devilsadvocate
New Member
Little bit of preface and apologetics for philosophy:
The reason I'm writing "The Good, The Bad, and The Ugly" is mostly because sometimes I feel just a bit too much scientism radiating off from the proponents of skepticism. While I wholeheartedly agree that science is special, it is far from clear why this is so. There is no branch of science that tells us what is science and what is not, and the very question "Why is science special?" is not answerable by science itself. Thus, we cannot avoid philosophy, and I feel that we shouldn't even try to. It is important to know at least something about the challenges the definition of science faces and of the breakthroughs philosophy of science has made, in order to better understand the practice and the special epistemic status of science.
Ideally, I would want to make condensed reviews, "The Good, The Bad, and The Ugly", for most of the well-known contributors to philosophy of science, in order to show that all of the proposed philosophies can be persuasive in their own right, but that they also leave us with loose ends and perhaps with more serious problems, as well as to show that the central issue, "What is science and what isn't?", is still largely left unresolved. I don't know if I will actually be making many more of these, as I just happened to have lots of lecture notes on this particular philosopher, and as such it was easy enough to make this "review". If I will, I'll probably make a blog where I put them, as well as post them here on the League of Reason forums for criticism and discussion. In any case, here is one for arguably the most inspiring philosopher of science, Karl Popper.
Karl Popper (1902-1994) was and is hugely influential philosopher of science, especially among scientists. Popper, in his early academic years held high interest in Einstein's theory of relativity, Sigmund Freud's and Alfred Adler's psychological theories, as well as in Marxist theory of history, all of which at one time were touted as scientific by the advocates of these theories. These were all taken very seriously in Vienna, where logical positivism was founded and on which intellectual circles Popper grew up. Popper's philosophy was influenced by his disillusionment in the latter of the three theories, criticism to logical positivism and verificationism, and by the problems Hume had resurfaced for induction.
The Good
Popper realized Adler's psychological theory, in which self-struggle against feelings of inferiority was portrayed as being motivation for human actions, could be used to explain contradictory behaviour: Helping a damsel in distress was ailment to inferiority by means of elevating oneself's ego for being chivalrous, beating a woman was blamed on the same motive in confines of his theory, though realized more grimly by establishing domination and control. Freud's theory was likewise able to be made fit any observation, and marxists could always invent auxiliary explanations a posteori to save their theory, as they did to explain why revolution took place in Russia rather than in Germany - They could not have taken into account Lenin's charisma and political prowess. This analysis lead Popper to the observation that is very fundamental to his philosophy of science, verification is cheap.
Popper, then, doesn't want to say Adler, Freud or marxists did not explain observations, or that the respective theories did not fit the data well. He remains unimpressed because evidence is too easy to come by. In his view observation is theory dependent and almost anything can as such be seen to verify the theory in question, while almost nothing can be seen as falsifying it. Cleverly formed pseudoscience can be even confirmed by contradictory data, as we saw with Adler.
The difference Popper found in Einstein's theory was that it was informative in the way Adler, Freud or marxism wasn't. General relativity made a bold claim that light will bend in reaction to gravity and this prediction put the whole theory in danger, had Eddington's experiment turned out differently. In this way Popper portrays science and scientists as almost heroic, willing to put their lives work at risk for truth. The difference between pseudo-science and science, according to Popper, is that scientific practice does not aim to form theories that fit the existing data well or that explain every possible outcome, that is easy, but to formulate them as such that they reveal new information that is readily falsifiable. Real science is risky, and bold claims that run contrary to common knowledge are virtues, while conformity can be seen almost as a vice.
To sum up Popper's central finding: There's asymmetry between verification and falsification - No amount of verification will make a theory true, but even a single falsifying experiment will make the theory false.
The Bad
The first thing to notice about Popper's philosophy is that it does not define science. Falsifiability in itself is not sufficient to call theory or hypothesis scientific. The claim, "If it rains tomorrow, I am Batman", is falsifiable, but, I'm sure we'll agree, shouldn't be counted as scientific one. All scientific theories must be falsifiable, but not all falsifiable statements are scientific. This is not a big problem, or problem at all for that matter, but important enough to mention as Popper is often read as solving the demarcation criteria. He doesn't, he merely gives one necessary condition.
