Herman Cappelen and Douglas G. Winblad
If anything unifies different philosophical methodologies it's some sort of reliance on intuitions. It's remarkable, therefore, how rarely we attempt to justify their employment in philosophy.
The intuitions philosophers care about are typically judgements about whether specific (hypothetical or actual) cases are cases of a certain kind. Some philosophical topic such as reference, knowledge or personal identity is under investigation. A theory is proposed and is then tested against our intuitions about specific cases that bear on the topic. In general, if our intuitions contradict what a theory implies about whether, say, S refers to x, or knows that p, or is identical to T, this counts against the theory. If on the other hand, our intuitions match what a theory tells us about particular cases, this usually counts in favor of the theory.
All procedures of this sort rest on a principle like I:
This paper is about whether I can be justified. We examine two models, the Concepts Model (CM) and the Belief Model (BM). In our view, neither of them provides a solid foundation for I as it is traditionally applied in philosophy.
As we are construing them here, concepts are supposed to be intermediaries between words and the world; they're that in virtue of which at least some subsentential expressions have the references or extensions they have, or would have, if they had any at all.
According to CM, an epistemologist evaluates candidate specifications of necessary and sufficient conditions for knowing by determining whether they're compatible with her intuitions about whether specific cases are instances of knowing. This procedure is legitimate because her intuitions are reliably guided by the tacitly known concept of knowing.
Criticism of CM
The plausibility of CM will depend upon what one takes concepts to be. We'll argue that whether one chooses to embrace an internalist or an externalist construal of concepts, serious difficulties arise when one tries to justify I with CM. Externalism and internalism can both be local or global. Global externalism has it that every term's extension is dependent on environmental factors; global internalism tells us that every term's extension is determined by a sense that's independent of such factors. The intermediate positions are local externalism and local internalism. There can be different forms of both, distinguished from one another on the basis of which expressions are held to express concepts that need to be understood in an externalist, and which in an internalist way. We'll first examine the global externalist version of CM, and local externalist versions that include philosophical concepts in the class of concepts that are supposed to be understood in externalist terms (what we call "localphil externalism"). We'll discuss two externalist theories, Kripkes and Putnams. Burge's view can be treated similarly.
According to Kripke, (1) names and natural kind terms have no senses. Senses do not serve as intermediaries between these terms and the world, which are related, instead, by initial "baptisms" which are causally linked to subsequent uses of the terms. Since concepts, as understood here, are at the level of senses, Kripkean externalism says that certain terms do not express concepts. It follows trivially that for anyone attracted to a localphil or global version of Kripkean externalism, CM is unavailable as a justification for I. We take this point to be obvious, but important and often overlooked. Given the influence of Kripkes views, it makes the search for an alternative justification for I pressing especially since arguments for Kripkean externalism rely heavily on I.
Even though Putnam is famous for the slogan "meanings just ain't in the head" (2) one who uses certain terms properly needs to have more in her head on Putnam's view than on Kripke's. In particular, Putnam holds, in order to use a natural kind term correctly (and to know what the term means) one must associate a "stereotype" with the term (he sometimes calls stereotypes "concepts").
Putnam holds that stereotypes "include" certain properties that are conventionally thought to be possessed by what's in the extension of the term. But it isn't necessary that all or most or even most normal things in the extension possess these properties (there can be tigers without stripes); nor does a thing's possessing them ensure that it lies in that extension. XYZ possesses the stereotypical properties of water, but isn't water.
Note that if our intuitions about Cs are guided by a stereotype that includes properties P1...Pn, and the stereotype leads us to judge intuitively that entities with P1...Pn are Cs, then, since by Putnams admission there can be Cs that don't have P1..Pn, and there can be entities with P1...Pn that aren't Cs, our intuitions won't necessarily be reliable.
However, it's important to notice that intuitions so guided could be reliable. It might be the case that all (or most) entities with P1...Pn in our environment are Cs and that there are no (or extremely few) Cs that don't have P1...Pn. This may well be the case for water, tigers, cats and many other natural kinds. When it's the case that is, when the stereotypes happen to pick out that which they are stereotypes of our intuitions, guided by the stereotype, will be contingently reliable.
Would contingently reliable intuitions make it possible for an externalist to use CM in a defense of I? According to CM, philosophers care about intuitions because they're engaged in conceptual analysis. Analysis of a concept C, we've assumed, is an attempt to come up with necessary and sufficient conditions for being a C. Contingently reliable intuitions won't enable you to do this. If your intuitions about Cs are reliable because of certain contingent facts, for instance that all and only Cs fit the stereotype of Cs, they will, at best, enable you to discover materially adequate conditions, not necessary and sufficient conditions. To see why this is so, suppose someone suggests a set, S, of necessary and sufficient conditions, for being a C. To test S using her intuitions about specific cases a philosopher would want to consider merely possible cases as well as actual cases (merely possible cases, of course, are often the only ones philosophers consider). Now, among the possible situations in which S should be tested are those in which the stereotype doesn't pick out Cs and those in which there are Cs that don't fit the stereotype (we know there are such situations because the reliability of the stereotypes is contingent). We know that with regard to such worlds, our intuitions (guided as they are by the stereotype) are necessarily misleading as a test for S. So contingently reliable intuitions won't help a proponent of CM.
