The Modern Critique of Biological Race in Forensic Anthropology

Human genetic diversity follows a clinal pattern. Allele frequencies at any given locus vary continuously across geographic space, with adjacent populations sharing more genetic variants than distant ones. There are no sharp genetic discontinuities that correspond to the traditional racial categories. When Richard Lewontin published his analysis of protein polymorphism data in 1972, he showed that approximately 85 per cent of human genetic variation occurs within populations and only about 15 per cent between populations: the within-group variation is far larger than the between-group variation.

Lewontin's finding has been replicated with successively more powerful tools. Human Genome Project-era microsatellite studies (Rosenberg et al. 2002, using 377 microsatellites in 1,056 individuals from 52 populations) showed that self-identified population groups can be statistically recovered in cluster analyses, but these clusters reflect continuous geographic gradients rather than discrete categories. The number of clusters returned by the analysis depends on the K parameter chosen by the investigator, not on a biological fact about human populations. Subsequent genome-wide SNP studies (Li et al. 2008; Reich et al. 2009) confirm the clinal structure and show that the boundaries between "races" are geographic and historical accidents, not biological realities.

Cranial morphology follows the same pattern. Phenotypic variation in skull shape, facial dimensions, and nasal morphology is continuous across human populations and correlates primarily with geographic distance and ancestry (in the population-genetic sense), not with discrete categorical membership. The fact that a discriminant-function analysis can statistically separate groups of skulls from documented skeletal collections does not mean those groups represent discrete biological categories. It means the groups differ, on average, in measurable ways. The discriminant function exploits those average differences. It does not imply that any individual skull falls into a discrete natural kind called "Caucasoid."

The same discriminant geometry that sex estimation from the skull and mandible uses is what population-affinity analysis uses, applied to a different population-level question. The AABA position statement (2020) states this directly: "Race is not a biological reality; it is a cultural, social, and legal construct. Variation in human biology does not map onto discrete categories (races) but rather exhibits clinal, mosaic patterns of change across geographic space." The statement also notes the historical entanglement of the racial-typology tradition in forensic anthropology with colonial-era classification projects, some of which were used to justify discrimination. This history is not irrelevant to the scientific critique: it shaped which skeletal collections were assembled, which measurements were taken, and which typological frameworks were applied to the data.

The parallel point, equally important for forensic practice, is that the critique of biological race as a discrete category does not eliminate the empirical finding that populations differ in cranial morphology in ways that can be statistically characterised. What it eliminates is the inference from that finding to the claim that those differences reflect bounded natural kinds rather than fuzzy geographic clines.

The American Association of Biological Anthropologists (AABA) 2020 position statement on race and racism in biological anthropology was the product of a committee that included forensic anthropologists, evolutionary biologists, and population geneticists. Its forensic-anthropology implications were direct: the statement endorsed the shift from racial typological language to population-affinity language and explicitly discouraged the use of the terms Caucasoid, Mongoloid, and Negroid in scientific and forensic reports.

The Society for American Archaeology's companion statement (2020) addressed the use of racial categories in archaeological and bioarchaeological contexts, reinforcing the position that skeletal analysis should report population affinity in terms of similarity to documented reference groups, not membership in racial types.

SWGANTH (Scientific Working Group for Forensic Anthropology), the US standards body for forensic anthropological practice prior to its OSAC successor, had already been moving toward the population-affinity framing in its 2013 guidelines. The SWGANTH guideline on biological profile reporting recommended using the term "population affinity" rather than "race" and recommended reporting the metric basis for the affinity assessment (i.e., the FORDISC reference groups compared, the posterior probability, the typicality probability) alongside the morphoscopic observations. This was not presented as a political position but as a methodological recommendation driven by the need for testimony to survive Daubert scrutiny.

The NIST forensic anthropology subcommittee, operating under the OSAC (Organisation of Scientific Area Committees) for Forensic Science, issued guidance in 2020 explicitly recommending that forensic anthropology reports avoid racial typological terminology. The NIST guidance noted that "Caucasoid," "Mongoloid," and "Negroid" are not operationally defined terms with reference standards: no forensic laboratory has a validated measurement protocol that maps specimens into these categories with known false-positive and false-negative rates. FORDISC, by contrast, does have such metrics: it is a validated discriminant-function procedure with published classification accuracy data for each reference group comparison.

The UK Forensic Science Regulator's guidance on expert witness testimony, and the Crown Court's admissibility framework under the Forensic Science Regulator Act 2021, do not specifically address racial typology in forensic anthropology, but the general Daubert-equivalent reliability requirement (that expert testimony must be based on methods with demonstrated validity and known error rates) applies. A UK expert witness who testifies that "the skull has Caucasoid features" without operationalising what "Caucasoid features" means, how they are measured, and what the false-positive rate is, faces a difficult cross-examination. An expert who testifies that "metric analysis using FORDISC 3.1 against Howells European reference samples returned a posterior probability of 0.77" is on significantly more defensible ground.

