[hist-analytic] Carnap on Philosophy

Danny Frederick danny.frederick at tiscali.co.uk
Mon Aug 24 14:58:51 EDT 2009

Hi Steve/Bruce,


I do not think there is any sharp distinction between physics and
metaphysics. It seems pretty clear that the two of you regard metaphysical
theories as being ones which are not empirically testable. But whether a
theory is empirically testable depends not just on the logical content of
the theory (i.e., what it says) but also on what other theories we conjoin
with it and on what technology we have developed. In ancient times, for
example, atomism was a metaphysical theory, being completely untestable. But
with the development of further theories and models, like the kinetic theory
of heat, and with the development of microscopes, atomism became empirically
testable. It passed from metaphysics into physics.


Bruce asks (and Steve accepts the questions): 'by what method or procedure,
are true metaphysical propositions ascertained? What shows them to be true?'


But the questions are misplaced. There is no method or procedure by which
the truth of ANY proposition may be ascertained or shown. We can test
theories; but even if they survive refutation, that does not show them to be


You may say: but how do we test metaphysical theories? I have just given the
answer: we develop them, or develop auxiliary theories, and we develop
experimental techniques, which may eventually enable us to test them. Grand
new scientific theories often start out as untestable metaphysics.
Relativity theory had to wait fourteen years before it could be tested.


Steve says: 'Metaphysical positions are taken not because they possess the
mark of truth or evidence but because one position is better than another.'


This is true in that no position possesses the mark of truth. But for many
metaphysicians AND SCIENTISTS metaphysical positions have been adopted
because the theorist concerned DID think that it possessed the mark of
truth. Kepler was a Pythagorean mystic for whom the sun was symbolic of the
deity, so it just had to occupy the centre of the universe (not just the
solar system). Descartes' physics was based on metaphysical intuition, as
was Leibniz's rejection of it in favour of fields of force (see Popper's
'Philosophy and Physics,' in which he shows how metaphysical disputes
inspired scientific theories; it is in his 'The Myth of the Framework').


Steve says: 'Whether, for example, causation is singular or a matter treated
within the compass of regularity views will not be decided on an evidential


Why not? Someone may develop a theory involving singular causation during
the course of scientific progress, and the theory may survive testing.
Admittedly, at first glance any theory postulating singular causes looks as
if it would be merely an ad hoc way of explaining away some difficulty with
another theory. But there seems to be no reason why such a claim of singular
causality should not, in combination with other theories about the marks of
causes, for example, turn out to be testable. When the motions of Uranus
refuted Newton's theory, the scientists said there must be another planet
with just the properties required to explain Uranus' divergences from
Newtonian predictions. It may seem totally ad hoc. But in fact, a planet
with the required properties would have been observable at certain times. So
they looked for it. And they found it: they called it 'Neptune.'


Science often progresses via untestable ('metaphysical') stop-gaps that
eventually turn out to be testable. For various examples see Feyerabend's
'Against Method' and Lakatos' 'Falsification and the Methodology of
Scientific Research Programmes.'


I am currently reading, bit by bit, chapter 3 of Bruce's book available on
this site. It is a very pleasant and worthwhile read; but as should be
evident from some of my statements above, I disagree with it. I reject the
notion of analytic truth entirely (but not for Quinean reasons). I will post
some criticisms in a message to the list later (I hope by the end of this


Best wishes,



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rbjones.com/pipermail/hist-analytic_rbjones.com/attachments/20090824/87e01855/attachment-0002.html>

More information about the hist-analytic mailing list