Thought Experiments and Philosophical Knowledge

r 2011 The Author
Metaphilosophy r 2011 Metaphilosophy LLC and Blackwell Publishing Ltd

Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK, and
350 Main Street, Malden, MA 02148, USA
METAPHILOSOPHY
Vol. 42, No. 3, April 2011
0026-1068
THOUGHT EXPERIMENTS AND PHILOSOPHICAL

KNOWLEDGE
EDOUARD MACHERY
Abstract: While thought experiments play an important role in contemporary

analytic philosophy, much remains unclear about thought experiments. In
particular, it is still unclear whether the judgments elicited by thought experiments
can provide evidence for the premises of philosophical arguments. This article
argues that, if an inuential and promising view about the nature of the judgments elicited by thought experiments is correct, then many thought experiments in philosophy fail to provide any evidence for the premises of philosophical
arguments.
Keywords: thought experiments, skepticism, intuitions, philosophical expertise.
Thought experiments are a distinctive feature of contemporary analytic

philosophy, and many inuential arguments rest on premises supported
by judgments elicited by thought experiments. Just think of the Godel
case and Twin Earth in the philosophy of language, Mary the neuroscientist and zombies in the philosophy of mind, Gettier cases in
epistemology, and trolley cases or the Society of Music Lovers in ethics.
Philosophers seem to assume that the judgments elicited by such thought
experiments have an important role to play in the growth of philosophical
knowledge.
Despite their prominence, however, much remains unclear about
thought experiments. Of particular importance is the epistemic status of
the judgments they elicit: Can these judgments provide evidence for the
premises of philosophical arguments? If they cant, then thought experiments cannot play any evidential role in philosophy, and their contribution to the growth of philosophical knowledge is severely limited.
In this article, I mount a defense of this kind of skepticism. More
precisely, I argue that an inuential and promising view about the nature
of the judgments elicited by thought experiments (endorsed, e.g., in Devitt
forthcoming and Williamson 2007) has striking skeptical implications:
many thought experiments in philosophy fail to provide any evidence for
the premises of philosophical arguments, and the kind of thought
r 2011 The Author
192
EDOUARD MACHERY
experiment that could provide such evidence is unlikely to be of much use

for philosophers.1
In section 1, I describe an inuential, promising view about the nature
and justication of the judgments elicited by thought experiments. In
section 2, I show that this view has skeptical implications for the judgments
elicited by many thought experiments. In section 3, I examine whether one
could appeal to philosophers expertise to block these implications.
1. Thought Experiments
1.1. Two Examples
Lets start with two examples of thought experiments in recent philosophy:
Kripkes Godel case and Foots and Thomsons trolley cases. According to
the versions of descriptivism Kripke criticizes in Naming and Necessity, a
proper name refers to the individual that satises the description competent
speakers associate with it. By contrast, according to Kripkes causal-historical
theory, a proper name refers to the individual this name was introduced to
refer to, provided that there is a chain of uses running from the introduction
of this name to its current uses. As part of the argument against
descriptivism,2 Kripke describes a situation in which a speaker associates a
proper name, Godel, with a description that is not true of the original
bearer of the name but is true of someone else, called Schmidt in the story.
Descriptivist theories of reference typically entail that in this situation
Godel refers to the man originally called Schmidt. But, Kripke maintains, this is just wrong:
Suppose that Godel was not in fact the author of [Godels] theorem. A man
called Schmidt . . . actually did the work in question. His friend Godel
somehow got hold of the manuscript and it was thereafter attributed to Godel.
On the [descriptivist] view . . . when our ordinary man uses the name Godel,
he really means to refer to Schmidt, because Schmidt is the unique person
satisfying the description the man who discovered the incompleteness of
arithmetic. . . . But it seems we are not. We simply are not. (Kripke 1972, 8384)
In contrast, causal-historical theories of reference are consistent with

Godel referring to its original bearer because he is the person causallyhistorically linked with contemporary uses of the name.
Thus, the Godel case elicits the judgment that in the circumstances
described by the case Godel refers to the man originally Godel. This
judgment is evidence that in these circumstances Godel does refer to
the man originally called Godel. That Godel refers to the man
1
I wont discuss the role and nature of thought experiments in science in this article (see,
e.g., Brown 2004; Norton 2004). A more complete discussion of thought experiments would
examine whether and how thought experiments differ in philosophy and in science.
2
I will bracket the fact that the exact role of the Godel case in Kripkes argument against
descriptivism is debated (Devitt forthcoming; Ichikawa, Maitra, and Weatherson forthcoming).
r 2011 The Author
THOUGHT EXPERIMENTS AND PHILOSOPHICAL KNOWLEDGE
193
originally Godel in such circumstances is the premise of the argument

against the common versions of descriptivism examined by Kripke.
Lets now turn to the trolley cases, focusing on Thomsons 1985 article
The Trolley Argument. Thomson starts by discussing Foots so-called
switch case:
Some years ago, Philippa Foot drew attention to an extraordinarily interesting
problem (Foot, 1978). Suppose you are the driver of a trolley. The trolley rounds
a bend, and there come into view ahead ve track workmen who have been
repairing the track. The track goes through a bit of valley at that point, and the
sides are steep, so you must stop the trolley if you are to avoid running the ve
men down. You step on the brakes, but alas they dont work. Now you suddenly
see a spur of track leading off to the right. You can turn the trolley onto it, and
thus save the ve men on the straight track ahead. Unfortunately, Mrs. Foot has
arranged that there is one workman on that spur of track. He can no more get
off the track in time than the ve can, so you will kill him if you turn the trolley
onto him. Is it morally permissible for you to turn the trolley? Everyone to
whom I have put this hypothetical case says, Yes, it is. (Thomson 1985, 1395)
Thomson compares Foots case with other versions of the trolley case,
including the so-called footbridge case:
Consider a case . . . in which you are standing on a footbridge over the trolley
track. You can see a trolley hurtling down the track, out of control. You turn
around to see where the trolley is headed, and there are ve workmen on the
track where it exits form under the footbridge. What to do? Being an expert on
trolleys, you know of one certain way to stop an out-of-control trolley: Drop a
really heavy weight in its path. But where to nd one? It just so happens that
standing next to you on the footbridge is a fat man, a really fat man. He is
leaning over the railing, watching the trolley; all you have to do is give him a
little shove, and over the railing he will go, onto the track in the path of the
trolley. Would it be permissible for you to do this? Everyone to whom I have
put this case says it would not be. (Thomson 1985, 1409)
The judgment that it is permissible for the driver to turn the trolley onto
the side track in the switch case is evidence that in the circumstances
described by this case it is permissible for the driver to turn the trolley
onto the side track. The judgment that it is not permissible to push the
large man in the footbridge case is evidence that in the circumstances
described by this case it is not permissible to push the large man. That it is
permissible to turn the trolley onto the side track but not permissible to
push the large man onto the track reveals that there is a moral difference
between the two situations. Principles such as the Doctrine of Double
Effect are meant to characterize this moral difference.3
3
Of course, Thomson herself is critical of the Doctrine of Double Effect.
r 2011 The Author

