1.1 Introduction

To many biologists, science and philosophy may appear an odd couple without much in common. Perhaps the word “philosophy” will even bring to mind endless arguments and speculation about whether the chicken or the egg came first, without ever getting anywhere. After all, are philosophers not still arguing over the same things as Aristotle and his fellow Greeks? Well, yes. But biologists too are concerned with the questions that occupied Aristotle: what living beings are and where they come from; how they develop, function, and interact with one another; and why there are so many forms and how those forms should be classified. There has been tremendous progress in biology, of course. But it does not appear that biologists will ever run out of questions. This is because good science does not only reveal new things about the world; it also reveals that there are things we did not even know we could know. So we want to know more.

Not all research is equally effective in promoting the advancement of science, and it is therefore useful to reflect on what works and what does not. In fact, while biologists may think of themselves as busy enough just doing science, they have been and still are preoccupied with metascientific questions and issues, which is what philosophy of science is about: how to think about genes or ecosystems, the nature of species, the causes of evolution, the value of experimentation versus observation, and if molecular or evolutionary biology is the most fundamental of the biological sciences, to take just some examples. Perhaps even more significantly, twentieth-century genetics and molecular and evolutionary biology were all shaped by attempts to ensure that the biological sciences meet the criteria of a “hard science,” mapped on ideals in physics.

This ideal no longer seems as appealing as it once did. Over the last decades, studies in philosophy of science have revealed that scientific aims, methods, models, and concepts are much more diverse than previously envisaged by either scientists or philosophers. The biological sciences have been important for reaching this conclusion because biologists turn out to be very flexible in their scientific exemplars. What goes on in ecology and molecular or evolutionary biology can at times bear limited resemblance to each other, let alone to traditional exemplars in physics. Yet, the biological sciences are hugely successful and influential. This success is difficult to account for if the standards of mid-twentieth-century physics were the only way to ensure knowledge.

The present volume was put together because we believe that being aware of the rich conceptual and methodological issues of science can make biologists better at what they are doing – as students, teachers, scientists, and professionals. Philosophy of science is not the preoccupation of armchair philosophers or retired scientists, but it is rather central to any scientific endeavor. The problem is that, more often than not, biologists are not educated and encouraged to pursue this kind of reflection. The present volume is an attempt to support them in doing this.

1.2 Science and the Philosophy of Science

While biologists study life, philosophers of science study science (see Lewens Reference Lewens2015 for an accessible introduction to philosophy of science). Not unlike biologists, some philosophers are motivated by the big picture, whereas others are obsessed with a particular problem (Box 1.1). Many philosophers of science work on particular “study systems,” such as molecular biology or evolutionary theory. This requires familiarity with the aims, methods, and knowledge of each particular area, which is one reason why philosophers need to engage with scientists. Another reason is that science is shaped by human abilities, interests, and values. Sometimes it is not possible to understand science without understanding the scientists themselves. This work often draws on the history of science to reveal how events actually played out or how facts and values influence each other over time (see Chapter 13). Being scrutinized can feel uncomfortable for scientists, in particular if they believe that science is – or should be – free of such biases. As illustrated by several of the chapters in this book, this belief is not only mistaken but can actually be detrimental to science itself. The questions, methods, and models that a scientific community considers exemplar are shaped in part by shared attitudes and beliefs. Ignoring these social aspects of science can make biologists less well equipped to identify and solve scientific problems, and it can make them struggle to handle controversies between scientists and between science and other parts of society (see e.g., Chapters 7, 12, and 14)

For philosophy of science to become useful to biologists, scientists and philosophers need to find ways to communicate, share ideas and results, and perhaps occasionally work together. As biologists who have attempted to work with people from other disciplines will testify, collaboration is easier said than done. One main hurdle is simply ignorance about each other’s work. Another is to become familiar with terminology and habits of mind that are often specific to particular disciplines. These hurdles can be overcome. Nevertheless, just as it will take time and effort for a cancer researcher to figure out if insights from evolutionary theory will be useful to her, it will take time and effort to figure out if philosophy of science will be useful to you. We hope that this collection of chapters will be helpful for those who are willing to dedicate their time. At the end of the book, we provide suggestions for further reading on both general topics in philosophy of science, as well as topics that are likely to be of particular interest to biologists. In the last chapter we also make concrete suggestions about how topics from philosophy of science could be taught to biology students.

