Defeaters in Epistemology

The concept of epistemic defeat or defeasibility has come to occupy an
important place in contemporary epistemology, especially in relation
to the closely allied concepts of justified belief, warrant, and
knowledge. These allied concepts signify positive epistemic appraisal
or positive epistemic status. As a first approximation, defeasibility
refers to a belief's liability to lose some positive epistemic status,
or to having this status downgraded in some particular way. For
example, a person may be epistemically justified in believing some
proposition p at one time, but then the belief might become less
justified or even unjustified at some later time. Moreover, beliefs
may also be prevented from having or acquiring some positive epistemic
status in the first place. So more generally, defeasibility refers to
a kind of epistemic liability or vulnerability, the potential of loss,
reduction, or prevention of some positive epistemic status. A defeater
is, broadly speaking, a condition that actualizes this potential. This
article begins by outlining two general types of defeaters:
propositional defeaters and mental state defeaters. Propositional
defeaters are conditions external to the perspective of the cognizer
that prevent an overall justified true belief from counting as
knowledge. Mental state defeaters are conditions internal to the
perspective of the cognizer (such as experiences, beliefs,
withholdings) that cancel, reduce, or even prevent justification.

1. The Concept of Defeasibility
a. Defeasibility: Legal, Moral, and Epistemic

The language of defeasibility is not unique to epistemology. In fact,
its use in epistemology is arguably derived from its use in legal and
moral discourse. For example, H.L.A. Hart (1961) borrowed the term
"defeasibility" from its prior uses in property interests and applied
it to contracts. Hart explained that though contracts were comprised
of an offer, acceptance and consideration, contracts may still be void
or voidable due to some exception such as fraud or incapacity. In
making this application to contracts, Hart noted that there is no
specific term in the English language to refer to exceptions to a
basic legal rule (Hart, 1961, p. 145; cf. Boonin, 1966). The
defeasibility of legal rules is analogous to the defeasibility of
moral rules in ethics or moral philosophy. While there may be
obligations to do X, many ethical theories add that at least some of
these obligations are only prima facie duties. They can be overridden
by other factors and thus are no longer morally binding. Moral rules,
like legal rules, are subject to being defeated in particular
circumstances or under particular conditions.

Talk of defeasibility in the legal and moral context translates into
epistemic defeasibility in at least one obvious way. If we think of
positive epistemic status as normative, then this status will – like
moral and legal rules – be subject to being overridden by other
factors. In circumstance C we may be epistemically justified to
believe p, just as we are legally or morally justified to perform
action A in circumstance C. In other circumstances C*, though, we may
no longer be epistemically justified to believe p, just as we are not
legally or morally justified to perform action A in circumstance C*.
This is particularly evident in deontological conceptions of epistemic
justification, according to which we have various intellectual
obligations and certain epistemic principles forbid believing p under
certain circumstances, for example when p is not likely to be true or
when p is likely to be false. But even if we think of justification
simply in terms of having adequate evidence, justification will be
variable. Chisholm (1966, 1989, pp 52-69), for example, notes that
while evidence e may make h evident, another evident proposition, d,
may defeat the tendency of e to make h evident because the conjunction
of e and d does not make h evident. In other words, there may be a
loss of justification when new evidence is added to an existing
evidence base.
b. Defeaters in Epistemology: Basic Distinctions

Defeater theories are generally distinguished by how they construe
what does the defeating and what gets defeated.

(i) While some philosophers construe defeaters as conditions external
to the perspective of the cognizer (true propositions), others
construe them as conditions internal to the cognizer (mental states
such as experiences or beliefs). Hence, while some philosophers might
regard the true proposition "There is a blue light shining on the
widgets" as a defeater for a belief about the color of the widgets,
others would regard the subject's belief that "There is a blue light
shining on the widgets" as the defeater. What does the defeating in
the first case is a certain fact (the obtaining of which is
independent of a cognizer's beliefs or perspective). What does the
defeating in the second case is a mental state of the cognizer.

(ii) Philosophers who construe defeaters as true propositions usually
take defeaters to be conditions that prevent an overall justified true
belief from counting as knowledge. So if the true proposition "There
is a blue light shining on the widgets" is a defeater it would prevent
my belief that "This widget is blue" from being something I know, even
if this belief is justified and true. On the other hand, philosophers
who take defeaters to be mental states of the cognizer tend to see
them as defeating the justified status of a belief, either by
downgrading the degree of justification or by canceling the justified
status of the belief altogether. In this case, having a defeater for
my belief that "This widget is blue" entails that this belief, even if
true, is no longer justified or justified to the same degree. Of
course, if justification (to some high degree) is necessary for
knowledge, defeaters that defeat justification may also prevent a true
belief from counting as knowledge.
2. The Gettier Problem and Propositional Defeaters
a. The Tripartite Definition of Knowledge and the Gettier Problem

One of the primary tasks of epistemology is the examination of the
nature of knowledge. One aspect of such inquiry is the analysis of
those conditions that are severally necessary and jointly sufficient
for knowledge. There have been three fairly widespread and
long-standing intuitions concerning knowledge in the Western
philosophical tradition. First, a person S's knowing some proposition
p entails that p is true. Second, though more controversially, S's
knowing that p entails that S believes or assents to p, perhaps
firmly. Third, knowledge is not equivalent to true belief. Knowledge
has a certain surplus value over true belief. The ancient Greek
philosopher Socrates indicated this surplus value metaphorically by
speaking of knowledge as true belief that has been "tied down" or
"tethered." Much of the work of epistemologists in the second-half of
the twentieth century has been devoted to examining candidates for
this epistemological tether, a plausible condition (or set of
conditions) that can transform a true belief into knowledge. The term
"justification" is commonly used to designate this condition. A
justified belief is roughly one that has a positive tie or strong
connection to the truth goal of believing, something like
"../evidence/">evidence, grounds, reasons, or processes of belief
formation that are in some sense indicative of the truth of the
belief. The so-called traditional or tripartite definition of
knowledge as justified true belief expresses all three of the above
intuitions.

However, owing to Edmund Gettier's arguments (Gettier, 1963),
epistemologists have generally recognized that justified true belief
accounts of knowledge suffer from a basic defect or inadequacy.
Gettier argued that there are cases in which an individual could
plausibly be said to have a true belief that is justified but which
fails to constitute knowledge. For example, I might be justified in
believing that "either Jones owns a Ford or Brown is in Barcelona"
because I validly deduce it from a justified belief "Jones owns a
Ford." If Jones does not own a Ford but Brown happens to be in
Barcelona, I will have inferred a true justified belief from a false
justified belief. However, it seems counterintuitive in this case to
suppose that I know that Brown is in Barcelona, even if the belief is
true and justified.

One of the early proposals to handle the Gettier Problem involved
adding a fourth condition to knowledge that excludes inferences from
or dependence on any false beliefs (Shope, 1983, pp. 81-118). But
Gettier cases can be generated where there is neither an inference
from nor dependence on any false beliefs (Steup, 1996, pp. 15-16). So
other strategies must be employed to deal with Gettier
counterexamples. One of these strategies employs the concept of
defeasibility or defeaters (Lehrer and Paxson, 1969; Swain, 1974;
Shope, 1983).
b. Defeasibility Analyses and Propositional Defeaters

Defeasibility analyses of knowledge come in a variety of different
specific versions. The generic idea is that a person S knows p only if
there is no true proposition, d, such that if S were to believe d (or
d were added to S's evidence for p), S would no longer be justified in
believing p. In other words, the existence of certain unpossessed
evidence prevents a person from actually knowing p if this unpossessed
evidence would result in a loss of justification were the person to
acquire the evidence, be aware of it, or recognize it. So according to
defeasibility theories, it's a true proposition that does the
defeating, not a believed proposition. Following Bergmann (2006, p.
154), I'll refer to these kinds of defeaters as propositional
defeaters. So according to defeasibility analyses of knowledge we must
adopt the view that:

[PD] S knows that p only if there is no propositional defeater d
for S's belief that p.

Consider the so-called "Fake Barn" scenario, an often-cited
Gettier-type case used by Alvin Goldman (Goldman, 1976, pp. 772-73).
Suppose Henry is driving through a Wisconsin town, admiring the
scenery. He sees a barn and believes "there's a barn." Unbeknownst to
Henry, this Wisconsin town is full of papier-mâché barn facsimiles,
which look like real barns when viewed from the road. However, the
structure Henry happens to look at is a genuine barn. He just happens
to glance in the direction where one of the few real barns is located.
His belief is true since he's looking at a genuine barn. He also
appears justified in holding this belief. Henry believes what seems to
him to be the case. He has no reason to believe that anything is
suspicious about his perceptions, much less that he's in a town mostly
populated with fake barns. He also knows that barns are fairly common
in this part of the state. Nonetheless, it seems that, however
justified Henry may be in holding this belief, he doesn't know that
there is a barn present. He is of course lucky to believe what is true
in this circumstance, but it's precisely this feature of the situation
that raises doubt about whether he knows there is a barn before him.
Had he looked at any other time, his eyes would have landed on a fake
barn and his resultant belief would have been false. Knowledge would
seem to require that it not be a matter of epistemic serendipity that
one's belief is true.

