1. Introduction
The autonomy of biology from the physical sciences has been a major theme in philosophy of biology. It was important in the vitalism versus mechanism debate and in the organicist movement, and it continues to feature in contemporary debates. Recent works on biological explanation defend the distinctness of biological explanations (Braillard and Malaterre Reference Braillard, Malaterre, Braillard and Malaterre2015; Bock Reference Bock2017; Fang Reference Fang2022), and there is widespread criticism of reductionist approaches in biology, with some authors advocating a return to organicism, according to which biological explanations must appeal to the organism as a whole (Nicholson and Dupré Reference Nicholson and Dupré2018; Dupré Reference Dupré2021; Reiss and Ruse Reference Reiss and Ruse2023; Mossio Reference Mossio2024, and references therein).
This contrasts with the widespread use of reductionist approaches in contemporary biology, especially in cellular and molecular biology, where “the dominant role played by physical and chemical principles” (Weber Reference Weber2005, 25) is not seen as problematic and is arguably part of their success. Despite widespread consensus that biological systems are, or are realized by, physical systems, the question of how biology relates to physics is not settled and continues to motivate discussions about the autonomy of biology, reductionism, and the role of physical explanations in biology.
Vitalism took living systems to be fundamentally different from nonliving matter, either because they contain something primitively lifelike or because they are not governed by physical laws. Following the demise of vitalism, some philosophers viewed biology as a “temporary science” that would in time be reduced to physics (Nagel Reference Nagel1951; Smart Reference Smart1963). Notoriously, Oppenheim and Putnam (Reference Oppenheim, Putnam, Feigl, Maxwell and Scriven1958) proposed a hierarchical model of the sciences that was supposed to correspond to reductive relations between them that were ultimately grounded in microphysics. The idea of the autonomy of biology was developed by Mayr and others in response to this and other strong forms of reductionism. Today, almost everyone agrees that biology is a legitimate science, with its own concepts, methods, problems, and theoretical approaches.
Yet, given that vitalism is wrong, it is also the case that the entities and processes of actual biological systems are just as subject to physical laws as all other systems in the universe. Organisms and other biological entities have physical components organized in certain ways, and biological processes always involve, depend on, supervene on, or are realized by a physical basis. Biological function is both constrained by the laws of physics and enabled by physical processes. Hence this article argues that, contrary to some strong autonomy theses defended in the recent literature, some biological explanations are physical explanations and that the scales and laws that are explanatorily relevant for a particular phenomenon must be decided by scientific discovery, not a priori.Footnote 1
There is not, in fact, one thesis of the autonomy of biology but several. It is important to distinguish the different senses of autonomy, but unfortunately, they are not made explicit and so are often conflated. This article precisely formulates five distinct “autonomy of biology” theses and evaluates each one individually.
The next section distinguishes three moderate autonomy theses and summarizes the arguments for them. Section 3 presents a much stronger autonomy thesis according to which physical laws are mostly irrelevant to biology and argues against it. Section 4 presents the strong autonomy thesis that biological phenomena must have biological explanations and argues against it on the basis of two case studies of physical explanations of biological phenomena. Section 5 concludes the article.
2. The autonomy of biology: Moderate theses
Historically, an important motivation for arguments in favor of the autonomy of biology was to dispel the worry that biology might not be a legitimate science and to make the case that biology is “as much a science as the physical sciences” (Mayr Reference Mayr, Depew and Weber1985, 43). This is unlikely to be opposed by anyone in the twenty-first century.
§1. Scientific status : Biology is a genuine science like physics and chemistry.
There were extreme reductionists in the 1950s and 1960s who thought that biology was not a genuine science. For example, J. J. C. Smart (Reference Smart1963, 57) argued that biology was “not a theory of the same logical sort as physics though with a different subject matter,” being mostly descriptive, and equivalent to just “physics and chemistry plus natural history.” Such views were plausible to others, such as Ernst Nagel (Reference Nagel1951), but the extensive predictive and explanatory success of biology in the decades since renders them quaint. Biology has also been deemed not a proper science for failing to be sufficiently universal, by allegedly dealing with particular, spatiotemporally restricted entities and events taking place on Earth. There is no reason to think that this is true. Despite the current lack of examples of extraterrestrial life, astrobiology is a thriving discipline, and at least some biological theories, notably evolution by natural selection, are thought to be universal (Dawkins Reference Dawkins and Bendall1983; Sterelny and Griffiths Reference Sterelny and Griffiths1999).
