My reading of Hume and Popper is that they tend to re-create and build upon Pyrrhonism without ever giving any credit to Pyrrhonism.
Your objection to Popper has a Pyrrhonist answer. From a Pyrrhonist perspective, you are lumping two different things together in what you call "belief." One of those things is a theory - a story of causality. The other is an expectation of future events based on past experience. In Pyrrhonism, this first kind of belief is called a "dogma," and the second kind is common sense based on experience. Pyrrhonism rejects belief in dogmas and endorses common sense based on experience. Even though the Pyrrhonists identified that induction was unjustifiable - long before Hume ever did - they also granted its practicality, explicitly saying it was perfectly reasonable to infer from smoke there is fire, or that scars are formed by wounds. They even had a technical term for this kind of inference: "epilogismos."
Going back to your criticism of Popper, conjectures are theories. If believed in, they're dogmas. However, your example of experiments involving dropping a ball are subject to epilogismos, allowing it to be proper to develop an expectation that repeating these experiments will provide no new information about phenomena.
In general I'm sympathetic to broadly pragmatist, Jamesian take on belief--belief is manifested in behavior, and we should have a default skepticism about putative distinctions in belief that don't show up in behavior; we should suspect that we're dealing with merely verbal disputes.
So on the reading of Pyrrhonists where they say we shouldn't have "beliefs" about the future, but we should make plans for the future based on expectations about what will or won't likely happen--yes get on the airplane, no don't jump out the 10th story window--I find myself worrying that the sense of "beliefs" in which they're saying we shouldn't have them is one that I was never interested in the first place. I worry that Pyrrhonism so understood differs only verbally from commonsense views on which it's fine to have beliefs about the future.
As for Hume, I tend to think his reading of Pyrrhonism was the one where Pyrrhonists really can't even have expectations about the future (not just where they can't have a distinctive sort of belief in dogmas), and he thought that was a bad way to be. Here he is explicitly distinguishing himself from Pyrrhonism in the enquiry:
"But a Pyrrhonian cannot expect, that his philosophy will have any constant influence on the mind: or if it had, that its influence would be beneficial to society. On the contrary, he must acknowledge, if he will acknowledge anything, that all human life must perish, were his principles universally and steadily to prevail. All discourse, all action would immediately cease; and men remain in a total lethargy, till the necessities of nature, unsatisfied, put an end to their miserable existence."
That leaves the question of in what lesson *he* draws from his skeptical considerations. This is a big one, and certainly people differ on it. But my reading of Hume is that he's really not a skeptic. He's interested in showing that our practice of induction lacks a kind of justification it would've been nice to have, but isn't at all recommending that we give it up in light of that. As I read Hume, the big theme in his work is that lots of stuff you might have hoped to be able to give a priori, knock-down arguments for--whether in epistemology or metaphysics or ethics--just need to be treated as assumptions. But he's happy making those assumptions, and isn't advocating we try to do without them.
I don't know whose reading of Pyrrhonism you're getting the idea that Pyrrhonism says you shouldn't have beliefs about the future or that Pyrrhonism differs only verbally from common sense. There are plenty of quotes available from Sextus Empiricus that the Pyrrhonists uphold common sense.
Whether Hume misunderstood Pyrrhonism or intentionally misrepresented Pyrrhonism in order to make his view appear novel and not subject to the historical criticisms of Pyrrhonism, we'll never know. However, that famous statement of his about Pyrrhonism can be rejected as an incorrect understanding.
My concern, though, isn't about Hume. It's about your criticism of Popper.
I know Pyrrhonists were happy to talk about relying on appearances; they certainly didn't recommend the lethargy Hume worries about. But in my reading of Sextus, I find myself genuinely unclear whether he's recommending something different from what we (today--where we rely not just on common sense, but on scientific theorizing for predicting the future) usually do, or not. I see conflicting strands. Here's a kind of dilemma:
Horn 1. The Pyrrhonist is happy with the whole range of techniques that go beyond everyday common sense--statistics, scientific theorizing, etc.--for forming expectations about the future. They just say that we shouldn't believe in dogmas. On this horn of the dilemma, it's not clear that they have much of a target; we can just agree that we don't accept any of that dogmatically, but merely provisionally, we're open to being proven wrong, etc. I don't think many people want to deny this.
Horn 2. Or instead, the Pyrrhonist thinks expectations about the future can legitimately be based on common sense, but not all this systematic theorizing that goes beyond that. Is general relativity a "theory of causality" in the sense that the Pyrrhonist says we shouldn't accept, even provisionally? On this horn of the dilemma, the Pyrrhonist has a real critique of how we ordinarily go about things, but I'm going to prefer sticking with our actual methods for forming expectations about the future, rather than the stripped down ones the Pyrrhonist is OK with.
