Professional Documents
Culture Documents
THE SKEPTIC
AND THE VERIDICALIST
Yuval Avnur
Scripps College
Shaftesbury Road, Cambridge CB2 8EA, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467
www.cambridge.org
Information on this title: www.cambridge.org/9781009462297
DOI: 10.1017/9781009243308
© Yuval Avnur 2023
This publication is in copyright. Subject to statutory exception and to the provisions
of relevant collective licensing agreements, no reproduction of any part may take
place without the written permission of Cambridge University Press & Assessment.
When citing this work, please include a reference to the DOI 10.1017/9781009243308.
First published 2023
A catalogue record for this publication is available from the British Library.
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
Elements in Epistemology
DOI: 10.1017/9781009243308
First published online: December 2023
Yuval Avnur
Scripps College
Author for correspondence: Yuval Avnur, yuval.avnur@gmail.com
1 Introduction 1
2 Veridicalism 8
References 50
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
The Skeptic and the Veridicalist 1
Although the whole of this life were said to be nothing but a dream, and the visible
world nothing but a phantasm, I should call this dream or phantasm real enough, if,
using reason well, we were never deceived by it.
Leibniz (1923)
1 Introduction
Skepticism about the external world has been formulated in many ways through
the centuries. It has been taken to concern not only knowledge, but justified
belief, rational degrees of confidence, doubt, and certainty; it has taken the form
of a theory, a challenge, a paradox, a way of life, and an invitation to doubt.
These days, skepticism is usually associated with the question,
Can you know that external, ordinary objects such as tables exist?
This concerns what there is, or what things there are. Sitting at your table, you
look down and see its surface, and feel its pressure under your hand. Is the table
really there, or is it part of a vast computer simulation, or perhaps an idea in the
mind of Descartes’ demon? If you conclude that you don’t know whether some
such skeptical scenario is playing out, then you are supposed to conclude that,
therefore, you don’t know that there is a table there. This is nothing special about
tables, so the conclusion generalizes: you don’t know whether any external objects
exist. This standard formulation ignores the question of what things are, or the
ultimate nature of the things whose existence is in question. The nature of things
such as tables is reasonably taken to be a question of metaphysics or science. In
fact, skepticism calls into question whether the metaphysics and science of ordin-
ary objects even have a known subject matter! The possibility that you are in
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
a small child, or our ancestor in the ancient world perhaps, knows a lot about
what there is. This knowledge is significant to their lives, and they know how to
act in light of this knowledge. And yet, we can imagine, they either have no
beliefs about what things are beyond how they appear (in the case of the child),
or else they have a bunch of false beliefs about this (in the case of our ancestor,
who perhaps thought that all things were ultimately made of water, or fire).
Whatever it is that one worries about when one worries about the skeptical
problem – that we know nothing about the world – surely these worries don’t
apply to the cases we are imagining, as described! Yet, all these people lack, it
seems, is knowledge about what things are. So, how could veridicalism, which
posits that we are roughly in the same position, since we know what there is but
not what things are, possibly fail to satisfy us? It shows that we know at least as
much as the child or the ancestor, and their situation does not seem worthy of
philosophical panic.
The bulk of the arguments in this Element aim to show that this attitude,
though initially appealing, is mistaken. The child’s and the ancestor’s ignorance
about what things are is not the same as skeptical ignorance about what things
are. As we will see, the former is a matter of immaturity (whether of the
individual or of the science of their time). The latter is a matter of necessary,
complete hopelessness. And, as we will see, it is equivalent in important ways to
ignorance about what there is. It is true that not all kinds of ignorance about
what things are constitutes an epistemic disaster. But a particular, skeptical sort
of ignorance about what things are is different. And so, one can know a lot about
what there is while knowing very little about the external world.
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
Whether skeptical hypotheses raise doubts about what there is or what things
are is, as veridicalism helpfully shows, a matter of metaphysics. But if this table
could be anything from a simulation to a demonic idea, then the mere conclusion
that you know that there is a table is still compatible with almost total ignorance
about the external world. As I will suggest, this also serves to clarify what
a skeptical scenario is, and its relation to other metaphysical hypotheses about
the world. Skepticism isn’t really only about knowledge of what there is, rather
it is standardly formulated that way because of a standard metaphysical picture.
Here is an overview of what is to come. In the rest of this seciton, after
I introduce the standard, contemporary skeptical argument, I describe and set
aside some more familiar, nonveridicalist antiskeptical strategies.
In Section 2, I introduce the veridicalist strategy and some of the different
ways it has been pursued. This strategy avoids skepticism about what there is
only by accepting skepticism about what things are. I then distinguish veridic-
alism from some other, externalist and idealist theories, and describe skepticism
about what things are and how veridicalism is committed to it.
4 Epistemology
Sections 3 and 4 are where most of the action (and novel argumentation) occurs.
Section 3 presents arguments for thinking that skepticism about what things
are leaves us with as little knowledge about the world as skepticism about what
there is. One important consequence of this is that the veridicalist antiskeptical
strategy fails – at least, if you were ever worried about standard skepticism
about what there is, then your worry should not be assuaged by veridicalist
skepticism about what things are.
In Section 4 I consider some objections to my arguments: veridicalism at least
grants us knowledge that the world we see around us is reality; skepticism about
what things are is not a radical skepticism at all, but rather mere scientific or
metaphysical humility about the ultimate nature of things; it doesn’t matter what
things ultimately are pragmatically or for other values; the problem of skepticism
is only about the existence of things, and other concerns about the nature of things
are a change of topic. In reply to this last objection, I discuss some of the history of
skepticism, from Sextus to Hume and Moore, to show that the exclusive focus on
the existence of ordinary objects is a relatively recent phenomenon, and that
skepticism about what things are has often been taken to be a concern.
Finally, in Section 5 I describe some more general lessons that can be learned
from the way in which veridicalism fails to solve the skeptical problem. One can
know what there is without knowing much about the world. Some other anti-
skeptical strategies besides veridicalism fail for this reason, and furthermore the
standard argument is inadequate as an expression of the skeptical problem.
I also make some observations about what a successful antiskeptical strategy
would have to look like.
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
and people. But, in fact, there are no such things – at least according to the
standard description of the scenario. Let us focus on such global skeptical
hypotheses, which on a standard metaphysics call all of our knowledge (or
justified beliefs) about the external world into doubt.1 For examples, the hypoth-
eses that your entire life has been like one long dream, that you are a brain in a vat
being stimulated by a computer in an otherwise empty world, and that you are in
the scenario depicted in the film The Matrix since birth are all global skeptical
hypotheses. Let “sk” stand for your favorite such hypothesis:
(I will remind the reader what each of (1)–(4) is each time it is used in a new
section.) This is one standard formulation of the argument, but a few qualifi-
cations are helpful here. First, arguably the dream version of sk raises
a different set of issues, which I address in Section 2. Second, one might
reasonably construe the “standard” skeptical argument as the argument for
(1), leaving the existence of ordinary objects as a further, different matter. Set
those arguments aside here, since our concern is the relation between global
skeptical hypotheses and the existence of ordinary objects. Third, there has
been some discussion in the literature about so-called underdetermination
principles which generate (3) and (4) in an alternative way. Veridicalism can
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
1
I briefly address one veridicalist’s (Chalmers’) view on nonglobal skeptical scenarios in my
review of his book, Avnur (2023).
2
This formulation closely follows Chalmers’ (2018, 1).
3
One recent and helpful discussion of the relation between underdetermination principles and the
sort of closure principles that are typically taken to motivate (1)–(4) is Pritchard (2016, especially
ch. 2). The underdetermination principle states, roughly, that if two propositions are (known to be)
incompatible, then in order to know either one of them one’s evidence must favor that one over the
other. Briefly, the underdetermination version of the skeptical argument is:
(1*) My evidence does not favor the proposition that there are tables over sk.
(2*) In order to know that there are tables, my evidence must favor the proposition that there is
a table over sk. (by the underdetermination principle)
(3) I do not know that there are tables. (from (1*) and (2*))
So,
(4). Generalizing from tables, I do not know that there are any ordinary things.
6 Epistemology
(4) is false still fail to avoid the sort of epistemic disaster that makes (4)
problematic. It is not necessarily enough to show that we know what things
there are.4
Why is (4) such an epistemic disaster? The ignorance (4) posits implies that
your cognitive life, in relation to the world, is a joke: for all you know, you are
thinking about the world and your place in it completely wrong. You don’t
even know that there are tables, and the most you can know is that there are
tableish sensory experiences, which is a nightmarish relation to have to
reality. In other words, (4) makes explicit just how little you know about
the world.
