Removing children from family trios (N=24) did not alter the results, which were pretty much identical. Variance components are provided in table 1 and can be seen also at this link: https://docs.google.com/spreadsheets/d/16m9MZUgrhv_Bv8-bN6BRN46s1AckgNeWmAHQgYnbfK8/edit?usp=sharing
Please credit my work as:
Piffer, Davide; Dall'Olio, Giovanni Marco (2015): Using Vcftools to calculate Weir and Cockerham's variance components for 1000 Genomes phase 3 data. Figshare.
http://dx.doi.org/10.6084/m9.figshare.1450792

Removing children from family trios (N=24) did not alter the results, which were pretty much identical. Variance components are provided in table 1 and can be seen also at this link: https://docs.google.com/spreadsheets/d/16m9MZUgrhv_Bv8-bN6BRN46s1AckgNeWmAHQgYnbfK8/edit?usp=sharing
Please credit my work as:
Piffer, Davide; Dall'Olio, Giovanni Marco (2015): Using Vcftools to calculate Weir and Cockerham's variance components for 1000 Genomes phase 3 data. Figshare.
http://dx.doi.org/10.6084/m9.figshare.1450792

Thanks. I mentioned and cited the results.

The final copy is attached for publication.

[Versions removed -- to preclude confusion -- but view-able at OSF.]

Removing children from family trios (N=24) did not alter the results, which were pretty much identical. Variance components are provided in table 1 and can be seen also at this link: https://docs.google.com/spreadsheets/d/16m9MZUgrhv_Bv8-bN6BRN46s1AckgNeWmAHQgYnbfK8/edit?usp=sharing
Please credit my work as:
Piffer, Davide; Dall'Olio, Giovanni Marco (2015): Using Vcftools to calculate Weir and Cockerham's variance components for 1000 Genomes phase 3 data. Figshare.
http://dx.doi.org/10.6084/m9.figshare.1450792

Thanks. I mentioned and cited the results.

The final copy is attached for publication.

References should include a doi, when it's an electronic journal and it has a doi.

In this instance you don't even mention the journal name, making it impossible to cite it:

Piffer, D. (2015a). Estimating the genotypic intelligence of populations and assessing the impact of socioeconomic factors and migrations. Journal? Doi?

This is how you should cite it:

Piffer, D. (2015a). Estimating the genotypic intelligence of populations and assessing the impact of socioeconomic factors and migrations. The Winnower. DOI: 10.15200/winn.142299.93508

Also these articles have doi but you don't report it:

Piffer, D (2015b): A review of intelligence GWAS hits: their relationship to country IQ and the issue of spatial autocorrelation. Figshare. http://dx.doi.org/10.6084/m9.figshare.1393160

Piffer, D. and Dall'Olio, G. M. (2015): Using Vcftools to calculate Weir and Cockerham's variance components for 1000 Genomes phase 3 data. Figshare. http://dx.doi.org/10.6084/m9.figshare.1450792

Finally, you give the wrong reference for this paper, whose final version was published on Mankind Quarterly and not on IBC. This is the correct reference:

Piffer, D. (2013). Factor Analysis of Population Allele Frequencies as a Simple, Novel Method of Detecting Signals of Recent Polygenic Selection: The Example of Educational Attainment and IQ. Mankind Quarterly, 54, 168-200.

There are other omissions and incorrect statements.
When intra-individual variance was partitioned out, roughly the same percent of genetic variance was located between subpopulations as between individuals within subpopulations. The decomposition is shown in Table 4.10 below. Interestingly, Piffer and Dall'Olio (2015) found similar results when decomposing variance between major human races using 1000 Genomes phase 3.

1. Why do you not report the table in Piffer and Dall'Olio (2015)? So much work was spent on this just to give a brief (and totally inaccurate) mention?
2. That is an incorrect statement. The percent of genetic variance between subpopulations is much bigger than that between individuals, after variance within individuals has been accounted for. I have provided columns in my table showing variance in relative terms (%). % variance of between populations: Weighted Fst is a/(a+b+c) or within populations (B=relative proportion or b/(a+b+c). The average Weighted Fst is 0.0807 (roughly 8%), but the average B is 0.0049 (0.49%). So the variance between populations is roughly 20 times greater than that between individuals if we partial out the between gametes within individuals variance.

There is also an incorrect part in this statement, relative to improper citation of Minkov et al: "Third, frequency differences in specific cognitively associated alleles have been shown to predict aggregate national and racial IQ differences (Piffer, 2015a; Piffer, 2015b; Minkov et al., 2014).".