1.3 Kuhn and Popper As Caricatures

Perhaps no philosopher of science is more familiar to scientists than Karl Popper. Popper was concerned with the big questions in philosophy of science, and his work has had a long and lasting intellectual impact, not the least on scientists (Lewens Reference Lewens2015). His idea that scientific hypotheses can never be proven, only falsified, is commonly introduced to beginners in the natural sciences as the fundamental feature of science. Falsification not only separates science from non-science but Popper also meant that the repeated failure to falsify hypotheses can account for the growth of scientific knowledge. Another philosopher of science who is likely to be mentioned in introductory science classes is Thomas Kuhn, famous for introducing the idea of a paradigm shift (Kuhn Reference Kuhn1996). Scientists tend to be more ambivalent toward Kuhn since he emphasized that science is a collective, social endeavor where scientists sometimes appear irrational. But Kuhn’s concept of paradigm shifts can be interpreted as a radical theory change introduced in the face of repeated falsification of established theories. For many biologists, this view of how science works – steadily securing knowledge through hypothesis testing and rarely interrupted by radical theory change when major hypotheses are disproven – may be the entire philosophy of science they are exposed to during their studies, perhaps even for their entire academic career.

No shame on Popper and Kuhn, and scientists are often taught a caricature of their work (like we just did!), but this is not really enough to understand how science works. Believe it or not, philosophy of science has progressed! While falsifiability remains an important litmus test for a scientific hypothesis, it is now widely recognized that the building of knowledge through falsification of a priori hypotheses is a poor characterization of many successful sciences, including biology. Scientific knowledge and understanding is generated through much more diverse standards and activities than envisaged by these early philosophers of science. There are good reasons for this diversity. The world is immensely complex, and humans are limited beings. Thus, it is reasonable that different scientific questions demand different approaches or methods. However, a diversity of scientific standards does not imply an absence of standards. It is important to understand what works and why.

In practice, biologists tend to pick up most ideas of what science is and how it works from fellow biologists, typically those who work on similar problems using similar methods. But if there is no universal standard of science, this can make it difficult to recognize or understand the importance of research that uses different standards or, for that matter, the limitations of one’s own approach. Such failure can lead to inefficient science, missed opportunities for scientific breakthroughs, or even long and fruitless controversy. In what follows, we reflect on three features of science – its aims, methods, and concepts – to make a case for why biologists can benefit from insights gathered from the philosophy of science.

1.4 Scientific Aims

What are the aims of science? A short list would likely include description, classification, prediction, and explanation. Biologists describe and classify new species, molecules, and biological processes; they predict the effects of human activities on biodiversity or the spread of disease; and they explain how cells work and why populations evolve. A main reason for these activities is that many biologists ultimately strive to understand living systems, such as cells, organisms, and ecosystems. This understanding has practical consequences for technology, medicine, and many other features that make up societies, and it is therefore important far outside academic circles.

A phenomenon can be said to be understood when one can give it a satisfactory explanation (see Chapter 2).Footnote ¹ Given that we explain phenomena all the time, it will perhaps come as a surprise to learn that it is neither obvious what it means to explain something, nor what, if anything, that makes scientific explanations different from everyday explanations. The traditional point of view on behalf of philosophers of science is that scientific explanations consist of statements that demonstrate that the phenomenon to be explained follows from natural law (Woodward Reference Woodward and Zalta2017). This account of explanation is heavily influenced by physics, and biologists hardly find it very appealing since there is a widespread skepticism toward the existence of biological laws.