Defeasibility analyses of knowledge attempt to relate the problem of
accidentally true belief to the existence of some bit of relevant
unpossessed evidence. That is, it is in consequence of lacking some
relevant evidence, of being less than ideally situated with respect to
the evidence, that a person ends up luckily believing what is true.
This is illustrated in the Fake Barn scenario. In that case, there is
a true proposition such that, if Henry were to believe it, he would
not have been justified in believing that the object he sees is a
barn. The true proposition would be something to the effect that "in
this town nearly everything that looks like a barn isn't actually a
barn." Call this proposition D, and call his barn belief B. If Henry
were to believe D, he would not be justified in his belief that B.
Alternatively, we might say that if D were added to Henry's actual
evidence E (the evidence of his senses and relevant background
beliefs), he would no longer be justified in holding the belief that
B. Given E, Henry is justified to believe B, but given the conjunction
of E and D, Henry is not justified in believing B. For Henry to know
that there's a barn present, it must not be an accident that this
belief is true. This in turn requires that Henry's justification be
indefeasible.

We should underscore that there being a propositional defeater for
Henry's belief that "there's a barn" does not entail that Henry is
actually unjustified in believing "there's a barn" or that he's
irrational or unreasonable in holding this belief. The point about
justification is a counterfactual one: Henry would not be justified in
believing "there's a barn" if he were to believe "in this town nearly
everything that looks like a barn isn't actually a barn" or if this
fact were added to his evidence. The counterfactual truth about
justification entails that Henry doesn't actually know "there's a
barn," not that he's unjustified in believing it. Of course, if we're
thinking of knowledge as simply justified true belief, we might speak
of Henry's justification being defeated in some way because the
justification is insufficient for knowledge (Lehrer and Paxson, 1969).
The target belief may be justified, but the justification is
"defective" (Marshall Swain, 1981, p. 148) because it fails to make
his true belief knowledge. Steup (1996, p. 15) captures this point by
speaking of the epistemizing potential of a person's justification
being defeated, and contrasts this with saying that a person's
justification is defeated. While Shope (1983, p. 47) speaks of S's
actual justification being defeated, by this he simply means that the
justification fails to be enough – together with the satisfaction of
the truth and belief conditions – for knowledge. And so also with
other authors who use similar language at this juncture. So we should
say that a propositional defeater for S's belief that p doesn't entail
that S is no longer justified in believing p, only that S's
justification isn't sufficient (along with true belief) for knowledge.
Technically, then, we should speak of knowledge being defeated (Audi,
1993, pp. 185-213) or warrant being defeated (Plantinga, 2000, p.
359-60), where warrant is the property that transforms true belief
into knowledge.
c. Constraints on Propositional Defeaters

As widely discussed in the early literature on defeasibility theory
(Lehrer and Paxson, 1969; Annis, 1973; Swain, 1974), the main
challenge facing defeasibility analyses of knowledge is to specify the
relevant range of true propositions that can function as defeaters. It
is generally acknowledged that not just any true proposition
(suggestive of a defect in justification) is an efficacious defeater.
There are genuine defeaters, but there are also misleading defeaters.

In the famous so-called Tom Grabit case (Paxson and Lehrer, 1969), I
see a man who looks to me like Tom Grabit remove a book from a library
bookshelf, slip it under his coat, and escape the library. I believe
that Tom Grabit stole a library book. As it happens, the man I saw was
indeed Tom Grabit, and he did steal the book. However, let's suppose
further that Tom Grabit's mother claims that on the day in question
Tom was out of the country but that Tom's identical twin brother John
was at the library. Here it seems that there is a true proposition
such that if I were to believe it, I would not be justified in
believing that Tom Grabit stole a library book. The true proposition
is "Tom Grabit's mother is testifying that. . . ." Call this true
proposition D, the ostensible defeater. It would seem that, like in
the case of Fake Barn, there is a propositional defeater for the
target belief. I may in fact have a justified true belief that there
is a barn over there, but the justification is defective and so my
justified true belief does not constitute knowledge. The true
proposition D is such that if I were to believe it (or add it to my
evidence), I would no longer be justified in believing that Tom Grabit
stole the library book. But now suppose that Mrs. Grabit is actually a
compulsive liar and Tom's twin brother is the product of Mrs. Grabit's
demented imagination. Tom Grabit is not out of the country and he has
no twin brother. Given this expansion of facts, our intuition may now
be that I do know that Tom Grabit stole a library book, that Mrs.
Grabit's testimony does not actually defeat my knowing that Tom Grabit
stole the book.

While we might say that there is a propositional defeater for my
belief that Tom Grabit stole the library book, we can say one of two
possible things about the defeater's lack of defeating efficacy.

First, the defeater in the Tom Grabit case is clearly misleading. It
is perhaps natural to say that it misleadingly suggests that that the
target belief is false or that the evidence for the target belief
isn't good. The defeater is a true proposition, for it is true that
Mrs. Grabit said that Tom's twin brother, not Tom, is in the library,
and that Tom is out of the country. The problem is that this true
proposition suggests that my belief that Tom Grabit stole the book is
false or that I shouldn't be relying on the evidence of my senses. It
also suggests other false propositions, for example that Tom Grabit
has an identical twin, that Tom was not at the library, or depends on
the false assumption that Mrs. Grabit is sane and her testimony
reliable. At all events, what is required is a genuine as opposed to
misleading defeater, and such a defeater will not presuppose, suggest,
or depend upon some falsehood (Klein, 1976, 1981).

Secondly, we might say that the potential defeating effect of D is
neutralized or defeated by some further true proposition, D*, such
that if I were to believe D* I would not be justified in believing D.
In this case, the true proposition, D*, is that Mrs. Grabit is a liar
and mentally deranged, whereas D is simply the fact of her testimony.
It seems that D defeats my belief that Tom stole the library book
because if I believed D, I would cease to be justified in believing he
stole the book. But if I were to believe D*, I would not be justified
in believing the content of Mrs. Grabit's testimony. In other words,
the total evidence set includes D and D*, but D* defeats D. A genuine
defeater must be undefeated by any further evidence (Barker, 1976;
Pollock, 1986; Swain, 1974).

Other epistemologists suppose that what defeats knowledge is
unpossessed evidence that most of the members of the person's society
or social group are aware of. We can use the example provided by
Gilbert Harman (1973, pp. 143-44). Suppose that a political leader has
been assassinated. A reporter who is a witness to the assassination
dictates details of the event to his news agency so that the story may
be included in the day's final edition of the paper. Jill picks up the
paper and reads the story and believes that the political leader has
been assassinated. However, before Jill picks up the newspaper and
reads the story, loyalists to the political leader declare on
nationwide television that the bullet actually struck and killed
someone in the political leader's entourage. Jill reads the true story
in the paper but misses the false report on television. Harman
contends that in this hypothetical situation Jill doesn't know that
the political leader has been assassinated. Some epistemologists
(Swinburne, 2001; Pollock 1986) contend that Jill's not having
knowledge in this case is the consequence of there being a true
proposition (suggestive of a defect in justification) that is widely
believed in Jill's society. (Advocates of this view would also seem
committed to saying that if the Tom Grabit example were altered so
that Mrs. Grabit testified in a public venue to the alleged
whereabouts of Tom and the existence of Tom's identical twin brother,
then her testimony would be a genuine defeater for someone's knowing
that Tom stole the book, even if Mrs. Grabit were lying or deranged).

Alternatively, we might suppose that the crucial factor that
determines whether a true proposition (suggestive of a defect in
justification) is an efficacious defeater is if the unpossessed
evidence is the sort of thing that is easily accessible. We can take
another example from Harman (1973). Suppose your good friend Donald
tells you that he's going to Italy for the summer. You take him to the
airport and see him off. He left in June, but in July he decides to
send you several letters informing you that he's actually in San
Francisco. This is not true. He's simply trying to fool you. He sends
the letters to another friend in San Francisco who is instructed to
send them to you one at a time, as if they were sent from Donald,
complete with a San Francisco postmark. You've been gone for a couple
of days, though, and your mail has piled up. There are two letters in
the stack from Donald. You haven't looked at them yet and so you
believe that Donald is in Italy. This is true, but there's evidence of
which you are not aware that would justify you in believing that
Donald is not in Italy. It might be argued that in this case, the
information contained in the unopened letters constitutes a genuine
defeater for your belief that Donald is in Italy since the information
is near at hand, readily available to you, even though in fact you're
not aware of it.

