, , , ,

In part 1 I presented a metaphor of a toy maker making toy soldiers. At the end of the process the toy soldiers were all alike because certain forces were at work ensuring similarity. However, in the natural world things are not so neat and tidy. The forces involved are far looser than in the toy maker example and result in far greater variation than identical toy soldiers.

Unlike eternal kinds, historical kinds are not likely to ground many, if any, exceptionless generalizations. The copying processes that generate them are not perfect, nor are the historical environments that sustain them steady in all relevant respects… Historical kinds typically have naturally and irreducibly vague boundaries… Because the occurrence of causitive factors accounting for similarities can be more or less regular or irregular, and because the number of grounded similarities characterizing a substance can be larger or smaller, there are two different continua from richer to poorer along which historical substances can range. These reflect (1) the reliability of the inferences supported, and (2) their multiplicity. Substances vary widely in both these directions. If the substance is sufficiently impoverished in both of these dimensions, whether there exists a real kind at all can be a vague matter. (On Clear and Confused Ideas, p. 24 – 26.)

Millikan visualizes this with the following illustration:

Suppose that we were able fully to describe every individual in the natural world by locating it at a point within a huge multidimensional logical space each dimension of which represented contiguous determinates from a different determinable range. And suppose that the world were such that every point in this huge logical space was equally likely to be occupied. That is, between any two actual individuals there would run any number of paths in the space along which other individuals might be randomly realized. So there might be animals that were half way between rabbits and snails and others three fourths or nine tenths of the way, along any of innumerable different paths on incremental similarity, and likewise between computers and mountains, between newspapers and rivers, between eggplants and egg beaters, and between these and whatever you might dream up in your imagination.

Defining ‘kind words’ would have no point in such a world, for no knowledge could be accumulated around them.

As a matter of empirical fact, logical space is almost entirely empty. Huge gaps exist separating tiny clumps or clots of actual individuals or carving out miniature humps, peaks or ridges on the surface of these clumps. There are, for example, well over two million separate extant species of animals, but they don’t generally fade into each other, let alone into shoe horns and alarm clocks. Our most basic kind words do not name classes. They name these clots or various protrusions from these clots.

[A species such as] Weasels form a ‘clump; in that they have very numerous subtle and less subtle common properties, or largely overlapping properties, often with clean gaps separating these properties, or diagnostic sets of these, from the properties of other species.

The clump that is a commonly-named species may exhibit few if any properties common to all and only all of its members. In most species, for example, every distinctive gene has alleles. There can always be malformed and mutilated individuals, and many don’t survive for good reason. Yet for every species there are traits without end that are possessed by nearly every member, others that are possessed by most members, others possessed by many, and so forth.

Historical kinds need not be isolated clumps; they may correspond just as to easily discernible protrusions, bulges, or bumps. Biological species, for example, tend to be bulges, having emerged through rather unstable and narrow necks from earlier species, and the Gothic style emerged from earlier styles and then influenced later styles, merging with them broadly to form a sort of bump or rounded peak within architectural forms [I’d like to see a social constructivist argue that Gothic architecture isn’t real]. Historical kinds or subkinds may merge seamlessly with kinds that are significantly different along specifiable dimensions. (“On Knowing the Meaning,” Mind, 119 (473).)

Just as species are bulges in the animal property space, races are bulges on the human property space. Yet these bulges are real features of the world. I am going to try to illustrate this with a diagram, but keep in mind this is just for visualization purposes—it is not meant to be an accurate depiction of the massively multidimensional human property space! It’s just supposed to be a picture of a clump with bulges.


However, I like this image because it illustrates the points Millikan is trying to make. For one, there is much overlap between the “clumps” and they can fade into each other. Species might not fade into shoe horns and alarm clocks, but individuals can be located halfway between two racial groups, such as Barack Obama. Social constructivists often argue that the fuzziness of the boundaries is evidence that the clusters are not real. However, despite the fuzziness of the boundaries, the clumps are real discernible phenomena. Fifty-thousand years ago there was just one clump, but as the races emerged out of the human Pangaea, major distinct identifiable divisions have appeared.

Social constructivists also argue that there is no way principled reason why, say, northern and southern Europeans would be members of the same race rather than different races. However, the subclumps that are the races can have their own sub-sub-clumps as in the following:


This image is just the first with a couple of clumps added at the top. This would capture whatever differences exist between say, Japanese and Vietnamese, or northern vs. southern Europeans. We can visualize this as different ethnicities being smaller protrusions within the larger racial group. The clumps are clearly part of the larger major division of the property space despite having their own recognizable distinctiveness. (I could have put small clumps off of all the major clumps to make this general.) Perhaps the divisions could be even more fine-tuned; perhaps Slavs deserve their own clump, and the different Slavic groups even finer-grained divisions. None of this disproves that they are all part of the larger racial division.

Many people who deny that there are races nevertheless agree that there are “populations.”  As I see it, these people think that only the smallest subclumps exist and deny that there can be larger clumps on which they protrude.  There does not have to be either races or populations exclusively, there can be both.

Finally, consider this last image:


This image is just the second image with a clump added at around 4 o’clock. What to make of this clump? Does it deserve to be considered a major clump or not? The fact is, we may not know what to make of some groups, whether to consider them an ethnic or racial group or something else. Scientists might argue over this as like whether some group of animals should be considered a species or sub-species. There may not be a fact of the matter whether something is a race or not, or how many races there might be. The best we could probably do is to keep an eye on it, notice how it changes over time. If it grows in distinctiveness we may decide to call it a race, like a hill may be judged a mountain if it continues to grow. None of this affects the fact that the major groupings exist, and are real features of the world. Social constructivists would like to call everything, the major divisions and the minor, “populations” or the like. But this does a disservice to our understanding of reality; there are definitely major divisions among humans that are different from other divisions and it would be to misrepresent reality to believe otherwise in order to further whatever political purposes social constructivists obviously hope to promote through their distortion of reality.

Often the issue becomes a legal matter and demands a clear-cut answer, for example whether some individual belongs to a certain group and so is eligible for privileges.  In such a case a definite answer might be needed where reality is fuzzy.  As Millikan states, language and the law are digital while nature is analog.  For example, I have heard of cases where mountains are eligible for certain government restrictions whereas hills are not.  At some point a decision such as a mountain is a landform that is over over 1000 feet is made. Likewise, we might have to set a date in time for a population split to be considered a race, or a certain number of genetic base-pair differences.

A final point is that one’s position in the property-space does not determine one’s race. If an individual received massive cosmetic surgery in order to appear to be of a certain race they would not thereby be a member of that race because races are a matter of the possession of a certain history and this individual’s history determines them as a member of their real race, not imitated one. Likewise, were a pair of sub-Saharan Africans to have a child who, though an astronomically miraculous coincidence just happened to possess an incredible number of mutations that just happened to exactly mirror a Scandinavian’s DNA, this child would not thereby be a Scandinavian. Races are a matter of history and this individual would not be the result of Scandinavian history. Morphology and genetics are results of history, and we can use them to fallibly infer history, but it is the history and not the genetics or morphology that determine race. (Philosophy nerds will know that I am referring to Swampman here.)

But this does raise the issue of the conventions by which language attaches to the world, and how we conceptualize reality, which will be the topic of part 3.

I am excited to see that Millikan has an upcoming article called “What do Thoughts do to the World? Deflating Socially Constituted Objects,” in Studies in the Philosophy of Sociality, Volume I, where I expect she will do a far better job on the subject than I can do here.