3 ms·
The article portrays Popperian falsifiability as a negative thing, when in fact it is exactly the kind of defining criterion that the author is hunting for.
by jll29 4y ago
The article portrays Popperian falsifiability as a negative thing, when in fact it is exactly the kind of defining criterion that the author is hunting for.
Except it is a criterion of a scientific _theory_ , not a fact derived from it; because the natural sciences are entirely empirical disciplines (today), theories can only be _disproved_ , not proved, by experiments.
Progress - in the form of convergence towards truth among the community of experts (Peirce) - is achieved by revising theories. Like with software, there comes a point when tweaking the old implementation becomes frustrating, and it's time for a re-write from scratch - that'd be a Kuhn-style change of paradigm a la Einstein's theory of relativity. Different from the realm of software, older theories (e.g. Newtonian physics) remain valid and often useful for some restricted scope.
The author Peter Vickers appears to have written this blog post motivated by a recent line of work that resulted in a book that came out entitled "Identifying Future-Proof Science":
https://academic.oup.com/book/44533?login=false https://academic.oup.com/book/44533?login=false (2022, Oxford University Press); putting it in my shopping basket/on my 2023 reading list & curious to hear from anyone who has read it already.
Crazy idea: I propose someone collect a list of facts - e.g. a wiki - that actual scientists can sign individually (e.g. owners of academic email addresses that do not contain a .stud. sub-string).
My contribution of a fact: "On earth, any fruit of _malus domestica_ (commonly called "apple") that is dropped will fall down" (which means from the height of any person towards the centre of earth) as long as (small print to establish the boundary conditions of Newtonian physics).
- pdonis 4y ago> theories can only be _disproved_ , not proved, by experiments While it is technically true that a theory cannot be "proved", that view ignores the huge range of predictive power in our various scientific theories. General relativity and quantum mechanics have extraordinary predictive power: we can make quantitative predictions that match experimental results to ten or more decimal places. But predictive power gets worse by many orders of magnitude as you go down the scale from "harder" to "softer" sciences, to the point where we don't have any real predictive power at all in many areas of social science. > a Kuhn-style change of paradigm a la Einstein's theory of relativity Actually, no, this wasn't a Kuhnian change of paradigm. Why? Because a Kuhnian paradigm shift requires that the two paradigms are incommensurable. And that's not true for relativity vs. Newtonian mechanics. Relativity contains Newtonian mechanics as a particular approximation (low speeds and weak gravitational fields). So Newtonian mechanics is not incommensurable with relativity. By contrast, the shift from Aristotelian physics to Galilean physics (I say "Galilean" since Galileo was actually the one that started the shift, before Newton was even born) was a Kuhnian paradigm shift, because Aristotelian physics and Galilean physics are incommensurable. Aristotelian physics contains teleological concepts that don't have any counterparts whatever in Galilean physics, not even as approximations. > older theories (e.g. Newtonian physics) remain valid and often useful for some restricted scope. With a true Kuhnian paradigm shift, this is not true: the old theory just gets thrown away. Aristotelian physics wasn't used by anyone after Galilean physics became the dominant paradigm, not even in a restricted scope. This is why Kuhn claimed that there is no such thing as scientific progress. He was wrong, of course: once a particular discipline actually becomes an experimental science (which Aristotelian physics was not), your statement about older theories does hold--Newtonian physics, as you say, being a prime example. But that is not what Kuhn was claiming happened in paradigm shifts.
- dumpsterdiver 4y ago> While it is technically true that a theory cannot be "proved", that view ignores the huge range of predictive power in our various scientific theories. It doesn't though. Acknowledging one thing does not mean that all other things are ignored.
- pdonis 4y ago> Acknowledging one thing does not mean that all other things are ignored. That's true in the abstract, but in the particular case we are discussing, the view that a theory cannot be proved goes along with the view that what is important in science is falsification: we rule out competing theories until only one is left standing. But that is not what actually happens. What actually happens is that, as we get more and more data in a particular domain, we refine our models to have better and better predictive power. The more efficiently we can do that (because we are able to conduct controlled experiments with very accurate measurements involved), the "harder" the science is (the "hardest" by this criterion being physics). Nothing ever gets falsified and thrown away. The other problem with the emphasis on falsification is that scientists still use theories that, by Popper's criterion, were falsified long ago. The obvious example is Newtonian physics. By Popper's argument, we should have abandoned it as soon as we had experimental and observational data, in the late 19th and early 20th centuries, that contradicted it. But what actually happened was that one branch of physics developed new theories, relativity and quantum mechanics, while other branches of physics kept right on using Newtonian physics, because, even though it was "wrong", it gave answers that were accurate enough for the purpose. In other words, its predictive power, while less than that of relativity and QM, was good enough for what it was being used for.
- naasking 4y ago> Actually, no, this wasn't a Kuhnian change of paradigm. Why? Because a Kuhnian paradigm shift requires that the two paradigms are incommensurable. And that's not true for relativity vs. Newtonian mechanics. Relativity contains Newtonian mechanics as a particular approximation (low speeds and weak gravitational fields). So Newtonian mechanics is not incommensurable with relativity. I don't think that's an adequate analysis. Newtonian mechanics to General Relativity was indeed a Kuhnian paradigm shift because it required giving up simultaneity, messed with causality, and did away with the notion that space and time were simply an immutable Euclidean background on which particle interactions took place. Instead, spacetime was itself malleable by the entities that inhabited it, and was described by hyperbolic geometry. All of these were tough pills to swallow and required significant revisions to how people thought about reality. That Newtonian mechanics is embedded within GR is one data point to consider when deciding whether something is a paradigm shift, but it's not the only point that matters. By your argument, quantum mechanics was also not a paradigm shift because it reduces to classical mechanics in some limit due to decoherence. I think any interpretation of Kuhnian paradigm shifts that excludes arguably the two biggest revolutions in science, ever, cannot be correct.
- jacquesm 4y ago> I propose someone collect a list of facts Didn't Cyc have a whole bunch of those?
- yamrzou 4y agoIs Popperian falsifiability itself falsifiable?
- avereveard 4y agoThe earth gravity field isn't uniform even in the context of Newtonian physics, the apple will not directly fall toward the center of the earth.