Forensic facial reconstruction is the process by which a face is reconstructed from a deceased individual’s “skeletal remains through an amalgamation of artistry, forensic science, anthropology, osteology, and anatomy.” This technique is used largely only when skeletal remains are the only evidence at the scene of a crime. Perhaps most famously, it was used in the facial reconstruction of Mitochondrial Eve—but just how accurate is forensic facial reconstruction?
The average person may believe that forensic facial reconstruction is successful and a good proxy for what an individual may have looked like. This is bolstered by the fact that people hear of stories in which facial reconstruction are successful, never hearing of the countless number of cases in which the “reconstruction” fails to identify anyone.
Numerous papers in the literature, though, do show that forensic facial reconstruction does have a high success rate. For instance, Lee et al (2011) and Wilkinson et al (2006) show that this method has a considerable chance of getting facial morphology in the ballpark of how they look. However, this was done on live subjects and is of no use for Mitochondrial Eve/deceased individuals. Facial reconstructions of the deceased are, for the purpose of this article, what we need to look at, not studies looking at live people. (There are also hurdles for facial recognition systems.)
One of the biggest hurdles for the accuracy of forensic facial reconstruction is that average facial tissue thickness cannot be inferred (most importantly, the lips, cartilage, skin and fat). Due to this, with no prior information to look at, the look of the skull will be subjective with the “reconstruction” looking somewhat similar due to chance. Further, the main facial features are largely determined by the shape of the skull.
Now to the fun part: Is this what Mitochondrial Eve really looked like?
There are a number of features that are problematic to infer from these facial reconstructions. Lips, ears, skin, craniofacial muscles—all are extremely hard, or next to impossible, to predict if the only thing we had was a skeleton. Further, since there are few tested relationships between soft and hard tissue for modern humans “it is clear that the use of facial approximation techniques on ancestral skulls of modern Homo are fundamentally flawed, as previously reported by Montagu” (Stephan, 2003). Since soft tissue quickly decays, it is left up to artistic interpretation. Further, attempting to map ape morphology since we diverged a few million years before is misleading, due to the fact that hard and soft tissue relationships are not likely to be the same for apes and our hominin ancestors.
Now that we know the so-called “reconstruction” of Mitochondrial Eve is not what she really looked like, there are a few more problems with this method I’d like to go over.
Forensic facial reconstruction is used when the remains of an unidentified individual are discovered. If the bones of the deceased are all that forensic artists have to go off of, the finished product may be extremely subjective/biased. People who believe that forensic facial reconstruction truly works may say “It works all the time. If it didn’t, how would it be able to solve crimes?” Success rates for the identification of individuals ranges from 50 to 100 percent (Stephan, 2003: 196), and so, the belief that “reconstructions” are largely accurate continue to persist.
However, like with the case of the famous stereotype threat with a modicum of unpublished studies, the “success” of forensic facial reconstruction is also skewed by non-reporting of unsuccessful cases
It is also rare for forensic facial approximation to be to be better than chance, with 403 incorrect facial identifications out of 592 identification scenarios in one study (Stephan and Henneberg, 2001). Facial reconstruction has the greatest accuracy if there is any knowledge of past injuries for the individual, a photo, and soft tissue. Obviously, in the case of Mitochondrial Eve, we don’t have a photo nor do we have soft tissue and knowledge of past injuries are a non-factor. So it seems that if some people claim to know what the first AMH looked like, it’s probably “just a guess”, and a pretty bad one at that due to the no knowledge of hard and soft facial tissue in the hominin lineage.
The hardest part about facial reconstruction is reconstructing soft tissues accurately since they quickly decompose. This is even more of a problem for people who lived hundreds of thousands of years ago. We’ve never seen any alive so we don’t know what they may have looked like to be able to infer what Mitochondrial Eve would have looked like. Things like lips/mouth, skin, hair, and ears are largely up to artistic interpretation, which are subjective in nature. Craniofacial morphology has also changed in the past 200 thousand years, which may be due to a decrease in testosterone/androgen receptors.
If we can’t identify humans with facial recognition better than chance, what makes anyone think that we can even have the slightest idea of how Mitochondrial Eve looked—when some of the most important parts of the phenotype aren’t around to observe and thus subjectivity then comes into play. Any “reconstructions” you come across, you should take with a grain of salt. It’s next to impossible to know what ancient hominins may have looked like due to the absence of soft tissue, and so any phenotype that a so-called “reconstruction” may give is, largely, up to the interpretation of the individual artist.
With our current technology, it’s next to impossible to ascertain what Mitochondrial Eve—or any other ancient hominin for that matter—may have looked like.
Lee, W., Wilkinson, C. M., & Hwang, H. (2011). An Accuracy Assessment of Forensic Computerized Facial Reconstruction Employing Cone-Beam Computed Tomography from Live Subjects. Journal of Forensic Sciences,57(2), 318-327. doi:10.1111/j.1556-4029.2011.01971.x
Stephan, C. (2003). Anthropological facial ‘reconstruction’ – recognizing the fallacies, ‘unembracing’ the errors, and realizing method limits. Science & Justice,43(4), 193-200. doi:10.1016/s1355-0306(03)71776-6
Stephan, C. N., & Henneberg, M. (2001). Building Faces from Dry Skulls: Are They Recognized Above Chance Rates? Journal of Forensic Sciences,46(3). doi:10.1520/jfs14993j
Wilkinson, C., Rynn, C., Peters, H., Taister, M., Kau, C. H., & Richmond, S. (2006). A Blind Accuracy Assessment of Computer-Modeled Forensic Facial Reconstruction Using Computed Tomography Data From Live Subjects. Forensic Science, Medicine and Pathology,2(3), 179-188. doi:10.1385/fsmp:2:3:179
Wilkinson, C. (2010). Facial reconstruction—anatomical art or artistic anatomy? Journal of Anatomy,216(2), 235-250. doi:10.1111/j.1469-7580.2009.01182.x