Defining Science

Over at Built on Facts, Matt Springer is easing his way back into blogging by asking “What is Science?”. He offers a simple one-sentence definition:

Science is the testing of ideas.

That’s all. Every technicality I can think of is avoided so long as the person doing the science is honest. Create fair and objective tests, try not to fool yourself or anyone else, don’t be wedded to your hypothesis, basic things like that. Be dishonest and I doubt there’s a definition in the world that some sufficiently clever pseudoscientist can’t wriggle out of. Test your ideas and be honest about it. That’s about it.

Shockingly, he immediately gets a bunch of comments pointing out some loophole or another that would allow some form of pseudo-science to squeak through. Nobody ever could’ve predicted that.

The problem here is that there are really two ways to come at the question of defining science, and Matt’s simple definition leans strongly toward one of them. Thinking about the different approaches, it also occurs to me that there’s some resonance between what’s going on here, an endless debate in science fiction circles, and the construction of Internet communities.

First things first, though: the problem Matt’s running up against is that there are two ways to come at the question of what science is, in that different people have different goals in mind when they ask the question. One group is trying to define science so as to exclude things that they find objectionable. The other is trying to define the common elements of science, so as to include all the things that they like.

Matt’s definition leans strongly toward the latter. He’s trying to find a way to bring theoretical, experimental, and observational sciences all together under a single definitional umbrella. He doesn’t phrase it that way, but if you look at what he says, particularly that last paragraph, it seems clear that that’s the spirit of the definition. He wants to bring together all of the different scientific approaches that work, and basically celebrate their effectiveness.

The people objecting to his definition are coming at the question with the other goal in mind: they’re interested in defining science not so much to celebrate to good as to cast out the bad. You can see it in the tone of most of the comments: they’re primarily interested in finding and eliminating arguments that quacks, astrologers, or cdesign proponentists could use to fit in under Matt’s definition of science.

This reminds me a lot of the endless arguments in science fiction fandom about where to draw the line between “science fiction” and “fantasy.” These rage without end in various corners of the Internet, and inevitably devolve into detailed arguing about fiddly edge cases– are Anne McCaffrey’s Pern books fantasy because they have telepathic dragons, or science fiction because the dragons are supposed to be genetically engineered by space travelers? Is Star Wars science fiction because they have space ships and laser guns, or fantasy because of the wibbling about the Force? And so on.

There’s a comment by somebody famous in the field, who I will confidently assert to be Ursula LeGuin, knowing that Jonathan Vos Post will correct me within nanoseconds if I’m wrong, pointing out what’s wrong, which is really the same as the difference between the inclusive and exclusive definitions of science. The problem with arguing about where to draw the line between science fiction and fantasy is that you end up spending all your time arguing about a handful of books way out in the borderlands, and stop thinking about the vast bulk of the individual categories. In many ways, it’s more productive and illuminating not to argue about the boundary cases, but to think about the common elements shared by the vastly larger number of works in the center of each of the fields.

I think there’s something to be said for that in terms of defining science as well. When I think about the question I tend to use a more operational formulation. My definition of science is a process with four steps: 1) identify an interesting phenomenon in the world, 2) come up with a model of why that thing might be that way, 3) test your model through observation or experiment, and 4) tell everyone you know the results. It’s essentially the same thing Matt’s saying, only with numbered steps (perhaps not surprising, as we’re both physicists). My interest in defining science lies primarily in trying to find the elements that make it the most effective means of producing knowledge ever devised, and the cornerstone of modern civilization

Other people come at this from a much different angle. Their interest is not so much in figuring out what works, but in sweeping away what doesn’t work. There tends to be a political flavor to most of these definitions, as the impulse to go in this direction often comes from specific grievances with various harmful and unscientific approaches and world views. These end up adding a lot more detail to the definition, trying to close off gaps in the boundary through which non-science might approach science, and thus gain some undeserved credibility.

In the end, I think we do need both approaches, because there are people out there who try to claim the status of “science” in bad faith. We need ways to stop them from both abusing that status to harm others, and from reflecting poorly on the mass of people who do good science.

Whenever I watch these arguments, though, I end up being reminded of one of the many very smart things Teresa Nielsen Hayden has said about the moderation of Internet communities (which she does professionally for Boing Boing, and recreationally for Making Light). Paraphrasing a bit, she said that the rules for moderation should be as simple as possible, because the more complicated you make them, the more the system starts to look like a game. And when the system looks like a game, people will try to play it like one, and you’ll spend all your time adding exceptions and closing loopholes.

The arguments that spring up in Matt’s comments have some of the same flavor to me. The whole business of adding qualifiers and finding exceptions, and adding qualifiers to the qualifiers to close out the exceptions, and so ad infinitum just feels like people gaming an increasingly arbitrary system of rules that become more and more divorced from what’s really important. And that way lies madness.

