F 343 Philosophy of Mind, Spring 1996

Professor: Kathleen Akins

Identification and Identity

Nicole Friedrich

Tuesday, March 5^th, 1996

In Sensations and Brain Processes, J.J.C. Smart claims that sensations are identical with brain processes, and that this identity is a contingent truth, knowable only a posteriori. Within his essay, Smart offers eight objections to his theory and responds to each. After outlining Smart’s project, I will offer an objection based on the intuitions behind objection #2 that will demonstrate that his thesis is inconsistent.

Smart equivocates two common uses of the word ‘identity’. One use, say identity₁, can be represented by the biconditional: a = b « " P(Pa ® Pb). Read from left to right, it asserts the indiscernibility of identicals; read from right to left it asserts the identity of indiscernibles. That is, the identity of individuals a and b implies and is implied by their sharing of all their properties P. An example of identity₁ is: water = H₂0. The other use, identity₂, is not an identity in the sense of identity₁; instead, it is really just an identification of an individual thing with a set of properties. The aggregation of properties, P, are held by individual a, such that a can be identified by P. An example of identity₂ is: on March 5, 1996, Glen Clark is the Premier of B.C. With this distinction in mind, I will show that Smart cannot consistently maintain that what he means by identity₂ is in fact identity₁, and also maintain that the identity is only a contingent truth. That is, either Smart must concede that the identity, if true, is a necessary truth (even though known only a posteriori), or Smart must concede that he doesn’t really mean identity₁ at all.

Smart cannot bring himself to believe that there is any ‘mental’ thing over and above a brain process. That is, he believes that all so called mental processes are reducible to brain processes (which are entirely physical processes). According to Smart, these ‘nomological danglers’ are simply figments of our understanding. And according to Occam’s razor, since they serve no purpose, we should not even posit them. Smart is neutral on what mental states actually are, and much of his project is to reduce the psychological properties to topic neutral language, which makes no ontological commitments to mental processes. Topic neutral language is language with which we can refer to a mental process while remaining noncommittal about (and thus, not presuppose) whether the mental process is material or immaterial. Eventually, to fully achieve this inter-theoretic reduction we must identify for every psychological property a physical property.

Smart, however, does not believe that our language of mental or psychological states must become obsolete, for he maintains that terms referring to mental processes do not mean the same thing as those referring to brain processes. Smart does not claim that statements of mental processes have the same logic as, or can be translated into, statements about brain processes (Smart: 170) However, he asserts that, even though they have different meanings, they each in fact refer to the very same thing; and that thing turns out to be a physical process. For instance, according to Smart, the (mental) process of seeing the leaf as green is equivalent to some brain process. Smart leaves it to scientific discovery to find out which mental process is identical to which brain process.

Smart maintains what is called the ‘type identity theory’, for he maintains that types of sensations are identical to types of neural processes. That is, mental process types are identical to brain process types. More specifically, one particular type of mental process = one particular type of brain process. According to Smart, just as lightning is an electrical discharge, a mental process is a brain process. Smart clarifies that he intends for the ‘is’ here to assert strict identity (Smart: 171), and distinguishes between strict identity and correlation. He then rejects the view that mental states are merely correlated with brain states, since correlation implies distinction ¾ as something cannot be correlated with itself.

In Naming and Necessity, Saul Kripke has argued that such type-type identities, if true, must be necessarily true (true in all possible worlds). Kripke distinguishes between rigid and non-rigid designation. Non-rigid designators are descriptive and may refer to different things in different worlds ¾ they are contingent references. For example, in the actual world, ‘the first President of the U.S.’ refers to George Washington. However, in another world it may refer to Thomas Jefferson (if in that world Thomas Jefferson was the first President of the U.S.). Rigid designators are not descriptive, but are simply references that pick out specific things within possible worlds. However, unlike non-rigid designators, rigid designators refer to the same things in all possible worlds ¾ they fix a reference across possible worlds. For example, typically we use the term ‘water’ to refer to that kind of stuff that we are acquainted with in this world (and call water). However, if we stipulate that ‘water’ is the term we wish to use to refer to that same kind of stuff in all possible worlds (whether or not any token of that type of stuff exists in other worlds) then we are using ‘water’ as a rigid designator. Likewise, if we use the term ‘H₂O’ to refer to the same kind of stuff ¾ in this case, hydrogen hydroxide ¾ in all possible worlds, then it is also a rigid designator. Kripke has shown that an identity between rigidly designated terms, if true, is necessarily true. Thus, when we assert the identity ‘water = H₂O’ (using water and H₂O as rigid designators), then this identity is a necessary truth. That is, water is H₂O.

