he says that the debate is really ‘about whether or not salience matters’ (n.19, 119). His discussion shows no care in distinguishing between the ascriber's context and the subject's. In several of his examples, error possibilities are salient in both contexts, so the experiments are not properly controlled. It is hard to design a good experiment to test a theory if you do not pay attention to what the theory says.
Another chapter illustrates the use of experimental methods to distinguish between conceptual competence and conceptual performance with examples where the application of the concept ‘intentional action’ is apparently sensitive to normative judgments about the agent. Alexander concludes that survey methods are of little help in determining whether the influence of some factor in an application of the concept was part of conceptual competence or just of conceptual performance. ‘What is needed instead, for example’, he proposes, ‘are neuroanatomical accounts of the cognitive processes and mechanisms responsible for our folk psychological judgments and evolutionary (or other teleological) accounts of the work that our folk concepts are supposed to be doing’ (69). However, those accounts will be of limited use in determining the bounds of conceptual competence if ‘conceptual competence’ is ill-defined. Alexander explains it thus: ‘The central idea is that certain factors (e.g., resource limitations or interference from other cognitive processes) can influence a person's use of a given concept without influencing her knowledge of that concept or being reflective of the meaning of that concept’ (60). Suppose that on seeing my son spill his milk I judge that he did it intentionally. Presumably, my visual perception is not itself part of my competence with the concept ‘intentional action’, since it involves another cognitive process, even though the latter's role is hardly just ‘interference’. Rather, my conceptual competence is meant to be something I bring to perception. Is it some mechanism for determining whether what I perceive is an intentional action? Such accounts are suspiciously verificationist. The mechanism will be less than 100% accurate. I could change it without changing the meaning of ‘intentional action’. Or does my conceptual competence consist in a list of analytic truths about intentional action written in my brain? The analytic-synthetic distinction is notoriously problematic and ambiguous. ‘Conceptual competence’ as Alexander presents it is so unclear that trying to determine its parts through neuroanatomical or evolutionary investigations would be a waste of time. It is hard to design a good experiment to test an ill-defined theory.
If some Experimental Philosophers neglect traditional philosophical skills in their enthusiasm for experimental methods, that is not an essential feature of Experimental Philosophy. In principle, it could disappear as the movement matures – although the divisions between experimentalists and theoreticians in natural science hint that some trade-off between the two sorts of skill is not easily avoided. A more central worry is that the project of Experimental Philosophy, as characterized by Alexander, does not withstand scrutiny.
What the Experimental Philosophy revolution is supposed to change – systematize, restrict, or abolish – is a philosophical method: the use of philosophical intuitions as evidence. Alexander's starting-point is that such a method is obviously widespread in, and distinctive of, contemporary philosophy (1, 11). The systematic deployment of elaborate hypothetical cases is indeed an eye-catching feature of much recent analytic work. But what Experimental Philosophers target is neither the systematicity nor the elaboration. Nor, officially, is it the hypothetical nature of the cases. For many of them can be replaced by real life cases. For example, hypothetical Gettier cases are famously used to refute the equation of knowledge with justified true belief. Experimental Philosophers have argued that verdicts about such cases are too culturally variable to carry weight in epistemology (although recent work by Jennifer Nagel has cast doubt on the robustness of their results). But there are also real-life Gettier cases; stopped clocks sometimes really do show the right time. Epistemologists can easily use them instead to make the same point. Unsurprisingly, that does not satisfy Experimental Philosophers. They insist that verdicts on real-life cases involve philosophical intuitions just as much as do verdicts on hypothetical cases. If so, the judgment ‘He does not know that it is noon’ may involve a philosophical intuition just as much as does the judgment ‘In the hypothetical case, the agent does not know that it is noon’. But ‘He does not know that it is noon’ is not distinctively philosophical; it is a judgment in ordinary language for which there is perceptual evidence: you see him at noon setting his watch by the stopped clock.
Although philosophical intuitions are often treated as non-inferential, one must be careful about the relevant sense of ‘inferential’. Just as the judgment ‘He does not know that it is noon’ somehow derives from information such as ‘He is relying on a stopped clock’, so the judgment ‘In the hypothetical case, the agent does not know that it is noon’ somehow derives from information such as ‘In the hypothetical case, the agent is relying on a stopped clock’. Although one normally reaches such verdicts without conscious deductive or inductive argument, the same applies to vast numbers of unproblematic judgments in natural science and everyday life, including many non-perceptual judgments. For instance, conscious deductive or inductive argument is not how scientists usually make their overall judgments as to which of several rival theories is best confirmed by a mixed body of evidence. I doubt that the reader used conscious deductive or inductive argument to reach the reasonable belief that there are no hobgoblins (I certainly didn't). Few of the statements made in Alexander's book appear to be based on conscious deductive or inductive argument. The sense in which verdicts on cases in philosophy are non-inferential covers far too much to characterize a distinctive philosophical method.
The criterion that the cases be philosophically significant is equally unhelpful. Any judgment whatsoever has potential philosophical interest, because it will be inconsistent with some whacky philosophical theories.
If the judgments that involve philosophical intuitions are in ordinary language, concern examples, and are not based on conscious deductive or inductive argument, so are vast numbers of uncontroversially unproblematic everyday and scientific judgments, including many made by Experimental Philosophers themselves. Who imagines that philosophy would be improved by a ban on examples, or an insistence that it be conducted entirely in technical jargon? Nor can all judgments in philosophy be based on conscious deductive or inductive argument; some premises are needed too.
Alexander's own gloss on ‘philosophical intuitions’, ‘what we would say or how things seem to us’, does no better. In both everyday and scientific situations, when I say that P, I would say that P, and (if I am sincere) it seems to me that P. If I am not idiosyncratic, we would say that P, and it seems to us that P. If I believe that I am not idiosyncratic, I believe that what we would say, and how things seem to us, is that P.