A more promising idea is that explanation is linked to causality, manipulability, and control (e.g., Woodward Reference Woodward2003).Footnote ² It will feel natural to biologists to think of causes as difference-makers (Illari & Russo Reference Illiari and Russo2014). Rain causes seeds to germinate because if it had not rained the seeds would remain dormant. Loss of genetic variation causes population extinction because if it were not for the loss of genetic variation the population might have adapted to the environment. One view of scientific explanation is that it is achieved when the information provided by the explanation allows one to answer a range of such what-if-things-had-been-different questions (e.g., Woodward et al. Reference Woodward2003; Strevens Reference Strevens2008; Potochnik Reference Potochnik2017). For example, an explanation for how ATP is generated may refer to biochemical features of glycolysis. This explanation reveals something about the causal tapestry of the world; the molecular detail makes it possible to grasp the consequences of a change in the concentration of pyruvate or the chemical structure of the reacting molecules. According to some philosophers, this is what it means to understand how ATP is generated, and the more what-if-things-had-been-different questions about ATP production we can answer the better we understand it.

Not all explanations in biology are mechanistic like this, however, and many explanations in biology look more like historical explanations (see Chapter 10). An explanation for the extinction of dinosaurs may refer to a meteorite that struck the earth and caused long-term changes in the earth’s climate. Nevertheless, the reason why this explanation generates understanding is similar to the case of ATP; reference to the meteorite and its effect on climate makes it possible to grasp what would have happened to the dinosaurs if the meteor had not have struck the earth, or if it had been smaller, or if there had been no competing mammals around. There may be other kinds of scientific explanations, but being able to give answers to what-if-things-had-been-different questions appears to at least be one important feature of many scientific explanations.

A good thing about this notion of explanation is that one need not take truth too seriously. What really is “out there” may forever be out of reach, but representations of the world can be sufficiently good approximations that enable one to foresee what would have happened if things had been different. It is not always possible to support the explanation through active intervention, of course (this is difficult for the dinosaur extinction, for example). But scientists can nevertheless ensure that their theories are empirically justified – or true enough – by imagining and studying a range of different situations. This is why it is important that scientific theories are falsifiable; if a theory makes no falsifiable claims, it also appears impossible to predict the consequences of an intervention.

Another helpful feature of the causal theory of explanation is that it brings attention to the fact that scientists need to manage causal complexity (Potochnik Reference Potochnik2017). Biological systems are enormously complex, and any representation of a living system will only capture some of its actual causes. This is in itself not a problem. In fact, too much detail makes it harder to grasp what would have happened if things had been different. A diagram of all causal interactions in a cell would describe the cell but not explain how it works. To explain phenomena, scientists leave things out (abstraction) and make assumptions that are false (idealization). Abstraction and idealization play positive roles in explanation because they foreground the causal relations that are of interest – idealization makes the phenomenon appear as if it were produced by the focal causes alone (Potochnik Reference Potochnik2017). As a result, how one thinks about biological processes influences which of the myriad of actual causes of a particular phenomenon that are picked out as being explanatory causes.

It may be helpful to illustrate this feature of idealization using a nonbiological example. Consider the frequent delays of trains arriving into Stockholm Central. One possible cause of these delays is that late departures of trains that are not headed toward Stockholm can propagate through a jammed train system – a kind of cascading effect. To see if this can explain the arrival delays into Stockholm C, transport planners may benefit from assuming that trains run at a constant speed unless they have to stop to let other trains pass. Plugging in real data on train speeds and how they vary seems unnecessarily complicated. Doing so might even make it harder to grasp how improving departure punctuality of trains throughout Sweden will affect arrival times of trains bound for Stockholm C.