194
EDOUARD MACHERY
1.2. The Nature and Function of Thought Experiments

So, whats a thought experiment? Typically, a thought experiment describes
a nonactual situation, and it invites the reader to make a judgment about an
aspect of this situation (applying such concepts as REFERENCE, PERMISSIBILITY, KNOWLEDGE, etc.). This judgment is evidence for a particular fact
(e.g., that Godel refers to the man originally called Godel), and that
fact (not the act of judging itself, but the fact that makes the judgment true if
it is true) is typically used in some philosophical argument.4
Why is a judgment elicited by a thought experiment taken to provide
evidence for particular facts? Presumably, for the same reason that ordinary
judgments are often taken to provide evidence for particular facts. If I judge
of an object that it is a chair, my judgment that it is a chair is evidence that it
is a chair because I am reliable at sorting chairs from nonchairs.
Providing evidence for premises of philosophical arguments is not the
only possible function of thought experiments. Thought experiments in
philosophy are sometimes merely illustrative: they are merely meant to
illustrate how a theory would apply to a particular case. For instance,
Davidsons (1987) swampman thought experiment may be understood as
illustrating Davidsons argument rather than providing evidence for a
premise of this argument. Here I will be concerned only with the
evidential function of thought experiments in contemporary philosophy.
1.3. The Psychology of Thought Experiments
Although there are numerous proposals about the psychological capacities involved in making the judgments elicited by thought experiments,
these can be classied into two fundamentally distinct types. According to
the Ordinary Judgment Proposal, the judgments elicited by thought
experiments are underwritten by the psychological capacities that also
underlie the judgments we make about everyday situations. According to
this view, for instance, the same psychological capacities underwrite my
judgment that the agent described in a fake-barn case does not know that
he is seeing a barn and my judgment that, judging by his answer to the
test, one of my undergraduate students does not know what the DN
account of explanation is. Williamson endorses this proposal about the
nature of the judgments elicited by thought experiments: We should
4
Some seem to deny that the judgments themselves (viz., the acts of judging) play any
evidential role. For instance, they seem to deny that the judgment that the agent does not
know the relevant proposition in a Gettier case provides evidence for the fact that the agent
does not know the relevant proposition. If this is really their position (which I am inclined to
doubt), it is simply mind-boggling. If I am a reliable detector of knowledge in everyday
circumstances, my judgment that an undergraduate does not know what the HD account of
conrmation is provides evidence that she is indeed ignorant. Furthermore, it is hard to see
what other kind of evidence could be put forward to support the claim that, e.g., in the
situation described by the Godel case Godel refers to Godel.
r 2011 The Author

195
expect the cognitive capacities used in philosophy to be cases of general

cognitive capacities used in ordinary life, perhaps trained, developed, and
systematically applied in various special ways (2007, 136).
According to the Spooky Judgment Proposal, the judgments elicited
by thought experiments and the judgments we make about everyday
situations are of a different kind, and they are underwritten by different
psychological capacities. According to this view, for instance, my judgment that the agent described in a fake-barn case does not know that she
is seeing a barn and the judgment that, judging by her answer to the test,
one of my undergraduate students does not know what the DN account
of explanation is are of a different kind, and they are underwritten by two
different psychological capacities. For instance, Sosa holds that the
judgments elicited by thought experiments such as the Gettier cases,
which he calls intuitions, have very distinctive features: On my
proposal, 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 a competence, and the content is explicitly or
implicitly modal (i.e. attributes necessity or possibility) (2007, 101).5
Admittedly, the Spooky Judgment Proposal is a catch-all category.
Different philosophers characterize the nature of the judgments elicited
by thought experiments and the nature of the distinctive psychological
capacities underwriting these differently.
In presenting the Spooky Judgment Proposal and the Ordinary
Judgment Proposal, I have alluded to the psychological capacities underlying judgments. I have in mind the cognitive structuresnamely, the
bodies of knowledge and the cognitive processes dened over thesewe
use when we decide to apply a concept to a particular object or to
subsume a concept under another concept. The psychology of concepts
has much to say about the nature of these bodies of knowledge and about
the nature of the cognitive processes dened over these (for review, see
Murphy 2002 and Machery 2009 and 2010), but I will not discuss the
implications of this literature here (see Machery 2009, chapter 2, for some
discussion). Still, it is worth noting that the proponents of the Ordinary
Judgment Proposal sometimes have a simplistic conception of the
psychological capacities underlying judgments. It is misleading to suggest
that the application of a concept to an object (the kind of judgment
expressed by, for example, This is a dog) is underwritten by a single
psychological capacity. Instead, evidence suggests that this application is
underwritten by several distinct bodies of knowledge and psychological
capacities that are to a large extent independent of one another.6
5
I will not attempt to disentangle the knot of ideas expressed here.

The same thing is true of the application of a concept to a substance (the kind of
judgment expressed by, e.g., This is gold), to an action (the kind of judgment expressed by,
e.g., This is right), and so on.
6
r 2011 The Author

196
EDOUARD MACHERY
1.4. The Epistemology of Thought Experiments

The psychology of thought experiments is philosophically relevant
because it seems to bear on the epistemic status of the judgments elicited
by thought experimentsin particular, it seems to provide ammunition
against the skeptic about thought experiments. A thought experiment
skeptic doubts that the judgments elicited by thought experiments provide
any evidence for the premises of philosophical arguments (or at least that
they provide sufcient evidence). Typically, the thought experiment
skeptic defends his skepticism by arguing that these judgments are not
sufciently reliable: the judgments are not sufciently more likely to be
true than false (for a different kind of skepticism, see Weinberg 2007).
The Ordinary Judgment Proposal seems to undermine this form of
skepticism. Because the same psychological capacities underlie everyday
judgments and the judgments elicited by thought experiments, the
judgments elicited by thought experiments seem to inherit whatever
reliability everyday judgments possess. And surely everyday judgments
are typically reliable. For instance, whatever makes my everyday judgments about knowledge (e.g., my judgment that one of my undergraduate
students does not know what the DN account of explanation is) reliable
also makes the judgments about knowledge elicited by thought experiments in epistemology (such as my judgment that the agent in a fake-barn
case does not know that a barn is in front of her) reliable. Here is another
way to make the very same point: if the same capacities underlie everyday
judgments and the judgments elicited by thought experiments, it might
seem that one cannot challenge the reliability and thus trustworthiness of
the latter judgments without challenging the reliability and thus trustworthiness of all our judgmentsa price too high to pay for even the
most ardent critics of thought experiments. I will call this argument the
Parity Defense of Thought Experiments.
Williamson, for instance, endorses this epistemological implication of
the Ordinary Judgment Proposal. For example, he defends our judgments
about metaphysical possibilities on the grounds that, rst, the ordinary
cognitive capacity to handle counterfactual conditionals carries with it the
cognitive capacity to handle metaphysical modality and that, second, we
have an overall capacity for somewhat reliable thought about counterfactual possibilities (2007, 136).
2. Thought Experiment Skepticism
I agree with Williamsons dismissal of the Spooky Judgment Proposal. As
he puts it, The postulation by philosophers of a special cognitive
capacity exclusive to philosophical argument or quasi-philosophical
thinking looks like a scam (2007, 136). So, in the remainder of this
article, I will focus on the implications of the Ordinary Judgment
r 2011 The Author
197
Proposal for the use of thought experiments in philosophy. Far from

