HUME: THE PROBLEM OF INDUCTION
 Hume holds that causal reasoning is, and must be, based on experience. Distinguish two
(C1) All effects have causes
(C2) All events have causes
The truth of (C1) is based on a relation among ideas: it is part of the idea of an effect that it be
caused. That is, ‘cause’ and ‘effect’ are correlative terms, just as much as ‘husband’ and ‘wife’. A
husband is someone’s husband, and a wife someone’s wife. Likewise, ‘effect’ means something
like ‘an event brought about by some other event’. Hume is interested in (C1) only to the extent
that it might be confused with (C2), not in its own right. Now (C2) is a matter of fact: its denial
is logically possible. Hume points out that what we identify as the cause and as the effect are
in fact no more than two distinct events; there is no more reason why a given event should a
priori be the effect of a given event as cause than any other event. Hume thinks that the point
is readily granted in the case of unfamiliar things. Isolated on a desert island, a castaway ﬁnds
an unfamiliar brownish-green kind of fruit growing on a tree. The fruit could equally well be
nourishing or poisonous, and no amount of a priori reasoning will serve to settle the question.
As with unfamiliar things, so too with familiar things. Hume, to his credit, takes an example
that seems undeniable: the motion of a billiard-ball being (apparently) caused by the impact of
the cue-ball. He argues that our immediate response is in fact a learned response, and it is merely
the force of habit (‘custom’) that makes it seem to be undeniable.
 Yet deny it Hume will. He holds that all causal inference has the following pattern:
As have been followed by Bs
Therefore: This A will be followed by a B
The inference is clearly not deductive, since it is at least possible that this A not be followed by
a B. Therefore, the conclusion is a matter of fact, and causal inference is inductive. Now what
reason or principle justiﬁes this inductive inference? The only principle that seems capable of
doing the job is what we may call the Uniformity-of-Nature Principle ([UNP]), namely that the
future will be like the past. Thus inductive inference includes a hidden premiss:
As have been followed by Bs.
The future will be like the past.
Therefore: This A will be followed by a B.
It’s clear that [UNP], if true, justiﬁes the inductive inference by providing the missing premiss.
Now for the big question. Is [UNP] true?
 Hume holds that there is no justiﬁcation for accepting [UNP]. Pay attention what he says!
He doesn’t say [UNP] isn’t true. He doesn’t even say he thinks it’s unlikely. He may even accept
it—and, in Enquiry 5, Hume will assert that in some sense we cannot help but accept it. Yet that
isn’t the same as having a justiﬁcation for accepting [UNP].
 To support the contention that there is no justiﬁcation for accepting [UNP] Hume reasons as follows:
[H1] The proposition that the future will be like the past is concerned with either (a) rela-
tions among ideas, or (b) matters of fact and existence.
[H2] There is no contradiction in supposing that the future will not be like the past.
[H3] Thus the claim that the future will be like the past is not based on relations among
[H4] Thus the claim that the future will be like the past is a matter of fact.
[H5] Thus the claim that the future will be