On burning armchairs

According to its (official?) twitter account, experimental philosophy, aka x-phi, is “an interdisciplinary field that uses the collection of empirical data to shed light on philosophical issues.” In what is often cited as “the first experimental philosophy study” Jonathan Weinberg, Shaun Nichols and Stephen Stich presented groups of Westerners and East Asians with some well-known epistemological problems and concluded there is a difference in terms of what they think counts as knowledge. Since 2001 countless similar, questionnaire-based and non-questionnaire-based studies have been published. Experimental philosophers contrast this kind of empirical work with what they call “armchair philosophy”, which is, as they suggest, deficient in some way. Hence their logo, the burning armchair.

There are many different accounts of what armchair philosophy is supposed to be exactly. It’s usually claimed that it’s non-empirical, or a priori, or has to do with analysing concepts, or uses intuitions as evidence, or searches for metaphysical necessities, or relies on introspection. In any case, the general idea is that many philosophers just sit in their armchairs and think, and there must be something wrong with trying to solve philosophical problems this way.

It’s also often claimed that armchair philosophy is mainstream. For example, in their “Experimental Philosophy Manifesto” Joshua Knobe and Shaun Nichols argue that the rise of analytic philosophy in the 20th century “heightened interest in more technical questions that could be addressed from the armchair”*, which they regard “as a highly regrettable development.” Then they advertise x-phi as the remedy to this situation.

What can be wrong with their idea? My beef with experimental philosophers is that what they call armchair philosophy is, I think, a caricature. They seem to be opposed to something that doesn’t exist. The problem they address is therefore a pseudoproblem, and the solution they offer doesn’t solve anything.

To see why, let me first point out there are dozens of straightforward examples of relying on empirical data to make philosophical points in contemporary analytic philosophy. Take, say, Peter Singer’s argument for veganism. The argument is based on biological data concerning animals’ capacity for pain and suffering and their behavioural needs. Or consider the fine-tuning argument for the existence of God, advanced by people like Robin Collins. Its premises are full of detailed empirical data from physics, having to do with values of physical constants and initial conditions of the universe. Or think of Robert Kane, who argues that quantum events in our brains can explain how libertarian free will is possible. Or take Thomas Kuhn who appeals to historical data concerning scientific discoveries to make his point about progress in science. There’s an awful lot of 20th century philosophy that is both very armchair (since philosophers don’t carry out any experiments themselves) and very empirical (since they rely on empirical results obtained by others) at the same time.

Here experimental philosophers might reply: “Ok, this may be mainstream philosophy, it may be done from the armchair, but it’s not what we refer to as armchair stuff. If you look into other instances of philosophical practice, especially in epistemology, metaphysics, philosophy of mind or philosophy of language, you’ll discover cases of purely conceptual, a priori arguments. And that’s what we’re opposed to.”

This is, admittedly, a hypothesis about philosophical practice that is worth testing. However it’s important to remember it’s only one of many possible positions, and one that is not terribly popular outside x-phi. Philosopher of language Martin Davies argues that there is a whole spectrum of views on the role of the empirical in semantics, and consequently on the relation between philosophy and science:

Towards one end of a spectrum of possible views is a position that says that the only questions about language and mind that are susceptible of rational investigation are questions that belong to the sciences of language and mind. According to this position the business of the philosophy of language and mind is simply to hand all the substantive questions over to cognitive science. We might call this first position cognitive scientism. (…)

Towards the opposite end of the spectrum is a position that maintains that the philosophy of language and mind offers a distinctive methodology for investigating a class of substantive questions about the notions that figure in our everyday thinking about linguistic and mental matters. The occupants of this position say that cognitive science has little or nothing to contribute to the philosopher’s project of plotting the contours of our conceptual scheme. Let us call this second position philosophical isolationism.

