Does science advance?

Karl Popper ( ) “There is no more rational procedure than trial and error – of conjecture and refutation: of boldly proposing theories; of trying our best to show that these are erroneous; and of accepting them tentatively if our critical efforts are unsuccessful.”

Friedrich von Langenfeld, Cautio Criminalis (1631):  On being accused, a woman might run or stay; if she ran, that proved her guilt; if she stayed, the devil had kept her so she could not get away.  If the woman had led an bad life, she was guilty.  If she had led a good life, she was a witch trying to appear virtuous.  In prison, if she was afraid, it was from knowing her guilt.  If she was not afraid, she must be confident in her dark magic.

Scientific theories can’t be confirmed, only falsified  Consider “All swans are white.”  One could never prove that claim, no matter how many swans one finds (recall Hume).  Popper thinks confirmation is easily biased in favor of existing beliefs.  But it takes only one non-white swan to disprove or falsify it.  So Popper claims the demarcating mark of science is falsifiability, not confirmation.

Falsification: Modus Tollens If H, then O. O is false. So: H is false. This is a logically valid argument form.

Predictions So the more falsifiable a theory is – the more it risks – the better it is. E.g.: Einstein’s theory of general relativity predicted a very specific bending of light around the sun (1.17”).

Falsifiable conjectures  F = G(Mm/r 2 ) is falsifiable: the law could have been F = G(Mm/r 1.3 ) or any other value.  “Metals expand when heated” is falsifiable: the next bit of copper heated might not expand.  “Humans evolved from primates” is falsifiable: we could find human fossils older than any other primates’, or our DNA could crucially differ from primate DNA.  “The planets orbit in ellipses around the sun” is falsifiable: they could have been circles or squares, or around the earth.

Unfalsifiable hypotheses  “The gravitational force is some function of some variables.”  “For every house, there is a buyer.”  “Everything happens for a reason.”  “Your dreams are actually about sexual desire for your father.”  “Communism will eventually overthrow capitalism.”  “Your love life will take a turn for the better.”

Popper’s model of science 1. Make bold explanatory conjectures. 2. Deduce unexpected (but falsifiable) predictions from them. 3. Make repeated observations and experiments to test conjectures. 4. Reject what is falsified; work with what survives.

Problems for the simple view of observation

Do you believe it because you see it? Or see it because you believe it?  One common-sense view of science (traceable back to Francis Bacon ( )) is that science is based on facts: on undeniable, publicly verifiable, objective observations.  However…

What do you see here? (What’s the datum?)

A problem for ‘facts’  What we see depends on the context in which it appears, or on what we already believe (the background, our prior beliefs, our theories).  Consider the claim ‘the earth does not move.’ 500 years ago, almost everyone would agree that this claim was an obvious ‘fact’ that we could see was true. To reject it, we need theories such as inertia to revise our ‘observation.’

Facts depend on theories  So we can’t simply appeal to ‘facts’ as the basis for science or common sense.  What I see partly depends on my beliefs, beliefs assumed in the words I use to express what I see, and what parts I choose to pay attention to.  We have to be on guard against the possibility that supposedly simple observations of ‘fact’ are theory-laden: they wouldn’t be “seen” if not for prior beliefs.

Popper: Science is not knowledge  Science, then, is not Nature’s Laws, nor empirical ‘Facts’.  “We must not look upon science as a ‘body of knowledge’, but rather as a system of hypotheses which in principle cannot be justified, but with which we work as long as they stand up to tests, and of which we are never justified in saying that we know they are ‘true’ or ‘more or less certain’ or even ‘probable’.” (Popper, Logic of Scientific Discovery, 1938)

Falsificationism’s flaws But Popper’s falsificationism 1) gives no reason to say theories are true, which is how we practically use theories, and 2) can’t definitively falsify theories, because you can always shift blame to another hypothesis or assumption, rather than the one tested – and this sort of ad hoc rescuing happens all the time in science!

