Thursday, January 24, 2013

What Is Science? - Still No Answer!

We had a fun meeting last night thanks to Rufina Kim [WTF Is Science?]. A bunch of students showed up along with Steve Livingston, the new co-Chair of CASS (Committee for the Advancement of Scientific Skepticism), and David Bailly, Chair of The Association for Science and Reason (Skeptics Canada).

Unfortunately we were not able to come to an agreement on "What Is Science."

Now we have to meet again in a couple of weeks!

We talked about whether there was a scientific method and whether falsifiability is part of the definition of science. The Wikipedia article on falsifiability is a good place to look for background information. Here are two sections from that article to get you started.
Paul Feyerabend examined the history of science with a more critical eye, and ultimately rejected any prescriptive methodology at all. He rejected Lakatos' argument for ad hoc hypothesis, arguing that science would not have progressed without making use of any and all available methods to support new theories. He rejected any reliance on a scientific method, along with any special authority for science that might derive from such a method. Rather, he claimed that if one is keen to have a universally valid methodological rule, epistemological anarchism or anything goes would be the only candidate. For Feyerabend, any special status that science might have derives from the social and physical value of the results of science rather than its method.


In their book Fashionable Nonsense (published in the UK as Intellectual Impostures) the physicists Alan Sokal and Jean Bricmont criticized falsifiability on the grounds that it does not accurately describe the way science really works. They argue that theories are used because of their successes, not because of the failures of other theories. Their discussion of Popper, falsifiability and the philosophy of science comes in a chapter entitled "Intermezzo," which contains an attempt to make clear their own views of what constitutes truth, in contrast with the extreme epistemological relativism of postmodernism.

Sokal and Bricmont write, "When a theory successfully withstands an attempt at falsification, a scientist will, quite naturally, consider the theory to be partially confirmed and will accord it a greater likelihood or a higher subjective probability. ... But Popper will have none of this: throughout his life he was a stubborn opponent of any idea of 'confirmation' of a theory, or even of its 'probability'. ... [but] the history of science teaches us that scientific theories come to be accepted above all because of their successes." (Sokal and Bricmont 1997, 62f)

They further argue that falsifiability cannot distinguish between astrology and astronomy, as both make technical predictions that are sometimes incorrect.
There's no such thing as a universal scientific method and falsifiability doesn't describe how the scientific way of knowing actually works.

I made up an example of a Professor of English whose research focuses on how the English language actually sounded in the time of Geoffrey Chaucer (about 1370)¹. Is she doing science? If not, what kind of way of knowing is she using?

1. This is roughly the time of World Without End. If the characters actually spoke in 14th century dialect we probably wouldn't have understood a word.


  1. I haven't read the comments on the previous thread, and I'm sure someone has mentioned it before, but just in case:

  2. You really should read Wittgenstein. And then, in a few years, and with a bit of work, and a bit of luck, you might come to see how funny what you've just written is.

    1. Please feel free to enlighten us on Wittgenstein. We need a good laugh.

    2. What, in 15 minutes? LOL. Would you also like me to tell you how to speak Japanese? These things take time and hard work I'm afraid. Here:

      Reasonable commentary into the bargain.

    3. Whereof one cannot speak, thereof one must be silent.

    4. That was his early philosophy. He thought something quite different in his later work. Now the ladder is no longer nonsense, it's good exercise - but only if one needs the (that particular type of) exercise.

    5. So reading Wittgenstein like reading the Koran, where earlier suras are superseded by later ones ?

      Or is Wittgenstein revealing His Word progressively, the revelation being levelled at the comprehension and culture of the people to whom it was first given, as recounted by His prophet Luther, PBUH.

    6. Not at all, he just thought something through, and thought it through, many argue, better than anyone else in the world had managed, but then changed his mind and explained how and why he was wrong.

      Re this topic, the kind of question being discussed, and the mistakes being made, were fairly central to his later philosophy. That's why I brought it up here.

