What is a natural thing?

[BroBrown]

What do scientists mean when they say that an explanation or entity must be natural in order for it to be scientific?

[Mokele]

As opposed to super-natural, mostly for practical reasons. If we hypothesize a natural cause for something, we can test that hypothesis by manipulating said cause (or by looking for 'natural experiments' where the cause varies naturally). If we hypothesize a supernatural cause for something, since that cause cannot be detected, manipulated, or quantified, we cannot test the hypothesis, ergo it's worthless as it cannot advance our knowledge.

 

That doesn't necessarily mean we're limiting ourselves forever, only limiting ourselves right now. 400 years ago, if you claimed that there was a strange energy given off by certain materials that could cause cancer, it would be untestable. But once we discovered radiation and had ways to quantify it and control it, it became testable.

[BroBrown]

As opposed to super-natural, mostly for practical reasons. If we hypothesize a natural cause for something, we can test that hypothesis by manipulating said cause (or by looking for 'natural experiments' where the cause varies naturally). If we hypothesize a supernatural cause for something, since that cause cannot be detected, manipulated, or quantified, we cannot test the hypothesis, ergo it's worthless as it cannot advance our knowledge.

 

Can anyone manipulate the cause of the universe's expansion?

 

Using "natural experiment" or "cause varying naturally" in the definition of the word "natural" is... I forget the name of that fallacy.

[Mokele]

Can anyone manipulate the cause of the universe's expansion?

 

No, but we can observe consequences, which would be different for different causes.

 

Using "natural experiment" or "cause varying naturally" in the definition of the word "natural" is... I forget the name of that fallacy.

 

Those aren't part of the definition, they're my explanation of how science works when manipulative experiments aren't possible.

 

For instance, we cannot manipulate the temperature in a 200 square mile region precisely to test the theory that species get larger in warm climates. But we *can* look at an exceptionally wide-ranging species like the tiger salamander and see if there's a correlation between the average annual temperature and salamander size at various locations with different temperatures.

 

I'm sure there's a thousand overly-elaborate philosophical papers on the subject, but in plain terms, we have to be able to detect something, to measure it, etc. (or to do so for a closely linked proxy) in order for it to be useful in science.

 

If we cannot detect it, cannot measure it, we cannot test the hypothesis.

[mooeypoo]

If we cannot detect it, cannot measure it, we cannot test the hypothesis.

Actually, this should be expanded to say that if we cannot detect it, cannot measure it directly or indirectly by its effects on other phenomena, we cannot test the hypothesis.

 

We can't truly detect all types of planets that orbit distant suns, for instance, and we can't directly measure them, but we CAN see the effects their gravity makes on the star (wobbling, etc) and its light. So, our technology isn't QUITE up to par with detecting the actual planet (yet!) but the fact we see its effects consistently also allows us to indirectly measure it and test the hypothesis.

 

Same goes for dark matter, for example.