Localphil and Global Internalism
It appears, then, that CM is available as a justification of I only to someone willing to defend an internalist treatment of philosophical concepts. Someone attracted to externalism and CM might be attracted to localphil internalism (i.e., local internalism that include philosophical concepts in the class of concepts that are supposed to be understood in internalist terms).
Localphil Internalism and CM
In order to be philosophically respectable, a localphil internalist of the sort just described needs two things: a) some principled way of determining which concepts are to be understood in an externalist way, and which in an internalist way, and b) an argument to show that the relevant philosophical concepts lie in the latter category.
We know of no attempt to do this, (3) and we are pessimistic about the prospects, given that the localphil internalist endorses externalist arguments for other terms. The commitments the localphil internalist undertakes by accepting these arguments render it difficult for her to make an exception of philosophical concepts, or so we shall argue.
Philosophy On Twin Earth
In "The Meaning Of 'Meaning'" Putnam constructs Twin Earth cases for terms such as 'pencil', 'chair' and 'pediatrician' and seems to suggest that externalist arguments can be applied in a very wide range of cases. He doesn't discuss whether these include philosophical concepts, but if he were to make such concepts an exception he'd have to claim that in order to understand or grasp a philosophical concept, C, we must have in our minds a set of necessary and sufficient conditions for being C. Is this claim plausible? We can think of no evidence for it and considerable evidence against it.
Non-philosophers are unable, if asked, to specify what philosophers would accept as necessary and sufficient conditions for, say, reference and knowledge. In light of this, the localphil internalist could try some version of what one might call the Platonic Assumption, namely, that we all know the "right" necessary and sufficient conditions; it is just that this knowledge is buried very deep down in our minds and is therefore hard to make explicit.
This is a desperate move for a number of reasons, only one of which we'll mention here. A localphil internalist would have to explain why she assumes this for philosophical concepts, but not for names, natural kind terms, artifact terms, and so on. Why assume that deep down we know what knowledge, reference, freedom, and the like are, but not what pencils, pediatricians, chairs and tigers are? If the view is unacceptable in the latter cases, isn't it equally unacceptable in the former?
If the localphil internalist doesn't endorse the Platonic Assumption and takes what people associate with philosophical terms (the philosophical analog of Putnam's superficial features) at face value (as she does when constructing other Twin Earth examples), it isn't difficult to construct externalist arguments involving philosophical concepts. Suppose someone let's call her 'Alice' is asked what she associates with 'reference' or 'a word referring to something', and answers that for a word to refer to something is "for it to be about something, to stand for something." Now, suppose this (and maybe a few other vague characteristics we can elicit from her) is what's "in her mind" when she uses 'reference' and 'refer'. Suppose also that Kripke has given a correct account of what reference is; in particular, he has correctly described the extension of 'reference' in the actual world. Now imagine another world, just like ours, except that in it, the extension of 'reference' fits Frege's description of reference. Twin Alice, when asked what she associates with 'reference', would give exactly the same answer as Alice. She would say things like "For something to refer is for it to stand for something, to be about something." But on Twin earth, the extension of 'reference' is not the same as it is in our world. The extensions of 'reference' for Alice and Twin Alice differ; but what is in their minds is the same. The localphil internalist thinks these sorts of cases imply externalism for non-philosophical terms; it would seem therefore, that she's committed to the same sort of conclusion when philosophical concepts are involved in the absence, of course, of a good reason for thinking otherwise.
We now consider the remaining alternative, global internalism. There are two reasons why this alternative is unlikely to prove compelling to many philosophers who use I. First of all, many want to employ I but, for reasons given by Kripke, Putnam and Burge, find global internalism an unattractive position. Secondly, as we have seen, externalists support their position in large measure by appealing to intuitions about cases. The intuitions many take to render some version of externalism plausible are common, if not universal. This fact is highly significant in the present context; for it justifies us in saying that if global internalism is true, intuitions in philosophy are not wholly trustworthy or reliable. If they were, the truth of internalism wouldn't be accompanied by externalist intuitions. But then it follows that a global internalist construal of CM doesn't provide us with a straightforward basis for I.
We suspect many users of I will either prefer externalism or prefer to let the argument for I be neutral between internalism and externalism. It's worth emphasizing that there's an important motivation for letting the argument for I remain neutral. If you have to rely on I to decide between internalism and externalism (and most arguments, at least for externalism, do), you'd want an independent justification for I to avoid the circle that would result if you based your favored position on I and I on your favored position.
