73. For another study that explored both the contributions of early androgen exposure and parental socialization, see Constantinescu, Moore, Johnson et al. (2018).
74. McGlone (1980): 226.
75. Voyer (1996): 64.
76. Baron-Cohen (2003): 105.
77. Baron-Cohen (2003): 105–6.
78. Sporns (2011).
79. Smith (2017).
80. E.g., Sporns (2006); He, Chen, and Evans (2007). The model was initially proposed in Watts and Strogatz (1998).
81. The idea was made famous by Stanley Milgram. His data didn’t actually prove his point in the real world, though the idea behind it makes mathematical sense. Stanley Milgram, “The Small World Problem,” Psychology Today, May 1967: 60–67.
82. E.g., Kansaku, Yamaura, and Kitazawa (2000); Gur, Alsop, Glahn et al. (2000); Clements, Rimrodt, Abel et al. (2006).
83. Yan, Gong, Wang et al. (2011): 452.
84. Ingalhalikar, Smith, Parker et al. (2014): 825.
85. Hänggi, Fövenyi, Liem et al. (2014): 10.
86. Ruigrok, Salimi-Khorshidi, Lai et al. (2014). See Appendix 3 for details.
87. Ingalhalikar, Smith, Parker et al. (2014): 826.
88. Ingalhalikar, Smith, Parker et al. (2014): 826.
89. Satterthwaite, Wolf, Roalf et al. (2015).
90. Karl Friston, quoted in Yeo, Krienen, Sepulcre et al. (2011): 1127.
91. The method for observing functional connectivity is known as blood-oxygen-level-dependent (BOLD) imaging.
92. Satterthwaite, Wolf, Roalf et al. (2015): 2383.
93. Satterthwaite, Wolf, Roalf et al. (2015): 2390.
94. Tunç, Solmaz, Parker et al. (2016).
95. de LaCoste-Utamsing and Holloway (1982).
96. For an account of the pre-MRI years of the battle, see Holloway (2017).
97. Ardekani, Figarsky, and Sidtis (2013), which also contains a review of the reasons that previous studies found so many discrepant results. A subsequent study, Luders, Toga, and Thompson (2014), with a sample of 24 males and 24 females, argued that individual differences in brain size account for apparent sex differences in callosal anatomy, an argument made previously in the literature. Ralph Holloway, who started the debate with de LaCoste-Utamsing and Holloway (1982) and was invited to summarize the state of knowledge in the 2017 special edition of Journal of Neuroscience Research, concluded as follows: “I would like to end this commentary by breaking out in a happy dance, thanks to the most definitive article to date by Ardekani et al. (2012), which showed that, when males and females (N=316) were analyzed, the female CC [corpus callosum] was larger than that in males. A separate analysis (N=74) performed with a younger sample matched for brain size showed the same thing, and significantly so. It should be pointed out that a later article by Luders et al. (2014) with smaller samples and using different methods found no significant sex differences, but it does not mention the results of Ardekani et al. (2012). The recent article by Ingahalikar et al. (2014) on connectome differences between males and females clearly reinforces our findings on the CC. We always did our best to take the high road on these issues, and it is gratifying that many authors today mostly stay with the facts rather than political issues, which should not be playing a role in scientific discourse. May I say: ‘we told you so’?” Holloway (2017).
98. Ardekani, Figarsky, and Sidtis (2013): 2518. The text omits embedded citations contained in the original quotation.
99. Lombardo, Ashwin, Auyeung et al. (2012). Specifically, total white matter volume, total gray matter volume, and total cerebrospinal fluid volume were entered as covariates in the statistical analyses.
100. The regions in question were the right temporoparietal junction/posterior superior temporal sulcus (RTPJ/pSTS), planum temporale/parietal operculum (PT/PO), and posterior lateral orbitofrontal cortex (plOFC).
