Hi all.
The subject of "knowledge" comes up here a lot, off and on, and elsewhere.
I shall add my naive tuppence, a simple approach to the matter, which
satisifies me, at least till somebody shoots it down. As I have many
anti-fans in these newsgroups, that should not take too long.
No doubt these naive ideas have already been expressed before, at
fairly intolerable length, in various philosophical tracts, and have
been much debated in all directions. Alas, a philosophobe like myself
is hardly likely to know any detail thereof.
As was mentioned recently, the "standard" attitude toward knowledge
is that of
"justified true belief".
~~~~~~~~~~~~~~~~~~~~~
Various odd examples are given to show that none of this can be dropped,
and there have even been attempts to add various fourth words into it.
IMHO, this approach is all wrong. Trying to crystallize a concept
into such a jewell-phrase is bound to be trying, and even if it becomes
useful as having proverb status, this one is still badly off.
For a start, knowledge does neither have to be true, nor a belief, nor
any combination of these. (So we can just ignore the "justified"!)
For instance, I know that Zeus married his sister. It is not part of
my belief, but it is part of my knowledge. For a Greek peasant of the
hellenic dark age it was both knowledge and belief. For most moderns
it is neither. For the Greek peasant child, it might be part of
his beliefs but not knowledge, (maybe at a bit of a stretch!)
So, as we use "knowledge" in all cases, what are we to do?
It seems to me, the obvious and simple answer is always to define
knowledge in a CONTEXT. Knowledge, is facts from some (possibly
implicit) background canon of agreed facts; only some of which may
be known by any particular person, (and indeed some of which may be
unknown by ANY of the people involved, e.g. if they are only on record.)
So, in one context I can say: "I know that Thor makes thunder", and be
correct. But I can say the same in another context, and be lying.
Seems simple enough so far. And, I dare say, would be a very simple and
effective approach to the concept of knowledge in an AI system, should
such folk possibly be concerned about such a matter.
Everyone will clamour – "What about *absolute* knowledge, in general,
based on absolute truth". This gets close to "unassailable beliefs"
again, which we know is a dodgy issue. So don’t necessarily expect
that "absolute knowledge" is going to be very meaningful.
But an obvious approach might be, "the context in which the agreed
canon of facts are all those TRUE statements about the universe, as
determined by what people ultimately find out about it".
I doubt though, that this could stand up to much scrutiny. Simple
middle-sized things that no longer change might qualify – e.g. it is
probably absolute knowledge that Paris was the capital of France between
1600 and 1990. But we daren’t go much further than these brute facts.
It was recently "(absolute?) knowledge" that gravity was an attractive
force with an inverse square law; but not altogether so any more. It
is doubtful if "electrons are charged particles" will be knowledge in
200 years time, indeed in some contexts it is already quite doubtful.
It is doubtful if "there was a big bang" or "evolution is true" and many
other such can be knowledge at all, merely on the grounds of ambiguity,
an ambiguity that can hardly be resolved except by changing the statements
so much that they become something else altogether. At the moment, in
the current academic context, they are agreed facts; but are hardly
likely to qualify as "absolute".
So summarising then:-
1) To have knowledge, one must have a bunch of people in some context;
2) and a (possibly extendable) canon of facts agreed between them;
3) then you KNOW something if you are in the context, and it is one of
those facts, and you have immediate(ish) mental access to the fact.
Vital note – there are no remarks here about TRUTH or BELIEF, either way.
Too simple really.
Bound to be major problems there…
——————————————————————————-
Bill Taylor W.Tay…@math.canterbury.ac.nz
——————————————————————————-
You are what you remember.
——————————————————————————-
math…@math.canterbury.ac.nz (Bill Taylor) writes:
>As was mentioned recently, the "standard" attitude toward knowledge
>is that of
> "justified true belief".
> ~~~~~~~~~~~~~~~~~~~~~
>Various odd examples are given to show that none of this can be dropped,
>and there have even been attempts to add various fourth words into it.
>IMHO, this approach is all wrong. Trying to crystallize a concept
>into such a jewell-phrase is bound to be trying, and even if it becomes
>useful as having proverb status, this one is still badly off.
Welcome to the club. That makes two of us.
>Too simple really.
>Bound to be major problems there…
Well you have just announced that most of what counts as philosophy
is based on a falacious notion of ‘knowledge’. The philosophers
won’t greet this news with any great enthusiasm.
In article <6hh0qv$2…@cantuc.canterbury.ac.nz>,
math…@math.canterbury.ac.nz says…
> …
> As was mentioned recently, the "standard" attitude toward knowledge
> is that of
> "justified true belief".
> ~~~~~~~~~~~~~~~~~~~~~
I will jump the gun here and recommend that we drop first two terms
leaving only belief. I will argue that knowledge is belief in both pure
and mathematical sense (i.e. Dempster-Shafer’s Theory). The two are
equivalent when both deal with empirical (observed) data, but quickly
part company when the data is derived (extrapolated). One can also add
that belief based on derived data can become knowledge upon verification
through observation. In other words all knowledge is belief, but not all
belief is knowledge. Ok, enough with my little commentary.
> …
> So summarising then:-
> 1) To have knowledge, one must have a bunch of people in some context;
You are referring to accepted (shared) knowledge. I do not think this is
a requirement for I can know things (have knowledge of) which may not be
in context shared by others.
> 2) and a (possibly extendable) canon of facts agreed between them;
Ok, we can call these facts evidence data.
> 3) then you KNOW something if you are in the context, and it is one of
> those facts, and you have immediate(ish) mental access to the fact.