Citation of Piffer is correct because I had actually correlated allele frequencies to IQ scores. But Minkov et al. never did that. Their SNPs have only INDIRECTLY been associated with IQ at the within population level, unlike the SNPs reported by Piffer which were associated with g. As the authors state: "The serotonin-transporter gene is not an IQ gene, however. Holmberg and Lesch (2011) point out that S-allele carriers do not perform better on some typical IQ tasks, such as mental rotation.
In a study of 572 Chinese college students, He et al. (2010) did not find a clear association between 5-HTTLPR alleles and performance on standard IQ tests, such as Raven’s Advanced
Progressive Matrices and a Chinese Version of Wechsler’s Revised Adult Intelligence Scale."
I will publish after these errors and inaccuracies will be corrected.

Davide,

Yes, I didn't know how to properly cite those papers. I have made the corrections.

As for Minkov et al., If you read closely what I said, you will see that I said nothing incorrect. I also provided a clarifying footnote which you must have missed:

[95] These alleles are not necessarily general cognitive ability associated ones. For example, Minkov et al. (2014) note that: “The literature reviews and analyses by Dobson and Brent (2013) and Holmberg and Lesch (2011) suggest that S-allele carriers outperform L-allele carriers on a variety of cognitive tasks, such as financial decision making, probably because of their hyper vigilant decision-making style… The serotonin-transporter gene is not an IQ gene, however.” The frequency differences nonetheless evidence historic selection pressure on cognitively related traits in general; if there generally was such selection pressure, it would likely, though not necessarily, have acted also on general cognitive ability. The alleles discussed by Piffer (2015a) have been shown to be reliably associated with educational attainment and education achievement scores.

I didn't say IQ scores because at the time I wrote this, they had not been reliably so associated. I changed the last sentence to:

"The alleles discussed by Piffer (2015a) have been shown to be reliably associated with educational attainment, educational achievement test, and intelligence test scores."

As for, Piffer and Dall'Olio (2015), with the statement "Piffer and Dall'Olio (2015) found similar results when decomposing variance between major human races using 1000 Genomes phase 3," I meant similar results with respect to: "They found that between 96.7 and 99.6% of the variance was located within individuals." That is, you both found that most variance was within individuals.

I will clarify the point by changing this to: "Interestingly, when decomposing variance between major human races using 1000 Genomes phase 3, Piffer and Dall'Olio (2015) also found that most variance was located within individuals."

As for your question:

"1. Why do you not report the table in Piffer and Dall'Olio (2015)? So much work was spent on this just to give a brief (and totally inaccurate) mention?"

What I said was not "totally inaccurate", perhaps misleading. I did downplay your results. This was because they fly so far in the face of theory that reporting them makes my discussion look suspect. If you want me to add the table, then request that I do as a reviewer. Doing so as an editor, oversteps your role.

Davide,

Yes, I didn't know how to properly cite those papers. I have made the corrections.

As for Minkov et al., If you read closely what I said, you will see that I said nothing incorrect. I also provided a clarifying footnote which you must have missed:

[95] These alleles are not necessarily general cognitive ability associated ones. For example, Minkov et al. (2014) note that: “The literature reviews and analyses by Dobson and Brent (2013) and Holmberg and Lesch (2011) suggest that S-allele carriers outperform L-allele carriers on a variety of cognitive tasks, such as financial decision making, probably because of their hyper vigilant decision-making style… The serotonin-transporter gene is not an IQ gene, however.” The frequency differences nonetheless evidence historic selection pressure on cognitively related traits in general; if there generally was such selection pressure, it would likely, though not necessarily, have acted also on general cognitive ability. The alleles discussed by Piffer (2015a) have been shown to be reliably associated with educational attainment and education achievement scores.

I didn't say IQ scores because at the time I wrote this, they had not been reliably so associated. I changed the last sentence to:

"The alleles discussed by Piffer (2015a) have been shown to be reliably associated with educational attainment, educational achievement test, and intelligence test scores."

As for, Piffer and Dall'Olio (2015), with the statement "Piffer and Dall'Olio (2015) found similar results when decomposing variance between major human races using 1000 Genomes phase 3," I meant similar results with respect to: "They found that between 96.7 and 99.6% of the variance was located within individuals." That is, you both found that most variance was within individuals.

I will clarify the point.

As for your question:

"1. Why do you not report the table in Piffer and Dall'Olio (2015)? So much work was spent on this just to give a brief (and totally inaccurate) mention?"

What I said was not "totally inaccurate", perhaps misleading. I did downplay your results. This was because they fly so far in the face of theory that reporting them makes my discussion look suspect. If you want me to add the table, then request that I do as a reviewer. Doing so as an editor, oversteps your role.