There are of course other variations on genuine defeaters. We might
throw a deontological spin into the defeasibility account. We might
suppose that unpossessed evidence defeats knowledge only if the
evidence is the sort of thing the person should believe and would
believe if certain intellectual obligations were satisfied. At all
events, all these defeasibility formulations are ways of placing
constraints on propositional defeaters. They each recognize that while
there are many true propositions that seem to indicate a defect in
justification (that is, such that if S were to believe them, S would
cease to be justified in his original belief) only some of these
entail an actual defect in one's justification, actually defeat the
person's knowing the target proposition.
3. Mental State Defeaters and General Epistemology

While defeasibility accounts of knowledge take defeaters to be facts
external to the perspective of the cognizer, another approach to
defeaters construes them as items internal to the perspective of the
cognizer, as mental states such as experiences, beliefs, or
withholdings. For example, on a particular day I see a person who
looks like Tom Grabit steal a book from the library. Based on my
sensory perceptual experience and my memory beliefs about what Tom
Grabit looks like, I believe that Tom Grabit stole the library book.
Later that day Tom Grabit's mother tells me that Tom is out of town
but that his kleptomaniac identical twin was at the library at the
time in question. Unlike the case of propositional defeaters, the
defeater here is information I actually possess, something I learn or
come to believe. It may not even matter that Mrs. Grabit is in fact a
liar or delusional, unless of course I have reason to believe that
this is true. Following Bergmann (2006, pp. 154-55), I'll refer to
these kinds of defeaters as mental state defeaters. (Some
philosophers, for example Alston 1986, p. 191, refer to these as
"overriders" and reserve the term "defeater" for propositional
defeaters. This terminological point is worth noting, but nothing
substantive rides on this).

Epistemologies that incorporate mental state defeaters typically take
them to defeat justification (Alston 1989, pp. 238-39; Bergmann, 2006,
pp. 155-56) or some species of rationality (Plantinga, 2000, pp.
357-66; Bergmann, 1997a, pp. 68-78). However since these positive
epistemic statuses are typically regarded as necessary for knowledge,
mental state defeaters may at least indirectly play a role in
defeating knowledge, not simply by preventing a person from coming to
know p but also by canceling a person's state of actually knowing p.
If S's knowing that p entails that S's is justified to degree N in
believing p, then if S ceases to be justified in believing p (or the
degree of justification for S's belief is significantly lowered), then
S ceases to know p. So we can think of mental state defeaters as
defeating one's actual justification and knowledge. We can refer in a
general way to a no mental state defeater condition for knowledge:

[MSD] S knows that p only if S does not have a mental state
defeater for S's belief that p.

Note that [MSD] only claims that knowledge requires the absence of a
mental state defeater, a defeater constituted by a person's
experience(s), belief(s), or other propositional attitudes. It doesn't
specify or delimit the range of what mental states will actually count
as defeaters. Would, for example, my simply taking a belief to be
defeated count as a mental state defeater? Or must I justifiably take
a belief to be defeated? Or must there be some kind of logical
relation between my beliefs and the defeatee? Similarly, must mental
state defeaters be occurrent states or can they be merely
dispositional? Advocates of [MSD] disagree about these issues, as
we'll see below. But the general idea behind mental state defeaters is
a fairly bipartisan epistemological insight, as may be shown by its
place in the broader landscape of contemporary epistemology.
a. Internalism, Externalism, and Mental State Defeaters

Epistemic internalists typically recognize that mental state defeaters
can defeat justification (Pollock, 1974, 1984, pp. 200-202, 1986, pp.
29-30, 37-58; Chisholm, 1989, pp. 55-60; Swinburne, 2001, pp. 28-31).
For the internalist, the endorsement of [MSD] is largely a consequence
of justification supervening solely on the perspective of the
cognizer. Just as the subject's beliefs and experience confer
justification on beliefs, they can also remove or downgrade
justification. If we also suppose that justification is necessary for
knowledge, the internalist will endorse a principle similar to [MSD].
Of course, for the internalist [MSD] is not an alternative to [PD].
[MSD] doesn't address the Gettier problem but only concerns
evidentialist intuitions about justification. [PD] is still needed by
internalists to handle Gettier cases. But note also that the
explication of [PD] seems to depend on certain counterfactual claims
about mental state defeaters and justification, for we must suppose
that if S were to believe d (or we were to add d to S's evidence),
then S would no longer be justified in believing p. This presupposes
that one's actual evidence can defeat one's justification. In this way
[PD] presupposes the type of conceptual framework employed by [MSD].

Many externalists have endorsed [MSD]. For example, some reliabilists
(Goldman, 1986, pp. 62-63, 111-112) include a non-undermining
provision in their accounts of justification or knowledge. In
consequence of such a provision, while reliability of belief formation
may be a necessary condition for knowledge, it's also necessary that a
person not (justifiably) believe that his belief was formed in an
unreliable manner. Alston (1988a, pp. 238-239) contends that
truth-conducive justification can be overridden by justified beliefs
that p is false or the justified belief that the belief that p is
based on inadequate grounds. According to Plantinga (1993a, pp. 40-42,
229-37; 2000, pp. 359-66), while warrant depends on the proper
functioning of our truth-aimed cognitive faculties, one aspect of this
proper functioning is a sub-system (called a defeater system) that
adjusts or revises our beliefs in the light of new experiences and
beliefs. Nozick (1981, p. 196) contends that knowledge requires that
the subject not believe that her belief doesn't track truth. In each
of these cases, the otherwise externalist theory advocates at least
one internal condition for knowledge, roughly that the subject does
not have a negative epistemic evaluation of her beliefs.
b. Coherentism, Foundationalism, and Mental State Defeaters

The idea that mental state defeaters can cause justified beliefs to
become unjustified (and the correlated [MSD] condition) is compatible
with coherentism and foundationalism, and is arguably entailed by some
versions of each.

From a coherentist viewpoint, coherence (of some form) among our
mental states confers justification on our beliefs. Very roughly
stated, I am justified in believing A if and only if A coheres with my
current experience and body of beliefs. It follows that I will become
unjustified in holding some belief A if the belief A loses its
coherence with my experience or body of beliefs. But a belief's losing
coherence with our experience and/or our beliefs is a particular way
of unpacking the idea of mental state defeaters. For example, I might
at time t recall the foyer of a certain Victorian house in
Springfield, Massachusetts having certain structural features, and
there's no incoherence at time t between my beliefs about the foyer
and the rest of my experience or beliefs. However, upon subsequently
revisiting the house at time t* I see that it's not at all as I
remember it. My present sensory experience is incompatible with my
memory beliefs about the foyer and so my former beliefs about the
foyer now become unjustified. Upon being appeared to catly, I may
believe that there is a cat in front of me. This belief may cohere
with everything else I believe and am currently experiencing at the
time, so it's a justified belief. But suppose that when I reach out
for the cat my hand goes through it, or when I move a couple of feet
to the right or left the cat disappears and then reappears when I move
back into place. My belief that there's a cat in front of me no longer
coheres with the larger network of my beliefs. In this scenario I have
lost my justification for supposing that there's a cat in front of me.

Mental state defeaters also play an important role in many versions of
foundationalism, specifically versions of so-called modest
foundationalism (Alston, 1976, 1983; Audi, 1993). Foundationalist
theories of justification, motivated largely by the justification
regress problem, terminate chains of justification in foundational
beliefs that are immediately justified. Immediately justified beliefs
are beliefs that are justified in some way other than their relation
to or dependence on other justified beliefs. Strong versions of
foundationalism restrict foundational beliefs to beliefs with various
epistemic immunities (from doubt, error, or revision) or beliefs that
are ostensibly maximally justified. These versions of foundationalism
have little or no place for the idea that subsequent mental states
might cause immediately justified beliefs to become unjustified (or
less justified). But this idea is important to modest
foundationalists, who argue that the regress problem may be avoided if
chains of justification terminate in beliefs that are prima facie
immediately justified. I can be immediately justified in believing
that there is a cat in front of me, even if I subsequently lose this
justification by realizing that I'm looking at a papier-mâché cat. My
justification is in the first instance prima facie and thus capable of
being overridden, cancelled, nullified, or downgraded by new
experiences or additions to my beliefs.

Audi (1993, pp. 105-112, 141-53) notes that one of the core intuitions
behind coherentism is really the idea of "negative epistemic
dependence," that a belief's justification is liable to being
overridden or undermined and so should not remain unaffected by
incoherence if it should arise. A belief that is justified at time t
independent of its relation to other beliefs need not be such that it
remains justified (or justified to the same degree) regardless of the
other beliefs a person forms. The idea of mental state defeaters
allows the foundationalist to incorporate a valuable insight in
coherentist theories of justification without having to subscribe to
the stronger thesis that coherence confers justification.
4. Prominent Features of Mental State Defeaters
a. Newly Acquired State Defeaters and Newly Acquired Power Defeaters

Mental state defeaters may defeat beliefs at the time the defeater is
acquired or they may do their defeating at some later time when they
acquire the power to defeat. Bergmann (2006, pp. 155-57) designates
the first a "newly acquired state defeater" and the latter a "newly
acquired power defeater."