Experimental philosophers generally think that armchair philosophy as a practice is based on the assumption of isolationism (and that isolationist distinct methodology is purely a priori). There are two problems with this position though. First, even if it’s correct and armchair philosophy is, or tries to be, a priori in some sense, it’s not clear why being a priori should be understood as its flaw. Take mathematics, the ultimate example of a priori activity. I think we can all agree it’d be pretty daft to tell mathematicians to burn their armchairs, put on lab coats and start conducting empirical experiments. So if it’s ok for mathematicians to stay in their armchairs, why is not ok for philosophers? Experimentalists have answers to this question, but they all seem inadequate. For example, they’d say things like “when philosophers use the method of cases, they assume they can know people’s intuitions without leaving their armchairs, but this is impossible”. However philosophers usually don’t try to appeal to people’s intuitions, so they don’t need to care what they are, and the so-called method of cases doesn’t even exist.

Secondly, the more basic problem is that whether philosophy is actually done within the isolationist framework is very controversial, at best. I think there are good reasons to reject this idea. There are quite powerful arguments, put forward by philosophers like Quine or Williamson, for the claim that this kind of philosophical activity, sometimes called conceptual analysis, is not even possible. If those arguments are sound, there is simply no way to peel off the empirical from the conceptual. And when we examine the alleged typical examples of conceptual analysis, it turns out, in my experience at least, there’s always some sort of empirical knowledge that plays an important role in the argument. It’s not necessarily scientific knowledge, very often it’s a kind of knowledge that we gain from everyday experience of the world, but it’s empirical nonetheless.

If experimental philosophers disagree with this, they should have a response to the arguments I’ve just mentioned. The problem is they don’t, as far as I know. They seem to just assume that a priori conceptual analysis is not only possible, but widely practised in contemporary philosophy, and that it’s basically identical with what they call “armchair philosophy”. I agree with Max Deutsch, who argues this picture of philosophy is heavily skewed:

It is now common to hear the dispute between xphiles and those who defend analytic philosophy and its methods characterized as one over whether philosophy can be pursued “from the armchair.” (…) the characterization of the dispute in terms of those for or against armchair philosophy is misleading in at least two ways. First, the characterization wrongly suggests that armchair philosophy is unscientific, or unconcerned with empirical results related to its subject matter. Second, the characterization unfairly casts xphi as a curative—a pro-science balm designed to counteract the tendency to simply sit in an armchair and think.

Since typical survey-style xphi methods are clearly empirical, casting xphiles as opposed to armchair philosophy suggests that armchair methods are not empirical. But this is not true. By definition, a priori methods are not empirical. But sitting in an armchair does not prevent one from appealing to things one has learned a posteriori.

Now: what happens when you try to solve the non-existent problem of lack of empirical data? One thing that might happen is that you start collecting irrelevant empirical data. And this seems to be exactly the problem with a lot of x-phi. For example, experimentalists spend a lot of time studying people’s judgments about knowledge – they try to determine how they differ across demographic groups and what influences them. But when it comes to explaining how their results can help us answer traditional epistemological questions, like “What is knowledge?”, the answers are either lacking or confused. And this makes me think that while armchair philosophy definitely “uses the collection of empirical data to shed light on philosophical issues”, it’s far from clear whether experimental philosophy does it too.

* Note that there are actually two different, incompatible claims being made here. One is that there is only a narrow range of philosophical problems that can be addressed from the armchair and analytic philosophers are too focused on those problems, while largely ignoring everything else. The other claim is that the problems that analytic philosophers address from the armchair cannot, or shouldn’t be, addressed from the armchair. These two claims seem a bit jumbled together in the Manifesto. Anyway, my point is that both are false.

18 replies on “On burning armchairs”

Perhaps, then, while we’re NOT warranted in exploring or “plotting the contours of our conceptual scheme” without taking into account things experienced or experienceable by ordinary human beings bumping about in our world, we ARE warranted in doing so without taking into account the “results” of cognitive science (should there be any that are agreed as being “the best cognitive science has to offer”).


Good point: I didn’t draw a clear line between empirical knowledge and scientific empirical knowledge in this paragraph. I think you may well be right, and Davies seems to agree too. He believes that equating “the domain of science with the domain of rational inquiry concerning the natural world” would be “a gratuitous scientism”. And he suggests that some kind of intermediate position between cognitive scientism and philosophical isolationism is correct. Myself, I’m not sure to what extent we can, or will be able to, equate the two. But I definitely think we should reject the idea that philosophers are, or even try to be, a priori isolationists when they practise philosophy (their metaphilosophical views about their own practice are a different story though).