If biology were a subfield of another science, such as physics, then it could seem to fail to be a proper science in the sense intended in the preceding paragraph. However, biology is not a subfield of physics in the way that, for example, optics is. Nevertheless, biology does fall within the scope of physics insofar as all systems at all scales obey physical laws, and this is as equally true of biological systems as it is of geological ones. Biological systems and processes involve physical systems and processes at many scales to which physical laws apply. This does not make biology a subfield of physics or less than fully scientific.
A second moderate autonomy thesis is methodological:
§2 . Methodological distinctness : Biology has its own concepts, methods, problems, and theoretical approaches.
The methodological distinctness of biology is explained by several distinctive features of biological systems: (1) the uniqueness and variability of living things, (2) their high complexity, and (3) the presence of historically acquired information in organisms (Mayr Reference Mayr, Depew and Weber1985, Reference Mayr1996, Reference Mayr2004).
Biological species are not classes of identical things but “variable populations consisting of uniquely different individuals” (Mayr Reference Mayr1996, 101). Even genetically identical clones raised in the same environment differ phenotypically in traits ranging from appearance to personality and life span, due to developmental variation, epigenetic differences, and the stochastic nature of cellular processes (Kirkwood et al. Reference Kirkwood, Feder, Finch, Franceschi, Globerson, Klingenberg, LaMarco, Omholt and Westendorp2005; Vogt et al. Reference Vogt, Huber, Thiemann, van den Boogaart, Schmitz and Schubart2008; Bierbach, Laskowski, and Wolf Reference Bierbach, Laskowski and Wolf2017).
In general, biological systems tend to be much more complex than most nonbiological systems. They have a high degree of hierarchical organization, complex causal and regulatory networks, and high interdependency among their parts. The functioning of organisms involves the coupling of processes at very different scales (Ladyman and Wiesner Reference Ladyman and Wiesner2020, 128). Specific methodologies are required to investigate these kinds of complexity.
Biological systems are also unique in the degree to which they store historically acquired information. While many nonbiological systems also preserve information, biological systems have evolved mechanisms for storing information. As products of evolution, living beings “carry a lot of the history of life with them,” and this history “is encoded in both mechanisms and structure” (Ladyman and Wiesner Reference Ladyman and Wiesner2020, 128). The presence of historically acquired information in living systems legitimates “why” questions that are inappropriate in other scientific contexts and gives an important explanatory role to historical narratives in evolutionary biology.
That differences in the nature of the subject matter should motivate methodological differences is not surprising. But these differences are mainly a matter of degree, resulting in a difference in emphasis and preponderance, rather than a sharp methodological break between biology and the physical sciences (Hull Reference Hull1974, 133). Neither is biology an entirely historical discipline; nor is historical contingency unique to it—consider the importance of history for geology and cosmology and, more generally, the relevance of initial conditions for the study of any dynamical system.
Another moderate autonomy thesis is formulated in terms of explanatory reduction:Footnote 2
§3. Explanatory irreducibility : Biology cannot be explanatorily reduced to physics.
The idea that biology would one day be explanatorily reduced to physics and chemistry was popular in the 1950s and 1960s. For example, Nagel (Reference Nagel1951, 38) thought that the autonomy of biology made sense only as a temporary research strategy, allowing the discipline to be “cultivated as an autonomous branch of science, at least during a certain period of its development” (emphasis added). Francis Crick (Reference Crick1966, 10) famously said that the “ultimate aim of the modern movement in biology is to explain all biology in terms of physics and chemistry.” Today, in contrast, explanatory reductionism is almost universally opposed (Pradeu Reference Pradeu, Barberousse, Bonnay and Cozic2018, 451).