The transport planners will feel satisfied if there is a good fit between their model and actual arrival times into Stockholm C. They could claim that they now understand why trains are delayed, because they can explain it in terms of cascading effects of delayed departures of trains bounded for other destinations. However, imagine that it turns out that, contrary to what the model predicts, the actual arrival times are unaffected by such departure delays. One possible explanation for this mismatch between model and reality is that train drivers adjust the speed to compensate and ensure that trains headed for Stockholm have a free pass. This may appear to imply that train speed is a cause of punctuality but not of delays. But this cannot be the case because slow-downed trains can also jam tracks and propagate delays.

What is happening here? Firstly, note that train speed initially appeared unable to account for the arrival delays because it was idealized away from the model. Secondly, it is only when the proposed model was unable to account for the phenomenon that we looked for another cause. This is why train speed appeared as a possible explanation for why trains were not delayed, but not an explanation for why they were delayed. But there is no fundamental causal asymmetry here. Interventions on either departure times or train speeds can cause arrival delays of trains bound for Stockholm C because both can result in interference between trains. As a result, a satisfactory explanation for the late arrivals may require the use of multiple different idealizations, each one suitable for picking out the contribution of a particular cause or set of causes. Trying things out and keeping what appears important may eventually allow more complex representations that have greater explanatory power.

The challenges that transport planners face are also faced by biologists. Biological phenomena are produced and sustained by many factors, and these are often causally intertwined. As a result, there can be several legitimate explanations of the same phenomenon, each drawing on only some of its causes. These explanations are often sufficiently different to happily coexist. One example is the distinction between what biologists commonly refer to as ultimate and proximate explanation. Roughly speaking, ultimate explanations are considered historical explanations that trace events that occur within a population or a lineage, whereas proximate explanations are considered mechanistic explanations at the level of the individual. An ultimate explanation for why mammals maintain a high body temperature may, for example, refer to its fitness benefits in cold climate, which implies that this trait became increasingly common and sophisticated as a result of natural selection. A proximate explanation for the same phenomenon may refer to the autonomic, neuronal, and molecular mechanisms that underlie the ontogenetic development of endothermy.

Following the highly influential work of Ernst Mayr (Reference Mayr1961), it is customary in evolutionary biology to consider that causes that feature in proximate explanations should not be invoked to explain evolutionary adaptation (e.g., Dickins & Rahman Reference Dickins and Rahman2012). However, a closer look at the rationale for this distinction reveals that it relies on an idealization of the evolutionary process that foregrounds fitness differences and screens off other putative causes of adaptive change (see Walsh Reference Walsh2015; Pocheville Reference Pocheville, Uller and Laland2019; Uller & Helanterä Reference Uller and Helanterä2019). This reflects that the main agenda for evolutionary biology has been to understand the role of natural selection in adaptive evolution, not the role of development, physiology, or behavior. The assumptions made in evolutionary theory tend to turn the latter into constraints; they can account for the absence of adaptive fit but not its presence.

This line of thought is so common to biologists that many take it for granted. However, a comparison to the explanations for the delayed train arrivals is a reason to treat this conclusion with caution. That is, that one particular idealization of evolution by natural selection privileges genes and natural selection does not imply that there is an inherent causal asymmetry in evolutionary processes (Laland et al. Reference Laland, Sterelny, Odling-Smee, Hoppitt and Uller2011). The role of proximate causes in adaptive evolution is in fact one of the most persistent controversies in biology (Amundsen Reference Amundsen2005). Contemporary examples include the disagreement over the explanatory role of development, plasticity, extra-genetic inheritance, and niche construction in evolution (see Laland et al. Reference Laland, Uller, Feldman, Sterelny, Muller and Moczek2014, Reference Laland, Uller, Fellman, Sterelny, Muller and Moczek2015). One possible reason that these issues are difficult to resolve is that the genetic representation of evolution is commonly taken at face value, rather than being understood as an idealization designed to explain evolutionary phenomena in terms of natural selection. An increased awareness of the relationship between idealization and explanation may reduce the risk that causes that are idealized away become permanently neglected, facilitate capitalization of insights from other disciplines, and put a restraint on unproductive scientific controversy.