Although I will mostly be setting this aside, external world skepticism is not
all about tables, it’s also about people. As Stroud put it when contemplating the
skeptical conclusion:
You know nothing about things, or people, beyond your sensory experiences.
Though the existence of other people is a central issue for those who worry
about classic skeptical scenarios, as we will see in what follows, there is some
question whether veridicalists mean to be addressing such scenarios at all. If
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
they intend to address only those scenarios in which other people still exist (as in
The Matrix), then they don’t address a significant part of what has been
worrying philosophers about (1)–(4). I will leave this point aside, though, to
see how veridicalism fares even with respect to a narrower set of skeptical
hypotheses, and even setting the existence of other people aside.5
The veridicalist strategy here would presumably be to reject the claim that the proposition that
there is a table is incompatible with sk, and so to maintain that one can know it without possessing
the requisite evidence that favors it over sk. Accordingly, (2*) is not supported by the underdeter-
mination principles after all, and can be rejected. Other strategies may go the route of rejecting the
principle, or rejecting (1*). But the veridicalist, it seems, can accept these and still hold that (2*) is
false. Much of what I say about veridicalist objections to (1)–(4) will apply to this version as well,
and I note that at least two recent proponents of the veridicalist strategy, Valberg and Chalmers,
explicitly target (2) rather than (2*).
4
For the record, Stroud, in a different essay, seemed to agree, at least about the point that
what things there are does not settle what sort of world we are in. See Stroud (1986,
263–264).
5
See Helton (forthcoming).
The Skeptic and the Veridicalist 7
To be sure, some people, and even some philosophers, don’t find (4) so
worrying at all. What I will argue in the next section is consistent with that: if
one finds standard skepticism about what there is disturbing, then one’s worries
should not be assuaged by veridicalism.
Philosophers try to avoid (3) and (4) by rejecting (1) or (2). The most
direct strategy is to reject (1): show that your information is sufficient to
rule out sk. Denying (2) is indirect in the sense that it targets what seems
to follow from our ignorance about sk, rather than sk itself. Though I will
be focusing on a particular kind of indirect approach, and setting aside
direct approaches, the direct approaches will come up again in a couple of
contexts.
First, there are well-known strategies, discussed in the next section and
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
So, our focus is on the indirect strategy, which accepts (1) but rejects (2). But
even here there is plenty of variety, and plenty to set aside. One familiar indirect
strategy is to deny (2) for epistemological reasons. Implicit in the justification
for (2) is the idea that, since the existence of tables implies the falsity of sk
(recall that sk is a global skeptical scenario), knowing that there are tables
implies knowing that not-sk. But if that is the justification for (2), then one way
to avoid commitment to (2) is to avoid commitment to such conditions for
knowledge. Maybe it isn’t always the case that knowing something, p, and
knowing that it entails something else, q, puts you in a position to deduce or
otherwise know q. The considerations involved in this kind of strategy, often
called “denying closure,” are all epistemological, since they involve what is
required to know something. That is, this strategy does not challenge the idea
that if there are tables, then sk is false. It does not challenge the underlying
metaphysics of objects. So, for the most part, we will set these strategies aside.
However, as with direct strategies, these epistemological indirect strategies will
come up again at the end, though this time not as beneficiaries. As we will see,
much of the trouble for veridicalism is also trouble for some epistemological
indirect strategies as well.
2 Veridicalism
In this section I describe a few versions of the veridicalist antiskeptical strategy
and explain its commitment to skepticism about what things are. They each
deny (2):
(2) If I do not know that ~sk, then I do not know that there are tables,
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
on the basis that there can be tables even if sk is true, because what appear to be
tables in any global skeptical scenario are tables. So, accordingly, we don’t need
to know whether or not we are in a skeptical scenario in order to know that there
are tables.6 Crucially, veridicalism is compatible with (1):
Part of the appeal of the view is that it need not bear the burden of explaining
how we know ~sk.
Veridicalism can seem far-fetched. How could something in the mind of
a demon, or in a simulation, be a table? Tables are real things, we want to say,
and while simulations themselves can be real, what they are simulations of are
6
Notice that this does not require rejecting closure for knowledge. For a different take on the
relation of (2) to skeptical hypotheses, see Roush (2010) and Avnur, Brueckner, and Buford
(2011).
The Skeptic and the Veridicalist 9
not, by virtue of being simulated, real. Let us now see how a few different
philosophers motivate an alternative, veridicalist metaphysics.
fact that it has the causal role of a table amounts to the fact that it causes, in some
sense of “cause” that we will grant the structuralist, other demonic ideas to
occur as well as the relevant experiences in the mind of the victim. Oddly, when
you knock on the wood of the table, on this view, you are knocking the demon’s
mind, and when you taste the coffee that is in the cup on the table, you are tasting
the demon’s mind. And if you are in the Matrix, you are tasting a computer chip
in a certain state!
Whether one is in a simulation or a demon victim will make some difference
to the causal profile of the table. If the table is a demonic idea, then a demon
7
Bousma is discussed in Chalmers (2018, 627–629).
8
Sklar (1985, 60–161) and Vogel (1990, 661) both consider, and reject, a sort of structuralist
veridicalism view as well, as Chalmers points out.
9
See Chalmers (2018, 24–27) for some important details, including whether mass is identified with
the property that actually realizes the requisite causal role, or whether it is identified as the
property of having a property that plays that role. Though this is an important distinction for
Chalmers, it won’t play an important role here.
10 Epistemology
could, if it wanted, make the table disappear. Some (perhaps those more
comfortable with theistic views according to which God constantly sustains
all objects) might find this to be perfectly compatible with what they usually
take the causal profile of tables to be, while others might find this to be
inconsistent with how they assume tables behave. Presumably, though, all
have tables in mind. So, in some respects, the causal profile of tables permits
of some flexibility. Perhaps there is some range of different causal profiles that,
if satisfied, is sufficient for tablehood. So, different “kinds” of tables will have
different total causal profiles. But structuralist veridicalism holds that your
ignorance about whether sk is true does not imply that you do not know that
there are tables, because you know that something within that range exists. You
still know that they have a certain mass, that they have four legs, that you can put
cups on them, and so on; so (2) is false. On this view, sk is a metaphysical
hypothesis about the ultimate nature of things like tables, and perhaps also about
the specific causal profile of the table. If you are a demon victim, there is still
a table here, but the table is ultimately a demonic idea that would vanish if the
demon willed it. If you are in the Matrix or a brain in a vat (BIV), then the table
is ultimately a computational state and could vanish according to the program,
or by a glitch. And so on for other sk’s.
There are metaphysical objections to structuralism, independent of its
implications for skepticism, that I will set aside.10 What I will argue in what
follows is that structuralism (and other versions of veridicalism, and some
other denials of (2)), even if true, would not suffice to solve the global
skeptical problem.
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
Valberg’s (2007) view is not intended to solve all relevant skeptical problems,
but only the standard sort. One cannot do his nuanced phenomenological
observations justice in summary form. But, roughly, he observes that the objects
of experience are arrayed around you in a spatial expanse with you (the subject)
at the center. He calls this array, or that within which things appear to you, your
“personal horizon” (Johnston 2010, ch. 2, following Valberg in most details,
posits a similar phenomenon, calling it your “arena of presence.”) According to
10
Vogel (2019) presents some, focusing on whether concepts such as distance and caus-
ation can be construed structurally. See Chalmers (2018, 18–19) for replies. But Vogel
agrees with Chalmers that structuralism, if true, would solve the global skeptical problem
(pp. 3–4), so on this point I differ from them both. Another sort of problem for
structuralist veridicalism might be developed from Korman’s (2015, 39) argument that
our ordinary object concepts are not so undemanding as to apply to whatever is out there
in our environment.
The Skeptic and the Veridicalist 11
Valberg, if this is a dream, then the table you see in front of you is a table “in the
dream,” and this is an extrinsic, rather than intrinsic fact about the table (2007,
30, 85, 93). As he puts it, the “this” in “this is all a dream” refers to your personal
horizon, which in this scenario is a dream. You can know, in the ordinary way,
that there is a table, then, without knowing whether this is a dream; you can
know what the thing appearing in front of you is without knowing what that
within which it appears is. “Whether THIS [your present personal horizon]
is a dream does not bear, one way or the other, on whether this is a cup (say),
an external object . . . dream skepticism is transcendent . . . It leaves my
knowledge of this totality, whatever it is, untouched. It leaves untouched,
then, precisely the knowledge that skepticism about the external world calls
into question” (110–111). On this view, skepticism is “both vindicated and
disarmed” (113). What is left is a “transcendent” skepticism about whether
the world around you, in which things like tables appear, is a dream. But this
does not bear on whether you know that there is a table, on Valberg’s view.