There are three main arguments against explanatory reductionism. First, the argument from functional and teleonomic explanation states that biology is explanatorily autonomous from physical science due to the prevalence of functional and/or teleonomic explanations, which find no counterpart in physics. For example, it is hard to explain the mechanical properties of wings without reference to their function of providing lift and thrust in self-powered flight. If functional explanations are ineliminable and unique to biology, then biology cannot be explanatorily reduced to physics. Second, the argument from multiple realizability states that biological kinds are multiply realizable, meaning that many different arrangements of physical components can realize the same biological kind; thus they cannot be reduced to physical kinds (Fang Reference Fang2018, Weiskopf Reference Weiskopf2011). Third, according to the argument from dual causality, biological phenomena uniquely require both proximate and evolutionary/historical/ultimate causation. The first can be equated to nomological-deductive explanations (Bock Reference Bock2017), whereas the latter reflect an organism’s evolutionary history. Because evolutionary causation is unique to biology, biology is not reducible to the physical sciences.
The next two sections examine stronger, and more problematic, autonomy theses.
3. Strong autonomy I: The irrelevance of physical laws for biological explanations
A stronger autonomy thesis is the thesis that physical laws are mostly irrelevant for biological explanations:
§4 . Irrelevance of physical laws : Physical laws (and perhaps laws in general) are irrelevant for biological explanations.
One reason that physical laws are thought to have little relevance for biological explanations is because physical goings-on are assumed to take place at different levels or scales from biological entities and phenomena. This might be the result of equating physics with microphysics. For instance, Mayr (Reference Mayr, Depew and Weber1985, 45) claimed that “disturbances at the level of elementary particles are ordinarily of no effect whatsoever at the higher levels of biological integration,” and Gilbert and Sarkar (Reference Gilbert and Sarkar2000, 3) argue that “when you have an entity as complex as the cell, the fact that quarks have certain spins is irrelevant.” Physical explanation is often accepted “at the cellular-molecular level” (Mayr Reference Mayr2004, 36) but is claimed to have no relevance for higher levels of integration, such as multicellular organisms, and particularly for their evolution.
However, physical laws pervade biology at all scales, as Green and Batterman (Reference Green and Batterman2017) acknowledge. Physics is not confined to molecules; macroscopic organisms and their parts are also physical entities, and their physical properties afford them their functionality. Even evolutionary explanations of highly contingent events rely on physical laws. Ghiselin (Reference Ghiselin and Ruse1989) gives the example of a paleontologist trying to reconstruct the feeding habits of fossil cephalopods. It is no use relying on taxonomic generalizations—though all extant cephalopods are carnivorous, early cephalopods might not have been. Therefore the paleontologist must rely on physical laws (Ghiselin Reference Ghiselin and Ruse1989, 63), focusing on the mechanical properties of skeletons, teeth, shells, and so on, which condition their possible functionality.
Another argument for the irrelevance of physical laws for biological explanations is based on the alleged irrelevance of laws in general for biology. Dupré (Reference Dupré2021, 7) claims that in philosophy of biology, “the concept of law has largely been abandoned, and replaced by an analysis of scientific understanding through models,” and Fang (Reference Fang2022, 153) argues that the biological sciences do not rely on either biological or physical laws but instead “typically use models to explain biological phenomena.”
However, such models themselves rely on physical laws. For instance, Fang (Reference Fang2022, 136–42) discusses a model of intraspecific competition in access to food (Senior et al. Reference Senior, Charleston, Lihoreau, Buhl, Raubenheimer and Simpson2015). The modeling of “nutrient excesses and deficits when eating nutritionally imbalanced foods” (Senior et al. Reference Senior, Charleston, Lihoreau, Buhl, Raubenheimer and Simpson2015) relies on the chemistry that underlies metabolic processes. More broadly, the model addresses how organisms deal with trade-offs involving energy acquisition and expenditure. These trade-offs are determined largely by biophysical constraints on “the uptake, transformation, and expenditure of energy and materials from the environment” (Carazo Reference Carazo2022), which are governed by physical laws. Hence it is not the case that all models in biology dispense with laws; physical laws are essential background assumptions in many such models.Footnote 3
Evolutionary explanations in biology also rely on physical laws and principles. Because the functionality of biological systems is determined by the physical properties of the system and its environment, functional explanation always presupposes physical principles. For example, the evolutionary history of bird flight relies on physical principles of aerodynamics that govern the functionality of wings and other traits.