  3. I can help with this definition, use it don't use it.....

    "Science is the study of what causes what!"

    If the world stuck to this definition of science half of the crap called "science" these days would not even exist.

  4. I made up an example of a Professor of English whose research focuses on how the English language actually sounded in the time of Geoffrey Chaucer (about 1370). Is she doing science? If not, what kind of way of knowing is she using?

    That's me and the work I'm paid for doing. I'm not female, to be sure, but many of my colleagues are. I'd defend the view that linguistics (including historical linguistics) is a branch of science.

  5. I always find the description of the scientific method as one of falsification as being a little odd; its certainty not the sole tool used in any field of science I've ever worked in/been familiar with. Its also not how I was taught (and continue to use) falsification in science.

    I do consider it a critical tool, thorughbeit, in a different way. Any hypothesis/theory must be falsifiable to be scientific - i.e. if it is wrong, it is possible through experiment/observation to prove the idea false.

  6. Listen up punks, here's the definition of what science is.

    Science consists of two things:
    1. selection of general/universal laws of nature via the scientific method [see def'n below] which is a competition between theories based on the fit between observable quantities and the testable predictions that theories emit; and

    2. Induction; or, application of the general principles thus found, often via analogy.

    Example of 1. Observe positions of some planets, input them to Newton's theory of gravity, make testable predictions, compare them to observations. See if Newton's theory beats, e.g., Aristotelian model.
    Example of 2. Apply Newton's theory of gravity to solve for angle theta at which cannon should be tilted so that cannonballs fired from cannon will fall on enemy troops.

    Engineering consists entirely of 2, induction. No scientific method, no competition between theories that model universal laws.

    Medicine consists of applying 2. induction, and also 1, scientific method, particularly while diagnosing the cause of diseases. However, medicine is NOT science because: it does apply the scientific method to competing hypotheses of causes of diseases, but the hypotheses describe specific instances of disease, not general/ universal principles.

    Ditto plumbing: applies 2, induction, and uses 1, scientific method, but still NOT science. Plumbers may apply the scientific method to competing hypotheses of the source of a water leak, but again, the hypotheses describe specific instances of water leaks, not general/ universal principles of hydrology etc.

    A competition between theories which are EXPRESSED AS ALGORITHMS that input observable quantities [X] and output predictions that can be compared to other observable quantities [Y].
    (Note: X and Y must be independent, if not you get circular logic and patholocial science.)

    The winner of the competition between theories is judged by two criteria.

    1. Predictive power: fit between testable predictions and observable quantities.

    2. Simplicity of theory / Kolmogorov-Chaitin complexity of algorithm / Minimal Description Length (MDL). [Occam's Razor goes here].

    Complication: Predictive power is judged by three criteria.

    1a. Accuracy: distance between prediction and observation.
    1b. Specificity of prediction. [Does theory accommodate all conceivable data? Or does it exclude some?]
    1c. Precision of observed data. [Are observed data known with great experimental precision or crappy and approximate?]

    1a, 1b, and 1c are competing interests, and one might argue that the trade-off between them will be forever subjective. However, I disagree, as I explain below.

    1. Continuing...

      1a, 1b, and 1c are competing interests, and one might argue that the trade-off between them will be forever subjective. However, I disagree.

      I believe that improvements in statistical theory will make this trade-off totally objective in the future, especially if we apply Kolmogorov information theory to scientific theories expressed as ALGORITHMS.

      Here is an example of an OBJECTIVE method of deciding between the competing interests of 1a, 1b, and 1c: the chi-squared statistical test, which is used to discriminate between different polynomial models of varying complexity: for example, consider fitting predictions Yi given some observed inputs Xi, where the predictions Yi should match observed values Y'i.

