My wife discovered a remarkable poster when she was visiting Myanmar. Hanging in most domestic kitchens, it advises against certain food combinations that, it is claimed, could make you ill or even kill you outright.
But where do such remarkable beliefs come from? And how can they remain embedded, sometimes for hundreds of years, in the face of all the evidence? With my Oxford colleague, the philosopher, neuroscientist and general know-all Anders Sandberg, I set out to find out. The result, presented as a paper at the annual Oxford Symposium on Food and Cookery, took us on a fascinating journey that encompassed the development of beliefs of all kinds, from the political and religious to the social and practical. Here are a couple of extracts:
“A cuisine is a way of structuring what is food or not, what is safe to eat or not, and what is appropriate to use or not. Within these bounds, no culture uses all of the edible food in its environment to make up its cuisine. Sometimes this is for good reason – some edible foods may be likely to be contaminated by pathogens, leading to food poisoning, while others may need treatment to remove toxins before becoming edible. Deep down this is linked to ‘the omnivore’s paradox’ – we need to subsist on many kinds of food, and hence need diversity, yet each new ingredient is a potential risk.
These practical reasons, however, do not entirely account for the prevalence of food taboos. As the British biochemist and virologist N.W. Pirie noted: “An interesting feature of such taboos is their persistence for centuries in communities which know that the tabooed food is eaten without harm by others.”
There are many possible reasons for the existence and persistence of food taboos:iv • Food taboos for certain members of the society and to highlight special events
• Food taboos during pregnancy and food changes over the course of the menstrual cycle
• Food taboos as an ecological necessity to protect the resource
• Food taboos in order to monopolize a resource
• Food taboos as an expression of empathy
• Food taboos as a factor in group-cohesion and group-identity
We can divide this list into objective reasons (health, ecology), subjective reasons (religious, empathy), and social reasons (group cohesion and identity). Importantly, such rationalizations do not need to be correct in order to be effective. An ingredient that is generally believed to be dangerous may in fact be quite safe. A food taboo designed to increase group cohesion may become a source of conflict within and between groups. Preservation of a resource may have unforeseen knock-on effects on other resources. More generally, the overall world-view of a culture may lead to mistaken conclusions about any or all of the factors listed above.”
” may refer to such mistaken conclusions as ‘superstitions.’ A superstition in this context is not necessarily about supernatural things, but can be defined as an erroneous belief (i.e. doesn’t conform with the actual facts) that causes some behaviour, often ritualistically. Often this is defined as not conforming with the official or best-known model of the world. Here we will use superstition to refer to warnings about food that empirically does not have a commensurate actual risk.
One way in which such superstitions may become embedded is through emphasis on the role of danger. As noted by Mary Douglas:
“It is obvious that a person when he finds his own convictions at variance with those of friends, either wavers or tries to convince the friends of their error. Attributing danger is one way of putting a subject above dispute. It also helps enforce conformity.”
We will return to this in building an opinion model of food superstitions.
Food superstitions may also originate as form of costly and hence hard-to-fake signalling of group loyalty or other desirable behaviour unrelated to the food. An interesting deliberate example was the rock band Van Halen’s contract with concert venues that contained a ban on brown M&M candy in the back stage area: while inconsequential, it was a reliable test that the other complex and safety-relevant instructions were likely to be followed. It is not hard to imagine that the rule could be misinterpreted as a taboo or a superstition by others (indeed, this is required for the M&M trick to work), and then propagated to fans wanting to imitate. The actual origin of a superstition may be very different from how it functions later.
This is not to say that superstitions are always irrational. If it is hard to tell safe from unsafe mushrooms and the cost of eating the wrong one is high, claiming all mushrooms are dangerous is a practical way of avoiding risk. Ascribing the ill effect of eating certain mushrooms with alcohol to an erroneous theory of digestion still produces the correct behaviour. It is common to hear claims that food superstitions have some factual basis [Douglas, 2003], but many clearly do not, but nevertheless do cause irrational behaviour, sometimes with detrimental nutrition and health effects.vii Avoiding boiled water due to it being a ‘hot’ food when treating people with fever causes avoidable harm.
Another cause may be cultural copying. Human survival is dependent on many fairly opaque practices, and copying behaviours in detail rather than experimenting is in many cases the smart choice.viii Hence some random superstitions may persist because it would be counterproductive to investigate them.”
“When it comes to a search for potential explanations, we are the beginning of a journey. What we offer here is very preliminary, and open for discussion. We have opted for two paths. The first is to examine the statistics of the food pairings in terms of categories into which the individual foods may most readily be placed (meat and eggs, vegetables, seafood, fruits, spices, milk and milky foods, sweeteners, contributors of sourness, and medicines). Preliminary results are encouraging, especially in revealing differences between believed ‘killer’ food pairings and those believed only to cause some form of illness.
Our second path has been to develop a simple model of how beliefs may originate, spread and then become fixated in a culture even when there is little or no evidence for their validity. This path is wider and more important than may appear at first sight, since similar processes may be involved in the development of all sorts of belief – political, social and religious beliefs being among them.
There are many ways in which relatively unsupported beliefs may become embedded. One strong candidate is the illusion of control effect: people often think they have more control over a situation than they actually have. Accidental correlations are taken as signs of causal impact, and learned as ways of control a situation. In this case the actual risk of food is largely random, and the decision to eat it has little effect on personal risk. However, imagined patterns are easily found and can then be transmitted and reinforced socially.
We will start with a sub-model of how observing food poisoning may originate a belief among people that a certain food pairing is risky. Then we will present a sub- model of how beliefs spread in a social network. After that we will combine them into a joint model of how food taboos emerge, that will later be extended by introducing other factors such as cultural broadcasts.
Observational evidence
Actual observations of people falling ill after eating an ingredient offer evidence that the ingredient is unsafe. If we assume that there is a case of food poisoning, the people hearing about it will update towards the ingredient being unsafe; if they think it is unsafe and they hear somebody survived eating it, they may update in the opposite direction.
The ideally rational person updates beliefs according to Bayes’ rule in statistics. This is a vast oversimplification, but can show how even such persons may become concerned about a food combination … …”
For the mathematics, I am afraid you will have to refer to the original article. If you can’t get a copy, let me know and I will send you one.