Earlier this month, Discovery News published an article about infanticide in Ancient Rome. Viegas (2011) summarized the argument: “Infanticide, the killing of unwanted babies, was common throughout the Roman Empire and other parts of the ancient world”. This article was based on a publication in the Journal of Archaeological Science, which argues that new evidence shows that infanticide was common throughout the world prior to effective contraception, and that it was widely practiced throughout the Roman empire.
Mays and Eyers (2011) publication on infanticide uses perinatal skeletal remains from the Roman British site of Hambleden. Perinatal means that the infants are aged around the time of birth. The remains of the infants were found buried near a larger cemetery, although they were not part of it. The original excavation in 1927 found 97 skeletons, but only a third were studied. Each individual was measured to determine age. Mays and Eyers (2011) argued that if the remains belonged to infants who died in childbirth or from complications there would be a wider age range than those who were killed. In order to determine this, they compared the age distributions of the Hambleden remains against a natural infant death distribution from the medieval site of Wharram Percy in Britain, and an infanticide distribution from Roman period Ashkelon site in Israel.
The Hambleden distribution is more reflective of infanticide than natural death. However, Mays and Eyers do not argue that this is reflective of widespread infanticide, rather that it adds to the evidence that it was a practice in ancient societies. Viegas (2011) does connect infanticide with low income and high female disparity, with it serving as a method of conserving resources. This is given as a potential reason for the infanticide. This argument is also supported by an earlier article in which Mays (1993) argued that female infanticide must have been common due to the high number of adult males found in cemeteries during the Roman period.
However, as noted by a response to this article written by Harrsch (2011), further exploration of the context of Hambleden is needed before we can fully determine what was practiced. Harrsch (2011) attempts to reconstruct a context using knowledge of Roman events in Britain in the period contemporaneous with Hambleden, as well as looking at Roman laws during this period regarding infanticide. She found that there were numerous laws and scholars discussing the problem of infanticide, indicating that the practice did exist although it was outlawed and looked down upon. While Harrsch (2011) does support the conclusion that there was a high amount of stress and strife in this period, an argument for infanticide due to conserving resources. She also notes that contraception and abortive practices were also known in Roman times, eliminating the need for infanticide. In response to Mays (1993) argument that infanticide of females is evidence by a high number of males, Lewis (2007) notes that this doesn’t take migration, warfare and differential burial into consideration.
Gowland and Chamberlain’s (2002) article investigating infanticide in Roman Britain is particularly revealing. They looked at the age distributions of a variety of skeletal collections of perinatal remains that were reported to be due to infanticide given the distribution of ages close to birth. They found a number of biases in the investigation of the collections including the techniques for estimating the age of perinatal skeletons and the statistic techniques used to understand the age distribution. Instead of traditional regression statistics which often mimic the reference sample, they used a Bayesian regression. They concluded that the peak in mortality shown by earlier investigations was due to errors in technique and statistics. Using the Bayesian method they calculated that most of the reported infanticide collections were reflective of natural birth curves.
So when assessing infanticide it is important to look at the entire context of the site, glean as much evidence as possible about the community, assess the age of the individuals using statistics that won’t bias the outcome, and be careful in making widespread generalizations. Also, read the original paper. Mays and Eyers (2011) analysis does not draw the types of conclusions that Discovery News summarizes, and it gives a more balanced analysis of the material. Since the remains were buried, it may be safer to conclude that they were given a separate burial due to their age and not separate due to being killed at birth. Infanticide is a loaded label, and before we identify an massive empire with having this trait we need more evidence. The Roman empire was vast, and had a high level of diversity. Before conclusions are drawn, evidence needs to back them up. Infanticide may have been practiced at Hambleden, but this does not mean that the entire Roman empire condoned the practice.
Viegas. 2011. Infanticide common in Roman Empire. Discovery News. http://news.discovery.com/archaeology/infanticide-roman-empire-110505.html
Harrsch. 2011. Widespread Roman Infanticide not Substantiated by Hembleden Studies. Roman Times. http://ancientimes.blogspot.com/2011/05/widespread-roman-infanticide-not.html
Gowland and Chamberlain. 2002. A Bayesian Approach to Ageing Perinatal Skeletal Material from Archaeological Sites: Implications for the Evidence for Infanticide in Roman-Britain. In Journal of Archaeological Science 29(6)
Lewis. 2007. The Bioarchaeology of Children. Cambridge University Press: UK