Are Epistemic Reasons Perspective-Dependent?
Davide Fassio
From the Philosophy Symposium in
Gordes – 5 November 2016
Draft version – Please do not
cite without permission
Introduction
In recent years, the notion of reason has attracted
great interest in several domains of normative philosophy. Many philosophers
take reasons to be the fundamental building bocks of normativity,[1] and traditional debates framed in terms of
reasons helped to illuminate long-standing issues in practical philosophy,
epistemology, and other normative domains. A common distinction in the philosophical
literature is that between explanatory
(or motivating) reasons and normative
(or justificatory) reasons: we appeal to the former when we attempt to explain
actions and attitudes (e.g., the reason why the policeman shouted at the
prisoner is that the prisoner was trying to escape); we appeal to the latter to
justify them (e.g., a reason to believe that the suspect is the culprit is that
his finger-prints are on the gun). The focus of the present paper is on
normative reasons. Normative
reasons (hereafter simply reasons) are considerations that count in favour of or against a certain attitude or action. They contribute to determine what
a subject is permitted, ought or ought not to do, believe, desire, feel, admire
and so on.
This paper is concerned with a specific subset of
reasons, namely, epistemic reasons,
reasons to believe. The main question I investigate in this paper is the
following: which set of entities can count
as epistemic reasons? Some possible answers given by philosophers are that
epistemic reasons are the set of known facts, facts that a subject is in a
position to know, propositions that a subject rationally believes, or a subset
of one’s mental attitudes (beliefs, rational beliefs…). In order to simplify
the discussion, I shall assume factualism about normative reasons,
the view according to which normative reasons are facts or true propositions.[2] This is the most popular view in contemporary
philosophy. We can thus reformulate our question in factualist terms. We ask: which set of facts can count as reasons to believe (epistemic reasons)? All facts or only those dependent on the perspective of the subject?
The actual perspective of the
subject or the potential one?
Skorupski in The
Domains of Reasons argues that epistemic reasons are facts that belong to
the epistemic field of the subject, where a fact belonging to the subject’s
epistemic field is one dependent on the potential perspective of the
subject. In this paper, I compare Skorupski’s view to other contemporary views
holding that facts that can count as epistemic reasons belong to narrower sets
than a subject’s epistemic field. I provide arguments which seem to favour
Skorupski’s view over these alternatives. However, against Skorupski, I also
tentatively argue that the set of facts that can count as epistemic reasons
might be even wider than the facts in the epistemic field of the subject,
including facts which do not even belong to the potential perspective of the
subject.
This is the plan of the paper. In §1, I briefly introduce Skorupski’s
view about epistemic reasons, with a focus on the notion of Epistemic Field. In §2, I introduce and challenge two alternative
views. These views are: i) that reasons are actual-perspective dependent facts
and ii) that they are facts that the subject is in a position to know. I also
show that the view of Skorupski seems to be able to overcome these objections. However,
in §3, I consider possible problems for Skorupski’s view, and tentatively argue
that epistemic reasons can also be facts not belonging to the subject’s
epistemic field. In particular, I argue that epistemic reasons are true propositions
not necessarily belonging to the epistemic field of a subject.
1. Skorupski on epistemic reasons
According to Skorupski, epistemic reasons are “reasons to believe –
think, judge, conclude, and so on” (2010, 36). Epistemic reasons are distinct
from practical and evaluative reasons (reasons to act and to feel). These three
kinds of reasons are irreducible and constitute an exhaustive trichotomy (i.e.,
there are no other kinds of reasons). As I said before, Skorupski takes epistemic reasons to be facts, where
this term is interpreted in the nominal sense: a fact is a true proposition, or
the truth of a proposition. He distinguishes this sense from a more substantive
one.
Examples of epistemic reasons considered by Skorupski
are the following:
The fact that the freezer door is open is a reason to
think the ice cream will melt.
The fact that the governing party is alienating so
many interest groups gives one reason to believe that it will not be
re-elected.
All the components we've had so far from that supplier
have had this particular flaw, so there's some reason to conclude that all of
them have it.
Data gathered from underwater sea currents give us
some reason to believe that global warming is taking place.
Reasons are facts standing in a specific relation,
that Skorupski calls reason relation.
Reason relations can be thought as having five places (2010, 36-37): A fact (p), a time (t), a reason of degree of strength d,
a subject (or actor) for which
that fact is a reason (x), and a
response φ (an Action, Belief, Feeling):
R(p, t, d,
x, φ) – p
is at time t a reason of degree d for x to φ.
According to Skorupski, epistemic reasons have a
special feature that distinguishes them from other kinds of reasons: they must
be knowable facts (2010, 41). More precisely, Skorupski argues that there is a
restriction on the facts that can constitute epistemic reasons for a subject.
These facts must be in the epistemic field of the subject for which they
constitute reasons:
“The facts that are knowable by an actor x, and can thus be epistemic reasons for
x at a time t are limited by what x
could at t know of by reflection or
by further (physically possible, spatio-temporal) inquiry, or just by ‘stumbling
across’ them. These are the facts that are epistemically accessible to x at t:
x could come to know them by skill or
by luck. Call them x's epistemic
field at t.” (2010, 43)
Skorupski defines epistemic field as follows: x's epistemic field at t, ex,t,
is the set of facts that are accessible to x
at t—that x can come to know, discover, at t, whether or not x has
come to know them.[3] The following example well
illustrates how Skorupski conceives the epistemic field:
“Suppose the fact that there are small and distinctive
scratches on the window sill, together with other facts which a careful
inspection of the scene of the crime would reveal, is a reason to think that
the criminal escaped through the window. Sherlock Holmes says to Watson ‘Well,
I think there's some reason for us to conclude that the criminal escaped
through the window, don't you?’ But Watson is baffled, because he hasn't
noticed those highly significant scratches. Nevertheless, Holmes is quite right
to speak in the first-person plural. Those scratches are a reason for Watson to
think the criminal went that way; Watson could have noticed them if he'd been
more careful or lucky—the fact that he hasn't noticed them just is the fact
that he’s missed a reason for concluding something about the criminal's escape
route” (2010, 43).
Skorupski gives three arguments for the existence of
such an accessibility constraint on epistemic reasons:
If epistemic reasons are not restricted to the
epistemic field of the subject, “what gives me reason to believe that p obtains could be just that fact, the
fact that p—unrestrictedly for any
fact at all. […] But in many cases our reason for believing some fact to obtain
can only be inferential. In these cases the fact itself is not a reason for
believing that it obtains” (2010, 43).
“Furthermore, we take it for granted that there can be
propositions that are true even though there is no reason to believe that they
are true. So in the case of any such true proposition that p, which there is no reason for us to believe, the fact that p cannot itself be a reason to believe
that p” (ibid.).
“Facts that one cannot know cannot be reasons to
believe”. Consider, for example, the fact that exactly three gulls died on the
West Beach at St Andrews on 26 June 56 bce. This fact cannot itself be a reason
for me now to believe that three gulls died on that beach then, for the simple
fact that I cannot know it. (Ibid.)
Personally, I don’t find convincing any of these
arguments. For what concerns 1, there may well be cases in which the reason to
believe p is p itself (for example, the fact that it is sunny now is a reason
for me to believe that it is sunny). Furthermore, that p is a reason for believing p
doesn’t exclude that other facts can be reasons for believing it as well. And the
reasons to believe p that we actually
possess may not include p itself. Therefore,
the fact that p itself can be a
reason to believe p even though we
don’t possess it, and thus cannot believe p
for that reason. Argument 2 assumes that there are cases in which there can be
propositions that are true even though there is no reason to believe that they
are true. This is not obvious, and should be distinguished from the much more
plausible claim that there can be propositions that are true even though we have no reason to believe that they are true. Argument 3 presupposes that we can
know every fact that can be a reason to believe. I seriously doubt the validity
of this accessibility constraint, and I will come back to this point in §3.2.