Typically when we think of mental state defeaters we think of
situations where a person S justifiably believes p at some time t but
then at some later time t* S acquires a mental state d (some new
experience or belief) that causes S's belief that p to be unjustified
at t*. Here S's belief that p is unjustified from the time S acquires
the mental state d. In the morning I hear the weather report and
there's a prediction of showers late in the morning. Later in the
morning I hear a pitter-patter against the window facing my backyard.
Looking through my blinds, I see some dark clouds in the sky and water
drops against my window. I justifiably believe at time t that it's
raining outside. But suppose that several minutes later my wife walks
in the front door (dry as a bone) and says that my next-door neighbor
is spraying water over our fence on to the back of our house. It would
seem that I'm no longer justified in believing that it's raining
outside. At time t I was justified in this belief but at time t* I'm
no longer justified in this belief because I have acquired evidence at
time t* that defeats my prior belief. This is a newly acquired state
defeater.

In other cases, though, a mental state d may be acquired at time t but
not do its defeating work until some later time t* when it acquires
the power to defeat. Bergmann (2006. p. 156-57) designates this kind
of defeater a newly acquired power defeater. Bergmann's illustration
is helpful. My younger brother quietly tells me that when my sister
comes into the room and informs everyone that my cousin Maggie is
downstairs in the basement, this is really code for "Maggie is at her
boyfriend's house." As he explains, no one wants Maggie's father, who
is present, to know that Maggie is at her boyfriend's house. My sister
then enters the room and says she was just talking with Maggie
downstairs, which I know really means that Maggie is at her
boyfriend's house. As it happens, I already believe this because
earlier in the day Maggie's boyfriend told me that Maggie would be
visiting him at his house. So I have a justified belief that Maggie is
at her boyfriend's house, even before my sister suggests this through
code. Now suppose that shortly after the announcement, my older and
very reliable brother tells me that my younger brother was just trying
to fool me with the code story. There was no plan for my sister to
speak in code about Maggie. In this scenario, it looks like I acquire
a mental state at a particular time that only subsequently acquires
the power to defeat a belief of mine. I believe B (Maggie is at her
boyfriend's house) at time t when I acquire the belief M (my sister
has said that Maggie is in the basement), but the belief M does not
defeat the belief B at time t. My belief M only gains the power to
defeat my belief B after my older brother informed me that my younger
brother was engaged in high jinx with me. This allows me to take my
sister's comment as indicative of the actual whereabouts of Maggie,
thereby defeating my prior belief Maggie is at her boyfriend's house.

Of course, in both the case of a newly acquired state defeater and a
newly acquired power defeater the defeater may not be a complete
defeater, that is, it may not render a belief wholly unjustified.
While defeaters are normally thought of as rendering a belief
unjustified or irrational, depending on the specifics of the
evidential situation they might merely render a belief less justified
than it was before the acquisition of the defeater or before it
acquired its defeating power. For example, suppose that when my wife
tells me that our neighbor is spraying the backside of our house with
his garden hose my wife has the kind of look she gets when she's
trying to fool me about something. At the time, I can't fully accept
what she says, but it's not obvious that she's trying to pull my leg.
Perhaps her testimony in this case lowers the degree of justification
for my belief that it's raining outside, rather than renders this
belief wholly unjustified. So we should distinguish between complete
and partial defeat/defeaters.
b. Diachronic Aspects of Mental State Defeaters

The above account of mental state defeaters construes them as mental
states that defeat a belief at some particular time. This way of
thinking about defeaters is naturally suggested by the correlated
synchronic view of justification, namely of some person's being
justified in believing p at some particular time t. But we can extend
this view of defeaters by viewing their defeating power – like
justification generally – through time or diachronically. (On the
nature and significance of synchronic and diachronic justification,
see Swinburne, 2001, pp. 152-91).

First, although mental state defeaters are naturally thought of as
rendering unjustified (or less justified) a person's prior justified
belief, mental states at some time t may also prevent a person from
coming to hold a justified belief at some later time t*. We might call
this the forward-looking defeating potential of mental states. Suppose
that my wife enters the house moments before I hear the pitter-patter
and see the water drops against my window. She informs me about my
neighbor's bizarre behavior of spraying the backside of our house. The
subsequent perceptual evidence that would otherwise justify my belief
that it's raining outside will not do so in this case. The potential
justification conferring power of this evidence acquired at time t* is
antecedently neutralized by what I know or justifiably believe
beforehand at time t. We might say that my wife's testimony
constitutes a preventative justification defeater. More generally, at
any given time t our experiences and set of justified beliefs will
prevent us from being justified in holding some other belief(s) at
some subsequent time t*. Thus all mental states have some
forward-looking defeating potential. Of course, we typically don't end
up holding such beliefs (because we take them to be unjustified for us
given the rest of what we believe), but if we did they would be
unjustified by virtue of our other mental states.

Secondly, the defeating power of some mental state over an
antecedently held belief can be said to continue into the future. Call
these continuing defeaters (Bergmann, 2006, p. 158). The natural way
of thinking about this is to take the case where someone continues to
hold the defeated belief (or continues holding it with the same degree
of firmness), despite the acquisition of a mental state defeater for
the belief. Suppose that some of Kurtis' neighbors accuse Kurtis' wife
Cathleen of having an affair with a married neighborhood man. Cathleen
denies this and Kurtis justifiably believes that Cathleen is telling
the truth. Later that day Kurtis sees Cathleen in a romantic embrace
with a neighborhood man behind a tree in the local park. Kurtis has
acquired a defeater for his belief that Cathleen is an honest wife,
but through a variety of rationalizations he continues to believe that
Cathleen is an honest person. Kurtis' seeing Cathleen romantically
involved with another man causes his belief in her honesty to be
unjustified. Kurtis' memory of what he saw (or his belief that he saw
it) continues to cause his belief in Cathleen's honesty to be
unjustified, though he nonetheless persists with this belief. So here
we have a case where a memory or belief state continues to make
another belief – the subject persists in holding – unjustified. The
defeater has continuing defeater power over a persisting belief.

Of course, the idea of a preventative justification defeater allows us
to think of the defeating power of a mental state continuing into the
future, even if the person gives up the defeated belief. Perhaps I
give up my belief that it's raining outside after my wife tells me
that my neighbor is spraying my house with a garden hose. In this
case, at time t* d (the awareness of my wife's testimony) is a
defeater for a belief I had at the earlier time t but don't have any
longer. Now at time t* it makes no sense to speak of d as defeating my
actual belief that it's raining outside, because I no longer hold this
belief at t*. But we can still speak of d's continuing power to
prevent me from forming the justified belief that it's raining
outside.
c. Defeater-Defeaters

Mental state defeaters can of course be subsequently defeated by other
mental states, and we can say that all mental state defeaters are
continuing defeaters until they are defeated. That is, they continue
to render a belief unjustified or less justified until their defeating
force is neutralized. It's common to speak of mental states that
defeat mental state defeaters as defeater-defeaters (Pollock 1986,
pp.45-58; 1970; Plantinga, 1993a, pp. 231-37; 1993b, pp. 216-221;
1986). Suppose I justifiably believe T, Tom Grabit stole a library
book. Now suppose I get a defeater D for the belief that T, namely
Mrs. Grabit tells me that Tom is thousands of miles away and his
identical kleptomaniac twin was at the library at the time in
question. If I subsequently learn that Mrs. Grabit is a compulsive
liar and deranged, then I have acquired a defeater D* for the original
defeater D. I have acquired a defeater-defeater. While D rendered my
belief that T unjustified, D* restores my justification for believing
T.

Notice that in this particular example that D* doesn't render my
belief that D unjustified, even though it restores my justification
for believing T. I'm still justified to believe D, namely that Mrs.
Grabit said such and such. What is defeated here is the power of D to
defeat my prior belief that Tom Grabit stole the library book. Take
another example. Suppose I see what appear to be blue widgets coming
down an assembly line. I believe that these are blue widgets. I then
discover that the widgets are being illuminated with a blue light.
This gives me a defeater for my belief that the widgets are blue. If I
subsequently pick up a widget outside the range of the blue light,
view it under normal lighting conditions, and see that it's blue, the
defeating force of "these widgets are being illuminated with a blue
light" is neutralized, but not in such a way that I cease to be
justified or rational in believing that the widgets are being
illuminated by a blue light. So when it comes to defeater-defeaters my
justification for holding the originally defeated belief can be
restored without causing the defeater against this belief to be an
unjustified belief. Defeater-defeaters might do that of course, but
they need not. Perhaps I discover that what I thought was a blue light
shining on the widgets is not a blue light at all or perhaps I learn
that Mrs. Grabit actually did not say what I thought she said. In
these cases the defeater-defeater causes my belief in the original
defeater to be unjustified.