Well, knowledge does have to do with context, empirical evidence, and
mental access to both. But, I don’t think that this is enough. Knowledge
seems to be more than just knowing something of the empirical evidence in
a context. One also needs the mental comprehension of the relationships
of empirical evidence and context and the ability to extrapolate based on
those relationships (to get new beliefs).
To summarize, given a context and a set of beliefs, knowledge are the
empirically verifiable beliefs which are mentally comprehended with in
that context to give rise to new beliefs (and possibly new context).
Alex Solntsev
————-
All disclaimers apply
Did you talk to your computer today?
In article <6hipth$…@ux.cs.niu.edu>, rick…@cs.niu.edu (Neil Rickert) writes:
> math…@math.canterbury.ac.nz (Bill Taylor) writes:
>>Too simple really.
>>Bound to be major problems there…
> Well you have just announced that most of what counts as philosophy
> is based on a falacious notion of ‘knowledge’. The philosophers
> won’t greet this news with any great enthusiasm.
I’ll match your ad hominem with a reductio:
If philosophers are a community in uniform agreement that he’s
misguided, then by his criteria don’t they _know_ he’s misguided?
Regards,
Larry Mayhew may…@wku.edu
math…@math.canterbury.ac.nz (Bill Taylor) says…
>As was mentioned recently, the "standard" attitude toward knowledge
>is that of
> "justified true belief".
> ~~~~~~~~~~~~~~~~~~~~~
>Various odd examples are given to show that none of this can be dropped,
>and there have even been attempts to add various fourth words into it.
>IMHO, this approach is all wrong…
[stuff deleted]
>So summarising then:-
>1) To have knowledge, one must have a bunch of people in some context;
>2) and a (possibly extendable) canon of facts agreed between them;
>3) then you KNOW something if you are in the context, and it is one of
> those facts, and you have immediate(ish) mental access to the fact.
>Vital note – there are no remarks here about TRUTH or BELIEF, either way.
It seems to me that your approach to formalizing knowledge replaces
the "true" in "justified true belief" by "belonging to the canon of
agreed-upon facts". The other parts of the phrase "justified true
belief" are still relevant, though.
To take your example, you consider the claim "Hera, queen of the
gods, is both wife and sister of Zeus" be knowledge (of the canon
of Greek mythology). However, if I don’t remember whether Hera is
Zeus’ sister, then the claim is not knowledge. Also, if the only
reason I think Hera was Zeus’ sister is because they look like
brother and sister in the Walt Disney movie of Hercules, then that
really is not a good reason for me believing it. I wouldn’t say
that I *know* that Hera and Zeus were siblings. So, even if you
replace "true" by "belonging to some canon", the issues of what
you believe (to belong to the canon) and what justification you
have for believing it still come up.
Daryl McCullough
CoGenTex, Inc.
Ithaca, NY
math…@math.canterbury.ac.nz (Bill Taylor) writes:
>As was mentioned recently, the "standard" attitude toward knowledge
>is that of
> "justified true belief".
> ~~~~~~~~~~~~~~~~~~~~~
>>IMHO, this approach is all wrong. Trying to crystallize a concept
>>into such a jewell-phrase is bound to be trying, and even if it becomes
>>useful as having proverb status, this one is still badly off.
It might be clearer to say: the standard idea is that the locution
S knows that p
can be analyzed as the conjunction of three conditions:
(B) S believes that p
(T) p is true
(J) S is justified in believing that p
Note this is not really an analysis of the term "knowledge".
Quite apart from such things as Gettier cases, which show these are not
jointly sufficient, I would say there are still issues akin to ones you
raise about the context-dependence of this idiom. These figure
prominently in late Wittgenstein, who makes some quite radical claims
about context-dependence that are now widely rejected.
To understand this, note that it is quite generally true that
repeatable word-forms only take on determinate significance as used in
particular contexts. This is especially obvious for indexical terms
like "here", "now", "I", and "yesterday". Taking this into account, we
might represent the linguistic meaning of a sentence form as a
*function* from contexts to the statement that would be made (or the
propositions that would be expressed) by using it in the context. (And
the meaning of a *word* or construction would be explained derivatively
as the contribution it makes to the meaning of the containing
sentence-form). But then it may be that this function is only partial;
i.e. there may be contexts in which word-forms that can take on
significance is other combinations or in other contexts cannot be
combined so as to make any statement at all. Such was the case,
suggested Wittgenstein, about many hypothetical uses of the word "know"
considered by philosophers.
As one example, Wittgenstein thought that although "He knows that __"
can generally be used combined with "I am in pain" to make a statement,
and "I know that ___" + "he is in pain" as well, still it made no sense
at all (expressed no proposition in any context) to combine "I know
that ___" with "I am in pain", except to make a joke. Presumably this
was because there was no significant possibility of *not* knowing;
also, no answer to the question "how do you know", and no such thing as
giving a procedure by which one arrived at the knowledge. Of course,
this means it would be equally wrong to say one *didn’t* know; rather,
neither the assertion nor the denial were meaningful, ones own pains are
simply outside the range of first-person knowledge claims.
In _On Certainty_ he suggests that the sentence by which Moore
attempted to claim to know he had a hand did not in fact express any
proposition. Witt argues that only in very special circumstances could
a human being make a significant statement using such words as "I know
I have a hand". To give an example: Suppose I have been in an
explosion, come to in a hospital bed with no feeling in my arms, and
with my head immobilized so I can’t get a look at my extremities. In
these conditions I might undertake an *investigation* of some sort into
whether my right hand was still there. If I assert, "I still have my
right hand", the patient in the next bed can say "Are you sure?" or
"how do you know", and I say such things as "I *know* I have my right
hand, because by looking to the right I can get a glimpse of
its reflection in the window".
But, the idea goes, in most ordinary circumstances, we don’t understand
what it even means to issue a putative claim to know one has a hand.