4. Strong autonomy II: The primacy of biological explanation
Another strong autonomy thesis is the claim that biological phenomena must have biological explanations.
§5 . Primacy of biological explanation : Biological phenomena must have biological explanations.
Although this view is seldom explicitly defended, it is often implied in discussions of the “failure” of reductionism in biology (e.g., Kaiser Reference Kaiser2011; Nicholson Reference Nicholson2014; Dupré Reference Dupré2021). For example, Dupré (Reference Dupré1993, 93) claims that “biological phenomena must be explained in terms of biological laws and principles.” Similarly, Pradeu (Reference Pradeu, Barberousse, Bonnay and Cozic2018, 451) says that “we cannot adequately explain biological processes by means of physicochemical theories and terms.” While it’s unclear whether these claims should be taken to mean that biological phenomena must be explained only in terms of biological terms and principles, both authors certainly suggest that physical explanations are not typically adequate and are, at best, complementary to biological explanations.
Dupré says that reductive explanation may play a role in accounting for “how things of a certain kind do what they do; but they typically do not help us to understand or to predict what, among the behaviours of which it is capable, a complex thing will do” (Dupré Reference Dupré1993, 106). He is right that the explanation and prediction of a predator’s behavior require an understanding of biological and ecological principles, but facts about its heterotrophic metabolism and adenosine triphosphate production mechanism in cells are relevant for the explanation of why it needs to eat, why it has certain evolved traits that allow it to feed on other organisms from which it obtains organic compounds for oxidation, and so on. Physical and reductive explanations are not limited to explaining the how, that is, the functioning of biological mechanisms.
Indeed, reductive explanations are typical in biology, and often they are the only kind of explanation that is available; for example, there is no correct explanation of how plants make sugars other than the mechanism of photosynthesis understood as a redox reaction. In the next two subsections, we consider two case studies that show that some biological explanations are physical explanations, in the sense that they explain certain biological phenomena in terms of physical laws. Weber (Reference Weber2005, 29) argues that this kind of explanation is in fact the goal of much of contemporary biology. Before moving on to the case studies, three points are important to note.
First, the notion of explanatory reduction has two senses that need not always coincide. One is that of the reduction of an explanation in one theoretical framework to an explanation in another, for example, from thermodynamics to statistical mechanics. This is the kind of reduction of biology to physical science that is at issue in the arguments about explanatory irreducibility. The second sense of explanatory reduction is when the behavior of a whole is reduced to behavior of its parts. This sense applies to the reduction of thermodynamics too, but also to many other reductive explanations in science to which the first sense does not apply. Such explanations are found in biology, for example, the collective behavior and coordination of social insects is explained in terms of their individual traits and interactions.
Second, reductive explanation essentially involves interactions among the parts. Some critics of reductive explanation mistakenly take it to require considering only the parts in isolation and ignoring their interactions. For example, Mayr (Reference Mayr2004, 80) claims that “by failing to consider the interaction of the components, [reduction] fails to fulfil what it promises.” Powell and Dupré (Reference Powell and Dupré2009, 62) argue against a systematic adoption of reductionism on the grounds that “rarely can explanations be given solely in terms of the properties of isolated components,” and they point out that biologically interesting causal powers are often “grounded not in the internal structures of the entities we analytically distinguish, but in the relations between them” (63). Meincke (Reference Meincke2019, 2) characterizes the nonreductionist approaches of systems biology as arguing that organisms “cannot be understood by looking at their parts only; it is the specific interplay of the parts” that needs to be studied.