101. Lutchmaya, Baron-Cohen, and Raggatt (2001).
102. Knickmeyer, Baron-Cohen, Raggatt et al. (2006).
103. Chapman, Baron-Cohen, Auyeung et al. (2006).
104. Catani, Jones, and ffytche (2005).
105. Lutchmaya, Baron-Cohen, and Raggatt (2002).
106. Shaywitz, Shaywitz, Pugh et al. (1998); Rojas, Bawn, Benkers et al. (2002); Pelphry, Morris, and McCarthy (2005).
107. Lombardo, Ashwin, Auyeung et al. (2012): 679.
108. See, for example, Rubin, Yao, Keedy et al. (2017) and Koscik, Dan, Moser et al. (2009).
109. For the most complete available catalog of areas of the brain that are differentially activated by men and women in response to emotional perception, see Filkowski, Olsen, Duda et al. (2017), a meta-analysis of 56 studies limited to those that report direct contrasts between men and women participating in the same visual emotion-eliciting task within each study. The meta-analysis focused on which regions were activated, with broad characterizations of the results rather than specific hypotheses about how they related to phenotypic differences. Here is the authors’ overall conclusion (omitting references embedded in the text):
Here we have assembled 56 human functional imaging studies that each reported a reliable sex difference in neural activity elicited by emotional cues in the visual modality. This analysis reveals regions of the brain that are differentially recruited across men and women across a range of emotion-evoking experiments, including frontal, insular, and medial temporal cortex, as well as amygdala, brainstem, and high-order thalamus. This pattern of differences is consistent with a perspective in which men more strongly engage volitional control processes when faced with emotional cues, recruiting frontal cortical regions, perhaps resulting in dampened amygdala reactivity. Medial PFC activity in men could also be a result of enhanced reactivity to reward cues, although valence-specific effects cannot be determined in the current effort. Women, by comparison, show evidence of enhanced subcortical sensitivity to emotional cues, consistent with an evolutionary bias toward harm avoidance. (p. 930).
110. Declarative memory (“knowing what”) is distinct from procedural memory (“knowing how”), the kind of memory that lets us tie our shoelaces or touch-type without consciously thinking about it. Declarative memory is sometimes labeled explicit memory, and procedural memory is sometimes labeled implicit memory.
111. These examples are taken from Kurdi, Lozano, and Banaji (2016). The photos are in the public domain and available at www.benedekkurdi.com. Another common form of stimulus also uses pictures, but exclusively of faces. Other types of stimulus, used rarely, are films, words, autobiographical recall, script-driven imagery, prosody, and odors. Stevens and Hamann (2012).
112. Hamann (2005): 291.
113. Hamann (2005); Karama, Lecours, Leroux et al. (2002).
114. This is the predominant finding. For a literature review of studies of sex differences in amygdala volume, see Kaczkurkin, Raznahan, and Satterthwaite (2019), which notes that amygdala volumes increase rapidly in females in early puberty before peaking and decreasing, while males show increasing amygdala volumes throughout puberty. Marwha, Halari, and Eliot (2017), a recent meta-analysis that found no significant difference in amygdala size after correcting for total brain volume, disputes this, reporting that “our findings do not support a marked change in the sex difference at puberty, for either raw or corrected AV.” Quotation taken from p. 15 of the accepted manuscript.
115. Baird, Wilson, Bladin et al. (2004).
116. Goldstein, Seidman, Horton et al. (2001).
117. Karama, Lecours, Leroux et al. (2002).
118. Hamann, Herman, Nolan et al. (2004). Hamann (2005) notes that the previous study, Karama, Lecours, Leroux et al. (2002), had not found greater male activity in the amygdala, but that the same authors had conducted a follow-up using a method more sensitive to rapid changes in amygdala activity and had obtained results replicating the Hamann study’s finding of greater left amygdala activity in men.
119. For a review of the literature, see Seidlitz and Diener (1998).
120. One common instruction given
to judges in such experiments is to classify as gist “any fact or element pertaining to the basic story that could not be changed or excluded without changing the basic story line,” while classifying all other facts or elements as peripheral. This wording is drawn from Heuer and Reisberg (1990): 499. The authors continued: “Thus, for example, a description of a slide as showing ‘a woman crossing an intersection’ would be considered basic. To describe the same scene as ‘a person outside’ would be too general, while ‘a red-haired woman with a handbag on her left arm stepping onto the first third of the cross walk’ would be regarded as too specific. The judges agreed on almost all items; in cases of disagreement, items were replaced. Some sample peripheral items were the color of the door in front of which the mother stood in a particular slide; whether the boy was carrying a lunchbox or a soccer ball; and so forth. Some sample central items were whether the mother and son were going to the gas station, to school, or shopping; whether the son watched the repair, or played in the back room; and so forth.”
