Re: knowledge and belief

[Steven Belknap <sbelknap@UIC.EDU>]

>If an epistemology is simply the grounds for the knowledge, then
>what is the difference between:
>
>        la sokrates djuno le du'u la djan klama le zarci kei fo le nu viska
>        Socrates knows that John goes to the market by having seen him.
>
>        la sokrates jinvi le du'u la djan klama le zarci kei fo le nu viska
>        Socrates thinks that John goes to the market by having seen him.
>
<snip>
>I'm not sure who you're referring to. I don't think certainty has much
>to do with it. It's more a matter of whether or not the reported beliefs
>coincide with one's own/established beliefs or not. When you're
>reporting your own beliefs, then "know"/"think" can be used to mark
>the degree of certainty, because obviously the basic difference
>collapses for the first person: normally your own beliefs coincide
>with your own beliefs. But when reporting someone else's beliefs,
>we use "know" for beliefs that coincide with our own, and "think"
>for those beliefs which we don't necessarily share.
<snip>
>
>[Fuzzy logic is an interesting topic on its own, but I think
>more or less orthogonal to this.]
>
>co'o mi'e xorxes

<djuno> and <jinvi> are not synonyms. At least I hope they are not! I would
use <djuno> to refer to a mathematical fact, which can be absolutely true,
as mathematics is an artificial construction founded on assumed postulates.
I might also use <djuno> to refer to a religious fact, which is true
because it is asserted as true by an authoritative source. I might also use
it to refer to a Wittgensteinian philosophical assertion, (but then some of
you may recall that Wittgenstein himself considered philosphizing to be a
form of mental illness.) If I were in the military I would use <djuno> to
pronounce my certainty in the pronouncements of my drill sargeant. I would
*not* use <djuno> to refer to an assertion I was making about the real
world, but would instead use <jinvi> for these assertions. Fuzzy modifiers
could then be used to adjectivize either <djuno> or <jinvi>.

Thus, if I was somewhat uncertain as to the exact proof of the Pythagorean
Theorem, then I might modify djuno to reflect my uncertainty about the
exact proof, while still holding that the Theorem might be provable to be
absolutely true. Error is of course still possible in Mathematics, but an
essential property of Mathematics is that there are absolutely true things.
It is not possible to prove the second law of Thermodynamics or the Theory
of special Relativity, for these refer to the real world, and are
potentially incomplete or inaccurate. The difference between these two
gismu is the difference between references to a Platonic world and a real
world.

Jorge's use of know and think is interesting. I do not doubt that this is
how he uses these words, but this usage is very different from my own. When
I use these words to describe my knowledge about an assertion, I consider
the predicates know/believe/think/suspect/uncertain to form a fuzzy
continuum from certainty to uncertainty. This is imperfect in English as
the degree is not reducible to a numeric value, but the ordering was
absolutely consistent among the three native English speakers who were in
my lab. (So this is an ordinal, but not an interval scale.) Interestingly,
the one non-native English speaker (native Mandarin speaker) did not agree
with this analysis, although his English is quite fluent. He opined that
know/believe/think were synonyms!

Consider these sentences:

I know that flying saucers exist because I have flown on one.
I believe that flying saucers exist because I saw one.
I think that flying saucers exist because John told me he saw one.
I suspect that flying saucers exist because there are many unexplained
things that happen.
I am uncertain whether flying saucers exist
I suspect that flying saucers don't exist because most of the alleged
sightings have been explained.
I think that flying saucers don't exist because no credible sightings have
been reported.
I believe that flying saucers do not exist because extraordinary claims
require extraordinary evidence.
I know that flying saucers do not exist because the Bible forbids it.

In my usage in this context, "to know" means there is no doubt, "to
believe" means there is a trace of doubt, "to think" means I have slight
doubt, and invite correction. lojban is attractive to me because it seems
possible to express these gradations of certainty using fuzzy operators
which directly modify a predicate, rather than requiring multiple separate
predicates.

This is why I suggested that fuzzy logic might be useful.

-Steven



Steven Belknap, M.D.
Assistant Professor of Clinical Pharmacology and Medicine
University of Illinois College of Medicine at Peoria