In India, the criminal justice system's treatment of forensic anthropology expert testimony is governed by the Bharatiya Sakshya Adhiniyam (BSA) 2023, which replaced the Indian Evidence Act 1872. The BSA Section 39 governs expert opinion evidence and requires the expert to explain the basis of the opinion. An Indian forensic anthropologist testifying in a sessions court or a High Court who uses terms like "Caucasoid" without explaining what the term means, how it was determined, and why it is relevant to Indian casework (where Caucasoid, Mongoloid, and Negroid categories do not map cleanly onto Indian subpopulations) is in an analogous position of methodological vulnerability.

M. Katherine Spradley and Rebecca Stull's 2018 paper in the Journal of Forensic Sciences, "The Population Affinity Reporting Standards in Forensic Anthropology: What Is Currently Being Reported and Suggested Standardization," surveyed forensic anthropology case reports from US medical examiner contexts to assess how ancestry was being reported in practice. The findings were striking: a significant minority of reports still used racial typological terms (Caucasoid, Mongoloid, Negroid), many reports did not specify which FORDISC reference groups were used, and many did not report posterior probabilities or typicality probabilities at all. Reports stated conclusions (the individual is Caucasoid) without stating the evidence basis for the conclusion.

Spradley and Stull's proposed standardisation specified five elements that should appear in every population-affinity statement in a forensic anthropology report: (1) the method used (metric, morphoscopic, or both); (2) the reference groups compared; (3) the statistical output (posterior probability, typicality probability); (4) the direction of the finding (highest affinity to group X); (5) a statement of the limitations and uncertainty range.

This standardisation has been progressively adopted in leading US forensic anthropology laboratories and endorsed in the OSAC Forensic Anthropology subcommittee's standards documents. Adoption in other jurisdictions is uneven. A 2021 survey of Australian forensic anthropology reports (Taylor and colleagues, 2021, Journal of Forensic and Legal Medicine) found that Australian practitioners varied considerably in their reporting language, with some still using typological terms. Indian forensic anthropology report language has not been systematically surveyed in the published literature, but case report citations in Indian medicolegal journals suggest that typological terminology remains common in practitioner-level reports produced for courts.

The practical implications of non-standardised reporting are real. A defence barrister or advocate in a murder trial who retains a forensic anthropology expert reviewer can challenge a case report that states "the skull is Caucasoid" on the grounds that the term has no defined operational basis in Indian casework, that it does not correspond to any Indian population group that could narrow a missing-persons search, and that the FORDISC analysis (if one was performed) has not been reported in a way that allows the court to assess its reliability. The Spradley-Stull standardisation pre-empts exactly this challenge.

The transition from typological to population-affinity language is sometimes misunderstood as a move toward less precise or less useful testimony. In fact, the population-affinity framework demands more specificity: which reference groups were loaded, what the posterior probabilities were, what the typicality probability indicates, and how the morphoscopic assessment corroborates or qualifies the metric finding.

Consider two report paragraphs describing the same FORDISC analysis:

Typological language: "Examination of the cranium reveals features consistent with a Caucasoid individual. The nasal aperture is narrow and the nasal bone shows the characteristic pinched form associated with European racial type. It is this examiner's opinion that the decedent is Caucasoid."

Population-affinity language: "Metric analysis using FORDISC 3.1 was performed on eighteen Howells craniometric variables measured on the case cranium. Reference groups loaded for analysis were American White Male, American Black Male, American Hispanic Male, American Indian Male, Japanese Male, and Chinese Male from the Forensic Data Bank and Howells world sample. The analysis returned a classification of American White Male with a posterior probability of 0.62 and a typicality probability of 0.14, indicating that the cranium is a moderately typical member of the American White Male reference group. The next-best classification was American Hispanic Male with a posterior probability of 0.21. Morphoscopic trait scoring revealed: anterior nasal spine pronounced (score 3), nasal bone shape pinched, inter-orbital breadth narrow, inferior nasal aperture with a sharp sill. These morphoscopic findings are consistent with the metric classification. It is this examiner's opinion that the cranial morphology shows highest affinity to the Howells European reference samples. Race is not reported; population affinity is expressed in terms of metric similarity to documented reference groups."

The second paragraph is longer and substantively more useful. It specifies which reference groups were compared and which were excluded, quantifies uncertainty as a posterior probability, provides independent morphoscopic corroboration, and pre-empts cross-examination on racial terminology by explicitly addressing it in the report.

In the United States, the Daubert standard (Daubert v. Merrell Dow Pharmaceuticals, 1993; further developed in Kumho Tire v. Carmichael, 1999) requires federal courts to evaluate whether expert testimony is based on a scientifically valid, tested method with a known error rate that is generally accepted in the relevant scientific community. A forensic anthropologist using racial typological terminology in federal court who cannot cite a published error rate for the typological classification is vulnerable to exclusion under Daubert. Leading US forensic anthropology training programs, including those at Texas State, the University of Tennessee, the University of New Mexico, and the University of North Texas, have moved away from typological language in both training and casework.