Ultimately, my criticism of veridicalism is compatible with this latter claim
about transcendent skepticism, but I will argue that Valberg’s treatment of
the more standard (what he calls “immanent”) skepticism does not constitute
an adequate solution to it.
Valberg applies the same strategy to other global skeptical hypotheses, such
as the BIV, but most of his remarks focus on dreams. So, a note about dream
hypotheses is in order. One might think that, in one sense of “dream,” dream
experiences are, by definition, nonveridical. Valberg, for various reasons,
rejects this view, and holds that if you are dreaming that there is a table here,
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
and thus must be interpreted to be largely correct (and the same applies to one’s
assertions) (2006, 236). So, to say that your belief that there are tables is
false is to misunderstand your belief. However, this applies only to one’s
simplest beliefs – the simplest cases get “special weight” in this interpret-
ation (2006, 237) – so that nonsimple, and more theoretical beliefs about the
inner nature of tables are not immune to error due to interpretation consid-
erations. And, at least on one plausible reading of Davidson, your belief that
you are not a BIV (assuming you have such a belief) is not one of those
simpler beliefs that are immune to error. Thus, this view is at least compat-
ible with (1), since it does not guarantee that the truth-conditions of the
belief that ~sk are met. And yet, your belief that there are tables must be, on
this view, true on any correct interpretation, so (2) is false.11 Davidson was
a veridicalist12 on this standard interpretation of his view.13
As with Chalmers’ structuralism, there have been some objections to
Davidson’s view. For example, Craig (1990, 213), McDowell (1994, 17), and
Brueckner (2010, 181) have argued that Davidson’s veridicalism swaps ignor-
ance about whether the belief that there is a table is true for ignorance about
what the belief that there is a table is about. Is it a belief about computer states,
demonic ideas, or the sort of mind-independent physical object we typically
assume it to be? Davidson, the objection goes, makes us ignorant about the
contents of our beliefs. One might wave this worry away if one accepts, as many
do, Putnam’s (1975, 227) saying that “meanings ain’t in the head.” Some
ignorance about the contents of our beliefs may be the norm, rather than some
skeptical catastrophe. But, more importantly for us in this context, not all
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
11
Klein (1986) suggests that there is an important gap between securing the truth of one’s belief
that there is a table, and knowledge that there is a table, so that (2) is not so easily secured. But,
presumably, ordinary knowledge attributions are among those simpler, nonphilosophical beliefs
that should be charitably interpreted. At any rate, I won’t pursue this, or other, objections to
Davidson’s appeal to charity here.
12
For interpretations of Davidson as a veridicalist – as accepting (1) (or at least not denying (1))
and rejecting (2) – see Brueckner (2010, 180), Button (2013, 142), Craig (1990), Klein (1986),
and McDowell (1994, 17 fn 14).
13
Though he doesn’t address Valberg, Chalmers rejects Davidson’s veridicalism, as well as
Bowsma’s (1949) ordinary language appeal to “illusion,” which he considers to be a version
of veridicalism.
The Skeptic and the Veridicalist 13
First: (2) is false because there are (knowable) things such as tables even if sk
is true, and even if (1) is true; therefore, the argument from (1) & (2) to (3) (I
do not know that there are tables) & (4) (I do not know that there are any
ordinary things) is unsound. Second: (1) shows merely that we do not know
something about the ordinary things (like tables) around us, for example
whether they are simulations.
neutral terms, not only can we articulate skeptical worries about specific
scenarios, but also distinguish, say, BIV from Matrix and demon scenarios
(2005, section 8). Since all versions of veridicalism are compatible with (1),
and countenance sk as an open, conceptual possibility, they all accept that there
are some semantically neutral terms with which to specify some of sk’s details,
and what exactly we might be getting wrong about the nature of ordinary
objects.14
14
For more on why Chalmers thinks we can specify these details about sk, while some of our other
terms, such as place names like “Tucson” are not neutral, see his (2005, section 9, note 1) and
(2018, pp. 2–4) Valberg also thinks we possess neutral concepts enough to specify and distinguish
different sk’s. For example, in contrast to the dreaming scenario: “the [BIV] hypothesis (properly
understood) is . . . first, that there is a transcendent world, a world outside [this personal
horizon]; second, that there is in the transcendent world a brain maintained and stimulated in
a vat; and third, that the transcendent brain-in-a-vat setup is responsible for there being such a thing
as [this horizon], with the world (including [me] and [this] brain) internal to it” (p. 118, fn 12). So,
clearly ‘vat,’ ‘stimulated’, and ‘transcendent’ can be used neutrally. He also does not think that the
14 Epistemology
An alternative view rejects all neutral terms with which to formulate sk. If
there are no neutral terms, then in any world – even one created by Descartes’
demon – everything counts as physical and mind-independent, and we cannot
express what we are getting wrong about objects in a demon scenario. Such
a strategy may appear similar to, or just a more extreme version of, veridicalism,
but it rejects (1) instead of (2) – arguably, it must accept (2), since (2) would then
have a false antecedent. It rejects (1) because, on this view, it is a priori that the
table is mind-independent, not an idea, not a computation, and so on, so it is
a priori that sk is false.
One such strategy, which rejects (1), is Putnam’s (1981, 14) content
externalism, on which concepts such as brain, vat, and even causation, are
not semantically neutral (Putnam, 1994, 207). Rather, what they express is
determined, roughly, by what you are actually in contact with – recall that, in
contrast, on Davidson’s view this is true only of one’s simplest beliefs, not all
of them. Putnam argued that, therefore, whatever scenario you are in, you are
not a BIV, and thereby rejected (1). Putnam (1994, 287) and others who
follow his antiskeptical strategy (for example Tim Button (2013, 117–48)
and A. W. Moore (1996, 224–230; 2011, 51–52) also hold that you possess no
concepts with which to describe what it is about the table that you are
ignorant of due to (1). Your skeptical ignorance about the table is ineffable,
because you lack the conceptual resources to specify it.15 So if you are
somehow deceived, or are misconceiving the world, you cannot say how. It
seems to follow that no real skeptical problem about the table due to sk
remains.16
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
dreaming and BIV scenarios are “pseudo-hypotheses,” as the Putnamian approach, described later
in this section, suggests (121).
15
Note that Button’s view on whether we possess the requisite concepts is more complex than
I have sketched here; see chapter 19 of his book, on the “cosmopolitan” view, and the book’s
Coda.
16
Sk is merely a “bare formal” possibility, as Button (2013, 131–132) puts it. See also Button
(2013, 147,48) on “metaphysical” skepticism, which holds “nothing more than a nebulous sense
that the world is mysterious.” In contrast, veridicalists hold that there are specifiable and
meaningful aspects of things that we fail to know about the table, due to (1).
17
See Nagel (1986, 73), Wright (1992, 93), and Sundell (2016, 248).
The Skeptic and the Veridicalist 15
articulate the basic details about sk in its various versions. As we will see, my
argument ultimately can be used to support such Putnamian approaches, since
one upshot is that only views that reject (1) can solve the problem that the
standard argument poses.
A different, older view that might also be confused with veridicalism is
Berkeley’s idealism. According to Berkeley, things such as tables exist, but
they are ideal – ultimately, ideas in the perceivers’ minds (including God’s).
Thus, their ultimate nature is discoverable: such objects are all mind-dependent.
This differs from veridicalism in a number of ways. To begin with, veridicalism
does not reject (1), and it is committed to the view that we do not know the
ultimate nature of things. The ultimate nature of things depends on what
scenario we are in, and we don’t know which scenario we are in. In contrast,
Berkeley does reject (1) and rejects the view that we don’t know the ultimate
nature of things. For example, Berkeley thinks we can rule out the BIV scenario,
on which tables (according to veridicalism) are mind-independent, computer-
simulated objects. He thinks we know that ordinary things are ultimately
ideal.18
might have in mind when thinking about what that thing is. For example, if you
tell me I will inherit some furniture, but I don’t know whether it is a table or
chair, that is one sort of ignorance about what it is. If you tell me it is a table, but
I don’t know whether it is wooden or plastic, that is another sort of ignorance
about what it is. If you tell me it is a table, but not whether it is an idea in some
other being’s imagination, then that is yet another sort of ignorance about what it
is, one that only veridicalism makes possible, and one that seems intuitively
much more extreme, or at any rate unordinary.
