Scientific theories and empiricism

[devilsadvocate]

The thread "When does scientific theory stop being a theory" got me thinking about empiricism and scientific realism. Especially the post about how theories explain strings of facts (observations). I'm thinking amongst other things Hume's challenge for induction. Hume wasn't arguing that induction is only so reliable, he was arguing that past events are absolutely no indication at all for future events. After all we merely observe strings of facts, not causation itself. In this regard, are we like Russell's chicken? Every day the farmer comes with a bucket of seeds. The chicken seems justified enough in her belief supper follows farmer, but as we know from grander scheme of things, eventually the farmer will come with an axe instead of a bucket.


So I pose this as a form of challenge, if you will: How justified are scientific theories on observations alone? Try to stay on observable evidence alone, careful not to assume any metaphysics.

 

[CosmicJoghurt]

I fail to see how scientific theories can be safely derived from inductive reasoning. Deductive reasoning gives us a real, non-assumptial understanding of the facts, and therefore I can't seem to find anything wrong with it being used as solid evidence... But not inductive reasoning, oh no. Again, that's my layman view.

 

[Case]

I support this fruitful conversation with the following artist's impression thereof.

 

[Your Funny Uncle]

The moment gravity fails and we all fly off the Earth into space we'll know that Hulme was right. Until such time I reckon the scientific method seems pretty reliable.

 

[Squawk]

You're farmer/chicken analogy is subject to a flaw which science actually works to counter. Let me expand on this.

A theory is only as good as the assumptions on which it is based. There are certain assumptions that we must make in order to use any form of reasoning. For one, the notion that reality exists, and two, that we can reason about reality. We should try to eliminate any other assumption as far as possible.

Beyond that we begin to make inferences. The chicken cannot make a sound inference of the farmer/food situation because it's treating the information in a vacuum. It has no knowledge of supermarkets, or even of cognition, of the farmers true purpose.

From this we see that past events are a predictor of future events if we can establish from first principles a cause and effect relationship, even if that relationship is probabilistic rather than strictly deterministic. If a given assumption (reality exists) is actually false (and it's possible), then all bets are off. Easiest example here would be if we actually live in a matrix and an alien being (the designer of the matrix) is capable of changing the fundamental constants that define our reality).

In short, a theory is only as good as the assumptions that underpin in. The more assumptions you can remove, the stronger the theory.

Might be worth noting that there is a branch of apologetics call presuppositionalism which asserts that the only justification we have for reason is God. It's a failed position since it assumes its conclusion (reason), then attempts to justify it, but you try getting a presuppositionalist to realise that. Head, wall, pain.

 

[devilsadvocate]

There are certain assumptions that we must make in order to use any form of reasoning. For one, the notion that reality exists, and two, that we can reason about reality. We should try to eliminate any other assumption as far as possible.

I constructed the argument from the viewpoint empiricism is the accepted epistemology. I believe it is possible to have sound reasons, deduced from observations (which incidentally count as, at least when carefully performed, justified evidence in this epistemology), that reality exists. This is not on the table (at least doesn't have to be) nor is the justification for empiricism in general. You can think of the question in this way: IF we adopt empiricism (and we agree reality exists to boot), WHAT can we justify to believe in. All justification has to be deduced from observations and observations alone.

Might be worth noting that there is a branch of apologetics call presuppositionalism which asserts that the only justification we have for reason is God. It's a failed position since it assumes its conclusion (reason), then attempts to justify it, but you try getting a presuppositionalist to realise that. Head, wall, pain.

Isn't same kind of logic given as an answer to the problem of induction most often? In so many variations the attempted justification for induction is of the general form "Induction has worked so great in the past, so of course it's silly not to trust it". That is of course just circular logic. Good luck trying get people to recognize this as well.

Before you're ready to throw me in the loony bin, I want to make what I'm claiming here clear: I am NOT saying we shouldn't trust induction, what I am asking is can we have any rational justification for induction, given only strict empiricism as epistemology.

 

[Squawk]

Before you're ready to throw me in the loony bin, I want to make what I'm claiming here clear: I am NOT saying we shouldn't trust induction, what I am asking is can we have any rational justification for induction, given only strict empiricism as epistemology.

No, we can't. That's exactly what I intended to spell out, and I'm pretty sure you already know the answer is no given what you posted. I won't throw you in the loony bit at all for it, because it's a position I agree with.

However, in addition to arguing that it cannot be justified, I would then argue that in order to even start to comprehend the world you must accept it as being true unconditionally (ie, it's an axiom), because you cannot even trust the thoughts you have if you don't accept it axiomatically.

 

[ImprobableJoe]

Too many big words... I assume that someone in this thread is LYING TO ME!!!! :lol:

 

[Andiferous]

<3 Joe.

I still prefer deduction.

 

[impiku]

Hume wasn't arguing that induction is only so reliable, he was arguing that past events are absolutely no indication at all for future events. After all we merely observe strings of facts, not causation itself.

Reality is regular but to infer logical necessity from induction is absurd. It is a non-sequitur. I think I posted something like this on this forum somewhere, probably free will vs determinism thread. I said that the matter simply transcends the scope of mere empiricism. It's a rudimentary mistake to simply enumerate "causal" events based on observation and argue for causal determinism, it just doesn't cut it.

How justified are scientific theories on observations alone?

Whether it is epistemically justified or not is beyond me. I would have to read/learn more to have an opinion but it is justified pragmatically, and it is this reason why I value science.