While there are good reasons why a biological phenomenon like adaptation can have several explanations, biologists may sometimes wish to determine which of a number of different explanations is the most satisfactory (see Chapter 3). Consider cichlid fish, famous for the ability to evolve very similar morphologies in different lakes (Seehausen Reference Seehausen2006). Evolutionary biologists have demonstrated that this convergence happened because the local habitat and foods are often similar in different lakes, which favors a limited set of life styles such as bottom-dwelling grazers and open water predators (e.g., Muschick et al. Reference Muschick, Indermaur and Salzburger2012). Thus, natural selection explains the convergent evolution of cichlid fish. But biologists have also pointed out that some of the recurring features of these fish, such as the shapes of bodies and jaws, tend to be plastic (Schneider & Meyer Reference Schneider and Meyer2017). That is, those characters respond to the habitat or diet that individual fish encounter during their lifetime. Some biologists believe that plasticity has contributed to the striking convergence between different species because plasticity can make some features more likely to become selected than other, perhaps equally fit, phenotypes (see West-Eberhard Reference West-Eberhard2003). In terms of explanation, the question is whether or not an explanation for adaptive convergence in terms of plasticity and natural selection is better or more satisfactory than an explanation in terms of natural selection alone (Uller et al. Reference Uller and Helanterä2019).

Picking the best of two explanations is easy if only one of them is backed up by empirical evidence (although negative evidence is not always enough to give up a hypothesis, as explained in the next section). Beyond this, there may be nothing inherent in these explanations for adaptive convergence that make one better than the other; perhaps picking the best explanation is simply a matter of one’s interest (assuming both explanations are empirically justified). However, scientists sometimes prefer one explanation over others if it applies to many different phenomena rather than a few, if it is easy to understand, or based on its elegance or simplicity (Ylikoski & Kourikoski Reference Ylikoski and Kuorikoski2010). Explanatory standards such as these are important since the criteria for picking the best explanation influences what biologists consider knowledge (Chapter 3). Such explanatory standards can vary between biological disciplines (e.g., those concerned with mechanistic versus historical explanations), but these differences may not be well recognized by biologists. As a result, it can be difficult to assess the quality and scope of biological research that lies outside of one’s immediate expertise.

These examples from evolutionary biology reveal exactly how difficult it is to do science. We have discussed scientific explanation at some length here because we believe that it is a good illustration of how philosophy of science can be helpful to biologists, but similar issues also come up for other scientific aims, such as classification (Chapter 11). In our opinion, the philosophy of science that is usually taught in introductory classes does not pay sufficient tribute to the challenges that scientists face when they attempt to produce knowledge and understanding of a world that is enormously complex. How biologists choose to represent the world influences the questions they consider worthwhile, how they organize their research, and the answers they look for. This means that scientific advancement requires the flexible use of methods and concepts.

1.5 Scientific Methods

It is probably obvious that biologists do not follow a single universal scientific recipe, but rather several more or less distinct approaches (see Chapter 9). Biological science certainly appears to frequently stray away from a strict method of falsification. A careful look behind the scenes of scientific papers that communicate the results of a test of a well-defined hypothesis will often reveal a process that looks very different to what we have been told science should look like. This is one reason why p-values are so problematic; scientists’ formulation of hypotheses often develops in parallel with observation, practice, and data collection rather than in a strictly ordered fashion. It may be tempting to conclude that failure to adhere to the strict rules of hypothesis testing makes some biological research fundamentally flawed. A more optimistic view is that scientific practice simply reflects that there is a diversity of scientific methods, all of which may be appropriate. Regardless, it is important to examine how different scientific methods achieve scientific aims (e.g., understanding), what tools are available to meet these aims, and how well those tools actually work in practice. Biologists are highly engaged in these issues, in particular with respect to the appropriate use of statistics, lab- vs field-based methods, the use of model systems, and so on. These discussions may benefit from a greater attention to the literature on philosophy and history of science, which often has a fair bit to say on the matter (e.g., Chapter 10; Chapter 9).

