In section 3.3 of “Philosophy of
Natural Science,” Hempel argues that crucial tests are not sufficient enough to
prove a given hypothesis or to disprove them. Hempel states what some may believe why a crucial test can
prove or disprove a hypothesis’. If
there are two competing hypothesis which involve the same subject and no
available evidence favors one or the other, then there exists a test, which
will produce conflicting outcomes for the different hypotheses. This test is the so-called crucial test
would then presumably refute one hypothesis while supporting the other.
(Hempel, 26) Hempel then presents his
side of this argument using an example of past
experiments involving the nature
of light. He describes how Foucault
performed an
Experiment involving the velocity of light through air and
water. This experiment was meant to
show whether light consists of waves or extremely small particles as presented
by Newton. Foucault’s experiment was
performed and the resulting outcome was used to refute Newton’s view of light
as small particles traveling at a high velocity. Hempel believed that this test was not strong enough to
completely support or refute either view of light. The experiment relied on the assumption that light as waves would
travel faster in air than in water.
However Hempel argues that the conception of light as streams of
particles was too indefinite to assume that it would travel slower in air
without additional assumptions about the motion of particles and their surrounding
medium. (Hempel, 26) So even though the
results may seem to support and prove the wave hypothesis, it doesn’t
necessarily disprove the particle theory.
In fact Einstein later proposed a theory, which eliminated the classical
wave theory, using support from an experiment by Lenard in 1903. (Hempel,
27) But again as in the previous
example one of the hypotheses was not definitely refuted, in this case being
the wave theory. This further supports
Hempel’s argument that experiments of this kind cannot completely refute or
support either of the competing hypotheses.
In
section 3.3 Hempel presents two experiments that provide sufficient support
that crucial tests cannot definitively prove one hypothesis while refuting the
other. Each of these experiments shows
how even though they may support one hypothesis it is not conclusive, and that
they cannot eliminate the rival hypothesis.
Hempel also draws on his earlier writing to help support this argument. He says that hypotheses or theories cannot
be conclusively proved by any set of data no matter how extensive. (Hempel,
28) You can never observe all cases of
any event so there is always a possibility that the hypothesis could be refuted
by an unknown case even if all observed cases had been supporting the
hypothesis. No matter how careful an
experiment is or how many cases it may cover there is always the chance that
another conflicting case exists. This
is strong evidence in support of Hempel’s argument in which he states that “a
crucial experiment is impossible in science.”
He does agree that experiments such as the ones he presented can provide
evidence in support of one claim or denial of another it can never be a
concrete and complete proof of two competing hypothesis.
Hempel, Carl. “Philosophy of Natural Science.” Prentice
Hall. 1966.