What Use Are Real-World Cases for Philosophers?

  • Katherine Furman (University of Liverpool)


In this paper I provide a defence of real-world cases as a legitimate part of the philosopher’s toolkit, in addition to the austere thought experiments and fictional cases that are more commonly used. I argue that thought experiments are effective because they streamline out extraneous details that might distract the philosopher from the principle under investigation. But in doing so they run the risk of inadvertently removing relevant information, thus preventing the philosopher from latching on to salient philosophical relationships. Fictional cases operate as extended thought experiments—removing what is hopefully irrelevant, but potentially at the cost of information that the philosopher needs. Cases from the real world are thus the only place that we can be sure that nothing important hasn’t been inadvertently lost, and so they are philosophically important.

Click the links below to download an audio version of this article's content:

How to Cite:

Furman, K., (2021) “What Use Are Real-World Cases for Philosophers?”, Ergo an Open Access Journal of Philosophy 7. doi:



Published on
22 Oct 2021
Peer Reviewed

1. Introduction

Philosophy emphasises abstracting away from the particularities of the world. If a purported piece of philosophy strays too far in the direction of the practical, its philosophical credibility is at risk (Srinivasan 2018: 1410; Jenkins 2014: 264). Jenkins, for instance, describes a situation from her student-days when she told a class-mate about her work on metaphysics and gender, to which her interlocutor responded that this is not “philosophical” (2014: 264). Finding examples of this kind of attitude in print is tricky, given that those who hold these views are unlikely write about them, but anecdotes like Jenkins’s abound.

It is not clear why concerns about the credibility of real-world orientated philosophy persist, given that criticisms against the armchair vision of philosophy are old (see Williams 1972; 1981; 2002). Regardless of this issue’s persistence, philosophical use of material from the world still requires a defence if it is to be taken seriously. In this paper I offer such a defence.

To start, it is fairly uncontroversial that the clarity and rigour provided by philosophical thinking can be useful for achieving better understanding in certain real-world cases (Srinivasan 2018: 1410; Kamm 2009: 19–20). In the sciences, for instance, philosophers often perform important clarificatory work when conceptual confusion arises (Kitcher 2011: 253). Somewhat more controversially, it can be argued that philosophers have an obligation, either professionally or ethically, to spend a portion of their work-time devoted to real-world cases to assist with this important clarificatory work (Jones 2006). I will not address either of these claims in this paper—I will not defend real-world philosophical engagement as a kind of professional public service.

Instead, my target is the naysayers—those who believe that no philosophical gains can be achieved by considering the real world. Taking this as my target, I focus on defending the claim that philosophical benefits can be achieved by paying careful attention to real-world cases: that is, over and above the thought experiments that are already a standard part of the philosophical practice, or the more detailed fictional cases that are often advocated for when thought experiments are found to be insufficient (see, for instance, Nussbaum 1990). I do not argue that thought experiments or fictional cases should be excluded from philosophical methodology, just that real-world cases ought to be included. Material from the real-world should be available to philosophers as a legitimate part of their philosophical toolkits.

The main line of argument I pursue is as follows. Thought experiments (or ‘streamlined hypothetical cases’) are effective because they allow for potential ‘confounders’ to be removed. In the sciences, a confounder is any factor that interferes with the relationship between the cause under study and its purported effect.1 In philosophy, a confounder is something that interferes with one’s intuitions. I take ‘intuition’ here to mean what Kamm describes as a “judgement about a case”, where that judgement is reason-driven and not merely a gut-feel emotional response (Kamm 2009: 23).2 By ‘interfere with one’s intuitions’ I mean any factor that distracts one’s intuitions from the philosophical principle under consideration.

Excluding potential confounders is useful, but it comes with a risk. The risk is that in the process of streamlining out potential confounders, important ‘support factors’ might be unwittingly lost. A ‘support factor’ in the sciences is any factor that is required for the cause to achieve its effect. Oxygen is an important support factor in an experiment examining the causal relationship between matches and combustion (Cartwright & Hardie 2012: 62). In philosophy, the ‘support factors’ are those that need to be present for the key factor of interest to ‘do its work’; those factors that are required to keep one’s intuition on the target principle. For instance, in the standard Trolley Problem (which I will discuss in more detail below), a support factor might be that the respondent needs to imagine herself some distance away from the person that they are considering sacrificing. Evidence suggests that once distance is taken out of the scenario, and the respondent is required to imagine themselves physically closer to the person that they might sacrifice, they lose focus on the principle under consideration (whether it is permissible to sacrifice one to save five), and become distracted by how repulsive they find direct physical violence (Kahneman 2009: 79; Singer 2005). Distance, it turns out (if Kahneman and Singer are correct), is a support factor in the standard Trolley Problem.