Another peculiar break with falsification is that scientists only occasionally and reluctantly abandon a hypothesis when the data fail to support it. For example, a frequent failure to demonstrate that offspring of males with exaggerated sexual signals were fitter than other offspring did not generally make behavioral ecologists abandon the “good genes” hypothesis (Roughgarden Reference Roughgarden2009, pp. 213–224). This widespread practice is difficult to make sense of if biologists really believed in falsification as a method to scientific progress. Eventually a hypothesis may of course fail to be confirmed and become abandoned, but this will not result in a wholesale rejection of the theory. For example, the biologists that did conclude that there was little if any evidence for good genes rarely used this to argue against the theory of sexual selection (but see Roughgarden Reference Roughgarden2009). This is not irrational behavior on behalf of the scientists. Instead, examples such as these demonstrate that scientific theories are organized into more fundamental theoretical frameworks that are not revised on the same basis as a more specific hypothesis. Philosophers of science have introduced and use many different concepts to make sense of this feature of science, including thought styles (Fleck Reference Fleck1979), paradigms (Kuhn 1962), research programs (Lakatos Reference Lakatos1978), and problem agendas (Love Reference Love2008; see Chapter 7).

An example from population biology can illustrate this point. During range expansions, populations on the front line are often small but rapidly growing. The combination of bottlenecks and rapid population growth can make particular genetic variants become very common even if they do not bring any fitness advantage (Excoffier et al. Reference Excoffier, Foll and Petit2009). The formulation of this idea – sometimes referred to as allele surfing – relies on a set of population genetic principles (see Charlesworth & Charlesworth Reference Charlesworth and Charlesworth2010). These principles derive from abstraction and idealization of complex biological processes, which are considered appropriate to solve a particular problem or kind of problem. Biologists may refer to both allele surfing and the set of population genetic principles as “theories,” but it is only the former that can be falsified (theoretically by demonstrating that the conclusions do not follow from the premises, and empirically by demonstrating that allele surfing is not something that actually happens in real populations). Principles relevant to some discipline, methodology, or problem are typically not falsifiable. In fact, idealizations are commonly made despite full knowledge they are false (Potochnik Reference Potochnik2017). Putting these “theories,” such as population genetic theory, to the test means to assess how well those theories deliver satisfactory explanations, not to try to prove them wrong. In the next section, we argue that this is one reason why conceptual analysis can advance science.

Despite these reasons to doubt the classic view of the scientific method, many biologists adhere to its core features including, of course, the notion that experiments are the key to scientific knowledge and understanding. However, not even experiments are fundamental to all biologists. Many evolutionary biologists, for example, rely heavily on observation and “traces” of past events to reconstruct what happened and explain why it happened (Currie Reference Currie2018; Chapter 10). Even molecular biology – perhaps the “ideal” reductionist science of biology – makes use of nonexperimental inference, for example, when species comparisons are used to substantiate claims that the activity of transposable elements (“jumping genes”) causes genome evolution (Bourque et al. Reference Bourque, Burns, Gehring, Gorbunova, Seluanov and Hammell2018). In her chapter, Cleland (Chapter 10) argues that experimental science is not as reliable as popularly thought, and historical science is not so unreliable. This may be welcome news to some biologists, and worrying to others. Regardless, it will be useful to become aware of the possible advantages and disadvantages of different scientific methods.

1.6 Scientific Concepts

Progress in the biological sciences is commonly driven by new discoveries and technological breakthroughs. The history of genetics provides many good examples, including PCR, high-throughput sequencing, and CRISPR-Cas. Nevertheless, data alone is often not enough, and conceptual analysis too can advance science (see Chapter 5, Chapter 8). This should perhaps not be surprising since concepts organize research agendas and feature in models, theories, and explanations. Since these are core features of science, understanding concepts may help to understand phenomena, or at least make science more effective at understanding phenomena. Concepts should not be confused with terminology, since a single concept can have several terms and the same term can be used for several concepts.