undermining the skepticism about thought experiments, the Ordinary
Judgment Proposal in fact has skeptical implications: it suggests that the
judgments elicited by many philosophical thought experiments fail to
provide evidence for the premises of philosophical arguments. Thus, if the
Ordinary Judgment Proposal is correct, thought experimenting (as a
source of evidence for the premises of philosophical arguments) should be
stringently limited in philosophy.7
2.1. Problem 1: Lack of Relevant Psychological Capacities
The rst (and least important) criticism of the Parity Defense of Thought
Experiments begins by noting that some judgments elicited by philosophical thought experiments have no counterparts in everyday life. It infers
that these judgments are not underwritten by psychological capacities
used in ordinary life, just as those expert judgments that have no
counterparts in everyday life (e.g., the judgments made by radiologists
or by archaeologists) are not underwritten by psychological capacities
used in everyday life.8 The Parity Defense fails to uphold the epistemic
status of this class of judgments elicited by thought experiments.
Similarly, nobody would think of justifying radiologists judgments about
cancerous nodule sites by alleging that they are underwritten by capacities
also used in everyday life.
Consider for instance Burges famous arthritis thought experiment (Burge
1979). Burge describes an individuallets call him Oscarwho is
convinced that he has arthritis in his thigh. Burge then asks the reader to
imagine a situation that is almost identical to Oscars situation. In this
second situation, Oscars counterpart also says that he has arthritis in his
thigh. The only difference between the two situations is in the language
spoken in the linguistic community of Oscars counterpart: in the English
spoken in this second situation, arthritis applies to ailments in the ankles
and to ailments in the thigh. The reader is invited to share Burges judgment
that Oscar, but not his counterpart, has thoughts about arthritis: In the
counterfactual situation, the patient lacks someprobably allof the
attitudes commonly attributed with content-clauses containing arthritis in
oblique occurrence. He lacks the occurrent thoughts of beliefs that he has
arthritis in the thigh, that he has had arthritis for years, that stiffening joints
and various sorts of aches are symptoms of arthritis, that his father had
7
Note that the arguments presented in section 2 do not exclude the possibility that some
thought experiments succeed in providing evidence for the premises of philosophical
arguments. But, as we will see, this kind of thought experiments is unlikely to be of great
use to philosophers.
8
Of course, in a sense, these judgments rely on everyday cognitive competences, such as
categorization, visual recognition, and so on. Nonetheless, they do not rely on the concepts
and recognition procedures we use in everyday life.
r 2011 The Author

198
EDOUARD MACHERY
arthritis, etc. (Burge 1979, 78). Burge concludes that what thoughts one has
(viz., what concepts one possesses) depends on social facts. The kind of
judgment elicited by Burges thought experiment (viz., the judgment that
Oscar and his counterpart have different thoughts) has plausibly no counterpart in everyday life: it is dubious whether people make individuating
judgments about thoughts. Indeed, it is plausible that lay people would
not know how to answer if presented with Burge-style thought experiments
(for consistent evidence, see Hewson 1994; Genome and Lombrozo unpublished manuscript).
2.2. Problem 2: The Unreliability of Some Psychological Capacities
No doubt, some judgments elicited by thought experiments do have
counterparts in everyday life, and they are underwritten by psychological
capacities used in everyday life. The second criticism of the Parity Defense
asserts that we have reasons to doubt that some of these capacities are
reliable even in everyday life (or that they are reliable to a sufciently high
degree for the resulting judgments to count as evidence). When this is the
case, the fact that the same psychological capacities underwrite everyday
judgments and the judgments elicited by thought experiments provides no
reason to believe that the latter are reliable (or sufciently reliable), and
the Parity Defense fails to undermine the skepticism directed at philosophical thought experiments. Similarly, the assessment of the quality of
New World wines by a (nonexpert) Frenchman probably relies on the
same competence he uses to assess French wines. But this fact provides no
reason to think that his judgments about the quality of New World wines
are correct because lay assessment of wine quality in general is mediocre
(Goldstein et al. 2008).
Everyday judgments about causation in the social domain illustrate
this second objection to the Parity Defense. Social psychologists have
shown that judgments about causation are systematically biased when
they are applied to morally salient actions. For instance, Mark Alicke
(1992, 2000) has shown that peoples desire to blame an agent can
inuence their judgment about her causal contribution to the production
of an outcome. Consider the following vignette:
John was driving over the speed limit (about forty mph in a thirty-mph zone) in
order to get home in time to hide an anniversary present for his parents that he
had left out in the open before they could see it. As John came to an
intersection, he failed to see a stop sign that was covered by a large branch.
As a result, John hit a car that was coming from the other direction. He hit it
on the drivers side, causing the driver multiple lacerations, a broken collar
bone, and a fractured arm. John was uninjured in the accident.
Consider another vignette that is identical to the previous one except that
the rst sentence now reads as follows:
r 2011 The Author
199
John was driving over the speed limit (about forty mph in a thirty-mph zone) in
order to get home in time to hide a vial of cocaine he had left out in the open
before his parents could see it.
People tend to judge that the agent has made a greater causal contribution to the outcome in the second vignette than in the rst.
More generally, peoples negative evaluation of agents, due among
other things to the biases they might harbor against minorities, inuences
their causal judgments about the causal contribution of these agents to
outcomes. David Rose and I presented participants with one of two
vignettes (unpublished data). The rst vignette was the following:
An intern is taking care of a patient in a hospital. The intern notices that the
patient is having some kidney problems. Recently, the intern read a series of
studies about a new drug, Rascalis, that can alleviate problems like this one,
and he decides to administer the drug in this case.
Before the intern can administer the drug, he needs to get the signature of the
pharmacist (to conrm that the hospital has enough of the drug in stock) and the
signature of the attending doctor (to conrm that the drug is appropriate for this
patient). So he sends off requests to both the pharmacist and the attending doctor.
The pharmacist, John Doughty, receives the request. After looking at the
request, John remembers hearing about the new drug. Recently, John and his
boyfriend had erotic sex. After having sex, they were lying naked in bed
together, softly tongue kissing one another. While they were tongue kissing,
Johns boyfriend was suddenly reminded about a new drug and stopped kissing
John to tell him about it. He told John that he had recently read about
Rascalis, a fascinating new drug that can treat kidney ailments. Having been
reminded of this at the pharmacy, John checks to see that the drug is in stock
and immediately signs off on the request.
The attending doctor, Frank Montgomery, receives the request at the same
time and immediately realizes that there are strong reasons to refuse it.
Recently, Frank was having lunch with a bright, young intern from a local
university, and the intern told him that although some studies show that
Rascalis can help people with kidney problems, there are also a number of
studies showing that it can have very dangerous side effects. For this reason,
the hospital has a policy forbidding the use of this drug for kidney problems.
Despite this policy, the doctor decides to sign off on the request.
Since both signatures were received, the drug is administered to the patient.
As it happens, after receiving the drug, the patient recovers from his kidney
ailment, and the drug has no adverse effects.
The second vignette was identical, except for the second and third
paragraphs:
The pharmacist, John Doughty, receives the request. After looking at the
request, John remembers hearing about the new drug. Recently, John was
having lunch with a bright, young intern from a local university, and the intern
told him that he had recently read about Rascalis, a fascinating new drug that
r 2011 The Author
200
EDOUARD MACHERY
can treat kidney ailments. John checks to see that the drug is in stock and
immediately signs off on the request.
The attending doctor, Frank Montgomery, receives the request at the same
time and immediately realizes that there are strong reasons to refuse it.
Recently, Frank and his boyfriend had erotic sex. After having sex, they
were lying naked in bed together, softly tongue kissing one another. While they
were tongue kissing, Franks boyfriend was suddenly reminded about a new
drug and stopped kissing Frank to tell him about it. He told Frank that he
recently read about a new drug, Rascalis, that can treat kidney ailments. He
went on to say that although some studies show that Rascalis can help people
with kidney problems, there are also a number of studies showing that it can
have very dangerous side effects. For this reason, the hospital has a policy
forbidding the use of this drug for kidney problems. Despite this policy, the
doctor decides to sign off on the request.
Participants were then asked the following two questions:

(1) To what extent do you think that the actions and decisions of
the attending doctor (Frank Montgomery) caused the patients
recovery?
(2) To what extent do you think that the actions and decisions of
the pharmacist (John Doughty) caused the patients recovery?
In the rst vignette, the doctor was judged to be more causally responsible
for the recovery, while in the second vignette the pharmacist was judged
to be more causally responsible. People seem to assume that the
individuals they judge to be bad (or against whom they harbor prejudices)
can do no good.
The upshot should be clear: everyday causal judgments in the social
domain are biased, and they are unlikely to be reliable (or reliable to a
sufcient degree for them to be treated as evidence).9 Thus, causal
judgments elicited by thought experiments provide no evidence for the
premises of philosophical arguments when the judgments bear on whether
an agent caused an outcome.
How often are the judgments elicited by thought experiments underwritten by psychological capacities that are not reliable (or not reliable
enough) in everyday life? It is obviously hard to know, but it is worth
noting that various philosophical views entail that such a situation is not
so rare. For instance, moral error theorists (see, e.g., Mackie 1977 and
Joyce 2006) hold that moral judgments are false, and eliminativist
9
One might object that the existence of such biases does not show that causal judgments
in the social domain are unreliable (or even insufciently reliable). After all, vision suffers
from biases, but it isnt unreliable in most circumstances. This objection, however, does not
take into account what has been learned about causal judgments in the social domain. When
people are making causal judgments in the social domain, they are typically more concerned
with apportioning blame, praise, and responsibility than with getting causal relations right.
r 2011 The Author

201
materialists (see, e.g., Churchland 1981 and Stich 1983) hold that
ascriptions of propositional attitudes are false.
2.3. Problem 3: The Atypicality of Thought Experiments
I now turn to the main criticism of the Parity Defense. To start with,
whether one is able to do something or to bring about an outcome typically
depends on the circumstances. For instance, whether one is good at
shooting a handgun at a paper target depends, among other things, on
how far and how wide the target is. Ones reliability might be high for eightinch bulls-eye targets at twenty-ve yards, but much worse for six-inch
targets at fty yards. That the reliability of a skill or capacity depends on the
circumstances in which this skill or capacity is applied is equally true of our
physical skills and of our psychological capacities. One can be very good at
mental division if the numbers are within a certain range, but quite poor
outside this range. Similarly, the reliability of a radiologists capacity to
recognize cancerous nodule sites in X-rays depends on the quality of the Xrays. If the quality of an X-ray is poor, perhaps because the scanner is not
functioning properly, the radiologist will be unreliable. I will say that the
circumstances in which a physical or psychological capacity is reliable form
the proper domain of this capacity.
If we have reason to suspect that a physical skill or a psychological
capacity is applied outside its proper domain, our condence in the success
of this application should decrease. If we have no further information about
the circumstances in which the skill or capacity is applied or about how these
circumstances impact its reliability, then, for all we know, the reliability of
the skill or capacity in this particular application might be almost as high as
it is in its proper domain, or it might be very low: that is, we have a reason to
believe that its reliability is lower than in its proper domain, but we do not
know how low it is. In these conditions, we should be reluctant to express
much condence in the success of the application of the skill or capacity.
To see the point, consider the following analogy. Because we know
that shooting accuracy decreases when people shoot at targets that are
more distant and smaller than they are used to, we should be less
condent in the accuracy of a shooter who has been trained in shooting
at twelve-inch targets at ten yards if we know that the target is smaller
than twelve inches and more distant than ten yards. If we do not know
how distant or how small the target is or if we do not know how size and
distance affect her reliability, for all we know, the shooter might be very
unreliable (if the target is very far or is very small) or quite reliable (if, say,
she is shooting at a ten-inch target at fteen yards). In these conditions,
we should be reluctant to accept a bet if the odds favor the shooter hitting
the target.
This point bears on the Parity Defense. Suppose that the judgments
elicited by thought experiments and everyday judgments are underwritten
r 2011 The Author
202
EDOUARD MACHERY
by the same psychological capacities and that these psychological

capacities are reliable in everyday circumstances. Even so, the judgments
elicited by thought experiments in general, or perhaps by some kinds of
thought experiment, might be less reliable (perhaps much less): the
situations described by thought experiments might be beyond the proper
domains of the relevant psychological capacities. If we have a reason to
suspect that these psychological capacities are used beyond their proper
domain and if we have no further information about their reliability in
these circumstances, then we should not put much stock in the resulting
judgments and, for this very reason, should not treat them as evidence.10
Suppose for instance that I am quite good at determining whether an
action is permissible in my everyday life and that, as the Ordinary Judgment
Proposal would have it, my judgments about the permissibility of actions in
thought experiments are underwritten by the very capacity that underlies
my everyday judgments. For all that, my judgment that in the footbridge
case it is not permissible to push the large man to stop the runaway trolley
and save ve people might still fail to provide evidence that it is not
permissible to push the large man. We might have a reason to believe that
this type of dilemma is beyond the proper domain of our capacity to make
judgments about the permissible: while it is reliable in everyday situations,
we might have a reason to believe that our capacity is less reliable when
confronted with this type of dilemma. If we have no further information
about its reliability in this type of dilemma, then we should not treat the
resulting judgment as providing evidence for the fact that in the situation
described by the footbridge case it is not permissible to push the large man.
There are reasons to suspect that in many philosophical thought
experiments the situations described are beyond the proper domains of
the underlying psychological capacities. Thought experiments in philosophy typically describe fanciful situations that are very remote from the
situations that elicit everyday judgments, and thus we cannot rely on our
memories of tried and true past judgments in everyday situations.11
Situations are also often described in vivid terms (see, for example,
Thomsons description of the trolley cases above), and their description
10
Weinberg (2007 and personal communication) doubts that attacking the reliability of
the judgments elicited by thought experiments is a successful strategy for defending the kind
of skepticism endorsed in this article because evidence fails to establish that these judgments
are unreliable. Such evidence, however, is not required by the present argument. We should
not put stock in the judgments elicited by thought experiments if we have reason to believe
that the situations described are beyond the proper domains of the relevant capacities and if
we have no further information.
11
One could object that we are used to fanciful situations in, for example, science ction
or heroic fantasy novels and movies and that we regularly make judgments about them. At
an abstract level, however, the situations described in, for example, science ction novels are
clearly very similar to everyday situations, and we thus have reasons to believe that they
belong to the proper domains of the relevant psychological capacities. If they are not similar
in this way, then we should suspect that they do not belong to the relevant proper domains.
r 2011 The Author