I argue that consulting real-world cases helps us to check that important support factors have not been inadvertently removed in the process of streamlining. Given that fictional cases are extended thought experiments (a claim that I will defend later in this paper), cases from fiction will not be able to fulfil the function of checking that support factors have not been excluded, because they run the same risk as more austere thought experiments—support factors might unknowingly be removed by mistake. Real-world cases are the only remaining place to look in order to make sure that all the relevant support factors have been included. Real-world cases fulfil an important philosophical function and should be part of philosophical methodology.

A few caveats before proceeding. First, many of the examples in this paper are taken from moral and political philosophy because this is an area of philosophy in which thought experiments are used to test out hypotheses and principles, while in other areas of philosophy thought experiments are often focussed on examining conceptual entailments (Wilson 2016: 127).3 This is important, because in this paper I focus on cases (thought experiments, fictional cases, real-world cases) in their capacity as experiments, by which I mean, as tests for hypotheses and principles (Wilson 2014: 13). I ignore the other philosophical functions that cases play—such as illustrations at certain points in the exposition of an argument (Brown & Fehige 2014) or the educational role that they play in training new philosophers to be sensitive to identifying morally relevant factors (Nussbaum 1990), because these are not really cases of method. That said, even though most of my examples are from moral and political philosophy, I don’t take my argument to be restricted to these areas, and toward the end of this paper I will look at cases in epistemology and in philosophy of biology.

2. Starting with Thought Experiments

Williamson (2011: 215) argues that one of the salient features of contemporary analytic philosophy is the importance of thought experiments in its methodology. I take thought experiments to be the major methodological rival to real-world cases, and so we must understand them before drawing comparisons. That is the focus of this section.

Fischer describes thought experiments as “schematized hypothetical scenarios in which only a few details are filled in, and all the other details are left out” (1995: 4). Similarly, Wilson describes them as “toy ethical cases that are designed to simplify an ethical problem along a number of dimensions, thus making the problem more philosophically tractable” (2016: 128). Relatedly, Elgin (2014) argues that this streamlining process in philosophical thought experiments should be understood as analogous to scientists’ lab experiments, and she is echoing a relatively popular position in the literature (also see Wilkes 1993; Fischer 1995; and Wilson 2016). Elgin describes this commonality between philosophical thought experiments and scientific lab experiments as follows:

It is a controlled manipulation of events, designed and executed to make some particular phenomenon salient. … Important properties and relations are often masked by the welter of complexities that embed them. In experimenting, a scientist isolates a phenomenon from many of the forces that typically impinge on it. To the extent possible, she eliminates confounding factors. She holds most ineliminable factors fixed, effectively consigning them to the cognitive background of things to be taken for granted. This enables the effect of the experimental intervention on the remaining variables to stand out. Through such a strategy, she casts into bold relief factors that might be typically hidden from view. (2014: 222)

An illustrative example of streamlining out potential confounders in the sciences is that of a controlled trial (Elgin 2014: 222–223). Imagine that we want to test a new headache medication and we know that various factors impact on how quickly individuals recover from headaches, independently of whether or not they receive any treatment. Factors to consider include age, sex, weight, whether the patient is a smoker, how many hours the patient sleeps a night, etc. In an ideal test, we would want there to be two groups, an experimental group (the group that receives the treatment) and a control group (the group that does not receive the treatment), and the members of the two groups would be identical in terms of the things that are relevant to the effect—they would all be the same age, sex, weight, they would all be non-smokers (or smokers), and they would all sleep the same number of hours each night, etc. The only thing that should differ between the groups is whether or not they receive the treatment, and this allows for the causal relationship between the treatment and the effect to be isolated. At least, this is the case for a particular type of methodologist. The point is to streamline out confounders and to focus just on the factors that are relevant for the hypothesis being tested.

