Where Do We Go Next? — III

Toolbox[for a brief explanation of this ongoing series, as well as a full table of contents, go here]

The tools of the trade

We have already discussed at some length one important — if controversial — philosophical tool: the deployment of intuitions. I find it interesting that rarely critics and defenders of the use of intuition in philosophy bother to look at the broader literature on intuitions in cognitive science, which is actually significant and covers fields as diverse as chess playing, nursing and the teaching of math. I have discussed some of this literature elsewhere (Pigliucci 2012), but a quick recap may be useful in this specific context.

The very word comes from the Latin intuir, which aptly means “knowledge from within.” It is no longer considered a mysterious faculty of the human mind, as it has been the proper object of study in the cognitive sciences for a while (Hodgkinson et al. 2008). The first thing to realize is that there doesn’t seem to be any such thing as a generic ability of intuition. That is, when people say that they are “intuitive” they are likely fooling themselves, or they are mislabeling a more specific aptitude they have developed. Intuition, as it turns out, is domain specific: the more we acquire expertise and familiarity with a subject matter, the more we develop correct intuitions about that subject matter. Intuition — as generally understood — is simply a form of rapid (Kahneman 2011, following up on an initial suggestion by William James) subconscious processing of information, which is offered to our slow, conscious mental processing for further checking and refinement.

This is why the literature on expertise — and in particular philosophical expertise — becomes relevant (Ross 2006; for a philosophical take on it, see Selinger and Crease 2006). Expertise develops in distinct phases, almost regardless of the specific field, moving from novitiate to proficiency to actual mastery (in an academic context, think of the roughly analogous progression from undergraduate to graduate studies and then to the professional level). The latter may require about a decade to achieve, and results in complex webs of structured knowledge in one’s mind, a phenomenon that makes it possible for experts to quickly assess a situation (or an argument), see the pitfalls, and develop a solution (or a counter-argument).

Perhaps the best studies on expertise have been conducted on chess masters (e.g., Charness 1991; Gobet and Simon 1996), in part because the task is clearly delineated and in part because it is straightforward to measure one’s proficiency in that field. When a chess master is faced with a set situation (i.e., a chess problem) that developed organically via an actual game she will have little difficulty quickly arriving at the best moves. Interestingly, when asked why she deployed certain moves rather than others, the master will often reflect on the question and slowly reconstruct a logical answer: she did not consciously, explicitly, go through those steps, because she reacted intuitively. But she is nonetheless capable of providing a reasonable justification of what she did. Even more interestingly, it has been demonstrated that chess masters’ intuitions often fail whenever they are presented with a set situation that did not develop from an actual game, i.e., with an artificial scenario that could not possibly have contributed to their store of subconscious knowledge. Of course, I am not saying that there is a direct equivalency between these studies on intuition and what philosophers call by that name. But if philosophical intuition is — at the least in some sense — something akin to the more general phenomenon that goes by the same name (as, for instance, XPhi supporters often claim), then this sort of literature ought to be taken into account.

Perhaps it is instructive at this point to briefly go back to XPhi and the issue it raises about the use of intuitions in philosophy. After having drawn a parallel with the deployment of intuitions in other fields, such as arithmetic and geometry, Sosa (2009, 101) provides a working definition of intuition for the purposes of philosophical inquiry: “to intuit that p is to be attracted to assent simply through entertaining that representational content. The intuition is rational if and only if it derives from competence, and the content is explicitly or implicitly modal (i.e., attributes necessity or possibility).” Please notice the emphasis on competence, that is, on expertise.

In response to XPhi papers showing the effect of cultural differences and/or “framing” on people’s intuitions about philosophical concepts such as moral desert, free will (but, apparently, not knowledge! See Machery et al. 2015), etc., a few observations may now be made in addition to our previous discussion. First, such differences may result from the lay subjects not having thought about those topics much (unlike, presumably, professional philosophers) — i.e., they are not experts on the subject matter at hand. Second, it is not really surprising that different background conditions, including individuals’ assumptions about the cases being presented to them, will cause variation in the responses (a point also made by Sosa); research on intuition has shown, as we have seen in the case of chess masters, that different conditions legitimately elicit different intuitions, and that even experts can be laid astray when faced with highly artificial scenarios. But of course that is precisely what careful philosophical unpacking of concepts is supposed to explore and deal with.

Moreover, according to Goldman (2007) the use of “intuition” as a term to describe how philosophers use hypothetical examples in their reasonings is actually of fairly recent vintage, tracing to Chomsky’s methodological discussions in linguistics. Goldman has also spent a significant amount of effort thinking about intuitions in philosophy. He examines a number of possible “targets” of philosophical intuitions: Platonic forms, natural kinds, Fregean concepts, concepts in the psychological sense, and “shared” concepts. I find his discussion of Fregean concepts particularly illuminating. He defines these as “abstract entities of some sort, graspable by multiple individuals. These entities are thought of as capable of becoming objects of a faculty of intuition, rational intuition.” He cites Bealer (1998) as explicating what it means to grasp a concept by rational intuition: “[W]hen we have a rational intuition — say, that if P then not not P — it presents itself as necessary; it does not seem to us that things could be otherwise; it must be that if P then not not P.”