However, observe that even if all facts that are epistemic reasons can be
known, this still doesn’t imply that all these facts are in the epistemic field
of a subject. Facts in the epistemic field of someone are a subset of those
that it is possible to known.[4]
2. Two
alternative views and why Skorupski’s view is better
In this section, I introduce and challenge two
alternative views to Skorupski’s one. These views are: i) that reasons are
actual-perspective dependent facts and ii) that they are facts that the subject
is in a position to know. The two views are introduced in §2.1. In §2.2 I
provide several objections to these views. In §2.3 I show that the view of
Skorupski is apparently untouched by these arguments.
2.1 Two
alternative views
Skorupski’s view is not amongst the most popular ones in
contemporary epistemology. The main contemporary views hold that epistemic reasons are things within the actual
perspective of the subject, not merely its potential one. If we assume
that reasons are facts, an instance of this view, defended amongst others by
philosophers such as Tim Williamson and Clayton Littlejohn, is that:
(ERK) epistemic
reasons are known facts
Known facts are a subset of the facts that are in the
epistemic field of a subject. If S knows p,
then p is also in the epistemic field
of the subject; but that q is in the
epistemic field of the subject doesn’t imply that S knows q.
According to another, slightly less popular view,
certain facts are epistemic reasons even if the subject doesn’t possess (know
or believe) them. But the subject must be in a position to know these
facts, where being in a position to
know involves the possession of attitudes that provide propositional
justification for the subject (Gibbons,
2013; Lord, 2015; Schroeder, manuscript; Smithies, 2012):
(ERPK) epistemic
reasons are facts that the subject is in a position to know
For example, according to Smithies:
“One is in a position to know a proposition just in
case one satisfies all the epistemic, as opposed to psychological, conditions
for knowledge, such as having ungettiered justification to believe a true
proposition. By contrast, one knows a proposition just in case one satisfies
not only epistemic, but also
psychological, conditions for knowledge, such as believing a true proposition
on the basis of one’s ungettiered justification to believe it.” (2012, 268)
Consider some example: a creationist teacher has
sufficient justification to make assertions from evolutionary theory, although
she does not know evolutionary theory because she does not even believe it for
religious motivations (Smithies 2012, 268-9). This teacher is in a position to
know what she asserts – for example, she has just to abandon “bad reasons” and
take at face value only evidence. Here it is another example: the perception of
an object, in absence of defeaters, provides sufficient reasons to believe that
there is an object even though the subject doesn’t form that belief – for
example, because she is distracted by other thoughts.[5]
Again, facts that one is in a position to know are a
subset of the facts that are in the epistemic field of a subject. If S is in a
position to know p, then p is also in the epistemic field of the
subject; but that q is in the
epistemic field of the subject doesn’t imply that S is in a position to know q. Smithies’s quote well illustrates
this point. One is in a position to know p
if one has an ungettiered justification to believe p, and one has such justification if one knows other facts or
perceives things from which it is legitimate to conclude that p. Being in a position to know
presupposes that the subject can come to know from his own mental states by
simple reflection or self inquiry, or by abandoning ‘bad reasons’. But p can be in the epistemic field of a
subject even if this subject doesn’t possess mental states from which she might
come to know that p by simple
reflection or self-inquiry. Both p
and other reasons to believe that p
can be completely unpossessed by the subject, though easily accessible (e.g.,
by inquiring further).[6]
An example well illustrates the above difference:[7] suppose that Silvie believes on good evidence
(the testimony of Mary) that there are five people at the party, but also
believes that Jack told her that Mary is unreliable on these matters (an
undercutting defeater of the former evidence), and she also believes that Jack
is very reliable. As a matter of fact, Mary is very reliable but Jack is not.
In this scenario, the defeater (Jack’s testimony) is sufficient for Silvie not
being in a position to know that there are five people at the party. But the
fact that there are five people at the party may still be in the epistemic
field of Silvie. For example, suppose that the party is just next-door of Silve’s
house and she can easily go to check how many people are at the party.
2.2. Arguments
against reasons as known facts
I will now introduce five arguments against the view
that reasons are known facts. The same arguments can extend to every view
holding that reasons are entities dependent on the actual perspective of the
subject (believed propositions; beliefs…).
Argument 1 – epistemic reasons
shouldn’t depend on our actual information set
If all epistemic reasons are
facts dependent on the actual perspective of the subject for which they
constitute reasons, we could modify the set of reasons there are for us to believe
simply by modifying our information set. This is obviously problematic, as
illustrated by the following example:
(Wishful Thinking) A mother is presented with overwhelming decisive evidence
that her son committed a crime. However, if she manages to forget this
evidence, as well as evidence that there is this evidence, there is absolutely
no reason for her to believe that her son is culpable.
This conclusion sounds seriously counterintuitive. It
seems obvious that the mother cannot get full epistemic justification to
believe what she wants simply by cancelling from her mind the evidence
supporting the unwanted conclusion. Notice that we are here concerned with a
strictly epistemic assessment of the mother in the present circumstance. Maybe by
inducing herself to forget evidence this mother is doing something good or
beneficial for her from a prudential point of view. This is utterly irrelevant from
an epistemic perspective. In other words, we are here interested in what are
the epistemic reasons of the mother
(as opposed to prudential, moral, aesthetic…). If we assume that only known
facts can constitute epistemic reasons, the answer is straightforward: the
mother’s efforts to forget good evidence succeed in substantially modifying her
set of epistemic reasons. They succeed to the point that for the mother now there is no reason at all to believe
that her son is culpable, and she is fully justified to believe that he is not.
I hope you will agree with me that this conclusion is deeply implausible. It
seems obvious that there are reasons for the mother to believe that her son is
culpable also after having forgotten the facts constituting evidence for the
relevant proposition.
The (Wishful Thinking) case exemplifies a situation in
which by subtracting knowledge we
change the set of epistemic reasons for a subject. There are also cases in
which we can reach equally counterintuitive consequences by adding new evidence/knowledge rather
than subtracting it. Here it is an example:
(Justified Irrationality) Jacques’s schoolteacher told him
that the Modus Ponens rule [A; A->B |– B] is actually a fallacy. For all
Jacques knows, this rule is invalid. Jacques knows that p, knows that p implies q, considers whether q. However, if reasons depend on the
subject’s actual perspective, there is absolutely no reason for Jacques to
believe q.[8]
This conclusion sounds
absurd: it seems that deducing q from
the known premises p and p -> q is a disposition constitutive of epistemic rationality. There seems
always to be epistemic reasons for an epistemic agent to deduce a conclusion
that follows by modus ponens from known premises. However, if all epistemic
reasons are actual-perspective dependent (e.g., known facts), our partial (possibly
misleading) information fully determines the set of reasons there are for us to
believe.
Argument 2 – The view doesn’t square
well with a range of data.
If we hold that epistemic reasons are limited to facts
dependent on the actual perspective of subjects, we find difficulties in
explaining a range of data about evidence possession, defeaters and knowledge.
Consider in particular the following points:
Unpossessed evidence. As many philosophers pointed out,
in our ordinary talks we often use a notion of evidence that is ‘unpossessed’
by subjects. Consider the following example from Skorupski (2010, 46): there
was enough evidence available for the police to have made a case against a
suspect (e.g., if they had conducted a disciplined inquiry), but this evidence
went unnoticed to the police. If, as it seems very plausible, evidence constitutes
epistemic reasons, then we must admit that there are epistemic reasons that we
don’t possess.
Unknown defeaters. Similarly, many philosophers have
argued that some of the facts that constitute epistemic defeaters are not facts
that we know.[9] These defeaters are reasons not to believe a
certain proposition. For example, the mere presence of many fake barns in a
countryside constitutes a reason not to believe that the barn that I see in
front of me is a real barn, even if I’ve no idea that there are fake barns
around. The view that reasons are only known facts cannot accommodate the
existence of unknown defeaters. On the contrary, if we assume that epistemic
reasons are not limited to facts we know, we would have an intuitive and
straightforward explanation.