121. Cahill and van Stegeren (2003).
122. The initial identification of this effect was Cahill, Haier, White et al. (2001). For a review of the studies, see Hamann (2005).
123. Stevens and Hamann (2012).
124. Nolen-Hoeksema, Morrow, and Fredrickson (1993); Nolen-Hoeksema, Parker, and Larson (1994).
125. Thomas, Drevets, Whalen et al. (2001); Williams, Barton, and Kemp (2005).
126. Andreano, Dickerson, and Barrett (2013). The sample consisted of 25 females and 16 males.
127. Blackford, Avery, Cowan et al. (2011); Schwartz, Wright, Shin et al. (2003).
128. Blackford, Allen, Cowan et al. (2012).
129. Barrett and Armony (2009); Sehlmeyer, Dannlowski, Schöning et al. (2011).
130. Shin, Wright, Cannistraro et al. (2005).
131. Andreano, Dickerson, and Barrett (2014).
132. Brohawn, Offringa, Pfaff et al. (2010); Ramel, Goldin, Eyler et al. (2007).
133. Cahill, Uncapher, Kilpatrick et al. (2004); LaBar and Cabeza (2006).
134. Andreano, Dickerson, and Barrett (2014): 1393.
135. Herrera, Wang, and Mather (2018).
136. Kaczkurkin, Raznahan, and Satterthwaite (2019).
Part II: “Race Is a Social Construct”
1. In By the People, I defined it this way: “By the American project, I mean the continuing effort, begun with the founding, to demonstrate that human beings can be left free as individuals, families, and communities to live their lives as they see fit as long as they accord the same freedom to everyone else, with government safeguarding a peaceful setting for those endeavors but otherwise standing aside.” Murray (2015): xiii. The book argues that the American project is effectively already dead because of proximal causes other than slavery, but I believe that the ultimate cause was slavery and its aftermath.
2. Boas collected his writings on race in Race, Language, and Culture (New York: Macmillan, 1940), available online at monoskop.org. For a detailed description of Boas’s opposition to the biological interpretation of race and his allies in sociology, see Degler (1991): chapters 2 and 3.
3. Montagu (1997).
4. Lewontin (1972): 397. In the introduction, I noted that many aspects of the sameness premise have long since been disproved by ordinary social science evidence. Similarly, it has been widely argued since Lewontin published his dictum in 1972 that it was scientifically unjustified. For a full review of the technical literature that followed Lewontin (1972), see Rosenberg (2018). Nontechnically, Lewontin’s mistake was to conclude that if 85 percent of genetic variance is within populations, then it follows that the effects of the remaining 15 percent must be trivial. The effects could be trivial, but there was no a priori reason to assume so. Sewall Wright’s measure of genetic population differentiation, FST, proposed in 1943, is one of the most commonly used statistics in population genetics. Ranging in theory from 0 to 1.0, it expresses the ratio of variance among subpopulations to the total variance. But a value of 1 requires that each subpopulation be fixed for a different allele—an extreme seldom seen in practice—and anything less than fixation in each subpopulation produces far lower values of FST. Wright himself suggested guidelines for interpreting FST: Values from 0 to .05 indicate “little” genetic differentiation, from .05 to .15 indicate “moderate” differentiation, from .15 to .25 indicate “great” differentiation, and values of .25 and higher indicate “very great” differentiation. Wright (1978). If one applies these guidelines to humans, .15 is on the borderline between “moderate” and “great.” The meanings of .15 for FST and in Lewontin’s analysis are somewhat different, but Lewontin’s version of .15 is also not obviously trivial. In 1982, three geneticists from the University of Michigan and the University of Pennsylvania raised a more specific statistical issue. “There are a few marker loci which yield almost categorical separation of some human populations, but most loci vary modestly for allele frequencies across populations. The central taxonomic point is that for purposes of classification, which is a direct measure of taxonomic separation (distributional nonoverlap), a large number of small differences is equivalent to a small number of large differences.” Smouse, Spielman, and Park (1982): 445–46. The authors proceeded to demonstrate the applicability of their model to seven tribes of South American Indians. In 2003, an eminent British geneticist at Cambridge, A. W. F. Edwards, published “Human Genetic Diversity: Lewontin’s Fallacy.” He wrote that “this article could, and perhaps should, have been written soon after 1974 [when Lewontin published The Genetic Basis of Evolutionary Change].” Edwards then pointed out what statisticians had understood since the 1920s, that it was fallacious to analyze data “on the assumption that it contains no information beyond that revealed on a locus-by-locus analysis.… The ‘taxonomic significance’ of genetic data in fact often arises from correlations among the different loci, for it is these that may contain the information which enables a stable classification to be uncovered.” Edwards (2003): 801, 799.