Goldman himself, however, prefers to talk about the last two types of targets, psychological and shared concepts, as particularly relevant to philosophical discourse. A psychological concept is a mental representation by a particular individual, and that individual will possess certain intuitions about the correct and incorrect usage of that concept relatively to how she understands the concept itself. This becomes useful once Goldman generalizes from psychological to shared concepts, which originate when there is substantial agreement within a community of individuals on the (correct and incorrect) usage of a given psychological concept. Within that community, people may then decide that some individuals — by virtue of their specific training — are better, more reliable, at deploying a certain concept. These individuals are acknowledged as experts in the usage of that concept, and their intuitions about the concept become more valuable than other people’s intuitions. You see where this is going: if the concept is, say, free will, then the community of experts is made of philosophers who have thought hard and long about free will (which excludes not just laypeople, but also philosophers who do not have expertise in philosophy of mind and metaphysics). [18]

What I have proposed so far, then, is to recast the debate about philosophical intuitions within the more general assessment of expert intuitions, about which there is a significant cognitive science literature. It also makes sense to survey philosophers’ take on what intuitions are and how they are deployed within their profession, which is precisely what Kuntz & Kuntz (2011) have done. They point out a crucial distinction that is often lost in discussions about intuition: that philosophers actually use their intuitions for two different purposes, in a “discovery” and in a “justification” mode — only the latter being typically addressed by critics of philosophical intuitions (but see my earlier discussion of Love 2013, and his framing in terms of complementary “images” of science).

Kuntz & Kuntz conducted an online survey of 282 philosophers, focusing on what they think about intuitions in their profession. To begin with, about 51% of participants said that intuitions are useful for justificatory purposes in philosophical analysis, while a whopping 83% said that are useful for exploratory analysis. Moreover, about 70% of respondents said intuitions are not necessary for justification. These statistics paint a picture that is somewhat at odds with the common criticism of the “ubiquitousness” of intuitions as “data” in philosophical discourse. The same authors provided their subjects with seven different accounts of intuitions, and it is instructive to see that two of these were given the highest rank by a majority of respondents: “Judgment that is not made on the basis of some kind of observable and explicit reasoning process” and “An intellectual happening whereby it seems that something is the case without arising from reasoning, or sensorial perceiving, or remembering.” Another of the accounts on offer received by far the lowest ranking: “The formation of a belief by unclouded mental attention to its contents, in a way that is so easy and yielding a belief that is so definite as to leave no room for doubt regarding its veracity.” [19] The first two accounts (but not really the latter) are compatible with cognitive scientists’ definition of intuition and the target of their empirical research. So, intuitions — re-conceptualized as they normally are in the cognitive science literature — remain an important tool for the philosopher, a tool that is characterized by the same pros and cons as intuitions in any other field of inquiry or, indeed, in everyday life. Crucially, it seems that a majority of philosophers uses intuitions just the way they are supposed: in an exploratory rather than justificatory fashion. While in science the justification is anchored by empirical evidence, in philosophy it is the result of “unpacking,” i.e., carefully and explicitly analyzing the initial intuition (moving from Kahneman’s system I to his system II, if you will).

Intuition, of course, is not the only tool available to the professional philosopher. Others include the method of analysis, counterfactual thinking, reflective equilibrium, and thought experiments. These are all actually related to each other (and to intuitions!), so a linear discussion of each in turn is by necessity a bit artificial. Nonetheless, I think it will be useful to complete my analysis of what philosophical inquiry consists of and how it is conducted.

Beany (2009) provides a convenient overview of the so-called method of analysis, defining it as “a process of isolating or working back to what is more fundamental by means of which something, initially taken as given, can be explained or reconstructed,” and in that sense its applicability clearly goes well beyond what nowadays is referred to as “analytic” philosophy (Chapter 2). Socrates, for one, was certainly doing analysis in Beaney’s sense. As the author correctly points out, it is misconceived to think of, say, Wittgenstein’s criticism of logical atomism, or of Quine’s rejection of the analytic-synthetic distinction as blows to the method of analysis in philosophy, since such criticisms were aimed at very narrow conceptions of that method. Indeed, Beaney identifies three different components of philosophical analysis, all of which are likely applied in combination in the course of actual philosophical practice: decompositional (aiming at unpacking the components of a concept and analyzing them individually), regressive (working back to first principles), and interpretive (translating a concept into a logically more rigorous form). Much of this goes back to the Greeks, and in fact Beaney traces it to the early influence of geometry, which made a crucial impression on thinkers from Plato on, though the development of what we call Euclidean geometry is actually a result of this, not a cause (Euclid’s Elements date from circa 300 BCE, after Plato and Aristotle).

Beaney remarks that regressive analysis was the dominant form of the method in ancient Greece, and that we had to wait until the medieval period to see the development of interpretive analysis. We then see all three approaches deployed in Buridan’s Summulae de Dialectica (1330-40; see Zupko 2003, 2014). Even so, the decompositional approach to analysis obtained its most famous formulation with Descartes, in Rules for the Direction of the Mind (1684), where he says: “If we perfectly understand a problem we must abstract it from every superfluous conception, reduce it to its simplest terms and, by means of an enumeration, divide it up into the smallest possible parts” (Rule 13). As Beaney points out, it is not by chance that Descartes admitted to be influenced by geometry, to which he of course made the novel contributions that made him justly famous: “Those long chains composed of very simple and easy reasonings, which geometers customarily use to arrive at their most difficult demonstrations, had given me occasion to suppose that all the things which can fall under human knowledge are interconnected in the same way” (Discourse on Method, 1637 /2000). From there, decompositional analysis continued its good run well into early modern philosophy with Kant.

By the 20th century, according to Beaney, both so-called analytical and continental philosophy (Chapter 2) had gone beyond decompositional analysis, with the continentals’ phenomenological approach being analogous to conceptual clarification, while Hegel can be thought of as employing regression. We have to remember that analytic philosophy in the strict sense is a new beast that originated with Frege, Russell and others, and which depends on logical analysis as made possible by contemporary logic (especially predicate logic), with the term “analytic” once again reminding us closely of geometry, more so than previous uses of the decompositional approach.

