[hist-analytic] Frederick's conception of the A Priori

Danny Frederick danny.frederick at tiscali.co.uk
Sun Apr 5 12:12:20 EDT 2009

Hi Roger,

I'm afraid you seem to have misunderstood most of what I said.

My reference to the 'justificational' practices of a cult was not
rhetorical. It raises a serious problem and one that is fundamental to
epistemology, namely, when is a practice of accepting statements one that
generates knowledge and when is it not? This is a slight generalisation of
Popper's 'problem of demarcation.' If you merely refer to the rules of the
'language game' of science, with its 'institutionally defined standards of
justification,' why should I regard that game as an epistemic or scientific
one rather than a game of some other kind? This is an outstanding question
for you to answer. Until you have provided an answer you have given no
reason for us to think you are engaging in epistemology rather than
sociology, social psychology or anthropology.

Popper, incidentally, said that his methodological rules were the rules of
the game of science; but he also argued that the game of science was worth
playing for epistemic reasons, because it was the only way of obtaining
better explanations. This has no reference to justification. Indeed, one of
his rules is that we never stop testing a theory, no matter how successful
it has been in the past. That rule acknowledges that justification is
unachievable. But theories can fare well or badly in tests and they can have
or lack other desirable features. Taking account of all this we can compare
and rank theories; and those that come out higher in the ranking are called
'better corroborated.' But the best-corroborated theory today might be
refuted tomorrow. So corroboration must not be confused with any kind of
epistemic justification.

I agree that this is a kind of scepticism. But I deny that it is 'out of
line both with ordinary and scientific norms.' It is not out of line with
scientific norms: scientists do continue testing and amending theories and
coming up with new ones. Although relativity theory is still the dominant
theory for the very large, there is all manner of work currently going on in
physics which violates this or that assumption of relativity. People like
Dennett say that relativity is 'proved beyond reasonable doubt;' but Dennett
is not a scientist and appears ignorant of what is going on in science. But
I agree with you that Popper's approach does violate lots of 'ordinary
norms,' such as those of astrologers, Marxists, psychoanalysts, mystics and
other windbags. So much the worse for them!

It is simply a fact that we are not in a position to ascertain which of two
statements is more likely to be true. What is the point of any philosophical
discussion that pretends that this is not so? It is just daydreaming. But
even though we cannot ascertain truth, probable truth or verisimilitude, we
can often ascertain when one theory is a better explanation than another,
for example, if it is simpler, more comprehensive, more precise and entails
surprising new predictions which survive severe testing. The history of
science gives us examples of such progress.

You ask: 'What purpose does it serve to use the term justification as you
are doing, for a standard which nothing will ever meet?' I would respond
with a question to you: what purpose does it serve to use the term
'justification' in epistemology for a standard which cannot be shown to
justify anything epistemically? It is misleading. Why not call a spade a

I agree that the notion of epistemic justification is worthless. That is why
I don't use it. Can you say that your use of 'justification' in epistemic
contexts does not imply or at least suggest that we can attain
demonstrations of truth, probable truth or verisimilitude? And since we
cannot attain such things, wouldn't it be better to drop the term
'justification' from epistemic contexts? How many people who use the term
are deluding themselves and others? I would guess it is a lot of them. For
some people this delusion seems important.

I am not a holist of any kind, and certainly not a radical one. I referred
to Duhem's argument, not Quine's overstatement of it in 'Two Dogmas.'
Duhem's position has been called 'largism.' Again, it is just a fact: the
writings of Duhem, Popper, Kuhn, Lakatos and Feyerabend give lots of
examples. It usually takes a big conjunction of theories to get a testable
consequence; and this makes it difficult (I think impossible) to demarcate a
priori from empirical propositions (rather than empirical from a priori
knowledge). This, incidentally, has nothing to do with any thesis about the
indeterminacy of meaning (that is also a Quinean thing).

I explained how Popper's demarcation criterion is unaffected by the Duhem
argument in an earlier email. Popper's demarcation of science from
pseudo-science was precisely a response to Duhem's argument. To save you
searching, here is what I wrote:

'Finally, going back to Popper, I said that the first step in his
demarcation of empirical science was the notion of falsifiability as
inconsistency with some (true or false) observation statements. This
distinguishes those systems of statements which are candidates for empirical
science from those that, for the time being at least, are not (i.e., those
which are logical, mathematical or metaphysical). The second step is to
distinguish science from pseudo-science. This is a matter of procedures. Any
theory can be saved from falsification by amending the other statements used
to derive empirical predictions from it or by rejecting any conflicting
observation statements. Such procedures are the mark of pseudo-science,
EXCEPT where the modifications to background knowledge that are made to save
a particular theory themselves lead to novel predictions which survive
severe testing.'

Thus a proposition or theory that is untestable at time t may be testable at
time t+n, so its status as empirical (and thus as a candidate for science)
will change. But whether a testable theory IS a part of science depends on
how severely it has been tested and what the results were.

I agree with you that there is a distinction between the 'formal' and the
empirical sciences. But that should be obvious from the fact that I offered
a demarcation between empirical and a priori knowledge. But where does this
leave metaphysics? It is not derivable from logical or mathematical axioms
and it is not empirically testable either. Yet in future, some of it may
become testable, like the theory that matter is composed of atoms, which was
an ancient metaphysical theory which became a part of science only
relatively recently.

I hope I have managed to reduce some of the misunderstanding.

Best wishes,


More information about the hist-analytic mailing list