Find Enlightenment

Identity And Universals: A Conceptualist Approach to Logical, Metaphysical, and Epistemological Problems of Contemporary Identity Theory
Chapter 4, Metaphysical Conditions Versus Epistemological Criteria
by Carolyn Ray

Date: 11 Nov 98
Copyright: Carolyn Ray

Condition: RD = A metaphysical fact or circumstance. A necessary condition is one that must obtain for another circumstance to be possible; a sufficient condition is one that guarantees another circumstance.

Criterion of identity: RD = An epistemological principle used as a guide for distinguishing one object from another and from its environment, and as a guide for determining whether a single object has endured through time and change. An adequate criterion of identity is one that can be relied upon by itself, though there may be others that could be used as well. A required criterion is one that must be used for successful judgment.

Metaphysics is the study of the way things are. Epistemology is the study of the methods and tools we use to acquire and assimilate knowledge. A metaphysical question refers to how reality is; an epistemological question refers to how we know how reality is. The two branches of philosophy are connected, but their objects are distinct. The literature on identity does not always make this distinction. Equivocations and other confusions result.

Some identity theorists seem to think of epistemological questions about identity as less serious, less important, less hard-core than metaphysical questions. I speculate that this is the case because they have accepted the idea that human "knowledge" is unreliable, whereas metaphysical reality is not subject to human fallibility. Whatever the reason, many theorists choose to associate themselves explicitly with what they feel is the more powerful discipline, denying that epistemological questions are even interesting. There are some identity theorists who are unable to stick with this decision, and end up doing epistemology, while denying that they are doing it. In this chapter, I will examine some of the terms that are subject to confusion.

Criteria of Identity

One of the major projects in mainstream identity literature is to specify criteria of identity. Thus it would be wise to understand what identity theorists mean by 'criterion,' and what I mean by it. To that end, I investigate according to the method of objectivity, and ask, "What facts of reality give rise to the concept CRITERION?

People make judgments about themselves, about others, about their environment. In general, when people consider the facts before them and either deliberate or draw on deliberations they have engaged in, in the past, they are making judgments according to some rule. The rule is created with respect to some standard against which other things are then measured. The rule might be propositional in form, or it might be like a "checklist" of required items; if the items on the checklist are present, then the rule is fulfilled. A simple example of the use of such a rule might be, "Go collect some wood for tonight's fire;" the checklist is short: "wood, burnable (dry)," measured against the standard of what is usually burned in the fireplace. A more complex example of the use of a rule might be, "Hire someone who is qualified for the job;" the checklist is long: "knows C++, has UNIX experience, understands object oriented programming, has at least 1 year experience," etc., etc. , measured against the standard junior level programmer. These are the facts that give rise to the concept CRITERION.

A good criterion of identity is one that is based on the facts of reality as we observe them. There may be features of the world that we will not account for because we are not yet aware of them. That is acceptable: our concepts and principles are only useful insofar as they help us to live in our world, and for that omniscience has never been required. We can also change our criterion if it turns out to be inadequate based on new data about the world.

A criterion, then, is part of one's epistemological framework. It is not a metaphysical condition. Whether it is the use of the word 'criterion' that leads identity theorists to confuse metaphysical issues with epistemological ones, or whether some prior confusion of metaphysics with epistemology inspires them to use the word 'criterion' when discussing metaphysical conditions, this much is certain: there is a great deal of confusion over the criteria of identity. Were identity theorists to use the word 'criterion' consistently to designate metaphysical conditions, we could simply make a note of this specialized usage as we analyze their metaphysical projects; but nothing so simple happens in the literature. To demonstrate the extent of the confusion, I turn now to a number of passages from contemporary and historical authors.