Let me now turn briefly to the use of counterfactual thinking. In his Presidential address to the Aristotelian Society in 2004, Timothy Williamson (2005) pointed out that so-called “armchair philosophizing” is chronically seen as a virtue by what he labeled “crude rationalists” and, symmetrically, a vice by what he referred to as “crude empiricists.”As Aristotle himself would have readily agreed, wisdom must lie somewhere in between. Williamson makes exceedingly clear what the problem is when he says, citing the example of analytic metaphysicians (but, really, it could be any branch of philosophy): “[they] want to understand the nature of time itself, not just our concept of time, or what we mean by the word ‘time’” (p. 2), which means that we must pay attention to the empirical, although the idea of philosophizing about it is precisely that the empirical by itself isn’t going to provide us with a satisfactory answer either.

Williamson presents a detailed discussion of Gettier-type cases in the epistemology of truth (see Chapter 6) as instances of the usefulness of counterfactual thinking, which eventually brings him to the observation that “examples [used to explore our intuitions] are almost never described in complete detail; a mass of background must be taken for granted; it cannot all be explicitly stipulated” (p. 6). He then suggests — rightly, I think — that the deployment of counterfactuals is not distinctive of philosophy: both everyday and scientific reasoning make use of them all the time. For instance, if we say “there are eight planets in the solar system” we are implicitly assuming the counterfactual that if there were more planets in our neighborhood we would have discovered them by now. The fact that we do not have discovered additional planets does not logically imply that there are none, so the counterfactual conditional plays the role of allowing us what we understand to be a provisional conclusion, revisable at any time in light of new evidence. The very same role is played by counterfactuals in philosophical reasoning: they imply a “as far as we can tell given what we know” condition. Which leads Williamson to argue that intuitions and counterfactuals in philosophy, along the lines of those famously deployed in discussions of Gettier cases, are examples of human judgment no different from the judgment we reach in other applications, it is the specific subject matter, not the method, that is philosophic: “we have no good reason to expect that the evaluation of ‘philosophical’ counterfactuals … uses radically different cognitive capacities from the evaluation of ordinary ‘unphilosophical’ counterfactuals. We can evaluate [these counterfactuals] without leaving the armchair; we can also evaluate many ‘unphilosophical’ counterfactuals without leaving the armchair” (p. 13), which ought to take at the least some of the bite out of the standard criticisms of armchair philosophizing. To reiterate: philosophical intuition is not a special cognitive ability, and does not, therefore, demand special defense or scrutiny.

Indeed, Williamson’s conclusion so nicely dovetails with my main thesis in this book that it is worth (to me) to quote him again in some detail (p. 21): “Both crude rationalism and crude empiricism distort the epistemology of philosophy by treating it as far more distinctive than it really is. They forget how many things can be done in an armchair, including significant parts of natural science … That is not to say that philosophy is a natural science, for it also has much in common with mathematics.” Exactly.

I turn next to another staple of the philosopher’s toolbox: reflective equilibrium. Although the term has reached wide popularity in philosophy as deployed by John Rawls (1971) in his A Theory of Justice, reflective equilibrium is a general feature or method of philosophical reasoning, formalized in recent times by Nelson Goodman in his Fact, Fiction and Forecast (1955) in the context of inductive logic (though Nelson didn’t use the specific phrase “reflective equilibrium”). Daniels (2003) defines it fairly clearly in this fashion: “The method of reflective equilibrium consists in working back and forth among our considered judgments (some say our ‘intuitions’ [20]) about particular instances or cases, the principles or rules that we believe govern them, and the theoretical considerations that we believe bear on accepting these considered judgments, principles, or rules, revising any of these elements wherever necessary in order to achieve an acceptable coherence among them.” Notice Daniels’ emphasis on a coherentist approach to epistemology, as opposed to a foundationalist one. As we discussed especially in Chapter 3, in the context of Quine’s “web of belief,” coherentism seems a much more sensible way of approaching knowledge and judgment, provided that at the least some of the elements of our epistemological web are empirical facts, for the very compelling reason that it is empirical facts that allow us to move from conceptual space (where often there are a number of equally logically coherent scenarios or approaches to a problem) to the world as it actually is (where I still presume most of us will agree that things are either one way or another, but not both).

I see reflective equilibrium in conceptual space as in a way analogous to inference to the best explanation in empirical space. Neither is a perfect approach, nor does it provide any guarantee of success, but they are very sensible tools for navigating both spaces. Inference to the best explanation, for instance, suffers whenever one has not conceived of better alternative scenarios, in which case one is stuck with a “best of a bad lot” situation; it also suffers whenever insufficient or low quality data is all that is available. Analogously, reflective equilibrium doesn’t work very well if one fails to consider better (i.e., more coherent with the available facts and assumed notions) scenarios, or if one’s knowledge of the relevant elements is insufficient or faulty. Nevertheless, these are the sort of problems that negatively affect (and impose limits upon) any kind of human reasoning, about either matters of fact or relations of ideas (or anything else), as Hume would put it.

Daniels (2003) helpfully distinguishes between a narrow and a wide form of reflective equilibrium, though I think it is better to treat these as two points of reference along what is essentially an epistemic continuum (analogous, I suspect, to the difference between Duhem’s and Quine’s theses — respectively narrow and wide — as discussed in Chapter 2). Daniels’ example of narrow reflective equilibrium is the case of rationing of medical care according to the age of the patient. At first glance, one might think that this is similar to rationing by, say, sex, or ethnic background, which would, presumably, be unethical. However, further reflection shows that age is a different biological phenomenon from the other two (for instance, because we all age, but people can’t change sex without medical assistance, and simply don’t change ethnic background). Rationing care by age, therefore, does not have to be discriminatory (and it may thus be morally acceptable), and it could very well turn out to be a highly sensible practice in terms of both efficacy and cost. This qualifies as a narrow type of reflective equilibrium because a large number of background assumptions and a lot of factual knowledge have been left unchallenged in order to focus on a fairly specific debate.