Indeed I did not request that you add the table. I suggested it and I clarify it now, I suggest that as a reviewer. I do not think your paper would benefit from hiding results. You can present the table and then discuss why the results are so paradoxical but burying them is not very honourable. As editor I can only request that you make the changes to the references section. As a reviewer and a human being who has spent a lot of time working out the decomposition of Fst components, I would not approve of a paper who does not report the table.

Indeed I did not request that you add the table. I suggested it and I clarify it now, I suggest that as a reviewer. I do not think your paper would benefit from hiding results. You can present the table and then discuss why the results are so paradoxical but burying them is not very honourable. As editor I can only request that you make the changes to the references section. As a reviewer and a human being who has spent a lot of time working out the decomposition of Fst components, I would not approve of a paper who does not report the table.

I originally started to, which was why the title still reads:

"Table 4.10. Total genetic variance partitioned into variance between subpopulations, among individuals within subpopulations, and within individuals in a Japanese sample and in 1000 Genomes"

As can be seen, I forgot to delete the "and in 1000 Genomes".

Do you mind if I include just the big three: Africans, Europeans, and East Asians? Since I have three Japanese groups, doing so would be more symmetrical.

Indeed I did not request that you add the table. I suggested it and I clarify it now, I suggest that as a reviewer. I do not think your paper would benefit from hiding results. You can present the table and then discuss why the results are so paradoxical but burying them is not very honourable. As editor I can only request that you make the changes to the references section. As a reviewer and a human being who has spent a lot of time working out the decomposition of Fst components, I would not approve of a paper who does not report the table.

I originally started to, which was why the title still reads:

"Table 4.10. Total genetic variance partitioned into variance between subpopulations, among individuals within subpopulations, and within individuals in a Japanese sample and in 1000 Genomes"

As can be seen, I forgot to delete the "and in 1000 Genomes".

Do you mind if I include just the big three: Africans, Europeans, and East Asians? Since I have three Japanese groups, doing so would be more symmetrical.

Ok that should be fine

Indeed I did not request that you add the table. I suggested it and I clarify it now, I suggest that as a reviewer. I do not think your paper would benefit from hiding results. You can present the table and then discuss why the results are so paradoxical but burying them is not very honourable. As editor I can only request that you make the changes to the references section. As a reviewer and a human being who has spent a lot of time working out the decomposition of Fst components, I would not approve of a paper who does not report the table.

I originally started to, which was why the title still reads:

"Table 4.10. Total genetic variance partitioned into variance between subpopulations, among individuals within subpopulations, and within individuals in a Japanese sample and in 1000 Genomes"

As can be seen, I forgot to delete the "and in 1000 Genomes".

Do you mind if I include just the big three: Africans, Europeans, and East Asians? Since I have three Japanese groups, doing so would be more symmetrical.

Ok that should be fine

Updated. Please publish.
...

Version with updated publication date below.

To give the editor more time to publish I updated the publication date to (edit) June 20th.

The paper never covered Diamond 1994 very well.

"There are many different, equally valid procedures for defining races, and those different procedures yield very different classifications."

Sure I did -- all throughout sections I and II. I distinguished between natural and artificial classifications, noted that Diamond-like ones were artificial, and pointed out that races have general been understood to represent natural classifications in the sense of ones which:

(1) cut out genealogically defined divisions (lines of descent)
(2) which index propinquity of descent (meaning overall relatedness)

Diamond's "races" are simply molecular polymorphs. They surely exist -- are real in the commonsense sense -- and the concept is valid and meaningful. But it seems dishonest to say that these are what "races" are, when races were rarely to never said to be these by anyone except Diamond. Now you could argue that there is no reason to focus on natural divisions (such as races) instead of artificial ones (such as morphs). But this is a false dichotomy -- since there is no instead of. Sure morphs are interesting -- especially e.g., sexual ones. But that doesn't entail that races, so defined, aren't.

I'm still waiting for your argument. I will post it over at HV. If you can't think of a sound one, you should grant my position. We can then write up a summary paper on the matter.

"There are many different, equally valid procedures for defining races, and those different procedures yield very different classifications."

See also section "V-E. Onto-epistemology Arguments", in which I discussed a variant of the argument above.

All classifications are "artificial", so I don't get the distinction you are trying to make:

(1) cut out genealogically defined divisions (lines of descent)
(2) which index propinquity of descent (meaning overall relatedness)

(1) can only be done arbitrarily. A line of descent in evolutionary biology may be defined as an: 'ancestral-descendant sequence of populations'. Here's a quote from the zoologist G. G. Smith:

"Certainly the lineage must be chopped into segments for purposes of classification, and this must be done arbitrarily . . . because there is no non-arbitrary way to subdivide a continuous lineage.” (Principles of Animal Taxonomy, 1961, p. 165)

According to the theory of evolution, there is only one lineage (universal common descent: all organisms are related) and there is no way to objectively split this lineage because it is continuous.

