Nutrition science has traditionally relied on population-level evidence, especially evidence from observational studies. However, it is facing what could be called a ‘credibility crisis’ (Penders et al. 2017; Jukola 2021). Critics have questioned the reliability of the evidence originating from observational studies and demanded randomized controlled trials to back up claims about the link between food and health. Further, the aim of implementing effective public health interventions and providing individual guidance based on population-level evidence has been called into question (e.g., Ordovas et al. 2018). Recently, Personalized Nutrition (PN) has arisen as a challenger to the traditional population-based approach to nutritional evidence and advice. It aims at providing more effective interventions by utilizing genetic, nutritional, medical, etc. information. This talk addresses the question: What are the epistemic and sociopolitical implications of the trend towards PN? In order to provide answers, I start by drawing on Longino’s (2013) account of local epistemologies to outline the epistemic landscape of PN and to show how it differs from that of the so-called traditional nutrition science. Despite its recent proliferation, PN lacks a commonly agreed-upon definition (e.g., Bush et al. 2020). I suggest that there are multiple ways of conceptualizing PN and, consequently, of delineating what its central research questions and methods are. For example, there are differences in which physiological, genetic, or clinical parameters researchers focus on (e.g., Drabsch & Holzapfel 2019). Second, I assess the potential ethical and political implications of PN. By applying the so-called Coupled Ethical-Epistemic Analysis (Katikireddi & Valles 2015) as a tool for analysing the entanglement of epistemic and non-epistemic aspects of research, I hypothesize that at least some dominant conceptualizations of PN lead to effective public health interventions being undermined with undesirable consequences. This concerns views of PN that focus on genetic variation or epigenetic marks, while overlooking the effects of social and environmental factors, when differences between health outcomes are explained. PN may overemphasize the responsibility of at-risk individuals for their own health to the detriment of interventions targeting social determinants of health.

Philosophers have the impression that evolutionary medicine is plagued by naive adaptationism (e.g., Murphy 2005, Valles 2011, Méthot 2012), leading to poor science through the proliferation of untestable ‘just-so stories’ and to poor medicine through not considering alternative explanations with different medical implications. It is therefore predictable that, given the centrality of optimality modelling to contemporary nutritional ecology, these criticisms will be applied, mutatis mutandis. It is also predictable that because it utilises the concept of ‘mismatch’ between mechanisms governing dietary choice and nutritional environments, nutritional ecology will be criticised for invoking an ‘environment of evolutionary adaptedness’ (Buller 2005). Such criticisms are understandable, because naive adaptationist reasoning occurs in both ‘popular nutrition’ – self-help books and lifestyle gurus – and in traditional nutrition science. However, the solution for this is not less evolutionary thinking, but more, as nutritional ecology demonstrates. Valles (2011) argues that evolutionary medicine is committed to ‘empirical adaptationism’ (Godfrey Smith 2001), the view that forces other than natural selection can be neglected in the explanation of organismic form. One of Valles’ examples is the now exploded ‘promiscuous primate’ hypothesis about menstruation (Profet 1993). But the work which refuted that hypothesis, and which used adaptationist reasoning to establish constraints on the evolution of endometrial reabsorption, is equally part of evolutionary medicine (Strassman 1996). In a similar vein we show that the prominence of optimality analysis and related methods in nutritional ecology does not reflect on a commitment to strong empirical adaptationism. It reflects both ‘methodological adaptationism’, a powerful tool for revealing constraints on natural selection, and ‘explanatory adaptationism’ – an explanatory focus on the observed degree of adaptation. Nutritional ecology examines the interactions of animals with nutritional environments (Raubenheimer and Simpson 2012). It makes extensive use of the idea that nutritional regulatory phenotypes and their teleonomic goals (‘intake targets’) are 1) an important determinant of evolutionary fitness, 2) therefore reflect a history of natural selection and 3) thus contain information about the interplay of optimization and constraint in evolution. It is explicitly and importantly a multi-scale theory, which studies the adjustment of organisms to their nutritional environments on scales ranging from minutes (e.g., homeostasis) to lifetimes (e.g., phenotypic plasticity), across generations (epigenetic inheritance) and in evolutionary time (gene selection). Since the causal interaction of species with their environments is bidirectional, nutritional ecology also makes use of gene-culture coevolution and niche construction when explaining nutritional phenotypes. Nutritional ecologists have examined these issues extensively in laboratory and field studies. Understanding the actual ways in which nutritional ecology is ‘adaptationist’ will provide a safeguard against what we might call ‘naive anti-adaptationism’, the failure to appreciate the methodological sophistication with which researchers use optimality analysis and related methods. Our discussion is therefore an example of the idea at the heart of this symposium – that nutrition science is a rich and productive field for interaction between philosophy and science.

