discusses the view that we must distinguish the act of perceiving from the object perceived, and
that the former is mental while the latter is not. His argument against this view is obscure, and
necessarily so, since, for one who believes in mental substance, as Berkeley does, there is no valid
means of refuting it. He says: "That any immediate object of the senses should exist in an
unthinking substance, or exterior to all minds, is in itself an evident contradiction." There is here a
fallacy, analogous to the following: "It is impossible for a nephew to exist without an uncle; now
Mr. A is a nephew; therefore it is logically necessary for Mr. A to have an uncle." It is, of course,
logically necessary given that Mr. A is a nephew, but not from anything to be discovered by
analysis of Mr. A. So, if something is an object of the senses, some mind is concerned with it; but
it does not follow that the same thing could not have existed without being an object of the senses.
There is a somewhat analogous fallacy as regards what is conceived. Hylas maintains that he can
conceive a house which no one perceives, and which is not in any mind. Philonous retorts that
whatever Hylas conceives is in his mind, so that the supposed house is, after all, mental. Hylas
should have answered: "I do not mean that I have in mind the image of a house; when I say that I
can conceive a house which no one perceives, what I really mean is that I can understand the
proposition 'there is a house which no one perceives,' or, better still, 'there is a house which no one
either perceives or conceives.'" This proposition is composed entirely of intelligible words, and
the words are correctly put together. Whether the proposition is true or false, I do not know; but I
am sure that it cannot be shown to be selfcontradictory. Some closely similar propositions can be
proved. For instance: the number of possible multiplications of two integers is infinite, therefore
there are some that have never been thought of. Berkeley's argument, if valid, would prove that
this is impossible.
The fallacy involved is a very common one. We can, by means of concepts drawn from
experience, construct statements about classes some or all of whose members are not experienced.
Take some perfectly ordinary concept, say "pebble"; this is an empirical concept derived from
perception. But it does not follow that all pebbles are perceived, unless we include the fact of
being perceived in our definition of "pebble." Unless we do this, the concept "unperceived pebble"