For present purposes, it is not necessary to come up with a complete account
of “knowing what” – though see Dasgupta (2015) for one recent account that is
compatible with my use of “knowing what.”19 Failing to know something about
18
See Chalmers (2022, 68–71) for a discussion of the differences between veridicalism and
idealism.
19
According to Dasgupta’s (2015, 468) account, “I ‘know what’ something is iff I can identify it by
its nature.” Arguably, if you don’t know whether the table’s “inner nature” is computational,
ideal, etc., you fail to know “its nature” in the sense Dasgupta develops, the “essence” of a thing.
16 Epistemology
a table, or its nature, due to (1) constitutes one sort of ignorance about what the table
is. So even if you know that whatever there is “here,” which causes your table
experience, counts as a table, you do not know what kind of table it is – a demonic
table, a simulated table, ad so on – according to (1). Since this ignorance derives
from ignorance about sk, let us call it ignorance about “whatsk” something is.
According to veridicalism, then, you don’t know whatsk the table is, and that is just
to say that the aspect of it that you fail to know about is the aspect that you are
ignorant about due to (1): whether it is an idea, a simulation, and so forth. If we
accept (2), we can express ignorance about the table due to (1) as (3). If, instead, we
reject (2) and accept veridicalism, then we are still committed to ignorance about
tables, but the structure of this ignorance changes, from (3), to our failing to know
whatsk the table is. The structure has changed from what there is (are there tables?)
to whatsk things are (are tables simulations?).
It is worth emphasizing the compatibility of ignorance about whatsk the table is
with some other knowledge of what the table is, so as not to overstate veridical-
ism’s commitments. For example, you might still know that the table is not a sofa,
is made of wood, has mass, and that you can put a cup on it, according to
veridicalism. For, wood, mass, cups, and impenetrability are, on the structuralist
version of veridicalism, analyzed in terms of their causal roles, which remain the
same on sk. On Valberg’s version, wood, mass, cups, and impenetrability are all
there in the dream. Whether they are in the dream is an extrinsic feature of them.
And similarly, beliefs about the table’s having these properties must be interpreted
to be true, on Davidson’s view. Thus, since the property of being made of wood,
for example, is instantiated here whether or not sk is true, that aspect of the table is
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
not included in whatsk it is. However, it is also important to note that this
knowledge of what the table is does not extend very far. You still don’t know
whether things like mass and wood are demonic ideas, for example. That is just
more ignorance of whatsk things – in this case mass and wood – are. Note, too, that
since tables, we are assuming, could have demonic causal profiles (as discussed
earlier in this section), ignorance of whatsk the table is implies ignorance of
whether, say, it would vanish if Descartes’ demon willed it.
For those who don’t like this talk of “what” things are, similar remarks apply to
the “nature” of a thing. Insofar as it is in the nature of a table that it is not a sofa, then
according to veridicalism you can know the nature of the table. But there is another
part of its nature that you cannot know due to your ignorance about sk, or (1): is the
nature of the table that it is a demonic idea, a computational simulation?20
20
I avoid the use of “essence” because whether it applies depends on the specific version of
veridicalism at issue. If a table is anything that has table-like causal roles (what Chalmers calls
“role” structuralism), then sk does not specify part of the essence of a table, since the table could
exist regardless of whether it is material, ideal, or computational. If instead tables are whatever in
The Skeptic and the Veridicalist 17
Consider the different skeptical claims at issue. Skepticism about what there is,
I cannot know whether this table is a BIV-table, a demon-table, and so on. (In
other words, I cannot know whatsk the table is.)
Intuitively, it would be strange to find the first conclusion worrying while not
finding the second conclusion worrying. To be sure, some people don’t find
either one worrying or problematic. For such people, the standard skeptical
argument, (1)–(4), does not constitute a philosophical problem, but perhaps (at
most) a curiosity or trivial puzzle. For such people, there is no use for the
veridicalist “solution” to the skeptical problem, because there is no problem. As
I explain further later, my claim is ultimately a conditional one: if you find
skepticism about what there is worrying, then you should be just as worried by
skepticiam about whatsk things are. Intuitively, they seem to express the same
basic ignorance about the world. The second conclusion about whatsk things are
seems like a mere reformulation of, rather than an improvement over the first
conclusion about what there is. This is because they are both implied by the
following basic skepticism:
(2) or veridicalism. Once we notice that veridicalism does not avoid basic
skepticism, it becomes hard to feel relieved by swapping the standard skepti-
cism about what there is for veridicalist skepticism about whatsk things are.
Basic skepticism is just what worried us about standard skepticism in the first
place if standard skepticism was ever worrying at all: you do not know about the
world beyond its sensory effects on you. This is the external world after all. It
was only due to some metaphysical assumption about what tables could or
couldn’t be that we expressed such worries as worries about what there is, rather
than what things are. But that’s just how we expressed the same, basic worry
about the external world. In the next section, Section 3, I vindicate this intuition
with some arguments.
the actual situation in fact has the table-like causal role (what Chalmers calls “realize” structur-
alism), and if in fact tables are ideal, then it is part of the essence of tables that they are ideal,
because nonideal things can’t be tables. By focusing on nature instead of essence, we remain
neutral on these two versions of verdicalism.
18 Epistemology
concludes with skepticism about what there is, and the veridicalist antiskep-
tical strategy leads to skepticism about whatsk things are. And I have suggested
that these two skeptical conclusions are intuitively equally problematic because
both are committed to basic skepticism: you don’t know what is causing your
experiences, what is “beyond” your subjectively accessible mental states, or in
the external world. In this section I will support this intuition with some
arguments.
Admittedly, it is a slippery matter what “problematic” in “equally problem-
atic” amounts to, or what one should find worrying in the first place. So the
first task is to get clearer on this. (1)–(4) is valid, and its premises look
plausible, at least at first. The argument poses a problem if the conclusion,
(4), is worrying or unacceptable from a philosophical perspective. As I have
suggested, (4) is worrying because it implies that our relation to reality is
a joke, because it implies that we have very little knowledge about the world
around us. One might attempt to avoid this, and to assuage the worry, by
showing that the argument (1)–(4) is unsound. But if such a refutation of (1)–
(4) commits us to something just as worrying as (4), by also implying that our
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
relation to reality is a joke, or that we have just as little knowledge about the
world as (4) implied, then that refutation fails to solve the problem presented
by (1)–(4). This is what I will be arguing. In other words, my conclusion here
is a conditional one:
If you found the standard skeptical conclusion (4) about what there is
worrying, you should not be significantly less worried by the veridicalist
skeptical conclusion about whatsk things are.
21
“In Avnur, Yuval (forthcoming) Veridicalism and Scepticism. Philosophical Quarterly.
I raise some similar problems for a narrower class of veridicalist strategies, focusing on
structuralism.”
The Skeptic and the Veridicalist 19
long as you know that one of them is, you know that there is a table. The only
difference, then, is that the veridicalist claims that you know that there is a table;
both the veridicalist and the skeptic think that all you know is that something
satisfies the disjunction, and that you cannot know which disjunct is true. But
how can veridicalism posit more knowledge about the world than the standard
22
I am here presenting Chalmers’ (2018, 650) version of veridicalism, or “realizer structuralism,”
according to which whatever actually realizes the relevant causal role is a table. It is arguable
that, on Valberg and Davidson’s views, being a table is the even less specific property of being
anything that satisfies the disjunction. That is, for example, for Valberg, a table in the dream is
a table, but so is a table outside of a dream. There are advantages to either kind of veridicalism.
Briefly, the advantage of the first kind is that a replica of an object (say, inside a holographic
model) that plays the same causal role as the real object will not count as a real object. And that
seems right. On the other hand, on that view, if you happen to be a BIV, then tables are computer
states, and nothing else could be a table – not a demonic idea or even a regular, non-sk table. It
seems odd, though, to say that a nonsimulated thing with the causal role of a table is not a table,
since it seems intuitively to be a paradigm case of a table as we all (even we who are BIVs)
conceive of tables.
20 Epistemology
skeptic, if there is no possibility about the table, no disjunct, which can be ruled
out according to the veridicalist but not the skeptic? The difference seems to be
merely how we describe the possibilities that we cannot rule out.23
The veridicalist might respond by saying that on her view one possibility that
can be ruled out is that there are no tables. This is denied by the skeptic who
accepts (3) and (4). But this is a difference about how possibilities are described,
not which possibilities can be ruled out. According to the veridicalist, the claim
that we can rule out the possibility that there are no tables is just the claim that we
can rule out the possibility that nothing plays the causal role associated with
tables. However, the skeptic who accepts (4) can also rule that out. Something,
either the demon, the BIV computer, the robot overlords in the Matrix, your
dreaming brain, or something else, plays a role that we associate with tables.