In the United Kingdom, expert testimony admissibility in the Crown Court is governed by the Criminal Procedure Rules, the Law Commission's 2011 review of expert evidence, and the Forensic Science Regulator's Codes of Practice. The 2023 Forensic Science Regulator Act (which gave the Regulator statutory powers) reinforces the requirement that expert evidence be based on validated, reliable methods. The Council for the Registration of Forensic Practitioners (CRFP) competence framework for forensic anthropology expects practitioners to use methods consistent with current scientific standards. The shift to population-affinity language is consistent with these expectations, though no specific UK regulatory guidance on racial typology in forensic anthropology reports existed as of 2024.

In India, the BSA 2023 (Bharatiya Sakshya Adhiniyam, replacing the Indian Evidence Act 1872) governs expert testimony under Section 39. There is no Daubert-equivalent gatekeeping mechanism in most Indian courts: the reliability of expert methodology is assessed post-admission by the trier of fact rather than pre-admission by the judge. However, in high-profile cases (the Aarushi-Hemraj Noida 2008 case, the Sheena Bora case 2015), defence teams have challenged the methodology of forensic experts on cross-examination with increasing sophistication. As Indian forensic anthropology practice professionalises through AIIMS and other institutions, the adoption of population-affinity language and OSAC-aligned reporting standards is an incremental process rather than a wholesale transition.

In the European Union, forensic anthropology practice is coordinated partly through the ENFSI (European Network of Forensic Science Institutes) Forensic Anthropology Working Group. ENFSI does not have a specific position statement on racial typology terminology as of 2024, but its best-practice manuals for forensic anthropology (most recently updated 2018) recommend metric-based reporting with statistical outputs rather than typological classification. German, Dutch, French, and Scandinavian forensic anthropology practices have broadly adopted population-affinity language, partly driven by ENFSI coordination and partly by domestic laboratory accreditation requirements under ISO 17020 / ISO 17025.

The population-affinity framework retains the scientifically valid content of ancestry assessment while discarding the typological scaffold that is not defensible. The useful content is: a quantified statement of which reference populations this skull most resembles metrically, with the degree of resemblance stated as a probability, plus a morphoscopic corroboration or qualification, plus an honest statement of what the reference populations do and do not represent.

The discarded content is: the inference that the skull "is" a member of a natural biological kind, the use of terminology (Caucasoid, Mongoloid, Negroid) that implies discrete biological categories, and the suppression of uncertainty behind a confident typological label.

A complete ancestry statement in a modern forensic anthropology report covers five elements. First, the method: which measurements were taken, which FORDISC version was used, which reference groups were loaded. Second, the statistical output: posterior probability for the classified group, posterior probability for the next-best group, typicality probability. Third, the morphoscopic assessment: which Hefner traits were scorable, what scores were assigned, whether morphoscopic and metric findings concord. Fourth, the affinity statement: the skull shows highest metric and morphoscopic affinity to [reference group], with posterior probability [value] and typicality probability [value] indicating [degree of fit]. Fifth, the limitation statement: the reference sample for this population group is [size]; the analysis was performed on [number] of the 28 available Howells variables; a typicality probability of [value] indicates [adequate/marginal/poor] fit within the reference group.

The 2020 AABA position statement and the Spradley-Stull standardisation address primarily US casework. The US forensic anthropology caseload is dominated by individuals of European, African, Hispanic, and Indigenous American ancestry, and the FDB reference data reflect this. The methodological challenges differ in international casework, particularly in South Asian contexts.

Take a case skull recovered from a rural area in Tamil Nadu, South India. Running it through FORDISC 3.1 against the available reference groups yields a moderate posterior probability for one group (perhaps American White Male at 0.48 or Howells Hindu at 0.44) and low typicality probabilities across all groups. The Howells "Hindu" sample is from North India; the FDB groups are from the US. Neither is an appropriate reference for a South Indian individual. Reporting this as "highest affinity to American White Male" would be technically accurate as a statement of FORDISC output but misleading to an investigator trying to identify a missing person.

The correct approach in this scenario involves four steps. First, report the FORDISC result with full statistical output, including the low typicality probabilities, explicitly noting that the reference groups available do not adequately represent South Indian populations. Second, supplement with available South Indian craniometric data (Pan 1924, Manjunath 2013) to provide a regional calibration check, noting whether the case skull's measurements fall within the published ranges for South Indian skeletal series. Third, score morphoscopic traits (anterior nasal spine, nasal bone shape, malar tubercle, inter-orbital breadth, palate shape) and note whether the pattern is consistent with published descriptions of South Indian cranial morphology. Fourth, state explicitly in the report that a definitive population-affinity classification is not possible with current reference data and that DNA ancestry analysis would provide additional information.

Explicit acknowledgment of reference-database gaps is scientifically defensible. It is more useful to an investigator than a confident but misleading classification, and it is better positioned to survive cross-examination than a classification that turns out to be based on an inappropriate reference sample. Where available, DNA ancestry analysis and personal identification by comparative radiography can provide independent lines of evidence.

The same logic applies to East African casework (Somali, Ethiopian, Eritrean individuals are poorly represented in FORDISC reference groups), to South-East Asian casework outside Vietnam, and to Middle Eastern and Central Asian populations where dedicated reference databases are limited.