Eli Hirsch and Criteria

Hirsch, a temporalist identity theorist, asserts that his book is concerned, not with epistemological questions, but with conceptual questions. Immediately, it is difficult to see what the distinction is, since he seems to believe that concepts are tools we use to acquire knowledge about the world; he apparently thinks of it as the difference between asking how we can be sure (epistemology) that we are reidentifying the same object under suboptimal conditions, and what we are doing (conceptually) under optimal conditions when we observe an object continuously for some stretch of time. At any rate, this unclarity pales in comparison to his confusion of epistemological (conceptual) questions with metaphysical ones. He claims to be concerned with the "analysis of our identity concept" and to be interested in developing identity criteria (Hirsch, Concept, p. 3). Of physical objects he says,

Any physical object has a career which stretches over a period of time, a career which we can think of as comprised of a temporal succession of momentary stages. The successive parts, or stages, of an object's career must hang together in some distinctive way; otherwise there would be nothing to prevent us from arbitrarily combining into a single career the early stages of one object with the later stages of a different object. Evidently not just any succession of object-stages corresponds to a single persisting object....where this fundamental category is to be understood as loosely comprising items which can straightforwardly be said to occupy space and to persist through time (Hirsch, Concept, p. 3).

Here, he seems to make straightforwardly metaphysical claims, and those claims seem to involve actual spatiotemporal continuity. He is not talking about our judgments regarding reality, but reality itself. He goes on to say:

We want to understand the nature of the unity-making relationship which binds the successive stages of the career of a single persisting object. When one first reflects upon this question, an idea which might readily come to mind is that contiguous stages of a single career must be qualitatively very similar and spatially very close....This very simple kind of analysis is much too simple.... (Hirsch, p. 7)

Notice that Hirsch intends to deny the truth of the speculation on the grounds of oversimplification, but not on account of its focus on metaphysical rather than conceptual questions. Evidently, then, what he means by 'conceptual' is what is normally meant by 'metaphysical.' Let us grant him this peculiar usage, keeping in mind that it is a special usage, and see what happens.

When Hirsch critiques the foregoing "analysis of persistence" (which he calls the "Simple Continuity Analysis"), he notes that it can be given (1) a strong reading, such that the qualitative changes can be divided up into a series of changes as small as you like (in the sense of approaching a limit), or (2) a weak reading, such that it tolerates "noticeable jumps." And now he makes another claim regarding the nature of his project:

I am not concerned with the question "What is the ultimate nature of a change which is continuous in the strong sense?" but rather with the question "Should we require of an object that its changes should be continuous in the strong sense?" (Hirsch, p. 12)

It seems his project is now about the kinds of judgments we are justified in making. His examples support this contention. While he acknowledges that many changes are continuous in the strong sense, he asserts that many are not; of the latter, he gives two kinds of examples: miraculous events (such as a cat's eyes suddenly changing from brown to green without any intermediate shades), and ordinary cases of part loss or part addition. Ignoring the miraculous events as outside the scope of a theory about reality, let us consider an example of the second kind to see if we can determine exactly what his project is.

Hirsch asks us to think of a tree that takes up 30 cubic feet of space, but has a 2 cubic-foot branch chopped off and then takes up only 28 cubic feet. Our ordinary thinking, he says, dictates that we say that there was no point at which the tree took up 29 cubic feet. The tree suffered a "noticeable" jump with respect to volume and shape. The jump in shape and volume he illustrates with two pictures--one of a tree with two branches and one of a tree with one branch--asserting that "clearly there was no time between t and t' at which the tree's shape could be pictured in some way intermediate between [the two pictures]" (Hirsch, Concept, p. 14). Though we can grant that the tree never took up 29 cubic feet, an accurate picture of a tree with its branch sawed part-way through represents such an intermediate. Shape is not the same thing as volume; they are the values that result from two different ways of measuring. Hirsch relies on our "noticing" only one kind of measurement, and on our tendency to picture the tree in discrete space-time segments, to convince us that there is metaphysical discontinuity. But measurement is an epistemological process that may describe the metaphysical facts from any number of angles (the tree's circumference increases as the blade cuts through it, for example). Either he has confused the epistemological project with the metaphysical, or the project is actually epistemological, though he may not realize it. So perhaps my original hypothesis that Hirsch uses the word 'conceptual' in the way most people use the word 'metaphysical' is incorrect. But since he insists that there really are cases in which discrete jumps occur, his own disavowal of interest in metaphysical questions is also incorrect. At bottom, the problem with Hirsch's discussions of criteria of identity is not merely a matter of using words differently than I would use them. It is a matter of constant equivocation. It is never clear what subject he is addressing.