To move from a narrow to a wide exercise in reflective equilibrium one can contemplate the famous example of Rawls’ questioning of the very tenets of utilitarian ethics, which may have been treated instead as a background assumption in the previous instance. Wide reflective equilibrium, in other words, brings under scrutiny some of the broader axioms of our thinking. Just as in the case of the difference between (narrow) Duhem’s and (broad) Quine’s approaches in philosophy of science and epistemology, one needs to keep in mind that most actual applications will be on the narrow side of things, as only occasionally it pays off to broaden the circle of questioning that far.

Naturally, there are a number of standard criticisms to the practice of reflective equilibrium, most of which, I think, seem to miss the point. For instance, it is argued that reflective equilibrium depends on judgments (say, in ethics) that are founded on certain “intuitions” which are in themselves questionable or can be challenged. This is certainly true — with the caveats about philosophical “intuition” mentioned above — but that very same criticism applies to pretty much any human judgment, in both conceptual and empirical matters. Yes, any given judgment can be challenged, and if so then it needs to be defended, unpacked, argued for, etc.. In other words: welcome to philosophy!

A second common criticism of the method of reflective equilibrium is a special instance of the general problem with coherentist views of truth: as Hare (1973) wrote in response to Rawls, fictional scenarios (say, in a novel) can also be made coherent, but one is hardly thereby justified in accepting them as truthful. But presumably philosophers will always be concerned with how things stand in the real world, i.e. their judgments will be, as much as possible, informed by — and anchored to — our best understanding of empirical facts. Although such facts themselves are revisable, theory-dependent, etc., it is the full web of beliefs that makes reflective equilibrium a dynamic practice, whose particular status at any given moment can always be reassessed if new facts and/or arguments come to light. Compare that with the static description of a fictional scenario in a novel and the difference ought to be obvious.

A theoretically more serious, but also much less impactful in practical terms, concern is that it is actually surprisingly difficult to provide a philosophically satisfying account of the very concept of coherence. What is required for reflective equilibrium is stronger than the simple lack of straightforward logical contradictions, and is more akin to a Bayesian-type judgment as deployed in scientific inferences to the best explanation. I think, however, that the project of providing a good account of coherence can proceed of its own accord without philosophers having to await for a final outcome in order to practice reflective equilibrium, as long as they are very clear about why they think a certain set of beliefs and empirical facts is more “coherent” than another, and are willing to defend such judgment when challenged.

Another classic tool available to philosophers (and to scientists) is the thought experiment. Examples in the sciences abound, as is well known, from Newton’s bucket to Schrödinger’s cat. In fact, the very term (“Gedankenexperiment” in German) appeared to have been introduced by 19th century physicist Ernst Mach, though the approach is much older than that. As J.R. Brown (2002) points out, one of the most appealing (and, as it turns out, wrong) thought experiments was advanced by Lucretius in De Rerum Natura, where he attempted to show that space is infinite by conjuring a hypothetical scenario in which we are able to shoot a spear through the universe’s boundary. The logic is solid: if the spear goes through the boundary, then it is not really a boundary; if it bounces back then there is an obstacle that must itself be located in a portion of space beyond the alleged boundary. But Lucretius didn’t know about the possibility of spaces that are both unbounded and yet finite (Einstein 1920). The fact that the experiment eventually failed is no argument against the use of thought experiments in general: first, because other such experiments succeeded; second, because plenty of empirical experimental results are also eventually overturned by further discoveries; and third, because we still learn something about how to think about the specific matter in question (in this case, space and infinity) by contemplating why exactly the outcome of the experiment (thought or empirical) had ultimately to be rejected.

Just like the case of intuitions, however, one may reasonably ask on what grounds we rely on thought experiments, i.e., what, exactly, are the epistemological foundations of this approach to theorizing? There is a large literature on this topic, and a good recent summary of the major positions has helpfully been provided by Clatterbuck (2013). There are at least three accounts that try to make sense of how thought experiments work: they may represent an example of Platonic knowledge; they may be a pictorial form of standard inductive arguments; or they may represent a special type of induction.

Starting with the suggestion that thought experiments give us access to a Platonic realm of ideas, so that we can somehow gain a priori knowledge about the physical world in just the way we gain mathematical knowledge, the argument in defense of this position is put forward by J.R. Brown (1991), and Clatterbuck reconstructs it this way:

Premise 1. Mathematical Platonism is true.
Premise 2. Mathematical Platonism entails that numbers are outside of space or time.
Premise 3. Mathematical Platonism entails that we can have intuitive knowledge of 
Premise 4. Realism about laws is true.
Premise 5. Realism about laws entails that laws of nature are outside of space and time.
Premise 6. If we can have intuitive knowledge of numbers which are outside of space and 
 time, then we can also have intuitive knowledge of laws which are outside of 
 space and time.

Conclusion: We can have intuitive knowledge of the laws of nature.

To unpack: Brown assumes two controversial positions in metaphysics, namely mathematical Platonism (see my skepticism about it in the Introduction) and realism about laws of nature (see Cartwright’s and others’ skepticism about that, in Chapter 4). From these two premises, and from what they logically entail, he arrives at the conclusion that we can obtain intuitive knowledge of the laws of nature, and suggests that thought experiments are one way to obtain such intuitive knowledge (there goes “intuition,” again!). The argument is valid, but it is an altogether different issue to establish the soundness of its premises.

Most scientists, I wager, would have no problem with P4, and therefore must accept also P5. P2 and P3 are tightly connected with P1, so the latter may be the obvious target of criticism. But, as we have already seen, a good number of mathematicians and philosophers of mathematics do not find mathematical Platonism to be a bizarre idea at all. Still, even if we provisionally accept mathematical Platonism, for the sake of argument, Clatterbuck suggests that the real stumbling block is actually P6: it embodies an argument from analogy, whereby mathematical truths are taken to be of the same kind as physical laws. This is metaphysically questionable, since mathematical truths are necessarily such, while physical laws appear to be only contingently true (i.e., in no possible world a given mathematical truth becomes untrue, while there are many possible worlds where our natural laws do not hold). Another way to put the point is that we routinely arrive at mathematical (and logical) truth simply by thinking about a mathematical (or logical) problem, but we usually need observations and/or experiments to gain any knowledge of the physical world — and this is because the latter is a contingent subset of a number of logically possible worlds, a subset that can only be pinpointed empirically.