Thought experiments in philosophy are meant to do something similar. The intention is to factor out potential confounders. I take the standard Trolley Problem to be a classic example of a thought experiment in moral philosophy. In this case, a runaway trolley is hurtling down the tracks, where it will kill five people. As a passer-by, you happen upon a switch, which allows you to divert the trolley down a neighbouring track, where it will only kill one person. The philosophical question is: do you sacrifice the one in order to save the five (Foot 1967)? Importantly, obvious potential confounders have been factored out. For instance, things that might trigger implicit biases have been excluded—such as the race, age and gender of the individuals on each of the tracks, etc. The set of possible actions has also been significantly curtailed (Wilson 2014: 14)—you can only allow the trolley to continue or you can divert it; the five cannot escape, nor can you warn the one on the neighbouring track. Additionally, other potentially morally relevant factors that might distract one’s intuition have been excluded. By streamlining these extraneous factors out of the case, we are able to focus on just that which is relevant to the philosophical investigation at hand—whether it is permissible to sacrifice a few in order to save many.

For an even clearer analogue between controlled laboratory experiments and philosophical thought experiments, consider Rachels’s two Bathtub cases:

In the first, Smith stands to gain a large inheritance if anything should happen to his six-year old cousin. One evening while the child is taking his bath, Smith sneaks into the bathroom and drowns the child, and then arranges things so that it will look like an accident. In the second, Jones also stands to gain if anything should happen to his six-year-old cousin. Like Smith, Jones sneaks in planning to drown the child in his bath. However, just as he enters the bathroom Jones sees the child slip and hit his head, and fall face down in the water. Jones is delighted; he stands by, ready to push the child’s head back under if it is necessary, but it is not necessary. With only a little thrashing about, the child drowns all by himself, “accidentally,” as Jones watches and does nothing. (1975: 79)

The only thing that is meant to differ between the cases of Jones and Smith is that Smith kills the child while Jones merely allows the child to die—the intention being that the philosopher can focus exclusively on that distinction. This is analogous to the ideal test of the headache treatment, in that the only thing that is meant to differ between the test group and the control group is that the one receives the treatment and the other does not. Being able to streamline out potential confounders and focus just on that which is philosophically relevant is obviously useful.

3. Support Factors and Thought Experiments

Having described what is good about thought experiments, in this section I describe what I take to be the central methodological problem with thought experiments: that they run the risk of inadvertently streamlining out support factors when confounders are removed.

So far, I have argued that thought experiments play a useful role in philosophical methodology because they allow for extraneous factors that might otherwise interfere with philosophical intuitions to be streamlined out. However, in this process, required support factors may be inadvertently removed. To borrow (and somewhat adapt) an example from Fischer (1995) to illustrate this problem in the sciences, we can imagine a scientist who is very eager to study the relationship between matches and combustion. In their eagerness to be thorough, the scientist decides to perform the experiment in an airless vacuum to make sure that all potential confounders have been excluded. However, oxygen (an important support factor) has been inadvertently removed in the process, and so the scientist incorrectly concludes that there is no relationship between matches and combustion (Fischer 1995: 10).

The ‘Ticking Bomb’ case makes it clear that the problem of unwittingly excluding support factors is a real risk for philosophical thought experiments, and not just for scientific laboratory experiments. Waltzer (1973) first introduces the Ticking Bomb case in his discussion of the dirty hands problem. The reader is asked to imagine a scenario in which a terrorist has been captured. Authorities have good reason to suspect that the terrorist knows the location of a bomb (or a number of bombs) that will go off shortly. The question posed is whether it is permissible to torture the terrorist to get the information about the location of the bombs, and save the lives of innocent people who will die if the bombs are not located in time (Waltzer 1973: 173). The issue that the thought experiment highlights is that we typically think torture is unacceptable under all circumstances, but this is one case in which that intuition does not hold—overwhelmingly respondents say that it is permissible to torture the captured terrorist.

Bufacchi and Arrigo (2006) argue that our intuitive response to the Ticking Bomb case (that it is permissible to torture the terrorist to save the innocents) rests on relevant details having been stripped away for the purposes of creating the thought experiment. Their point is that once those details are reintroduced, so that the case more closely resembles actually torturing a captured terrorist, our response to the case are reversed and we no longer accept that torture is permissible (2006: 359).