Hi Tomasz,
Thanks for these helpful thoughts on experimental philosophy. I just wanted to follow up and ask some further questions about how you are seeing this.
Shaun Nichols and I argue in our manifesto that contemporary experimental philosophy is doing something very similar to what one finds in a lot of very traditional work in the history of philosophy. For example, as you note, there has been a lot of work in experimental philosophy on judgments about knowledge across cultures. This work indicates that people from numerous different cultures make extraordinarily similar judgments – indicating a very surprising degree of cross-cultural convergence.
In the way that we were thinking of it, this type of work is not at all a radical departure from traditional philosophy but is instead extremely continuous with what one finds, e.g., in Book 1 of Locke’s Essay Concerning Human Understanding.
I wasn’t sure whether you were disagreeing with that claim or whether you were just disagreeing with our claim that 20th century analytic philosophy tended not to focus on these more empirical issues in the way that, we suggest, previous philosophy so often did.


Thanks for your comment Joshua – to answer your question, I think I disagree with both. First, I don’t think there was any kind of anti-empirical turn in the 20th century philosophy, at least if we use the term “philosophy” in the contemporary sense. The concept of philosophy seems to have changed at some point in the late 19th century (I wrote more about this here), and this may create the false impression.
Second, I don’t think there is too much continuity between experimental philosophy and what you call traditional philosophy. If I remember correctly, Locke didn’t do, or appeal to, any cross-cultural studies on people’s judgments about what counts as knowledge. Or any (proto-)psychological studies whatsoever. He mostly appeals to empirical knowledge he gained from personal, everyday experience. I’d say this makes him more like a typical 20th century armchair analytic philosopher then like an experimental philosopher. As for the “very surprising degree of cross-cultural convergence”, this seems controversial. You say there’s a lot of convergence, but other experimental philosophers, like Machery, argue there’s very little. But irrespective of who’s right about this, I cannot see how this kind of work is supposed to be “extremely continuous” with Locke’s work or other traditional philosophy stuff.


Hi Tomasz,
I want to better understand the point you are making about differences between what contemporary experimental philosophers do and what philosophers like John Locke did. From this most recent comment, it sounds like you are saying that both are interested in cross-cultural differences in people’s judgments but that they differ in that they use different methods. Contemporary experimental philosophers use systematic quantitative studies, whereas Locke used more informal approaches (such as relying on reports from travelers). Is that correct? If it is, I completely agree with what you are saying.
(Since it isn’t possible to convey one’s tone on a blog using facial expression, etc., I should say explicitly that this is intended as a friendly question. I really am just asking for further information about your thoughts on this.)
Regarding the minor point about convergence, you are right that there is controversy about the broader question regarding the overall degree of cross-cultural convergence in philosophical intuitions, but I don’t think there is much controversy about the more narrow question regarding cross-cultural convergence in knowledge attributions specifically. See, for example:
But obviously, that is not the important issue here. Regardless of what turns out to be the right answer when it comes to this empirical matter, the metaphilosophical questions will still be more or less the same.


“It’s not necessarily scientific knowledge, very often it’s a kind of knowledge that we gain from everyday experience of the world, but it’s empirical nonetheless.”
This is such an important point that seems to get lost nowadays in the reactions and counter-reactions borne by scientism at large and the science-envy of so much of contemporary analytic philosophy in particular.