A second possibility is that thought experiments are nothing but standard arguments accompanied by pictures, a position that has been defended, for instance, by Norton (2004). According to him, what Galileo, Einstein and others have been doing in this respect is simply to present an inductive or deductive argument, peppered with pictures to make it more vivid to the intended audience. The point being that the mental “pictures” (e.g., Galileo’s two bodies of different weight falling separately or linked to each other; Einstein’s beams of light) are not necessary for the argument to go through, they are just embellishments with rhetorical but no epistemic force. Essentially, Norton is being parsimonious here: if we assume that thought experiments are just standard arguments plus pictures (where the latter play no additional epistemic role), then we don’t need to invoke Platonism.

Clatterbuck, however, thinks that Norton is being too parsimonious, missing an important part of the action. Imagine a concrete problem in physics, say, the dynamics of the fall of two bodies of different weight, either linked to each other or not; we can begin by abstracting away all irrelevant or distracting factors, such as air friction that interferes with the fall of the bodies; we arrive at certain conclusions concerning the problem, in this case that Aristotelian physics applied to the hypothetical experiment leads to a logical contradiction, as Galileo did; finally, we generalize our findings to the real (i.e., not idealized) world. What we have as a result of this procedure, of course, is a thought experiment. But notice that the idealization of the circumstances played a crucial role in the construction of the experiment. That is, according to Clatterbuck, the “picture” part of the inductive reasoning we just deployed is not simply a pretty but ultimately dispensable accessory, it is a crucial aspect of what’s going on. The result is something different from, say, enumerative induction, since we don’t have to “observe” more than one case to infer our conclusions: the (idealized) case is sufficient by itself to yield a logically valid inference. The upshot is that thought experiments embody a type of inductive reasoning, but that such reasoning requires idealization or abstraction to yield the conclusion from a single instance to a generally valid class of cases. In a sense — when done well — thought experiments are more powerful than standard enumerative induction, which after all is based on always fallible collections of observations. There is much more to be said about thought experiments, including interesting discussions about what, exactly (if any!), is the difference between a thought experiment in science and in philosophy. Nonetheless, thought experiments have clearly yielded many fecund lines of inquiry, and will certainly remain an essential part of the philosopher’s toolbox.


[18] Goldman does add the caveat that many philosophical discussions are about folk-ontological concepts, not technical terms. I submit, however, that philosophical discussions of this type begin with folk-ontological concepts, but then mold them in a way that transforms them into technical terms. It is at that point that philosophical expertise becomes paramount and that laypeople’s understanding of the newly molded concept becomes pretty much irrelevant to philosophical inquiry. Unfortunately, the frequent lack of appreciation that philosophers and laypeople may be using the same words but meaning different things ends up generating a lot of unnecessary confusion.

[19] As a philosopher of science, I also found interesting that colleagues in my field turned out to think that the role of intuition in justification in the course of their practice is far less important than philosophers who work on epistemology, ethics, metaphysics and philosophy of mind.

[20] As one can see, it is really difficult to get away from the i-word in philosophy, no matter how hard one tries!


Bealer, G. (1998) Intuition and the Autonomy of Philosophy. In: M. DePaul and W. Ramsey (eds.), Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical Inquiry. Rowman & Littlefield, Lanham, pp. 201–240.

Beany, M. (2009) Analysis. Stanford Encyclopedia of Philosophy (accessed on 28 August 2014).

Brown, J.R. (1991). The laboratory of the mind: Thought experiments in the natural sciences. Routledge.

Brown, J.R. (2002) Thought experiments. Stanford Encyclopedia of Philosophy (accessed on 11 September 2014).

Charness, N. (1991) Expertise in chess: the balance between knowledge and search. In: K. Anders Ericsson and J. Smith (eds.) Toward a General Theory of Expertise: Prospects and Limits. Cambridge University Press.

Clatterbuck, H. (2013) The epistemology of thought experiments: A non-eliminativist, non-Platonic account. European Journal for Philosophy of Science 3:309-329.

Daniels, N. (2003) Reflective equilibrium. Stanford Encyclopedia of Philosophy (accessed on 11 September 2014).

Descartes, R. (1637 / 2000) Discourse on Method and Related Writings. Penguin Classics.

Descartes, R. (1684) Rules for the Direction of the Mind (accessed on 28 August 2014).

Einstein, A. (1920) Relativity: The Special and General Theory. Chapter XXXI (accessed on 22 November 2015).

Gobet, F. and Simon, H.A. (1996) The role of recognition processes and look-ahead search in time-constrained expert problem solving: evidence from Grand-Master-level chess. Psychological Science 7:52-55.

Goldman, A. (2007) Philosophical intuitions: their target, their source, and their epistemic status. Grazer Philosophische Studien 74:1-26.

Goodman, N. (1955) Fact, Fiction, and Forecast. Harvard University Press.

Hare, R.M. (1973) Rawls’ Theory of Justice. Philosophical Quarterly 23:144-155; 241-251.

Hodgkinson, G.P., Langan-Fox, J. and Sadler-Smith, E. (2008) Intuition: a fundamental bridging construct in the behavioural sciences. British Journal of Psychology 99:1-27.

Kahneman, D. (2011) Thinking, Fast and Slow. Farrar, Straus and Giroux.

Kuntz, J.R. and Kuntz, J.R.C. (2011) Surveying Philosophers About Philosophical Intuition. Review of Philosophy and Psychology 2:643-665.