Forgetting reasons for which one believes
doesn’t destroy knowledge. We forgot the reasons for which we believe many things we know. For
example, I have long forgotten the reason on the basis of which I believe that
America was discovered in 1492ac. If we think that only known facts can
constitute epistemic reasons, we should conclude that in all such cases we have
knowledge without reasons – or alternatively, that we lose knowledge of such
things in the moment in which we forget the reasons for which we believe them.
Both these claims are problematic. It seems perfectly possible to know such
things, even though we don’t know anymore the reasons for which we came to
believe them. If reasons do not depend on their possession (viz., their knowledge),
then we can maintain that there are reasons to believe such things that the subject
doesn’t actually possess (though once possessed).
Argument 3 - From epistemic modals to
perspective-independence of ER
A very popular view in epistemology is that evidence supporting the truth
of some proposition constitutes reason to believe that proposition. That the
suspect’s fingerprints are on the knife is evidence that the suspect committed
the crime, and also a reason to believe that the suspect committed the crime.
Another claim on which epistemologists and
semanticists widely agree is that epistemic modals such as ‘must’ and ‘might’
quantify over a domain of possibilities compatible with one’s evidence – or, in
different terms, that evidence constitutes the base of epistemic
modals. For example, if I say that tonight it might snow, I am
saying that it is compatible with the evidence that tonight it will snow; or
alternatively, that given the evidence there is a chance that tonight it will
snow.
Now, many have observed that some epistemic modals have
a more expansive base than total
knowledge of a subject or a group (e.g., Hacking 1967;
Anderson 2014). For example sometimes we include in the base of an epistemic
modal also merely knowable truths. According to Anderson (2014, 600):
“As examples already present in the literature
suggest, the relevant epistemic base can extend beyond what is known by the
relevant group to include what would be known were proper attention given to the
available evidence. Consider this case from Ian Hacking:
Imagine a salvage crew searching for a ship that sank
a long time ago. The mate of the salvage ship works from an old log, makes some
mistakes in his calculations, and concludes that the wreck may be in a certain
bay. It is possible, he says, that the hulk is in these waters. No one knows
anything to the contrary. But in fact, as it turns out later, it simply was not possible for the vessel to be in
that bay; more careful examination of the log shows that the boat must have gone down at
least thirty miles further south. (1967, p. 14)
Many find it intuitive that although no one in the
relevant group knew ~p, it was false
for the mate to assert ‘Might p’
because ~p was available in some
important sense to the relevant group.
This is evidence that there is a use of the epistemic modal such that
the relevant base is a more expansive base than the total knowledge of the
relevant group.”
As Hacking’s example and Anderson’s discussion well
illustrate, the set of facts relevant to the base of epistemic modals is not
exhausted by facts dependent on the actual perspective of one or more subjects,
such as the set of the propositions a subject actually knows. The widest set of
epistemically relevant circumstances can transcend the actual perspective of
the subject.
Form the above considerations it is possible to build
the following argument supporting the perspective independence of epistemic
reasons:
1) Bases of epistemic modals are constituted by evidence
(either total evidence or subsets of it).
2) Bases of epistemic modals can include facts beyond
the subject’s perspective.
4) Evidence
supporting the truth of some proposition constitutes reasons to
believe that proposition.[11]
C) Epistemic Reasons include facts beyond the subject’s
perspective (from 3 and 4).
Argument 4 – From excusable ignorance
In order to introduce the present argument, let me
first provide a rough characterisation of the notions of justification and
excuse. It is commonly held that an excuse is the type of defense that admits the violation of some relevant
undefeated norm and the consequent absence of sufficient reasons to φ, but points to factors that
rationalize the subject’s norm violation, explaining why the subject φ-ed. On the contrary, a
justification is a defence that points to conditions indicating that the
subject complied with the relevant norms in her context, i.e., she φ-ed for sufficient reasons or didn’t
φ impermissibly.[12]
The present objection relies on the observation that
knowledge-denials are often used as excuses, including as excuses for believing
something. This fact is at odds with the view that the reasons there are for a
subject to believe something are all and only the things she knows. According
to this view, if something is not known, it cannot constitute a reason to
believe. Thus knowledge-denials would deny the existence of a sufficient reason
to believe. This type of defence would indicate that the subject was right to
believe a certain proposition and should be counted as justified. On the
contrary, a knowledge-denial couldn’t count as an excuse for believing
something; this would involve an admission that there were sufficient reasons
for the subject not to believe a certain thing, but the subject was not aware
of them. However, if epistemic reasons are known facts, there cannot be
epistemic reasons one is not aware of.
The following case provides an example of a
knowledge-denial used as an excuse:
(Hangzhou Meeting)
Matteo and Xi should attend an important meeting in
Hangzhou. Matteo must go to Hangzhou from Rome. In Rome there are flights only to
Shanghai and to Beijing. Matteo ignores that [Hangzhou is closer to Shanghai
than to Beijing] (q). For all he knows,
the opposite is true. His secretary, who is usually quite reliable, told him
that Beijing is closer, and this is all he knows about the matter. On that
basis, Matteo forms the belief that he can reach Hangzhou faster if he flies to
Beijing (p), and flies to Beijing. As
a result, he arrives late at the meeting. At his arrival Xi, informed of the
misunderstanding and quite upset for the delay, engages in the following
dialogue with Matteo:
Xi: Why did you believe that you could reach Hangzhou
faster by flying to Beijing? This is obviously not the case!
Matteo: You are right. My apologies. I should have not
believed that. I didn’t know that Hangzhou is closer to Shanghai. For all I
knew, the opposite was true...
Matteo’s reply to Xi
clearly sounds like an apology, an implicit confession of and excuse for some wrong.
Matteo seems to beg for excuses adducing an explanation that rationalizes why he
acted wrongly (namely, that she inadvertently and blamelessly ignored the
relevant information). However, if epistemic reasons
depend on the actual perspective of subjects, Matteo’s response to Xi should not
sound as an excuse, but rather as a full justification. After all, Matteo is
claiming that he didn’t know that Hangzhou is closer to Shanghai than to
Beijing (q). So this proposition is
not a reason for him to believe not-p,
and should thus be irrelevant for his justificatory or exculpatory status. According
to the view under consideration, assuming that there are no other facts known
to Matteo indicating that not-p,
there are no sufficient reasons for Matteo to believe that not-p, and there are sufficient reasons to
believe p (namely, the known fact
that his secretary told him so). Matteo’s beliefs are in perfect conformity to his
reasons (i.e., coherent with all and only the facts he knew). Contrary to the
intuition about the case, the present view predicts that Matteo should be fully
justified to believe p, and shouldn’t apologize. However, since Matteo’s
excuses for believing p are perfectly in order in this circumstance, we
should conclude, by modus tollens, that he is not fully justified to
believe p, and thus that there are sufficient reasons for him to refrain
from believing p and to believe not-p in the circumstance.[13]
Therefore, epistemic reasons cannot be limited to facts that Matteo knows. Actual-perspective
views of epistemic reasons are false.
A related problem for the ‘reasons
= knowledge’ view is that Matteo seems to adduce as excuses facts unrelated to (what
this view counts as) reasons whether to believe p, such as the ignorance
of q (“I didn’t know
that Hangzhou is closer to Shanghai”).[14] According
to this view, in order to count as excuses, Matteo’s defences should point to a rationalizing
explanation of why he didn’t believe in accordance to the sufficient reasons there were for him to
believe. Since according to this view reasons are only known propositions, an
excuse for believing for insufficient reasons would amount to an excuse for not
believing compatibly to what Matteo knows. An instance of such an excuse would
be a reasonable explanation why he didn’t draw proper conclusions from things
he already knew. Unfortunately Matteo cannot make use of this type of excuses,
since his belief that p was perfectly coherent with all what he knew
about whether p. In short, according to this view, on the one hand,
ignorance of q cannot count as an excuse for not believing p; thereby
Matteo’s reply should sound as inappropriate and insufficient to excuse him. On
the other hand, there cannot be considerations Matteo could avail himself of in
his circumstances which could count as excuses.[15]
Argument 5 – Argument from the
Unity Thesis
Recently several philosophers have been interested in
principles connecting different kinds of reasons. In particular, some have argued more or less
explicitly for claims entailing or supporting the following principle
(Unity Thesis) A fact can count as an epistemic
reason if and only if it can count as a practical reason.