According to Wittgenstein such a sentence really expresses part of the
*background* within which particular claims can show up as significant
candidates for knowing or not knowing, but is not normally itself one
of those candidates. He also suggested that what was part of the
background can shift over time.
So the idea is that certain putative propositions about what we know
actually misfire, the sentences fail to make a claim by attempting to
express explicitly what is really part of the implicit practical
background of our conduct that forms the framework for other
significant propositions. In another context, Wittgenstein suggested
that in unreflectively pitying someone who is hurt and reflexively
moving to comfort him, I am taking up a practical orientation towards
the man, but am not "of the opinion" that, say, he has a mind. In this
sense Witt thought it nonsense even to say I *believe* he is in pain;
rather, I *act* towards him as if he were in pain, without intellectual
acts believing propositions, much less attempting to justify them.
Bert Dreyfus has made the idea that most of human understanding exists
in the unformalizable background (not the explicitly thematized
propositions in the foreground) the cornerstone of his critique of AI.
John Searle also essays some ideas about "the Background" in his work
on Intentionality. I gather, however, he regards views of the sort
Wittgenstein advanced as fallacious for e.g. confusing the question
of whether a statement would have a point in a context with whether it
would be true (now something like the consensus view, actually).
ander…@pitt.edu (Anders N Weinstein) writes:
>It might be clearer to say: the standard idea is that the locution
> S knows that p
>can be analyzed as the conjunction of three conditions:
> (B) S believes that p
> (T) p is true
> (J) S is justified in believing that p
>Note this is not really an analysis of the term "knowledge".
>Quite apart from such things as Gettier cases, which show these are not
>jointly sufficient, I would say there are still issues akin to ones you
>raise about the context-dependence of this idiom. These figure
>prominently in late Wittgenstein, who makes some quite radical claims
>about context-dependence that are now widely rejected.
When I say "I know that X", I mean something very different from when
I say "I believe that X." However, it cannot be that
I know that X
is equivalent to
I believe that X, I believe that X is true, and I believe that I am
justified in believing that X,
for typically "I believe that X" already carries those other
implications.
Yet, when I say "I know that X", I am saying something about my
internal state, and I am not saying anything about whether others
would judge X to be true, or would judge me to be justified in making
my assertion. It is possible that I might claim "I know that X",
even when I am aware that the rest of the world judges X to be
false. Likewise there are times that I might assert "I know that X",
yet not be surprised to discover I was mistaken.
And what do you make of assertions such as:
John knows he has AIDS, but he does not believe it.
I can’t believe I ate the whole thing.
I don’t know what kind of weather we will have today, but I believe
it is going to rain.
In article <6hh0qv$2…@cantuc.canterbury.ac.nz>, math…@math.canterbury.ac.nz (Bill Taylor) writes:
> Hi all.
I was sorry to see your message appear, because it confuted one of my
favorite criteria for cranks and newbies: Someone who originates a
thread with a subject in all caps!
> For a start, knowledge does neither have to be true, nor a belief, nor
> any combination of these. (So we can just ignore the "justified"!)
> For instance, I know that Zeus married his sister. It is not part of
The question here is whether such an example provides much insight
into knowledge. It’s usually treated as a contrivance of connivance:
We connive to speak as if the myths are true because speaking thus
is simpler. We’re prepared to elaborate, of course, if some
innocent ignoramus wanders into the conversation and takes us at our
word.
I think you need more than such an example to support the claim that
"I know X even though X is false" is unproblematical.
More important, I’d understand your account better if I understood how
it deals with such issues as these:
(a) My forever undisclosed knowledge of my private thoughts.
(b) My participation in multiple contexts. I expect this is
one root of the skeptical impulse (see (d) below).
(c) How "I know X" performs its function of "I hereby warrant for
you that X is true.
(d) Where the power of ordinary (and philosophical) skepticism comes
from–given your account of knowledge. If one supposes that
knowledge is somehow concerned with truth, skepticism has a
role to play in the epistemic picture. On your account, that’s
no so clear. "Dad, I’m out of sync with my peer group" somehow
doesn’t have the flavor of "Dad, I think you’ve been lying to
me about Santa Claus."
(e) Our unwillingness to agree, simply because there was a community
of agreement, that the Nazis knew Jews to be the scum of the earth.
In other words, what’s the distinction between knowledge and a
community of prejudice?
(f) Our commonsense distinction between what we think we know and
what we actually know. Don’t we ordinarily apply this distinction
to communities, too?
> 2) and a (possibly extendable) canon of facts agreed between them;
> 3) then you KNOW something if you are in the context, and it is one of
> those facts, and you have immediate(ish) mental access to the fact.
> Vital note – there are no remarks here about TRUTH or BELIEF, either way.
Right, since your project is to severe any necessary relationship between
truth and knowledge.
You do, however, speak about facts which are
_agreed_. So these people are all agreed about some fact F. One might
ordinarily parse this as: Each person agrees F is true. Or: Each
person asserts that F is true (or simply: each person asserts/believes F).
But all of these appear ruled out because they would reestablish the
connection between "…knows X" and "… believes X true." So what does
"agree about F" mean, when it doesn’t include belief-in-the-truth-of?
> Too simple really.
> Bound to be major problems there…
Well, if you elaborate your account a bit, you wind up approximately
where the cultural relativists have been for ages. You’re just
another one of those ancient Greek sophists. ;-)
But that makes you modern, too, because one way of reading
the sophists is pretty modern: Forget truth–All that really
matters is spin (public relations).
Regards,
Larry Mayhew may…@wku.edu
may…@wku.edu wrote:
> The question here is whether such an example provides much insight
> into knowledge. It’s usually treated as a contrivance of connivance:
Is that the same as a marriage of convenience?