Those arguing that names can be rigid designators need not insist that names are always rigid designators. Clearly, for instance, ‘water’ is used to describe that which is not just H₂O. It could be used to describe that impure stuff we typically drink. Our original choice of names is contingent. Water could have been, and is (in other languages), referred to by other names. The name chosen as the referent is not important. For when we assert A = B, we do not assert that the names or their meanings are identical, rather, we assert that those things which the names A and B refer to are identical. Although the name is contingently chosen, once chosen it fixes a reference. Furthermore, it does not matter that different people in different discussions (contexts) may use different names (or different symbols) to fix a reference, as long as those within the same discussion fix the reference the same way.

This is really just one part of a two stage argument against the type-type identity theory. For, once it is established that the identity, if true, is necessarily true, Kripke argues that the type identity theorist owes the world an explanation for the apparent contingency of the identity (an aspect of the identity that Smart certainly cannot deny). However, Kripke argues that this explanation is not possible due to the indubitable nature of mental processes.

Within his book Consciousness, William Lycan poses a number of arguments against Kripke’s position. Lycan accepts much of Kripke’s necessary identity thesis (Lycan: 14), stating "Kripke’s essentialist thesis is plausible for the case of sense-data and other putative mental objects." (Lycan: 79) However, Lycan claims that the identity is genuine (and subject to Kripke’s thesis) only when it refers to objects, but not when it refers to events or processes. In fact, "If there (really) are phenomenal individuals such as sense-data, then materialism is false right there; no further reasoning is needed." (Lycan: 18) And Smart clearly distinguishes between the object of sensation versus the experience of sensation, stating that "[he is] not arguing that the after-image [(object)] is a brain-process, but that the experience of having an after-image is a brain-process." (Smart: 173)

Lycan challenges Kripke to define mental states as events, and attempts to show why he believes that this cannot be done. Curiously however, Lycan proceeds to defend token-token identities, not type-type identities. (Lycan: 10-12) Indeed, Lycan attributes to Kripke the conclusion that "the token Identity theory is false." (Lycan: 12) and even names a section "Kripke’s Argument against Token Identity" (Lycan: 10) ¾ even though Kripke explicitly states that his argument is directed at type-type identities and may not apply to token-token identity theories. (Kripke: 242) In addition, although it provides us insight into his thinking, the example Lycan provides bears no similarity, likeness, or analogue to a brain or mental event. It does not follow, from his basketball-bouncing example, that all events are not rigidly designatable. Furthermore, Kripke does not have to show that all events can be rigidly designated, but only certain types of events; namely, events that are repeatable function-like process types (analogous to brain and mental process types). Of which, any single example would demonstrate that Lycan was mistaken.

It turns out that some events¾ specifically, repeatable function-like process types¾ can be rigidly designated. Chemical reactions provide nice examples. Consider the following chemical processes, both of which are type-type identities.

[4Fe + 3O₂ ® 2Fe₂O₃] = rust formation process

iron + oxygen ® iron(III) oxide

The type of chemical process referred to by ‘4Fe + 3 O₂ ® 2Fe₂O₃‘ just is the type of chemical process referred to by ‘rust formation’.

[C₆H₁₂O₆ ® 2C₂H₅OH + 2CO₂] = fermentation process

glucose ® ethanol + carbon dioxide

The type of chemical process referred to by ‘C₆H₁₂O₆ ® 2C₂H₅OH + 2CO₂‘ just is the type of chemical process referred to by ‘fermentation’.

Clearly, repeatable function-like processes (analogous to brain and mental processes) can be rigidly designated, and if rigidly designated (and true), are necessarily true identities. Thus, Kripke’s thesis need not be restricted to objects, but can apply to certain event types as well.

