and θέσις, a placing under. The Latin
suppositio exactly corresponds in
formation and meaning, and hypothesis may very well be explained by
Aristotle (1) supplies us with the following account of the preliminaries of reasoning. If we have a point
of departure, as we needs must, which does not admit of proof itself,
because it is the basis of proof, and is not an axiom, it is to be
called a θέσις, position. A thesis or position may either state
thing is, in which case it is a definition; or it may state that it
is, or that it is not, and then it is an hypothesis.
The enunciations prefixed to the theoretical propositions of Euclid,
together with his definitions, are the theses which render his
demonstrations possible. That the radii of a circle are equal is the
result of his definition of a circle, and is a thesis of which he makes
continual use; that in a given proposition in the enunciation of which
he has laid it down that the angle A B C is equal to the angle A B D, we
are presented with an hypothesis, a supposition, on which he reasons. In
the demonstration the angles are equal by hypothesis.
The word, however, without essential change of meaning, is used for the
most part nowadays in a more especial sense. We mean by an hypothesis a
supposition not proved, and perhaps not provable by itself, adopted in
order to explain a set of phenomena. The history of physical science is
the record of a succession of such. The Ptolemaic system was an
hypothesis, so that of Tycho Brahe, so the Copernican.
We must not identify such hypotheses, as does Reid, with mere
conjecture. Doing this, he treats hypothesis of every sort with
unmitigated scorn, and pronounces that we should "despair of ever
advancing real knowledge in that way." (2) This, however, is to fly in the
face of all history. The right employment of hypotheses has continually
furnished the steps by which real knowledge has been gained and a true
theory arrived at.
How has hypothesis done this, and what is such right employment? The
answer to the first question is that an hypothesis furnishes us with a
reason for trying one set of experiments sooner than any other, and with
some principle of selection in experiments we must be furnished if we
are to make a beginning at all. We use our hypothesis aright when we use it for this reason, keeping steadily in
mind its provisional nature, and holding ourselves in readiness to
abandon it the moment the result of our experiments is adverse to it. It
ought too to be credible in itself, not at variance with any knowledge
that we possess, or any ascertained facts, whether in the sphere with
which we are at present engaged, or in any other.
Hypothesis thus used is plainly a path of progress. Even should it
ultimately prove untenable, it will have led to many experiments and
much ascertainment of fact. If, on the other hand, it be found to
explain all the phenomena, not only those which first stirred inquiry,
but the additional ones which such inquiry has brought to
view, if it do this with as much clearness in what might at first seem
adverse as in favourable cases, and if nothing else can explain them,
then the hypothesis may be regarded as established, and as providing us
with the true theory of the facts to which it has been applied. The
animal spirits and the vortices of Descartes have disappeared from the
world of science, as has the astronomical system of Ptolemy, while that
of Copernicus, at first as much an hypothesis as they, has taken rank
with established truth. But surely the Ptolemaic system led to much
important astronomical observation which would hardly have been made
without its guidance.
As hypothesis ought not to be identified with mere conjecture, so ought
it not to be with theory, which is a word of larger meaning (see
(1) Analyt Post. L. I. c. 2.
(2) REID on the Intellectual Poems, Essay I. cap. III.