Please try to think through what I wrote. I clearly distinguished between "arbitrary" and "artificial". And I distinguished between various senses of arbitrary. Try re-reading:

I-E. Natural Biological Divisions
II-D. What the Core Biological Race Concept Does Not represent
II-F. Clarification on the Meaning of "Arbitrary" and "Objective" in Context to Natural Divisions
III-B. Biological Races and Biological Reality
IV-E. THRs and Biologically Objective Races

Yes, species were once thought of as being "non-arbitrary"/"real"/"natural" in a deep ontological sense. They were thought to represent independent creations (like the elves and dwarves of Middle Earth) and to be permanent in the sense of unchanging across generations. And, yes, post-Darwin it was realized that no biological divisions were "non-arbitrary"/"real"/"natural" in this species realist sense. Hence I noted:

This last point is important. Post-Darwin species were, in a sense, ontologically demoted. The result is that all lineage segments are now conceived of as unreal in the species realist sense and, hence, not "real natural kinds" understood thus. At the same time, though, what it is to be biologically natural was reconceptualized in a genealogical sense without the "natural kind" presumptions; hence: biological natural divisions. We see this meaning shift play out in Darwin's writing...The upshot of these considerations is that races can now be said to be as real – or unreal, meaning depending – as species. Relatively speaking, then, the ontological status of "Negroid,” “Caucasoid,” and so on has moved up quite a bit since the age of the Enlightenment! Of course, Darwin recognized this. In Descent of Man, having noted that the distinction between species and intraspecific race was fundamentally arbitrary, he pointed out that human geographical groups could be considered to be either. This is because, with his theory, the ontological gulf between specific and intraspecific variation vanished...Given the shifting nature of understandings, explicitness of meaning is necessary for coherence. If one proclaims that “races are unreal” where “unreal” is used in the archaic sense of permanent or extra-mental, one should make one's meaning clear... On the other hand, they are not biologically real in the way that species were once thought to be. To better make sense of the issue, it would be helpful if, in the future, biological race anti-realists offered examples of the types of biological things that, by their understanding, are "really" biological real.

So, yes, you could say that species, races, and all other biological divisions and entities are "unreal", "unnatural", and "arbitrary" in the way that Linnaean species were not thought to be. I said as much.

I just asked for consistency.

And for clear definitions of terms. I am very clear on how I define the terms I use. I expect the same of others. I say for example:

Since, from our contemporary perspective, species realism – and taxonomic realism in general – is false, one could say – and it sometimes is said – that all natural divisions are arbitrary. They are just points along an evolutionary continuum. This being the case, though, does not entail that natural divisions are empirically arbitrary in the sense of made willy-nilly. If some criteria are specified, if the criteria are empirical, and if we consistently group according to these, groupings are by definition empirically non-arbitrary...

Now, again, I am not inclined to waste time with word games or to struggle over terms. We can just agree that races as divisions of a species are not non-arbitrary/real/natural in the species realist senses -- and were never thought of being so -- and that they are non-arbitrary (in arrangement)/real/natural in the senses I mean. We can then discuss which senses have more currency. Agreed?

Krom: Above though has nothing to do with what Diamond (1994) was saying. I understand that if we are talking about genetics, what he wrote is now outdated because we can now show with advances in genetics that x is overall more similar to y than z etc in terms of their genome. Diamond's criticism though extends to non-genetic criteria of race: why should we choose genetics over something else? And non-genetic criteria will produce a different racial classification.

Nothing at all? Let's look at the argument:

As it turns out, this seemingly unassailable reasoning is not objective. There are many different, equally valid procedures for defining races, and those different procedures yield very different classifications. One such procedure would group Italians and Greeks with most African blacks... As a result, if you classified golden whistlers into races based on single traits, you would get entirely different classifications depending on which trait you chose... A method that could in principle overcome these problems is to base racial classification on a combination of as many geographically variable genes as possible. Within the past decade, some biologists have shown renewed interest in developing a hierarchical classification of human populations...

According to Diamond, discordant variation potentially posses a problem for racial classifications. Because there are discordances one could, in principle, make all sorts of classifications "based on single traits". (These would be, by definition, artificial ones.) Diamond notes, though, that one could overcome this problem by using combinations of variable genes. (These would happen to produce, by definition, natural classifications.) Now, we know that when one does use "a combination of as many geographically variable genes" ones gets, at a certain level of focus, the traditional "major races" which happened to be said to represent natural divisions (in the propinquity of descent/overall relatedness sense). So, what problem do you see?