(Other skeptics might deny that there is even any causation, but here we are
considering a skeptic who accepts (1)–(4) but accepts that something causes her
experiences.) She knows this in just the same way that the veridicalist does: while
granting (1) (I do not know that ~sk) and not knowing whatsk the thing causing the
table experience is. So, the very same possibility can be ruled out by someone who
accepts (4). The only difference is that the skeptic who accepts (4) does not say
that anyone who can rule this out thereby knows that there are tables, because she
does not hold the same, permissive view of what counts as a table. To be clear, the
standard skeptic and the veridicalist agree that this possibility, which we can rule
out, is one in which there are no tables, because playing the causal role associated
with tables is necessary for being a table. They disagree only about whether
whatever is actually playing that role counts as a table, even if it is an sk-entity.
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
So, they disagree about how to describe the remaining possibilities. But there is
no possibility that veridicalism says we can rule out, but which standard skepti-
cism does not. Therefore, we possess the same information about the world
according to veridicalism as we do according to standard skepticism, and this
implies that veridicalism fails to solve the problem posed by (1)–(4). Another,
related way to reach this conclusion is by appeal to skeptical neologisms, next.
23
Note that this applies as well to an “underdetermination” skeptic, who holds that you do not
know that there are hands because your evidence does not favor the hypothesis that there are
hands over the hypothesis that you are in an sk scenario being stimulated in way that mimics
seeing hands. This kind of skeptic also agrees with the veridicalist that there is either a BIV table,
a demonic table, or a non-sk table (and so on).
The Skeptic and the Veridicalist 21
your table experiences with any regularity. Both the veridicalist and the standard
skeptic claim that we have knowledge that something is causing our table
experiences and has a causal profile similar enough to what we usually think
tables have. Even the skeptic knows that, if we are deceived by the demon, then
our experiences are caused by the demon. The only difference is that the
veridicalist claims that whatever this is counts as a table. But the skeptic could
come up with a clever term, “tabbies,” which she stipulates to apply to the thing
that actually causes her table experiences. She has thereby turned knowledge that
something caused her table experience into knowledge that there are tabbies. So,
she accepts (4), and yet now she also knows what things there are: tabbies. She
knows as much about tabbies as the veridicalist claims to know about tables.
Now our clever skeptic – who I call a skeptic because she accepts (1)–(4) – who
knows that there are tabbies, may no longer qualify as a skeptic about the external
world, because she claims to know what there is, namely tabbies (though she still
holds that we don’t know whether there are tables). Has she thereby avoided the
skeptical disaster implied by (4)? Compare what this clever skeptic, who accepts
(4), claims to know and what the veridicalist claims to know. They both know that
something causes their table experiences, but the veridicalist calls it a table, while
the skeptic calls it a tabby. But knowledge that there is a table and knowledge that
there is a tabby amount to the same information about the world. Metaphysically,
there is no difference between tabbies and the veridicalist’s tables: they exist in all
the same possible (global) scenarios, and regardless of whether sk is true.
Veridicalists hold that the clever skeptic has a false theory of tables, as we
ordinarily think of them, because tables are tabbies, even though this skeptic
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
says that sk is incompatible with tables, but is compatible with tabbies. But that is
a disagreement about what counts as a table; there is no substantial difference
between their claims to knowledge about what is going on around them, given
that that they both accept basic skepticism. The information they claim to have
about their environment is the same, expressed differently.
We can now see why the veridicalist antiskeptical strategy is not an adequate
solution to the skeptical problem posed by (4). Granting veridicalism, (4) is false,
and skepticism about what there is is avoided. (4) posed a problem by raising
a worry about our knowledge about the world. But veridicalism’s solution to the
problem posed by (4) is adequate only if the knowledge we have about the world
according to veridicalism is substantially better than the knowledge we can have
about the world according to (4). But someone who accepts (4), by merely
stipulating new terms such as “‘tabbies,” can claim to have essentially the same
knowledge as the veridicalist claims we have about the world, only using different
terms to express it. So, the knowledge we have according to veridicalism is not
substantially better than the knowledge we can have according to (4). It follows
that veridicalism doesn’t adequately solve the skeptical problem posed by (4).
22 Epistemology
Another thing follows from this: the standard skeptical argument, (1)–(4), is
not an adequate expression of the skeptical problem, since refuting the standard
argument is insufficient for solving the problem. We will return to this important
upshot later.
a conditional claim: if you find standard skepticism (1)–(4) worrying, then you
shouldn’t take solace in veridicalism. This is compatible, of course, with the
antecedent’s falsity. Second, and more substantively, it is implausible that our
ordinary thinking about the world is vindicated by veridicalism. We ordinarily
assume (or are disposed to believe) various things about tables that are not
vindicated by veridicalism: that they are not demonic ideas, or that they are not
items in a vast simulation brought about by robot overlords. This is clear from
the fact that skeptical scenarios initially appear to contradict what we assume
about the world, and surely you would be surprised if you somehow learned
that your entire life, and the world that stimulates your senses, is actually
a computer interface in a video game played by some being outside the
universe you know. Some of our ordinary beliefs are vindicated (there are
tables), and some of our ordinary beliefs (tables are not figments of some
creature’s imagination) are not. The question remains whether saving only the
first kind of belief is sufficient to undermine any worry we might have about
the second – more on this later.
To be sure, other assumptions about tables, for example that they are not
mostly empty space, have turned out false, and learning this has not amounted to
a full-blown skeptical crisis. I address this rebuttal, and the difference between
learning something new about a table’s material nature and our ignorance of
whether its nature is as according to sk, later, in Section 4.
Moreover, it is difficult to square the preceding dissolution with the claim,
established earlier, that veridicalism fails to posit any substantial knowledge or
information about the world that (4) fails to posit. Whether to call the things
causing our experiences “tables” rather than “tabbies” is a paradigmatically
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
empty question: not a question about which scenario you are in, but how to
describe that scenario.25 So, even if veridicalism saves some of our ordinary
thinking about the world, it is unclear how this could matter. Veridicalism does
not show that we know any more about the world, or about tables, than we do
according to standard skepticism. This is due to the way (4) is avoided, namely
by a very permissive account of what could count as a table. How could the way
your knowledge is formulated or described make a difference to how worried
25
This notion is originally due to Parfit (1984, 235). He illustrates it in Parfit (2011, 435): When
about to cross the channel, suppose you know that you will feel seasick, but you wonder whether
being seasick counts as being in pain – this is a question about the extension of the term “pain,”
not a question about how you feel. Here, you know what will happen, though you don’t know
how to describe it. This is in contrast to a case in which, about to undergo some medical
procedure, you don’t know whether you will be in pain. In the channel case, there aren’t two
different possibilities: you know which possibility will hold, just not what to call it. As Parfit puts
it, “It matters whether, while receiving the medical treatment, I shall be in pain. And it matters
whether, while crossing the Channel, I shall be seasick. But it does not matter whether, in feeling
seasick, I can be said to be in pain.”
24 Epistemology
you are about your knowledge of the world? Recall that veridicalism is still
committed to basic skepticism, according to which you don’t know what causes
your experiences. If that is what was worrying, then veridicalism cannot help.
Turn now to the interpretation according to which veridicalism solves the
problem raised by (4) because our ordinary beliefs, or more of them than
according to (4) anyway, turn out to be knowledge after all. On this view, the
number of ordinary beliefs that are true and constitute knowledge is higher
according to veridicalism than it is according to (4), and this is enough of an
epistemic difference to assuage the problem raised by (4). For example, you
know that there is a table made of wood, and a cup on the table, and that the table
and cup are in a room, and you even know what city the room is in, and so on.
These are ordinary beliefs that constitute knowledge according to veridicalism,
but not according to standard or even basic skepticism (though the latter leaves
this open rather than denying it). Earlier, I argued that this is not epistemically
significant, since it does not show that your information or knowledge about the
world excludes any possibilities (it only changes how those possibilities are
described), and that the very same items of knowledge can be known by
a skeptic who accepts (4) and who also invents a clever neologism. Here, the
objection is that the superior quantity of ordinary beliefs that count as know-
ledge is itself valuable, and this is an advantage that the veridicalist has over the
standard skeptic.
However, the comparative “score,” as I will refer to the matter of how many
of our ordinary beliefs constitute knowledge, is, first, not so straightforward,
and, second, not clearly significant.
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
ear plugs. But you don’t know why the noise is happening or what its source is.
