Most people engage in what philosophers call “inductive” reasoning. People make accurate predictions about the future on the basis of conjunctions they have observed in the past. For example, I assume the sun will rise tomorrow because in the past I have observed it rise. Scientists, in particular, rely on extrapolations of this sort in the construction of their theories and in the detection of the laws of nature.

Implicit in such reasoning is the assumption that nature is uniform or regular, one can legitimately contend that events in the past will resemble events that will occur in the future. One interesting problem consists of how one knows that the future will resemble the past in this way. What is the basis for this view? In this post Machine Philosophy discusses some of these issues.

Conjunction and Inductive Argument

Even those who accept the regularity view are compelled, sometimes in the statement of that view itself, to assume that that regularity view itself is false.

Events have causes in the sense of regular and special antecedents.

Do you intend to apply that to the future as well?

Yes.

But you have not experienced the future.

No.

Then you must be using an argument. You’re saying that b will continued to follow a in the future because b has followed a in the past.

But unless a is connected with b by something more than mere proximity of place or time, there’s no ground for that argument.

Without some other connection, there is no reason for saying that past uniform sequence must be followed by future uniform sequence.

When we’ve argued from past uniformities to their continuation in the future, our expectations have been verified.

But the uniform sequence of verification of your expectation is just another sequence. It has no special privileges when we’re asking why one should argue from any past sequence to future ones.

It’s a matter of probability. Two or more events frequently occuring together is more likely to be maintained than broken.

You’re either repeating the old assumption whose basis is in question, that the past is a guide to the future, or else it’s is false.

If a and b are unconnected, there’s no more reason to expect them to continue together than there is to expect an unloaded penny to continue turning up heads because it has just done so ten times in a row.

We ‘postulate’ the uniformity of nature, namely that the same cause is always followed by the same effect.

But what’s the status of that postulate itself?

It can’t be an arbitrary assumption, because it comes from experience.

But it’s not a conclusion derived from experience, because the argument would be circular. If we did not assume ‘same cause, same effect’, we wouldn’t argue that the same cause in the past would be followed by the same effect.

What part does this assumption play in the argument? It is the principle of the inference, just as the principle of syllogism is the rule involved in any syllogistic reasoning.

Now if the principle of an argument or inference can be without necessity, then the ‘inference’ to which it applies is not really an inference.

It’s not even cogent.

The principle of uniformity does not have the necessity of geometric demonstration, since the terms viewed as causes and effects themselves remain vague. But going from past sequences to future ones is itself an argument, and the principle of the argument, like the principle of inferences generally, must be more than a chance proximity of symbols or characters.

Therefore, the connections between causes and effects being used to predict their future sequences are always assumed to be intrinsic.

Tags:   Conjunction · Induction · Philosophy of Science · Regularity View · Uniformity of NatureNo Comments