Theory vs Law?

January 15, 2005 10:59 AM

In science, what is the difference between a theory and a law? Why is something called a law rather than a theory?
posted by furiousxgeorge to Science & Nature (18 answers total)

hasn't this been discussed here before? unfortunately, i can't find it.

iirc there were two lines of thought. one being that a law is an observed correlation between measured variables, while a theory provides a possible explanation (in causal terms) for that relationship. the other being that they were sometimes interchangeable, and reflected common usage/confidence/age.
posted by andrew cooke at 11:06 AM on January 15, 2005

There's a huge about of confusion about scientific theory. For instance, creationists often "remind" us that evolution is "just a theory." They mean "theory" in the standard sense (which is similar to "a guess"). In science, a theory is based on evidence, not guesswork.

Here's a web page that has pretty clear definitions of law, theory and (though you didn't ask) hypothesis.
posted by grumblebee at 11:14 AM on January 15, 2005

As I've come to understand it, 'law' is an outdated kind of nomenclature, although still used for some very well understood things, like describing not how gravity works, but how to calculate its effects. Describing how gravity works would take the form of a theory - a very complex theory - but the Law of Gravity is a simple equation.

Nowadays, 'law' has pretty much fallen out of usage. If Maxwell discovered his 'laws' of thermodynamics today, they'd be theories.
posted by Dipsomaniac at 11:42 AM on January 15, 2005

Thanks, cleared it up for me.
posted by furiousxgeorge at 11:50 AM on January 15, 2005

I was taught that a law is "a terse statement of experience." The thing separating it from a theory is its terseness. Compare the ideal gas law (PV=nRT) with the general theory of relativity.
posted by grouse at 12:20 PM on January 15, 2005

Thought I'd jump in even though the asker is satisfied.

I don't think the notion is as outdated as discussed above. One can, for instance, discover a law of nature. One cannot in the same way discover a theory. Scientists (and philosophers!) form hypotheses which, when well-confirmed, become theories. Theories are attempts to describe laws of nature.

Take Newton's theory of gravity. It is, of course, a theory that tried to describe the law of gravity. Unfortunately, it fails in several cases. It describes only part of the law. Einstein's theory of gravity is a much better theory. It successfully covers many more phenomena than Newton's theory does. It describes the law (whatever that is!) much better.

Since most confirmation theory is about predicting what will happen, it becomes much, much harder to talk about confirmation when talking about the past. This, in my opinion, is why the evolution/creation debate is so persistently about theory. The dataset is crazy, based on what people dig up, rather than what will happen. One cannot prepare experimental conditions as the experiments have already been done and the results very poorly preserved. Thus, it is much harder to argue that anyone is successfully describing a law. The normal notion of confirmation is on holiday.

Of course, we can do experiments to show what is going on now and what will happen with gene X in environment Y, and we should simply infer that since it can be so well demonstrated, it has always been going on. But many people resist this extension of the law to the distant past. In some ways it is an argument about how much time there has been, not what the law is.
posted by ontic at 12:26 PM on January 15, 2005

My physics teacher explained the difference by saying that something becomes a law when it has virtually unanimous confidence that it is correct in every way.

A theory is still a very strong statement, but may lack confidence from scientists in some areas, and there may be other explanations that work nearly as well.

"Intelligent Design" is a myth because all of it is wholly unprovable. "Evolution" is a theory because there may be some cases which don't fit well to the theory [I wouldn't doubt there are, nature is too complex to *always* fit into a single scientific theory] [I would doubt any myth based on a God influencing things, though]

Something like the law of gravity is a law because, generally, there had been no conventional experiments for hundreds of years showing that a large mass does *not* attract other masses. Now, with all the knowledge we have, there may be cases in which we can show this doesn't work, but they are always artificial.

Well, that's my best explanation.
posted by shepd at 2:36 PM on January 15, 2005

shepd: I don't think that is correct. Evolution is a theory and it will always be a scientific theory instead of a physical law because it is too complex to be summed up in a single equation.
posted by grouse at 2:52 PM on January 15, 2005

To expand on what grouse said: Evolution is a theory because there is still debate over the details. Really, they're pretty picky details, but Dawkins and Gould used to go at it pretty hot and heavy on them. Neither one would have challenged the general idea of evolution by natural selection.

