The Digging into JRA blog authors’ opinions are not necessarily identical to those of the featured JRA volume contributors.
What would you assume the Romans ate, pork? Not quite. When it comes to textual accounts about the diet in Roman Italy, pork is terribly overrated, claims Mamoru Ikeguchi. In his study “Beef in Roman Italy”, the scholar from Kurume, Japan, grapples with the best ways of measuring meat consumption, arriving at data that positions beef, rather than pork, as the meat most widely consumed by the Romans.
When making assumptions about the Roman meat diet, historians have mostly trusted available literary evidence, such as the famous Roman cookbook by Apicius De re coquinaria, which clearly presents pork as preferred over mutton, goat’s meat, and beef. Such reasoning is not scientific enough, Ikeguchi argues, as it limits the evidence to rather elitist sources, shedding little to no light on the diets of the middle and lower classes. The way to go, according to the scholar, is by turning to zooarchaeology, the study of dead animal remains:
“Zooarchaeological evidence, which has not been given enough attention by historians, helps fill the gap between texts and the reality” (JRA 30 , p. 7).
But how do you quantify meat? More importantly, how do you measure all the meat consumed as far back as the 5th century B.C. in all of ancient Italy? Clearly, there is not only one answer. The two dominant methods for assessing the meat consumed so far have been through collecting data of NISP (Number of Identified Specimens, practiced for example by Tony King of the University of Winchester, UK) and MNI (Minimum Number of Individuals, practiced for example by Michael MacKinnon of the University of Winnipeg). Translated into plain English, what you do with NISP is count dead animal bones and then multiply them by the accepted ratios of meat weight between domesticates. This has its virtue in objectivity: there is obviously no human factor to counting bones. However, as the saying goes, sticks and stones can break anybody’s bones, especially those of dead animals, and bone fragmentation can significantly mess up the data. Therefore, counting NISP alone is a no go. With MNI, the procedure is to attribute the bones that have been dug up to particular species and then count up the species. But not only is this too complicated to do, the method is also rather subjective, as data will depend heavily on the human factor, meaning the methods employed in excavation on each site by each different faunal specialist studying the bones.
Since none of the existing methods of quantifying meat consumption is sensitive enough on its own, Ikeguchi’s solution is to combine both NISP and MNI counts while devising a normalizing unit between them: that of quasi weights (kg). Quasi weights are based on NISP by assuming the same meat weights as in the MNI-based weights. This way, the sample size of each data set, such as how many dead animal remains were collected in total at a particular site, is neither downplayed nor given excessive significance. The scholar further contextualized the meat data, dividing the whole meat weight of each species by the number of centuries to which they belong and then adding up all sites into a chronological progression, as well as mapping out the trends in meat consumption on a geographical basis. As you might well expect, that makes for a bunch of data: 6 sets of tables, 1 map, and 10 sets of graphs, as featured in the JRA article. And guess what all that data is saying? Pork has been unjustifiably overrated as a staple of the Roman meat diet, with beef displaying a far wider record of consumption, especially in the rural and suburban areas of Northern and Southern Italy.
Now what? Pork is overrated and beef is underrated–who cares? Well, raw data makes far more sense when inserted into its historical background. First and foremost, the prevalence of beef as a staple of the diet in the rural and suburban regions of the North and South accounts for its lesser presence in literary sources such as Apicius’ cookbook. While pig was raised for butchering alone and produced tender meat of high quality, cattle such as cows, as well as goats and sheep, were first kept around for their labor, for dairying, and for their wool, until they were so old and useless that they would be butchered. Needless to say, their meat would have become subpar by that time, and was thus cheaper than pork (12 denarii vs. 8 per Italian pound). It is no great surprise, then, that pork was preferred by the elites and was less accessible to the lower classes. Furthermore, we learn roughly how much of the population belonged to the beef-preferring class, which is more than 60% of all the studied sites in Northern and Southern Italy. Even in central Italy, 36% of the sites studied had a beef-eating pattern, in strong competition with the pork-oriented ones at 34%. Finally, the keen competition between pork and beef in the central regions of Roman Italy indicates that those were the better-off regions at the time. An interesting detail highlights a steep increase in rural beef consumption near the turn of the era, towards the end of the Republic and beginning of the Empire, when rural settlements are recorded to have shrunk, apparently leading to migration and the need to get rid of many cattle.
In short, Ikeguchi has a beef with beef consumption in ancient Italy. But he goes further, showing that by relying on the textual and visual sources alone historians risk significant bias in their results. After all, literacy in the past has a notorious correlation with socio-economic status. It is for this and other reasons that considering less conventional historiographies, such as that of dead animal remains, is of crucial importance for re-creating narratives of the past.
For access to the original charts, accounts of Roman herding systems along with contemporary attitudes toward the value of various domesticates, and even literary evidence (surprise!) supporting Ikeguchi’s beefy argument, follow the link to order M. Ikeguchi, “Beef in Roman Italy,” JRA 30 (2017) pp. 7-37, or even the whole volume 30 of JRA: http://journalofromanarch.com/order.html