It seems purely semantic. If we understand races to be natural divisions (defined by propinquity of descent), we don't have a problem with trait discordances. If we understand them to be e.g., morphs (defined in terms of singular trait similarities) we don't have races in the original sense.

I noted:

We might, of course, wonder why we should – given that nature does not hand us the concept – think of it this particular way. Why not, for example, conceptualize “race” just as we conceptualize “morph”? This line of inquiry mistakenly reifies the term “race.” “Race”, here, is the concept. To think of it differently is to think of a different concept. We might more properly ask, though: why think of this particular concept? The obvious reason is that there is something out there – not well captured by other concepts – that we would like to describe.

Now, while on the topic, I noticed that Diamond made a couple of obvious errors too:

1. "Thus all human populations, no matter how different they look, belong to the same species because they do interbreed and have interbred whenever they have encountered each other." -- Depends on the species definition. In the 90s the phylogenetic species concept was rather popular, why was he oblivious to this?

2. "Hence they are classified into two different races, or subspecies (alternative words with identical meanings), termed the myrtle and Audubon races, respectively" -- As I showed, this is clearly not the case.

Duplicate. Deleted.

Here's a quote from the zoologist G. G. Smith:

"Certainly the lineage must be chopped into segments for purposes of classification, and this must be done arbitrarily . . . because there is no non-arbitrary way to subdivide a continuous lineage.” (Principles of Animal Taxonomy, 1961, p. 165)

By the way, that was G.G. Simpson. I quoted him on a related matter:

"In relation to natural divisions in general, this point was noted by Simpson (1961):

"The point will be discussed later, but even here it is advisable just to mention that such arbitrary subdivision does not necessarily produce taxa that are either 'unreal' or 'unnatural,' as has sometimes been stated. A simple but, at this point, sufficient explanatory analogy is provided by a piece of string that shades continuously from, say, blue at one end to green at the other. Cutting the string into two is an arbitrary act, but the resulting pieces are perfectly real section of the string that existed as natural parts of the whole before they were severed."

Yes, but this was not available in 1994. In the last 21 years it has been made possible e.g. the Human Genome Project. What I bolded was Diamond's concern in 1994 and the reason he wrote "different procedures yield very different classifications"

.

And yet I said:

NofR: One variant of the bio-statistical argument runs: since the level of genetic differentiation between populations is low, individuals can not accurately be assigned to natural divisions... This argument has been shown to be unsound. For the explanation why, readers are referred to the discussions of Mitton (1977; 1978), Risch et al. (2002), Edwards (2003), Witherspoon et al. (2007), Gao and Martin (2009), and Tal (2012). What is odd is that the idea of taking into account multiple indexes when making a racial classification is hardly new. Blumenbach in 1806 passingly noted that human racial classifications, being classifications in a natural system, should be based on “all bodily indications alike." Darwin noted that human racial classifications should be based on a full pedigree; moreover, he discussed the importance of correlated variation when it came to classification in general. In 1950, William Boyd showed how one could use multiple genetic loci to make such a classification and noted that one should use all possible genetic loci when doing so. And Lewontin (1978), in reply to Mitton (1977), agreed that it was obvious that one could divide humankind into biological races using multiple loci.

If such was not possible prior to 1994, why did Lewontin concede to Mitton that this could be done in 1978 and how was Dobzhansky (1970) able to discuss Boyd's (1950) multi-locus analysis? The deeper point is that Buffon, Blumenbach, Darwin and others argued that one should use correlated variation -- or an ensemble of characters -- when it came to picking out natural divisions and races as these.

My point though above was that Diamond's "different procedures yield very different classifications" extends to non-genetics and this is why human races cannot be shown to be "real" because why should we choose genetics over non-genetic criteria? Why not define race by something else? You do no cover this in the paper.

But I do! I say:

NofR: Next, we must clarify what it would mean for a biological concept to be a biological race concept. Such a biological concept need not always be called “race,” nor does the term “race” need always refer to the said concept, but any biological race concept must have a reasonable claim to the term. Obviously, there is a subjective element involved when it comes to assessing what constitutes a “reasonable claim.” Nevertheless, examples of biological concepts which qualify (microgeographic race, etc.) and do not qualify (sexual morph, life form, etc.) come readily to mind. Biologists and others who discuss race in relation to biology have frequently referred to the former, but not the latter, as “race.”...