Hi Joshua,
First of all: no worries about the tone! Everything you wrote looked perfectly friendly, and anyway I don’t mind if someone chooses to be more confrontational, as long as the discussion doesn’t descend into pure name-calling. Being friendly is always more helpful though. Let me also assure you that even though I disagree with a lot of what you say I really appreciate your comments and I don’t mean anything I write to sound snide.
With respect to the convergence or lack thereof, we seem to agree it’s not very important for this debate. Anyway, thanks for clarification, I’m not entirely up to date with all the results, I thought that some people still found the 2001 Weinberg et al. study I link in the post convincing, but now I understand it’s universally rejected because of poor methodology, right?
Now back to Locke. Initially I wasn’t sure which bits of Book 1 you had in mind and what kind of analogy you were trying to make, and because of that I didn’t express myself very clearly, so let me clarify. Once again: I don’t agree there is an “extreme continuity” between Book 1 of Locke’s Essay and the x-phi studies you mentioned. In fact I think there’s little similarity between them – so little that it doesn’t make sense to say that Locke’s work is more like x-phi than like contemporary armchair epistemology.
First, we must ask: what is Locke up to? He wants to prove that there are no innate principles. And his argument is: if innate principles exist, then there are some “to which all mankind give a universal assent”, but there are no such principles, therefore innate principles don’t exist. He carefully specifies what “assent” would have to be like in this case, then he takes a number of principles that people believe to be universally accepted and innate, like “it is impossible for the same thing to be and not to be”, “white is not black”, “men should keep their promises”, “God exists” etc. and shows there are groups of people that don’t assent to them (in the way he has specified). What are those groups? He focuses on children, illiterate people, “idiots” and “savages”, and then while discussing atheism he also mentions people in different parts of the world, like Brazil, China or the Caribbean.
What are experimental philosophers up to? They are interested in what they call “knowledge ascriptions” in different cultures. That’s the first big difference: Locke wants to examine whether people accept different logical, metaphysical and moral truths, experimentalists want to examine people’s judgments about what counts as knowledge in particular situations. Why do they want to study these judgments? Well, at least initially the motivation was the belief that philosophers’ job is to explain people’s intuitions, or to use them as evidence for their theories about knowledge. If this is true philosophers need to know what those intuitions are, what shapes them and whether they’re reliable. That’s the second big difference: Locke is not interested in explaining knowledge-intuitions or using them as evidence for anything. He’s interested in innateness. Perhaps experimentalists could move their goalposts a little bit and argue they now want to study knowledge ascriptions for some different reasons? If this is true, I have no idea what those new reasons are, so it’s impossible for me to compare Locke’s project and the new x-phi project, if there is such a thing.
To be fair, there are some similarities between Locke and x-philsophers. Both are trying to make a philosophical point about knowledge, and both are interested in examining people’s beliefs across different cultures. But that’s pretty much it. There are more important similarities between Locke and armchair epistemologists. Just like Locke, they are uninterested in people’s intuitions about what counts as knowledge. Just like Locke, they appeal to all sorts of empirical evidence whenever they need to. On the other hand, x-philosophers are interested in intuitions and they only work with one particular kind of empirical evidence. This is why I don’t agree with your point about “extreme continuity”. Does that make sense?


Hi Tomasz!

Thanks so much, this is super helpful. I hadn’t quite understood your position before, but I think I have a much better understanding now.

I completely agree with the way you describe what Locke is up to, and I also completely agree that existing metaphilosophical work on experimental philosophy describes it as not being engaged in that sort of project at all but instead being engaged in a more or less unrelated project (the exact one you lay out here). However, in my view, what’s going on here is not that experimental philosophy is a radical departure from what people did more traditionally in the history of philosophy. Rather, I think that what experimental philosophers actually do day to day in their research is very continuous with these more traditional projects, and the issue is just that the metaphilosophical literature is getting increasingly out of touch with what experimental philosophers actually do.

Clearly, Locke was interested in questions about the nature of human understanding (e.g., about the existence of innate principles), and he was interested in cross-cultural differences insofar as these differences could shed light on questions about the nature of human understanding. Just as you say, the metaphilosophical literature on experimental philosophy makes it sound as though experimental philosophers aren’t interested in that all but are just engaged in some deeply different sort of project. But I don’t think that this is an accurate portrayal. If you look at what experimental philosophers actually do day to day, it’s really much more continuous with these more traditional projects than one would think if one just tried to guess what we do by reading the metaphilosophical literature.