To focus in on just one factor that Bufacchi and Arrigo (2006) highlight as being salient and excluded, the case assumes that the captured terrorist will provide accurate information about the location of the bombs. That information will then allow the relevant authorities to locate and defuse the bombs, thus saving the lives of innocents, which makes torture permissible. However, pre-existing evidence on coercive interrogation techniques show that torture leads to false confessions and inaccurate information being offered in the majority of cases. This is often because prisoners will say whatever they believe the interrogator wants to hear in order to put an end to the torture. Alternatively, savvy prisoners may intentionally give false information in order to mislead their captors and keep their plot intact. For example, the Japanese captured a US fighter pilot in August 1945, and after “rough interrogation” the pilot told his captors that the US intended to drop atomic bombs on Kyoto and Tokyo (when the truth was Hiroshima and Nagasaki), thus misleading the Japanese and ensuring that the plan went forward unimpeded. In yet another alternative, in scenarios in which members of organisations are likely to be captured and tortured, and they are aware that this is the case, those higher up within the organisation might intentionally give individuals in lower ranks incorrect information. The idea is that false testimony will then be offered to their enemies when they are predictably captured and tortured (Bufacchi & Arrigo 2006: 361–362). Once we recognise the high probability of false testimony in torture scenarios, it no longer seems that the bombs will be located and that any lives will be saved. Thus, it is no longer permissible to torture the prisoner. The excluded details, once reintroduced, reverse our moral judgment of the case.

Some might argue that this is an unfair reading of the Ticking Bomb case; that it should be read as examining conceptual possibilities, and that it is a secondary question whether we should ever torture anyone in practice. However, I think that Bufacchi and Arrigo’s discussion of the case shows us that the unreliability of the information acquired from torture is so pervasive that it should be part of what we understand the practice of torture to be. Thus, it doesn’t make sense to consider cases in which torture produces truthful testimony, regardless of whether we think it is permissible to torture people in practice or not.

4. In Search of an Appropriate Supplement

Given the problem of excluding potential support factors in the process of streamlining, something other than philosophical thought experiments is required. In this section, I assess whether fiction can provide a suitable alternative to the very austere thought experiments that are typically used in philosophy. In this section I argue that cases from fiction cannot fulfil this function, because they are subject to the same risks as thought experiments, and that, as such, real-world cases should be used. I conclude this section by noting that precedent exists in the sciences for using real-world cases as an alternative to experimental methods.

Elgin (2014) argues that cases from fiction ought to be included in philosophical methodology, because they provide more detail than very schematized thought experiments. Fictional cases are also more manageable than examples taken from the real world, due to the streamlining process that authors subject reality to in the production of fiction. Fictional cases thus seem like the perfect solution: more detailed, but still manageable (Elgin 2014: 232). The example that she uses to illustrate this point is that of Jane Austen’s novels. Her argument is that they provide substantially more detail than a standard philosophical thought experiment, but by focusing on only three or four families in a boring English village (Austen’s characters rarely venture into London where there is a lot going on, there are no peasant revolutions in her novels, etc.), Austen is able to remove extraneous factors and focus in on that which is relevant: the relationships. Elgin makes the point as follows:

Austen devises a tightly controlled thought experiment. Restricting the factors that impinge on her protagonists enables her to elaborate on the effects of those that remain…. Real families, however, are affected by too many forces for the social and moral trajectories exhibited by Austen’s characters to stand out. Too many other factors impinge on them; too many descriptions are available for characterizing their lives. Any sociological study would be vulnerable to the worry that unexamined factors played a non-negligible role in the interactions studied, that other forces were significant. (2014: 233, emphasis added)

As the italicized text makes clear, Elgin argues for the inclusion of fictional cases over real-world cases because she is concerned that important support factors might go unnoticed when studying cases taken from the real world. Her worry is that we will be so distracted by all of the other things going on in the case (real families might spend time in London, or find themselves in the midst of a peasant revolution, etc.) that we will fail to pick out that which is really important. However, as I have already argued in this paper, missing out on support factors is more of a problem for those making use of highly schematized thought experiments, where almost all factors have been intentionally excluded from the description, thus substantially increasing the likelihood that relevant information will have been excluded in the process. Further, given the streamlining process involved in fiction (which is precisely what Elgin thinks is good about these cases), fictional cases cannot provide a suitable alternative to thought experiments, because they also involve the risk of unwittingly streamlining out support factors.