NofR: Prior to being incorporated into natural history in the 18th century, the term "race" referred to breeds and lineages. “Race” was employed, in a related fashion, in several different discourses. In context to animal husbandry, it was used to describe strains of animals produced through linebreeding...This notion was integrated into natural history largely by Comte de Buffon, Immanuel Kant, and Johann Blumenbach to make sense of a particular sort of intraspecific variation, which was sometimes called the “constant variety.”

NofR: Generally, granting that races are intraspecific groups that differ genetically/genealogically on average, which is what they are now invariably said to be, they could be: (1) artificial divisions such as morphs or forms, (2) natural divisions, (3) spatiotemporally defined populations, like Iowans and North Carolinians, which, while differing genetically on average, do not cut out genealogical/genomic divisions, or (4) genealogy-defined divisions which are not so linebred that members are more overall genetically similar to other members of the same division than to members of other divisions. This is the manifold of possibilities. Proponents of the race concept...

Must I spoon feed you the argument!

There is this concept:

"organismic groups which differentiated from one another as a result of historic patterns of filiation; they are groups, which due to histories of sufficient linebreeding, form intraspecific natural divisions, ones which can be identified based on the correlations between the organisms' inherited characters"

You agree that these types of groups or divisions exist, no?
You agree that these types of groups or divisions were often called races, no?
You agree that the groups or divisions traditionally called "major (biological) human races" correspond with these types of groups or divisions, no?

[Please answer the above.]

So what possibly is the dispute?

The argument is:

(a) there is some valid, meaningful, and utile natural historian concept, which describes a type of biological variation not well described by other concepts
(c) which, based on historic term usage, has a reasonable claim to the name race
(d) by which there are human races
(e) which roughly correspond to the divisions historically called races

Sure you could call another concept "race". I note that. The reason you would call this concept race, though, is because this what it was often called. But you ask:

Why should we choose genetics over non-genetic criteria? Why not define race by something else?

And the answer is: Because "race", the term, originally meant -- at the time of introduction -- lineage/genealogical division/strain/breed. It was always tied to the idea of genes in the sense of genealogy. So if you were to use the term in a historically consistent manner, you would identify races with lineages in some sense. That leaves but two options, which I discussed:

NofR: We imagine that few to no researchers have conceptualized biological races so loosely as to include mtDNA lineages or all descendants of Genghis Khan, even though doing so would be consistent with the historic extra-scientific usage of the term, as expressed by the phrase “the race of Caesar.”... Regardless, we are confident that few would consider extended families which are so little linebred that many members could not be, based on propinquity of descent, assigned exclusively to this or that division to make for good biological races...As suggested above, we might further conceptualize biological races as representing overlapping ancestrally defined groups. For example, in the case of humans, the set of individuals descended from Charlemagne and that descended from Confucius could be said to represent two overlapping biological races, with membership assigned on a hypodescent basis. These divisions would overlap because while many individuals share only one or the other lineage many share both... such undoubtedly constitute meaningful biological divisions of a sort, they do not constitute, as we are characterizing them, natural divisions, since they are not being defined in terms of overall genetic similarity. Such divisions could be said to represent artificial biological races proper...With artificial biological race concepts, members are arranged by ancestry, but not overall ancestry. Hypodescent defined races would be an example of such. They can, insofar as the defining genealogies are accurate, be said to be no less biological than morphs or forms; yet the groups described would frequently not be natural in the sense of describing overall genetic relatedness. When lines of descent are sufficiently endogenous, that is, when our extended families are sufficiently linebred, natural division races emerge from artificial.

You either have races as artificial genealogical divisions or races as natural (genealogical) divisions.

I am not opposed to the idea of the "race" of Confucius. This, though, would not correspond wit the natural historian concept discussed in the 1700-1800s or the biological concept discussed after.

Also Sesardic who you quote is wrong when he claims multiplying the number of skeletal traits in forensics increases the accuracy of "race" identification. I can show this later in more detail, but just look up statistical noise. This is a known problem with FORDISC and why fewer measurements of the skull (about 10) are more reliable than 50+ Howells often used.

So... unlike with neutral genetic loci, with craniometric traits, some of which are under neutral selection, some of which are not, there is a nonlinear positive relation between number of characters and accurate classification into divisions defined by propinquity of descent ~ genomic similarity? Were did Sesardic as quoted claim otherwise?

And as far as I am aware we cannot demonstrate anything is "real"/"natural", this includes categories of being. Anyone claiming something is "real"/"natural" I think is mistaken because there is no way to show this.

If you wish to say that nothing can be known to be Krom-real, that is fine with me as long as you are consistent and clear about your meaning.

