Signs of an Educated Person: Falsifiability

How do you know whether a claim is true or false? One of the first things to consider is whether or not the claim can ever be proven true, or proven false — and if so, how? The ability to disprove an assertion is crucial to scientific reasoning, but it is equally important to day to day discourse, if it is to mean anything.

The question of whether an assertion can be proven false, is informed by the concept of “falsifiability.” The idea of falsifiability was made famous by Karl Popper, a Viennese-born (1902) philosopher who moved to England in 1946, for the last half of his life and work (d. 1994).

Falsifiability is the ability of a theory—a working framework for explaining and predicting natural phenomena—to be disproved by an experiment or observation.[1] The ability to evaluate theories against observations is essential to the scientific method, and as such, the falsifiability of theories is key to this and is the prime test for whether a proposition or theory can be described as scientific. __ RationalWiki

Cycle of Reasoning Cycle of Reasoning
https://explorable.com/falsifiability%5B/caption%5D

For Popper, a theory is scientific only if it is refutable by a conceivable event. Every genuine test of a scientific theory, then, is logically an attempt to refute or to falsify it, and one genuine counter-instance falsifies the whole theory. In a critical sense, Popper’s theory of demarcation is based upon his perception of the logical asymmetry which holds between verification and falsification: it is logically impossible to conclusively verify a universal proposition by reference to experience (as Hume saw clearly), but a single counter-instance conclusively falsifies the corresponding universal law. In a word, an exception, far from ‘proving’ a rule, conclusively refutes it.

Every genuine scientific theory then, in Popper’s view, is prohibitive, in the sense that it forbids, by implication, particular events or occurrences. As such it can be tested and falsified, but never logically verified. Thus Popper stresses that it should not be inferred from the fact that a theory has withstood the most rigorous testing, for however long a period of time, that it has been verified; rather we should recognise that such a theory has received a high measure of corroboration. and may be provisionally retained as the best available theory until it is finally falsified (if indeed it is ever falsified), and/or is superseded by a better theory. _ Stanford Encyclopedia on Karl Popper

Pseudo-scientists — such as those who currently profit from the milieu of climate apocalypse — refuse to be held to the strict discipline of falsifiability and scientific reasoning. This makes the entire enterprise of climate apocalypse — from the scientific to the political — meaningless, except as a gauge of the people who espouse radical economic and political transformation based upon unfalsifiable pseudoscience.

Political propaganda — such as we see from Russian state media, Chinese state media, or British state media such as the BBC — is another example of essentially worthless unfalsifiable assertion. Since state propaganda is not held to the discipline of being disprovable, it is only a tool of power — meant to influence gullible and undisciplined minds.

Most discussion on the internet does not come close to the level of falsifiability. Strict scientific falsifiability requires a good deal of mathematics and esoteric knowledge. Common speech and writing requires only reliable documentation — and the understanding that sometimes one must take his best shot, until more and better evidence comes in. The poorer the evidence, the more important it is to collect better evidence before making a significant commitment to the idea — depending upon the time allowed by the situation.

If someone is approaching from a dark alley, holding a metallic object that glistens dimly from a streetlight, time may dictate a quick decision as to fight or flight.

No matter how much schooling a person has had, if he is not comfortably acquainted with the idea of falsifiability — and the levels of falsifiability in daily life — he cannot be said to be educated. Education often has nothing to do with schools or schooling, and the sooner a child learns that lesson, the brighter his prospects will be for an informed and liberated life.

Advertisements
This entry was posted in Philosophy and tagged . Bookmark the permalink.

6 Responses to Signs of an Educated Person: Falsifiability

  1. jabowery says:

    A better exoteric slogan than “That’s not falsifiable!” is “That’s not even wrong!“.

    A better esoteric slogan than “That’s not falsifiable!” is “That’s not elegantly predictive!”

    From “A Complete Theory of Everything (will be subjective)” by Marcus Hutter:

    Predictive power and elegance. Whatever the intermediary guiding principles for designing theories/models (elegance, symmetries, tractability, consistency), the ultimate judge is predictive success. Unfortunately we can never be sure whether a given ToE makes correct predictions in the future. After all we cannot rule out that the world suddenly changes tomorrow in a totally unexpected way (cf. the quote at beginning of this article). We have to compare theories based on their predictive success in the past. It is also clear that the latter is not enough: For every model we can construct an alternative model that behaves identically in the past but makes different predictions from, say, year 2020 on. Popper’s falsifiability dogma is little helpful. Beyond postdictive success, the guiding principle in designing and selecting theories, especially in physics, is elegance and mathematical consistency. The predictive power of the first heliocentric model was not superior to the geocentric one, but it was much simpler. In more profane terms, it has significantly less parameters that need to be specified.

    • alfin2101 says:

      Thanks. Hutter has some interesting ideas as a computer scientist. By sticking to mathematical and algorithmic languages, Hutter has made a name for himself.

      When such a person strays into the morass of common language and logic, he should not expect to be taken as seriously as in his day job.
      He should not assert that “predictive success” is separate from “falsifiability,” for example. What is a “successful theory or model” if not one that has not yet been falsified?
      His forays into “theories of everything” and “elegant theories in physics” can be inspiring and thought-provoking, but at the expense of rigour. More like poetry of science than science itself.

      He can be downright disappointing when he discusses approaches to artificial intelligence, for example: http://theconversation.com/to-create-a-super-intelligent-machine-start-with-an-equation-20756

      Unfortunately, this misguided, dead-end algorithmic approach to AI is all too common in those “cognitive scientists” whose grounding is almost exclusively in computer science.