> We connive to speak as if the myths are true because speaking thus
> is simpler. We’re prepared to elaborate, of course, if some
> innocent ignoramus wanders into the conversation and takes us at our
> word.
Ok, I’ll be the innocent ignoramus. I thought a better example
rather than mythology, would have been physical theories.
Starting with Newton: If I drop an apple on my toe, do I know
it will hurt? I think a fact is something which has already
happened. So I would say before the fact, I have a justified
true belief. In essence, I am predicting that the apple will
hit my toe and cause pain. I justify this prediction by making
the assumption that I exist in a deterministic space with
consistent physical laws. But is this a fact? I think not.
Einstein: Within this domain or context the predictions are
based upon considerations not accounted for within Newton.
What is true in one domain is not necessarily true within
another; different predictions are generated.
Knowledge hardly seems limited to just facts. I think we
include also our beliefs about relationships between these
facts. These relationships/rules are verified (seen to be
true) by observations which confirm consistency of prediction.
So knowledge is what we "know" to be true and what we expect
to be true. This expectation has been tested for relevancy.
Quantum: I think we know that there are sets of equations
which predict results with a high degree of experimental accuracy.
This is a domain in which the predictabiliy of Relativity loses
meaning somewhat like Newtonian physics loses relevancy to GR.
We find 7 or 8 major interpretations of the equations involved.
It seems unlikely that they are all true and it may be none of
them are true. But knowing the different interpretations is
considered part of having knowledge about quantum theory even
though the interpretations cannot be considered facts(aside from
the fact that they have existence;it is a fact that we have a
concept ‘unicorn’ whether or not one manifests in reality).
Most of the intrepretations, I believe, imply a probabilistic
or indeterministic universe. However, Bohm has produced a ‘hidden
variable’ model which is consistent with experimental quantum
data and is ‘deterministic’. So we do not know if the universe
is deterministic or indeterministic in an ultimate sense. Perhaps
it is transcendental like wave-particle duality.
Facts may not be relative but it seems like specifying or
describing a fact is. These descriptions appear to me to
need to specify a domain for relevancy and consistency of prediction.
Facts are represented by justified true beliefs as above.
Anyway I think this example has more flexibility than the Hera/Zeus.
So Below,
Stephen
In article <6hlf9v$…@ux.cs.niu.edu>, Neil Rickert <rick…@cs.niu.edu> wrote:
>When I say "I know that X", I mean something very different from when
>I say "I believe that X." However, it cannot be that
> I know that X
>is equivalent to
> I believe that X, I believe that X is true, and I believe that I am
> justified in believing that X,
Sure, but this paraphrase does not accurately reflect the commitments
of the JTB theory. The JTB theory has it that the truth of what you,
NR, claimed when you said "I know that X" will stand or fall with the
conjunction of the following three
(1) NR believes that X
(2) X [ is true ]
and (3) NR is justified in believing that X.
These may be assessed from a third person perspective as well as your
own first person perspective.
Moore’s paradox was to explain the oddity of "p but I don’t believe p",
which is logically consistent but cannot be significantly asserted
(ditto for "I believe falsely that p"). I take it the way out is to
recognize that there may be "fine structure" distinctions in the
content of a proposition that are invisible at at the point of
first-person present tense "freestanding" uses (the distinct contents
are as it were collapsed into one), but are discerniable when we
consider the same function (predicate) as it occurs in both
first-person and third-person uses. You can also distinguish the
contents from the first person by considering differences when the
contents occur embedded (not "free-standing") in the antecedents of
conditionals or in negations.
That is, for example, in your context the value of some "p" is always
equal to that of "I believe that p". But the value of "IF p THEN p"
does not equal "IF I believe that p THEN p". Nor will "PAST(p)" always
equal "PAST(I believe that p)". In that sense the two are
distinguishable in inferential role hence content, even if the
distinction cannot show itself in freestanding first-person present tense uses.
I could point you to some interesting references on this issue if
you are interested.
>for typically "I believe that X" already carries those other
>implications.
[ i.e. I believe that X, I believe that X is true, and I believe that I am
justified in believing that X, ]
First, I am not sure this is always the case. A detective might say "I
believe with all my heart he’s guilty, but I just can’t prove it". It
is not clear to me the detective even believes he is *justified* in this
case.
Secondly, you need to get clearer on what the sense is in which the
statement "carries these implications". Not everything one can deduce from
the act of your making a statement is an implication of the *content* of
the statement. For example, you might question someone sharply and thereby
*express* your impatience, and someone else can therefore infer that
you are impatient, but you are not stating or describing your impatience.
In that sense you do not assert you have a justification when you
claim that p.
>Yet, when I say "I know that X", I am saying something about my
>internal state, and I am not saying anything about whether others
"Internal" is misleading here, I would suggest.
>would judge X to be true, or would judge me to be justified in making
>my assertion.
No, but you are still putting your neck out with respect to the truth
of p and your justification for it; you make yourself vulnerable to
refutation if it turns out on further showing that not p, for example,
or that your justification depended on a faulty inferential step.
>my assertion. It is possible that I might claim "I know that X",
>even when I am aware that the rest of the world judges X to be
>false.
Of course it is, so what?
>false. Likewise there are times that I might assert "I know that X",
>yet not be surprised to discover I was mistaken.
That strikes me as odder, I’m not sure what you mean. Anyway, it does
not much matter to the JTB if you are *surprised* or not, it only matters
that you are in fact *mistaken* in that case.
>And what do you make of assertions such as:
> John knows he has AIDS, but he does not believe it.