According to Plantinga (1986), some beliefs can, by virtue of their
own degree of warrant, defeat defeaters that come their way. When a
belief has this power, Plantinga designates it an intrinsic
defeater-defeater against some ostensible defeater. I write a letter
to the chair of my department trying to bribe him to write a highly
exaggerated letter on my behalf for an NEH fellowship. The letter
mysteriously disappears from the chairperson's office. I have a motive
to steal it, the opportunity to do so, and I have been known to do
such things in the past. Moreover, a reliable member of the department
claims to have seen me hanging around the chairperson's office about
the time the letter must have been stolen. Given the evidence, my
colleagues believe that I stole the letter. Perhaps they are justified
in believing this. However, I believe that I spent the day in the
woods and so could not have stolen the letter. My memory belief has a
great deal of nonpropositional warrant for me. So despite the
counter-evidence, I'm justified to believe that I was in the woods and
didn't steal the letter. Here it seems that the ostensible defeatee
actually operates as a defeater-defeater. Plantinga of course isn't
suggesting that an actually defeated belief restores warrant to itself
by defeating an acquired defeater. It's not as if my belief that I
didn't steal the letter was actually defeated at some point in time
and its justification subsequently restored. The idea is rather that
the original belief prevents or insulates itself from being defeated
because the defeating potential of counterevidence is antecedently
neutralized by the degree of warrant had by original belief. So I
never actually acquire a defeater for my belief that I was in the
woods or that the belief that I didn't steal the letter (Sudduth,
1999, pp. 180-82).
5. Variations on Mental State Defeaters

Advocates of mental state defeaters (and the corresponding no mental
state defeater condition) differ on some crucial points regarding
mental state defeaters.
a. The Epistemic Status of Defeating Beliefs

One of the issues of debate between adherents of [MSD] is whether
beliefs that function as mental state defeaters must have some
positive epistemic status to have defeating power, specifically if
they are to defeat beliefs that do have some positive epistemic
status. Plantinga (2000, pp. 364-65, 2002, pp. 272-75) contends that
irrational and unwarranted beliefs can defeat beliefs that are
(otherwise) rational and warranted. Suppose I believe that I'm made of
flesh, blood, and bone. I then come to believe – due to some cognitive
disorder – that my head is made of blown glass. According to
Plantinga, given that I come to hold this second belief I now have a
defeater for the prior belief, even if the defeater was formed by way
of cognitive malfunction. In other cases, my belief may be rational
but nonetheless unwarranted, and yet it might still function as a
defeater for a warranted belief. Using another example from Plantinga
(2000, pp. 363-65), suppose I believe that you were born in Yankton,
South Dakota. Your uncle, whom I believe to be a reliable person, told
me this. My belief is warranted. But then one day you inform me in all
seriousness that you were actually born in New Haven, Connecticut and
you provide a reasonable explanation for why your uncle thinks
otherwise. Absent any reason to suppose that you're trying to fool me
or are delusional, I have a defeater for my belief that you were born
in Yankton, South Dakota. However, suppose that your parents actually
lied to you about where you were born. In that case, your belief that
you were born in New Haven, Connecticut would not be warranted (given
Plantinga's understanding of warrant), and neither would my belief
that this is where you were born. So the defeater in this situation
would be an unwarranted belief of mine. (Note that it also follows
from Plantinga's account of defeaters that a belief D can defeat a
belief A with no warrant, and that D can defeat a belief A that has
more warrant than D).

Now in the above cases I acquire what Plantinga calls a "rationality
defeater." By virtue of acquiring the defeating belief D I'm no longer
rational in believing A. This is a consequence of an internal aspect
of cognitive proper functioning, what Plantinga specifically
designates internal rationality. Plantinga distinguishes between the
proper functioning of our cognitive faculties "downstream" from
experience (internal rationality) and the proper functioning of our
cognitive faculties "upstream" from experience (external rationality)
(Plantinga 2000, pp. 110-12). The former refers to the appropriate
belief response to phenomenal imagery and doxastic experience, whereas
the latter refers to proper functioning in the production of
phenomenal imagery and doxastic experience. Internal rationality will
include coherence among our beliefs and drawing the appropriate sort
of inferences from what we believe. So to say that I have acquired a
rationality defeater D for my belief A is to say that a certain
doxastic response is called for given that I have a sensuous or
doxastic experience of a certain sort. Perhaps I'm externally
irrational in forming D (e.g., because I'm suffering from paranoia,
dementia, or some kind of mental illness), but I'll still be
internally irrational to continue holding A given that I hold D.

Alston (2002) has argued that Plantinga's position is
counter-intuitive, and that only beliefs with positive epistemic
status can defeat beliefs that have positive epistemic status, and a
belief D can defeat belief A only if D has greater warrant than A. The
efficacy of a defeater depends on the relative positive epistemic
status of each of the beliefs being compared. Bergmann (2006, pp.
164-66) argues that Alston's rebuttal to Plantinga is plausible as an
account of belief revision or how we ought to change our beliefs.
Since Plantinga parses his own account of defeaters in this way,
Alston's criticism is applicable to Plantinga's position. However,
Bergmann maintains that Alston's argument doesn't undermine the notion
that irrational or unjustified beliefs can defeat justification. My
belief that I have hands is unjustified if I believe (however
irrationally) that I'm a brain in a vat, even if it's more reasonable
as a policy of belief revision to give up the belief that is less
rational or less warranted.
b. Subjective and Objective Contours

Another issue, related to the first, concerns the relationship between
having a mental state defeater and believing that one has such a
defeater.

Plantinga suggests that, ordinarily at least, having a defeater
involves one seeing or taking it that one's belief is defeated. But
would this be sufficient for having a defeater?

Alston's criticism above entails that merely taking one's belief to be
defeated isn't sufficient for defeat, because one might irrationally
or unjustifiably take one's belief to be defeated. This is presumably
the case when, due to my irrationally believing that my head is made
of blown glass, I take my belief that my head is made of flesh, bone,
and blood to be defeated. Alston and some other externalists would
argue that only truth-conducively justified or reliably produced
beliefs can be defeaters. However, since the truth-conducivity of
grounds of belief and reliability of belief formation are not
introspectively accessible facts, it is possible for an otherwise
internalist no-defeater condition to be parsed with an external or
objective component arising from the demand that defeaters be drawn
from the subject's stock of justified beliefs or knowledge.

Internalists too may impose a similar requirement, so even if it's
necessary that the subject take his belief to be defeated (in order to
have an efficacious defeater), it will also be necessary that
defeating beliefs have positive epistemic credentials of some sort. If
my belief that Jack is a lifeguard is to be defeated by my belief that
Jack can't swim, then the latter belief must be rational or justified.
And for the internalist (unlike the externalist) that a belief has
this kind of status will itself be a matter that is introspectively
accessible.

Moreover, the internalist will likely require that there be the
appropriate kind of negative evidential relationship between the
defeater and the defeatee. That is to say, if belief d actually
defeats S's belief that p, then p will at least not be likely given d
and the relevant rest of S's beliefs. D must sufficiently lower the
evidential probability of p. If we suppose that criteria of inductive
(and deductive) reasoning are introspectively accessible, then an
internalist version of the no mental state defeater condition can be
internalist in this additional respect. It can require the absence of
a negative logical relation between d and S's belief that p, where
this is introspectively accessible and so can be determined upon
reflection. (Swinburne, 2001, pp. 28-31).

Bergmann (2006, pp. 160-63), however, argues for a more subjective
account of defeat, which he believes is at least suggested by both
Plantinga and Pollock. On Bergmann's view, a person S has a defeater
for his belief that p just if he consciously takes his belief that p
to be defeated, and a person S takes his belief that p to be defeated
just if S takes the belief that p to be epistemically inappropriate.
For the latter, S must simply take himself to have good reasons for
denying p or good reasons for doubting that the grounds of his belief
that p are trustworthy, truth-indicative, or reliable. It isn't
necessary that the person have what are actually good reasons for the
negative epistemic evaluation of his beliefs. It is only necessary
(and sufficient) that the person take himself to have such reasons,
and Bergmann places no restriction on what kinds of considerations
might play this role for the subject. So on Bergmann's view the no
mental state defeater condition (as requirement for knowledge) is
really a no believed defeater condition (Bergmann, 2006, p. 163).
Bergmann's no defeater condition, then, is strongly internalist since
one has introspective access to whether or not one takes a particular
belief to be epistemically inappropriate, even if there's no
introspective access to either the justificational status of a
defeating belief or the causal origin of one's taking a belief to be
defeated.
c. Conscious and Reflective Defeaters

Since mental state defeaters include beliefs and beliefs may be
occurrent or dispositional, it will be helpful to distinguish between
conscious and reflective mental state defeaters (Bergmann, 1997a, pp.
116-121). There is a distinction between defeating experiences or
beliefs of which one is aware at time t and defeating experiences and
beliefs of which one is not aware at time t but of which one would
become aware upon reflection. Similarly, there's a distinction between
consciously taking one's belief to be defeated and this being
something that one would do upon reflection. Accordingly, someone who
advocates [MSD] may suppose that knowledge requires either the absence
of conscious defeaters or the absence of a reflective defeater.