Love, A. (2013) Experiments, Intuitions and Images of Philosophy and Science. Analysis Reviews 73:785-797.

Machery, E., S. Stich, D. Rose, et al. (2015) Gettier across cultures. Noûs, online 13 August 2015, DOI: 10.1111/nous.12110.

Norton, J.D. (2004) On thought experiments: Is there more to the argument? Philosophy of Science 71:1139–1151.

Pigliucci, M. (2012) Answers for Aristotle: How Science and Philosophy Can Lead Us to A More Meaningful Life. Basic Books.

Rawls, J. (1971) A Theory of Justice. Belknap Press.

Ross, P.E. (2006) The expert mind. Scientific American, July 24.

Selinger, E. and Crease, R.P. (eds.) (2006) The Philosophy of Expertise. Columbia University Press.

Sosa, E. (2009) A defense of the use of intuitions in philosophy. In: M. Bishop & D. Murphy (eds.) Stich and His Critics. Blackwell.

Williamson, T. (2005) Armchair philosophy, metaphysical modality and counterfactual thinking. Proceedings of the Aristotelian Society 105:1-23.

Zupko, J. (2003) John Buridan: Portrait of a Fourteenth-Century Arts Master, Publications in Medieval Studies, University of Notre Dame Press.

Zupko, J. (2014) John Buridan. Stanford Encyclopedia of Philosophy (accessed on 28 August 2014).


Categories: Nature of Philosophy

59 replies

  1. If intuition, the application of experience, is a valuable tool for philosophy, how valuable is insight; The sudden connections and paradigm shifts which open up new vistas unexpectedly?

    Or is it not methodical enough?


  2. First, Massimo, I xerox your first comment on this thread. Even before I read it I had strong skeptical inclinations against the “is there data” gambit as you excellently call it.

    Second, this is one of the strongest parts of your mss so far.

    Third, I think there are different degrees of thought experimentation, and dependent on different dimensions of analysis. One dimension is what I’d call originality of thought. When Einstein at 16 first conceived of surfing a light wave and thought about what he’d see–a static electromagnetic field–that was a very original way of thinking about observers and physical circumstances. It was not well-constrained in terms of data and the like, but led to him to realize that c might be a constant for all observers. So originality of the thought experiment was primary. But then pose against that his EPR thought-experiment criticism of quantum explanation. While Einstein thought his criticism was sufficiently well-constrained to make its point, the various de- and re-constructions of it over the years shows that it is more a basic story that needs more detailed empirical and associated theoretical narrative for drawing conclusions of any worth. So original as it was, it needed (and needs) stretching by other narrative factors to be of use.

    What I tell my students is that the best thought experiment is a logically possible narrative that, in one telling, stretches a concept in one dimension to its breaking points to see the limits of that concept. But then, the experiment should allow a retelling that, with specified adjustments for it, stretches that same concept along another dimension of inquiry to see another breaking point. If the experiment proceeds along several such dimensions, the concept may become clearer in abstract outline. The trolley problem is what I have in mind as a classic example here–though actually I think it is a bad example in that it is horribly underdetermined as a moral narrative.

    Liked by 1 person

  3. The “badness” in this case would be the actual decision to drink and drive, I think. Both virtue and ethics and deontology would find that prima facie morally wrong.


  4. Back to last night’s aesthetics topic ….

    Sun-dried tomato alfredo (Classico jar), supplemented with fresh Asiago cheese, grated Parmesan and butter (all melted together, then added to the alfredo in the main pot), fresh ground black pepper, and basil, served over the usual whole wheat pasta, plus green beans and broccoli, and artichokes, plus diced pork loin and a touch of diced pepperoni.

    I use just enough on sauces to coat, not drown. All quite healthy, with the fiber, and no added salt. There’s enough in the pasta sauce jar.

    Eat your heart out …

    Liked by 1 person

  5. Socratic,

    Thank you for clarifying. Yes I think that would hold true in deontology and virtue ethics. And does the combination of virtue ethics, deontology, and consequentialism cover the domain of ethics/morality in your opinion?

    Meaning, if an act with seemingly ethical consequences has 4 dimensions and those dimensions all fall into one of those three general ethical systems, does that mean those three ethical systems are the only ethical systems relevant to moral behaviour? Or that they comprise all the ways of looking at moral acts?

    This isn’t a challenge on that front, I’m just trying to understand your idea completely.


  6. Synred,

    Yes it’s in that video but you were right, it’s not the mass of the new particles that would cause the weak interaction simply that we know within QFT any new wave function found would be to weakly interacting to alter your brain state.

    You can skip right to 14:00 min mark in the video for his summation. I assume you don’t need the introduction.


  7. Hi synred,

    “Does Carrol really say that? Ever hear of QM? And there is dark matter which does interact with us albeit not much.”

    Of course Carroll is an Everettian (Rupert, not Kenny) and if EQM was the case then that really would close the door on “could have done otherwise”

    “However, I don’t see how a bit of randomness helps you with ‘free will’. You still can’t ‘will what you will’ even if you can flip a qbit-coin.”

    Our minds are not single bits and the fact that a process contains random elements does not necessarily entail thar the process itself is random.


  8. Hi Robin,

    In what way could the process not be random? This is where I can’t get to logically. Can you give an account of what a non-random process would entail that would allow for “could have done otherwise?”


  9. Hi garth,

    “Carrol’s claim is that with the Higgs field confirmed, we now have a closed system in which we will find no new particles, or at the very least, no new particles with enough mass to interact with and alter our brain states, which closes the door on “could have done otherwise.””

    I can’t see why LFW would need new particles. Carroll is an Everettian, hence he believes we could not have done otherwise. But we don’t know if that is true or not and so we do not yet have enough information for the first part of the equation.

    Then you come to agency. Does physics say that for a process that I am ostensibly doing that I do not really do it?