The Unity Thesis expresses the thought that the set of
facts that can potentially be practical reasons for a subject (i.e., which
would be reasons provided the existence of some act for which these facts would
be normatively relevant) is identical to the set of facts that can potentially
be epistemic reasons (again, which would be reasons provided the existence of
some content to the truth of which these facts would provide epistemic
support). If a
certain fact can count in favor of forming a belief for a certain subject at a
certain time, it can also count in favor of performing an action for the same
subject at the same time, and vice versa. For example, that the ground is wet
is both a reason for someone to believe that it is raining and to take an
umbrella (provided that one doesn’t want to become wet). The Unity Thesis
implies that the only feature distinguishing a practical reason from an
epistemic one is the type of thing that a fact supports (whether it is a belief
or an action). Other features, such as the epistemic standing that a subject
has with respect to a fact, are irrelevant to that distinction.
Philosophers provided very different arguments in
support of this thesis. Some of them point to our ordinary habits of deliberation. Fantl and McGrath (2009, 74-75) observe that we don’t
segregate reasons by whether they are available for drawing practical or
theoretical conclusions. Such segregation would be irrational and ‘barmy’. In a
similar vein, Littlejohn (2014) observes that it would be
akratic to seclude types of reasons, i.e., to take a consideration to
potentially support a belief but not an action, and vice versa. Others defend
versions of the Unity Thesis pointing to considerations of theoretical
simplicity and coherence: endorsing the Unity Thesis would provide a general
unified account of reasons, leading to a general theory of normativity valid
for all normative domains (Alvarez,
2010; Gibbons, 2010: 335; Kearns & Star, 2009). Many also support this thesis
on the ground that it is implied by specific accounts of reasons, accounts reducing practical reasons to epistemic
ones (e.g.,
Kearns & Star, 2009; Thomson, 2008), or vice versa (e.g., Steglich-Petersen,
2011).[16] Specific arguments for principles connecting
epistemic and practical reasons provide at least indirect support for the Unity
Thesis (e.g., Kiesewetter,
2016; Littlejohn, 2014; Way & Whiting, 2016).
A second premise in the argument is another very
popular view in contemporary philosophy, namely, the idea that practical
reasons are facts not dependent on the actual perspective of the subject they
constitute reasons for. They
are possibly unknown facts.[17] According to this view, certain
facts count as considerations for performing or refraining from performing a
certain act even in contexts in which a subject is unaware of these reasons. In
particular, there are reasons for someone not to perform certain actions
regardless of whether she possesses (knows or believes) such reasons or not.
For example, there are moral reasons not to torture and kill innocent people
irrespective of one’s epistemic position (e.g., whether there are reasons for Hitler
to kill millions of innocent people doesn’t depend on what he knows or believes)
and there are
prudential reasons for someone not to start smoking (e.g., that smoking will
cause negative health effects) regardless of whether one is actually aware of these
reasons or not.
From these two premises it is possible to mount an
argument for the claim that epistemic reasons are not dependent on the actual
perspective of subjects for which they constitute reasons. The argument is quite
straightforward. If practical reasons
are not dependent on the actual perspective of the subject, and every
fact that can count as a practical reason can also count as an epistemic reason
(from the Unity Thesis), Then also
epistemic reasons are not dependent on the actual perspective of the subject.
2.3. Against reasons as facts one is in
a position to know
In the previous section I considered five objections
to the view that reasons are known facts. The same objections also apply to
every view holding that reasons are entities dependent on the actual
perspective of the subject (believed propositions, beliefs…). In this
subsection I will briefly show that variants of these arguments can be directed
against the view that epistemic reasons
are facts that the subject is in a position to know (ERPK). For
presentational purposes, I will assume here a notion of ‘being in a position to
know’ restricted to what a subject can come to know by simple reflection or
self-inquiry into her own mental states or by abandoning ‘bad reasons’ (e.g.,
misleading presuppositions, prejudices…). Similar objections apply to slightly
different characterisations of this notion.
Argument 1* - According to
argument 1, if all epistemic reasons are facts dependent on the actual
perspective of the subject, we could modify the set of reasons there are for us
to believe simply by modifying our information set. This may happen either by subtracting or by adding
knowledge. The
problem, illustrated by problematic
cases such as (Wishful Thinking) and (Justified Irrationality), also applies to
every view according to which epistemic reasons are facts indirectly dependent
on our actual information set, such as the view that reasons are facts we are
in a position to know. The problem can be illustrated by a variant of (Wishful
Thinking). Suppose that the mother who manages to forget her evidence that her son committed a crime is not
in a position to retrieve the lost information from what she knows. This
implies that she is not in position to know the forgotten information about her
son’s culpability. According to (ERPK), we should then conclude that there is
no reason for her to believe that her son is culpable. As in the original case,
this conclusion
sounds seriously counterintuitive. It seems obvious that the mother cannot get
full epistemic justification to believe what she wants simply by forgetting
evidence supporting the unwanted conclusion. Similarly, in (Justified
Irrationality) imagine that for all Jacques is in a position to know, the Modus
Ponens rule is invalid. Again, even though Jacques knows that p, knows that p implies q, and
considers whether q, if (ERPK) is
right, there is absolutely no reason for Jacques to conclude that q. The lesson is, once again, that if all
epistemic reasons are depend on our actual set of mental states, our partial
information fully determines the set of reasons there are for us to believe,
with obviously counterintuitive consequences.
Argument 2* - (ERPK) has the same
difficulties of (ERK) in accounting for the existence of unpossessed evidence. Recall
Skorupski’s police example in §2.2: there may have been evidence available for
the police to make a case against a suspect even though the police was not in
the position to know that evidence by simple reflection or self-inquiry. If
evidence constitutes epistemic reasons, then (ERPK) is false. (ERPK) has also
difficulties accounting for knowledge based on forgotten reasons, as long as
these reasons are definitively forgotten facts, ones that one is not in a
position to know by simple reflection or self-inquiry.
Argument 3* - It is easy to see that the
argument from epistemic modals to perspective-independence of epistemic reasons
applies equally well to (ERPK). Consider Hacking’s example quoted in the
previous section: in that example, more careful examination of the log is
required to verify that the boat must have gone down at least thirty miles
further south. Therefore, the mate of the salvage ship was not in the position
to know that only on the basis of what he was in a position to know (e.g., by
mere reflection or self-inquiry about what he knew). This indicates that
evidence constituting the base of epistemic modals can include facts which go
beyond what a subject is in a position to know. If evidence supporting the
truth of some proposition constitutes reasons to believe that proposition, then
epistemic reasons include facts that the subject is not in a position to know.
Argument 4* - It is also easy to verify that
the argument from excusable ignorance applies to (ERPK). According to (ERPK),
knowledge-denials could excuse a belief only if the subject is in a position to
know the relevant information. Otherwise these denials can only justify the belief.
However, as the (Hangzhou Meeting) example shows, there can be cases in which
knowledge-denials can excuse a belief that p
even if one is not in a position to
know p. In that example, for all
Matteo is in a position to know, Hangzhou is closer to Beijing than to
Shanghai, and on that basis he should be fully justified to believe that he can
reach Hangzhou faster if he flies to Beijing, not merely excused. This
contrasts with the clear intuition that Matteo’s apologies are perfectly in
order. Furthermore, if (ERPK) were right, Matteo’s ignorance of the fact that
Hangzhou is closer to Shanghai could never count as an appropriate excuse. The
only type of considerations that would count as excuses for Matteo’s belief
should point to a rationalizing explanation of why he didn’t believe in
accordance to what he was in a position to know.
