Jonathan
Adler has written an ambitious book. He
has set himself the task of rehabilitating the epistemological view called
‘evidentialism,’ by conceptual analysis. 
Evidentialism is the view that, in order for a belief to be properly
held, it must be supported by adequate reasons. This view, famously propounded by W. Clifford[1]
and just as famously criticized by William James[2],
has been unpopular in recent times.

Adler’s
strategy is to show that, in order for something to count as a belief in the
first place, it must be held for what the believer takes to be adequate
reasons. This is a constraint imposed
by the concept of belief, which can be seen in the fact that sentences like “It
is raining, but I don’t have adequate reason to believe it is raining” sound
paradoxical. The book begins with an
Introduction, distinguishing between intrinsic and extrinsic approaches to the
subject matter. An approach is
intrinsic if it finds the criteria for appropriate belief in the concept of
belief itself; it is extrinsic if it locates the source of those demands
anywhere else. This is similar to, but
distinct from, the internalist/externalist distinction in epistemology. The first chapter takes up this distinction
to argue that the twentieth century move to extrinsic theories has been a
mistake, and that therefore much of contemporary epistemology is misguided.

Chapter two
purports to show that one of the reasons philosophers have given for abandoning
the intrinsic approach is an inadequate reason, based on a misunderstanding of
what evidentialism requires. Many have
thought that if we say that we should believe only in proportion to our
evidence for the belief, then we are committed to the view that our beliefs are
under our direct voluntary control. 
After all, we can be obliged to do something that we can’t do; ‘ought’
implies ‘can.’ Here Adler makes a case
for a distinction between obligation and responsibility, which nicely solves
the problem; we can be responsible for our beliefs even we can’t voluntarily
adopt or abandon them.

If
evidentialism is true, it follows that we can talk about epistemic
responsibilities in a substantively normative way. Chapter three explores some of the ramifications of that
fact. The fourth chapter continues the
theme of answering the critics of evidentialism by exposing fallacies in
various particular arguments. Chapter five
is devoted to testimony, a topic of current interest, and one to which Adler
has already made an important contribution.[3] Chapter six addresses another popular
criticism of evidentialism, that it is committed to foundationalism. Foundationalism is the view that, in order
to avoid either circularity or infinite regress, we must admit that some
beliefs get their justification (or warrant, or other positive epistemic
status) from something other than evidence. 
Here Adler invokes the notion of background belief to show that one can
reject foundationalism without circularity or infinite regress. Any given belief, even ones that seem to be
held without support from other beliefs, will be supported by masses of
background evidence.

Chapter
seven shows the virtues of Adler’s evidentialism by showing how it can be used
to solve three paradoxes of belief, one of which—the Preface Paradox—has vexed
epistemologists for some time. Here
Adler invokes, besides his evidentialism, a distinction between partial belief
and full belief (a distinction already familiar in the literature). The idea is that in many cases of apparently
paradoxical belief, one or more of the beliefs in question, if believed at all,
is believed only partially. Full belief
is typically expressed by simple declarative sentences, such as “It is
raining”; partial beliefs are typically expressed as qualified sentences, such
as “I believe it is raining” or “I’m fairly sure it is raining.” Chapters Eight, Nine, and Ten address
various concerns about this distinction and how it fits into Adler’s
project. Chapter Eight argues that on
many occasions, we are constrained to believe fully, so it can’t be that all we
have is degrees of partial belief; Nine discusses a useful distinction between
degree of belief (which would necessarily be partial belief) and degree of
confidence in a belief (which could apply to full belief). Ten then invokes this distinction to defend
the compatibility of full belief with doubt. 
The final chapter explores ways we might, on this view, be able to
control and correct our belief systems.

Adler’s
book is impressive both for its ambitious project and the scope of its
application. He clearly has a good grasp
on a wide range of epistemological issues and theories. There are places where he admits his theory
is less than comprehensive, but this lack is by design. For example, he says nothing about the
various views of belief discussed in philosophy of mind and cognitive
science. This is because a good theory
of ordinary, everyday belief should not presuppose any substantive views on
these matters. Whatever belief is, we
have been wielding the concept since long before cognitive science was born.

Some gaps are
not so easy to forgive. Adler has
little to say about what is meant by “adequate evidence” for a belief, beyond
that it should be conclusive. However,
this notion cries out for explication. 
One natural way to understand conclusive evidence is that the evidence
logically entails the belief, but Adler rightly disavows this interpretation as
too strong. But what does it take,
then? The reason this matters is that
we must know what the evidential requirement amounts to before we can decide if
it is credible. This much Adler gives
us: Conclusive reasons need not rule out all logically possible alternatives;
that is, I can have conclusive reasons for believing it is raining even if I
cannot rule out, on my own resources, that I am a brain in a vat. Moreover, what counts as conclusive evidence
is sensitive to context; in some context, where the stakes are high (say, lives
are at stake), I must meet a higher standard of evidence. But he can’t leave it at that. When he discusses background evidence to
answer the foundationalist regress argument, he allows that ordinary inductive
evidence, and even coherence, can give support to beliefs that can contribute
to their grounding (180-181). But when
he discusses the compatibility of full belief and doubt, he denies that
inductive reasons against a belief count as evidence against it (259). Likewise, in chapter five, in his discussion
of testimony, Adler allows judgments of probabilty to be evidence for beliefs
(148-149), but in his discussion of the lottery paradox (188-189), probability
judgments are contrasted with evidence. 
But these are problems that may well be resolved with further inquiry.

On the
whole, Adler’s book is an impressive contribution to epistemology. I recommend it to academics that wish to
clarify their thinking on belief, assertion, and evidence. It should be especially useful in the
upper-level or graduate classroom, as a way to bring together many different
strands of contemporary epistemology.