In contrast, Dupré’s (Reference Dupré, Francisco and Arp2010, 34) characterization of reductionism in the following terms is much fairer: “The reductionist claim should be that the lynx is nothing but a collection of physical parts assembled in a certain way.” Similarly, Weber (Reference Weber2005, 18) characterizes reductionism as the belief that “once the parts of a system and their interactions are understood, there is nothing left for science to explain” (emphasis added). It is the aggregative result of the iteration of many interactions that features in many reductive explanations. This is no less true for reductive explanation of this second kind in physics. Anderson’s famous slogan “More is different,” about condensed matter physics, is true only because the numerosity in question includes interactions as well as individuals (Ladyman and Wiesner Reference Ladyman and Wiesner2020, 66–68). Complexity science is founded on the recognition of how generally this applies and how it illuminates emergence.
The third point to note is that it is often assumed that reductive explanations of the second kind in biology “bottom out” at the molecular level (Machamer, Darden, and Craver Reference Machamer, Darden and Craver2000) and that reductive explanation in biology must be sought in “the interaction of macromolecules” (Rosenberg Reference Rosenberg2006, 54). However, there is no good reason why biological phenomena cannot have explanations at scales both above and below the molecular. Green and Batterman (Reference Green and Batterman2017, 31) point out that many explanations in biophysics and mechanobiology appeal to explanatory properties above the molecular scale, such as tissue stiffness. But scales lower than the molecular can be important too—for example, quantum tunneling is important for respiration (Lane Reference Lane2011), and electron spin may be crucial for explaining navigation in bird migration (Arndt et al. Reference Arndt, Juffmann and Vedral2009).
4.1. Case study: Eye lens protein
Certain mutations in the human γ-D crystallin protein cause early-onset cataracts that manifest in childhood. The explanation of why these mutations cause cataracts in children is a physical explanation, as detailed shortly—the explanandum is biological, and the explanans involves only physical concepts (in addition to the biological concepts in the explanandum).
A biophysical study of the human γ-D crystallin protein shows that although single point mutations at site 23 do not produce any significant structural change in the protein, they dramatically alter its solubility profile. Specifically, the removal of proline from this region of the protein causes changes in the binding energy that are strongly temperature dependent. While the solubility of the native protein increases with temperature, the opposite happens with the mutant protein: It crystallizes when the temperature is raised (McManus et al. Reference McManus, Lomakin, Ogun, Pande, Basan, Pande and Benedek2007). The mutation has a negligible effect on the properties of the protein in the solution phase, whereas the properties of the solid phase are very different. This discrepancy is explained in terms of highly anisotropic interprotein interactions (McManus et al. Reference McManus, Lomakin, Ogun, Pande, Basan, Pande and Benedek2007).
This study is a good example of a biological phenomenon (how a particular mutation causes cataracts in children) that has a physical explanation: The binding energy of mutant γ-D crystallin protein strongly increases with temperature, resulting in a phase transition at physiological temperatures that causes lens opacity. The study also demonstrates that protein function is not entirely determined by its structure but depends also on interprotein interactions. The fact that the interactions between protein molecules are an important part of the explanation illustrates the point that reductive explanations rely on the interactions between parts.
Kaiser (Reference Kaiser2011, 469) suggests an alternative reading of what studying a system’s “parts in isolation” means in the context of reduction: It means “investigating them not in situ, that is, in the context of the system they are a part of, but detached from the system (e.g., in vitro).” She argues that, in systems involving complex integration and interdependence of parts, the insights that can be obtained from studying parts in isolation (in this sense) is limited (471). However, used in this sense, the study of parts in isolation is an essential feature of methodological reduction. In the aforementioned study, it would not have been possible to analyze the chemical potentials involved “without using temperatures that ranged well away from physiological” (Thurston Reference Thurston2007, 18878).