One problematic consideration Popper's theory faces is whether it is quality of the theory that makes it scientific, or whether theory's scientific status depends on how it is treated by it's practitioners. Behe in the Dover trial admitted that the definition of science that must be used in order to include intelligent design as sciencific theory would also include alchemy and astrology as such. This is a bit of atheist quote mining, because what Behe meant is that alchemy and astrology should be allowed in the scientific competition, not that they are likely contenders to win. Behe makes here an argument Popper would likely agree with: Alchemy and astrology in the sense he talks of them are sciences, albeit ones that have been falsified. We cannot simultaneously say that astrology, or creationism, give wrong predictions and are therefore false, and that they're unfalsiable and therefore not sciences. At least not if we are going with falsification criteria alone.
The question that then rises, is if the devout insistence, in face of piling evidence to the contrary, of the practitioners of astrology, or creationists, is condition to call the theory unscientific as long as the theory itself is falsifiable. Some astrologists and creationists do discard their respective theories when they find out the predictions do not pan out and some don't. If all scientists working on quantum mechanics only did it for the money and were unwilling to take any experiment as falsifying the theory, would we be willing to call quantum mechanics unscientific? Popper does not give a clear answer which one it is: Does scientific status depend on the theory itself or on how it's practitioners are behaving.
Another consideration is the case of sciences that are too complex to give accurate predictions. We're not going to throw medicine out the window just because there's some inherent uncertainty in the practice, nor do we do that to the theory of evolution because it can't predict precisely what sort of mutations will happen and take over in a population, or if indeed any will happen at all. How is this then different from astrologists admitting that they sometimes get the reading wrong? It may be that astrologist get their predictions wrong more often than practitioners of medicine, but this seems quite subjective evaluation. There's lots to be said on the subject, but Popper's proposed falsification criteria is not at least straight forwardly applicable to probalistic claims.
The Ugly
Good case can be made that Popper's criteria is too strict. As it stands, it is in danger to leave out genuine sciences. Advocates of copernican theory tried to measure parallax of the stars in vain, so they gave auxiliary explanation that stars were far more distant than what was thought - A hypothesis that could only be falsified centuries later. When it was observed by Boveaurd that Uranus did not follow an orbit newtonian mechanics dictated, the observation didn't lead Boveaurd to discard the theory. Instead he postulated there was another planet perturbing Uranus's orbit and some 25 years later Neptune was first observed. Likewise quantum mechanics gives wildly false predictions when it is applied to objects that are both massive and small, like blackholes or the early conditions of big bang. The problem is that inventing explanations to save a theory is not only what we observe scientists regularly doing, but fitting discordant observations by modifying the theory is the very frontier where scientific progress is made. How then, are astronomers and physicists different and better from marxists? If we were to take Popper's criteria seriously - along with his idea that conformity to observation is worthless -, and never invent auxiliary explanations in order to save a theory, may they be independently falsifiable or not, we would have no science left. This would be a devastating undertaking.
It needs to be said that Popper realized the problem himself, and instead of advocating what he called "naive falsificationism", he argued that auxiliary hypotheses are permissible as long as they are independently falsifiable. This, however, raises more problems. First, there might be infinite number of ad hoc falsifiable hypotheses that can be used to save the theory, in practice leaving a theory unfalsifiable. Second, auxiliary hypotheses cannot be tested without another set of underlying assumptions, leaving us with an infinite regress of falsifiable hypothesis. To this objection Popper says that ultimately the decision to discard a theory depends on scientists subjective evaluation of the problems the theory is facing. A compromise that seriously undermines the appeal of Popper's philosophy.