Fiction also involves an additional risk, which is that the authors might not be streamlining the world in ways that are philosophically useful. The author’s aim in producing a fictional case is typically to create a compelling narrative, not to create a rigorous thought experiment. Some of the extraneous and distracting factors that are left behind by the author might make for good fiction, but might not be conducive to philosophical rigour.

The only suitable alternative to thought experiments is cases from the real world. Williams (2002) makes a similar suggestion when he argues that “real history fills in the merely schematic picture”. For instance, in the Ticking Bomb case, it was only by checking the thought experiment against the real world that it became clear that salient factors had been removed in the creation of the thought experiment.

A potential criticism that could be made at this point is that we do not need to turn to the real world to check that support factors have not been accidentally excluded, we can do this just by having more thought experiments, with more of the potentially relevant factors varied across the experimental cases. The way that Frances Kamm uses trolley problems is a good example of this—she assesses a vast number of trolley problems in which very subtle things are changed in each of the permutations, so that potentially relevant and irrelevant factors are screened in and out of the thought experiment. For instance, it was by looking at the Footbridge case in contrast to the standard Trolley Problem that it became clear that distance was potentially a support factor present in the standard version of the problem and absent from the Footbridge case. No consultation of the real world was required. However, it still seems like the risk of inadvertently excluding potential support factors persists, because what is and is not varied across the permutations of the thought experiments (even if there are many of them) will still be dependent on the imagination of the philosopher, and they might miss out on something crucial. To reiterate Williamson’s point from before, “Any humanly compiled list of such interfering factors is likely to be incomplete” (2007: 185).

It is also useful to note that, maintaining the analogy that has been used throughout this paper between thought experiments in philosophy and laboratory experiments in the sciences, precedent exists for using real-world cases (or natural experiments) as an alternative to experimental methods. One area where this is particularly clear is in the history of psychology. Historically, there was a strong tradition of only studying psychological responses in laboratory settings, in order to maintain rigour and to ensure that irrelevant factors were excluded. However, concern grew within the field that the phenomena of interest might not occur in laboratory conditions, particularly aspects of individuals’ social lives. For instance, how would one study psychological aspects of friendship in a laboratory setting? Many psychological phenomena that we are interested in cease to exist in the context of the laboratory. Studies of individuals in their everyday environments were thus included to capture the phenomena that had been previously streamlined out of experimental methods (Dechesne & De Roon 2014: 186–188). That is, real-world cases were included as a methodological alternative to streamlined experimental cases.

Given that schematized thought experiments and cases from fiction both run the risk of unwittingly excluding support factors when confounders are removed, real-world cases should be accepted as a legitimate part of philosophical methodology.

5. Some Examples

In this paper I remain methodologically permissive about what philosophers ought to do in their work. There are many philosophical problems, and many ways to approach them. As such, it would be foolish to be prescriptive about how philosophers should make use of the items in their toolkits—the tool you should use depends on the problem you have. In this section, I offer some ideas of what productive philosophical engagement with material from the real-world can look like by discussing some examples from philosophical practice. I will describe two types of philosophical engagement with real-world cases: 1) checking conceptual commitments of thought experiments; and 2) generating new philosophical concepts and theories. This list is intended to be indicative, not exhaustive.

5.1. Checking Thought Experiments

One way to use real-world cases is as a check on thought experiments. Consider Judith Jarvis Thomson’s (1971) Violinist Case, and its relationship to recent philosophical thinking about real cases of pregnancy. Thomson’s thought experiment asks that you imagine waking up to find that your circulatory system has been plugged into the circulatory system of an unconscious famous violinist. It turns out that the famous violinist is suffering from kidney disease. The Society of Music Lovers has checked all the available medical records and determined that you are the only person whose blood is a match to his, so they kidnapped you during the night and hooked you up to the violinist so that your kidneys can circulate his blood. If you stay attached to the violinist for nine months he will make a complete recovery and you can both go back to your regular lives. If you detach yourself at any point in that time period the violinist will die (Thomson 1971: 48–49). Thomson’s argument is that nobody would think that you are morally required to remain attached to the violinist for nine months (although that might be a very nice thing for you to do). Similarly, no woman should be morally required to act as a human life-support machine for a foetus for nine months (Thomson 1971: 49–50).