For those inclined towards neutrality or externalism, there may be a way out of this quandry. I may be capable of being justified by the Belief Model suitably augmented, as we shall see. BM has two components:
According to BM, it's the externalist's beliefs about meaning, reference, and the way counterfactual discourse functions that guide her intuitions about Putnam's Twin Earth case and Kripke's Aristotle case. According to BM, these beliefs need'nt be constitutive of the concept of reference; there may be no set of beliefs the possession of which is necessary and sufficient for linguistic mastery of terms like 'reference'. It's consistant with BM that there are no internalized necessary and sufficient conditions for falling under a philosophical term.
It's also consistent with BM that many of the beliefs that guide intuitions are particular. We all have indefinitely many beliefs like "Aristotle might not have taught Alexander the Great." It's possible that people also have some very general pre philosophical beliefs about matters of philosophical interest. This would be the equivalent in BM of the Platonic Assumption. We see no compelling reason to think this must be the case or that if it is, these beliefs are true. Intuitions about cases that are not covered by the particularized beliefs can be explained by invoking analogical lines of reasoning. Whether the mechanism that produces these intuitions can be analyzed purely in terms of beliefs and operations performed on them, is a further issue we will not take up here.
If BM is correct, the reliability of one's intuitions depends on two factors:
How 1 and 2 are determined regarding our intuitions about some domain is a question that probably needs to be settled on a case-by-case basis. We suspect that in many cases, we are unlikely to find that our intuitions are generally reliable. The reason for this suspicion has to do with conflicting intuitions about cases, a phenomenon we think is more widespread and important than is usually recognized. Linguists who have studied intuitions in their own field have found evidence that they are theoretically biased and conflict with layperson's intuitions. (4) We'd like to see an empirical investigation of this sort of phenomenon in philosophy.
The phenomenon of conflicting intuitions in philosophy gives us further reason to prefer BM to CM. A proponent of CM might employ two strategies for explaining this phenomenon. One is to accuse some of the parties to the conflict of philosophical blindness, maintaining that their intuitions are based on an inadequate grasp of or access to the concept at issue. It's difficult to see how applications of this approach could garner a consensus; in the heat of an argument, what philosopher is willing to plead guilty to the charge of philosophical blindness? On what non-question-begging grounds can legitimate, "veridical" intuitions be singled out from the rest?
An equally problematic alternative is to say that differing intuitions indicate differing concepts, rendering the conflicts merely apparent. This strategy is bound to strike many philosophers as counterintuitive. A proponent of CM who uses it therefore owes us an account of what is wrong with this intuition an account which must avoid conflicting with the general presumption in CM that intuitions are trustworthy.
One who endorses BM can avoid these difficulties simply by saying that individuals who do not share the relevant intuitions fail to do so because they have different beliefs. (5) She can also point out that what look like philosophical stalemates from the standpoint of CM may well be resolvable, because at least in principle those involved in the disputes can discover where their beliefs differ, and then try to figure out which beliefs are true or more comprehensive. It may turn out in some cases that this can be done using empirical methods.
We want to end by mentioning four issues we think should be pursued further. First, the two models we have presented require additional elaboration. In particular, we'd like to see an answer to the question, What is the origin of the beliefs on which BM tells us intuitions depend? Second, more potential defenses of I (such as Rawlsian constuctivism and the attempt to base I on a dispositional theory of meaning) need to be evaluated. (6) Third, it should be determined whether the justification of I varies in different areas of philosophy. Finally, the phenomenon of conflicting intuitions deserves considerably more attention than it has received.
(1) Kripke, Saul. Naming and Necessity (Cambridge, Mass.: Harvard University Press, 1980).
(2) Putnam, Hilary. "The Meaning of 'Meaning'," in Philosophical Papers, Vol. 2 (Cambridge: Cambridge University Press).
(3) It is interesting in this connection to look at George Bealer's recent "On the Possibility of Philosophical Knowledge," in James E. Tomberlin, ed., Philosophical Perspectives, 10 (Oxford: Blackwell, 1996). We believe that Bealer's efforts to justify I fail.
(4) See, for example, William Labov's What is a Linguistic Fact? (Lisse: The Peter de Ridder Press, 1975).
(5) At some point, beliefs may differ enough to justify an attribution of conceptual difference. But BM does not force one into conceptual relativism the way CM is in danger of doing. And an advocate of BM who attributes different concepts to people who initially appear to share a concept does not render her position unstable in the way a proponent of CM does.
(6) We think that Kripke's arguments against the dispositional theory of meaning render the second of these alternatives problematic, and remain unconvinced by the various attempts to defend the dispositional theory. See Kripke's Wittgenstein on Rules and Private Language (Cambridge, Mass.: Harvard University Press, 1982). We also think that Rawlsian constructivism marks a serious departure from the use of intuitions as evidence for or against the truth of philosophical theories.