1
Concepts of Indiscriminability
This chapter is a study in the epistemology of identity. Of course, what is true of knowledge in general is true of knowledge of identity in particular, and for most purposes it would be both inefficient and prejudicial to focus on the latter rather than the former. Nevertheless, the relation of identity exhibits a unique formal character, which is reflected as in a distorting mirror by the formal character of knowledge of identity. The interaction of general features of knowledge with general features of identity needs special attention. The phenomenon of indiscriminability provides one way into this area, the way to be followed here. Since the indiscriminability of objects is less a route to knowledge that they are identical than a block to knowledge that they are distinct (or so it will be argued), it might be less misleading to say that this chapter deals with knowledge of non-identity.
Indiscriminability, unlike identity, is non-transitive. It is not always the case that when a is indiscriminable from b and b is indiscriminable from c then a is indiscriminable from c, for otherwise there could not be series in which the differences between successive members are both too small to be discriminated and yet add up to a discriminable difference between the first member and the last. The non-transitivity of indiscriminability is often treated as a specific and rather mysterious feature of sensory experience, from which startling philosophical conclusions may be drawn. The underlying aim of this chapter is to understand it as a general cognitive phenomenon.
Section 1 develops a cognitive model of discrimination.1 Section 2 uses the model to explain some formal features of indiscriminability, such as its failure to be transitive, Section 3 uses it to expound a sense in which discrimination is intentional. In section 4 this intentionality is observed to threaten an attempt to define a transitive notion of indiscriminability in terms of a non-transitive notion. Section 5 ties up loose ends.
1.1 Indiscriminability and Cognition
What is indiscriminability? Surface form indicates that things are indiscriminable if and only if it is not possible to discriminate between them. One may therefore expect an account of indiscriminability to comprise accounts of discrimination and, perhaps less importantly, of the relevant kind of possibility.
What is it to discriminate? The verb has an active meaning, in a more than purely grammatical sense. To discriminate is to do something. That is not, of course, to say that discrimination is a bodily (rather than mental) act, still less that it is whatever falls under a certain behavioural (rather than intentional) description. Discrimination has at the very least a cognitive component. For the processes involved in discrimination can also lead to ignorance or error. If I fail to discriminate between the lengths of two lines, one slightly longer than the other, there is something I have failed to find out. If, misled by perspective, I judge one line to be longer than another, and they turn out to be of equal length, my would-be discrimination was incorrect. Failures of discrimination are cognitive failures, so discrimination is a cognitive act.
Naturally, it is open to anyone to use the word âdiscriminationâ in a non-cognitive sense, perhaps defined in terms of differential responses to stimuli, where âstimulusâ and âresponseâ are somehow themselves understood non-cognitively. One might even say that a window can discriminate between a ball and a feather, if one breaks it and the other does not. However, that just looks like a metaphorical extension of the concept; we pretend that the window is a knowing subject. In literal terms, the window is at most the instrument of our discrimination. Derived senses of the word will not be pursued here.
What kind of cognitive act is discrimination? If discriminating could be assimilated to judging, then false or incorrect discriminations should be possible, as well as true or correct ones. They seem not to be. If I judge this vintage to be fuller-bodied than that, and they turn out to be the same vintage, poured from the same bottle, I have not falsely or incorrectly discriminated the vintage from itself; I have not discriminated at all. I may have seemed to discriminate the vintage in this glass from the vintage in that, but I have not actually done so. In general, a good way of refuting someone's claim to be able to discriminate between this and that is to show that this is that. The cognitive failures associated with discrimination occur in its absence.
Discrimination thus seems closer to knowledge than to belief; false discrimination, like false knowledge and unlike false belief, is a contradiction in terms. Indeed, there is good reason to push this likeness to the point of identity, for if we can characterize discrimination as knowledge, we shall be in a position to explain both why discrimination cannot be in error and why the alternatives to it are ignorance and error.
It would not be quite right simply to assimilate discriminating to knowing, for the former is a process and the latter â like believing â a state. One might use ârecognitionâ and âjudgementâ to express the processes analogous to knowledge and belief, respectively, since judgement but not recognition can be false. Just as knowledge entails belief (on most views), so recognition entails judgement. The hypothesis would then be that discrimination is a kind of recognition, not merely of judgement.