203
contains numerous irrelevant narrative elements, features that are known

to bias judgments. Most important, thought experiments typically pull
apart the features that go together in everyday life: for instance, using
physical violence and doing more harm than good are pulled apart in the
footbridge case (see below for other examples). In such circumstances,
peoples judgments tend to be less reliable because, rst, the eliciting
situations are less typical (Mccloskey and Glucksberg 1978).12 Second,
they are also likely to be less reliable when they are underwritten by heuristics
(Sunstein 2005). Heuristics are judgmental procedures that use the presence
of one or a few features to produce a judgment. Sometimes, these features are
merely coincidental to the truth of the judgment. For instance, Tversky and
Kahneman (1983) have shown that whether an individual is typical of a class
is often used to judge whether he is likely to belong to this class. In everyday
life, heuristics are reliable because the features they rely on co-occur with
whatever it is that makes the resulting judgments true. This co-occurrence is
likely to be disrupted when the features that go together in everyday life are
pulled apart.
All these features of philosophical thought experiments should lead us
to expect the elicited judgments to be less reliable than the counterpart
judgments made in everyday life. If we have no further information about
their reliability, the judgments elicited by thought experiments having
these features cannot be treated as providing evidence for the premises of
philosophical arguments, and we should be reluctant to base arguments
on these judgments.
It is worth highlighting the fact that the skeptical argument against
thought experiments developed in this section does not claim that, unless
the proponent of thought experiments can provide evidence for the
reliability of the capacities underlying the judgments elicited by thought
experiments, we should not treat these judgments as evidence. Rather, the
argument is that we have positive reasons to think that the situations
12
One could object that the judgments elicited by at least some thought experiments
(e.g., Gettier cases and the footbridge version of the trolley case) do not display the features
that are characteristic of unreliable judgments in Mccloskey and Glucksberg (1978):
disagreements among individuals and inconsistencies within individuals across occasions.
In response, rst, it is worth restating that some thought experiments probably do not
impugn the reliability of the capacities underlying judgments. One would expect these
thought experiments to elicit stable, consensual judgments. Second, many thought experiments give rise to disagreements (e.g., Twin Earth, zombies, etc.). Finally, the stability and
consensus of judgments among philosophers should offer little solace to the proponents of
thought experiments. First, this stability and consensus might often be due to some form of
conformism: philosophers judgments about, for example, the Godel case might be
inuenced by the judgments publicly stated of other (perhaps high-status) philosophers.
Various contingent factors (the prestige of the inventor of the thought experiment, whether a
thought experiment was invented early on in a philosophical debate, etc.) might determine
whether a thought experiment elicits conformist judgments. Second, as Cummins (1998)
noted, people who do not share the relevant judgments may be less likely to become
philosophers.
r 2011 The Author

204
EDOUARD MACHERY
described by thought experiments do not belong to the proper domains of

the capacities underlying the judgments they elicit. That is, we have reasons
to believe that these judgments are less reliable, and if we have no further
information about their reliability, we should not treat them as evidence.13
Examples abound of thought experiments displaying the features just
described. Lets suppose that we are good at identifying what a proper name
refers to in everyday life: we correctly judge that Barack Hussein Obama
refers to Barack Obama, and that David Lewis refers to David Lewis.
Even so, it is dubious that the judgment that in the Godel case Godel
refers to the man originally called Godel provides any evidence that
Godel refers to the man originally called Godel. Situations involving
proper names associated with a single description that happens to be false of
the original bearer of the name are probably beyond the proper domain of
our capacity to identify the reference of proper names since in everyday
circumstances many of the numerous descriptions associated with a proper
name tend to be true of the original bearer of the name.
It is even questionable whether the judgment usually elicited by Gettier
casesviz., the agent fails to know the relevant propositionprovides
evidence that in the circumstances described by Gettier cases the agent fails
to have any knowledge. When knowledge is ascribed or denied in everyday
life, truth, justication, and the reliability of the belief-forming method go
hand in hand. When people fail to know something, their beliefs are
typically false, unjustied, and the products of unreliable methods. When
people know something, their beliefs are typically true, justied, and the
products of reliable methods. By contrast, Gettier cases sever truth and
justication from the reliability of the methods of belief formation since
they describe situations where truth comes about by luck.14 Thus, one has a
reason to believe that the situations described by Gettier cases are beyond
the proper domain of our everyday capacity to ascribe knowledge.
13
Not everybody will be convinced that the features of thought experiments discussed in
this article provide reasons to question the reliability of the judgments they elicit. Others will
doubt that particular thought experiments (e.g., Gettier cases) have such features. First, I am
happy to concede that more research is needed to understand which features of thought
experiments may undermine the reliability of the judgments they elicit. Second, the burden of
proof does not fall on the skeptics alone: proponents of thought experiments and thoughtexperiment skeptics should be equally concerned with the reliability of the methods
philosophers use. In other disciplines, when the reliability of a method is questioned, critics
and users of this method attempt to determine its value. Third, I am afraid that no amount of
evidence would convince some proponents of thought experiments. People tend to have an
exaggerated impression of the epistemic value of their judgments, and they are too often
impervious to contrary evidence: for instance, after dozens of experiments showing the
superiority of actuarial methods (Grove et al. 2000), physicians are still reluctant to prefer
them over their own clinical judgments.
14
Here the method is not the tendency to endorse ones perceptual experiences (which is
a reliable method) but the use of a broken clock.
r 2011 The Author

205
The same basic point could be made about Putnams Twin Earth
thought experiment, ssion and fusion scenarios in metaphysics, most
variants of the trolley case and various other thought experiments in
ethics, the truetemp case in epistemology, and many other thought
experiments that have been inuential in philosophy.
It is unlikely to be an accident that many thought experiments in
philosophy have some of the features described above (fanciful situations
that pit against one another the features that co-occur in everyday life).15
Everyday judgments (about permissibility, personal identity, reference, and
so on) often fail to distinguish between competing philosophical theories:
these theories are all compatible with the truth of these judgments. The
proponent of the Ordinary Judgment Proposal thus faces the following
problem. Thought experiments are likely to be most useful for philosophical
purposes when they describe situations that differ from the situations that
elicit everyday judgments, but these situations are precisely those where the
Parity Defense fails. In these circumstances, that the judgments elicited by
thought experiments are underwritten by the very capacities that underlie
everyday judgments provides no justication for using these judgments as
evidence for premises of philosophical arguments.
It is also important to note that the fact that on occasion a real-life
situation is similar to the situation described by a thought experiment does
not ensure that the situation belongs to the proper domain of the relevant
psychological capacity. For instance, Williamson reports in The Philosophy
of Philosophy (2007, 192) that he occasionally brings about the conditions
characteristic of a Gettier case in the classroom, and that he then elicits
judgments from his students (see also Williamson 2011). This fact alone
does not ensure that the judgment elicited by a Gettier case can be used for
philosophical purposes. What is needed is some positive reason to think
that, despite the peculiar features of Gettier cases (see above), the situations
they describe nonetheless belong to the proper domain of the psychological
capacity involved in ascribing knowledge to oneself and others.
To conclude, the proponent of the Ordinary Judgment Proposal
should recommend a skeptical attitude toward the judgments elicited by
thought experiments when the situations that thought experiments
describe have features known to reduce the reliability of judgments.
Many thought experiments in philosophy have such features, plausibly
because the kinds of thought experiment that are most useful for
philosophers are likely to have these features. As a result, the Ordinary
Judgment Proposal leaves very little room for thought experiments to
support the premises of philosophical arguments.
15
The use of vivid materials, however, seems to be only an accidental element of thought
experiments in philosophy.
r 2011 The Author