I would go a step further and say that we cannot demonstrate that any concept of "real" is the real one or that any concept of "natural" is the natural one. Thus we can not say for sure that anything is not really real or naturally nature. Thus races/species/populations/morphs/ecotypes/etc. may or may not be really real and naturally natural. Since we can not know, we can only speak about "real" and "natural" in this or that lower case sense -- which is what I set out to do:

NofR: If our philosophers can not decide what it really means to be "biologically real,‟ nor determine the nature of a „natural kind,‟ we can not hope to do so, and therefore can not even begin to try to defend the real biological reality or true 'natural kindhood' of race – so we will not...Instead, we will simply sketch out a biological concept of race, defend its biological validity, explain how this concept applies to human beings, make clear the sense in which race is natural and real, and criticize the usual arguments presented by opponents of biological conceptions of race.

If you wish to argue against a "race capital-R Realist", you are going to have to find someone else. Now, you say:

You mention Phillip Kitcher in the paper, now he adopts philosophical pragmatism for this reason: species and races are to him not "real" or "natural kinds" but he defends them on the grounds they are useful biological categories.

Kitcher seems to vacillate with his usages, but he often refers to his pragmatic kinds as natural kinds[/i]. Consider some passages from the 2007 paper I cite:

PK(2007): So, for brevity, I shall call the position I have sketched a pragmatists account of natural kinds...This version of pragmatism about kinds can be defended by focusing on the pluralistic character of taxonomic practices in the sciences, especially within biology...My aim is to explore the consequences of this pragmatic approach to kinds for the naturalistic proposal about races outlines above...(Does ‘race’have a future?)

In this paper at least he is not contrasting "natural kinds" with "pragmatic kinds" but understanding the former as the later. Don't you agree? You continue:

So yes, this is mostly a semantics issue since you are defining "real" or "natural" as something that is biologically meaningful, when this does not make it "natural" or "real", over conceptual and artificial.

Let's just agree that Chuck-real/natural is not equivalent to Krom-real/natural. And that when Chuck establishes that races are Chuck-real/natural this doesn't establish them as Krom-real/natural. And vice versa. You continue:

If we all just adopted pragmatism, the race debate would be a lot easier to follow: it would just come down to the question of how useful is race (as a category) in capturing human biological variation..But my point was to highlight the concept of race is not one of "reality" but it being a useful research tool (which i dispute).

If you carefully read through my section I-K, you will see that I do not argue that the concept of race is real. Following Levin (2002), I noted that the "ontological status of concepts is another matter". I argue, rather, that races -- the referents (understood in line with the concept) -- can be real. In a parallel manner, I wouldn't argue that the concept of dinosaurs is unreal/real, just that dinosaurs -- the referents -- are.

In short, I can not concur with you. I believe that there is a commonsense sense of "real races" that is worth discussing. Specifically, it's worth discussing whether my concept references anything in existence. I noted that in the Linnaean frame race so understood referenced nothing -- thus it is not a truism to say that "races are real" (in this commonsense sense). Now, you apparently don't like my commonsense sense of real. Well, how should I describe what I mean? I say, for example that hydra, the animal marked by radial symmetry, is real but hydra, the many-headed monster, is not. Most people, I think would understand my meaning. But how would you, who wants to say that the reality of both types of hydra is indeterminable, phrase things (to communicate the point)?

Thanks for the Baker reference, though. Now, you conclude:

"Most anthropologists have abandoned the concept of race as a research tool and as a valid representation of human biological diversity." (Sauer, 1992)

Granting your epistemology, I don't get how this would work. In section V-B ("Human Biological Races and Scientific Consensus"), I discussed regional variation in the acceptance of concepts of race. In some countries it's high; in some, it's low. Would race concepts (?) -- or races the referents -- then be real in countries where there was sizable acceptance, but not in others? That sounds like an unwieldy epistemology of science, no?

I don't imagine that acceptance would be a problem, though, since I am concerned with a generic biological or natural historian concept of race, not an e.g., cultural anthropological one or one specific to humans. And there is zero chance that biologists would reject my concept as inutile -- at best they might deem it worthwhile to call it something else. Now, Kitcher seems to think that biologists could reject the concept for humans but not for non-humans. But that doesn't make any sense to me, at least when we understand the concept to be inherently generic (across species, as it was originally conceptualized). Do you disagree?

It sounds as if we are retreading old ground. Perhaps it would be easier to approach the matter in the form of a Socratic dialogue. You could ask questions and see if you could trap me in contradictions. If not, we could switch roles.

Unless you can quantify/specify the whole genome, you have an arbitrary classification because you're left with an indefinite measurement (and the genetic variation not measured could produce a different number of "races")...