In the thought experiment, Thomson relies on what has become known as the ‘foetal container’ view of the relationship between the pregnant woman and her foetus (Kingma 2019: 615; Purdy 1990). That is, that a woman is an incubator in which the foetus resides for nine months. This assumption is what makes the thought experiment work—being pregnant, according to the thought experiment, is like having a stranger plugged into your circulatory system. However, recent work on the relationship between mother and foetus argues that given the various ways that the anatomy of the foetus and the mother are integrated—for instance the foetus “resides not in the uterine cavity, but is implanted in the uterine wall, within the maternal deciduous tissue and is, at least in its early stages, completely covered by it” (Kingma 2019: 624)—the ‘foetal container’ image on which Thomson’s thought experiment rests makes less sense.

In this example we see material from the world being used to check the conceptual underpinnings of a thought experiment. One might respond that this just amounts to cases from the world acting as counter-examples, which is already standard philosophical practice. This would be fair, if this were the only useful role that real-world material plays in philosophical practice. In the next sub-section, I consider an additional function played by cases from the real world—that of generating new theories and concepts.

5.2. Generating Concepts and Theories

In this sub-section I consider two recent pieces of philosophical work that involve considerable engagement with material from the real world, and which indicate that the role of real-world material extends substantially beyond counter-examples. These are Quassim Cassam’s Vices of the Mind (2019) and Sabina Leonelli’s Data-Centric Biology: A Philosophical Guide (2016).

Cassam’s Vices of the Mind is structured around a series of real-world high-profile media cases, ranging from the Suez Crisis to Brexit, which he uses to explore epistemic vices—that is, blameworthy impediments to knowledge acquisition. He sees this partly as professional public service—helping us to avoid epistemic vices, and their associated political and social harms (2019: vii)—a form of philosophical engagement that I have explicitly avoided in this paper. But he also uses the cases to generate new philosophical concepts. For instance, he uses the case of Boris Johnson’s involvement in the Brexit Campaign to develop the concept of ‘epistemic insouciance’; a callous disregard for the truth (Cassam 2019: 79). While it is possible that Cassam could have developed this concept through the use of thought experiments and fictional cases, he didn’t and his engagement with cases from the real world is philosophically productive.

Cassam shows us one way of engaging with material from the real world. He uses material generated by others, in this instance, news items from journalists. Leonelli offers us a very different picture of what philosophical engagement with the world can involve. In her case, she spent substantial time working with biologists and paying attention to their practices—the way she describes her own work is as “empirical philosophy of science” (2016: 6). In particular, she argues that traditional philosophy of science has focussed on theory and explanation as the central units of analysis, missing the important role of data; the way it is handled and the various epistemic issues and value judgments involved in its production. Through close work with biologists, she was able to identify this philosophical gap and develop a philosophy of science to accommodate the role of data. Data-Centric Biology won the 2018 Lakatos Prize, a prestigious award in Philosophy of Science, thus indicating the extremely productive philosophical role that working with material from the real-world can play.

This section has offered some indication of the roles that real-world material can play in philosophy: it can act as a check on thought experiments, and it can be used to generate concepts and theories. It also showed that there are various ways of engaging with the material from the real-world; this can range from checking the biology, through to reading the news, or spending time with practitioners.

6. Criticisms

One potential criticism against the position presented in this paper is that streamlining also occurs when describing real-world cases. That is, even though the case is taken from the world, the process by which it is described before philosophical analysis can occur involves streamlining. Not all of the details can be included in that description. The writer picks out that which they take to be salient, excluding much of what actually happened as they do so. In that selection process, important support factors might be excluded. Someone critical of my position might thus argue that real-world cases are just as susceptible to the kinds of problems I have argued are applicable to very streamlined hypothetical cases (like the Trolley Problems) and to cases from fiction (like Jane Austen’s novels).