Some externalists advocate [MSD], specifically parsed in terms of the
subject S not taking his belief that p to be defeated. Alston (1988b)
appears to argue that the absence of a mental state defeater is not a
necessary condition for knowledge. However, it's fairly clear that
Alston has in mind a reflective defeater, not a conscious defeater,
much less a person S's consciously taking his belief that p to be
defeated. Alston asks us to suppose that there is some person who has
acquired substantial evidence that his sensory experience is a
radically unreliable guide to his physical environment, that he's been
the subject of a mad scientist's neurophysiological experiments for
several years. So the subject justifiably believes that his senses are
not to be trusted. However, as this person is about to cross a street
he seems to see a truck heading towards him, and he forms the belief
that a truck is approaching. His sensory perceptual system is working
fine, and a truck is approaching. Alston says that in this scenario
the person knows that a truck is approaching, despite having
overriding reasons for supposing that his senses are not reliable. It
would seem that the person has knowledge, despite having a mental
state defeater. Crucial to Alston's account, though, is his claim that
when the subject seems to see a truck approaching, he "momentarily
forgets" his skepticism and acts accordingly. This makes it clear that
the person in question does not consciously take his belief to be
defeated when he sees the truck approaching. Rather, we have a
reflective defeater, for the subject presumably would upon reflection
take his belief to be defeated or epistemically inappropriate. So
Alston's scenario can't plausibly be taken as a counter-example to a
no conscious defeater requirement for knowledge, especially if this
kind of requirement is parsed in terms of a subject not consciously
taking her belief to be defeated.

The fact that a no conscious defeater requirement is widely subscribed
to by both externalists and internalists counts in favor of its
intuitive plausibility. But Bergmann (1997a, pp. 127-39) argues
further that we have good reasons to reject the no reflective defeater
requirement for knowledge. His argument is based on the premise that
knowledge is incompatible with veritic epistemic luck but not
evidential epistemic luck. Veritic luck refers to a person being lucky
to believe what is true, given the evidence the person has. Evidential
epistemic luck refers to a person being lucky to have the kind of
evidence she has. The Political Assassination, Unopened Letters, and
original Tom Grabit case discussed above (in 2.c) are arguably
examples of evidential epistemic luck, whereas Goldman's Fake Barn
case is an example of veritic epistemic luck. Bergmann argues that
there are cases where a person has a reflective defeater for a belief,
but the situation is analogous to cases of evidential epistemic luck.
So we have reason for resisting the idea that knowledge requires the
absence of a reflective defeater.

Here's Bergmann's example (Bergmann 1997a, p 136). Due to a strange
cognitive disorder Chuck thinks that reports he hears between 4:15pm
and 4:30pm are highly unreliable. On a particular day, Chuck's alarm
clock wakes him up from an afternoon nap at 4:20pm. Immediately upon
waking up Chuck hears noises outside his window. He looks and sees
what appear to be city workers at work near a large hole in his front
yard. One of the men tells Chuck that they are there to do work on the
main waterline to Chuck's house, and that Chuck's wife was informed of
this the day before. Chuck believes what he's told, and the man is
telling the truth. However, if Chuck reflected on the matter, he would
believe that the man's report was unreliable, for Chuck would have
realized that he's being given this report between 4:15pm and 4:30pm
and that reports he hears during this time period are unreliable. If
Chuck reflected on the matter, he would consciously take it that his
belief about what these men are doing is defeated. But Bergmann argues
that most of us would be strongly inclined to say that in this
scenario Chuck actually knows what the men in question are doing on
his property, even though Chuck has a reflective defeater for this
belief. Chuck is certainly lucky here not to have evidence against his
belief, but in much the same way in some Gettier-type cases (e.g., Tom
Grabit case above in 2.c) the subject is lucky to have the evidence he
does and not have other evidence (that is misleading) but it's not a
matter of luck that the person believes what is true given the
evidence he has.
6. Taxonomy of Defeaters and Formalities of Defeat

Having considered the distinction between propositional and mental
state defeaters, something should be said about the formalities of
such defeaters. It's fairly common for epistemologists to distinguish
between two general ways beliefs may be defeated. There are defeaters
that are reasons for supposing that p is false, and there are
defeaters that are reasons that, if added to ostensible evidence for
p, would sufficiently lower the likelihood that p is true. According
to the first kind of defeater, we get reasons to believe the negation
of p (or that p is false). According to the second, we simply lose our
reasons for supposing that p is true. But let's look at the range of
defeater-types.
a. Primary-Type Defeaters: Rebutting, Undercutting, and No Reason Defeaters

(i) A rebutting defeater for some belief that p is a reason (in the
broad sense) for holding the negation of p or for holding some
proposition, q, incompatible with p (Pollock, 1986, p. 38). Mary sees
in the distance what appears to be a sheep in the field and forms the
belief that there is a sheep in the field. The owner of the field then
comes by and tells her that there are no sheep in the field. She has
acquired what is commonly designated a rebutting defeater for her
belief that there is a sheep in the field. She has acquired a reason
for supposing that there is no sheep in the field. Alternatively, she
might have walked up to the object and discovered that it was actually
a papier-mâché facsimile. Here she acquires a reason for believing
something incompatible with her belief that there is a sheep in the
field. These are of course examples of rebutting mental state
defeaters. There can also be rebutting propositional defeaters.
Perhaps Mary doesn't hear the owner of the field tell her that there
are no sheep in the field, but he has mentioned this to several people
in the neighborhood the day she believes there is a sheep in the
field. There is a true proposition that counts against the truth of
Mary's belief, even if it isn't a proposition she believes. (Of
course, as noted above in connection with defeasibility analyses,
there will be many true propositions that misleadingly count against
the truth of beliefs).

(ii) An undercutting defeater for some belief that p is a reason (in
the broad sense) for no longer believing p, not for believing the
negation of p (Pollock, 1986, p. 39). More specifically, it is a
reason for supposing that one's ground for believing p is not
sufficiently indicative of the truth of the belief. A person enters a
factory and sees an assembly line on which there are a number of
widgets that appear red. Being appeared to red-widgetly, the person
believes that there are red widgets on the assembly line. The shop
superintendent then informs the person that the widgets are being
irradiated by an intricate set of red lights, which allow the
detection of hairline cracks otherwise invisible to the naked eye.
Here the person loses his reason for supposing that the widgets are
red, rather than acquires a reason for supposing that they are not
red. Again, these are illustrations of undercutting mental state
defeaters. There can also be propositional defeaters of the
undercutting variety. The mere fact that the widgets are being
irradiated with a red light would be one such example. Or suppose that
Jason believes his tie is red. The fact that he is red-green
colorblind might be a propositional defeater for this belief. The fact
that someone is prone to perceptual hallucinations might be a
propositional defeater for some range of sensory perceptual beliefs,
and so forth.

(iii) A no-reason defeater is a reason for supposing that it's no
longer reasonable to believe p given that (a) one has no reason for
believing p and (b) the belief that p is the sort of belief that it's
reasonable to hold only if one has evidence for p (Bergmann, 1997a,
pp. 102-103). For example, Johnny believes that if he dies he will
immediately thereafter be turned into a zombie. Upon reflection he
can't locate any reasons why he believes this, but he realizes that
it's the sort of belief for which he ought to have some reason if he
is rationally to believe it.

Now in each of these three cases (parsed in terms of mental state
defeaters), the acquisition of a defeater makes it epistemically
inappropriate to continue holding a particular belief B given that (i)
there is evidence against B, (ii) reasons for B have become
neutralized, or (iii) there is a recognition that one has no reasons
at all for holding B though one ought to have such reasons.
Consequently, a person's belief is no longer justified (or – in the
case of partial defeaters – not as justified as it would be absent the
defeater). If knowledge entails justification, each of these kinds of
defeaters has the potential to defeat knowledge. If parsed in terms of
propositional defeaters, then the corresponding true propositions are
such that they prevent an overall justified true belief from counting
as knowledge.
b. Secondary-Type Defeaters: Defeaters for Grounds of Inferential Beliefs

There are also defeater-types that appear to be derived from (i),
(ii), and (iii), and which apply specifically to cases where beliefs
are based on other beliefs, that is, inferential or mediate beliefs.

(iv) A rebutting reason-defeating defeater is a rebutting defeater
against a belief, c, where c is a ground or reason for the belief that
p. Mark believes that his computer has a hardware problem that is
causing several operation errors. He believes this because his wife
tells him that Peter told her this and Mark knows that Peter is an
expert on computers. Later, though, Mark discovers that it was not
Peter but John who told his wife this, but Mark believes that John has
little knowledge about computers.