For example, you are right to say that Locke wanted to look at children, people with psychological disorders and people from other cultures as a way to figure out whether there are innate principles, but experimental philosophers sometimes do *exactly* the same thing. For example:

If I understand correctly, what I am saying is not so wildly different from the view you sketched in your recent comment. The only thing I am disagreeing with is the assumption that what experimental philosophers actually do in their day to day research is the very same thing that gets discussed in the metaphilosophical literature on experimental philosophy.


Hi Joshua, many thanks for your comment! It really helped me understand where I stand on this, and how we differ.
So we definitely agree on several things: that Locke was interested in questions having to do with human understanding, and that x-philosophers are interested in these questions too. Also that both Locke and x-philosophers are interested in learning about beliefs of people in different cultures, people with mental disorders, children etc. However note that armchair philosophers – as you call them – are not any different in this respect. They also often discuss beliefs of people in different cultures, people with mental disorders, or children. Take, say, Fodor’s stuff on how children learn a language, or Quine’s stuff on radical translation. Armchair philosophers like Fodor or Quine usually don’t rely on systematic quantitative studies (as I think they don’t need to), their empirical data comes from personal experience, or informal reports of other people’s personal experience. They also typically rely on more abstract and general empirical data, but in any case, their work has little to do with conceptual analysis. Just like Locke’s argument has little to do with conceptual analysis, even though Locke didn’t use any systematic empirical data. So on this very broad and vague account, all philosophers are similar.
However in my view, when we get into important details, it turns out there is a significant continuity between what you call traditional philosophy (TP) and armchair philosophy (AP), but x-phi is a radical departure from both, which has very serious (and dire) consequences. You reject this claim, which seems to be the gist of our disagreement.
Let me begin with a point that’s hopefully uncontroversial: anyone who wants to use empirical data to make a philosophical point should be prepared to answer three questions:
(1) What philosophical question do I want to answer?
(2) What empirical data do I want to collect?
(3) How can this kind of data help me answer my question?
Now: if you start asking this set of three questions for different examples of TP, AP and x-phi, you’ll see that x-phi always stands out. When x-philosophers’ answer to (1) is any of the traditional philosophical problems that TP and AP deal with, then their answer to (3) must be pretty outlandish, and very unlike anything traditional and armchair philosophers would say. If x-philosophers’ answer to (3) is not outlandish, then it turns out that their answer to (1) is not a typical philosophical problem. I mean, maybe you can call it philosophical in some cases, but it’s definitely not anything TP and AP are normally interested in addressing. So either way, TP and AP are not very continuous with x-phi.
Let me illustrate this with Locke. If Locke had been interested in asking random people in different cultures questions like “Do you agree that there is innate knowledge?”, and had tried to answer the question “Is there innate knowledge?” on the basis of their answers, then I’d agree there is a continuity between Locke and the Gettier-intuition study you linked. But clearly Locke didn’t do anything like that. I think he’d find this idea bizarre. Armchair philosophers would also find it bizarre. They’d say it’s a silly way to find out about the existence of innate knowledge – and they’d be right.
However x-philosophers study responses to Gettier-scenarios across cultures, and on the basis of those responses they hope to answer the question that Gettier tried to answer: “Is knowledge justified true belief?”. This enterprise seems equally silly to me.
Here x-philosophers might protest: “Wait, we’re not trying to answer this question! We’re only interested in the folk concept of knowledge”. If this is the case, then x-philosophers are not doing philosophy, at least not the kind of philosophy that traditional and armchair philosophers do. TP and AP are not interested in studying folk concepts of knowledge for the sake of studying folk concepts of knowledge. They are interested in studying knowledge itself. (Btw, I doubt whether x-phi surveys are methodologically suitable even for studying folk concepts, but this is less relevant for our discussion, so I’m putting my objections aside.)
So from my perspective, x-phi presents us with a dilemma: either we abandon our traditional philosophical questions and replace them with completely different questions (while still calling it “philosophy”), or we start answering the traditional questions in a highly irrational way. Neither of these options seems very attractive to me. And both constitute a radical departure.
This is why I don’t agree with your claim that metaphilosophy is out of touch with what experimental philosophers do (btw, I assume that by “metaphilosophy” you must mean just a fraction of metaphilosophy that is not x-phi-friendly). I think that while some of those criticisms miss the point, there are also some that are correct. This is not to say that x-phi research cannot be valuable in some way. I think it can. But I also think that x-philosophers are guilty of caricaturing armchair philosophy and grossly misrepresenting the nature of their own project.