An even bigger blow to Popper comes from his insistence that a failure to falsify a theory is not evidence for the theory. When Eddington made his famous experiment confirming Einstein's predictions, Popper did not see it as evidence for general relativity. Bayesian, for example would say that the experiment highly confirmed the theory because the outcome was contrary to prior beliefs, but Popper must claim, somewhat in face of common sense, that the experiment constitutes no evidence at all. By being consistent with his principle of falsification he gives solution to Hume's problem of induction, but the cost is high as we will see. He can give an answer to Hume's problem: The claim that Sun will rise every day is scientific hypothesis on the grounds it is easily falsifiable. What he is NOT saying, is that this theory is likely to be true because it survives a test every morning. The only implication is that the theory is not dead yet. He has to deny failure to falsify a theory is a reason to think the theory is more likely to be true, because by doing that he would play into Hume's grasp. The claim that past performance of surviving batteries of tests is indicator of future success of a theory, relies on induction.
Popper, then, faces serious problem on how to explain why we should prefer corroborated theories to untested ones. Certainly we should prefer theories that have survived falsification to theories that haven't, since we already know the latter to be false, but what explanation can he give to prefer established theories over untested ones? Why aren't all theories that are not known to be false treated equally? Popper seems to give no answer to the question, other than a very unsatisfying one - That is how science is done. What he seems to be saying is that scientists should prefer tested theories because there is nothing better to go on, but certainly untested theories are no worse, if we allow his own conclusion that surviving tests counts as no reason to think the theory is more likely true. In fact untested theories could, in some unscientific sense, be better. Finding Higgs boson by reading pig intestines is certainly cheaper than the experiments going on in CERN and arguably more fun. That is not to say that it will work, but as long as that theory remains untested, and we are still taking Popper seriously, it should be taken as likely to work as quantum mechanics.
In conclusion, Popper's theory is elegant and there seems to be something profoundly right about it. It's riskiness and falsifiability, not evidential support that makes theories scientific. On a little more deeper look, it is perhaps found that the criticism of verificationism is what seemed so right about it, rather than the persuasive weight he can accumulate to support his philosophy.
Demarcation criteria should be subject to the porridge test: Whether it is too cold, too hot or just right. Popper's falsificationism seems clearly too hot. It is difficult to interpret his criteria in a way that allows physics, for example, to pass the test and marxist theory of history not to. It fails to be descriptive, as actual scientific practice does not follow the theory. And it fails to be persuasively normative for reasons outlined in the "Ugly" section. It's simple appeal comes with loads of problems, some of which seem overwhelming.
The reason I'm writing "The Good, The Bad, and The Ugly" is mostly because sometimes I feel just a bit too much scientism radiating off from the proponents of skepticism. While I wholeheartedly agree that science is special, it is far from clear why this is so. There is no branch of science that tells us what is science and what is not, and the very question "Why is science special?" is not answerable by science itself. Thus, we cannot avoid philosophy, and I feel that we shouldn't even try to. It is important to know at least something about the challenges the definition of science faces and of the breakthroughs philosophy of science has made, in order to better understand the practice and the special epistemic status of science.
Ideally, I would want to make condensed reviews, "The Good, The Bad, and The Ugly", for most of the well-known contributors to philosophy of science, in order to show that all of the proposed philosophies can be persuasive in their own right, but that they also leave us with loose ends and perhaps with more serious problems, as well as to show that the central issue, "What is science and what isn't?", is still largely left unresolved. I don't know if I will actually be making many more of these, as I just happened to have lots of lecture notes on this particular philosopher, and as such it was easy enough to make this "review". If I will, I'll probably make a blog where I put them, as well as post them here on the League of Reason forums for criticism and discussion. In any case, here is one for arguably the most inspiring philosopher of science, Karl Popper.
Karl Popper (1902-1994) was and is hugely influential philosopher of science, especially among scientists. Popper, in his early academic years held high interest in Einstein's theory of relativity, Sigmund Freud's and Alfred Adler's psychological theories, as well as in Marxist theory of history, all of which at one time were touted as scientific by the advocates of these theories. These were all taken very seriously in Vienna, where logical positivism was founded and on which intellectual circles Popper grew up. Popper's philosophy was influenced by his disillusionment in the latter of the three theories, criticism to logical positivism and verificationism, and by the problems Hume had resurfaced for induction.
