Once more on theory and law in science

It is well known that the popular usage of “theory” as meaning an unproven idea differs from how scientists use the word. And that allows creationists to wreak all sorts of “only a theory” mischief.

Defenders of science, however, often over-react to this charge by asserting that, in science, “theory” means a well-tested and well-evidenced account. For example the US National Academy of Sciences makes a well-meaning but wrong statement, defining:

Theory: In science, a well-substantiated explanation of some aspect of the natural world that can incorporate facts, laws, inferences, and tested hypotheses.

In fact, scientists use the word to mean merely “explanation”, or perhaps “set of explanatory ideas”, and it carries little connotation either way as to whether that explanation is tested or proven. For example “string theory” is highly speculative and unproven. Similarly, we still talk about Newton’s theory of gravity even though we know that it is only approximately correct and that it has been superseded by Einstein’s theories.

When I point out such usages I’m sometimes told that these are instances of scientists being sloppy and that “theory” should only be used for well-tested explanations, with speculative explanations being called “hypotheses”. The trouble is, scientists in general don’t abide by such rules and it would be a heck of a job to persuade them all to do so. Further, the word “hypothesis” is generally used for a single-sentence or single-equation idea, whereas a “theory” or “explanation” would typically be a whole set of interlocking ideas.

Darwin’s theory of evolution is thus a “theory” because it is an explanation that takes more than one sentence to state (in its most minimal form it takes three sentences). Similarly, “string theory” cannot be a hypothesis because it involves more than one equation.

Nor does a “hypothesis” become a “theory” if it becomes proven, rather it would be more likely to be called a “law”. That’s because, like “hypothesis”, the term “law” is generally used for an idea that can be stated in one sentence or one equation. Thus a “theory” could contain several “laws”. For example, Boyle’s Law, Avogadro’s Law and the Ideal Gas Law are all part of the kinetic theory of gases.

And for that reason, a “theory” won’t ever be “promoted” to being a “law” (despite the misconceptions of those less familiar with science). No matter how proven it gets (and Darwinian evolution is about as proven as it is possible to be) a “theory” will always be a “theory” because an “explanation” will always be an “explanation”.

And, there again, the word “law” does not connote being fully proven and accurate. The Ideal Gas Law is very useful, but is only an approximation to real-world behaviour. Similarly, Newton’s Law of Gravity is useful and approximately correct in many circumstances, but is not fully accurate. The Titius–Bode Law, to pick another example, is widely regarded as highly dubious and as being only approximately true by accident; it is still called a “law”.

The lesson of all of this is that scientists usually do not connote how well-established or well-evidenced an idea is by which of these terms they use. Even if you think they should — as the textbooks sometimes tell you — they don’t. And you’re really not going to persuade them all to start. So, the best idea, when explaining to non-scientists, is to say that one cannot interpret any of these terms as indicating any level of proof or lack of proof.

By the way, mathematicians do pay more attention to such terms, and care more about the difference between a “conjecture” or a “hypothesis” and a “theorem” (which is what a conjecture becomes when proven). But then mathematicians are well known for being much more formal and pedantic in their approach than is typical of scientists. Having said that, though, note that Fermat’s Last Theorem was called that long before it was proven, and many conjectures, such as the Poincaré conjecture, are still called that after being proven, and indeed some conjectures are still called conjectures after being dis-proven. .

So, again, it really is a bad idea to over-interpret any of these words. If you want to know how well established and accepted a scientific idea is, ask about the evidence for it, don’t ask about the label attached to the idea.

I was prompted to write all the above by stumbling across yet another “only a theory” post; a post that is notable mostly for being about as wrong as one can get in the interpretation of such terms. The article is by a Sam Gerrans, a writer for the RT (“Russia Today”) network. Gerrans tells us that he believes: “that the Qur’an is from God, complete and protected”, which doesn’t suggest a science-friendly way of thinking.

Despite having no qualifications in science, Gerrans is convinced that he understands how science works better than Richard Dawkins, who he accuses of supporting: “Scientism: Religion of the new heresy-hunters”. The “heresy” here is the now-notorious tweet by a head teacher of a UK school, which said: “Evolution is not a fact. That’s why it’s called a theory! There’s more evidence that the Bible is true”. Even after reading an account by Dawkins, which clearly stated what science means by “theory”, Gerrans still failed to get it.

Gerrans writes:

“Evolution is, objectively, a theory. It may be treated by the scientific establishment as a fact, but that does not make it one.”

Oops, fail!

“The realm of theory is where the mind goes for an after-dinner glass of port and cigar and stretches out in a leather armchair in front of the fire and blows a few what-if-scenario smoke-rings around the sitting-room.”

An imaginative metaphor. Nice writing. Shame it is wrong!