My hypothetical critic makes a good point. A selection process does occur when real-world cases are described, and so there is a risk that philosophically relevant factors might be factored out in that process. But this is not just a problem that occurs when describing real world cases; it is likely that this would even be a problem for the philosopher experiencing a real-world case directly. For instance, we can imagine a philosopher being present for an interrogation in an actual ticking bomb case. Even when the philosopher is present in the room, there will be aspects of the experience that stand out as salient to them, and others that fade into the background of irrelevancy, and some of the factors that fade into the background might very well be philosophically relevant. This just seems to be a limitation to our ability to engage with the world—we are always streamlining out (hopefully) extraneous factors and doing so imperfectly, and so we are always at risk of inadvertently excluding support factors. The question, then, is whether this problem is more or less severe when describing real world cases for philosophical purposes than it is for streamlined hypothetical thought experiments and cases from fiction.

Part of the problem with the streamlining process that occurs in producing austere thought experiments and fictional cases (in contrast to the streamlining that occurs when describing real cases), is that the philosopher is entirely reliant on their imagination when attempting to figure out what is an extraneous confounder and what is a potential support factor. While we are still required to factor things out in our descriptions of real-world cases, and how this is done will be a matter of interpretation, at least the real-world places constraints on that interpretation, and we can test our interpretations against reality. These checks are absent in fictional cases and more austere thought experiments. As such, real-world cases are less likely than the alternatives to accidentally exclude important support factors. Thus, they still provide a valuable supplement to our philosophical methodology.

Another potential criticism may arise at this point, especially in response to my description in this section of the limited ways in which we encounter and describe the world. That is, it isn’t possible to empirically read the philosophy off the cases. I agree. But this isn’t a problem unique to philosophy. Most philosophers of science argue that is isn’t possible to empirically read scientific or social scientific theory off the data (that is, off material from the real world) (Douglas 2009; Longino 1990). Kuhn (1962/2012) argues that a substantial portion of scientific training involves teaching students to pick out what is salient in new cases by educating them in the “paradigm cases”. Martha Nussbaum (1990) makes a point about philosophy that is close to Kuhn’s point about science, when she argues that ethics education helps students to identify what is morally salient in new cases, and Iris Murdoch (1970) gestures toward something similar when she talks about developing ‘moral vision’. In philosophy, as well as in science and social science, it is a difficult task to look at the world and figure out what is important. It is also difficult to then generate theory and principles from that material. Being able to do this requires expertise and training.

7. Conclusion

In this paper I have argued that real-world cases should be included in philosophical methodology. I argue that thought experiments are effective because they streamline out extraneous factors and allow one to focus on just that which is philosophically relevant. However, the process of streamlining also runs the risk of overlooking important support factors. Cases from fiction are subject to the same risk. A methodological alternative to thought experiments and fictional cases is thus required, and this role is satisfied by real-world cases. Real-world cases face the problem that there may be too much going on to focus in on that which is philosophically relevant, and they are likely to be subject to confounders, but at least we can be sure that all of the requisite support factors are present.


Thanks to audiences at the Society for Applied Philosophy Annual Conference and the LSE Philosophy Department Cumberland Lodge Away Day for helpful feedback on earlier versions of this paper. Thanks also to Luc Bovens, Katie Steele and two anonymous reviewers for very helpful comments. Very special thanks to John Mee for extensive written comments and advice for how to improve this paper.


  1. I am following Elgin (2014) and Wilson’s (2016) use of terminology here. There is also the more common use of the term ‘confounder’ in the sciences which refers to a common cause that undelies a spurious correlation. This is not the intended use here. [^]
  2. This use of the term ‘intuition’ is different to how psychologists use it, by which they mean a response that is “automatic, quick, effortless, associative, and often emotionally charged … [and] not open to introspection” (Kahneman 2009: 72). The use in this paper tracks the standard use in moral philosophy. [^]
  3. Although, I concede that there are thought experiments in metaphysics and epistemology that also operate in this way. I am just less familiar with them, and so less able to discuss them in detail. [^]


Brown James R. and Fehige Yiftach (2014). Thought Experiments. In Zalta Edward N. (Ed.), Stanford Encyclopedia of Philosophy (Fall 2014 ed.). Retrieved from

Bufacchi Vittorio and Arrigo Jean M. (2006). Torture, Terrorism and the State: A Refutation of the Ticking-Bomb Argument. Journal of Applied Philosophy, 23(2), 355–373.