Thinking of defeaters in terms of argument forms, Pollock (1986, pp.
38-39) distinguished between reasons that attack a conclusion
(rebutters) and reasons that attack the connection between the
premises and the conclusion (undercutters). Rebutting reason-defeating
defeaters are distinct from both rebutting and undercutting defeaters
in Pollock's sense. In the language of argumentation, they attack
neither the conclusion nor the connection between the premises and the
conclusion. A rebutting reason-defeating defeater is a species of
rebutting defeater (as I defined it above), but it's a reason to
believe the negation of a belief, c, that functions as the ground or
reason of another belief p. In terms of argument forms, we can say
that a reason-defeating defeater is a rebutting defeater against a
premise in some argument. This kind of defeater is also distinct from
Pollock's undercutting defeater. In the case of rebutting
reason-defeating defeaters, it's not that the grounds fail to be
indicative of the truth of Mark's belief that his computer has a
hardware problem, but Mark comes to believe that one of his original
grounds for holding this belief is false. Like undercutting defeaters,
in acquiring a rebutting reason-defeating defeater we lose our reasons
for supposing that the target belief that p is true. As a result, the
grounds lose their power to confer justification on the target belief.
However, this comes about by way of acquiring reasons for supposing
that a ground of the target belief is false. (See Bergmann 1997a, pp.
99-103, for further discussion on the distinction between undercutters
and reason-defeating defeaters).

(v) If we continue to think of defeaters and defeat in terms of
argument structures then we can apply undercutting defeaters to more
complex grounds for belief, where a belief that p is based on some
further belief, q, that is in turn based on some other belief, r. An
undercutting reason-defeating defeater for some belief that p is a
reason for supposing that the grounds, r, for some belief that q fail
to be sufficiently indicative of the truth of q, but where q is itself
a ground for believing p. In terms of general logic, the premises of
arguments are often themselves supported by reasons, thereby creating
sub-arguments. Just as we can acquire reasons for the negation of a
premise in an argument, we can acquire reasons for supposing that the
premises of a sub-argument fail to be indicative of the truth of a
premise in some main argument. As with rebutting reason-defeating
defeaters, we lose our reasons for believing the main conclusion, p,
but here we do so by virtue of losing our reasons for believing a
premise, q, rather than by acquiring a reason for denying the premise
q.

(vi) A no-reason reason-defeating defeater is simply the application
of the no-reason defeater to the grounds of an inferentially held
belief. In (iii) a belief is defeated because it's not based on any
reason but is the kind of belief that is reasonable only if there are
reasons for it (or the person believes this is the case). However,
even where some belief that p is based on the belief that q, the
belief q may be such that it isn't based on any reasons but it would
be unreasonable to hold the belief that q unless it's based on
reasons.
7. Conclusion

This article outlined two general types of defeaters: propositional
defeaters and mental state defeaters. The former are conditions
external to the perspective of the cognizer that prevent an overall
justified true belief from counting as knowledge. The latter are
conditions internal to the perspective of the cognizer (such as
experiences, beliefs, withholdings) that cancel, reduce, or even
prevent justification. Propositional defeaters are designed to address
the problem of accidentally true belief, whereas mental state
defeaters arise from the variable nature of justification. Inasmuch as
justification is necessary for knowledge, mental state defeaters are
capable of defeating knowledge. This leads to the viewpoint that
knowledge requires the absence of any mental state defeater. So both
kinds of defeaters ultimately relate to conditions of knowledge, and
the article developed each in connection with their larger
epistemological territory.

This was followed by an examination of the complexities that arise in
developing no propositional defeater and no mental state defeater
conditions for knowledge. The defeasibility theorist must select from
among different criteria to locate the relevant range of true
propositions that are genuinely indicative of a defect in
justification that prevents knowledge. Advocates of mental state
defeaters face a range of other issues, from choosing more or less
subjective accounts of mental state defeaters, to choosing between
conscious and reflective types of mental state defeaters for the no
defeater condition for knowledge. Synchronic and diachronic aspects of
mental state defeat were also considered.

The latter part of the article outlined a taxonomy of defeaters that
highlights the difference between getting defeaters for beliefs and
getting defeaters specifically for beliefs based on reasons of varying
degrees of complexity. Here several of the dynamics that emerge within
the taxonomy of defeaters were pointed out. One of the more important
distinctions is between losing one's grounds for believing p and
acquiring reasons for believing the denial of p (or for believing
something incompatible with p). The article also considered several
ways in which a subject might lose his grounds for believing p. While
some of these involve a subject becoming unjustified in holding to
some reason, r, for his believing p, others amount simply to the
subject's reasons, r, losing their power to confer justification on
the target belief that p while the subject remains justified in
believing r.
8. References and Further Readings