Isn’t this is patently a somewhat paradoxical way to put it? I presume
it aims to express the sort of fragmentation of John’s psyche we call
"denial", in which not all of his beliefs are rationally integrated in
governing his action in the right way. Sure that can happen, but I do not
think we should base a theory of knowledge on this sort of case — it
is clearly a certain kind of normative failing of John’s rationality that he
refuses (presumably) to act in ways appropriate to what he knows.
But then we might say the part that does not believe it does not know it, and
the part that does know it also believes it.
In any case I don’t think we ought to rest an account of knowledge on
this sort of phenomenon. Concepts of epistemic conduct function by way
of reference to norms of rationality, e.g. the norm of inferential
integration of everything you believe in one system of knowledge.
Actual human reasoners will fall short of this norm in various ways,
but still the concepts work by reference to the ideal. You might
compare: you wouldn’t attempt to base an account of the design of a
computer program or automobile engine based on a *malfunctioning* i.e.
instance.
> I can’t believe I ate the whole thing.
This seems pretty obviously metaphorical, for something like "I am
amazed I really did eat the whole thing". I expect tests for belief will
show the person does in fact believe it.
> I don’t know what kind of weather we will have today, but I believe
> it is going to rain.
Not sure what the problem is here. Do you mean: because I have good,
but only probabilistic grounds for the belief? In that case it seems
fine. You can’t really *know* it’s going to rain if the probability is,
say, 80%, although you might be justified in expectation of rain.
Or do you mean, I have no grounds at all, I just have an unaccountable
hunch which has led to a very strong conviction on the topic? In that
case, we might indeed wonder why you are so irresponsible as to trust
your hunch if actually use it as the basis for a strong conviction
(presumably without having inductively validated the reliability of
similar hunches in the past). But still, it is possible. Like the
detective above, you can have a conviction (belief) but not claim
knowledge of the matter, sure, no problem for JTB as an account of
knowledge.
ander…@pitt.edu (Anders N Weinstein) writes:
- Hide quoted text — Show quoted text -
>In article <6hlf9v$…@ux.cs.niu.edu>, Neil Rickert <rick…@cs.niu.edu> wrote:
>>When I say "I know that X", I mean something very different from when
>>I say "I believe that X." However, it cannot be that
>> I know that X
>>is equivalent to
>> I believe that X, I believe that X is true, and I believe that I am
>> justified in believing that X,
>Sure, but this paraphrase does not accurately reflect the commitments
>of the JTB theory. The JTB theory has it that the truth of what you,
>NR, claimed when you said "I know that X" will stand or fall with the
>conjunction of the following three
> (1) NR believes that X
> (2) X [ is true ]
>and (3) NR is justified in believing that X.
>These may be assessed from a third person perspective as well as your
>own first person perspective.
>Moore’s paradox was to explain the oddity of "p but I don’t believe p",
>which is logically consistent but cannot be significantly asserted
>(ditto for "I believe falsely that p"). I take it the way out is to
>recognize that there may be "fine structure" distinctions in the
…
You are making a valient attempt to defend what I think is an
untenable characterization of what people mean when they say "I know
that X", or "He knows that X."
I suggest that "I believe that X" signifies some kind of emotional
committment, perhaps accompanied by some kind of intellectual
committment. And "I know that X" signifies that in some sense I live
my life in accordance with X. So "know that" is deeper but less
emotional and less a matter of intellectual committment than is
"believe that."
In article <6hlf9v$…@ux.cs.niu.edu>, rick…@cs.niu.edu (Neil Rickert) writes:
> ander…@pitt.edu (Anders N Weinstein) writes:
>>It might be clearer to say: the standard idea is that the locution
>> S knows that p
>>can be analyzed as the conjunction of three conditions:
>> (B) S believes that p
>> (T) p is true
>> (J) S is justified in believing that p
>>Note this is not really an analysis of the term "knowledge".
[snip]
> When I say "I know that X", I mean something very different from when
> I say "I believe that X." However, it cannot be that
> I know that X
> is equivalent to
> I believe that X, I believe that X is true, and I believe that I am
> justified in believing that X,
This isn’t the standard analysis; it is, as you say, an obviously
incorrect analysis. The standard analysis is:
"I know that X" is equivalent to:
I believe that X
X is true NOT: I believe that x is true
I am justified in believing X NOT: I believe I am justified …
> for typically "I believe that X" already carries those other
> implications.
Not if you do the analysis in the standard way. "I believe that X"
does not imply "X is true". Please note that I’m not defending the
standard analysis; I’m saying–let’s get it right before rejecting it.
Regards,
Larry Mayhew may…@wku.edu
In article <6hm1cn$…@ux.cs.niu.edu>, Neil Rickert <rick…@cs.niu.edu> wrote:
>ander…@pitt.edu (Anders N Weinstein) writes:
>I suggest that "I believe that X" signifies some kind of emotional
>committment, perhaps accompanied by some kind of intellectual
>committment. And "I know that X" signifies that in some sense I live
>my life in accordance with X. So "know that" is deeper but less
>emotional and less a matter of intellectual committment than is
>"believe that."
I would agree wholeheartedly that *belief* is a species of commitment.
According to a sort of minimal pragmatism about belief, you do not
really believe something unless you rely upon it in rationally planning
and guiding your conduct.
However, it is a rather special species of the genus, and more needs to
be said about this. With respect to your claim, I would say it simply
cannot be that "s knows that p" just denotes a *deeper* commitment than
belief. That just fails to reproduce the most basic logic of the
concept, for example, the truth condition, the fact that you can’t know
something that’s not true
If you are interested in a fascinating account of belief as a kind of
commitment, you really should look into Bob Brandom’s (long) book
_Making it Explicit_. Brandom puts forward a very highly developed
model of the structure of specifically *assertional* commitment, the
sort of commitment you undertake when you make an articulate public
claim in a socially constituted "game of giving and asking for
reasons". I think this account is very illuminating, although the book
is dense and I do not endorse all features of Brandom’s model.