Alexander proceeds on the assumption that, even if he cannot define ‘philosophical intuition’, we can recognize when one is being used as evidence, and thereby demarcate a tractably narrow class of cases. But although we may indeed be able to recognize when a judgment about an explicit hypothetical case is being used as evidence, that is not the point. As just seen, under dialectical pressure Experimental Philosophers have applied the term ‘philosophical intuition’ so broadly that it fails to capture anything useful. If Experimental Philosophers want to put their activities on a proper scientific basis, they would do well to drop misleading terms like ‘philosophical intuition’, and face up to their failure to identify any distinctive philosophical method to be transformed or overturned by their revolution.
Eliminating all the parts of Alexander's book that depend on talk of ‘philosophical intuitions’ does not leave nothing. His case studies need not be described in such terms. If the experiments have been properly designed and conducted, they still reveal unexpected and intriguing patterns in ordinary human judgments about philosophically central matters such as knowledge, intentional action, causation, and morality. What to make of those patterns is unclear, but simply ignoring them would be imprudent and incurious. Even if they turn out to result from various kinds of bias, a philosophical training is unlikely to render one entirely immune to such bias. Understanding a source of bias is a step towards correcting for it. To be credible, the diagnosis of bias should be backed by an appropriate psychological theory, independently tested on philosophically uncontentious evidence. Such a theory is more likely to come from cognitive psychology than from Experimental Philosophy, and to rely on experiments designed and conducted by psychologists with the requisite experience and know-how. The methodological moral for philosophers to draw will concern ways of correcting for bias of the psychologically identified kind. It will not be to avoid or reduce reliance on ‘philosophical intuitions’, because that term does not pick out any specific psychological kind.
Alexander's reliance on the ideology of ‘philosophical intuition’ is not the only problematic aspect of his discussion of the epistemology of philosophy. His arguments tend to instantiate all-purpose sceptical forms. His comments on evidence are an example. Consider a philosopher who uses a Gettier case to refute the theory that justified true belief is knowledge. The key premise of her refutation is that the subject does not know. One might suppose that she is using that premise as evidence against the theory. But Alexander claims that further evidence is needed for the premise. His point is that further evidence is needed to persuade those whom the premise does not initially persuade (104–7). But such an argument could be formulated about any premise whatsoever. The upshot would be that no premise is ever good enough to use as evidence, because further evidence would always be needed to persuade those whom it did not initially persuade. Although Alexander naturally does not endorse or even consider that conclusion, he gives no reason why his form of argument should work in the case to which he applies it but not in general.
When one puts forward a premise p, one usually hopes that one's audience will accept p. If they do not, one may use whatever further evidence one expects to carry weight with them. One may say ‘The Pope believes p’ or ‘Daniel Dennett believes p’, depending on circumstances. One may emphasize anything the challenger believes that favours p. But one is unlikely to get far by saying ‘I philosophically intuit p’ or ‘We philosophically intuit p’. Appeals to further evidence typically reflect the dialectical needs and opportunities of the moment rather than revealing the standing basis on which one believes p. Sometimes, when you fail to persuade your interlocutor, that reflects shortcomings on his part rather than on the part of your evidence. Alexander does not engage with the task of showing that it is otherwise when a philosopher uses a Gettier case as evidence.
Alexander also falls back on generic sceptical arguments in his discussion of philosophical expertise. For example, he sees an explanatory regress in the idea that students train their judgments against those of established experts, and worries that various kinds of bias may make us overconfident of the value of our training (94). Those are sceptical possibilities for any academic discipline. Alexander provides no serious evidence that they are more urgent in philosophy than elsewhere. To the suggestion that theoretical reflection may improve philosophical judgment, he responds ‘theoretical commitments are just as likely to contaminate as they are to clarify’. The next sentence backs off from the unsubstantiated ‘just as likely’ to a mere claim of possibility: ‘The fact that expert philosophical intuitions are theoretically informed doesn't ensure that they are more theoretically valuable than folk intuitions’ (95). Indeed; but in no academic discipline does the fact that expert judgments are theoretically informed ensure that they are more theoretically valuable than lay judgments. Alexander also objects that ‘If our theoretical commitments shape our philosophical intuitions, it is hard to see how our philosophical intuitions can help us independently assess the accuracy of those theories’ (ibid.). That is just a version of the old concern, familiar from the philosophy of science, that the theory-ladenness of observation undermines the rationality of theory testing. But a theory can get into trouble on its own terms; its influence on our observations does not entail that they will always favour it. Anyway, the theory under test need not be one of those with which our observations are laden. Alexander applies such all-purpose concerns about expertise to argue that we need Experimental Philosophy to test whether the purported experts in philosophy are any good. He does not explain why, if at all, such testing of expertise is more urgent in philosophy than in other disciplines.
Confronted with Alexander's restriction of generic forms of scepticism to philosophy, it is hard not to suspect some anti-philosophical bias. ‘Calls for change’, he reminds the reader, ‘are often met with resistance, especially by those heavily invested in the orthodox’ (89). But those heavily invested in revolution have their own reasons for resisting their opponents. Fortunately, Experimental Philosophy shows signs of outgrowing the sort of polemical philosophy-hating philosophizing from which it has not been entirely free. The best work reported in this book does not reflect any such anti-philosophical agenda.
Experimental Philosophers did not invent the idea of ‘philosophical intuition’. It belonged to the ideology of one faction of the ancien régime. Against that faction, their use of it was dialectically legitimate. For constructive purposes, however, it has outlived its utility. The psychological and sociological study of philosophy will make more progress once it ceases to work within a framework of obsolescent epistemology.