Cartwright Nancy and Hardie Jeremy (2012). Evidence-Based Policy: A Practical Guide to Doing it Better. Oxford University Press.

Cassam Quassim (2019). Vices of the Mind. Oxford University Press.

Dechesne Mark and de Roon Charlotte (2014). Perspectives on Social Knowledge. In Matheson Jonathan and Vitz Rico (Eds.), The Ethics of Belief (184–197). Oxford University Press.

Douglas Heather (2009). Science, Policy and the Value-Free Ideal. University of Pittsburgh Press.

Elgin Catherine Z. (2014). Fiction as Thought Experiment. Perspectives on Science, 22(2), 221–241.

Fischer John M. (1995). Stories. Midwest Studies in Philosophy, 20(1), 1–14.

Foot Phillipa (1967). The Problem of Abortion and the Doctrine of Double Effect. Oxford Review, 5, 5–15.

Jenkins Katharine (2014). ‘That’s Not Philosophy’: Feminism, Academia and the Double Bind. Journal of Gender Studies, 23(3), 262–274.

Jones Ward E. (2006). Philosophers, Their Context and Their Responsibilities. Metaphilosophy, 37(5), 623–645.

Kahneman Daniel (2009). Can We Trust Our Intuitions? In Voorhoeve Alex (Ed.), Conversations on Ethics (67–84). Oxford University Press.

Kamm Frances (2006). Intricate Ethics: Rights, Responsibilities, and Permissible Harm. Oxford University Press.

Kamm Frances (2009). In Search of the Deep Structure of Morality. In Voorhoeve Alex (Ed.), Conversations on Ethics (15–40). Oxford University Press.

Kingma Elselijn (2019). Were You a Part of Your Mother? Mind, 128(511), 1–44.

Kitcher Philip (2011). Philosophy Inside Out. Metaphilosophy. 42(3), 249–260.

Kuhn Thomas S. (2012). Structure of Scientific Revolutions (4th ed.). University of Chicago Press. (Original work published 1962)

Leonelli Sabina (2016). Data-Centric Biology: A Philosophical Study. Chicago University Press.

Longino Helen (1990). Science as Social Knowledge: Values and Objectivity in Scientific Inquiry. Princeton University Press.

Murdoch Iris (1970). The Sovereignty of Good. Routledge.

Nussbaum Martha C. (1990). Love’s Knowledge: Essays on Philosophy and Literature. Oxford University Press.

Purdy Laura M. (1990). Are Pregnant Women Fetal Containers? Bioethics, 4(4), 273–291.

Rachels James (1975). Active and Passive Euthanasia. New England Journal of Medicine, 292(2), 78–80.

Singer Peter (2005). Ethics and Intuitions. The Journal of Ethics, 9(3/4), 331–352.

Srinivasan Amia (2018). How to Do Things with Philosophy. European Journal of Philosophy, 26(4), 1410–1416.

Thomson Judith J. (1971). A Defense of Abortion. Philosophy and Public Affairs, 1(1), 47–66.

Thomson Judith J. (1985). The Trolley Problem. The Yale Journal, 94(6), 1395–1415.

Waltzer Michael (1973). Political Actions: The Problem of Dirty Hands. Philosophy and Public Affairs, 2(2), 160–180.

Wilkes Kathleen V. (1993). Real People: Personal Identity without Thought Experiments. Oxford University Press.

Williams Bernard (1981). Moral Luck. Cambridge University Press.

Williams Bernard (1983, November/December). The Uses of Philosophy. Donald McDonald (Interviewer). The Center Magazine, 40–49.

Williams Bernard (2002, October 17). Why Philosophy Needs History. London Review of Books, 7–9.

Williamson Timothy (2007). The Philosophy of Philosophy. Blackwell.

Williamson Timothy (2011). Philosophical Expertise and the Burden of Proof. Metaphilosophy, 42(3), 215–229.

Wilson James (2014). Embracing Complexity: Theory, Cases and the Future of Bioetics. Bioethics Review, 32(1/2), 3–21.

Wilson James (2016). Internal and External Validity in Thought Experiments. Proceedings of the Aristotelian Society, 116(2), 127–152.