A good example of the latter is “gene,” which in modern biology routinely refers to several different concepts (Griffiths & Stotz Reference Griffiths and Stotz2013; Kampourakis Reference Kampourakis2017). Biologists sometimes express the feeling that this multitude of meanings only creates problems and confusion. The solution, in their view, is a clear definition and a precise one-to-one map from concept to term. One illustrative example is many biologists’ frustration over the term “epigenetics” (e.g., Deans & Maggert Reference Deans and Maggert2015; see Baedke Reference Baedke2018 for a conceptual analysis). For example, the journal Cell Reports requires authors to adhere to a strict definition of epigenetic (and epigenetics) and does not allow authors to refer to epigenetic as a stand-alone term.Footnote ³ This certainly removes ambiguity with respect to what biological feature that epigenetic refers to in papers published in this journal. This can be a good thing. Nevertheless, there are reasons to be skeptical toward this puritanism since concepts are more than placeholders of specific biological entities or phenomena. For example, the concepts of “gene,” “genetic,” or “genetics” do not have a similar one-to-one identity as what Cell Reports demands for epigenetic. That biologists appear reasonably content with genetic, whereas the ambiguity in epigenetic is a major source of frustration and confusion, suggests that there is something to be learnt from understanding how biological concepts are formed, used, and transformed (see Chapter 5).

Conceptual analysis aims to clarify the use of concepts and the different roles that those concepts play in science. One role is to circumscribe a phenomenon or entity, which is often accompanied by a definition. This is how Cell Reports views the epigenetic concept. Another important role of concepts is to set a research or problem agenda (Love Reference Love2008). Arguably this too is a major role of the epigenetics concept, similar to one of the meanings of genetics. Since its inception, epigenetics has changed meaning several times, but it has consistently been used to refer to a set of biological problems, processes, and entities that are tied to a particular explanatory aim. To Waddington, who is often considered to have coined the term, this explanatory aim was to understand development (e.g., Waddington Reference Waddington1957). To molecular biologists today it may be to understand gene regulation (e.g., Allis et al. Reference Allis, Caparros, Jenuwein, Reinberg and Lachlan2015). To epidemiologists, it may be to understand the molecular mechanisms that can cause the lifestyle of parents to influence health and disease of their offspring (e.g., Mill & Heijmans Reference Mill and Heijmans2013). This use of a single concept to refer to a multitude of research agendas can obviously lead to confusion. But recognition that concepts such as epigenetics can be useful even without a consensus definition should make the situation easier to handle. Historical analysis of the gene concept actually suggests that the fuzziness of the concept sometimes has been a strength since it allowed formulation of a research agenda that was more inclusive than it otherwise might have been (Rheinberger Reference Rheinberger, Beurton, Falk and Rhienberger2000). Familiarity with such examples from the history of biology will make biologists better able to handle the conceptual complexity of contemporary biology and avoid some of the more common pitfalls.

One such pitfall arises from the fact that many scientific concepts are metaphors (see Chapter 6). Two familiar examples are membrane “pumps” and “channels.” These concepts make use of familiar objects to make it easier to grasp an important biological difference in cell membranes; a pump requires ATP to move ions across the membrane, whereas channels allow ions to freely flow from one side to the other. But it would be a mistake to attach to the biological pumps and channels some other properties that are distinctive of water pumps and channels. For example, water pumps need to be constructed, controlled, or programmed, whereas water channels may passively form and function without external control. It is not obvious that these differences between water pumps and channels also apply to membrane pumps and channels; applying them may even be misleading. Human brains are, however, prone to make such associations. This is why it is often pointed out that the “cost of metaphor is eternal vigilance.” One widely discussed example where vigilance may have slipped is the concept of a genetic program. As Susan Oyama (Reference Oyama2000), Evelyn Fox Keller (Reference Keller2000), and others have demonstrated, the metaphorical use of “program” or “blueprint” exercised an enormous influence on research on genes, development, and evolution, and it continues to exercise an important hold over scientific and public understanding of biology. Some biologists shrug this off and claim that everyone always knew there was no “blueprint,” perhaps offering “recipe” as a better metaphor. Familiarity with the history and philosophy of science will make one wary of such recollections of the past, and encourage vigilance when new metaphors replace the old ones.