Harold Noonan and the Reliability of Criteria

An issue I want to touch on briefly is the way problems arise for establishing criteria of identity when one accepts the doctrine of logical possibility. I have claimed that such commitments lead to the unrestrained generation of spurious puzzle cases. Harold Noonan's treatise, Personal Identity, is a solid example of the problem. Noonan would not easily accept my understanding of the word 'criterion,' because it is based on reality as he and I both understand it. A good criterion of identity, he thinks, must include every "logical possibility" that one can think of, where 'possible' is defined as 'can be described by an internally consistent statement'. And apparently Noonan's view is that logic, having little to do with the real world as it is, must not be unduly constrained by facts.

Noonan critiques several criteria of personal identity through the same basic approach: state the criterion, imagine a case in which it does not hold, and pronounce the criterion inadequate in the sense that it does not cover every case. In general, the imagined cases are imaginary: the Bodily Criterion, he claims, is inadequate because he can imagine a case in which a person's body does not survive but the person does. This procedure would seem to contradict his implicit "better doctrine," which is evident in the first passages of his book: must be regarded as a condition of adequacy on any account of what personal identity consists in that it not entail that personal identity is unknowable, or not knowable in the ways we ordinarily take it to be, or leave it completely mysterious how it can be known in these ways....

....Personal identity, as we know it in our everyday lives, is in fact constituted by bodily identity (Noonan, pp. 2-3).

Clearly, he has a sense that there is a connection between metaphysics and epistemology, and that the generation of metaphysical hypotheses must not lead to the invalidation of good epistemological procedures. Given this attitude, one might reasonably expect Noonan to reject as absurd, for example, the above objection to the Bodily Criterion. His acceptance of the doctrine of logical possibility persuades him to make personal identity mysterious by contradicting the facts that provide us with actual criteria and thereby make it knowable.

It is perhaps significant in this regard that Noonan's survey of criteria of identity states all of the criteria in terms of conditions which are both necessary and sufficient--and in his terms, this means stating the conditions as an equivalence formula. In other words, they are stated in the form,

Person2 at t2 is the same person as Person1 at t1

IF AND ONLY IF Person2 has the same body as Person1 had.

In other words, Noonan interprets the Bodily Criterion, allegedly set forth by thinkers throughout the history of identity theory, as meaning that bodily identity is both a necessary and a sufficient condition for personal identity. In fact, he claims that his project is to examine previous hypotheses regarding, and eventually to decide, what is constitutive of personal identity. And this, he says, is a metaphysical project; while "evidential" questions are not completely irrelevant, "it is important to be aware at the outset that this is not what philosophers are interested in when they debate the problem of personal identity" (Noonan, p. 2). This sentiment should come as a surprise, for example, to Bernard Williams, given that his Problems of the Self is all about what evidence we have for making judgments about who is who. Noonan is quite familiar with Williams's work; hence we should probably interpret this claim as a normative one: perhaps other identity theorists do epistemology, but they should not.

So here we have a clear case of an identity theorist using the word 'criterion' to denote metaphysical conditions. So far as we have seen, Noonan is still working within his metaphysical framework, at least in the sense that he feels that logical possibility is indicative of metaphysical possibility. Again, we will grant him his usage, and see what happens. In the next section we will see that he does not stick to the project.