In this case, it would be natural to say that, although you know what to do, you
don’t understand what is going on, because you don’t understand why the noise
is happening. Similarly, the veridicalist posits knowledge of how to manage
your experiences and adjust your behavior, but without any understanding of
why any of it is happening.
To be sure, some might think that such a lack of understanding is simply
a natural limitation of human cognition, or perhaps of the scientific enterprise.
Constructive empiricists about science, for example, hold that a theory can be
evaluated only for its empirical adequacy – how it matches observation – and
nothing more (Van Fraaseen 1980). Accordingly, science cannot tell us why
things are happening, in just the same sense that veridicalists cannot tell us why
26
Things are made more complicated here by the fact that, as Nick Treanor (2014) has pointed out,
what looks grammatically like “one” fact can often contain a conjunction of many.
26 Epistemology
things are happening. This suggests that the lack of understanding that veridic-
alism posits is no more troubling than the lack of understanding we are left with
in science according to constructive empiricists. I take up comparisons between
skepticism about whatsk things are and scientific humility in more detail in
Section 4.2. For now, note that constructive empiricism is a rather skeptical
view of what science can tell us about the world. (Van Fraassen himself was also
skeptical about induction and inference to the best explanation.) And the
general outlook that humans can never understand why things are happening
around them is, of course, rightly regarded as a skeptical outlook.
Let us return to the objection that knowledge of whatsk things are is not
necessary for navigating around furniture, and set the notion of “understanding”
aside for now. Consider other ways that sk might be significant for our lives
besides navigating around furniture. Certainly, if I thought that I may well be in
a vast simulation, this might affect lots of my thinking about my life, the
difference I can make by my actions, and potentially death (whether there is
more life once you “die” in the Matrix or in the demon’s mind is a different
question from whether there is more life once you die on the more common-
sense or “natural” hypothesis). Anyone who reflects on one’s life, and for whom
these reflections matter in their life decisions and attitudes, will not be indiffer-
ent to sk, or whatsk things are. This is quite independent of whether we call such
reflections or their outcomes “understanding.” Admittedly, this is not a matter of
navigating around furniture, but why should navigation be the only benchmark
for epistemic significance? The veridicalist antiskeptical strategy itself does not
provide an answer, since it is not a sort of pragmatism about epistemic value,
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
and since skepticism is, after all, a problem concerning epistemic value. I take
up this issue again below, in Section 4.
The second point against appealing to the score is that it seems, on reflection,
not to matter. I’ve argued earlier that our information about the world, or the
depth of our knowledge about it, is not significantly improved by replacing (2)
with veridicalism. Rather, merely the form of our ignorance due to (1) changes.
This is so despite the fact that, according to veridicalism, more of our “ordinary”
beliefs about what ordinary things there are turn out to constitute knowledge.
So, the only benefit, if there is one, of more of our ordinary beliefs turning out to
be knowledge is the number, or score, of things we got “right.” However, what
was disturbing about (4), originally, is the implication for our cognitive relation
to the world, or the information we have about it, rather than the score, or
number of beliefs that turn out to be knowledge. So, it is hard to see why the
score itself would matter. The veridicalist rejection of (4) is a matter of book-
keeping of the little knowledge we have, rather than an addition to the little
knowledge about the world we have according to (1).
The Skeptic and the Veridicalist 27
In other words, “what” we know about the world is not the same as “how
much” we know about the world. “How much” we know depends on how we’re
counting, or how we formulate what we know. True beliefs, when their contents
are more broadly construed or more easily satisfiable, are easier to come by, but
that is not due to the improved informational situation, or even shifting epi-
stemic standards (as is the case with contextualist solutions to skepticism).
Rather, it is merely a reinterpretation of what our beliefs are about.27
To illustrate these points, consider an analogy. Suppose that you are con-
cerned about your knowledge about the world due to your failure to know the
exact color of any of the dinosaurs – there are educated guesses out there, and
well-supported hypotheses, but, not being an expert, you don’t regard yourself
as knowing. Then, your friend convinces you that, properly understood, birds
count as dinosaurs. (Never mind whether this is correct, just suppose you are
convinced.) So now, you think to yourself, you do know the color of some
dinosaurs. Should you no longer be concerned? It depends on what concerned
you about not knowing the color of any dinosaurs before. If it was that you
failed to know the color of velociraptors, brontosauruses, and the other large
reptilian things that you thought of as dinosaurs, then surely you shouldn’t feel
any better now. You count yourself now as having more items of knowledge
concerning the color of dinosaurs, but this doesn’t improve your knowledge
about the things that concerned you before. The same is going on with veridic-
alist rejections of standard skepticism about what there is. For what concerned
us, intellectually, about lacking knowledge of what there is – basic skepticism –
is still the case, even if we can count more of our beliefs as items of knowledge
about “tables.”
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
If one was worried about one’s knowledge about the world on the basis of
(1)–(4), one must have been assuming a false metaphysics of things like tables,
according to veridicalism. Similarly, the worry about our ignorance of the color
of dinosaurs was based on a false assumption about what counts as a dinosaur.
But whether correcting these assumptions assuages the corresponding worry
depends on whether the worry itself depended on the assumption, or whether
instead merely the expression of that worry, as a worry about tables and
dinosaurs, depended on those assumptions. And in both the skepticism and
dinosaur cases, it is the latter. Expressing the skeptical worry as a matter of what
there is rather than a matter of whatsk things are is a result of assuming
a standard, nonveridicalist metaphysics of tables. But the root of the worry,
basic skepticism, remains. Learning that birds are dinosaurs doesn’t assuage the
27
See Treanor (2014) for a more nuanced discussion of this general point about counting true
beliefs.
28 Epistemology
worry about brontosauruses, and learning that demonic ideas can be tables
doesn’t assuage the worry that we don’t know whether the cause of our table
experiences is an evil demon. Rather, learning veridicalism should lead who-
ever was worried about (1)–(4) to reformulate the worry as one about whatsk
things are, or, more plainly, basic skepticism. And all you need in order to
motivate that worry is (1).
Let us sum up the case against the veridicalist antiskeptical strategy. If
veridicalists are right about tables, then skepticism about what there is (and
(4)) misdescribes our ignorance, or mis-expresses the worrying implication of
basic skepticism.28 Given basic skepticism, which veridicalism accepts, the
remaining question is not whether there is a table, but whether there is a certain
kind of table here, a nondemonic, nonsimulated table. Why bother with such
a question? Why should we reformulate the worry about what there is rather
than abandon it? If one was ever worried about our knowledge about the world
due to (1)–(4), then one should ask the new question because the veridicalist
answer to the old question doesn’t assuage the worry, it merely changes its
terms. Basic skepticism, the root of the original problem, is still there. If (4) was
worrying because it suggested that we don’t know much about the world (as
Stroud’s formulation suggests), swapping (4) for skepticism about whatsk there
is doesn’t help, because it doesn’t improve our knowledge about the world.
Rather, the terms in which the little knowledge we had (about the cause of our
experiences) is expressed changed when we accepted veridicalism. We didn’t
realize, before accepting veridicalism, that ideas could be tables, just as we
didn’t realize birds are dinosaurs when we asked what color dinosaurs are. The
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
mistake committed by (1)–(4) was not a mistaken evaluation of the state of our
knowledge, but a mistaken interpretation of “table.”
28
Craig (1990, 213) offers a similar diagnosis of Davidson’s veridicalism: it goes wrong in its
“over-concentration on a single sceptical formula, ‘It could be that most of our beliefs are false.’”
Craig leaves it at that, without arguing that there is a formulation that is just as epistemically
devastating, or that showing that most of our beliefs are true does not suffice to substantially
improve our epistemic position.
The Skeptic and the Veridicalist 29
veridicalism hold that we know that this is reality, even if we don’t know ~sk.
According to Chalmers, for example, we can know that this is reality despite not
knowing whether we are in some global simulation (or other sk’s). This follows
from his so-called simulation realism (2022, 106), the view that “If we’re in
a simulation, the objects around us are real and not an illusion.” This is based on
various conceptions of “reality” on which something is real if it exists, has
causal powers, is mind-independent, is nonillusory, and is genuine (see 2022,
ch. 6 for details). And, he argues, a global simulation will meet all of those
criteria – though other skeptical hypotheses may not satisfy all of those. So,
according to Chalmers, the table is real and you can know that it is real even if
you don’t know that this is not all a simulation. And that, one might think, is
a significant bit of knowledge about the external world that the skeptic who
accepts (4) denies. Note here that, on this view, it is simply not possible for the
thing playing the causal role associated with a table to fail to be real (unless you
are in a local or nonglobal skeptical scenario, which we are setting aside).