Hello again!

Thanks once again for your response. I really feel like we are making a lot of progress here, and I’m very grateful for your continued engagement.

The key point I was making was about what experimental philosophers actually do. I take it that this is a purely empirical question – a question that could be resolved just by looking at published papers in the field – and I am disagreeing with a specific view about that question that comes out of some of the metaphilosophical literature.

I think we both agree about what is going on in the work we are discussing from more traditional philosophy. In the early modern period, for example, philosophers were interested in whether people had innate ideas, and they tried to address this question by looking, e.g., at whether people’s judgments varied substantially across cultures.

In your recent comment, you suggest that it would be a massive departure from this tradition if someone ran a study asking people whether there was innate knowledge and then using the results to argue that there is innate knowledge. I completely agree, and I also agree that the metaphilosophical literature makes it sound as though we do things like that.

My point is just that we literally *never* do that. There is a lot of discussion in experimental philosophy about whether people have innate knowledge, and as far as I know, there has never been a single published experimental philosophy paper that uses that type of argument.

Rather, what experimental philosophers actually do to make progress on this question is exactly the same thing that the early modern philosophers did. For example, if you look at the two papers linked in my previous comment, one looks at whether children have the same intuitions that adults do, and the other looks at whether people from different cultures have the same intuitions. Both papers then use these sorts of data as evidence for a claim about whether people have innate knowledge.

In other words, there is no disagreement between us about whether the thing you describe would be a big departure from traditional philosophy. My point is just that experimental philosophers don’t actually do that thing. Then, further, I’m trying to suggest that the thing that we actually do is very, very continuous with traditional philosophy.

Thanks again for continuing this conversation!


Thanks Joshua, after your last comment I think I’m beginning to see the light in this discussion! You say our disagreement can only “be resolved just by looking at published papers in the field”, and I fully agree. Let’s look at the papers then. So far you’ve linked three: Machery at al. 2017, Yuan&Kim (manuscript) and Leslie et al. 2006. Let’s discuss them in turn.

First, Machery et al. 2017. You say x-philosophers never try to find out about the existence of innate knowledge by asking random people in different countries whether they agree there is innate knowledge. I believe you and I never tried to suggest otherwise. My thought experiment about Locke doing this was meant to serve two purposes: first, I wanted to show under which conditions Locke would count as continuous with x-phi (Machery et al. 2017, more specifically). Second, I wanted to make an analogy between trying to find out about the existence of innate knowledge this way and what Machery et al. do in their paper, namely trying to find out whether the Gettier judgment is true by asking random people in different countries whether it’s true.

You seem to agree with me that it’d be absurd to try to find out whether there is innate knowledge this way. If you agree, may I ask you: do you also agree that it’s absurd to try to find out whether the Gettier judgment is true this way? If not, why? What is the difference between the two cases? For me, there are two options: either we conclude that Machery et al. are doing something highly irrational, or we ignore what they say about intuitions being used as evidence in the “Philosophical Implications” section of their paper and we conclude their aim was just to show how people in different countries respond to Gettier-type scenarios. If we choose this option, the question arises: what is the philosophical significance of learning about how people in different countries respond to Gettier-type scenarios? Either this paper is a piece of bad philosophy (this is what the metaphilosophical remarks in the text imply) or a piece of social psychology whose relevance to philosophy remains unknown (at least based on what is stated in the paper). Do you agree it’s the former or the latter? If you go with the latter, why do you think we should ignore the metaphilosophical remarks? Or maybe you think there is a third option?