Discrimination involves states as well as processes. The state of being able to discriminate stands to the process of discrimination as the state of being able to recognize stands to the process of recognition. A more specific comparison might be with memory. There is the process of remembering that tomorrow is my birthday, active recall, and there is the state of being able to remember that tomorrow is my birthday. The process is a kind of activation of the state, an exercise of the ability which it is the state of having. However, if I am able to remember that tomorrow is my birthday, I already know that tomorrow is my birthday â I am not merely able to know it; in contrast, I may be able to discriminate between two things, even though I have not yet encountered them and have as yet no relevant knowledge, simply because my present cognitive capacities would permit me to discriminate if I did encounter them. The knowledge activated in discrimination need not pre-exist the process. Thus the closest we can reasonably come to assimilating discrimination to knowledge is the hypothesis that it is the activation of knowledge, where this activation may be described either as acquisition or employment, depending on whether the subject calls on knowledge already possessed.
What knowledge is activated in discrimination? What is its content? The identity of a and b rules out discrimination between them. It would be natural to explain this in terms of the content of the knowledge activated in discrimination, by the incompatibility of this content with the identity of a and b. The simplest hypothesis meeting these requirements is this: to discriminate between a and b is to activate the knowledge that a and b are distinct. One cannot discriminate between a and a because there can be no knowledge that a and a are distinct to be activated; knowledge entails truth. The idea of discriminating as activating knowledge of distinctness is explored in this and subsequent chapters; it will emerge as adequate to a wide variety of cognitive phenomena.
Not every kind of knowledge is relevant to every kind of discrimination. When we speak of discriminating between wines, we do not usually mean reading the labels on the bottles, although in special circumstances we might speak of this as a quick but not always reliable means of discrimination unavailable to the illiterate. In this respect the sense of the word âdiscriminationâ is context-relative; different kinds of knowledge are meant in different contexts. The word âknowledgeâ, in contrast, will not be used in this context-relative way; it will be read as stably covering all relevant kinds. Thus although discrimination between a and b entails knowledge that a and b are distinct, not all knowledge that a and b are distinct entails discrimination between a and b in the sense appropriate to a given context, since the knowledge may not issue from the right source.
There is no good reason to restrict the kinds of knowledge which constitute kinds of discrimination. The relevant sources may be sensory modalities, statistical techniques, the use of litmus paper or parish records. Any source which yields knowledge of the properties or relations of a and b may indicate that a has a property which b lacks, or stands in a relation to something to which b does not, thus revealing the distinctness of a and b by Leibniz's Law; identity entails sharing of properties and relations. Any source of knowledge of properties or relations can therefore correspond to a kind of discrimination.
There is also no good reason to restrict the kinds of object which can be discriminated. For, by the same reasoning, for any object a of whose properties or relations one can have non-trivial knowledge, there will be an object b such that one can know that a and b are distinct, and in that sense discriminate between them. One can discriminate between paints, or between painted surfaces or walls, or between their colours, or between one's experiences of those colours, for knowledge of distinctness is possible in all these cases. The subjects who discriminate will also be as various as the subjects who know, or at least who know facts of distinctness: people, animals, perhaps machines, and groups of these.
Given this account of discrimination, it is easy to define indiscriminability: a is indiscriminable from b for a subject at a time if and only if at that time the subject is not able to discriminate between a and b, that is, if and only if at that time the subject is not able to activate (acquire or employ) the relevant kind of knowledge that a and b are distinct.
This account is no doubt something of an idealization; as an attempt to state necessary and sufficient conditions for the correct use of the word âdiscriminationâ in ordinary English it would presumably go the way of most attempts to state necessary and sufficient conditions. Instead it displays a paradigm of discrimination, like pure water. What satisfy the account are clear cases of discrimination; what do not satisfy it but come near to doing so are less clear cases, which it may nevertheless be useful to think of in the same terms. A baby or animal may discriminate sugar from salt, even if it seems excessive to describe them as having the propositional knowledge that sugar is not salt. One might even say that it discriminates between London and Paris (its attitude to passers-by varies) while being reluctant to admit that it thinks of either city. In increasingly attenuated senses, flowers discriminate between day and night, litmus paper between acid and alkali, the window between the feather and the ball. The use of the word is not wholly inept, for such conceptualized sensitivity to difference bears some similarity to the cognitive paradigm. Knowledge is missing, but information may be received. This chapter and the next two, however, deal with more strictly knowledgeable forms of discrimination. They are more amenable to a certain kind of understanding, and the phenomena it discerns should have their analogues in the less articulate cases.
Knowledge is intentional. If, at a party, the sandal-wearer is the vegetarian and the tea-drinker is the George Orwell expert, I can know ...