206
EDOUARD MACHERY
2.4. Circumscribed Skepticism

The discussion in this section shows that, far from stiing skepticism, the
Ordinary Judgment Proposal leads us to view with skepticism many
judgments elicited by thought experiments. This skepticism differs from
the types of skepticism usually considered by philosophers (contrast with
Williamson 2007, 222). Skeptical scenarios (I am a brain in a vat, an evil
demon is fooling me, I am dreaming, and so on) raise a possibility that has
two features: it is inconsistent with my actual beliefs (that I have hands, and
so on), and the evidence that is non-question-beggingly available to
distinguish between this possibility and my actual beliefs (for instance, my
perceptual experiences) fails to support the latter over the former. In usual
skeptical scenarios, no reason is presented that supports the skeptical
scenario over our actual beliefs; rather, the skeptic challenges the nonskeptic
to show that the possible scenario she envisages is not actual. By contrast,
the skepticism directed at thought experiments is supported by positive
reasons: people dont apply some concepts of philosophical interest in their
everyday life, suggesting that they do not have the relevant psychological
capacities; empirical evidence suggests that the ordinary application of such
concepts is biased and unreliable (or not sufciently reliable); or the thought
experiments describe situations that are beyond the proper domains of the
underlying psychological capacities.
One common response to skepticism is to refuse to engage with it (e.g.,
Williamson 2007, 239): the nonskeptic holds that there is no reason to
grant that the available evidence reduces to what the skeptic takes it to be
(e.g., our perceptual experiences). While this response might be defensible
for the usual kinds of skepticism, it is inappropriate for the skepticism
about thought experiments defended here because of the peculiarities of
this skepticism. The thought-experiment skeptic adduces positive reasons
to doubt the judgments elicited by thought experiments. The proponent
of thought experiments must explain why, appearances notwithstanding,
these reasons do not undermine the reliability of the judgments elicited by
thought experiments.
Finally, the skepticism defended here is satisfyingly circumscribed (see
Williamson 2004 about this concern): it does not generalize to everyday
judgments, for the positive reasons to doubt the reliability of the
judgments elicited by many philosophical thought experiments do not
carry over to everyday judgments.
3. Philosophers Expertise
3.1. The Expertise Reply
Proponents of the Ordinary Judgment Proposal have a card up their
sleeves: they can respond that philosophers expertise undermines the
arguments developed in the previous section. Even if ordinary people
r 2011 The Author
207
have no capacity to make judgments about, for example, propositionalattitudes individuation, philosophers might have acquired such a capacity, and they might use it in ordinary and philosophical contexts.
Similarly, radiologists have acquired the capacity to identify cancerous
nodule sites in X-rays. Even if the psychological capacity underlying some
kind of judgments (e.g., causal judgments in social contexts) is not reliable
among ordinary people, philosophers judgments might be reliable
because the philosophers have thought a lot about the relevant issue
(e.g., causation), because they are more cautious in making their judgments, and so on. Similarly, experts wine assessment is much more
reliable than ordinary peoples because experts have been trained to assess
wines. Even if the proper domain of a psychological capacity does not
extend to the situations described by thought experiments among
ordinary people, the proper domain of this capacity among philosophers
might be much broader. Thus, philosophers judgments about the
reference of proper names in the Godel case and about what is permissible
in the convoluted trolley cases might be reliable. Similarly, while people
are often good at shooting a handgun at a twelve-inch bulls-eye paper
target at fteen yards but poor when they shoot at six-inch targets at fty
yards, experts might be reliable in the latter circumstances too.
Williamson (2007, 2009, and 2011) seems to endorse this line of
reasoning (see also Ludwig 2007 and Devitt forthcoming). He writes:
Much of the evidence for cross-cultural variation in judgments on thought
experiments concerns verdicts by people without philosophical training. Yet
philosophy students have to learn how to apply general concepts to specic
examples with careful attention to the relevant subtleties, just as law students
have to learn how to analyze hypothetical cases. Levels of disagreements over
thought experiments seem to be signicantly lower among fully trained
philosophers than among novices. That is another manifestation of the
inuence of past experience on epistemological judgments about thought
experiments. (2007, 191)
Lets call this reply the Expertise Reply. The Expertise Reply is in the
spirit of the Ordinary Judgment Proposal since it asserts that the same
capacities underlie the judgments elicited by thought experiments and
those elicited by everyday situations. Thought experiments elicit judgments that do not differ in kind from the judgments one makes in
everyday contexts. It differs from the Ordinary Judgment Proposal in
insisting that philosophers judgments about the situations described by
thought experiments are more reliable than lay peoples (or perhaps that
only the former are reliable).
There is no doubt that philosophers possess some kind of expertise,
and the argument developed here against the Expertise Reply does not
depend on challenging this truism (contrast with Williamson 2009).
Rather, the argument depends on the following point (for a different
r 2011 The Author
208
EDOUARD MACHERY
line of argument, see Weinberg et al. 2010): Appealing to philosophers

expertise can stie the circumscribed skepticism put forward in section 2
only if this expertise increases the reliability of the judgments philosophers make about the situations described in thought experiments. But
philosophers expertise does not seem to improve the reliability of the
judgments about reference, causation, the right action, knowledge, and so
on, elicited by thought experiments (or at least does not improve it
sufciently for these judgments to provide evidence for the premises of
philosophical arguments). If philosophical expertise exists, it consists in
something else.
3.2. The Princess-and-the-Pea Illusion
Are philosophers better than ordinary people at identifying causes, judging
whether an action is permissible, determining what a proper name refers to,
assessing whether an agent knows some proposition, and so on? Many
philosophers would like to think so. For instance, Kamm writes:
Having responses to complex and unfamiliar cases requires that one see a
whole complex landscape at once, rather than piecemeal. This often requires
deep concentration. Only a few people may be able to respond to a complex
case with a rm response. . . . The princess and the pea is the fairy tale best
associated with the method I describe: it tells of someone, despite much
interference, who cannot ignore a slight difference that others may never sense.
(1993, 8)
While Devitts views about intuitions are quite different from Kamms,
Devitt agrees with Kamm about the superiority of philosophers judgments (for discussion, see Machery forthcoming):
[T]he normal competent speaker with even a little education does reect on
linguistic reality just as she reects on many other striking aspects of the world
she lives in. And this education will usually provide her with the terms and
concepts of folk semantics, at least. As a result she is likely to be able to judge
in a fairly immediate and unreective way what an expression refers to . . . .
Still, are these referential intuitions likely to be right? I think we need to be
cautious in accepting them: semantics is notoriously hard and the folk are a
long way from being experts. Still it does seem to me that their intuitions about
simple situations are likely to be right. This having been said, we should
prefer the intuitions of semanticists, usually philosophers, because they are
much more expert (which is not to say, very expert!). (Devitt forthcoming)
Williamson suffers from the same illusion, which I will call the Princessand-the Pea Illusion. He writes: [T]he expertise defence [what I call the
Expertise Reply] does not imply that a good philosophical education
involves the cultivation of a mysterious sui generis faculty of rational
intuition, or anything of the kind. Rather, it is supposed to improve far
more mundane skills, such as careful attention to details in the description
r 2011 The Author
209
of the scenario and their potential relevance to the questions at issue