Now to Yuan&Kim. They also study Gettier judgments across cultures, but their aim is at least partly different. They suggest that their results can be explained by the hypothesis according to which the concept of knowledge is innate. They say it’s only one possible option as the results are also compatible with a different hypothesis. But anyway, the problem of whether the concept of knowledge is innate is very different from the problem Gettier wanted to address, namely whether knowledge is justified true belief. And it seems to me that while testing Gettier intuitions across cultures is clearly the wrong way to address the latter, it might not be the wrong way to address the former (depending on many things, let’s not get into details right now). Would you agree with that?

Something similar is going on in Leslie et al., except there is no thought experiment previously used in the “armchair” literature. You say your results about children’s judgments exhibiting the Knobe effect can be explained by the hypothesis according to which the effect “is the product of an innate ‘‘moral faculty’’”, and that there is also a different hypothesis compatible with the results.

So to sum up: to the extent Yuan&Kim and Leslie et al. are roughly on the right track with the way they test their innateness hypotheses, I don’t have any quibbles about their work. I think it can be philosophically valuable and it’d count as continuous with Locke’s project. And if I’m right about all this, it follows that x-phi contains bad stuff I’ve been talking about, but also good stuff you’ve been talking about. Would you agree with that, or do you think there’s no bad stuff?


Hi Tomasz,

Definitely seems like we are converging here! Let’s talk further about the remaining points of disagreement.

I agree that lots of experimental philosophy papers are trying to look at intuitions about whether p as a way to figure out whether p is actually true, but I would describe the method most of them use in a slightly different way. Usually, there are two steps. First, one uses facts about the patterns in people’s intuitions to argue for some view about the underlying cognitive processes that explain those intuitions.  Then one uses claims about the underlying cognitive processes to argue for a view about whether p is actually true.

For example, Yuan and Kim argue that the explanation of the epistemic intuitions in our culture is not (a)  idiosyncratic facts about our specific culture but is instead (b) innate mechanisms or (c) facts about which epistemic intuitions are most helpful in achieving certain goals.

Then the idea is that claims like this one actually help us figure out whether certain intuitions are true. In other words, it’s not that the percentage of participants attributing knowledge in certain conditions directly helps us figure out what’s true. Rather, the percentages are used as evidence for claims about the underlying processes, and claims about the underlying processes are thought to help us figure out what’s true.

Of course, you might think that experimental philosophers are wrong to think that claims about the underlying processes can help us figure out what’s true – it’s certainly a controversial view that is worth disputing – but if you do think we are wrong about this, it still wouldn’t be that we are going wrong by directly inferring from the percentages to claims about the nature of knowledge itself.


Hi Joshua! Previously I asked if you agree there’s some bad stuff in x-phi, and your answer seems a bit evasive, but more on the “no” side (correct me if I’m wrong). I still think there’s a lot of badness out there.

You say you agree “that lots of experimental philosophy papers are trying to look at intuitions about whether p as a way to figure out whether p is actually true”, but you’d “describe the method most of them use in a slightly different way”. Actually you’re not describing it in a different way, you’re describing it the same way I do. I’ve never tried to suggest that x-philosophers just run intuition popularity contests, and always announce the most popular intuition as true. I agree they’re interested in what influences intuitions. But I don’t think this make their method any less dodgy.

Let me try another thought experiment. Suppose you want to figure out whether the Earth is flat. You discover there are some flat-Earthers, and they present you with their best evidence. There are also round-Earthers, and they present you with their best evidence. But you’re like: “I’m gonna ignore this evidence, I’m not really interested in evaluating it. I think I’d better ask random people in different countries whether the Earth is flat and try to learn something about the underlying cognitive processes behind their snap judgments. By knowing these processes I’ll figure out what the shape of the Earth actually is!”. Does this make sense to you?

It doesn’t to me, but x-philosophers do something exactly like that. They take a philosophical judgment, p, they ignore evidence in favour of p that is typically provided by the philosopher who claims that p, and off they go handing out questionnaires to learn how many people believe that p and what influences their judgment that p, or that not-p. Why do they do that? The answer is simple: because they mistakenly believe that philosophers expect us to accept p on the basis of its intuitiveness, not on the basis of arguments for p. The belief that philosophers use intuitions as evidence is the original sin of the movement.