On his view, to explain what it is to assert something is to explain
how the "score" in the game, i.e. yoor status in terms of obligations
and entitlements, is transformed with the assertion, i.e. to give a
kind of kinematics of social "deontic" statuses (commitments and
entitlements). For example, if you undertake commitment to p, others
can thereby take you to be committed to the consequences of p,
even if you have not drawn them all. A complicating twist is that
others may have different conceptions of the inference rules and can
use them in evaluating changes in your commitments.
He wants to explain the concept of "belief" derivatively, by analogy to
the concept of public assertional commitment in such a justification
game. I.e. belief is conceived as a kind of inner analog of the same
sort of state change, bearing the same sort of properties.
According to Brandom’s social normative pragmatics, one can explain the
concept of knowledge if one considers the interplay of first person and
third person perspectives in the social practice of playing assertional
language games. When you make a claim that p, remember, you make a
certain move in the game that changes the score (your state in the
game) in various ways. Now On Brandom’s view, when another player,
asserts "NR knows that p" he is doing several things in the game:
first, he is *attributing* one commitment to you, the assertional
commitment to p. Second, he is *attributing* entitlement to that
commitment to you as well – that is the traditional justification
condition. But, finally, he is also making a move of his own, namely,
performing an act of *undertaking* himself a commitment to p. That is
the "truth" condition, i.e. a matter of myself actually undertaking a
commitment and so responsibility for the consequences as well.
That story at least gets some of the logic right, while puporting to
explain the function of the term "knows" in the game in a broadly
pragmatist way.
In article <FQvlaQBByOdu@axp1>, <may…@wku.edu> wrote:
>> [Rickert] However, it cannot be that
>> I know that X
>> is equivalent to
>> I believe that X, I believe that X is true, and I believe that I am
>> justified in believing that X,
>This isn’t the standard analysis; it is, as you say, an obviously
>incorrect analysis. The standard analysis is:
> "I know that X" is equivalent to:
> I believe that X
> X is true NOT: I believe that x is true
> I am justified in believing X NOT: I believe I am justified …
Yes. To avoid further confusion I would recommend completely removing
the first-person pronoun and just consider an analysis of the
(schematic) predicate
(lambda(S) (S knows that p))
This is a common function that can be discerned in both first-person
and third person sentences, a predicate that can be applied to "I" or
"Jones" or "The man who shot Liberty Valance" or any other subject.
But it must have the same content in all of them, else you could not
use it as a link between first person and third-person statements in
rational inferential chains without committing a fallacy of
equivocation.
may…@wku.edu writes:
>In article <6hlf9v$…@ux.cs.niu.edu>, rick…@cs.niu.edu (Neil Rickert) writes:
>> When I say "I know that X", I mean something very different from when
>> I say "I believe that X." However, it cannot be that
>> I know that X
>> is equivalent to
>> I believe that X, I believe that X is true, and I believe that I am
>> justified in believing that X,
>This isn’t the standard analysis; it is, as you say, an obviously
>incorrect analysis. The standard analysis is:
Regrettably, you took my comment out of context.
I said (as quoted above, "it cannot be that …". So you haven’t
criticized anything I claimed. However, I also pointed out that when
I say "I know that X", I am making an internal judgement. I do not
need to consult any truth-recognizing oracles. So it cannot be that
I am applying any standard of truth external to myself.
Perhaps you are saying that the standard JTB analysis forbids the
assertion "I know that X". But in that case you would be agreeing
with me that there is something wrong with the JTB idea.
ander…@pitt.edu (Anders N Weinstein) writes:
- Hide quoted text — Show quoted text -
>In article <6hm1cn$…@ux.cs.niu.edu>, Neil Rickert <rick…@cs.niu.edu> wrote:
>>ander…@pitt.edu (Anders N Weinstein) writes:
>>I suggest that "I believe that X" signifies some kind of emotional
>>committment, perhaps accompanied by some kind of intellectual
>>committment. And "I know that X" signifies that in some sense I live
>>my life in accordance with X. So "know that" is deeper but less
>>emotional and less a matter of intellectual committment than is
>>"believe that."
>I would agree wholeheartedly that *belief* is a species of commitment.
>According to a sort of minimal pragmatism about belief, you do not
>really believe something unless you rely upon it in rationally planning
>and guiding your conduct.
>However, it is a rather special species of the genus, and more needs to
>be said about this. With respect to your claim, I would say it simply
>cannot be that "s knows that p" just denotes a *deeper* commitment than
>belief. That just fails to reproduce the most basic logic of the
>concept, for example, the truth condition, the fact that you can’t know
>something that’s not true
Of course it fails to reproduce "the most basic logic of the
concept." What I am arguing is that the supposed logic of JTB is
entirely bogus.
- Hide quoted text — Show quoted text -
In article <353E6EE3.4…@pacbell.net>, Stephen Harris <mulcy…@pacbell.net> writes:
> may…@wku.edu wrote:
>> The question here is whether such an example provides much insight
>> into knowledge. It’s usually treated as a contrivance of connivance:
> Is that the same as a marriage of convenience?
>> We connive to speak as if the myths are true because speaking thus
>> is simpler. We’re prepared to elaborate, of course, if some
>> innocent ignoramus wanders into the conversation and takes us at our
>> word.
> Ok, I’ll be the innocent ignoramus. I thought a better example
> rather than mythology, would have been physical theories.