The Good
Popper realized Adler's psychological theory, in which self-struggle against feelings of inferiority was portrayed as being motivation for human actions, could be used to explain contradictory behaviour: Helping a damsel in distress was ailment to inferiority by means of elevating oneself's ego for being chivalrous, beating a woman was blamed on the same motive in confines of his theory, though realized more grimly by establishing domination and control. Freud's theory was likewise able to be made fit any observation, and marxists could always invent auxiliary explanations a posteori to save their theory, as they did to explain why revolution took place in Russia rather than in Germany - They could not have taken into account Lenin's charisma and political prowess. This analysis lead Popper to the observation that is very fundamental to his philosophy of science, verification is cheap.
Popper, then, doesn't want to say Adler, Freud or marxists did not explain observations, or that the respective theories did not fit the data well. He remains unimpressed because evidence is too easy to come by. In his view observation is theory dependent and almost anything can as such be seen to verify the theory in question, while almost nothing can be seen as falsifying it. Cleverly formed pseudoscience can be even confirmed by contradictory data, as we saw with Adler.
The difference Popper found in Einstein's theory was that it was informative in the way Adler, Freud or marxism wasn't. General relativity made a bold claim that light will bend in reaction to gravity and this prediction put the whole theory in danger, had Eddington's experiment turned out differently. In this way Popper portrays science and scientists as almost heroic, willing to put their lives work at risk for truth. The difference between pseudo-science and science, according to Popper, is that scientific practice does not aim to form theories that fit the existing data well or that explain every possible outcome, that is easy, but to formulate them as such that they reveal new information that is readily falsifiable. Real science is risky, and bold claims that run contrary to common knowledge are virtues, while conformity can be seen almost as a vice.
To sum up Popper's central finding: There's asymmetry between verification and falsification - No amount of verification will make a theory true, but even a single falsifying experiment will make the theory false.
The Bad
The first thing to notice about Popper's philosophy is that it does not define science. Falsifiability in itself is not sufficient to call theory or hypothesis scientific. The claim, "If it rains tomorrow, I am Batman", is falsifiable, but, I'm sure we'll agree, shouldn't be counted as scientific one. All scientific theories must be falsifiable, but not all falsifiable statements are scientific. This is not a big problem, or problem at all for that matter, but important enough to mention as Popper is often read as solving the demarcation criteria. He doesn't, he merely gives one necessary condition.
One problematic consideration Popper's theory faces is whether it is quality of the theory that makes it scientific, or whether theory's scientific status depends on how it is treated by it's practitioners. Behe in the Dover trial admitted that the definition of science that must be used in order to include intelligent design as sciencific theory would also include alchemy and astrology as such. This is a bit of atheist quote mining, because what Behe meant is that alchemy and astrology should be allowed in the scientific competition, not that they are likely contenders to win. Behe makes here an argument Popper would likely agree with: Alchemy and astrology in the sense he talks of them are sciences, albeit ones that have been falsified. We cannot simultaneously say that astrology, or creationism, give wrong predictions and are therefore false, and that they're unfalsiable and therefore not sciences. At least not if we are going with falsification criteria alone.
The question that then rises, is if the devout insistence, in face of piling evidence to the contrary, of the practitioners of astrology, or creationists, is condition to call the theory unscientific as long as the theory itself is falsifiable. Some astrologists and creationists do discard their respective theories when they find out the predictions do not pan out and some don't. If all scientists working on quantum mechanics only did it for the money and were unwilling to take any experiment as falsifying the theory, would we be willing to call quantum mechanics unscientific? Popper does not give a clear answer which one it is: Does scientific status depend on the theory itself or on how it's practitioners are behaving.
Another consideration is the case of sciences that are too complex to give accurate predictions. We're not going to throw medicine out the window just because there's some inherent uncertainty in the practice, nor do we do that to the theory of evolution because it can't predict precisely what sort of mutations will happen and take over in a population, or if indeed any will happen at all. How is this then different from astrologists admitting that they sometimes get the reading wrong? It may be that astrologist get their predictions wrong more often than practitioners of medicine, but this seems quite subjective evaluation. There's lots to be said on the subject, but Popper's proposed falsification criteria is not at least straight forwardly applicable to probalistic claims.