As for the claim that reductive explanations are less adequate the more complex and highly integrated a system is (Kaiser Reference Kaiser2011, 471), this is not always the case. Which level of explanation is the most adequate should be determined by empirical investigation and depends on the specific features of the phenomenon under study. As the human γ-D crystallin protein explanation of childhood cataracts demonstrates, for some biological phenomena, a physical explanation is not only possible but correct.Footnote 4
4.2. Case study: Water transport in trees
Another interesting case of a physical explanation of a biological phenomenon is the cohesion-tension theory of water transport in trees (Dixon and Joly Reference Dixon and Joly1895; Tyree and Ewers Reference Tyree and Ewers1991; Niklas and Spatz Reference Niklas and Spatz2012). Trees require a specialized water transport system that is able to carry large amounts of water against gravity, from the roots up to the leaves. Water is transported from the roots up to the leaves through the xylem, which consists of long, thin capillary vessels composed of dead cells (tracheids and vessel elements). Xylem vessels are filled with a continuous water column, and the pull from above is provided by transpiration in leaves, which occurs through specialized structures, the stomata. This produces a negative pressure gradient that pulls the water upward against gravity. The continuous water column maintains its stability due to water cohesion and adhesion to the inner surfaces of the tracheids, and its tensile strength is remarkably high (Niklas and Spatz Reference Niklas and Spatz2012, 95–96).
The main principles involved in the physical explanation of water transport in trees are the following: the high cohesive force of water molecules, surface tension, evaporation, and hydraulic principles that govern the flow of liquids through pipes.
The high cohesive force between water molecules and surface tension results from physical properties of the water molecule, namely, its polarity (Bagcchi Reference Bagcchi2013). The asymmetric distribution of electrons across the oxygen and two hydrogen nuclei means that the oxygen end of the molecule is partially negatively charged, whereas the hydrogen end is partially positively charged. This causes water molecules to electromagnetically attract each other, forming hydrogen bonds between then, which results in a high cohesion among water molecules that generates a strong surface tension when in contact with a nonpolar surface like air (Bagcchi Reference Bagcchi2013, 9).
Water evaporation takes place in specialized structures in the leaves, the stomata, and depends on the water potential at the interface between the liquid and gaseous phases, which in turn is related to water vapor pressure (Niklas and Spatz Reference Niklas and Spatz2012, 95). Usually, the water vapor pressure is higher within the stoma than in the atmosphere, driving evaporation.
The physical explanation of water transport in trees is particularly interesting for two reasons. First, it is unexpected that the mechanism responsible for the transport of water many meters upward against gravity turns out to depend on the physical properties of water molecules and on purely physical hydrodynamical principles. Second, it is an example of a problem faced by certain plant lineages under evolutionary pressure to grow taller that was solved by evolution by countering one physical force (gravity) with another (electromagnetic forces between water molecules). This solution neatly illustrates the fact that physical laws both constrain and enable the functioning of biological systems. One reason why physical explanations play important roles in biology is because physical laws are often at the heart of both the problems encountered by evolving biological systems and the solutions they find for them.
5. Conclusion
Biology is autonomous from physical science in various ways. Biology is a genuine science (scientific status) that has its own concepts, methods, problems, and theoretical approaches (methodological distinctness). Biological explanations are not reducible to physics (explanatory irreducibility). However, the case studies show that there are physical explanations of biological phenomena, and physical laws are relevant to biological explanations in general (the associated autonomy claims irrelevance of physical laws and primacy of biological explanation are false). The case studies also illustrate that explaining things in terms of their components and their interactions is not the same as explaining things in terms of components behaving as they do in isolation (though studying components in isolation may be scientifically fruitful to some extent). However, whether there are reductive explanations is nothing to do with whether everything can be reduced to fundamental physics, because there are reductive explanations of biological phenomena to other biological phenomena, and reductive explanations within physics itself, that do not involve the entities and processes of fundamental physics. Furthermore, biological systems are made of physical components at many scales and obey physical laws at those scales, not just microphysical ones. Which scales and laws are explanatory in a given case is discovered empirically and cannot be known a priori. Biology and physics are much more integrated than is sometimes claimed.
Acknowledgments
MH was supported by a Jacobsen Postdoctoral Fellowship in Philosophy at the University of London. The authors are grateful to audiences at the Physics at the Interface workshop (Durham), the Workshop on the Philosophy of Biophysics (Bristol), and PSA 2024 (New Orleans).
Open access