As explained in the previous section, concepts also function to organize larger bodies of theory or research efforts. Such conceptual “frameworks” differ from scientific hypotheses or conjectures, even if both of them sometimes are referred to as “theory.” Biologists usually keep track of this distinction, but confusion over the structure of scientific concepts can be one reason for scientific controversy (see Chapter 12). The controversy over the role of development and other proximate causes in evolutionary explanations already discussed may be one example. Some biologists appear to think that this controversy will be resolved by proving the gene-centric conceptual framework right or wrong (at least, this is one way to interpret Noble Reference Noble2013 and Wray et al. Reference Wray, Hoekstra, Futuyma, Lenski, Mackay, Schluter and Strassmann2014, to exemplify with two opposite conclusions). However, there are reasons to doubt that falsification is an appropriate criterion here. Population and quantitative genetic theory are not conjectures but collections of principles, adopted to address particular kinds of problems. Rather than empirical falsification, a more realistic aim is to clarify the explanatory limits of such conceptual frameworks and suggest alternative representations that may be better suited to move beyond those limits. The literatures on developmental plasticity, extra-genetic inheritance, and niche construction are good illustrations of this endeavor (Laland et al. Reference Laland, Uller, Fellman, Sterelny, Muller and Moczek2015).

This endeavor is complicated by the fact that it is possible to explore the role of plasticity, non-genetic inheritance, and niche construction in evolution without giving up on idealizations that will grant a privileged role of natural selection and genes (see Section 1.4). One particularly good illustration is the large body of research on “plasticity-led evolution” that has followed since the publication of Mary-Jane West-Eberhard’s book Developmental Plasticity and Evolution (West-Eberhard Reference West-Eberhard2003). Much of this research represents plasticity as a property of genotypes (biologists call these reaction norms; Levis & Pfennig Reference Levis and Pfennig2016). As a result, any contribution of plasticity to adaptive evolution can be considered a consequence of natural selection on genes, rather than as a primary cause of adaptive change (see Uller et al. Reference Uller and Helanterä2019). Examples such as these illustrate that simply “extending” a theory to include new phenomena need not resolve contention. It is also important to be aware that biologists’ interpretative understanding of these phenomena will be dictated by their conceptual framework, that is, how they think about living beings.

1.7 Concluding Remarks

As biologists we feel that we have benefited from reading philosophy of science and engaging with philosophers. Not all biologists will feel the same, of course. But we do not believe that we are particularly unusual. Any field probably benefits from a diversity of perspectives, and this diversity tends to grow from encouraging reflection and critical assessment, not the least from scientists within the field. A little bit of philosophy of science is one way to make this happen.

The chapters in this book can be read in any order and where one would like to start depends on one’s interest. Some chapters are easier to digest than others, and not everything will be to everyone’s liking. Our advice to biologists is to look for issues that feel most relevant to their own work and begin there. If you are engaged in a field that is controversial, consider if the controversy could partly be dissolved through conceptual analysis or a more explicit formulation of the idealizations that are used for the phenomena you study. If you struggle to see the value of someone’s research, or even an entire field, consider if your opinion is a result of different aims, problem agendas, or methods, or if it is shaped by preferences, values, and beliefs. If you look for new and exciting ways to tackle your problem, identify causes that are currently screened off, and alternative concepts and metaphors that may prove fruitful. If you are engaged in public outreach, consider if your research can be communicated more effectively by emphasizing the process by which knowledge is generated. Above all, stay curious, not just about biology, but also about the nature of the biological sciences.