5. Stephen Jay Gould, “Human Equality Is a Contingent Fact of History,” Natural History, November 1984: 26–33.
6. Stephen J. Gould, “The Spice of Life: An Interview with Stephen Jay Gould,” Leader to Leader 15 (Winter 2000): 14–19.
7. Omi and Winant (1986): 65.
8. American Sociological Association (2003).
9. American Association of Physical Anthropologists, “AAPA Statement on Race & Racism,” March 27, 2019, physanth.org/about/position-statements/aapa-statement-race-and-racism-2019/.
10. Diamond (1994).
6: A Framework for Thinking About Race Differences
1. I focus exclusively on the genetics in hopes of avoiding the problem faced by Nicholas Wade, who began his book, A Troublesome Inheritance, with chapters recounting developments in population genetics and evidence for recent evolutionary adaptation (corresponding to my chapters 7 and 8). Then Wade devoted the second half of his book to a larger set of topics that linked phenotypic differences to possible evolutionary explanations. Wade explicitly warned the reader that arguments in these latter chapters were speculative and “fall far short of proof.” But the response to the book conflated the explicitly labeled speculations with the hard science in the first half of the book. See, for example, David Dobbs, “The Fault in Our DNA,” New York Times, July 10, 2014.
2. Okbay, Beauchamp, Fontana et al. (2016): 60–61.
3. Sanger was awarded the 1980 Nobel Prize for Chemistry, which he shared with American biochemists Walter Gilbert and Paul Berg. Gilbert had invented another method for sequencing DNA that subsequently was bypassed in favor of Sanger’s. Berg’s award was for work on the chemistry of nucleic acids, especially those involving recombinant DNA.
4. More precisely, the sequence was said to cover 99 percent of the euchromatic human genome with 99.99 percent accuracy. Euchromatin refers to the 95 percent of the active genome that does not stain strongly with basic dyes when the cell is not dividing. The sequencing did
not include a tightly packed form of DNA known as heterochromatin.
5. Cells in humans are diploid, meaning that the cell carries two copies of each chromosome.
6. There are some rare exceptions to this rule, in which one allele in a SNP is consistently passed on to more than half of the offspring. The technical term for this phenomenon is meiotic drive (Sandler and Novitski (1957)), and it also is a source of evolutionary change.
7: Genetic Distinctiveness Among Ancestral Populations
1. I stipulate a particular population so that I can use actual data from Phase 3 of the 1000 Genomes Project. The subsequent numbers in the discussion are based on the allele frequencies for the EAS (East Asian) aggregation, using data downloaded from the 1000 Genomes website, www.internationalgenome.org. The selection criteria for being included in the download were that VT = SNP (i.e., the variant type is a SNP, not another form of variation) and AA = [ACGT] (i.e., the ancestral allele is A, C, G, or T, which culls most SNPs with more than two alleles). The calculation of probabilities uses the Wright-Fisher model for a biallelic SNP.
2. I use a round number because the selection criterion AA = [ACGT] does not cull all SNPs with more than two alleles and because counts of SNPs don’t yet stand still. They’re still changing as new methods yield more detailed and accurate maps of the genome.
Human Diversity Page 51