  10. Hi garth,

    “In what way could the process not be random? This is where I can’t get to logically.”

    I have a program and it runs using a pseudo random number generator, but it would run the same way if I could plug in a true random number generator.

    Basically the program starts with a random string of bits and runs that string as a program on a runtime module. The ones that create something closest to a rectangle are retained and the others dumped. Then some of those programs are mutated by altering a single bit, or adding a new bit or transcribing a random portion of the bit string. The end of the process will be a computer program that draws a rectangle.

    But it is not deterministic because there are many ways for a computer program to draw a rectangle and this will not find one in particular.

    But it is not random because a random process would take more time than there is in the Universe to write that program with random bits.

    Hence I have a process which (if I used a true random number generator for it) would be neither deterministic nor random.

    Liked by 1 person

  11. Hi garth,

    “The compatibilist position begins with the premise that people “could not have done otherwise.” But they believe is it more or less useless to talk about life that way, hence compatibilist FW.”

    I don’t get that position at all. If someone believes determinism is true then they should just say so – that everything we have ever done or ever will do was already inevitable before we were born.

    There are advantages to that position. For example, since it makes no sense to feel guilt or responsibility for something that you could not have prevented, something that could not possibly be prevented, then we would never need to feel guilt or responsibility for anything we have done, or will do in the future.

    Because if determinism was true then everything would be something that we could not possibly have prevented – something that no one or nothing could have prevented.

    But some people want to be determinists but also hang on to the idea of alternative courses of action.


  12. Hi garth,

    “Can you give an account of what a non-random process would entail that would allow for “could have done otherwise?”

    Can I do what no scientist has ever done and give an account of how the human mind works? No.

    But when you have a claim like “any non-deterministic process is random” then all you need is one counter example to show that it is false.

    If I was claiming that LFW was the case then I would need to give an account of how it could happen, but all I am claiming is that on current knowledge we don’t know whether it is the case or not.


  13. Hi garth,

    “I’m not saying I agree but they are not being so dumb as to think their version of nothing satisfies the philosophers quest just that the philosophers quest is a bridge to nowhere.”

    As I have pointed out before, this idea that this is some sort of a “philosophers’ quest” is a faulty assumption.

    It is actually very hard to find any philosopher now or in the past who have interested themselves in this question. Leibniz touched on it briefly when he asked “Why is there something rather than nothing, or, given that there must be something why is it the way it is rather than some other way?”. But it was merely a step in the argument and the point was not to find the reason, merely to conclude that there was a reason. So Krauss is merely supporting a premise in Leibniz’s argument.

    Other than that, I have come up blank. Who are the philosphers that are or were on this quest?

    The very first misstep that Krauss made was to assume that this was an age old problem that has stumped philosophers. No, as far as I can tell it has never interested philosophers.

    So it may be a bridge to nowhere, but there are no philosophers on that bridge, only physicists.


  14. Hi Robin,

    We agree on compatibilist determinism. You said:

    “I don’t get that position at all. If someone believes determinism is true then they should just say so ”

    Agreed. In my case I am technically agnostic about determinism/indeterminism in the universe on the deep level, but I am operationally determinist in the manifest world regardless, because A) I see no way for any possible indeterminacy to make “could have done otherwise” any more plausible, and B) more importantly, I take Schopenhauer’s “can not will what I will” observation as empirically observable by anyone in the manifest image making it “useful” to talk about the world that way, especially in light of the point you made that such a view would relieve feelings of guilt, regret, shame, and I would add hate and revenge become irrational.

    So technically I am agnostic on determinism in the scientific image and operationally determinist in the manifest image for reasons A and B above. That’s where I’m at right now but I don’t think I’ll ever stop being malleable in this space.

    “If I was claiming that LFW was the case then I would need to give an account of how it could happen, but all I am claiming is that on current knowledge we don’t know whether it is the case or not.”

    I really think we do. Without evidence of, or even a conceptual description of how it would work, Occam won’t let us go there I think. Again I am malleable in this space but without even a conceptual theory to work off, for now I feel justified in eliminating LFW from my ontology.


  15. Hi Robin,

    Almost forgot, on Kraus, if philosophers agree with him on “nothing” why are they attacking his take on “nothing?” His book was aimed at the creationist claim that a universe can not come from nothing, not philosophers who agree that “nothing” is a bridge to nowhere.


  16. I don’t think I said philosophers agree with Krauss on “nothing”. They don’t. What Krauss was saying is incoherent.


  17. Revenge isn’t irrational under determinism. Revenge and retribution are just scratching an itch. Why would scratching an itch be irrational under determinism?

    That is the most harmful furphy about this debate. Retributive justice is socially harmful whether or not we have LFW.

    But now these people have made it much harder for people like me who argue for non-retributive justice because they are implying that it is now associated with their pet metaphysical theory.


  18. Hi Robin,

    “Revenge and retribution are just scratching an itch. Why would scratching an itch be irrational under determinism?”

    Scratching itches makes you more itchy. Sure the urge to scratch is strong, but knowing that it’s not in your best interest to scratch it helps you resist. The urge for revenge is strong and in the heat of the moment some people can’t resist and they scratch that itch, but upon reflection, one can talk them self down from the revenge rage by recognizing the reality of “couldn’t have done otherwise.” It works for me. Whenever I catch myself hating or moralizing I can always snap myself out of it with “could not have done otherwise.”

    And I can not see how an advocate for non-retributive justice would find this concept to be a roadblock rather than an ally. Is it because it is so hated and dismissed so it taints the non-determinists argument for non-retributive justice?


  19. Hi Robin,

    Take Hawking’s quote: ‘because there is a law such as gravity, the universe can and will create itself from nothing’. So Hawking cannot mean that the law of gravity is encoded in space and time if he thinks that it can call space and time into existence.