30 Epistemology
a simulation, then you might emerge from it, or be taken out the vat and
detached from the computer running the program. And then you would be apt
to judge that none of what you had experienced before was ever reality, even
though you thought it was during the simulation. Intuitively, such judgments are
true.29 The Chalmers-style veridicalist might reply that the meanings of words,
such as “tree” and “real,” change to match the actual environment as soon as one
emerges into the new environment, so that once you emerge, it is now true, but
wasn’t just a moment before your emergence, that the world you were experi-
encing isn’t real. But in order for this rebuttal to succeed, it must imply that
when you were in the vat, for example, your sentence “the things I experience
29
Chalmers does not apply his general veridicalist strategy to scenarios in which you have been
recently envatted or put into the Matrix, but rather only “global” scenarios. So presumably he
does not mean to apply his view to scenarios in which you were recently de-vatted, either.
However, we can still consider how a subject would or should react if a subject were suddenly
removed from its vat, or simulation. Our intuitions about the truth conditions of “what
I experienced as a table was a real table” if that were to happen are consistent, of course, with
the hypothesis that it will never happen.
The Skeptic and the Veridicalist 31
are real” expressed a true proposition, even though when you emerge from the
vat, your sentence “what I experienced while in the vat was not reality” is also
true. This would require the meaning of “real” to change suddenly, the instant
you emerge. There is some question, which I cannot settle here, about whether
our meanings can change so quickly due to our being in a new environment.30
But, furthermore, it seems that we are just as apt to judge, after emerging from
the vat, that when we thought, before, that the things that apparently happened
were reality (or “really” happened), those thoughts were false. So, on the current
reply, we’d have to be mistaken about what we meant before, which is at least
some significant cost to such a view of meaning.
But perhaps the preceding difficulty for the Chalmers-style view is surmount-
able. Suppose instead that “reality” is not a neutral term, so that, like “wooden,”
it applies to whatever realizes the right causal role. If so, then knowing that this
is reality does not amount to much. For, “real” applies even to simulations (if
you are in one). “Reality” is merely a label we can place on whatever explains
why our experiences have a certain regularity. But knowing this is no better than
what you can know consistently with (1)–(4), since even if (4) is true, you know
that something explains why our experiences have a certain regularity (it could
be a demon, robot overlords, etc.). So now it is hard to see why the ability to
claim knowledge that what you experience is “real” is any achievement at all.
This is not knowing much, since even a skeptic knows the same thing, even if
she doesn’t use the term “real” to describe it. Moreover, this view seems to
imply that, upon emerging from the vat, one could truly state, “I’ve been in
reality all along, and still am now.” This seems intuitively false.
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
humility about the nature of things. And still today there is much we don’t know
about what lies behind the manifest image. Consider what Chalmers calls “The
Computational Hypothesis”:
Let us call the computational hypothesis “ch” for short. Chalmers (2005, 10)
suggests that the Matrix (or BIV) scenario is an instance of ch, so ignorance of
sk is no worse than ignorance of whether ch is true (as some physicists
apparently hold).
The Skeptic and the Veridicalist 33
course, for H2O and the precise chemical composition of water. If you’re
ignorant of sk, then science, when it comes to the ultimate nature of things, is
not just incomplete, it is hopeless.
It follows that ignorance about sk is not just as acceptable as ignorance about
ch or H2O. For, what makes ignorance about ch and H2O acceptable is that
investigation is ongoing, that the scientific enterprise at any given time (at least
before it becomes “complete”) can’t be expected to tell us everything about the
fundamental nature of things. It is reasonable to accept the relative immaturity
of the science of fundamental nature. But if we are ignorant of sk, then the
immaturity of the scientific enterprise is irrelevant, since we can never know
what things ultimately are. Science can’t be expected to tell us anything about
the fundamental nature of things. This is worse than the truism that science can’t
tell us everything.
To be clear, some scientific hypotheses can arguably be known even if we are
ignorant of sk, given verdicailsm: those expressible in nonneutral terms. We can
know that the table is made of wood, and that wood comes from felled trees,
34 Epistemology
and that trees photosynthesize, given that these are nonneutral terms. So scien-
tists can investigate such claims regardless of our ignorance of sk. Of course, if
this were all happening in the Matrix, machines might be able to trick the
scientists into getting even those hypotheses wrong, since they can produce
any sequence of experiences, or data, they want to. But this is a matter of
messing with your life within the simulation in ways that are probably not the
global scenarios veridicalists like Chalmers have in mind (such scenarios are
more like The Truman Show than The Matrix).
However, the fact that some science may be possible even if we are ignorant
about sk does not address the main problem with the objection. The point is that
trees and photosynthesis may be, according to (1), ultimately demonic ideas.
This sort of ignorance is of a different sort, as I have suggested, than what mere
scientific humility suggests. Hypotheses about the fundamental nature of things
that are expressed in neutral terms are unknowable, if sk is unknowable. (On the
Putnamian view considered earlier, on which there are no neutral terms, there
are no such hypotheses; but that is not veridicalism, and it amounts to a rejection
of (1).)
Nor does the preceding imply that there is some clear-cut principle distin-
guishing scientific from skeptical hypotheses. One might regard it as acceptable
scientific humility that we can never know, say, whether this universe is
embedded within a larger one, or even whether there is a level of reality
“under” the ones that we can discover, which is entirely different from what
we could ever know about the inner nature of things. And if we cannot know
whether the universe is infinitely complex, perhaps we cannot know whether
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
scientific hypotheses are so strange that, given that they cannot be ruled out,
they can serve as skeptical hypotheses. Science fiction has, of course, supplied
much of the material for modern skeptical hypotheses, so this is hardly
surprising.32
the BIV scenario has it that tables are ultimately computational. But so does ch.
Suppose we are ignorant of whether such hypotheses as panpsychism and ch are
true (and suppose we treat ch as a metaphysical rather than a scientific hypoth-
esis). Why is this not disturbing in the way that ignorance about sk is? Here, one
cannot appeal to the fact that skeptical scenarios are (purported to be) unfalsifi-
able. For it could be argued that metaphysical views such as panpsychism are
also unfalsifiable. If so, then why couldn’t panpsychism be a version of sk?
Like Berkeley’s idealism, panpsychism is a positive view about what things
ultimately are. Skepticism about whatsk things are is a negative view about our
ability to know any such theory, or any of its alternatives. It may well be that our
inability to know whether panpsychism is true demonstrates a significant inabil-
ity of ours, namely to know the nature of things. If so, then panpsychism could,
being employed in this way, serve as one version of sk. Whether panpsychism is
as effective at highlighting our ignorance as the Matrix is depends on how deep
32
For related discussion on quantum phenomena in relation to skepticism, see Vogel (2019, 9–10).
36 Epistemology
ignorance of the intrinsic properties of substances. The substances that bear these
intrinsic properties are the very same unhidden substances that do indeed affect
us perceptually. But they affect us, and they affect other things that in turn affect
us, in virtue of their causal powers, which are among their relational properties.
Thereby we find out about these substances as bearers of causal powers, but we
find out nothing about them as they are in themselves. (2009, 203)
possible substances that bear causal powers, or “realizers,” clash with our
assumptions about the world so violently that our ignorance about them disturbs
us and undermines what is intuitively our knowledge about the world. As Rae
Langton put it,
There are some ways of filling in the details of what substances in themselves
are that make too much of a mockery of our assumptions about the world. “Who
ever promised me that I was capable in principle of knowing whether this table
is an idea in someone’s (maybe my) imagination” does not have the same
comforting ring to it as Lewis’s original. This is because that is not an expres-
sion of humility, it is an expression of skepticism: you have no idea what is
going on around you.
It should be conceded, though, that there is no bright line between skeptical
scenarios and such metaphysical hypotheses. This corroborates the point, made
earlier, that veridicalism helps to establish the metaphysical subject matter of the
skeptical problem: we can never know what things are. Still, this does not
seriously undermine the main concerns about veridicalism as a response to
skepticism. For, just as not every sort of ignorance about what things are consti-
tutes skepticism, nor does every sort of ignorance about what there is constitute
skepticism. Suppose you wake up in a basement one day, after a near-apocalyptic
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
war, wondering if any furniture in the world has been left undestroyed: “Are there
any tables?” This would not be a case in which you are suddenly in a skeptically
bad situation of ignorance about what there is. And yet, even veridicalists agree
that (3) is skeptically disturbing. The fact that some other ways of being ignorant
of what things are fail to be skeptically disturbing doesn’t show that all ways of
being ignorant about what things are fail to be skeptically disturbing.
