[FOM] Coherentism, pt. 1

A.P. Hazen a.hazen at philosophy.unimelb.edu.au
Sun Sep 28 01:30:30 EDT 2003

	In a post titled "Simpson on Tymoczkoism," I tried to characterize
the epistemological view called, in philosophertalk, "foundationalism":
roughly, the view that propositions are organized into a well-founded
hierarchy, with lower ones providing evidence for (or against) higher, but
not vice versa.  I said it was contrasted with "coherentism."  (Of course,
philosophers being what they are, there are lots of variants and hybrids
and subclassifications, so this is only a crude, broad-brush,
classification. Similar disclaimers are in order for what follows.)  In
this (2-part) post I try to say something more about coherentism.  This
really is meant to be relevant to FoM, but that may not be apparent until
part 2.

	F-ism and C-ism are views about evidence: what makes it
right/rational/appropriate to accept a proposition.  In many cases we can
answer the question "What makes you think P is true?" by citing some other
proposition, Q, as evidence for it.  Example: let P be "President Kennedy
was shot" and Q be "It was widely reported in the media that Kennedy was
shot, and the conspiracy needed for a hoax of such a nature would involve
too many people to be plausible."  Note that, in the example, we wouldn't
cite P as a reason for believing Q.  F-ism is the claim that, ideally, this
is the general case.

	C-ism looks at things  more abstractly, notes that there are
evidential connections between propositions (i.e.: believing one gives you
a reason to believe another), and says that a particular proposition ought
to be accepted if it "coheres" with all the others we accept: a
belief-worthy proposition is one that belongs to a system of belief-worthy
propositions, each of which is evidentially supported by others in the
system, but with no particular ranking, and with the possibility of
circularity: mutually supposrting propositions.  (I don't actually remember
any newspaer headlines from forty years ago: for me, now, that Kennedy was
shot is one of my reasons for believing that it was widely reported that he
had been shot!)

	The obvious objection is that the whole system seems to be free
floating, ungrounded.  Why believe real history and not "The Lord of the
Rings," given that the various propositions describing Tolkien's universe
form an intricate web, in which (for example) the translation given in one
chapter of an Elvish poem can be supporting evidence for the proposition
that Aragorn is in love with Arwen in a later one?

	Perhaps the most plausible response to the objection is to say that
we have some prior disposition to accept SOME propositions before we
consider the evidential relations between propositions.  (There may be lots
of supporting evidence for a proposition describing my current sensory
experience, but I feel a strong inclination to accept the proposition
"There are neighboring patches of green and white in my visual field" even
before considering the supporting evidence!  ...  Of course, other,
non-sensory, "intuitions" are also possible, and perhaps more relevant to
FoM.)  For any proposition, then, the test of belief-worthiness is  whether
it "coheres with" a system which also includes MOST of the propositions we
feel a STRONG antecedent urge to accept.  (This language immediately
suggests a formulation in terms of weighted sums, but-- alas-- few
epistemologists have gone details precisely enough to allow actual
mathematics.)		...		Of course, this formulation seems
to allow for a foundationalist structure, but it also allows for two
important departures from F-ism: (a) evidential relations can run in many
directions, so that the whole structure is not well-founded, and (b) ANY
antecedent disposition can be outweighed by enough other evidence, so there
is no analogue of the F-ists indubitably certain bottom stratum.

	This is the epistemological picture for which Quine coined the
metaphor of a "web of belief".  (Quine is by way of being a patron saint of

	Remark: Most versions of F-ism imply that certain kinds of
proposition, identifiable by their subject matter, intrinsically have to be
low-down in the evidential hierarchy-- they have the job of being evidence
for higher strata, and not the other way around.  C-ism typically rejects
this-- any proposition can provide evidence for any other.  (What with
after-images and floaters in my eyes, my visual field often contains things
I would  have difficulty classifying by color.  Was I really seeing white
next to green just now?  Well, the quad outside my office window is often
used as a backdrop for wedding photos, so the presence of a woman in a
white dress on the lawn is not unusual on weekends: this sociological
proposition, for me, is relevant supporting evidence that raises my
confidence of having described my sensory experience correctly.)

	Second remark: We are particularly interested in evidence for (or
against) propositions we are not sure of, and it's nice if we don't have
the same doubts about the evidence.  So it is tempting to think that there
is a hierarchy in which the things we are surer of act as evidence for the
ones we are less sure of, with things we are absolutely sure of at the
bottom.  Bertrand Russell, in "Human Knowledge: its scope and limits" wrote:
	"It is obvious that the conclusion of an argument cannot
	   derive from the argument a higher degree of credibility
	   than that belonging to the premisses: consequently, if
   	   there is such a thing as rational belief, there must be
	   rational beliefs not wholly based on argument."
(Considerations like this have made many philosophers lean toward
foundationalism.)  But Russell was wrong, as any crossword puzzle addict
knows.  You may not feel at all confident that "Persian King" in six
letters is "Xerxes", and even less  confident that  "Make a copy" in five
letters is "Xerox," but once you  notice that they have the same first
letter, each makes the other VERY credible.

	Next posting: Axioms, or what this has to do with FoM.
Allen Hazen
Philosophy Department
University of Melbourne

More information about the FOM mailing list