Argument 5* - A variant of the argument from
the Unity Thesis applies to (ERPK). First, notice that the standard view in
contemporary practical philosophy is that practical reasons for a subject are
facts potentially independent of what that subject is in a position to know. There
are reasons for one (not) to perform certain actions regardless of whether one
is in a position to know such reasons. For example, despite the fact that for
all Hitler was in a position to know it was perfectly right to torture and kill
millions of innocent people, there were no reasons for Hitler (or for anybody)
to do that. Now, if practical reasons are not dependent on what the subject is
in a position to know, and every fact that can count as a practical reason can
also count as an epistemic reason (from the Unity Thesis), also epistemic
reasons are not dependent on what the subject is in a position to know.
2.4. Skorupski’s view avoids problems 1-5 and 1*-5*
In the previous subsections I have shown that two
alternative views to Skorupski’s, (ERK) and (ERPK), are affected by five
problems. In this subsection I will show that the view of Skorupski is
apparently untouched by these arguments.
For what concerns arguments 1 and 1*, these arguments
rely on the thought that epistemic reasons shouldn’t depend on our information
set, neither the actual one, nor the one we are in a position to know. Skorupski’s
view seems not affected by this problem, at least prima facie. To see this, observe that in (Wishful Thinking), even
if the mother manages to forget her evidence that her son is culpable, we can
imagine that she may easily acquire again that or similar evidence which is still
in her epistemic field. For example we can imagine that she can receive again
the same information about her son’s culpability from the police. The same
point applies to arguments 2 and 2*: as long as unpossessed or fogotten
evidence is somewhat available to the subject (i.e., it is in her epistemic
field), it can count as epistemic reasons. Even though the police was not in
the position to know the evidence speaking for the suspect’s culpability, they
can easily discover the available clues that the suspect was the author of the
crime.
Skorupski’s view also avoids problems 3 and 3*. While
it seems obvious that we are not in a position to know some of the evidence
constituting the base of epistemic modals, it is far more contentious that this
evidence can transcend our epistemic perspective to the extent that is not even
in our epistemic field. Consider again Hacking’s example, since more careful
examination of the log would have been sufficient for the mate to see that the
boat must have gone further south, that evidence belongs to the facts in the
mate’s epistemic field. This type of cases doesn’t show that the base of
epistemic modals includes facts which are not in the epistemic field of the
subject.
Skorupski’s view seems also to be able to avoid the
argument from excusable ignorance. According to this view, knowledge-denials
can excuse a belief even if the subject is not in a position to know the
relevant information. It is sufficient that facts within one’s epistemic field support
that belief. In the (Hangzhou Meeting) example, we can imagine that Matteo
could have known that Hangzhou is closer to Beijing than to Shanghai by
engaging in further inquiry, for example by checking this information on his
iPhone. Thus this fact was in the epistemic field of Matteo, and thus it is
relevant for epistemic assessments and can count as an epistemic reason.
According to this view, while Matteo’s belief that he can reach Hangzhou faster
if he flies to Beijing is fully reasonable and excusable, it falls short of
justification.
It seems slightly more difficult for Skorupski’s view to
avoid argument 5*. A version of this argument applying to Skorupski’s view
requires the following premise: that practical reasons are not dependent on the
epistemic field of the subject for which they constitute reasons. Now, in
principle we can imagine that nothing in Hitler’s epistemic field spoke against
torturing and killing millions of innocent people. But of course this requires a
certain imaginative effort, and is not so intuitive as the cases supporting the
independence of practical reasons from what the subject is in a position to
know. I will come back to this point in the next section.
In the previous section I provided five arguments
against two alternative views to Skorupski’s one, namely, the view that reasons
are known facts, and that reasons are facts we are in a position to know. I’ve
also suggested that the view of Skorupski is apparently untouched by these
arguments. In this section, I will tentatively argue that epistemic reasons
extend beyond the facts within the epistemic field of a subject. In §3.1 I will
consider some potential problems for Skorupski’s view. In §3.2 I will assess
the prospects of a view according to which epistemic reasons are true propositions
possibly not included in the epistemic field of the subject, the ER=truths
view.
3.1. Possible problems for Skorupski’s view
Consideration
1
A fist set of considerations
against Skorupski’s view is that it doesn’t completely avoid complex variants
of the arguments considered in §3. Consider first a variant of argument 1*. According
to this view, the facts in the epistemic field of a subject fully determine the
set of reasons. However, these facts might constitute only partial and possibly
misleading information. Consider first a variant of (Wishful Thinking) in which
the mother manages not only to forget her actual decisive evidence, but also to
exclude every clue that her son is culpable from her epistemic field. For
example, she kills the only witness of the son’s culpability, make disappear
all traces of the dead body, brainwash her son… and after all this, she forgets
her decisive evidence that her son is culpable. Is there really no reason for
her to believe that her son is culpable at that point? Or imagine a variant of
(Justified Irrationality) in which no facts in Jacques’s epistemic field can
incline him to believe that Modus Ponens is
a valid rule. Following Skorupski’s view we should conclude that in this case there
is absolutely no reason for Jacques to believe q on the basis of her reflection on her beliefs that p and that p implies q.[18]
A variant of argument 2*
follows quite closely the modified (Wishful Thinking) case: completely and
definitively forgotten information does not belong to the epistemic field of
the subject, and thus cannot constitute reasons to believe. But then a subject who
learned p on the basis of
definitively forgotten data cannot know p
for such reasons. Suppose for example that a certain piece of information about
an historical event is stored in only one book. Jack learns that information.
Then he definitively forgets the reason on the basis of which he came to believe that information.
Furthermore the book is destroyed in a fire and nothing remains of that
information except in Jack’s memory. As a result, Skoruspki’s view implies one
of the following implausible consequences: either Jack knows that information
for no reason, or Jack’s loss of memory prevents him from knowing that information.
Furthermore, and even less plausibly, according to this view, reasons for Jacques
to believe that information cease to be such in the precise moment in which the
page of the book in which the information is written gets destroyed in the
fire.
For what concerns argument
4*, we can imagine a variant of (Hangzhou Meeting) in which there are no facts
in Matteo’s epistemic field indicating that Shanghai is closer to Hangzhou. For
example, a general internet blackout prevents Matteo from having access to that
information, some crazy guys burns all maps of China in Rome’s airport, and so
on. In these circumstances, Matteo should be fully justified to believe that
flying to Beijing is a faster way to reach Hangzhou, not merely excused. His
excuses would not make much sense. Furthermore, Matteo’s knowledge-denials
would turn from excuses to justifications in the precise moment in which Matteo’s
access to the relevant information is prevented (e.g., when the last map of
China in the airport is burned). This sounds quite odd. It seems obvious that
similar information’s accidental unavailability cannot transform an excuse into
a justification.
Consider a variant of argument
5*. Skorupski’s view presupposes an asymmetry in the extension of entities
which can count as epistemic and non-epistemic reasons. Some fact can count as
a practical reason without being an epistemic reason. For example, that there
is a treasure in my courtyard is a reason to dig there. However, suppose that
this fact is unknowable to me given the actual circumstances. Therefore, this
fact can count as a reason to act but not as a reason to believe. Now, Littlejohn (2014b) observes
that it would be akratic to seclude types of reasons depending on whether they
are reasons to believe or act. In particular, it is akratic to take a
consideration to potentially support an action but not a belief that one should
perform that action, and vice versa. For example, it seems akratic to take the
fact that there is a treasure in my courtyard as a reason to dig but not as a
reason to believe that I should dig, and vice versa.