Let us first recognize that I am not advancing a novel position.

Darwin: Naturalists try to arrange the species, genera, and families in each class, on what is called the Natural System. But what is meant by this system? Some authors look at it merely as a scheme for arranging together those living...I believe that something more is included; and that propinquity of descent, the only known cause of the similarity of organic beings, is the bond, hidden as it is by various degrees of modification, which is partially revealed to us by our classifications...The importance, for classification, of trifling characters, mainly depends on their being correlated with several other characters of more or less importance. The value indeed of an aggregate of characters is very evident in natural history. Hence, as has often been remarked, a species may depart from its allies in several characters, both of high physiological importance and of almost universal prevalence, and yet leave us in no doubt where it should be ranked. Hence, also, it has been found, that a classification founded on any single character, however important that may be, has always failed; for no part of the organisation is universally constant. The importance of an aggregate of characters, even when none are important, alone explains, I think, that saying of Linnaeus, that the characters do not give the genus, but the genus gives the characters; for this saying seems founded on an appreciation of many trifling points of resemblance, too slight to be defined...I believe that the arrangement of the groups within each class, in due subordination and relation to the other groups, must be strictly genealogical in order to be natural.

To summarize: For Darwin, a natural classification is one where individuals are arranged into divisions by propinquity of descent. Since full pedigree tables are unavailable, heritable characters are used to index genealogical relationship; correlated characters -- of little importance -- are used because in aggregate they better index ancestry.

(Note, in section II-H and elsewhere, I do note a slight complication with this formulation in light of some contemporaneous re-understandings of what it means to be a biological natural division. But I wave it away by noting that genomic similarity ~corresponds with propinquity of descent on the individual level and that one can just stipulate IBD (genomic similarity) or subordinate propinquity of descent to IBT. For the sake of simplicity, I will adopt the pedigree understanding of natural divisions. But one could just as well used an IBT genomic one.)

Now, your initial point is that our correlated characters imperfectly index true pedigree, just as say color board-measured color imperfectly indexes true color. This is no doubt true, and I note this point in no less than 3 sections (e.g., p. 43, p. 95-97, p. 134). From here, from the point that our correlated character based classifications are not perfectly reliable with respect to true ancestry, you seem to deduce that they are artificial. If you do, you confuse conceptual dichotomies: measure versus latent variable and artificial versus natural classifications. An artificial classification purports to arrange individuals by specific character resemblances e.g., color, a natural one by (genealogical) affinity. Due to almost inevitable non-zero measurement error actual classifications imperfectly correspond with intended (latent) ones. This is a non-interesting point.

I say "you seem to" because I am not 100% sure about your argument. You continue for example:

" Likewise, the list of differences could be infinite… any two entities can be arbitrarily similar or dissimilar by changing the criterion of what counts as a relevant attribute." (Murphy and Medin, 1985)"

This statement seems to suggest a different argument. The problem here is not that our measure is imperfect, per se, but that a system based on "overall similarity in characters" can produce inherently discordant classifications, because, after all, you could weight or define characters differently. If this is the real critique you missed the point. The classification is not based on "overall similarity in characters" in a phenetic sense, but rather is based on propinquity of descent. To this end characters are weighted by -- biased by -- the genealogical information content of the characters. The issue then reduces back down to one of whether our selected characters (our measures) index descent (our latent variable) -- a measurement issue of little theoretical importance. To be absolutely clear, the fix -- the weighting value -- would be propinquity of descent, since this is what our classifications, insofar as they are drawn in line with our concept, purport to arrange individuals by.

Now, I guess you could point out that there is no reason to choose "propinquity of descent" as a weighting variable. Why not, after all, weight by "propinquity of color". But this returns us back to section "V-E Onto-epistemology Arguments".

But you continue:

Even if you could show all that in regards to genetics, the classification (outside the constraint) would still be arbitrary. Look at the plum and lawnmower example. Can we show a putative plum is overall more similar to another plum, than a lawnmower?

Here again you conflate issues. I can show that classifications highly reliably arrange individuals by propinquity of descent and thus are natural in the Darwinian sense. I can't show that this arranging is the only or "True" way of arranging things -- because it isn't and I nor anyone else is claiming that it is.

Before responding to your other post, I will give you a chance to clarify. To summarize my points here:

(a) We are not concerned with perfectly reliable classifications so perfectly reliable indexes are not needed.

(b) Indexes are weighted to produce classification which arranges individuals by propinquity of descent.

(c) The a priori reason to weigh indexes this way is so that out classifications stand in accords with our concept.

(d) There obviously is no a priori reason to think about this instead of that concept -- we do because there is something out there which we wish to think about.