> Starting with Newton: If I drop an apple on my toe, do I know
> it will hurt? I think a fact is something which has already
> happened. So I would say before the fact, I have a justified
> true belief. In essence, I am predicting that the apple will
> hit my toe and cause pain. I justify this prediction by making
‘Fact’ was the term used in the original message and I think he meant
any proposition people agree about. In other words, universal props
could be facts in that sense. My use of it simply meant that I didn’t
want to quarrel over semantics. Maybe you do; but sooner or later we
need to address what he intended rather than how he said it.
> the assumption that I exist in a deterministic space with
> consistent physical laws. But is this a fact? I think not.
> Einstein: Within this domain or context the predictions are
> based upon considerations not accounted for within Newton.
[etc.]
> Quantum: I think we know that there are sets of equations
[etc.]
> Anyway I think this example has more flexibility than the Hera/Zeus.
But he was making a point with Hera/Zeus that you are not making with
your examples. He claimed that we can know X without believing X and
without X being true. This is a surprising claim and goes beyond
anything you say. Also, I think he has a lot riding on the claim, so
I spent some time on it.
My only point here is this: Your examples may contribute something to
the discussion, but they don’t contribute anything that I can see to the
Hera/Zeus business.
Regards,
Larry Mayhew
- Hide quoted text — Show quoted text -
In article <6hm1cn$…@ux.cs.niu.edu>, rick…@cs.niu.edu (Neil Rickert) writes:
> ander…@pitt.edu (Anders N Weinstein) writes:
>>> I believe that X, I believe that X is true, and I believe that I am
>>> justified in believing that X,
>>Sure, but this paraphrase does not accurately reflect the commitments
>>of the JTB theory. The JTB theory has it that the truth of what you,
>>NR, claimed when you said "I know that X" will stand or fall with the
>>conjunction of the following three
>> (1) NR believes that X
>> (2) X [ is true ]
>>and (3) NR is justified in believing that X.
>>These may be assessed from a third person perspective as well as your
>>own first person perspective.
> You are making a valient attempt to defend what I think is an
> untenable characterization of what people mean when they say "I know
> that X", or "He knows that X."
No, I don’t think he’s making any such attempt (valiant or not). He’s
not defending the analysis, but trying to explain what it is. This
comment is in that same vein: The analysis isn’t trying to capture
everything people mean. The analysis is trying to provide a set of
truth conditions, the conjunction of which has the same truth value
as the knowledge claim. It’s widely agreed that the analysis fails;
the disagreement is over whether (and if so, how) it can be repaired.
> I suggest that "I believe that X" signifies some kind of emotional
> committment, perhaps accompanied by some kind of intellectual
> committment. And "I know that X" signifies that in some sense I live
> my life in accordance with X. So "know that" is deeper but less
> emotional and less a matter of intellectual committment than is
> "believe that."
"I believe your calculation gives the wrong answer" may be said with
or without an emotional commitment. I don’t understand why you think
"I know I’m going to die shortly" has less emotional commitment attached
to it than "I believe I’m going to die shortly." But in any case, the
JTB analysis isn’t about trying to include or exclude such connections.
It’s about truth conditions. If the three conditions are true, then
the knowledge statement is true, and vice-versa.
A strategy for refuting the JTB is thus straightforward: Describe a
situation in which "P knows X" is true, but "P believes X" is false,
or X is false, or P is not justified in believing X.
Or describe a situation in which "P knows X" is false, but "P believes
X" is true, X is true, and P is justified in believing X.
Regards,
Larry Mayhew
In article <6hocmc$…@ux.cs.niu.edu>, rick…@cs.niu.edu (Neil Rickert) writes:
> may…@wku.edu writes:
>>In article <6hlf9v$…@ux.cs.niu.edu>, rick…@cs.niu.edu (Neil Rickert) writes:
>>> When I say "I know that X", I mean something very different from when
>>> I say "I believe that X." However, it cannot be that
>>> I know that X
>>> is equivalent to
>>> I believe that X, I believe that X is true, and I believe that I am
>>> justified in believing that X,
>>This isn’t the standard analysis; it is, as you say, an obviously
>>incorrect analysis. The standard analysis is:
> Regrettably, you took my comment out of context.
If this is true, I apologize, but I’m not sure it is true. Let me be
more explicit:
Your statement below clearly has the form: P cannot be equivalent to Q
>>> However, it cannot be that
>>> I know that X
>>> is equivalent to
>>> I believe that X, I believe that X is true, and I believe that I am
>>> justified in believing that X,
I think the context authorizes me to suppose that you are denying the
JTB model of knowledge. But (I say) the JTB model of knowledge does not
say or imply that P is equivalent to Q. So what you are denying is
an equivalence that no one has asserted. The JTB model asserts an
equivalence that looks sort of like the one you mention, but they’re
not the same.
I gave my reasons in the message you take exception to. Here I am only
concerned to single out what I believe you said and what I was saying
in response.
> Perhaps you are saying that the standard JTB analysis forbids the
> assertion "I know that X". But in that case you would be agreeing
> with me that there is something wrong with the JTB idea.
No I wasn’t saying that. I was saying that in the quote above you
believe you are discussing the JTB model of knowledge, but you are
not.
Regards,
Larry Mayhew may…@wku.edu
On 21 Apr 1998 02:38:23 GMT, math…@math.canterbury.ac.nz (Bill
Taylor) wrote:
>For a start, knowledge does neither have to be true, nor a belief, nor
>any combination of these.
Agreed.
Joshua Stern
JRSt…@gte.net
| When I say "I know that X", I mean something very different from when
| I say "I believe that X." However, it cannot be that
| I know that X
| is equivalent to
| I believe that X, I believe that X is true, and I believe that I am
| justified in believing that X,
Is the point here that if the JTB analysis were correct, "I believe I
know X" would be synonymous with "I believe I believe that X,
I believe that X is true, and I believe that I am justified in believing
X"? And hence with "I believe that X, and I believe that I am
justified in believing X"? I’m afraid this doesn’t sound so implausible
to me, at least if by "justified" we mean a strong enough sense of
justification.
