1. Introduction

Bechtel and Mundale (Reference Bechtel and Mundale1999) argue against the multiple realization of psychological states by neural states. They claim psychological states are not realized by type distinct neural states, since neural states across individuals and species are type identical. Part of their view involves showing that psychological and neural taxonomies coevolve. To support this, they provide evidence to show that researchers in neuroscience individuate neural states across individuals and species by finding commonalities.

Kim (Reference Kim2002) replies that Bechtel and Mundale support their view with the wrong kind of evidence. The evidence they offer concerns how neuroscientists individuate homologous structures across individuals and species. However, homologous structures are not the relevant kind of evidence. The relevant evidence concerns homoplasious structures across species, where the underlying structures are likely diverse. I argue here that Kim's criticism is mistaken. I think both kinds of structures are relevant to establishing multiple realization, and both provide evidence that neuroscientists can accommodate.

2. Kim's Problems with Bechtel and Mundale

Putnam (Reference Putnam, Capitan and Merrill1967) argued it is empirically unlikely that psychological states across species are type identical to neural states. The neural states underlying a human's and an octopus's pain are likely distinct. So the property of being in pain is not uniformly realized by the same kind of neural structure. The problem this creates is that, if psychological states are type identified with neural states, then they will be insufficiently abstract to characterize being in pain for all kinds of species. In that case psychologists would be unable to formulate laws about pain that apply across different structure types. To deal with the problem, Putnam argued, we need a more abstract notion of a psychological state. He suggested functional states are neutral on differences in the structural details of the states that realize them and appear suitable. By using functional properties psychologists can thus obtain their desired generalizations.

Bechtel and Mundale reply that neuroscientists, in fact, individuate neural states similarly across individuals and species. Putnam's mistake is to think that neuroscientists are committed to fine-grained principles of individuation. But neuroscientific research reveals that individuating criteria are broader than he suggests. As Bechtel and Mundale say,

They give as an example Brodmann's brain maps. The maps were developed by comparing brain regions in different species including the human, monkey, squirrel, and rabbit (Reference Bechtel and Mundale1999, 180). The maps show there are commonalities in brain regions across individuals and species in functionally important areas. In contrast to Putnam's claim, neural taxonomies correspond to one another. But since neural states are individuated as type identical, there is no evidence that psychological states are multiply realized.

Kim claims that Bechtel and Mundale appeal to the wrong kind of evidence. First, he says establishing multiple realization requires using evidence from two independent lines. Two individuals of the same species with type identical neural states underlying a psychological state are not proper evidence; “if the psychological property is realized by the same physical property in two individual organisms of the same species, this fails to provide relevant evidence for or against multiple realizability” (Reference Kim2002, 609). Second, when considering structures of different species, Bechtel and Mundale's case depends on homologies. Homologies are similar organs or characters in different species that are inherited from a common ancestor. For example, the cerebral cortex in humans and monkeys is homologous, since the two species share a common lineage. But homologies also are not a relevant kind of evidence. Kim says, “different species will tend to exhibit the same character if it is an inherited character that was present in their common ancestor…. It provides neither evidence for nor evidence against the multiple realizability thesis that brain structures are homologous” (Reference Kim2002, 609). The relevant kind of evidence is homoplasies. Homoplasies are similar organs or characters in different species that are not inherited from a common ancestor. An example is human and octopus eyes, since they have similar functions, but separate evolutionary histories. Returning to the brain case, Kim claims, “if brains had evolved several times in different lineages and still had the same physical-cum-structural properties, then we would have evidence against the multiple realizability thesis” (Reference Kim2002, 609). Although he thinks such an outcome is unlikely, his point is that homoplasies are the relevant evidence. It is homoplasies that can determine whether psychological states are multiply realized.

