physics eventually gets down to assumption and acceptance

[sanctus]

It seems to me that all physics is only descriptive and in addition there is no explication which doesn't start from assumptions.
Take for example all the electrodynamics, it starts from the assumption that there is an electric charge and then makes up the model which describes the phenomenas. Maybe particle physics explains where charge comes from, but there are other assumptions.
That's why I wonder is it possible to have an explication not based on principles (energy conservation and so on)? Is physics only about the best model which describes experimental facts?

I think it has to be so, because there is no way to know that a theory is the ultimate one, but I think it's anyway very surprising: we never know real reason for any phenomena.

[Tormod]

You ask some very good questions, but you forget that physics is used to predict stuff. Yes, physics is very much about creating working models for the universe we live in. But it is also a way to say that "if this is true, then that should happen if I do so and so". Then I can experiment and try it out. The reality of the real world will change my perceptions of it if my experiment turns out differently than expected.

So I'd argue that physics is used to build models because that is how we as humans understand things - we create models and study them.

We know the reasons for some phenomenons - but I am willing to accept that the definition of "reason" boils down to pure semantics (ie, is there a willed reason, is the world mechanical, can we accept causality as a reason etc).

Say for example that we know why light bulbs emit photons - it is because freely moving electrons change their energy levels when they pass through the varbon filament in the light bulb. The direct cause could be something else completely (turning on the switch, or the sun going down, or waking up at night needing light) - but the difference between "reason" and "cause" is not so simple to dissect.

[gpdone]

"Is physics only about the best model which describes experimental facts?"



For some this statement is true, unfortunately. Those that use Physics in that way impress themselves and others who miss the point Tormod makes quite well. Prediction of something for the first time or better prediction of something already somewhat predictable is the only way to truly profit from modeling anything. If one cannot demonstrate the ability to do anything practical or do something better one can not demonstrate at least to me that an explanation no matter how well put or elegant is worthwhile.

Lee

[Tormod]

Well put, Lee.

Sanctus, I guess I didn't really take you up on the real issue here, either. You actually ask two things:

1) Is everything we know based on assumptions and acceptance of these assumtions?

2) Can we ever know anything about why things happen?

To me these are related but we should try to define an angle of attack because they can also be discussed separately.

[sanctus]

I agree 1) and 2) are directly related. They are directly related if everything we know is based on assumption and acceptance of those, this implies that the basis is always an assumption and therefore it implies 2). If the answer to 2) is no it directly implies yes to 1) and vice-versa. But I believe 2) is a more fundamental question.

My answer to 1) is that it is yes, unless in mathematics, as this is the only science in which if something has been proved 2000 years ago, it still is right today and for the rest of time (unless our model of how to count turns out to be wrong....)

My answer to 2) is No, by nature we humans are doomed to observe, therefore we can only make models, but we can never say a model is the ultimate one. Even if we have observed the whole universe in every detail, we never know if there couldn't be something that happens only every 10^56654 years, as we can't observe it (and if we could who tells as that there isn't even a "rarer" event).

[pgrmdave]

Within science there is one basic, fundamental assumption - given the same conditions, the same thing will happen. Assuming that you perform a repeatable experiment, you expect the same results the next time. If this isn't true, than none of science could predict anything, but since it seems to have held true for thousands of years, we assume that it is true.

[Tormod]

Originally Posted by sanctus

My answer to 2) is No, by nature we humans are doomed to observe, therefore we can only make models, but we can never say a model is the ultimate one. Even if we have observed the whole universe in every detail, we never know if there couldn't be something that happens only every 10^56654 years, as we can't observe it (and if we could who tells as that there isn't even a "rarer" event).

This is called the Incompleteness theorem and was proved by Goedel a long time ago.

[Tormod]

Originally Posted by pgrmdave

Within science there is one basic, fundamental assumption - given the same conditions, the same thing will happen. Assuming that you perform a repeatable experiment, you expect the same results the next time. If this isn't true, than none of science could predict anything, but since it seems to have held true for thousands of years, we assume that it is true.

Yes, no, well, this is only partially true. We can say that science is based on empirical data, i.e. since something happened one way yesterday, we will expect it to happen the same way tomorrow.

However, we cannot prove that it will. We can only prove the likelyhood that it will happen the same way (ie, the probability). You can never have a 100% probability for anything in the physical world.

So a prediction is only that, a forecast based on a certain probability.

It's like the coin toss thing - toss a coin and there's a 50% chance that it will turn up heads. But flip it 10 times and I'll bet you it doesn't show heads 5 times. However, the more your throw it, the closer to 50% you will get. That is why science needs to test, retest, retest, etc ad naseum...because we need to verify that our predictions are based on the correct probability.

[pgrmdave]

Is there a way to calculate that probability?

[lindagarrette]

Tormod. Maybe you guys are stull confusing chaos theory with randomness. Probability is based on randomness, which means we don't have control over, or know all the variables. If we did, there would be no need for probability and statistics. In every observable case, however, cause and effect is the rule. That means, with perfect knowledge, every event is predictable. In a closed system, such as, for example, computer programs, the outcome will always be the same given the identical input. Randomness may not even exist within quantum reality. Determinism wil stand until someone can prove an uncaused event has ever occured. (My challenge, several times in these forums.) Linda