Which, I guess, is why I’m neither a philosopher nor a lawyer.


  1. #1 BiophysicsMonkey
    January 13, 2009

    “The problem with arguing about where to draw the line between science fiction and fantasy is that you end up spending all your time arguing about a handful of books way out in the borderlands, and stop thinking about the vast bulk of the individual categories. In many ways, it’s more productive and illuminating not to argue about the boundary cases, but to think about the common elements shared by the vastly larger number of works in the center of each of the fields.”

    This seems somehow related to the old maxim that “hard cases make bad law”.

  2. #2 Memoirs
    January 13, 2009

    Some people say that only physics is science. According to Lord Rutherford “there’s physics and then there’s stamp collecting”

    Unless problems can be mathematically defined it’s just words.

    For more discussion based on a lifetime of work see the synopsis in:

  3. #4 Mary Aileen
    January 13, 2009

    I’m not JVP, but I think that was Samuel R. Delany, not Ursula LeGuin.

  4. #5 Josh
    January 13, 2009

    This reminds of a very similar question which was at the heart of a raging debate at the wikipedia physics page about the definition of physics. Some wanted to define physics as more or less all of science (everything obeys the laws of physics, right?) while others wanted to include only the traditional stuff one would learn in typical physics courses. I dropped out of this debate when I realized my own definition was “physics is whatever physicists do.” There is a similar phrase amongst string theorists.

  5. #6 Sven DiMilo
    January 13, 2009

    ¨The only rules of scientific method are honest observations and accurate logic.¨
    – Robert MacArthur, ecologist par excellence

    p.s. I´m posting this comment from the Galapagos Islands! Santa Cruz, to be precise

  6. #7 milkshake
    January 13, 2009

    Science is about not fooling yourself (and others) when trying to find out how things work behind the scene. It is the opposite of salesmanship, mysticism and ideology

  7. #8 yogi-one
    January 14, 2009

    Science is the testing of ideas.

    Too dumbed down for me. Sort of like music=sound+rhythm. A jackhammer has sound and rhythm. Is it music? On the other hand, Edgar Varese has pieces that are entirely comprised of machine sounds, and they are considered the compositions of a genius.

    How about this idea: If I make snide comments about my wife’s/girlfriend’s appearance, will she start hating me? The methodology and the result we are looking for is clearly spelled out and we can rigorously test it by trying it many, many times. Does that make it science?

    There’s something much harder to define that is going on that makes the distinction of whether something is scientific or musical. The rub is that while it seems obvious to each of us individually, it is damn hard to define for all of us collectively. My feeling is that it’s too abstract, and any time you start trying to define an abstraction, the going gets rough. It fractals down endlessly into the “edge cases” you mention.

    I think there has to be some agreed upon features.

    One might be “no supernatural or miraculous explanations allowed.”
    And “no circular logic” – you cannot say something is because it is.
    Also, one feature is the testing has to be constructed so it avoids biases built into the methodology.

    I think the closest I can get is calling science an operational methodology, namely, the scientific method.

    For me, science is the application of the scientific method to a given phenomena.

    Then, what you have to do is nail down the definition of the scientific method.

    I think that would be far easier than trying to define the much more abstract concept called “science”.

    I’m not a scientist (I’m a musician with a layman’s interest in science), but that may give me a different perspective on this thing than a physicist or geologist might have.

  8. #9 MadGastronomer
    January 14, 2009

    yogi-one, I’m pretty sure that the experiment you propose would count as psychology or, possibly, sociology, so yes, science, even if the physicists call it “soft” science. Because science is that process.

  9. #10 Umlud
    January 14, 2009

    Of course there is another aspect of science: the accumulation of the knowledge created using science.

    Without previous work (the “standing on the shoulders of giants” if you will), then one could not easily go about “testing ideas” without having to re-invent the wheel each time (or rely only on one’s own work).

    And (in the spirit of #6’s comment) I’m writing this from the IAHS Ecohydraulics conference in Concepcion, Chile.

  10. #11 Thony C.
    January 14, 2009

    Other people come at this from a much different angle. Their interest is not so much in figuring out what works, but in sweeping away what doesn’t work.

    I thought about making the following comment on Matt’s original post but then didn’t. One of his critics remarked that his definition would permit astrology to be called a science. The simply answer to this is it was propagated as a science, was tested and failed and therefore ceased to be considered a potential science.

    In the Renaissance astrology was not only considered to be a science but was considered to be the most important of the mathematic sciences. Before somebody screams, ”what do you mean mathematical?” you actually need quite a lot of maths to produce a horoscope. There were chairs for astrology at all of the leading European humanist universities, even Galileo taught astrology in Padua. The supporters of astrology were however very aware of its very obvious shortcomings and undertook various attempts to shore up its scientific status.