Shifting the blame onto another theory  Our hypothesis: the hare is faster than the tortoise.  Our hypothesis: the hare is faster than the tortoise.  So we predict: the hare will outrace the tortoise.  So we predict: the hare will outrace the tortoise.  We then observe: the tortoise wins.  We then observe: the tortoise wins.  The hypothesis that the hare is faster than the tortoise is NOT thereby falsified because we could say another hypothesis was falsified instead. For example:  The hypothesis that the hare is faster than the tortoise is NOT thereby falsified because we could say another hypothesis was falsified instead. For example:  The hare did not stop for a nap.  The hare did not get run over while crossing the road.  The hare did not get entangled in a philosophical discussion about the rationality of scientific methods with his friend gopher before crossing the finish line.

A real example of shifting the blame  A neutron can decay into a proton and electron (beta decay).  In the 1920s physicists found the combined energies of the proton and electron were less than the original neutron’s. This seems to falsify the principle of conservation of energy.  Pauli suggested another particle is emitted: invisible, tiny, and electrically neutral. Fermi called it the “neutrino”.  At the time, there was no way to detect neutrinos, so Popper would have said “falsified.”  But neutrinos were detected in Conservation of energy is correct. (Or so it seems!)

Can we ever falsify?  If a prediction of a theory is found to be false, and we can’t think of a revised hypothesis that leads to new testable predictions, then Popper says we must conclude that the theory is false.  But it could be that the theory is true and the other hypotheses are responsible for the failed prediction, and those others can’t be tested yet.  So can we ever conclusively disprove a theory? And why believe current theories?

Thomas Kuhn ( )  Science is a series of ways of seeing the world: paradigms.  Paradigms are incommensurable.  Paradigms change in revolutions: science does not progress toward truth.

What is a paradigm? (1) An exemplary explanatory achievement that attracts followers and (2) gives them a model for work to do and puzzles to solve. Kuhn’s examples: Aristotle’s physics, Ptolemy’s astronomy, Newton’s Principia, Lavoisier’s chemistry, Einstein’s relativity

Normal science  A major theory will raise further questions: puzzles.  Scientists perpetuate a paradigm by working on these puzzles. They rarely falsify theories, and interpret, ignore, or explain away anomalies to protect the existing paradigm.  If anomalies persist and accumulate, eventually the paradigm falls into crisis.

Scientific revolutions  If a striking new theoretical achievement explains the outstanding anomalies, then a scientific revolution occurs.  The old anomalies are solved.  Many scientists (especially younger ones) convert to a new paradigm.  New puzzles arise, and normal science resumes … under a new paradigm.

Celestial paradigms Ptolemy’s model accounts for the moving planets, sun, and stars by saying they revolve around the Earth every 24 hours. Copernicus’s model accounts for this by saying the Earth rotates every 24 hours.

Paradigm shifts change what you see  “The physical referents of Einsteinian concepts are by no means identical with those of the Newtonian concepts that bear the same name.” (SSR, 102)  “Lavoisier saw oxygen where Priestley had seen dephlogisticated air and others had seen nothing at all.” (SSR, 118)

Paradigms are incommensurable 1. Scientific terms refer to things through a network of meanings. 2. If paradigms were commensurable, then terms would still refer to the same things across paradigms. 3. Paradigms change scientific terms. 4. So: paradigms are incommensurable.

Paradigm shifts are not rationally decided  “The choice is not and cannot be determined merely by evaluative procedures of normal science, for these depend upon a particular paradigm, and that paradigm is at issue.” (SSR, 88)  “Something must make at least a few scientists feel that the new proposal is on the right track, and sometimes it is only personal and inarticulate aesthetic considerations that can do that.” (SSR, 158)

Science does not make progress If paradigms are incommensurable, then science does not approximate truth over time. “Proponents of different theories are native speakers of different languages.” Scientific theories are more like fashions or religions: historical points of view.

Kuhn (1962)Evans-Pritchard (1937) 1 “Newton’s second law of motion behaves like a purely logical statement that no observation could refute.” (SSR, 76) 2 “Failure... discredits only the scientist and not the theory.” (SSR, 80) 3 “[Scientists] devise numerous ad hoc modifications in order to eliminate any apparent conflict.” (SSR, 78) 1 “Since [magic] transcends experience, it cannot easily be contradicted by experience.” (WOMA, 475) 2 “A witch-doctor is a cheat because his medicine is poor.” (WOMA, 194) 3 “Elaborations of belief free Azande from having to admit … the logical consequences of belief.” (WOMA, 24)

Maps as models