That said, as I sit here, a lot of the things that I still hear referred to as "laws" are either very terse foundational statements (like Newton's "laws") or can be expressed as constants or reliable equations (like Boyle's law).

Plus, "Clarke's Third Law" sounds a lot funnier than "Clarke's Third Conjecture".
posted by lodurr at 4:02 PM on January 15, 2005

Evolution is a theory because there is still debate over the details.

There is still debate over the details in all of science, so this does nothing to elucidate why one thing is a "theory" and another a "law."

A better answer would be to follow Karl Popper and say that in science we have nothing but theories - even the so-called "laws" are theories, and the only reason why certain concepts are known as "laws" is because there's an easy formula to state them. Everything in science is necessarily provisional - not just because of human fallibility, but also because induction is logically unjustified, and there's really no rational reason why, say, the "law" of gravitation shouldn't suddenly cease to be valid tomorrow - so strictly speaking all reference to "law" in science is an abuse of language.
posted by Goedel at 5:17 PM on January 15, 2005

I think some combination of what's suggested fits best.

For one, scientific "laws" exist mostly within physics and, to a degree, chemistry, because they are not only agreed upon, they are simple statements of fact -- whereas in sciences which deal more directly with our experience, this becomes much more difficult, as there are too many variables.

Evolution by natural selection is generally agreed upon, but that in and of itself isn't enough to make a law, and some of the exact mechanisms are still unknown/disputed, which is why scientists talk about "evolutionary theory"
posted by dagnyscott at 5:31 PM on January 15, 2005

you know, i never understood why we didn't just call it the "supertheory" of evolution just to make the creationists' jobs harder. i know, it's wrong, but i think making it harder for the IDers/creationists to mess with kids' heads is worth it.

i'm not sure "law" is an abuse of language, per se, as they're really just condensed statements of fact - loads and loads of observations get condensed into laws. there's no implication of "why" in a law.
posted by sachinag at 5:42 PM on January 15, 2005

Dipsomaniac got it. Basically, laws of science are equations. For example, F = MA is a law because when numerical numbers are plugged in, it is a mathematical law that F must equal MA in the math sense of things. However, how it is described is still a Theory, because, well, it doesn't involve numbers. And to be more exact, a Theory is basically a framework of ideas that best fits any phenomenon that occurs. IE, Theory of Gravity is our best description at how/why things fall to the ground. The equation (I forget it off the top of my head) is the Law of Gravity because it involves number as I described above.
I know it's been mentioned, but I'll repeat that we won't see the Law of Evolution because you can't exactly reduce evolution to mathematical terms. Well, maybe one day they will be able to by describing possible changes over time such as population genetics, but I don't see that happening anytime soon,
posted by jmd82 at 10:13 PM on January 15, 2005

I think it would be a good idea to start calling it the "Law of Evolution" in order to take some wind out of the sails of this silly creationist "it's only a theory" argument.
posted by sour cream at 1:45 AM on January 16, 2005

I agree with sour cream and sachinag. From now on, incorrect as it may be, I'm going to refer to the "Supertheory of Evolution" or the "Law of Evolution".
posted by Bugbread at 6:03 AM on January 16, 2005

bugbread, I doubt it will help any argument to lie. Eventually, the lie will be found out, and then you (and the rest of us evolutionists) will just lose credibility.
posted by grumblebee at 8:58 AM on January 16, 2005

Well, it isn't really a lie, just uncommon terminology, but point taken. I suppose the best thing to do is just say, "The theory of evolution, or, as perhaps it would be called in layman's English, the law of evolution". Then people will inevitably ask about the difference between the layman's English and scientific use of the term, at which point pontification can begin.
posted by Bugbread at 9:04 AM on January 16, 2005

bugbread, I doubt it will help any argument to lie. Eventually, the lie will be found out, and then you (and the rest of us evolutionists) will just lose credibility.

It isn't in the least a "lie"; F=ma is no more a "law" than is the "theory" of evolution. All of science is about theories, and the theory of natural selection is actually more in keeping with what we currently know to be true than are Newton's "laws" of gravitation, which are mere approximations to the field equations of general relativity.
posted by Goedel at 4:16 PM on January 16, 2005