The central aim of the nutrition sciences is to understand how nutrition impacts health. One problem supposedly plaguing this endeavor is nutritionism—a ‘reductive’ focus on the role of nutrient composition or isolated nutrients (e.g., macronutrients or vitamins) for explaining a food’s effects on health (Scrinis 2008; 2013). Methodologically ‘reducing’ foods to nutrients can foster adversarial debates about the purported health effects of isolated nutrients, obscure the complexity of food-organism interactions, and distort how nutrients produce different outcomes in the context of foods, processing techniques, or dietary patterns. Anti-reductionist critiques, most of which claim that foods or dietary patterns should be the fundamental explanatory levels, permeate nutrition research (Messina et al. 2001; Zeisel et al. 2001; Hoffmann 2003; Jacobs and Tapsell 2008; Fardet and Rock 2014; 2018; 2020; Mayne, Playdon, and Rock 2016; Mozaffarian, Rosenberg, and Uauy 2018; Rees 2019; Moughan 2020; Campbell 2021). Moreover, the claim that nutritionism is the “dominant ideology” is entering philosophy (Siipi 2013; Borghini, Piras, and Serini 2021). Amidst calls to reform nutrition research (Ioannidis 2013; 2018; Mozaffarian and Forouhi 2018; Hall 2020), this presentation contributes by clarifying whether the problem raised by nutritionism is less about which level is fundamental (nutrients vs. foods), and more about which level(s) provide adequate explanations of what in nutrition impacts health. 1) For instance, are explanations that aim to link the nutrients in foods or dietary patterns to specific health outcomes inherently flawed? 2) Relatedly, can nutrient-level research elucidate causal mechanisms that are often obscured in highly complex food- or diet-based investigations? 3) Moreover, can nutrients explain variations in organismal feeding behaviors, explaining why organisms select their foods? While the complexity of food-health interactions requires more than ‘one nutrient at a time’ approaches (Simpson, Le Couteur, and Raubenheimer 2015), answering the above questions entails evaluating whether/how nutrient-level research can generate integrative explanatory frameworks (Brigandt 2010). First, I analyze claims that nutrient reductionism could be useful if it offers mechanistic explanations that combine explanatory levels (Machamer, Darden, and Craver 2000; Ströhle and Döring 2010). For instance, nutritional ecologists propose that nutrient ratio variations, e.g., protein to carbohydrates, are the common threads among foods, meals, and diets that provide robust explanations of distinct outcomes—from metabolic regulation and biological fitness to obesity (Raubenheimer and Simpson 2016; 2019; 2020)—and of organismal feeding behaviors: organisms select foods largely based on nutrient content (Simpson and Raubenheimer 2012). Can this form of nutrient-level research offer ‘appropriate’ levels of complexity to clarify how dietary constituents interact, which properties of foods or diets reliably affect health, and the mechanisms involved (Solon-Biet et al. 2019)? I further analyze this proposal by looking at how nutrients modulate cancer risk and development (Theodoratou et al. 2017; Zitvogel, Pietrocola, and Kroemer 2017; Altea‐Manzano et al. 2020; Kanarek, Petrova, and Sabatini 2020; Salvadori and Longo 2021). Overall, I offer an epistemological evaluation of nutrient-level research and its potentially integrative explanations of what in foods and dietary patterns affects health.