The Ugly
Good case can be made that Popper's criteria is too strict. As it stands, it is in danger to leave out genuine sciences. Advocates of copernican theory tried to measure parallax of the stars in vain, so they gave auxiliary explanation that stars were far more distant than what was thought - A hypothesis that could only be falsified centuries later. When it was observed by Boveaurd that Uranus did not follow an orbit newtonian mechanics dictated, the observation didn't lead Boveaurd to discard the theory. Instead he postulated there was another planet perturbing Uranus's orbit and some 25 years later Neptune was first observed. Likewise quantum mechanics gives wildly false predictions when it is applied to objects that are both massive and small, like blackholes or the early conditions of big bang. The problem is that inventing explanations to save a theory is not only what we observe scientists regularly doing, but fitting discordant observations by modifying the theory is the very frontier where scientific progress is made. How then, are astronomers and physicists different and better from marxists? If we were to take Popper's criteria seriously - along with his idea that conformity to observation is worthless -, and never invent auxiliary explanations in order to save a theory, may they be independently falsifiable or not, we would have no science left. This would be a devastating undertaking.
It needs to be said that Popper realized the problem himself, and instead of advocating what he called "naive falsificationism", he argued that auxiliary hypotheses are permissible as long as they are independently falsifiable. This, however, raises more problems. First, there might be infinite number of ad hoc falsifiable hypotheses that can be used to save the theory, in practice leaving a theory unfalsifiable. Second, auxiliary hypotheses cannot be tested without another set of underlying assumptions, leaving us with an infinite regress of falsifiable hypothesis. To this objection Popper says that ultimately the decision to discard a theory depends on scientists subjective evaluation of the problems the theory is facing. A compromise that seriously undermines the appeal of Popper's philosophy.
An even bigger blow to Popper comes from his insistence that a failure to falsify a theory is not evidence for the theory. When Eddington made his famous experiment confirming Einstein's predictions, Popper did not see it as evidence for general relativity. Bayesian, for example would say that the experiment highly confirmed the theory because the outcome was contrary to prior beliefs, but Popper must claim, somewhat in face of common sense, that the experiment constitutes no evidence at all. By being consistent with his principle of falsification he gives solution to Hume's problem of induction, but the cost is high as we will see. He can give an answer to Hume's problem: The claim that Sun will rise every day is scientific hypothesis on the grounds it is easily falsifiable. What he is NOT saying, is that this theory is likely to be true because it survives a test every morning. The only implication is that the theory is not dead yet. He has to deny failure to falsify a theory is a reason to think the theory is more likely to be true, because by doing that he would play into Hume's grasp. The claim that past performance of surviving batteries of tests is indicator of future success of a theory, relies on induction.
Popper, then, faces serious problem on how to explain why we should prefer corroborated theories to untested ones. Certainly we should prefer theories that have survived falsification to theories that haven't, since we already know the latter to be false, but what explanation can he give to prefer established theories over untested ones? Why aren't all theories that are not known to be false treated equally? Popper seems to give no answer to the question, other than a very unsatisfying one - That is how science is done. What he seems to be saying is that scientists should prefer tested theories because there is nothing better to go on, but certainly untested theories are no worse, if we allow his own conclusion that surviving tests counts as no reason to think the theory is more likely true. In fact untested theories could, in some unscientific sense, be better. Finding Higgs boson by reading pig intestines is certainly cheaper than the experiments going on in CERN and arguably more fun. That is not to say that it will work, but as long as that theory remains untested, and we are still taking Popper seriously, it should be taken as likely to work as quantum mechanics.
In conclusion, Popper's theory is elegant and there seems to be something profoundly right about it. It's riskiness and falsifiability, not evidential support that makes theories scientific. On a little more deeper look, it is perhaps found that the criticism of verificationism is what seemed so right about it, rather than the persuasive weight he can accumulate to support his philosophy.
Demarcation criteria should be subject to the porridge test: Whether it is too cold, too hot or just right. Popper's falsificationism seems clearly too hot. It is difficult to interpret his criteria in a way that allows physics, for example, to pass the test and marxist theory of history not to. It fails to be descriptive, as actual scientific practice does not follow the theory. And it fails to be persuasively normative for reasons outlined in the "Ugly" section. It's simple appeal comes with loads of problems, some of which seem overwhelming.