    Firstly, it was believed that celestial influence (the driving force of astrology) was responsible for the weather, so in the 16th century many astrologers began to keep astrological-meteorological diaries in order to produce an empirical proof of astrology. The attempt failed of course but it laid the foundations for scientific meteorology. Other astrologers believed that the problem lay in the inaccuracies of astronomy and so devoted their lives to improving the standards of astronomy; two amongst many were Tycho Brahe and Johannes Kepler. They did not exceed in providing a scientific status for astrology but they did make substantial contributions to the evolution of the new astronomy.

    The purpose of testing is to establish whether a potential science is one or not. This was the case with astrology, which was extensively tested from about 1450 CE to 1660CE the result was failure and the loss of its scientific status.

  11. #12 Coriolis
    January 14, 2009

    Yogi, if you do that experiment right, it certainly could be science. If you wanted to you could try and figure out how many snide comments would it take to make your partner hate you. The problem is that there are so many variables here (how well you get along, how much your partner hates snide comments, etc.), that getting anything other than a statistical distribution from performing many experiments on different couples would be pretty hard. Maybe you’d be able to fit some empirical rule of thumb – i.e. on average 5 snide comments and you’re out the door with some distribution around it, but that’s about it I imagine.

    Like I commented on Matt’s place, apart from maybe political reasons, I just don’t see the use of anything other than the simple definition of science. He didn’t specifically mention what he means by testing – i.e. testing against experimental evidence, but that’s pretty obvious, and Chad fixed it anyways. All of the counter examples (astrology, bible, etc.) fail at the last step – testing against experimental evidence. But if all of a sudden you could use astrology to actually predict what would happen to you – then it’d be science.

  12. #13 Jeremy
    January 15, 2009

    I think a good way to think about this is by considering the birth of scientific disciplines. Bertrand Russell aptly points out that astronomy and psychology used to be philosophies but “…as soon as definite knowledge concerning any subject becomes possible, this subject ceases to be called philosophy, and becomes a separate science.” What has changed to allow these disciplines to join the sciences? In my opinion, it is only the ability to measure things which in turn allows us to ask the world questions with less bias than without measurements.

  13. #14 abb3w
    January 16, 2009

    I think you need to specify “competitive” testing, for starters.

    More formally: Science refers to the process of gathering evidence, forming conjectures about the evidence, developing a formal hypothesis which indicates how the current evidence may be described under the conjecture, competitive testing of all candidate hypotheses under a formal criterion for probable correctness, plus the body of hypotheses testing best thereby and which thereafter are referred to as “Theories”.

    The method of Testing used by Science is dependent on the philosophical assumptions that propositional logic is valid for formal inference, that joint affirmation on the Zermelo-Fraenkel Axioms is self-consistent, and that Reality is Relateable (with at most RE-complexity) to Evidence.

    The rest? Just details of the math.

  14. #15 Jonathan Vos Post
    January 16, 2009

    Re # 4 | Mary Aileen

    “I’m not JVP, but I think that was Samuel R. Delany, not Ursula LeGuin.”

    I thought it better for all concerned if I waited on the order of ten trillions naoseconds for someone else to have a turn. Mary Aileen is right. Nor is Chad wrong, because Ursula LeGuin has written on related definitional issues, since the Sciences she grew up on (given her famous parents) were Anthropology, Sociology, and Linguistics — yet the Ansible is Physics.

    Samuel R. Delany also takes a step up to meta-criticism, with a third approach that is neither inclusive nor exclusive, but Semiotic. His full essays on the matter (a set of 3, I think) were in the New York Review of Science Fiction.

    In that spirit, we may have a meta-theory which valorizes Science, and Pseudoscience, and Magic, in that each satify the same criteria, but in different universes (with different natural Laws) within the Multiverse.

    In that sense, “Macroscope” by Piers Anthony is both Science Fiction and Fantasy, because it is in a universe where Astrology works as well as Astronomy, for complicated reasons.

    One of the two novel manuscripts that I’m trying to finish writing this season is “Axiomatic Magic” — where Wizard-Professor Richard Feynman, at the California Institute of Thaumaturgy in Paradena, is the amateur sleuth trying to solve a magical murder on campus, where John Horton Conway is the prime suspect. Both magic and science work.

    The chapter I’m trying to complete today is about the boy-friend/girlfriend couple who intentionally get themselves locked in the Quantitative Alchemy Lab over the weekend, to finish their Final Exam at the last possible chance, by analyzing what each has in their Flagon. It’s trying to be funny, by being also a homage to a Dobie Gillis story (1959-1963 TV sitcom based on 1953 film based on 1940s and 1950s stories by Max Shulman).

New comments have been temporarily disabled. Please check back soon.