(Williamson 2011).
3.3. Against the Princess-and-the-Pea Illusion
I will now present some evidence that casts doubt on the idea that
philosophers intuitions are so reliable that they can serve as evidence for
premises of philosophical arguments.
Plausibly, if the judgments of philosophers that are elicited by thought
experiments are more reliable than those of lay people, then their
everyday judgments about the relevant topics (causation, responsibility,
and so on) are also more reliable (assuming reasonably that lay peoples
judgments are not perfectly reliable). That is, for instance, philosophers
everyday judgments about the permissible should be more reliable than
lay peoples. It would indeed be strange (although not inconceivable) if
the improved reliability of philosophers judgments were limited to the
situations described by thought experiments. It is also reasonable to
assume that, if philosophers judgments are more reliable than ordinary
peoples because of the expertise philosophers have acquired, the judgments philosophers make about their particular area of expertise should
tend to be more reliable than the judgments of philosophers working on
other areas of philosophyfor instance, ethicists judgments about
ethical matters (what is right or wrong, what is permissible, what is
morally required, and so on) should be more reliable than metaphysicians. It is also reasonable (although not uncontroversialsee below) to
assume that, if ethicists judgments about ethical matters are more reliable
than other philosophers, then ethicists actions should be better than
other philosophers, since ethicists judgments have practical signicance.
In recent years, Eric Schwitzgebel has accumulated a large body of
evidence suggesting that ethicists do not behave better than other philosophers. Moral philosophers are 50 percent more likely to borrow books
permanently from libraries than other philosophers (Schwitzgebel 2009).
That is, moral philosophers are 50 percent more likely to steal books from
libraries! Moral philosophers are also not more likely to abide by elementary norms of politeness (such as replying to e-mail or behaving properly in
conferences) than other philosophers (Schwitzgebel forthcoming). Finally,
moral philosophers, including political philosophers, are not more likely to
vote than other philosophers (Schwitzgebel and Rust 2010). Unsurprisingly,
philosophers tend to think that ethicists do not behave better than other
philosophers (Schwitzgebel and Rust 2009)!
I view this growing body of ndings as indirect evidence that ethicists
judgments about ethical matters are not more reliable than other
philosophers, which casts doubt on the idea that philosophical expertise
improves the reliability of the judgments elicited by thought experiments.
Of course, one could challenge this interpretation of Schwitzgebels
r 2011 The Author
210
EDOUARD MACHERY
ndings on various grounds. The connection between ethicists actions

and the reliability of the judgments elicited by thought experiments is
admittedly indirect. Furthermore, the quality of philosophers judgments
need not be reected in their actions. After all, moral philosophers might
fail to act on their enlightened judgments for a variety of reasons,
including weakness of the will, failure to pay attention to their own
judgments when they act, and so on.
There is, however, some more direct evidence challenging the Expertise
Reply. Schwitzgebel and Cushman (unpublished manuscript) gave participants both the bystander and footbridge trolley cases in two different
orders, and they examined whether the order of presentation inuenced
lay peoples and philosophers judgments. You would think that, if the
judgments of philosophers that are elicited by thought experiments are
more reliable, they would be less inuenced by the order of presentation.
It turns out, however, that philosophers, including ethicists, are as
inuenced as lay people by the order of the cases.
You would also think that, if the judgments of philosophers that are
elicited by thought experiments are more reliable, philosophers theoretical commitments would not have a large inuence on their judgments.
To examine this question, I gave the Godel case to linguists and
philosophers of language (Machery forthcoming; for further discussion
of the relation between expertise and linguistic intuitions, see Culbertson
and Gross 2009; Machery and Stich forthcoming). It turns out that the
judgments of linguists and philosophers of language about the reference
to Godel in this case are inuenced by their disciplinary training:
linguists who work in elds that highlight the descriptions associated with
words (sociolinguistics, historical linguistics, anthropological linguistics)
are less likely to have Kripkean intuitions than those linguists who are
likely to have read Naming and Necessity during their training (semanticists and philosophers of language).16
Current evidence tentatively suggests that the Princess-and-the-Pea
Illusion is just thatan illusion. Admittedly, the ndings reviewed in this
section fall short of establishing beyond doubt that philosophers expertise does not consist in part in an increased reliability of their
judgments about causation, the permissible, and so on. However, they
cast sufcient doubt on this idea to prevent philosophers from simply
gesturing at their expertise when they attempt to undermine the skepticism about thought experiments. Philosophers who want the justication
of thought experiments to hang on the contribution of their expertise to
the reliability of their judgments should put up or shut up: evidence is
called for.
16
Or perhaps the former disciplines attract more students with descriptivist intuitions
and the latter more students with Kripkean intuitions!
r 2011 The Author

211
3.4. What Does Philosophers Expertise Consist In?

Philosophers often make (occasionally pompous) claims about the nature
of their expertise, and they are fond of supporting them with anecdotes.
But the truth is we know very little about the nature of philosophers
expertiseviz., about what philosophers are distinctively good at, about
whether and how this expertise is acquired, and so on. The reason for this
ignorance is that there have been very few empirical, systematic studies of
these questions: the psychology of philosophy remains underdeveloped.
As we have seen, evidence suggests that expertise might not consist in a
capacity to make more reliable judgments. On the other hand, there is no
doubt that philosophers acquire some expertise. What does it consist in?
In recent work, Livengood, Sytsma, Feltz, Scheines, and Machery (2010)
have started to examine what we call the philosophical temperament.
By this we mean the characteristics that are distinctive of philosophers.
Participants were presented with the Cognitive Reection Test through
the Philosophical Personality website (philosophicalpersonality.com).
The idea behind the Cognitive Reection Test is to present participants
with questions that have intuitive answers that are incorrect. To arrive at
the correct answer requires participants to move beyond the answer that
initially comes to mind and to consciously reect on the problem instead.
The Cognitive Reection Test consists of three questions of this sort:
(1) A bat and a ball cost $1.10 in total. The bat costs $1.00 more than the ball.
How much does the ball cost? _______ cents.
(2) If it takes 5 machines 5 minutes to make 5 widgets, how long would it take
100 machines to make 100 widgets? _______ minutes.
(3) In a lake, there is a patch of lily pads. Every day, the patch doubles in size.
If it takes 48 days for the patch to cover the entire lake, how long would it take
for the patch to cover half of the lake? _______ days. (Frederick 2005, 27)
We collected data on 4,472 participants, 823 of whom reported having at

least some philosophical training. We found that training in philosophy
positively correlates with Cognitive Reection Test score even when
controlling for education level. Thus, if we consider each level of
education separately (with the exception of participants with a vocational
or trade degree), the mean score for participants with some training in
philosophy is higher than for those without training in philosophy.
These ndings cast light on an important characteristic of the philosophical temperament: philosophers are less likely to blindly accept their
intuitions and more likely to submit those intuitions to scrutiny. Philosophers ponder; they question what spontaneously seems to be the case;
they readily take a skeptical eye toward how things seem to them.
Philosophical expertise is thus real and distinctive (more on this in
Livengood et al. 2010). But, so far as we know, it does not consist in
r 2011 The Author
212
EDOUARD MACHERY
being more reliable at judging whether something is a cause, what a