In order to avoid akrasia, we
should take one of two opposite direction: either we consider practical reasons
as facts limited to one’s epistemic field, or we extend epistemic reasons
beyond facts within one’s epistemic field. I think that the latter option is
the less costly. As the treasure in the courtyard example shows, intuitively there
are reasons to act which are not in our epistemic field.[19] Here it is
another example: Suppose that Mani grew up in a community where people have no
clue of the negative effects of smoking. We can imagine, for example, that all
members of the society where Mani always lived smoke, and that this society
lived in a period in which there was no information about the dangers of
smoking. In this context, there are no facts in the epistemic field of Mani
indicating that smoking can cause negative health effects. Still, it seems
that, as long as smoking causes such effects, this constitutes a prudential reason for Mani not to
smoke.
Consideration 2 - Epistemic reasons are not closed under
conjunction
Skorupski’s view implies
that closure under introduction rule for conjunction fails for epistemic
reasons. This is because it is compatible with this view that p and q belong to the epistemic field of a subject S, but p&q does not. If epistemic reasons are only the facts within the
epistemic field of a subject, the following rule (Epistemic Field Closure) is
invalid (where “EFr” means that r is in one’s epistemic field):
(EFC)
EFp, EFq |– EF(p&q)
Consider some examples
illustrating the failure of (EFC). From Heisenberg’s uncertainty principle, a
scientist can easily come to know the position or the wave function of an
atomic particle but not both. Similarly, one can easily come to know that p and that [one doesn’t know p], but not their conjunction (Heylen 2016).
Now, if epistemic reasons
are restricted to the epistemic field of the subject, then both p and q constitute epistemic reasons for S, but not p&q. The following
principle is violated (Epistemic Reasons’ Closure under Conjunction):
(ERCC) If A belongs to the epistemic
reasons’ set of S and B belongs to the same set, then A&B belongs to the epistemic
reasons’ set of S.
This conclusion will sound
problematic to many holding that the set of evidence (epistemic reasons) is
closed under conjunction.[20] Furthermore,
if we assume a suitable form of evidentialism according to which epistemic
reasons are evidence, the failure of (ERCC) is incompatible with an account of
evidential support according to which propositions supported by evidence are
those most probable on one’s total evidence.[21] This simply
follows from the closure under conjunction of one’s total evidence when used as
probabilistic base to calculate the likelihood of a proposition. Suppose that S’s
evidence includes only p and q. Then the probability of r on S’s total evidence will be P(r|p&q). But p&q may not be part
of S’s evidence. And the probability of r
can be very different on p&q and on either p or q individually. If
one wants to preserve Skorupski’s account of epistemic reasons, one is forced
to adopt a compartmentalized model of evidential support, calculating
probabilistic support on proper partitions of total evidence that one is in a
position to jointly know. But at this point the upholder of this view will face
the further problem of determining which subset of evidence is relevant to decide
what one should believe.[22]
Furthermore, the closure
failure for the first relatum in the epistemic reason-relation can transmit to
the second relatum, the supported belief. Suppose that p is the only
reason for S to believe r and q is the only reason for her to
believe s. In this case, there will be reason for S to believe r
and to believe s, but no reason to believe r&s (the
only reason to believe this proposition being p&q). If S
believes what she has reason to believe, she ends up believing r,
believing s, but not believing their conjunction.
This violates plausible
principles of rational doxastic closure such as the following:
(RDC) If one rationally believes
that A and rationally believes that B, one also rationally believes that A and
B.
If something along the lines
of (RDC) is right, then someone may be irrational to believe only what she has
reason to believe. A rational subject shouldn’t believe only what she has
reason to believe. Both these conclusions sound quite problematic.
3.2. Epistemic
reasons = true propositions?
Until now I have considered three views relating
epistemic reasons to the perspective of agents, each in a different way and to
a different extent. I’ve also considered problems for each of these views. In
this section I want to explore the hypothesis that, differently from what these
views claim, epistemic reasons are not dependent on the perspective of
subjects, neither the actual, nor the potential one. More precisely, I will
assess the prospects of the view that the
total set of epistemic reasons is constituted by all and only the true
propositions, not only those belonging to the epistemic field of the
subject. I shall call this view the ER=T view. I will also try to address
some potential worries that this view might raise.
The ER=T
view avoids all the problems affecting the other considered views. This view
obviously and straightforwardly avoids problems 1/1*-5/5*. For example, it
allows that amongst the epistemic reasons of the mother in (Wishful Thinking)
there are also facts that she has completely and definitively forgotten. This
preserves the intuition that she persists having decisive reasons to believe
that her son is culpable even when these are forgotten. Similarly, in
(Justified Irrationality), even though for all Jacques knows the Modus Ponens
rule is invalid, he still has reason to believe in its validity (assuming that
this rule is true), and thus to believe that q. According to ER=T,
evidence can be unpossessed and definitively forgotten and still be epistemic
reasons for which the subject believes (and knows).
The view does not receive more support than Skorupski’s
one by the argument from epistemic modals. But it can easily avoid all variants
of argument 4. Even if there were no facts in Matteo’s epistemic field
indicating that Hangzhou is closer to Shanghai than to Beijing, the latter
proposition is true, and thus it is an epistemic reason for Matteo for so
believing. As a consequence, while Matteo is fully excusable for ignoring this
fact, he is not fully justified and his apologies are in order.
Finally, ER=T
avoids problem 5 and its variants in an elegant way: in conformity to the Unity
Thesis, it says that every fact that is a reason to act is also a reason to
believe and vice versa. That there is a treasure in my courtyard is both a
reason for me to dig there and a reason to believe that I should dig, and vice
versa; and that smoking is dangerous for health is both a reason to believe it
is and to stop smoking. More generally, this view holds that the set of all
facts constitute the set of all available reasons for a subject, where these
reasons can be taken as epistemic or practical depending on the type of thing
they support (an action or a belief).[23]
The ER=T
view also avoids the specific problem affecting Skorupski’s view. This view has
no problem accommodating reasons’ closure under conjunction: truth is closed
under conjunction. In this perspective, if p
is a reason for S, and q also is,
then p&q is also a reason. As a consequence, we avoid cases in which there is reason
for S to believe r and to believe s, but no reason to believe r&s. This view holds that every truth is a
reason to believe itself, as well as to believe other propositions it
evidentially supports. Within this framework, we will be still able to
distinguish the evidential support there is for a proposition from that that a
subject has. The latter will be assessed on the basis of the support that a
subset of total truths provide to a given proposition, where this subset is
constituted by propositions the subject knows or is in a position to know.
One might wonder whether the view I have just sketched
in not too extreme. Even assuming that epistemic reasons are facts which can
potentially not belong to the epistemic field of a subject, we can think, for
example, that epistemic reasons are still restricted to knowable facts, or to
facts that one can believe. Knowability and believability include a much wider
set of facts than those in the epistemic field of a subject. Such a restriction
would constrain epistemic reasons to the potential perspective of the subject.
This view would also be motivated by epistemic versions of the ‘Ought’ implies ‘Can’
principle. According to some philosophers, what we ought to do is restricted to
what we can do. Similarly, what we ought to believe might be restricted to what
we can believe (the ‘ought’ implies ‘can believe’ principle). Assume a case in
which the only reason to believe p is
r. If r is unknowable, it is also unknowable that we ought to believe p. This violates the ‘ought’ implies ‘can
believe’ principle.
I admit that a knowability view is a plausible
alternative to the ER=T view. Notice, however, that every view
limiting epistemic reasons to facts that the subject can know or have access to
is not exempt from the second problem affecting Skorupski’s view. There will be
cases in which epistemic reasons will not be closed under conjunction. An
obvious example is constituted by the case based on Heisenberg’s uncertainty
principle, where a scientist can come to know the position or the wave function
of an atomic particle but, as a matter of physical necessity, she cannot come
to know both. Failure of epistemic closure for epistemic reasons leads to the
problems considered at the end of §3.1.