* Alston, William. 2005. Beyond Justification: Dimensions of
Epistemic Evaluation. Ithaca: Cornell University Press.
o Alston provides a systematic analysis of various epistemic
desiderata and their implications for revising our approach to the
concept of epistemic justification.
* Alston, William. 2002. "Plantinga, Naturalism, and Defeat." In
James Beilby (ed), Naturalism Defeated? Essays on Plantinga's
Evolutionary Argument against Naturalism. Ithaca: Cornell University
Press, pp. 176-203.
o Alston examines Plantinga's evolutionary argument against
naturalism and offers criticisms of Plantinga's suggestion that an
irrational belief can function as a defeater.
* Alston, William. 1989. Epistemic Justification. Ithaca: Cornell
University Press.
o This is Alston's collection of previously published essays
on substantive and meta-questions in epistemology, including essays on
foundationalism and the concept of epistemic justification.
* Alston, William. 1988a. "An Internalist Externalism." Synthese
74: 265-83. Reprinted in Alston, 1989, pp. 227-45. Page references are
from reprint.
o Alston develops a theory of epistemic justification that
combines elements of externalism and internalism.
* Alston, William. 1988b. "Justification and Knowledge."
Proceedings of the World Congress of Philosophy, 5. Reprinted in
Alston, 1989, pp. 172-82. Page references are from reprint.
o Alston argues that justification (construed in both
internalist and externalist ways) is not necessary for knowledge. The
essay includes an argument for supposing that a person can know p even
though she has a certain kind of mental state defeater for her belief.
* Alston, William. 1986. "Internalism and Externalism in
Epistemology." Philosophical Topics, 14: 179-221. Reprinted in Alston,
1989, pp. 185-226. Page references are from reprint.
o Alston's examination of internalist and externalist
approaches to justification.
* Alston, William. 1983. "What's Wrong with Immediate Knowledge?"
Synthese, 55:73-95. Reprinted in Alston, 1989, pp. 57-78. Page
references are from reprint.
o Alston critically examines various objections to
"immediate knowledge" and argues that these objections rest on various
implausible assumptions about the character of immediate knowledge.
* Alston, William. 1976. "Has Foundationalism Been Refuted?"
Philosophical Studies, 29: 287-305. Reprinted in Alston, 1989, pp.
39-56. Page references are from reprint.
o Alston defends "minimal foundationalism" against the
criticisms of foundationalism raised by Frederick L. Will and Keith
Lehrer.
* Annis, David. 1973. "Knowledge and Defeasibility." Philosophical
Studies, 24: 199-203.
o Critical response to the defeasibility analysis provided
by Lehrer and Paxson in Lehrer and Paxson, 1969, and which examines
the nature of misleading or defective defeaters.
* Audi, Robert. 1993. The Structure of Justification. Cambridge:
Cambridge University Press.
o Audi's previously published essays on various topics in
epistemology, including his development and defense of moderate
foundationalism and the idea of "negative evidential dependence."
* Barker, John. 1976. "What You Don't Know Won't Hurt You."
American Philosophical Quarterly, 13: 303-308.
o Barker attempts to tackle the Gettier problem in terms of
a defeasibility analysis that distinguishes between genuine and
misleading defeaters.
* Beilby, James (ed). 2002. Naturalism Defeated? Essays on
Plantinga's Evolutionary Argument against Naturalism. Ithaca: Cornell
University Press.
o Essays discussing Alvin Plantinga's evolutionary argument
against naturalism, some of which discuss Plantinga's notion of
rationality defeaters.
* Bergmann, Michael. 2006. Justification without Awareness. New
York: Oxford University Press.
o Bergmann defends an externalist theory of justification,
which includes both a proper function and no mental state defeater
requirement.
* Bergmann, Michael. 2005. "Defeaters and Higher-Level
Requirements." Philosophical Quarterly, 55: 419-36.
* Bergmann, Michael. 2000. "Deontology and Defeat." Philosophy and
Phenomenological Research 60: 87-102.
o Bergmann argues that deontologism does not lend support to
internalism. Essay provides several helpful observations on defeaters.
* Bergmann, Michael. 1997a. "Internalism, Externalism, and
Epistemic Defeat." (PhD Dissertation: University of Notre Dame).
o Bergmann provides a detailed examination of the nature of
defeaters and their relation to internalist and externalist theories
of knowledge.
* Bergmann, Michael. 1997b. "Internalism, Externalism, and the
No-Defeater Condition." Synthese, 110: 399-417.
o Bergmann argues that the no mental state defeater
condition being necessary for warrant is compatible with externalist
theories of warrant. Section 4 contains an analysis of externalists
who endorse some version of the no mental state defeater condition.
* Boonin, Leonard G. 1966. "Concerning the Defeasibility of Legal
Rules." Philosophy and Phenomenological Research, 26: 371-78.
o Boonin examines the meaning of defeasibility in law and
its implications for legal analysis.
* Chisholm, Roderick. 1989. Theory of Knowledge. 3rd edition. New
Jersey: Prentice-Hall.
o Chisholm provides an internalist response to the Gettier
problem, as well as an account of defeasible justification influenced
by defeasibility in moral philosophy. First edition: 1966.
* Gettier, Edmund. 1963. "Is True Belief Knowledge?" Analysis, 23: 121-23.
o Gettier's famous paper in which he argues that beliefs can
be both true and justified and yet fail to constitute knowledge.
* Goldman, Alvin. 1986. Epistemology and Cognition. Cambridge, MA:
Harvard University Press.
o Goldman endorses a version of reliabilism with a no mental
state defeater requirement for justification.
* Goldman, Alvin. 1976. "Discrimination and Perceptual Knowledge."
Journal of Philosophy, 73: 771-91.
o Goldman discusses a causal theory of perceptual knowledge
and defeasibility analyses of knowledge. The essay includes the famous
"Fake Barn" scenario, a Gettier-type case initially suggested to
Goldman by Carl Ginet.
* Harman, Gilbert. 1973. Thought. Princeton, NJ: Princeton University Press.
o Text contains Harman's "Political Assassination" and
"Unopened Letters" Gettier cases.
* Hart, H.L.A. 1961. "The Ascription of Responsibility and
Rights." In Herbert Morris (ed), Freedom and Responsibility: Readings
in Philosophy and Law. Stanford, CA: Stanford University Press, pp.
143-48. Originally published in Proceedings of the Aristotelian
Society, 1948-49, 49: 171-94. Page references are from the reprint.
o This is Hart's classic discussion of the defeasibility of
legal rules.
* Klein, Peter. 1981. Certainty: A Refutation of Skepticism.
Minnesota: University of Minnesota Press.
o Klein presents his revised and detailed development of a
defeasibility analysis of knowledge.
* Klein, Peter. 1976. "Knowledge, Causality, and Defeasibility."
The Journal of Philosophy, 73: 792-812.
* Klein, Peter. 1971. "A Proposed Definition of Propositional
Knowledge." The Journal of Philosophy, 68: 471-82.
o Klein presents a defeasibility analysis of propositional
knowledge to handle the intuition that knowledge cannot be
accidentally true belief.
* Kvanvig, Jonathan L. (ed). 1996. Warrant in Contemporary
Epistemology: Essays in Honor of Alvin Plantinga's Theory of
Knowledge. Lanham: Rowman and Littlefield.
o A collection of essays on Alvin Plantinga's theory of
warrant by prominent contemporary epistemologists. See especially
articles by Peter Klein (pp. 97-130) and Marshall Swain (pp.131-146),
both of whom address defeasibility analyses of knowledge in relation
to Plantinga's theory of warrant.
* Lehrer, Keith and Paxson, Thomas. 1969. "Knowledge: Undefeated
Justified True Belief." Journal of Philosophy, 66: 225-37.
o Influential early defeasibility analysis of knowledge in
response to the Gettier problem, focusing on the problem of specifying
the relevant sub-set of true propositions that are indicative of a
defect in justification. The essay includes the widely discussed "Tom
Grabit" illustrations.
* Nozick, Robert. 1981. Philosophical Explanations. Cambridge, MA:
the Belknap Press.
o An externalist account of knowledge that requires that the
absence of a certain kind of mental state defeater, specifically that
a person not believe that his belief does not track truth.
* Plantinga, Alvin. 2002. "Reply to Beilby's Cohorts." In James
Beilby (ed), 2002, pp. 204-75.
o Plantinga responds to criticisms of his evolutionary
argument against naturalism. His detailed comments on rationality
defeaters are particularly relevant.
* Plantinga, Alvin. 2000. Warranted Christian Belief. New York:
Oxford University Press.
o Plantinga applies his externalist theory of warrant and
proper function to questions regarding the positive epistemic status
of Christian belief. In chapter 11 Plantinga provides a more developed
account of his view of rationality defeaters earlier introduced in
Plantinga 1993a.
* Plantinga, Alvin. 1996. "Respondeo" in Jonathan Kvanvig (ed),
Warrant in Contemporary Epistemology. Lanham, MD: Rowman and
Littlefield, pp. 307-78.
o Plantinga responds to various criticisms of his
externalist theory of warrant and proper function. Particularly
relevant here is Plantinga's discussion of defeasibility analyses of
knowledge in response to Klein and Swain, pp. 317-26.
* Plantinga, Alvin. 1995. "Reliabilism, Analyses, and Defeaters."
Philosophy and Phenomenological Research, 55: 427-64.
o An early version of Plantinga's evolutionary argument
against naturalism in which he provides some detailed reflections on
rationality defeaters, subsequently developed by Plantinga in
Plantinga 2000.
* Plantinga, Alvin. 1993a. Warrant and Proper Function. New York:
Oxford University Press.
o Plantinga's earlier discussion of rationality defeaters
and the defeater system (pp. 40-42, 216-37) in the larger context of
his theory of warrant as requiring the proper functioning of our
cognitive faculties.
* Plantinga, Alvin. 1993b. Warrant: The Current Debate. New York:
Oxford University Press.
o Plantinga articulates various inadequacies in contemporary
internalist and externalist theories of warrant. The appendix examines
Pollock's conception of defeaters.
* Plantinga, Alvin. 1986. "The Foundations of Theism: A Reply."
Faith and Philosophy 3, 3: 310-312.
o Plantinga responds to Philip Quinn's criticisms of
Plantinga's proper basicality thesis regarding theistic belief.
Plantinga presents the idea of an intrinsic defeater-defeater.
* Pollock, John. 1986. Contemporary Theories of Knowledge. Savage,
MD: Rowman and Littlefield.
o Pollock's account of justification utilizes a detailed
account of mental state defeaters.
* Pollock, John. 1984. "Reliability and Justified Belief."
Canadian Journal of Philosophy 14, 103:114. Reprinted in Moser, Paul
K. (ed). 1986. Empirical Knowledge: Readings in Contemporary
Epistemology. Savage, MD: Rowman and Littlefield Publishers,
pp.193-202.
o Pollock discusses how the acquisition of reasons for
supposing that a belief was unreliably produced defeat justification,
but that this does not commit the epistemologist to a reliabilist
theory of justification.
* Pollock, John. 1974. Knowledge and Justification. Princeton, NJ:
Princeton University Press.
* Pollock, John. 1970. "The Structure of Epistemic Justification."
American Philosophical Quarterly, monograph series 4: 62-78.
o Article contains Pollock's early reference to two kinds of
defeaters, Type I and Type II excluders, which later become rebutting
and undercutting defeaters.
* Shope, Robert. 1983. The Analysis of Knowing: A Decade of
Research. Princeton, NJ: Princeton University Press.
o Shope provides an overview of a dozen or so early attempts
to resolve the Gettier problem. Chapter two examines defeasibility
analyses.
* Steup, Matthias. 1996. An Introduction to Contemporary
Epistemology. Upper Saddle River, NJ: Prentice Hall.
o In chapter 1 Steup distinguishes between propositional
defeaters (what he calls factual defeaters) and mental state defeaters
(what he calls justificational defeaters) and considers their
implications for various issues in epistemology.
* Sudduth, Michael. 1999. "The Internalist Character and
Evidentialist Implications of Plantingian Defeaters." The
International Journal for the Philosophy of Religion, 45: 167-187.
o Sudduth argues that Plantinga's notion of a "defeater
system" (as a part of cognitive proper functioning) entails two
significant evidentialist conditions for warranted belief in God.
* Swain, Marshall. 1981. Reasons and Knowledge. Ithaca: Cornell
University Press.
o Swain attempts to address inadequacies in defeasibility
analyses by combining a reliabilist indicator view of justification
and a causal account of the basing relation.
* Swain, Marshall. 1974. "Epistemic Defeasibility." The American
Philosophical Quarterly, 11,1: 15-25.
o Swain examines defeasible vs. indefeasible justification
in relation to the Gettier problem and the analysis of knowledge.
* Swinburne, Richard. 2001. Epistemic Justification. Oxford:
Clarendon Press.
o A development and defense of epistemic internalism, with
chapters on Bayesian probability. Swinburne adopts a defeasibility
analysis to handle the Gettier problem (pp. 192-200), but also
incorporates mental state defeaters in his account of justification
(pp. 28-31).