Noonan's Necessary and Sufficient Conditions

We have seen that there is a great deal of confusion in the literature because of equivocation on the term 'criterion.' Let us see if the phrase 'necessary and sufficient conditions' fares any better. In the last section, we saw that Noonan relied on logical possibility to create the objection to the Bodily Criterion of personal identity. The objection to the Memory Criterion is stated somewhat differently:

Person2 at t2 is the same person as Person1 at t1 JUST IN CASE Person2 at t2 is linked by continuity of experience-memory to Person1 at t1 (Noonan, p. 12).

One of the objections Noonan takes as worthy of note goes like this:

Memory cannot occur as an ingredient in a definition of personal identity because memory already presupposes personal identity--as knowledge in general presupposes truth....We distinguish between veridical and apparent memory and accept without difficulty that people can seem (to themselves) to remember doing things which they did not do....But how is this distinction to be made if not by an appeal to personal identity? (Noonan, p. 13).

Given Noonan's stated metaphysical project, we must take 'definition' not as an epistemological term in the sense of condensing information about real persons or their identity, but rather as another way of saying 'conditions which are both necessary and sufficient.' Nonetheless, the objection itself is based on an epistemological concern: we do not know whether to call the person the same person, because sometimes we--or the person itself--can be mistaken with regard to memories. But now it is mysterious why Noonan, as a metaphysician, finds this persuasive. If his chief concern is to set out the metaphysical conditions that make personal identity possible, or to say what a person is, what is wrong with saying that, metaphysically, veridical memories constitute personal identity, and letting the epistemologists sort out how to tell one person from another? Moreover, circularity itself is an epistemological concern: when we search for knowledge, we do not want to be told what we already know in different words. Circularity in metaphysical terms, is a good thing, if we are looking for a single condition which is both necessary and sufficient. For what does it mean to look for such a condition, if not to look for conditions that are mutually--circularly--dependent upon each other? Noonan's 'criterion' says "Sameness of memory is both a necessary and a sufficient condition for personal identity, and personal identity is both a necessary and sufficient condition for sameness of memory. Will not any "criterion" of identity, if stated in these terms, be circular?

This particular debate, first waged between John Locke and Joseph Butler, will be handled in the later discussion on persons and memory; so I will draw no conclusions concerning the worth of the Memory Criterion now. Here, I am only interested in illustrating the confusion exhibited by even those who most emphatically deny an interest in epistemological issues.

I suggest that the problem occurs because Noonan considers Locke's project to be an attempt to give a definition of personal identity. Since Noonan thinks that supplying an equivalence statement about metaphysically necessary and sufficient conditions is giving a definition, it does not occur to him that these might be distinct endeavors with distinct results. If we literally try to understand the Memory Criterion, as given above, as a definition, even in the mathematical/modern logical sense, then it would fail miserably in point of being too broad. All sorts of animals answer the description "linked by an overlapping chain of direct memories"--it is on this, for example, that we rely when we train pets (and on which they themselves rely to survive). Compare this alleged definition to "a figure is square IFF it is composed of 4 equal sides and 4 equal angles;" this statement is considered a definition because all there is to a square is this condition. To put it another way, if you have one side of the biconditional, you get the other.

Thus, I contend that there is no reasonable sense in which we can understand such a statement of conditions as a definition; and since circularity is a problem for definitions and other epistemological statements, not for statements of metaphysical conditions, there is no reasonable sense in which we can understand this criterion to be viciously circular. The objection attacks a straw man, because of a confusion between metaphysics and epistemology.

Bernard Williams's Epistemological Project

Bernard Williams's work is straightforwardly epistemological. His concern is judgments about--and justifications of judgments about--personal identity. Yet he uses the terms 'necessary condition' and 'sufficient condition'; in fact, he uses them as though they are interchangeable with the word 'criterion,' and as though all of these words refer to epistemological concerns. Let us examine the way he uses these terms to work with identity puzzles.