The question arises, did Smart intend his identity to be rigidly designated? Lycan’s argument implies that Smart does not intend for his identity statements to be non-rigid. According to Lycan, relationalists (such as Armstrong, Lewis, and other liberals) can renounce chauvinism, whereas non-relationalists cannot. Lycan later claims that Smart cannot escape chauvinism, for "it is reasonable to continue to take Smart as an unabashed Type-Type theorist … [and] as we have seen, a Type-identity Theory of this kind … is crassly chauvinist[ic]." (Lycan: 20) Hence, according to Lycan, Smart is a non-relationalist. Later, however, Lycan claims that relationalists do not intend for their identity statements to be between rigidly designated terms. However, he then concedes that non-relationalists do not escape in this manner, and that an identity statement ‘brain process = mental process’ can be between rigidly designated terms. (Lycan: 14) Thus, even Lycan implies that Smart does not escape Kripke’s thesis by claiming that he does not intend for his identity statements to be rigid.

However, since his essay was written prior to much of the required modal infrastructure, it is likely that Smart did not intend for his proposed identity to be between two rigidly designated terms. However, it has been proposed that Smart would continue to hold this position, even armed with modern modal understanding and the distinction between rigid designators and definite descriptions. For instance, Frank Jackson says that Smart did not intend for the identity statement ‘Brain process type A = mental process type A’ to hold in all possible worlds. According to Jackson, Smart intended the identity to be between a rigid designator and a definite description; and did not intend to fix both references. Smart thought that ‘mental process type A’ may refer to different things in different worlds. However, according to Jackson, Smart should not have compared the identity of brain process types and mental process types to the identity of lightning and electrical discharge. Instead, he should have used the analogous example ‘Brain process type A is mental process type A’, just as ‘Plutonium is the most lethal poison known to man’. Where clearly, the plutonium example is not an identity between two rigid designators, but rather, between a rigid designator and a definite description.

However, if Smart intended for the proposed identities to be between a rigid designator and a definite description, then the ‘identities’ are not identities₁ at all. If we cash out the analogy offered by Jackson, a mental process turns out to be much different than what is expected. Consider the identity statement ‘Plutonium is Pu’, which is analogous to ‘water is H₂O’ in that both terms in the identity statement are rigid designators. Now, in the identity statement ‘Plutonium is the most lethal poison known to man’, Plutonium is a fixed reference, whereas ‘the most lethal poison known to man’ is a contingent description of plutonium ¾ a physically contingent (accidental) property of plutonium. However, if the identity statement ‘Plutonium is the most lethal poison known to man’ is genuinely analogous to ‘Brain process type A is mental process type A’, then, if we designate ‘Brain process type A’ as a fixed reference, then ‘mental process type A’ would be a physically contingent property of ‘Brain process type A’. That is, a mental process is just a contingent property of a brain process, and is not identical₁ to a brain process after all.

Thus, according to Smart (according to Jackson), a mental process is a predication of a brain process. Where a predicate is just a set of properties that may be contingently applied to (identified with or by) different individuals in different worlds. But is a mental process just a set of properties attributed to a brain process? If so, then clearly, unless the brain process is just a set a properties (attributed to a brain process), then a brain process and a mental process could never be identical₁. In fact, in his defense of Smart’s contingent position, Jackson posited that the term ‘identity’ may not be the correct word to use when describing Smart’s thesis.

This position has a number of serious implications. For instance, Lycan’s argument that mental events are not objects becomes irrelevant; for not only are mental processes not objects, but they are not events or processes either. A mental process is merely a descriptive predicate, applied to different processes in different possible worlds. And since the descriptive predicate (mental process type) is not identical to (but merely identified with) a rigidly designated brain process type, contrary to common understanding, Leibniz’s law does not even apply to Smart’s ‘identity’ theory.

Either Smart does not intend for his proposed ‘identities’ between brain process types and mental process types to be between two rigidly designated terms, or he does. However, I have shown that both routes are untenable. For if the identities are between two rigid designators then they, and the type identity theory as a whole, falls prey to Kripke’s two-part thesis. And, if not, then these strict ‘identities’ are not identities₁ at all.

Basically, Smart has two options. According to Jackson, if Smart genuinely meant identity₁, then he could and should simply concede that the proposed identity is necessarily true. However, if I was Smart, I would maintain the contingency of an identification (identity₂), and embrace functionalism.

Works Cited

Kripke, Saul A., "Naming and Necessity," in The Nature of Mind, edited by David M. Rosenthal, (New York: Oxford University Press, 1991).

Lycan, William G., Consciousness, (Cambridge, Mass.: MIT Press, 1987).

Smart, J.J.C., "Sensations and Brain Processes," in The Nature of Mind, edited by David M. Rosenthal, (New York: Oxford University Press, 1991).