Why does the spiciest food come from hot places? In a fascinating series of papers published two decades ago, Paul Sherman and colleagues argued that variations in spice use were a sign of cultural evolution in action. Since many spices have anti-microbial properties, adding them to recipes might reduce the risk of food poisoning, so communities that developed spicy cuisines could have an advantage in reducing infection rates. This intriguing hypothesis was termed “Darwinian gastronomy”. Could selection of spicy foods, deliberately or unconsciously, be an adaptation to help cultures survive in hot climates where risk of infection from food can extract a large toll in terms of health and survival?
To test this hypothesis, Sherman and colleagues collected traditional recipes from around the world, and showed that the average number of spices per recipe showed a significant relationship with average temperature. But there is a special problem with any analysis that looks for correlations between cultural and environmental traits. Nearby and related cultures tend to share many aspects of culture, including cuisine, and also have similar environments, including temperature and parasite load, generating incidental associations between the cultural and environmental variables. In other words, you might get a significant correlation between different traits – like spice and temperature – not because there is any causal link between them, but because related cuisines are found in similar environments. For example, Scandinavian cultures have similarities in their cuisines because they share a common heritage, and they also all co-occur in a cold part of the world. In contrast, cuisines in South East Asia are all more similar to each other than any of them is to Scandinavian food, and they are also all in a much warmer climate. Treating all of those cuisines as separate data points will make it look like there is a link between climate and cooking, even if there isn’t actually any functional connection.
We adapted analytical techniques from evolutionary biology that are designed to overcome these issues of proximity and relatedness, and used them to analyse recipes from all over the world. We gathered data from 33750 recipes from 70 cuisines from around the world, containing over 90 different spices. Average number of spices is related to average temperature – but only as much as you would expect from similarities due to proximity and relatedness. In other words, temperature doesn’t seem to be the driving force behind difference in spiciness.
But temperature was only ever considered a proxy measure for the proposed cause of variation in spice use which is infection risk. Using global data from the World Health Organisation (WHO), we show that spice use is related to risk of foodborne illness. But it’s also related to infections from other causes, such as mosquito-borne, water-borne and contact-based diseases. In fact, the average number of spices per recipe is also related to health outcomes that have nothing to do with infection risk, such as fatal car accidents, so there is a significant relationship between average life expectancy and spicy food. This doesn’t mean that spicy food shortens your life span (or makes you crash your car). Instead, it shows how many cultural variables all scale together, making disentangling causal relationships difficult, if not impossible.
In fact, general socioeconomic indicators (e.g. gross domestic product per capita) and overall measures of health outcomes (e.g. life expectancy) provide better predictors of the spiciness of cuisines than temperature or infection risk. Because so many aspects of culture are bundled together and scale with these general measures, this makes it very hard to identify meaningful causal relationships for differences in cultural traits like cuisine. However, we can rule out some hypotheses as having little support from this global dataset of recipes. Spicier foods are not explained by variation in climate, human population density or cultural diversity. Patterns of spice use don’t seem to be driven by biodiversity, nor by the number of different crops grown, nor even by the number of spices growing naturally in the area.
We can also test other predictions of the Darwinian gastronomy hypothesis. If spice use is driven by the advantages of adding antimicrobial ingredients, then other ingredients with antimicrobial actions should also be favoured. But we found no evidence that using vinegar or alcohol in recipes is related to infection risk. Spices should also be preferentially used in riskier dishes. While dishes with meat or seafood do seem to have more spices, they also have more ingredients overall, and the spiciness of meat dishes doesn’t seem to be influenced by the level of foodborne illness.
Spicier food is found in hotter countries, but our analysis provides no convincing reason to believe that this is primarily a cultural adaptation to reducing infection risk from food. This study illustrates the challenges of testing hypotheses about the drivers of cultural evolution, because it is difficult to disentangle meaningful relationships from incidental associations caused by neighbouring and related cultures sharing many aspects of their culture and environment in common.
While explaining patterns of spicy food may seem interesting but ultimately unimportant, the problem we have illustrated is common to many studies that draw conclusions by comparing different cultures. In particular, infection risk has been used to explain diverse cultural variation, not just cuisine but also social behaviour, traditionalism, religiosity, forms of governance, and even IQ. Yet these correlations are also largely caused by the statistical artefacts generated by similarities between related and neighbouring cultures. Testing evolutionary explanations for cultural phenomena requires a toolkit of analytical techniques that are suited to overcoming these problems.
Further information: Bromham L, Skeels A, Schneemann H, Dinnage R, Hua X (2021) There is little evidence that spicy food in hot countries is an adaptation to reducing infection. Nature Human Behaviour https://www.nature.com/articles/s41562-020-01039-8
Bromham L, Hua X, Cardillo M, Schneemann H & Greenhill SJ (2018) Parasites and politics: why cross-cultural studies must control for relatedness, proximity and covariation. Royal Society Open Science 5, 181100. https://doi.org/10.1098/rsos.181100
Sherman PW & Billing J (1999) Darwinian gastronomy: why we use spices: spices taste good because they are good for us. BioScience 49, 453-463