Williams lays out his project in "Personal Identity and Individuation," the first article in his collection Problems of the Self:

....Identity of body is at least not a sufficient condition of personal identity, and other considerations, of personal characteristics and, above all, memory, must be invoked...

....I shall try to show that bodily identity is always a necessary condition of personal identity... (Williams, 'Personal,' p. 1).

The above statements appear to announce the undertaking of a metaphysical project. However, Williams says he is concerned to refute the following two theses:

(1) least one case can be consistently constructed in which bodily identity fails, but in which the other conditions will be sufficient for an assertion of personal identity....

(2) ....there is no conceivable situation in which bodily identity would be necessary, some other conditions being always both necessary and sufficient (Williams, 'Personal,' p. 1)

Williams takes several lines of argument against the notion that bodily identity is not a necessary condition for (i.e., required for judgment of) personal identity. Note the use of the phrase 'sufficient for the assertion of personal identity' in the first thesis. Williams is here--and throughout his article--talking about criteria of identity in the sense I originally defined it: as epistemological principles. Whatever other evidence of personal identity we may have, we must always resort to evidence of the same body.

Thus we see that Williams, at least, uses the phrase 'necessary and sufficient conditions' to mean 'epistemological criteria.' Williams uses his widely-acclaimed Reduplication Argument to show that we have no way of making identity judgments on the basis of memory claims alone. The basic form of the Reduplication Argument considers two men who claim to remember being a deceased historical figure. If Charles and his brother Robert both claimed to remember being Guy Fawkes, we can easily see that, because the two men are not identical to each other, the expression 'has the same memories as' cannot possibly mean 'has the identical memories' but must mean 'has similar memories as.' We cannot decide which man, if any, is really identical to Fawkes, since they both just have memories similar to what we imagine Fawkes's memories were. But if we must describe Charles and Robert as merely having similar memories, we must surely do so if Charles alone were to make this claim; in the absence of Fawkes's body, we can claim only that that Charles has memories similar to Fawkes's.

Williams soberly concludes that the Reduplication Argument shows the absurdity of claiming to be able to make (epistemological) identity judgments without reference to the person's body, that "the omission of the body takes away all content from the idea of personal identity" (Williams, p. 10).

It is perhaps testimony to my hypothesis that the confusion between metaphysics and epistemology in identity theory has wrecked havoc in the field, that Harold Noonan draws a very different conclusion from Williams's thought experiment. For he believes that Williams has shown that the "intrinsic" (i.e., metaphysical) relationship between Charles and Fawkes is not altered by the existence of Robert! He writes:

Hence, Williams concludes, neither should one identify Charles with Guy Fawkes in the original case where there is no reduplication, for the absence of Robert from that case has nothing to do with the intrinsic relations between Charles and Guy Fawkes--the relations that obtain between them independently of what is true of other people--but it is absurd to suppose that whether a later person P2 is identical with an earlier person P1 can depend upon facts about people other than P1 and P2 (Noonan, p. 15)

This fantastic misreading of Williams's epistemological point may perhaps be traced to Noonan's assessment of identity as a purely metaphysical issue. Then again, perhaps Williams's unfortunate use of the phrase 'necessary and sufficient condition' in an epistemological context is at the heart of this and other misinterpretations. And, predictably, I speculate that the method of science fiction itself has contributed to the muddling of the issues.

It should be clear from this discussion that it is crucial to distinguish between metaphysics and epistemology, and that it would be extremely helpful to use the words 'criterion' and 'condition' consistently and non-interchangeably. It is my goal in this work to use these terms as self-consciously as possible; hence, I will use only 'criterion' in epistemological contexts, only 'condition' in metaphysical contexts; and I will label each use as being metaphysical or epistemological wherever this is not crystal clear from the context.

I have considered some confusions that result from conflating metaphysics and epistemology, and established a standard use of 'criterion' and 'condition.' I will rely on these terms in subsequent chapters.

Go to the next section.

Find Enlightenment at