      Here we could float various polynomial models:
      Constant: Yi = A
      Linear: Yi = A + BXi
      Quadratic: Yi = A + BXi + CXi^2
      Cubic: Yi = A + BXi + CXi^2 + DXi^3

      etc., you get the idea. An infinite number of models is conceivable. Obviously, more complex models will produce a better fit between predictions and data. However, Occam's razor demands that we don't want an over-complex theory; they're not robust. You have competing interests between model simplicity and accuracy of predictions.

      So you do a chi-squared test, which objectively compromises between the competing interests of model simplicity and predictive accuracy.

      Since this is possible in the case of the chi-squared test, I conclude that in the future, statistical theory and Kolmogorov information theory will resolve the competing interests of simplicity and predictive power.

    2. By the way, in case it's not clear:

      Intelligent Design is not science because, as a theory, it does not emit testable predictions. Expressed as an algorithm, "God makes it so" has the same algorithmic form as "I don't know." They both emit the same number of testable predictions: ZERO.

      Young Earth creationism would be falsified, if we compared the HONEST predictions of the Noah's Flood model against observed geological, paleontological, etc. facts, the HONEST predictions of YEC would not match up to the data.

      However, Young Earth & Old Earth creationism are PSEUDOSCIENCE, which I define as distinct from science.

      Pseudoscience is defined as a set of theories whose claimed predictions DO NOT follow from the hypothesis as described, i.e. the claimed predictions are not honest. For example: when Young Earth creationists look at 1,000 feet thick layer of chalk [mostly foraminifera fossils] at the White Cliffs of Dover, and creationists say, "Oh yes, our Noah's Flood theory predicts there should be a stack of foraminifera fossils in chalk 1,000 feet thick formed in 3 days or so during the year of Noah's Flood."

      Naw. That's not an honest prediction of your theory, expressed as an algorithm. If you lie about the predictions that come from your theory, expressed in its simplest form, it's pseudoscience.

    3. @Diogenes

      Expressed as an algorithm, "God makes it so" has the same algorithmic form as "I don't know."

      As well, "God makes it so" has the immediate detrimental effect of stopping any further investigation into the theory while the admission that "I don't know" opens up the theory to more investigation.

    4. Oberski: "God makes it so" has the immediate detrimental effect of stopping any further investigation into the theory while the admission that "I don't know" opens up the theory to more investigation.

      But that's a sociological judgment, and ID "scholars" of the (alleged) history of science would disagree, asserting that historically, science is supposedly based on Christianity.

      So then you that drags you into arguments over the history of science, e.g. does it matter that Newton believed in God? etc. They will never, ever concede that point.

      Steve Fuller, the pro-ID sociologist who testified at Dover and blathers in "Expelled", would disagree with you and say that the history of science shows that "God makes it so" inspired great geniuses, blah blah. Therefore ID deserves NOT to be expected to make testable predictions, as a form of "affirmative action" [Fuller's exact phrase] to correct for discrimination against theists.

      That actually is an accurate description of Fuller's beliefs: ID should not be expected to make testable predictions-- we should lower or eliminate our standards for ID-- but, says Fuller, we should pretend or act as if ID really is science anyway, because oh, it's "affirmative action" [yes, he did say that] for theists, they deserve it, they need affirmative action.

      John "Quote-Miner" West would argue also, so would Klinghitler, bringing up the names of dead white men (DWM's) who believed in God, all veddy intellectual.

      I don't want to argue about history of science. I want to define the scientific method, and make clear what happens if you redefine it.

      If you redefine the scientific method even a little, "my cow died" leads straight to: "My neighbor's a witch."

      That was in fact Cotton Mather's logic, and Mather was, in fact, the first American to promote Intelligent Design [see "The Christian Philosopher."]

    5. A really nice job, Diogenes! Of course what you define is the fundamental machinery of science -- how theories are built and tested. Science as done on the everyday life of scientists is mostly about the use of models derived from general theories to solve more specific problems (such as explaining any results of ongoing observations in terms of what we already know).

