70
PARTISAN REVIEW
phors in general. To take metaphors literally is, however, sometimes
to take them nonsensically, and in philosophy, as elsewhere, is fre–
quently the mark of a certain naivety.
In regarding some metaphysical statements as metaphorical we
have assumed that a statement may be fictive despite the fact that
it is asserted as cognitive. Are we justified in calling a statement meta–
phorical when it is proposed, like metaphysical doctrines, as literal
truth? The historical and psychological role of metaphors
in
religion,
philosophy and science supports an affirmative answer to this question.
If
the doctrines which provoke the endless disputes of metaphysicians
have no literal meaning, we must, in justice to their seriousness, look
for some implicit, metaphorically expressed meanings which form the
subject of their discussion. The issues involved in metaphysical argu–
ments may have serious scientific or logical import but sometimes
inadequate analytical methods compel the metaphysicians to express
these issues in analogical fictions. Metaphorical expressions of mean–
ing often antedate the clear analysis of their cognitive content. Num–
erous instances of this can be found in psychological studies of chil–
dren, and in primitive anthropomorphic accounts of natural phen–
omena. More recondite and perhaps also more important examples
are offered by the history of science. When electrical phenomena were
first made the subject of extensive scientific analysis, they were ex–
plained analogically or metaphorically by the theories of the highly
developed sciences of mechanics and hydrodynamics. These theories
persisted in electrical science despite the fact that many ·of their con–
cepts had no empirically determinable meaning in electrical phenom–
ena. It is only in recent decades that electrical theory was freed of
these mechanical fictions.
Fictions and Empirical Objects
Berkeley, Hume as well as early positivists, considered the empir–
ical
statements~
science to be limited to the description of the con–
cur~ence
or sequence of sensations. In opposition to rationalistic
claims to
a priori,
speculative knowledge, sensationalistic empiricism
(Mach, Pearson) sought to limit science to the immediate data of per–
ception. Science, within this restriction, must avoid speculation con–
cerning imperceptible entities. This limitation, however, excludes
from science not merely the usual metaphysical concepts, but also
such elements of scientific hypotheses as atoms, electrons, potentials,
and forces. Strictly speaking, it would eliminate even the idea of pro–
pagated light, since light is never perceived as
moving
from one place
to another. To circumvent the difficulties arising from their omission,
these concepts were tolerated in science as convenient, methodological