33
Granted, Langton ultimately rejects this skepticism, but the passage quoted here indicates that
she took ignorance of what ultimately generates our experiences as a potential source of skeptical
worries about the external world.
34
This was suggested to me by an anonymous referee.
38 Epistemology
physical, so that we remain ignorant of “what minds are.” Surely this would
solve the “problem of other minds,” even if it leaves open the nature of minds.
Why isn’t the view, offered by veridicalism, that there are tables, but which
leaves open whatsk tables are, just as satisfying a solution to skepticism?
The answer is that veridicalism leaves open much more about the nature of
tables than this view about minds does in leaving open whether minds are
physical. What’s worrying about the problem of other minds is, primarily, that
we don’t know whether other people are conscious. It is a certain conception of
other minds – as conscious – that enables us to express the worry as one about
minds. Likewise, what is worrying about the existence of tables, at least as it
comes up in the context of skepticism, is primarily that we don’t know whether
all things in the external world (that is, beyond our sensations) are ideas, or
simulations. It is a certain conception of tables – as nonsimulations and
nonideas – that enables us to express the worry as one about the existence of
tables. The view that minds exist but may not be physical still addresses the
question of consciousness, and so it solves the problem of other (conscious)
minds, so long as we accept that something physical could be conscious, of
course. The view that tables exist but might be ideas does not address the
question of whether tables are ideas, so it does not solve the skeptical problem.
Rather, it directly challenges the conception of tables that was assumed when
the problem was expressed as one about the existence of tables.
So, a better analogy to veridicalism about tables is the view that there are
other minds, which leaves open extreme eliminativism about consciousness.
That is, on this alternative view, there are minds but they might not be con-
scious. There is no problem of other “minds,” but only because this radical view
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
about minds makes it much easier to count as a mind than it is according to the
standard view that is assumed by those who usually pose the problem as one
about “minds.” On this alternative view of minds, even zombies and robots can
count as having minds, so the problem of other “minds” is better expressed as
the problem of other “consciousness.” This is what bothered us about other
minds in the first place, of course, and is a reformulation of the original problem.
For this reason, the alternative view of minds that leaves eliminativism open is
not itself a solution, but a rejection of the terms of the problem. Not just any
theory of mind is such that, as long as we end up knowing that there are other
minds, the problem of other minds is thereby solved. Such a theory must be
consistent with the conception of minds that led us to formulate the problem as
one about “minds” in the first place. What is the criterion for what must be
assumed about minds in order for such a theory to constitute a solution? Minds
need to be conscious. Otherwise, what we were originally worried about in the
problem of other minds will not have been addressed. We need to consult the
The Skeptic and the Veridicalist 39
assumptions about what minds are that made us express the worry as one about
minds in the first place.
Likewise, our worry in (3) about the existence of tables depended on the
assumption that tables cannot be ideas, or simulations. Once that assumption
is removed, the claim that there are tables no longer assuages the original
worry, but rather it changes the terms. So, the view that there are tables, but
they might be ideas or simulations, doesn’t help, given that what one was
worried about in (3) was that it implied that the things we call “tables” might
be mere ideas.
One might reply: what about a view according to which other people
have consciousness, but that doesn’t settle whether consciousness is phys-
ical? That would indeed solve the problem of other minds, assuming we
can know such a view to be true. But then, similarly, a view according to
which there are tables, which doesn’t settle the fundamental physics of
such things as tables, solves the skeptical problem, too. It’s just that this is
not equivalent to a view, like veridicalism, that determines that there are
tables but not whether they are BIV simulations, as I argued earlier. When
we worried about other minds due to the inaccessibility of others’ experi-
ences, we assumed minds have consciousness, not that they are physical
(or nonphysical). When we worried about tables due to (1), we assumed
tables aren’t simulations, not that this or that other hypothesis about the
nature of matter applied. Show me that other people are conscious, and I’m
satisfied about other minds. Show me that tables, or whatever causes my
table experiences, are not simulations, and I’m no longer worried about
https://doi.org/10.1017/9781009243308 Published online by Cambridge University Press
(4). But veridicalism doesn’t do that. Only a refutation of (1) (not just (2))
can do that.
— Vik, vik, vii ii! nyt on joulu! Aamulla tulee suntio kirkkoon ja
sytyttää pitkällä puutikulla kynttilät. Sitten meillä on oikein hauskaa!
— Vik, vik, vii, ii, ii, rientäkääpä pojat, että joudutaan ennen kotiin
kuin herrasväki, kehoitteli emohiiri. Tänään saamme kissaltakin
kunniallisen joulurauhan. Kylläisenä makaa se salin sohvalla.
Emojänis jatkoi:
Karta pellon aitaa, jos jonkun talon, tai torpan pihajänikseksi satut
joutumaan. Pellon aita olisi kyllä hyvä ystäväsi, vaan siinä piilee
vaaroja. Näes, torpan tai talon pieni poika saattaa virittää
hirttolangan pellon aidan rakoon, johonkin aukkoon, josta polkusi olet
pujoittanut ja lanka on niin hieno ja hajuttomaksi kuusen havuilla
sivelty, että sinä et sitä huomaakaan ennenkuin olet sen silmukassa.
Karta siis laittamasta polkuasi niin, että se pujotteleikse aidan raosta.
Katsele veräjä ja kulje siitä, se on varminta. Ja jos ei torpan mies
aukaisisi veräjää ensi lumen tultua, niin löytyy aina aidassa
suurempia aukkoja, jotka eivät ole vaarallisia.
Puputti oli taas jäänyt yksin latonsa kupeelle. Toveri oli lähtenyt
kerran yksin tepastelemaan läheiseen korpeen ja siellä pistänyt
päänsä loukkuun, jonka torpan mies oli virittänyt.
Ja nyt hän oli yksin. Suru painoi yksinäisinä pakkasöinä niin, ettei
kehdannut lähteä edes lämpimikseen hyppelemään. Ruokakaan ei
maistunut.
— No, ainahan sitä sellaista sattuu, eikä siinä suru auta. Taisi olla
ihan läheinen sukulainen, jos lie ollut ihan oma toveri, arveli harakka.
Tulipa sitten eteen aukeama, jonka laidassa kulki leveä tie. Puputti
loikkasi tielle, huomaamatta sitä ennenkuin oli jo melkein yli
pääsemässä. Mitä ihmettä? Tietä pitkin kulki rinnan kaksi mustaa
nauhaa, jotka olivat yhtä etäällä toisistaan. Puputti pyörsi
säikähtyneenä takaisin ja painautui rämeen juurelle. Mitähän se oli?
Hetkisen perästä uskalsi hän loikata tielle uudelleen. Piti ihan
käpälällään koettaa sitä mustaa nauhaa. Puuta se ei ollut, taisi olla
rautaa. Puputti töllisteli ja huomasi nyt pylväitä, joita myöten kulki
hieno rihma. Mitähän se oli? Taisi olla samaa rihmaa, jota poikaset
virittelivät metsään tallatuille poluille ja aidan rakoihin. Se lauloi
somasti tuulessa ja se näytti jatkuvan loppumattomiin.
— Älähän pelkää, en minä sinulle pahaa tee, usahti Mikki, joka oli
koiran nimi. Minä olenkin vain tällainen pieni kartanokoiran retus,
rakki, niinkuin sanotaan ja sellaisena jänöjussien ystävä. No, tulehan
pakinoille.
— Ota nyt pitkiä askelia, kehoitti Mikki ja niin mentiin niin että vilisi.
Mutta pojatkin pysyivät kintereillä. Puputin sydän oli seisahtua
pelosta, mutta hän koetti parastaan pysyäkseen Mikin perässä.
Pojat jäivät jo muutamassa pimeässä loukossa ja Puputti huohotti
niin, että henki oli katketa. Mikki naureksi seikkailulle, mutta Puputin
rintaa kouristi koti-ikävä niin rajusti, ettei voinut sanaakaan lausua.
Jos hän onnistuisi vielä ehjin nahoin pääsemään tästä oudosta
paikasta, niin ei koskaan enää lähtisi pahaa maailmaa katselemaan.
Oli ollut maaliskuun kantava hanki, kun Puputti lähti. Nyt oli
huhtikuun leuto yö, kun hän sai oikaista pitkin pituuttaan omaan
makuukseensa ja haukata ladosta hienoa heinää, jota torpan mies
oli jättänyt sinne hänen varalleen.
Updated editions will replace the previous one—the old editions will
be renamed.