Criteria of validity in empirical sciences

[Speakpigeon]

The unique criterion of the validity of a theory in empirical sciences, say, physics, is that the theory should produce results in line with our observation of the physical world, ranging from our direct visual observation of nature to experiments involving possibly, and increasingly so, complex installations, machines, apparatuses and sometimes a large team of scientists working months to agree on an interpretation of the results.

Do you agree with this presentation of this criterion, including with the suggestion that it is not only the main but that it is also the only criterion admissible in empirical sciences like physics.

If you think that there are other criteria necessary to assessing the validity of a theory in empirical sciences, what are they?
EB

 

[Lion IRC]

Should produce deterministic, identically repeatable results on demand.
None of this spooky random spontaneity of the gaps woo.

If it happens in the natural world we should be able to see it and replicate it like clockwork.

 

[steve_bank]

In physics deterministic does not mean predetermided or presestined.

Calculating distance traveled in a car using time and speed is a deterministic model. It is accurate within measurement uncertainties. No measurement is exact.

Imaging sitting in a restaurant from 10 to 1 o'clock, around lunchtime. Keep a record of time and how many people enter and leave each minute. Over a period of days you can build a statistical model. With enough data you assign a probability of someone entering or leaving the restaurant in any minute but you can never predict an exact time. It is called a stochastic process.

At the quantum level exactly when uranium will emit a particle, but we can predict a probability from observation. Whether there are hidden variables or unknown underlying deterministic process is not knowable by current science.

There are different views as where observed quantum probabilities are a measurement issue or whiter it is an aspect of nature.

Regardless of which it is for us humans some things appear mathematical deterministic as described above and some things seem probabilistic. If you waft to argue that I will sell you a system for winning at craps in Vegas.

 

[fromderinside]

steve_bank makes points Speakpigeon. Do you demand that space and time must be included for measurement to be empirical? After all they are aspects of our observation of the physical world. Would it not be sufficient to say "I observed this before that here then there" for one to say observation is empirically deterministic.

 

[Speakpigeon]

steve_bank makes points Speakpigeon.

About what, though?!

According to the evidence available, he was replying to Lion IRC. And I don't see any relation to the OP in his post. He doesn't even quote it.

Do you demand that space and time must be included for measurement to be empirical?

What does that mean to begin with?!
The OP is about the fact that empirical sciences should produce results in line with our observation of the physical world, ranging from our direct visual observation of nature to experiments involving possibly, and increasingly so, complex installations, machines, apparatuses and sometimes a large team of scientists working months to agree on an interpretation of the results.
Can you give specific examples of this scientific theory of space and time you seem to be talking about?

After all they are aspects of our observation of the physical world. Would it not be sufficient to say "I observed this before that here then there" for one to say observation is empirically deterministic.

And? Are you talking about "our direct visual observation of nature"?
EB

 

[Lion IRC]

In physics deterministic does not mean predetermined or predestined.

How do you get repeatability without deterministic certainty?
Is there an unseen ghost in the machine.

Measurement error is irrelevant because it's a known tolerance and nobody is trying to achieve repeatable measurement error. There's nothing supernatural about +/- 0.0001

 

[steve_bank]

In physics deterministic does not mean predetermined or predestined.

How do you get repeatability without deterministic certainty?
Is there an unseen ghost in the machine.

Measurement error is irrelevant because it's a known tolerance and nobody is trying to achieve repeatable measurement error. There's nothing supernatural about +/- 0.0001

Tail is wagging the fog. If measurements fall outside the error bounds the measurement is not repeatable. If within it is considered repeatable within stated bounds. If measurements fall outside predicted error then the experiment and error analysis is reviewd.

Some years back the Italians ran an experiment to measure the speed of a particle oven a long distance open air range. It was supposed to be C. I forget it was either measured greater or less than C. It was big science news, Tuned out it was an instrument calibration error.

In my very first college physics lab we measured a block of metal with a screw micrometer. After multiple measurements it was clear it was not exactly repeatable, the point of the experiment.

No experiment is ever exact and never exactly repeatable. Error analysis is performed to determine the window of uncertainty about a true value. In the case of a mechanical or electronic measurement with random variations you take multiple measures' and calculate the average. For random uncertainties the arithmetic mean is the best estimator of the true value. In any system there is always a form of noise, sorting that interferes with determining the true value.

When numerical results of an experiment are reported as a number with a bound of uncertainty. 10.1 +- .001. From high school science the number 1 has no meaning. 1.0 means 1.0 +- 0.0x


I spent 30 years applying science as an engineer.

Steve's the name and measurements are my game...so to speak.


My point was uncertainty is part of reality regardless if it is fact of nature or a measurement problem on our part. The uncertainty Principle so far is a demonstrated fact in many applications.