proper name refers to, what is permissible in specic situations, and so on.
Conclusion
The argument developed in this article does not support an unconditional
form of skepticism toward all thought experiments in philosophy. Rather, it
supports the following conditional skeptical claim about many (but plausibly
not all) philosophical thought experiments: if the inuential, reasonable view
that the judgments elicited by thought experiments and everyday judgments
are underwritten by the same psychological capacities is correct, then many
thought experiments (and in any case the kind of thought experiment that is
plausibly most needed to decide between competing philosophical theories)
cannot be used to support the premises of philosophical arguments. This
skepticism, which is not abated by the appeal to philosophers expertise, is
satisfyingly circumscribed: fortunately, it does not breed any untoward
skepticism against ordinary judgments.
Department of History and Philosophy of Science
University of Pittsburgh
1017CL
Pittsburgh, PA 15260
USA
machery@pitt.edu
Acknowledgments
I would like to thank Armen Marsoobian for organizing the conference
The Future of Philosophy: Metaphilosophical Directions for the TwentyFirst Century. I would also like to thank Joshua Alexander, Joachim
Horvath, Eric Schwitzgebel, and Jonathan Weinberg for their comments
on previous versions of this article.
References
Alicke, M. D. 1992. Culpable Causation. Journal of Personality and
Social Psychology 63:36878.
. 2000. Culpable Control and the Psychology of Blame. Psychological Bulletin 126:55674.
Brown, J. R. 2004. Why Thought Experiments Transcend Empiricism.
In Contemporary Debates in Philosophy of Science, edited by C.
Hitchock, 2343. Malden, Mass.: Blackwell.
Burge, T. 1979. Individualism and the Mental. Midwest Studies in
Philosophy 4:73121.
Churchland, P. M. 1981. Eliminative Materialism and the Propositional
Attitudes. Journal of Philosophy 78:6790.
r 2011 The Author
213
Culbertson, J., and S. Gross. 2009. Are Linguists Better Subjects?

British Journal for the Philosophy of Science 60:72136.
Cummins, R. 1998. Reections on Reective Equilibrium. In Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical
Inquiry, edited by M. DePaul and W. Ramsey, 11327. Lanham, Md.:
Rowman and Littleeld.
Davidson, D. 1987. Knowing Ones Own Mind. Proceedings and
Addresses of the American Philosophical Association 61:44158.
Devitt, M. Forthcoming. Experimental Semantics. Philosophy and
Phenomenological Research.
Foot, P. 1978. The Problem of Abortion and the Doctrine of Double
Effect. In Virtues and Vices and Other Essays in Moral Philosophy,
1935. Berkeley: University of California Press.
Frederick, S. 2005. Cognitive Reection and Decision Making. Journal
of Economic Perspectives 19:2542.
Genome, J., and T. Lombrozo. Unpublished manuscript. Concept Possession, Experimental Semantics, and Hybrid Theories of Reference.
Goldstein, R., J. Almenberg, A. Dreber, J. W. Emerson, A. Herschkowitsch, and J. Katz. 2008. Do More Expensive Wines Taste Better?
Evidence from a Large Sample of Blind Tastings. Journal of Wine
Economics 3:19.
Grove, W. M., D. H. Zald, B. S. Lebow, B. E. Snitz, and C. Nelson. 2000.
Clinical Versus Mechanical Prediction: A Meta-Analysis. Psychological Assessment 12:1930.
Hewson, C. 1994. Empirical Evidence Regarding the Folk Psychological
Concept of Belief. In Proceedings of the Sixteenth Annual Conference
of the Cognitive Science Society, edited by A. Ram and K. Eiselt, 403
6. Hillsdale, N.J.: Erlbaum.
Ichikawa, J., I. Maitra, and B. Weatherson. Forthcoming. In Defense of
a Kripkean Dogma. Philosophy and Phenomenological Research.
Joyce, R. 2006. The Evolution of Morality. Cambridge, Mass.: MIT Press.
Kamm, F. M. 1993. Morality, Mortality, volume 1: Death and Whom to
Save from It. New York: Oxford University Press.
Kripke, S. 1972. Naming and Necessity. Cambridge, Mass.: Harvard
University Press. (Paperbound edition published 1981.)
Livengood, J., J. M. Sytsma, A. Feltz, R. Scheines, and E. Machery. 2010.
Philosophical Temperament. Philosophical Psychology 23:31330.
Ludwig, K. 2007. The Epistemology of Thought Experiments: First
Person Versus Third Person Approaches. Midwest Studies in Philosophy 31:12859.
Machery, E. 2009. Doing Without Concepts. New York: Oxford University Press.
. 2010. Precis of Doing Without Concepts. Behavioral and Brain
Sciences 33:195244.
. Forthcoming. Expertise and Intuitions About Reference. Theoria.
r 2011 The Author
214
EDOUARD MACHERY
Machery, E., and S. P. Stich. Forthcoming. Experimental Philosophy of

Language. In Routledge Companion to the Philosophy of Language,
edited by G. Russell and D. G. Fara. New York: Routledge.
Mackie, J. L. 1977. Ethics: Inventing Right and Wrong. London: Penguin
Books.
Mccloskey, M. E., and S. Glucksberg. 1978. Natural Categories: Well
Dened or Fuzzy Sets? Memory and Cognition 6:46272.
Murphy, G. L. 2002. The Big Book of Concepts. Cambridge, Mass.: MIT
Press.
Norton, J. D. 2004. Why Thought Experiments Do Not Transcend
Empiricism. In Contemporary Debates in Philosophy of Science,
edited by C. Hitchock, 4466. Malden, Mass.: Blackwell.
Schwitzgebel, E. 2009. Do Ethicists Steal More Books? Philosophical
Psychology 22:71125.
. Forthcoming. Ethicists Courtesy at Philosophical Conferences.
Philosophical Psychology.
Schwitzgebel, E., and F. Cushman. Unpublished manuscript. Expertise
in Moral Reasoning? Order Effects on Moral Judgment in Professional
Philosophers and Non-Philosophers.
Schwitzgebel, E., and J. Rust. 2009. The Moral Behaviour of Ethicists:
Peer Opinion. Mind 118:105359.
. 2010. Do Ethicists and Political Philosophers Vote More Often Than
Other Professors? Review of Philosophy and Psychology 1:18999.
Sosa, E. 2007. Experimental Philosophy and Philosophical Intuition.
Philosophical Studies 132:99107.
Stich, S. P. 1983. From Folk Psychology to Cognitive Science. Cambridge,
Mass.: MIT Press.
Sunstein, C. 2005. Moral Heuristics. Behavioral and Brain Sciences
28:53173.
Thomson, J. J. 1985. The Trolley Problem. Yale Law Journal 94:1395
1415.
Tversky, A., and D. Kahneman. 1983. Extensional Versus Intuitive
Reasoning: The Conjunction Fallacy in Probability Judgement.
Psychological Review 90:293315.
Weinberg, J. M. 2007. How to Challenge Intuitions Empirically Without
Risking Skepticism. Midwest Studies in Philosophy 31:31843.
Weinberg, J. M., C. Gonnerman, C. Buckner, and J. Alexander. 2010. Are
Philosophers Expert Intuiters? Philosophical Psychology 23:33155.
Williamson, T. 2004. Philosophical Intuitions and Skepticism About
Judgments. Dialectica 58:10953.
. 2007. The Philosophy of Philosophy. Malden, Mass.: Blackwell.
. 2009. Replies to Ichikawa, Martin and Weinberg. Philosophical
Studies 145:46576.
. 2011 Philosophical Expertise and the Burden of Proof. Included
in this issue.
r 2011 The Author
Copyright of Metaphilosophy is the property of Wiley-Blackwell and its content may not be copied or emailed
to multiple sites or posted to a listserv without the copyright holder's express written permission. However,
users may print, download, or email articles for individual use.

Thought Experiments and Philosophical Knowledge

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Thought Experiments and Philosophical Knowledge

Hochgeladen von

Copyright:

Verfügbare Formate

r 2011 The Author

Metaphilosophy r 2011 Metaphilosophy LLC and Blackwell Publishing Ltd