The word "predict" is used in the main article because it seems to be commonly used in popperian literature and a suitable alternative has not yet come to my mind.
It is not intended to suggest that a theory must predict any particular set of happenings in the future. In fact, there are very good reasons to suppose that no scientific theory will ever be able to predict the future except over a very limited range of events and a limited period of time*.
What is meant by "predict" here is that a theory, to be worth anything much, must have something to say about potential observations outside the range of those used in arriving at the theory. These observations need not necessarily be of the future, or even made in the future. For example, a theory in climatology or geology may make a prediction that is tested by digging up rocks that were laid down 100 thousand or 100 million years before the theory was thought of. A theory could possibly even be tested by results that were obtained long ago and have lain forgotten in the library because there was no theoretical framework then available for their importance to be recognised. (This is probably unusual, however.)
The "prediction" and the resulting tests should, ideally, stretch the theory to the limit. That's the important thing.
* For example, the indeterminacy inherent in quantum mechanics. The strong dependence of non-linear systems on initial conditions, leading in many cases to chaotic behaviour, which is deterministic but in practice unpredictable. The vast amount of information that would be needed to track and predict the trajectories in time and space of all the particles and subsystems making up even a small part of the universe, making all but limited forecasting impractical.
Of course, you can argue that all these hide a fundamental determinacy and therefore predictability of the universe, currently hidden from us because of our ignorance. Of course you can, and maybe you'd be right. I have no theory of that, and I cannot predict the future.
On the other hand, these various factors may well be different faces of a universe that is either (a) undeterministic or (b) deterministic but theoretically unpredictable or (c) deterministic and theoretically predictable but practically unpredictable. Please reconstitute me when you have an answer to this problem (that will probably be a trivial technological feat then).
Copyright © 1999-2000 Richard Burnham