While here I have not the
space to fully address potential objections to the ER=T view, let me briefly observe
that the very same issues arising for this view also arise for
perspective-dependent views of practical reasons (and for normative reasons in
general). As there
are moral dilemmas and laws requiring impossible things and not accessible to
their addressees, so it shouldn’t be surprising that there are unknowable
reasons and reasons to believe things that we cannot believe. As I ought to pay
taxes even if I don’t have money, it might well be that I have a reason to
believe that I should dig in my courtyard even though I cannot know that there
is a treasure there. In this
perspective, an unknowable reason can still be a reason – though possibly a
defective one, one that fails to perform the role of reasons of guiding agents
to do or believe what they ought to do or believe. Subjects committed to such
reasons can be excusable and blameless (at least under certain conditions
and to a certain extent), as would be agents that have no access to or cannot
comply with moral, legal or prudential reasons.
We can push the analogy between the practical and
epistemic domains even further. As in the practical domain we have reasons to
perform uncountable actions, but we cannot do all what we have reason to do
(e.g., give some money to every beggar we met on our way, or save the life of
every starving child), in a similar way we cannot believe all what we have
reason to believe. In the practical domain some reasons are more important, in
the sense that it is more urgent to comply with. The same is the case in the
epistemic domain. While every truth is a reason to believe, we are more
blameworthy and less excusable if we are insensitive to epistemic reasons that
are closer to our perspective than to those who are farther. Suppose that I
look out the window and I see bright sun. That there is bright sun is a reason
to believe that it is not raining. Suppose also that the email I just received
is a reason to believe that I must reply to it as soon as possible. But I have
not yet opened the email box. While I do not possess this reason now, I can
easily come to know it by simply checking my email box. In this situation, I
have both reasons to believe that it is not raining and that I should answer
the email, but while I would be epistemically blameworthy for not believing the
former, I am epistemically excusable for not believing the latter. And I am also
fully excusable and totally blameless for not believing that in my courtyard
there is a treasure when there is no clue at all that this is true.[24]
5. Conclusion
Skorupski in The Domains of Reasons argues that
epistemic reasons are facts in the epistemic field of the subject. In this paper,
I introduced Skorupski’s view about epistemic reasons and I compared it to two
other views: i) that reasons are actual-perspective dependent
facts, and ii) that they are facts that the subject is in a
position to know. I provided arguments favouring Skorupski’s view over these
alternatives. However, I have also suggested that Skorupski’s view is affected
by other problems. I also considered an alternative view according to which
epistemic reasons are all the true propositions, including those which do not
belong to the epistemic field of a subject. The latter view has its own
drawbacks, but it easily avoids the problems of the other views. While the
present essay has no pretence to provide a decisive case for the independence
of epistemic reasons from the perspective of the subjects, my hope is that it could
succeed at least in casting some doubts on the enduring dogma, well entrenched
in contemporary epistemology, that epistemic reasons are perspective-dependent
entities.
References
Alvarez, M. (2010). Kinds of
Reasons: An Essay in the Philosophy of Action. Oxford University Press.
Botterell, A. (2009). A Primer on the distinction
between justification and excuse. Philosophy Compass, 4(1), 172–196.
Broome, J. (2004). Reasons. In R. J. Wallace (Ed.), Reason
and Value: Themes From the Moral Philosophy of Joseph Raz (pp. 2004--28).
Oxford University Press.
Broome, J. (2013). Rationality Through Reasoning.
Wiley-Blackwell.
Bykvist, K., & Hattiangadi, A. (2007). Does
thought imply ought? Analysis, 67(296), 277–285.
Bykvist, K., & Hattiangadi, A. (2013). Belief,
Truth, and Blindspots. In T. Chan (Ed.), The Aim of Belief (pp. 100–122).
Oxford University Press.
Carnap, R. (1950). Logical Foundations of
Probability. University of Chicago Press.
Carrara, M., & Fassio, D. (2010). Perfected
Science and the Knowability Paradox. In M. D’Agostino, G. Giorello, F. Laudisa,
T. Pievani and C. Sinigaglia (Ed.), New Essays in Logic and Philosophy of
Science (College Publications, Vol. 1). London.
Dancy, J. (2002). Practical Reality (Vol. 78).
Oxford University Press.
Duff, R. A. (2006). Excuses, moral and legal: a
comment on Marcia Baron’s “excuses, excuses.” Criminal Law and Philosophy,
1(1), 49–55.
Dutant, J., & Littlejohn, C. (2016). Just do it?
When to do what you judge you ought to do. Synthese, online first.
Fantl, J., & McGrath, M. (2009). Knowledge in
an Uncertain World. Oxford University Press.
Fassio, D. (forthcoming). Is there an epistemic norm
of practical reasoning? Philosophical Studies, 1–30.
Fassio, D. (2014). Taking Norm-Regulation Seriously.
In D. Fassio, J. Dutant, & A. Meylan (Eds.), Liber Amicorum Pascal Engel
(pp. 760–777). University of Geneva. Retrieved from
http://www.unige.ch/lettres/philo/publications/engel/liberamicorum/
Gardner, J. (2007). Offences and Defences: Selected
Essays in the Philosophy of Criminal Law. Oxford University Press.
Gibbons, J. (2010). Things That Make Things
Reasonable. Philosophy and Phenomenological Research, 81(2), 335–361.
Gibbons, J. (2013). The Norm of Belief. Oup
Oxford.
Heylen, J. (2016). Being in a Position to Know and
Closure. Thought: A Journal of Philosophy, 5(1), 63–67.
Kearns, S., & Star, D. (2009). Reasons as
Evidence. Oxford Studies in Metaethics, 4, 215–42.
Kelp, C., & Simion, M. (forthcoming). Criticism
and Blame in Action and Assertion. Journal of Philosophy.
Keynes, J. M. (1921). A Treatise on Probability
(Vol. 31). Dover Publications.
Kiesewetter, B. (2016). You ought to F only if you may
believe that you ought to F. The Philosophical Quarterly.
Kolodny, N. (2005). Why be rational? Mind, 114(455),
509–563.
Littlejohn, C. (forthcoming). A Plea for Epistemic
Excuses. In F. D. Julien Dutant (Ed.), The New Evil Demon Problem.
Oxford University Press.
Littlejohn, C. (2012). Justification and the
Truth-Connection. Cambridge University Press.
Littlejohn, C. (2014). The Unity of Reason. In C. L.
John Turri (Ed.), Epistemic Norms: New Essays on Action, Belief, and
Assertion.
Lord, E. (2015). Acting for the Right Reasons,
Abilities, and Obligation. In R. Schafer-Landau (Ed.), Oxford Studies in
Metaethics (Vol. 10). Oxford: Oxford University Press.
Markovits, J. (2010). Internal Reasons and the
Motivating Intuition. In M. Brady (Ed.), New Waves in Metaethics.
Palgrave-Macmillan.
Markovits, J. (2011). Why be an Internalist about
Reasons? Oxford Studies in Metaethics: Volume 6, 6, 255.
McDowell, J. (1994). Mind and World. Cambridge:
Harvard University Press.
McHugh, C. (2014). Fitting Belief. Proceedings of
the Aristotelian Society, 114, 167–187.
Parfit, D. (2011). On What Matters: Volume One.
OUP Oxford.
Raz, J. (1978). Practical Reasoning. Oxford
University Press.
Raz, J. (1999). Engaging Reason: On the Theory of
Value and Action. Oxford University Press.
Scanlon, T. (1998). What We Owe to Each Other
(Vol. 66). Belknap Press of Harvard University Press.
Scanlon, T. (2014). Being Realistic About Reasons.
Oxford University Press.
Schroeder, M. (manuscript). Knowledge is belief for
sufficient (objective and subjective) reason.
Setiya, K. (2014). What is a Reason to Act? Philosophical
Studies, 167(2), 221–235.
Skorupski, J. (2010). The Domain of Reasons.
Oxford University Press.
Smithies, D. (2012). The Normative Role of Knowledge. Noûs,
46(2), 265–288.
Sorensen, R. (1995). Unknowable Obligations. Utilitas,
7(2), 247–271.
Srinivasan, A. (2015). Normativity without Cartesian
Privilege. Philosophical Issues, 25, 273–299.