Keith Ramsay "Thou Shalt not hunt statistical significance with
kram…@aol.com a shotgun." –Michael Driscoll’s 1st commandment
may…@wku.edu writes:
>In article <6hm1cn$…@ux.cs.niu.edu>, rick…@cs.niu.edu (Neil Rickert) writes:
>> You are making a valient attempt to defend what I think is an
>> untenable characterization of what people mean when they say "I know
>> that X", or "He knows that X."
>No, I don’t think he’s making any such attempt (valiant or not). He’s
>not defending the analysis, but trying to explain what it is. This
>comment is in that same vein: The analysis isn’t trying to capture
>everything people mean. The analysis is trying to provide a set of
>truth conditions, the conjunction of which has the same truth value
>as the knowledge claim. It’s widely agreed that the analysis fails;
>the disagreement is over whether (and if so, how) it can be repaired.
No, this is completely wrong. The analysis (or misanalysis) does not
provide any truth conditions whatsoever. Weinstein very openly
evades all questions of truth conditions (criteria for truth).
may…@wku.edu writes:
>Your statement below clearly has the form: P cannot be equivalent to Q
>>>> However, it cannot be that
>>>> I know that X
>>>> is equivalent to
>>>> I believe that X, I believe that X is true, and I believe that I am
>>>> justified in believing that X,
>I think the context authorizes me to suppose that you are denying the
>JTB model of knowledge. But (I say) the JTB model of knowledge does not
>say or imply that P is equivalent to Q. So what you are denying is
>an equivalence that no one has asserted. The JTB model asserts an
>equivalence that looks sort of like the one you mention, but they’re
>not the same.
But what is JTB claiming? Is it claiming that the meaning of "knows
that" is such that the expression
John knows that X
may never be asserted? For the proponents of JTB will require that X
be true. They require that "the asserter believes X is true" is not
sufficient. And they deny that there are any general criteria for
identifying what is true.
This leaves us in the position that nobody ever has a basis for an
assertion with ".. knows that .."
rick…@cs.niu.edu (Neil Rickert) says…
>…I also pointed out that when I say "I know that X", I am
>making an internal judgement. I do not need to consult any
>truth-recognizing oracles. So it cannot be that
>I am applying any standard of truth external to myself.
There is a distinction between the meaning of a phrase and
the conditions under which it is uttered. If someone says
"I know that Chicago is the capital of Illinois" then they
have simply uttered a falsehood. Chicago is *not* the capital
of Illinois.
In another post, you say:
>But what is JTB claiming? Is it claiming that the meaning of
>"knows that" is such that the expression
> John knows that X
>may never be asserted?
That doesn’t follow from JTB!
>For the proponents of JTB will require that X be true.
No. JTB simply says that X must be true in order for
"John knows that X" to be true. A statement doesn’t
have to be true in order for me to assert it.
>They require that "the asserter believes X is true" is not
>sufficient.
It is not sufficient to establish the *truth* of
"John knows that X". In exactly the same way that
my belief that Chicago is the capital of Illinois
is not sufficient to establish the truth of whether
Chicago is the capital of Illinois.
>And they deny that there are any general criteria for
>identifying what is true.
So?
>This leaves us in the position that nobody ever has a basis for an
>assertion with ".. knows that .."
That doesn’t follow from JTB! According to JTB, John knows
that X is equivalent to (1) John believes X, (2) John is
justified in believing X, and (3) X is true. Therefore,
I have a basis for asserting "John knows that X" if
(1) I have a basis for asserting that John believes X,
(2) I have a basis for asserting that John is justified
in believing X, and (3) I have a basis for asserting X.
(To my mind, there is no distinction between "asserting
X" and "asserting that X is true".)
Daryl McCullough
CoGenTex, Inc.
Ithaca, NY
In article <6hocmc$…@ux.cs.niu.edu>, Neil Rickert <rick…@cs.niu.edu> wrote:
>may…@wku.edu writes:
>Perhaps you are saying that the standard JTB analysis forbids the
>assertion "I know that X". But in that case you would be agreeing
>with me that there is something wrong with the JTB idea.
Even if this were correct, still it would not necessarily show
any inadequacy in the analysis. It would just show that the truth
conditions of the claim were different from the assertibility conditions.
I think you are not sufficiently appreciating a certain Fregean point
about truth-conditional content, viz. that a proposition has the same
content when it occurs unasserted in the antecedent of a conditional,
and that present and past tense versions of a proposition are
semantically linked as well.
For example, it may be you can never assert "I believe falsely that p".
Still it is perfectly possible that you believe falsely that p. Another
person can assert it, you can say it in the past tense, and you can
even use it as an antecdent of a conditional or as a premise in a bit
of hypothetical reasoning.
It is similar for pairs like "p" and "I believe that p": perhaps there is
no difference in assertibility conditions, in that you will always
assert both in exactly the same circumstances. Still there is a
difference in truth conditional content, they cannot be semantically
equivalent for they clearly have different logical consequences.
Compare: "if it’s raining then the streets will be wet";
if I believe that it’s raining then the streets will be wet.
The main problem with simple verificationisms is that they
only take into account assertibitility conditions, they do not
take into account what Dummett (in Frege: Philosophy of Language)
calls "ingredient sense", the contribution a sentence makes to a
complex containing it. Brandom has a nice paper on this topic called
"Truth and Assertibility", I could look up the reference.
The upshot is quite significant, I think, e.g. you can’t simply say
that "p" really means "I believe that p" or anything like that.