      • jabowery says:

        While I can’t speak for Hutter I think the way falsifiability rationally enters into scientific discourse is in what Hutter would call the “parameters” of the system that add to the Kolmogorov complexity. Rationality usually requires taking ratios of commensurable quantities to maintain perspective. One man’s “fasification” is another man’s “experimental error” is another man’s “irrelevance”. Kolmogorov complexity incorporates erroneous predictions into the algorithmic description of the observations as uncompressible data that can be thought of as parameters. You don’t throw such “falsifying” observations away — you don’t even try to compress them. They count _very_ heavily against the model but do not logically “falsify” so much as render the resulting K program a lot more complex. An “experimental error” is handled by a description (minimum description length and minimum message length are related fields in the philosophy of science) of the experiment, consistent with the rest of the model, that explains the observation in terms of that model — thereby compressing the observation in a way that is less complex than merely recording a “falsification”. It imputes the error in the experimental method. One of the main ethical dilemmas in science is where the “burden of proof” from such an imputation resides. Everyone wants the other guy to test their theories.

        Now, having said that, I’ll agree that Hutter’s “top down” approach to AI is unlikely to yield much in the way of practical discoveries for the next decade or two. However, when young men ask me for advice in this area, I basically tell them that if they want to make money in the near term, they should study imputation of missing data, and if they want to be where the field will be when they’re at their professional prime, they should study Hutter. Hutter’s approach is the most rigorous mathematical approach to universal intelligence and is therefore in a position to take advantage of advances in mathematics in a way that other approaches are not.

      • alfin2101 says:

        Falsifiability in science is a harsh discipline, and many scientists are unwilling to pay their dues to something that may prove much of their life’s work to be a waste. All the worse for them when they are judged by a better informed history down the line.

        It is important to acknowledge that “information science” occupies something of a middle ground between mathematics and science, and may well have its own rules of falsification in many of its labyrinths of thought.

        Falsifiability in ordinary language and argument is far less formal than what one will find in science.

        I am far from an expert in the esoteric subspecialty realms of Kolmogorov complexity. My only comment on that topic would be to suggest that it does not hold the key to resuscitating the corpse of algorithmic artificial intelligence.

        If Kolmogorov complexity expands in the future to become the informational equivalent of “The Philosopher’s Stone,” it will indeed become an important area of study for ambitious young men. :-}

  2. William Newman says:

    “No matter how much schooling a person has had, if he is not comfortably acquainted with the idea of falsifiability — and the levels of falsifiability in daily life — he cannot be said to be educated.”

    As far as I know, nobody had really thought very carefully until about the middle of the twentieth century. (Part of the interest is how quantum mechanics and relativity revealed that the most famous hugely successful scientific theory was only correct in ordinary conditions, breaking down completely for tiny particles and near the speed of light.) I realize that reasonable people may differ passionately about what educated means, but I’m gonna claim that any definition of educated which excludes absolutely everyone before 1930 is highly suspect.

    (Admittedly lots of people before 1930 behaved as though they appreciated the importance of falsifiability even if they never thought about how to articulate it. But so do many people today who aren’t familiar with the modern systematization of the idea.)

    “What is a ‘successful theory or model’ if not one that has not yet been falsified?”

    A theory that “compresses” a large amount of observational data (that rival theories don’t) is successful in a vitally important way; for a pointer into the technical details of applying this intuition inpractice, see http://en.wikipedia.org/wiki/Minimum_description_length . It’s hard to justify this claim in a blog comment but you can take it from me, or spend some time chasing references from the MDL argument and checking for yourself: this is not only a natural generalization of Popper’s idea of a successful theory and its falsifiability, it is also an important generalization. One reason it’s important is that it neatly captures why it was still so useful to know Newton’s theory of mechanics even after people found results that were inconsistent with the theory in extreme cases. Another reason it’s important is because it makes it clear what it means for a probabilistic theory to be successful, and/or for a historical theory to be successful. Doing that is nontrivial (except perhaps if you’ve studied some information theory:-). Note that Popper himself (who like all well-known philosophers of science seems to’ve ignored Shannon’s work, and thus naturally ignored the Shannon+Turing+Kolmogorov coding idea that jabowery refers to) seems to have become bewildered for some years about the fundamental scientific meaningfulness of Darwin’s theory of evolution by natural selection (a theory which is both probabilistic and historical).

    • alfin2101 says:

      Information theory and complexity are useful and powerful tools used in modern science and technology. But discussion of Kolmogorov complexity etc. is much in the way of a red herring and non sequitur in regard to the blog article above. The point being made there has almost nothing to do with Popper’s ability to exhaustively describe and predict all possible past-present-future forms of falsifiability, and everything to do with being able to conceive of ways to falsify everyday beliefs, “news stories,” conspiracy tales, folklore, or any basic argument or theory.

      Popper’s ideas of falsifiability can provide a useful bridge between formal logic, scientific practise, and everyday reasoning.

      The sense of what is required for a person to be “educated” will naturally change over time with what is learned, discovered, and devised. Work done in the 16th century on the scientific method ( http://plato.stanford.edu/entries/francis-bacon/ ) supplied much that was needed for a later formalisation of Popper’s falsifiability concept. Many similar proto-concepts were no doubt floating around in ancient Greece, lost now in antiquity.

Comments are closed.