Definition of Race

I planned to write a more extensive treatment on race, how it differs from species, how it is created, and how it is undone, but that will have to wait, as I do more research and try to clarify my thought further.

While we wait for that to happen, here is my interim Definition of Race:

Let P be a human population.
If:

1. all members of P pass an objective test T.
2. non-members of P don't pass T.
3. the vast majority of matings between members of P produce offspring that pass T.
4. the vast majority of matings including at least one non-member of P produce offspring that don't pass T.

then population P is said to be a human race.

I will update this entry explaining each point in the definition.

Update I

1. Membership in a race cannot be subjective, i.e., dependent on the observer. It must be determined objectively, i.e., any two well-meaning observers will assign an individual to a race via an objective procedure, a procedure that does not depend on the observer.
2. The second rule simply says that a race is a maximal set of individuals who meet its membership criterion. It includes all such individuals, since none of the non-members meet the criterion.
3. A race cannot be a random temporary combination of features. It must have the capacity to reproduce itself into the future. And most of the offspring produced intra-racially must pass the objective test (and thus be members of the race). Otherwise, in the space of a few generations, even if it is genetically isolated, the race will cease to exist. But, we don't want to claim that all the offspring will meet the test, because offspring are random combinations of their parents' genes + mutations. But, as long as most repeat the pattern of their parents, each race will be a long-standing reality.
4. Finally, the genetic pattern of a race, which allows it to pass the objective test must be something that is difficult for the rest of mankind to "fake." It must be something non-trivial, that is unique to the race that can be replicated from non-members very rarely - if at all.

I will continue updating this entry, discussing how it is different from traditional racial taxonomies.

Update II

One thing to note about the above definition is that it proposes a program for discovering races:

1. Identify a candidate population
   
2. Determine what common features they have, and how they differ from other populations
   
3. Write a formula that takes a number of features and produces an output: "member of race", or "non-member of race".
   
4. Prove that criteria 1-4 are met
   
5. Done!
   

Notice, that in practice we can't literally prove something for an entire population. Objectively assessing every individual will be expensive. So, necessarily we will be limited to samples.

What happens when an individual turns up who we've thought is member of race A, but doesn't pass the objective test? There are two options:

1. If the objective test can be updated in a way that the race is preserved, i.e., it continues to define a race according to the definition, then update it. The new individual has just given us some extra information about the particular race. For example, suppose that there are two alleles for a gene: A and a, and one of the criteria for inclusion in our race is to be aa. If a mutant allele b is found in the population, then we can update our test to include ab and bb individuals as well.
2. If the former option is not feasible, then simply the individual does not belong to the race. Our intuition that he ought to be genetically similar failed us.

An important point to make is that our definition allows us to define a race without worrying about a racial taxonomy for mankind in general. We simply have to see that a population exhibits traits and can reproduce them, while the rest of mankind doesn't. Then, that population is a race.

One can of course develop a set of such races, and if he defines their objective criteria intelligently, these might encompass a great part of mankind.

It is of course possible, indeed likely, that some individuals will never be able to fit in a racial group. I suspect that this is the case for "racially hybrid" individuals. But, that remains to be explored.

The recent progress in identifying continental human populations without any a priori assumptions by examining large numbers of polymorphisms is strong evidence to the effect that the outlined program is feasible. Indeed, such studies will help generate distinct clusters of human biodiversity. Subsequently, by looking at the clusters, we can see what they mean and write a procedure that assigns individuals to clusters, hence defining race.

It is entirely possible that competing sets of "races" might emerge by the procedure we have outlined. It all depends on the set of features that one chooses to use for their objective test. Two sets of features might be correlated, and indeed one can already define a procedure for identifying, e.g., Black Africans either via genetics or anthropometrics. The two are not independent: if you have lots of alleles found in black Africans, you are likely to look like a Black African.

But, the main point is that our definition of race forces us to be objective about race. When we write, e.g., "Caucasoid", we will have to establish an objective procedure that determines whether one is Caucasoid or not, and meet the other criteria as well. So, it will no longer be possible to accuse thinkers on race that their terms are without content, because they will be precisely defined.

Posted by Dienekes at May 3, 2003 03:16 AM | PermaLink