Steglich-Petersen, A. (2011). How to be a teleologist
about epistemic reasons. In A. Steglich-Petersen & A. Reisner (Eds.), Reasons
for Belief (pp. 13--33). Cambridge University Press.
Talbot, B. (2016). The Best Argument for “Ought
Implies Can” Is a Better Argument Against “Ought Implies Can.” Ergo, an Open
Access Journal of Philosophy, 3.
Thomson, J. J. (2008). Normativity (Vol. 70).
Open Court.
Way, J. (2015). Reasons as Premises of Good Reasoning.
Pacific Philosophical Quarterly.
Way, J., & Whiting, D. (2016). If you justifiably
believe that you ought to Φ, you ought to Φ. Philosophical Studies, 173(7), 1873–1895.
[1] E.g., Alvarez, 2010; Dancy, 2002; Parfit, 2011; Raz, 1999;
Scanlon, 1998, 2014; Skorupski, 2010.
[2] Alvarez, 2010; Broome, 2004, 2013; Dancy, 2002;
Kolodny, 2005; Littlejohn, 2012; McDowell, 1994; Parfit, 2011; Raz, 1978, 1999,
Scanlon, 1998, 2014; Skorupski, 2010.
[3] Skorupski distinguishes epistemic reasons
from warranted reasons (or warrants): A Warranted Reason is self-accessible. It is
one that a person could come to know that she has it solely by means of
normative insight and self-examination, including attention to the way things
seem to her (2010, 108; see also 47, 107).
[4] Of course, the truth of this claim depends on the type of possibility
we have in mind. However notice that the epistemic field is restricted to those
facts that we could know at the present time by reflection, by further
inquiry, or by ‘stumbling across’ them. We would need a quite narrow notion
of possibility to make it plausible the entailment from ownership to the epistemic
field to knowability.
[5] According to Schroeder (manuscript), reasons are either
the contents of beliefs or of perceptual experiences.
[6] The notion of being
in a position to know is more similar to Skorupski’s warranted reason – though
there are important differences. For example warranted reasons are not factive,
and they are not facts but intentional objects (2010, §5).
[9] Again, a complete
list of the authors defending this claim is impossible. An incomplete list
includes Harman, Kelly, Skorupski, Sylvan, Ballantyne, Whiting, Gibbons and
Goldberg.
[10] Strictly speaking, 3
doesn’t follow deductively from 1 and 2. For example, one may conceive the
extension of facts constituting evidence as context-dependent, and take facts
counting as epistemic reasons as varying depending on the context of utterance
of epistemic modals. I find this view utterly implausible, since it introduces
a dependence of epistemic normative considerations on contextual factors such
as interests and conventions of speakers. However, for reasons of space, I
cannot discuss further the view and its problems. I will simply assume, for the
sake of argument, that the set of facts that count as reasons is not
context-dependent in this way. Notice however that even if this view were
correct, it would still be true that in some contexts total evidence includes
facts beyond the subject’s perspective. This would suffice to refute the view
that epistemic reasons are known facts.
[11] It is worth
mentioning that Skorupski (2010, §2.5 and ch.3) disagrees with the claim that
all evidence constitutes reason to believe. He distinguishes between epistemic
reasons and indicative evidence, which may even not be in the epistemic field
of the subject. I agree with Skorupski that some evidence is not in the
epistemic field of a subject, but I take this to indicate that there can be
reasons to believe which do not belong to this field. I will discuss this view
in §3.
[12] On the distinction between justification and excuse see, for example, Botterell, 2009; Duff, 2006; Gardner,
2007: ch.4-6; Kelp & Simion, forthcoming; Littlejohn, forthcoming.
[13]
Since Matteo’s excuses point to the unknown fact
that q, the latter is an obvious candidate to be such a sufficient
reason.
[15] Note also that Matteo’s excuses point to the
ignorance of q, that Hangzhou is closer to Shanghai than to Beijing.
This indicates that q was a reason for Matteo to believe p, even
though Matteo didn’t know q and didn’t form any attitude from it. More
in general, excuses for not F-ing which point to the ignorance of a proposition
r indicate that r was a reason for S to F. The epistemic domain
doesn’t seem to constitute an exception to this rule: S’s ignorance of r
excludes blameworthiness and makes rational for S to believe as she did, but it
doesn’t exclude epistemic norms’ violations and indicates that r was reason
about whether to believe the relevant proposition.
[16] Other views which
seem to imply the Unity Thesis are those defended by Setiya, 2014 and Way, 2015. See also McHugh, 2014: §VI.
[17] See, for example, Alvarez, 2010;
Broome, 2004; Parfit, 2011; Skorupski, 2010. A complete list of authors
endorsing this view would be too long to find a place in this paper. Very few
philosophers disagree with this view.
[18] A possible reply may
be that the Modus Ponens rule is a
fact known a priori or necessarily belonging to the epistemic field of the
subject. However, this response presupposes quite contentious assumptions about
what a subject is able to necessarily know or access.
[19] Some philosopher even
holds that there are unknowable practical reasons. See, for example Dutant & Littlejohn, 2016; Sorensen, 1995;
Srinivasan, 2015. Consider a specific
example in Markovitz (2011, fn 13): “Suppose I love
surprise parties, but hate parties I know about in advance. It seems like, if
my friends are throwing me a surprise party at their house, it would be best for
me to go there.” This case exemplifies a reason to act (to go to the party)
which is unknowable, for if known, it would cease to be a reason to go to the
party. Another example is the reason to inquire further into a fact which is
unknown, where it is in part its being an unknown fact which makes it a reason
to inquire further. While I find such kind of examples compelling, my argument
does not rely on the claim that there are, or there can be, unknowable reasons.
Not being in the epistemic field does not imply unknowability.
[20] Though one may argue
here that there are such cases. An example is provided by the preface paradox,
in which a subject is in position to know each single sentence in the book but
not their conjunction. Notice however that the failure of (EFC) would extend to
a multitude of cases having few analogies with the structure of the preface
case. If (ERCC) fails also epistemic reasons’ simple closure under modus ponens
fails ((Heylen, 2016). (MPC) is false:
(MPC) PKp, PK (p->q) |– PKq
If epistemic reasons are facts in the epistemic field of a subject, then
there are cases in which p and p -> q are part of S’s epistemic reasons, but q is not.
[21] The idea that
probabilistic epistemic support must take into account all the available
information is usually associated to Carnap and his Principle of Total Evidence
(Carnap, 1950: 211–13;
Keynes, 1921: 313).
[22] Skorupski seems to
be aware of this problem. See 2010, §2.4.
[23] However this view can
preserve a difference between reasons there
are for someone to φ and reasons that one
has and on the basis of which one
φ-s. This view allows for assessments related to the perspective of the
agent or the subject. Elsewhere I have argued that at least some of these assessments
are instrumental evaluations related to the regulation conditions of reasons
and norms, i.e., to the ways in which the subject is sensitive to and tries to
follow these reasons and norms. See Fassio, forthcoming,
2014. Other such perspectival
assessments concern excusability and blameworthiness. Another interesting
feature of the ER=T view is that it
considers all epistemic reasons ‘thinker-neutral’. The reasons there are for
some subject to believe a proposition are the same there are for all subjects
(though the reasons that each subject has
are of course different).
[24]
The
above considerations partially address also a further problem discussed by Bykvist &
Hattiangadi, 2007, 2013. There are
propositions that it is logically impossible to believe (or rationally
believe). Consider the following example:
(UP) A tree is
falling in the forest but nobody will ever believe that.
If (i) epistemic reasons are all truths, (ii) each truth constitutes a
reason to believe itself, and (iii) there are unbelievable truths such as UP
(for those who are not idealists there are such truths), then there are reasons
to believe unbelievable propositions. Again, it is unclear whether this is a
problem rather than a simple matter of fact: as there are laws requiring people
to pay taxes they cannot pay, there are reasons to believe unbelievable
propositions such as (UP) (though we are fully excusable for not believing
them).