      Also, "general" is a relative term. Is biology a science if its area of study is something as parochial as life on Earth? In the Milky Way alone here may be several billion planets that might support some form of "life", but we have no "universal theories" of life on a cosmic scale. By contrast, the applicability of the known laws of physics or chemistry extends as far as the observable universe. Still, I don't think anybody here doubts that biology is an important and successful discipline of science despite its local scope.

    6. That was in fact Cotton Mather's logic, and Mather was, in fact, the first American to promote Intelligent Design

      Didn't Cotton Mather advocate the use of dreams as visions as "spiritual evidence" in court trials?

    7. Piotr: Didn't Cotton Mather advocate the use of dreams as visions as "spiritual evidence" in court trials?

      He wanted to put certain limits on spectral evidence. He supported it with certain conditions.

      Mather was not technically a witch-hunter as he did not, SFAIK, accuse anyone of witchcraft while they were alive. He did accuse them of witchcraft after they had been executed. Real nice for their families.

      Mather was more of a witchcraft detective, he wanted to improve the methods of identifying witches.

      His evidence for accusing one woman of witchcraft (after she was dead) included things like a cow died, and there's no natural explanation.

      He applied the same god-of-the-gaps thinking to science, which is why I count him as the first American proponent of ID. In "The Christian Philosopher" he goes on and on about "the imbecility of reason", meaning that reason can't solve all kinds of problems.

      He describes the paradox of atomic theory: if we keep slicing a material smaller and smaller, is there a point where it stops being the same material? Since no mind can resolve the paradox, that proves Christ is our Savior.

      He goes on at great length about magnets and their behavior. No mind can ever explain magnetism, that proves Christ is our Savior.

      This is the same logic, of course, no just of ID, but also of ICP, the Insane Clown Posse in their song "Miracles". That's where they rap "Fuckin' magnets-- How Do they Work?" Since no mind can ever explain magnetism, that proves it is a miracle.

      The difference is that Mather didn't attack scientists per se (just Descartes, he hated Descartes, the atheist.)

      Whereas ICP, being a product of American evangelical culture, rap in their song that scientists are LIARS and should shut up when they lie about being able to explain stuff, like you know, magnets, rainbows, blah blah.

      Mather also promoted the first American fake creationist fossil: a tooth of (what was later identified as) a mastodon, which Mather interpreted as a Biblical giant's tooth, from Genesis 6. He wrote to the Royal Society to promote this proof of Noah's Flood.

      That was understandable ignorance on Mather's part. Since then, of course, creationists have perpetrated dozens of frauds, hoaxes, and dishonestly described real fossils (like Paluxy man-prints, Calaveras skeleton, the "Black Skull of Freiberg" etc.) to prove humans and dinosaurs coexisted.

      Modern creationists are much more dishonest than Mather, and hate scientists with far greater intensity.

  7. Just reading the opening thread it shows the old idea of science as a noun is being threatened.
    No universal scientific method means no method exists.
    Indeed it doesn't.
    It always was just a attempt to sure up conclusions. Be careful before conclusions are settled.
    If one is careful then ones conclusions can be confidently asserted.
    Just simple quality control was what the old ones tried to make investigation into nature etc to be.
    Then science became a noun and was turned into a special place of human thinking.

    There's no such thing as science.
    There is conclusions about things and these conclusions turned into other things useful but its all just people thinking.

    Methodology of sureing up conclusions could only come after the original huch/ apple on the head.
    Science doesn't create ideas but verify's for/against them.
    If it does that.

  8. Two words: Bayes Theorem.
    All valid reasoning is based on Bayes, including science.

    1. See Richard Carrier's YouTube video on the topic for an entertaining introduction.

    2. Two words: Bayes Theorem. All valid reasoning is based on Bayes, including science.

      Amazing. Does that mean that Watson & Crick were using Bayes Theorem even though they didn't know it?

      How about Darwin? Can you point me to examples in On the Origin of Species?