So that’s, in my view, one thing that’s fundamentally wrong with x-phi. It’s closely related to another problem: relentless caricaturing of armchair philosophy. Would you agree there is at least some level of caricaturing going on here, or do you think there’s no problem whatsoever?


Hi Tomasz,

I’ve really enjoyed this conversation, but I just got hit with a bunch of other obligations, so I’m afraid this will have to be my final comment. Happy to give you the last word.

First, a description of where I think we agree. Experimental philosophers run experimental studies that aim to provide insight into why people think the way they do. This can play two different roles in larger philosophical questions: (a) Sometimes the question is itself about how people think, and the empirical claims are directly relevant. (b) Sometimes experimental philosophers argue with understanding why people believe that p will help us figure out whether p is actually true.

If I understand correctly, you are arguing that (a) might be a reasonable approach but that (b) is a mistake. This does seem like a valuable line of attack in trying to criticize experimental philosophy, and I think it connects up in helpful ways with some more general questions in epistemology. My own more general view is that if I learn certain facts about why I believe that p, these facts might give me reason to change my credence in p. This more general view informs my view about the application of that type of argument to philosophical questions more specifically, but of course, that more general view might be disputed. If it turns out to be false, this would have obvious and important implications for questions about the use of this type of argument in philosophy.

By contrast, I don’t think that claims about what ‘armchair philosophy’ usually does would have any important implications for these sorts of arguments. For example, I have argued that if we gain further empirical insight into why people are drawn toward compatibilism, this insight can help us figure out whether compatibilism is actually true.  Perhaps I am right about this, or perhaps I’m  wrong. (You seem to think that I am wrong.) But whether I am right or wrong doesn’t hinge in any way on claims about what armchair philosophy usually does to understand the nature of free will. It only hinges on epistemological claims about whether learning new facts about the explanations of our attitudes toward certain propositions should change our credences in those propositions.

p.s. I’m sorry that I won’t be able to keep writing back about this, but if you respond to this comment, I will absolutely be sure to carefully read your response. Thanks again for the conversation! 


Hi Joshua, no worries and thanks again for all your comments, they helped me change my mind about parts of x-phi – I’ve written a new post to wrap it up, and there’s another one in the pipeline.

Regarding your final point: you suggest that I disagree with your claim that “if I learn certain facts about why I believe that p, these facts might give me reason to change my credence in p.” But I actually agree with it. However note that changing your credence in p is only possible when you already know what makes p true. And when p is not made true by facts about why you believe that p, learning those facts is pointless if your goal is to find out if p.

Going back to my example: if your p is “the Earth is not flat” and you wonder if your credence in p is well-adjusted to the available evidence for p, you don’t really need to learn how many people in the US and other countries are flat-Earthers and to what extent this is due to education, religion, the influence of celebrities, genetic factors etc. I mean, you may eventually learn something relevant if you dig deep enough, but this seems like a very roundabout and inefficient way of addressing the problem, at best. A more efficient approach would be to skip the psychology part and go straight into facts about the shape of the Earth.

That’s why I think that if x-philosophers started paying more attention to what armchair philosophers actually do (especially in terms of employing conceptual analysis and using intuitions as evidence), they’d understand that a lot of x-phi is about collecting irrelevant data. Armchair philosophers are typically very much in favour of the interdisciplinary approach to philosophical problems and using all kinds of empirical evidence in philosophy, as long as it’s relevant.


I think it’s also worth noting here that a lot of science is done from the armchair: theoreticians in all branches of science often work from their armchairs. Sure they are generally familiar with the present state of experimental work in their fields, but that doesn’t stop a lot of their work being done from the armchair, especially the work of coming up with new theories to explain the data.


Exactly. Even if you’re an experimental scientist most of what you do is armchair-based: you need to design the experiments, assess statistical significance, interpret results etc. And when armchair science gets more abstract it can be hard to distinguish it from philosophy: for example lines that people draw between theoretical physics and metaphysics often seem arbitrary. And yet for some reason many philosophers believe that armchair activity must be somehow detached from, or independent of empirical data.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s