    It is wrong to think of laws as agents that “call” stuff into existence. A “law” is a description of how something behaves. It is wrong to think of “how something behaves” as being different from the thing itself; indeed, the only attributes that a particle such as an electron have are how it behaves.

    So Hawking’s quote would mean something like: because of how space and time actually are at a fundamental level, space and time are self-contained and self-creating, they can be in existence without any prior cause.

    The term “law” is perhaps misleading, but much of the language is like that. The one thing Hawking is *not* saying is that there is some prior cause, the “laws”, that “exist” either Platonically or in some “other” space, which then cause or “call” the space-time into existence. Indeed, he’s trying to say the opposite, the absence of any prior cause is the whole point of the sentence.

    Seems to me that intuitions are mostly an obstacle to understanding the laws of physics.

    To quite an extent yes, a training in physics is very much a re-programming of the intuition to work with things that are highly counter-intuitive, thinking of particles as being waves for example, along with relativity and much else.

    Hi garthdaisy,

    Indeed there is no reason to believe we evolved any intuitions for understanding the laws of physics.

    There is every reason to suppose that we have evolved a good-enough appreciation of the everyday physics that was relevant to our survival in the environments we lived in. Thus, we do have a pretty good intuitive understanding of the flight of a thrown ball or spear, and of what would happen if you stepped over a cliff.

    The problem for our intuition is that, over an evolutionary timeframe, all speeds we encountered were vastly less than the speed of light, and so we’re intuitively useless at relativity, and all sizes we encountered were all vastly bigger than Planck’s constant, and so we’re intuitively useless at quantum mechanics.


  20. Coel,

    Well we used to think the stars and the sun revolved around us so even at Newtonian mechanics we are pretty intuitively useless. But yes, what goes up must come down seems pretty intuitive although I actually think it is learned by experience rather than intuited. But as Dan said, “intuition” is a slippery word.


  21. And because we require an atmosphere, we intuititively feel that tge natural tendency of things is to come to a rest.


  22. Hi garthdaisy,

    Well we used to think the stars and the sun revolved around us so even at Newtonian mechanics we are pretty intuitively useless.

    Though Earth-centred models are pretty useful for Earth-centred beings. We all use coordinate systems (e.g. longitude, latitude) and maps that are Earth-centred. In that sense our intuitive understanding is good-enough for most everyday purposes.

    But yes, what goes up must come down seems pretty intuitive although I actually think it is learned by experience rather than intuited.

    We should distinguish between intuition — what we (think we) know without conscious deliberation — and instinct, which is genetically programmed knowledge. Much of our intuition is indeed the result of experience.

    Liked by 1 person

  23. Hi Massimo,

    This may be over-pedantic and is anyhow a lost cause, but adopting my language-police hat:

    Rationing care by age, therefore, does not have to be discriminatory (and it may thus be morally acceptable), …

    Rationing care by age is the very essence of discrimination (wiki: “treatment or consideration of, or making a distinction in favor of or against, a person or thing based on the group, class, or category to which that person or thing is perceived to belong to rather than on individual merit”).

    But, the word “discrimination” does not — in itself — carry any moral connotations. For example, we discriminate against children in not allowing them to drive cars, and most of us would regard that as a good idea. But, nowadays the word seems to have acquired the automatic connotation of moral badness.

    Why does this matter? Well, in today’s offense culture, where people being offended or upset by speech is taken to trump notions of free speech, the word “discrimination” is becoming used as a weapon to silence people. Thus, if someone criticises a religion or says anything that might upset someone, the reply will be: “ooh, you’re discriminating against me”, with the implication that the speaker is morally at fault.


  24. Why would will, free or otherwise, require randomness? We don’t distinguish between courses of action to then randomly pick one. The reality is that there is no frame which can pre-determine the total input into any situation, prior to its occurrence, because a lot of information travels at the speed of light, i.e. all we see.
    So if input cannot be fully known, then neither can output and it is not the function of our will to be random, but to determine a course of action.


  25. Garth, Robin, Brodix,

    Would you mind dropping the free will debate? Not only it has previous little to do with the topic at hand, but I now have a long experience that guarantees that these discussions never, ever, go anywhere. Please move on.


    Interesting point about language, but I disagree. “Discrimination” often very much has a moral connotation, though it doesn’t have to. So my use of the phrase “does not have to be discriminatory [in the moral sense of the term]” is perfectly fine.

    Of course any time you treat one group differently from another one you are “discriminating” in the neutral sense of the term. But adopting that sense would lead to the weird phrase that welfare programs “discriminate” in favor of the poor.

    And I don’t give a hoot about today’s “offense culture”…

    Liked by 1 person

  26. Massimo,

    Per my prior question, what role does insight play in philosophy? If Einstein is imagining surfing a lightwave at 16, it would not seem he had the extensive experience to have sufficiently informed intuition, but possibly not totally encumbered by it, it might have given him the open mindedness to see from a fresh perspective.

    So does the role of simply trying to wipe the slate as clean as possible/push the rest button/start with a blank sheet, have any sort of established role in philosophy, or would that conflict with the whole principle of building on established foundations?

    Obviously the method of just pushing ideas to the point of reductio ad absurdum serves the same function, but can often take decades, or even centuries to play out. Such as with the premise of free will, or even string theory. Then so much momentum and generational angst is built up, that only the more obvious issues are even open to discussion, while some of the real problems have been so thoroughly built into the paradigm, that they are not open to question.


  27. Ah! ‘All worlds;. Not only is there no choice, but ‘you’ make all choices.


  28. In my view ‘you are the process’. You can ‘will’ but not ‘will what to will’. The turtles have to stop some place.


  29. Arthur,
    Not to push our luck, but I don’t think most people would have a problem with being “part of the process.” It is the connotation that our sense of will is irrelevant to the process, that most people’s intuition says, Huh?


%d bloggers like this: