Food taboos during pregnancy: meta-analysis on cross cultural differences suggests specific, diet-related pressures on childbirth among agriculturalists

Abstract

Pregnancy is the most delicate stage of human life history as well as a common target of food taboos across cultures. Despite puzzling evidence that many pregnant women across the world reduce their intake of nutritious foods to accomplish cultural norms, no study has provided statistical analysis of cross-cultural variation in food taboos during pregnancy. Moreover, antenatal practices among forager and agriculturalists have never been compared, despite subsistence mode being known to affect staple foods and lifestyle directly. This gap hinders to us from understanding the overall threats attributed to pregnancy, and their perceived nutritional causes around the world. The present study constitutes the first cross-cultural meta-analysis on food taboos during pregnancy. We examined thirty-two articles on dietary antenatal restrictions among agricultural and non-agricultural societies, in order to: (i) identify cross-culturally targeted animal, plant and miscellaneous foods; (ii) define major clusters of taboo focus; (iii) test the hypothesis that food types and clusters of focus distribute differently between agricultural and non-agricultural taboos; and (iv) test the hypothesis that food types distribute differently across the clusters of taboo focus. All data were analysed in SPSS and RStudio using chi-squared tests and Fisher's exact tests. We detected a gradient in taboo focus that ranged from no direct physiological interest to the fear of varied physiological complications to a very specific concern over increased birth weight and difficult delivery. Non-agricultural taboos were more likely to target non-domesticated animal foods and to be justified by concerns not directly linked to the physiological sphere, whereas agricultural taboos tended to targed more cultivated and processed products and showed a stronger association with concerns over increased birth weight. Despite some methodological discrepancies in the existing literature on food taboos during pregnancy, our results illustrate that such cultural traits are useful for detecting perception of biological pressures on reproduction across cultures. Indeed, the widespread concern over birth weight and carbohydrate rich foods overlaps with clinical evidence that obstructed labor is a major threat to maternal life in Africa, Asia and Eurasia. Furthermore, asymmetry in the frequency of such concern across subsistence modes aligns with the evolutionary perspective that agriculture may have exacerbated delivery complications. This study highlights the need for the improved understanding of dietary behaviors during pregnancy across the world, addressing the role of obstructed labor as a key point of convergence between clinical, evolutionary and cultural issues in human behavior.

Keywords: Agriculture; Anthropology; Evolution of human diet; Evolution of human subsistence patterns; Food taboos; Hunter-gatherers; Obstetric dilemma; Obstructed labor; Phenotypic plasticity; Pregnancy.

Figure 1. PRISMA flow diagram.

Figure 2. Percentage distribution of food types avoided during pregnancy.

The first bar in the graph refers to the Food type distribution among total taboos. The remaining two bars show the Food type distribution within agricultural and non-agricultural taboos. The percentages of Animal products (black), Plant products (grey) and Processed and miscellaneous products (white) differed significantly between agricultural and non-agricultural taboos, (X² = 58.433, df = 2, p < 0.001, Cramer’s V = 0.471; n = 263). Plant and processed products are more present among the taboos of agriculturalists, and Animal products are more present among those of non-agriculturalists.

Figure 3. Percentage distribution of taboos focus.

The first bar in the graph refers to the distribution of the focus on the taboo, as a percentage of the total number of taboos. The remaining two bars show the distribution of the focus within agricultural and non-agricultural taboos. No or unspecified physiological complications (white) were significantly higher among non-agricultural taboo focus, while fear of Big baby and/or difficult delivery (black) was significantly higher among agricultural taboo focus, X² = 40.682, df = 2, p < 0.001, Cramer’s V = 0.393 (N 263). No significant difference was found for the distribution of Varied physiological complications (grey) between subsistence patterns.

Figure 4. Percentage distribution of food types avoided during pregnancy between categories of Taboo focus.

Animal products (black) were significantly more present in the No or unspecified physiological complications category than in Big baby and/or difficult delivery. Percentages of Plant products and Processed and miscellaneous products were significantly higher in the Big baby and/or difficult delivery category than in No or unspecified physiological complications, X² = 35.002, df = 4, p < 0.001, Cramer’s V = 0.258(N = 263).

Figure 5. Distribution of food types between subsistence patterns within geographical regions.

Mosaic plot that shows which cells contribute most to the significance of the test of independence (Fisher’s Exact test) between subsistence patterns and taboo food types in (A) Africa, (B) Asia and (C) Oceania. The width of squares represents the numerosity of each subsistence pattern category, while the height the numerosity of each category of food type within each subsistence pattern. Pattern of blue (positive values of standardized Pearson residuals) show cells whose observed frequency is greater than would be found under independence. Pattern of red (negative values) show cells whose observed frequency is less than would be found under independence. The frequency of animal products is greater than would be found under independence among non-agriculturalists from (A) Africa and (B) Asia. White squares indicate positive (solid line) and negative (dotted line) values of standardized Pearson residuals of cells whose observed frequency does not significantly differ from the distribution of data under independence. Plant and processed products are thus more frequently mentioned than animal products among agriculturalists in (A) Africa, (B) Asia and (C) Oceania, even if the distribution does not diverge sharply from the indipendent distribution. Similarly, non-agriculturalists in (C) Oceania mention animal products more frequently than plant and processed products but with lower standardized Pearson residuals than in (A) Africa and (B) Asia.

Figure 6. Distribution of focus of taboos between subsistence patterns within geographical regions.

Mosaic plot that shows which cells contribute most to the significance of the test of independence (Fisher’s Exact test) between subsistence patterns and focus of taboos in (A) Africa, (B) Asia and (C) Oceania. The width of squares represents the numerosity of each subsistence pattern category, while the height the numerosity of each category of taboo focus within each subsistence pattern. Pattern of blue (positive values of standardized Pearson residuals) show cells whose observed frequency is greater than would be found under independence. Pattern of red (negative values) show cells whose observed frequency is less than would be found under independence. (A) Africa: the frequency of varied physiological complications is greater than would be found under independence among non-agriculturalists. White solid line squares indicate that fear of big babies, but also the less frequently mentioned no or unspecified physiological complications, show positive though not extreme values of standardized Pearson residuals among agriculturalists. (B) Asia: No or unspecified physiological complications are mentioned more frequently than under independence among non-agriculturalists. Conversely, this taboo focus is mentioned less frequently than under independence among agriculturalists (red square), where frequency of fear of big babies and varied physiological complications show positive though not extreme values of standardized Pearson residuals (white and continuous line squares). (C) Oceania: agricultural taboos show high frequency of fear of big babies, while non-agricultural taboos show high frequency of varied and no or unspecified physiological complications, even if this distribution does not divergence significantly from that expected under independence between variables.