“[Scientific] conclusions are based on observation and experimentation. … This is what distinguishes law from theory.”

When I started writing this article I had intended deconstructing all of this, but really, it writes itself.

“The problem comes when those who presume to speak for science forget the distinction between theory and law, and simply attack those who have not forgotten it.”

I mean, anyone who is reading me is going to realise how blatantly wrong this is just as well as I do.

“If you have proof: bring it. If you do not: acknowledge openly that you have theory — perhaps a well-honed, much loved theory, but a theory nonetheless.”

Which is why, instead, I ended up writing about how replies to this sort of stuff are often also somewhat wrong about the term “theory”.

“But for Dawkins to claim in the absence of proof that he has a right to his theory greater than Christina Wilkinson — a teacher in a religious school — has to hers, places him rather than her in the role of ideologue.”

Epic fail! But it does raise the question. If so many people are interpreting “theory of evolution” as implying that there is no evidence for it, should we change the term, however hard that would be? Or is it the case that the only people who still get this wrong are the sort who would think “that the Qur’an is from God, complete and protected”, and so are lost causes anyhow?

6 thoughts on “Once more on theory and law in science

  1. Ron Murphy

    I think the same flexibility applies to the term ‘hypothesis’. It can stand in for and in wait of a theory, if it generalises what would become a theory.

    The one that bugs me is ‘law’ – when used as ‘a law of the universe’, as if this means that it is the universe that is determining this law, or the law itself is an aspect of the universe, or the ‘law’ is written in the stars sort of thing. A law is a human model of some aspect of the universe. This becomes clear when considering Newton’s laws. They may have seemed like aspects of the actual universe, but when it later appears they are approximations, then it becomes obvious that like all models they are not the actual thing but a representation of it.

    Another perspective is one of simulation. Simulations are models too, but the notion of a dynamic simulation brings that home. Simulations can be ever more precise and accurate, but they never capture everything about a system. The only ‘complete, precise, accurate simulation’ of a system is the system itself. Anything else is a model.

    What’s more, a component of a system can never ‘completely’, or even closely, model the system that contains it. We cannot know everything about he universe – even if it were entirely classically deterministic, because we are components of it. And sub-system that attempts to model the system containing it must put itself into that model. It becomes recursive and he limiting factor is a lack of capacity.

    Our laws will never fully describe the universe.

    To get back to your main point, everything we say about the universe (or beyond) is a model of sorts, whether it’s the best scientific theories, or theology. The only difference is the extent to which they map onto the world as far as we can tell.

    The problem with religion (putting aside the quite ridiculous use of faith as a methodology) is that it’s claims are simply not backed up by evidence, and often defy counter evidence. And even the most elementary theism, which all religions must presuppose, is nothing more than one of countless speculative ontological claims with no reason to support it than any other. Any monotheistic speculative presupposition is no better than any other. I never here theologians speculating about hyper-gods in a hyper-natural realm that create their supernatural gods that create us. It’s the utter flakiness of the ideas that religion stands on that should be attacked.

    1. taurisstar69

      After reading your explanations about the limitations of modelling a system that contains the system i really could see it better in my head. This is all information that over time, I have accumulated but I never arranged it all like i have now. Thanks.
      But I also have to make the same correction when you used the word religion to encompass all of the avenues involving a higher level being, there exist many different forms of being a created being then just religion.

    2. taurisstar69

      After walking away I realized I mistakenly used the word created being as a reference to the differences in religion. The more accurate way to say what I was meant is: There are more ways to having a purposeful creator than Just religion.
      Also I use science because I am a scientist when looking at the world and we most assuredly do have a creator. This creator built Earth as well as created life. This creator abolishes darkness, nurtures life and if you look directly at his body for too long you will go blind. Which all are descriptions that match God’s descriptions of himself in the bible. Cool aye =)

  2. Disagreeable Me (@Disagreeable_I)

    Nice post, Coel. I’ve often tried to make the same point myself. It usually does not go down well. The mantra of a theory being a tested hypothesis is becoming too ingrained.

    One reason I like the kind of interpretation of “theory” you advocate is that it is continuous with the sense of the term in mathematics.

    Which is often highly appropriate, given the nature of String Theory, for instance.

  3. Peter Lyons

    And the prize goes to those on the side of …….
    Reason. Nobody here claiming absolute truth, whichever definition you use for law, theory or hypothesis. Just an honest search to find out how it all flipping works. Can bible thumpers say the same? (FYI the last question was rhetorical .)

  4. Henry Fitzgerald

    I would submit that the sleight-of-hand pulled by people who say “only a theory” doesn’t really have to do with any subtleties in the word “theory”. The culprit here is the word “only”.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s