3. The Different Kinds of Evidence

Let me begin with the first problem. Kim claims that facts about individuals of the same species do not count as evidence. First, this is not how the issue is normally understood. Intraspecies facts are often cited as reason for thinking psychological states are multiply realized. Block and Fodor (Reference Block and Fodor1972, 160), early on, claimed the brain can achieve a given psychological end by a variety of neurological means in individuals of the same species. In addition, others (Horgan Reference Horgan, Wagner and Warner1993, 308) have appealed to differences within individuals to show specifically that the strategy of species-restricted neurological reductions is implausible. Intraspecies differences are thought by many to clinch the case for multiple realization. Even putting this aside, though, it's not obvious why Kim thinks intraspecies facts should not count as evidence. That individuals of the same species may have type identical neurological structures is a contingent fact. Before considering it, there is no reason to think that individuals of the same species must have the same neurological structures (humans are not like issues of a newspaper). There are surely differences in the genetic makeup and environmental history among different individuals. For example, nobody would think that all people must have the same eye color merely because they were human. Facts about individuals are thus relevant to whether multiple realization occurs (and, I think, can be dealt with the way Bechtel and Mundale suggest). Since Kim thinks multiple realization across individuals is not an issue, the point in a sense is moot. Individuals are not a problem for Bechtel and Mundale's view, but one needs to be clear about why this is.

A similar point applies to Kim's claim about homologous structures across species. There is no reason to think that homologies across species must always be structurally identical. But since Kim says such homologies do not present a problem, we can focus on his other claims.

The last issue Kim raises concerns homoplasies. He says homoplasies are organs or characters shared between species that are not inherited from a common ancestor (Reference Kim2002, 608). The concept of a homoplasy, though, has different meanings among biologists.Footnote ¹ There are (at least) two related notions to be distinguished: analogies and parallelisms. Analogies are organs or characters with a similar function and different evolutionary origin. Their similarity is due not to common ancestry, but to convergence. A standard example is human and octopus eyes, which are similar organs despite being evolutionarily unrelated. On the other hand, parallelisms are similar organs or characters in species with a recent common ancestor with a genetic disposition for the trait, but in which the trait is absent. An example is the stalked eyes that occur intermittently in lineages of acalypteran flies (Mayr Reference Mayr2001, 225). The flies have a genetic disposition to produce them, but not all the members of the lineage exhibit the trait. The notion of a parallelism is a distinct phenomenon from an analogy and treated separately.

But researchers disagree about how to classify these notions. Some think that analogies and parallelisms are both kinds of homoplasies, since the relevant traits are not expressed in the ancestral species (Kent Reference Kent1992, 25). However, others classify parallelisms differently. In their view, since common ancestry is still involved at some level, such traits represent a kind of homology (Heckt and Edwards Reference Hecht, Edwards, Hecht, Goody and Hecht1977, 7).Footnote ² Depending on which criteria are privileged the classification comes out different. My concern is not to settle which interpretation of these notions is correct. I only want to note that one needs to be clear about which kinds of traits are being considered.

Kim's concern seems to be analogies, since he mentions bird wings and bat wings. The wings have similar functions despite not being inherited from a common ancestor. Let us limit the term “homoplasies” to analogies, then, and let us grant they are relevant evidence concerning multiple realization. Kim claims the wings provide positive evidence since there are differences in their structure. As he says, the sets of wings have the same function, but are “constructed from different materials” (Reference Kim2002, 609). So the question we have to answer is: is Kim right that homoplasies are evidence for multiple realization?

Establishing multiple realization requires showing that the same function is realized by different types of physical states. Two steps are involved in the process. (i) One has to show that the physical state types in question are distinct. (ii) One also has to show that the relevant functional properties are type identical. Claims about multiple realization can be challenged at either step. And, at this point, one has an option. Which option one chooses depends on the particulars of the case. In some cases, one may want to argue that the physical types are the same at an appropriate level of description. In other cases, where the structures are clearly distinct, it is open for one to argue that the structures have different functions.

To see this take a standard case of homoplasies: human and octopus eyes. The structures are similar in certain respects, since they consist of a spherical shell, lens, and retina. But in other respects they differ (Messenger Reference Messenger and Autrum1981). They have different kinds of visual pigments in their photoreceptors. The octopus eye has one kind of visual pigment while the human eye has four. Furthermore, the eyes have different kinds of retinas and methods for focusing light. The retina in the human eye contains an array of rods and cones, and light is focused onto the receptors by bending the lens to change its shape. The retina in the octopus eye contains rhabdomeres, and focusing light involves moving the lens backwards and forwards within the shell. The parts of the structures are different and operate in different ways. But the differences are not merely present between the structures. The differences show up as differences in output. For example, having only one kind of visual pigment means the octopus is color blind. The structure of its receptors enables the species to perceive the plane of polarized light. And the structure of its retina means that compared to the human eye little processing occurs there (there are no ganglion cells in the octopus retina, and most information processing occurs in the optic lobe). The differences result in the eyes responding to different kinds of stimuli, having different reaction times, etc. So despite the structures having similar functions, it is plausible to claim that the functions are in fact distinct. That the eyes have similar functions does not mean that their functions must be identical.

Bechtel and Mundale make a related point about hunger (Reference Bechtel and Mundale1999, 203). Comparing psychological states in species like humans and octopi, psychologists often ignore functional differences.Footnote ³ At some level of description being hungry is the same state in both species. But this is only at a very abstract level. At a more fine-grained level, differences in the mechanisms associated with hunger-related behavior between the species reveal differences in behavior. Food-seeking behavior in the octopus is different from human's concerning how food is acquired, the digestion process, and so on. The behavioral differences between the species mean it is plausible that the functional states involved are distinct.

The same point applies to the psychological states Kim is concerned with. To establish multiple realization, it needs to be shown that the psychological states are realized by diverse neural states. But Kim cannot merely assume that the psychological states are type identical. In reality, the comparison of homoplasious psychological states across species is more complicated than it is usually realized. Psychologists sometimes talk about humans and species like octopi sharing the same psychological states. However, they also recognize that there are important differences involved depending on how finely one identifies the relevant features. Like in the hunger case, they are aware that psychological states in humans are not necessarily the same states in other species, and that the states are often only superficially similar (Allen and Bekoff Reference Allen and Bekoff1997, 179–180). This is especially the case when the states belong to species from widely divergent taxa. Furthermore, where there are neural differences in such species psychologists rightly look for differences in functional states, since their interest is with neural states that have functional relevance. They know that at a detailed level the neural differences make for functional differences. The problem for Kim is that establishing multiple realization requires showing that the same psychological state has diverse realizations. But we can always disagree with the functional taxonomy, and claim there are psychological differences at another level of description. So what needs to be established is that the psychological states are type identical, but it seems it will only be shown that the states have superficial resemblances.Footnote ⁴

Some philosophers are fond of pointing out differences in the neural states that underlie the same psychological states. They emphasize that whether the neural states belong to diverse neural types is an empirical question. But the individuation of psychological states is often treated as a straightforward matter, as if folk intuitions were sufficient to settle the issue. I believe that the individuation of psychological states is also a matter for investigation. The category of homoplasious states does not pick out items as uniform as some would like to think it does. Merely claiming that psychological states are homoplasious is not enough to establish they are truly the same states in different species.Footnote ⁵ The above examples serve to illustrate why this is. For similar reasons, I think it is plausible that the psychological states Kim is concerned with are distinct.

4. Conclusion

Kim thinks that Bechtel and Mundale's case uses the wrong kind of evidence. The appeal they make to homologous structures is irrelevant, and they ignore evidence concerning homoplasious structures across species. I have argued that Kim is wrong on both of these points. In addition, I suggested that homoplasious states do not present a problem for Bechtel and Mundale's view, since they can plausibly deny that the psychological states are type identical. When we examine the psychological states more closely we find that there are differences in the relevant functional states. In that case homoplasies provide evidence that can be accommodated. For if the psychological states are type distinct, then the same state will not have diverse realizations.