It's difficult to evaluate the two horns you present on the basis of ancient Pyrrhonism. In antiquity, there was no knowledge of statistics or of the scientific method. - the basis of horn #1. However, I would not conclude from that "it's not clear that they have much of a target; we can just agree that we don't accept any of that dogmatically, but merely provisionally, we're open to being proven wrong, etc. I don't think many people want to deny this." It seems to me that we do have a substantial target and that we disagree about accepting things provisionally. So, that classifies me as one of those few people who want to deny the point.
Regarding the target, nothing about the scientific method or about statistics substantively affects the Pyrrhonist issues with ethical dogmatizing, and it is the issue of ethics that these days has people interested in modern revivals of ancient Hellenistic philosophies of life. As for accepting things provisionally, that sounds a lot like the Academic Skeptic way of thinking about the matter. I think one can get a functionally indistinguishable Pyrrhonist approach via a different way of thinking, following Popper. If a conjecture fits with the phenomena and has fended off a lot of refutation attempts, one can hold the conjecture as open to further inquiry. It may be the only conjecture that is still standing at the time, and as such, for practical purposes, seem to be provisionally accepted.
On horn #2, again, what's normal now did not exist in antiquity. The way I see it, the scientific method - at least as described by Strevens in "The Knowledge Machine" - is an attempt to overcome the Pyrrhonist/Epicurean objections via the primacy of empirical data over any other form of reasoning. (The ancients often disregarded empirical evidence in favor of other methods of coming to a conclusion). And although Sextus unfortunately didn't come out and say so, in practice at several points outright rejects the plausibility of some theories because they are contradicted by evidence. The best example of this is in his "Against the Astrologers," but other examples include his rejections of Stoic dogmas on providence and virtue.
I'm not sure I understand what you mean to say about horn 1. At the end of the first paragraph you seem to be saying you want to deny that we should accept things, even provisionally. But at the end of the second paragraph you seem to say that one can be functionally indistinguishable from a Pyrrhonist while, at least for practical purposes, provisionally accepting conjectures that have survived many attempts at refutation. Apologies if I misread you in either case.
If both of those paragraphs can stand, as I've interpreted them, then the distinction between "accepting" and "accepting for practical purposes" must be bearing a lot of weight. That's the kind of distinction I meant to be expressing something some skepticism about by mentioning James. Roughly, the Jamesian idea is that regular old "belief" already is "acceptance for practical purposes", so there's no real ground to be occupied in saying that we shouldn't believe any claims, but should merely accept them for practical purposes.
In my second paragraph, I am carefully trying to hairsplit. Note the "seem" in the final sentence: "...as such, for practical purposes, seem to be provisionally accepted." It's not the same way of thinking that the Academic Skeptics use, but produces the same actions. If there's only one theory standing, one is going to act like it is accepted, just like one treats other things that are commonly accepted. It's like saying non-dogmatically that there are gods and that they have foreknowledge (Outlines of Pyrrhonism III.2).
Again, there's a big difference - to a Pyrrhonist - between accepting that wounds result in scars and accepting a theory about how wounds result in scars. In antiquity, there were all sorts of crazy theories that were firmly believed in because there was no agreed-upon method for refuting them. To the best of my knowledge, the ancient Pyrrhonists were not exposed to a situation in which a single theory was the lone survivor of efforts to refute it and all competing theories empirically.
I haven't read the whole exchange between you and Dan, but I'm unclear about how the distinction between ghe mere application of commonsense and going beyond it applies here. Dan's example of dropping a ball for the millionth time is an experiment done to confirm commonsense, but others are not. Nonetheless, Dan is right that Scientists eventually move on from even the most theory-laden experiments which presently mark the frontier of knowledge. Scientists were at one time super busy smashing atomic nuclei together trying to form element 113, in order to verify the going theories about atomic stability. But once they showed that it was possible to create such an element, as our theories of atomic stability predicted, we moved on. The Periodic Table now features more than 118 elements.
So Dan's point seems quite unharmed by the Pyrrhonist distinction you draw.
On Popper-hatred, there are at least two factors worth naming and emphasizing over "science envy": (1) his personal reputation for irascibility and hypocritical tyranny, to the point local students used to joke about "The Open Society by One of Its Enemies" and (2) his vociferous antagonisms to Hegel and Plato, on the basis of what quite a few philosophers thought were indefensibly shallow grounds. He developed a reputation as an impolite, blowhard hack among many philosophers, for better and for worse, deserved or undeserved.
1. What is a characteristic of disciplines where political viewpoint diversity is beneficial? I can't image, say Mathematics or Computer Science, or Chemistry research and training to be much affected by political viewpoint diversity.
2. It seems like this viewpoint diversity is transient. Should you have fair representation of federalists and anti-federalists? Of Whigs and Jacksonian Democrats? Or theists and non-theists? These seem like fashions of a time rather than enduring facts about the world.
I think it's easier to point to examples than to find some non-trivial feature they all have in common. So here's a thing that I think happens in psychology more than you might imagine. Psychologists try to show that some psychological trait that we'll all view negatively--an authoritarian personality, or racial resentment, say--is associated with some political behavior, such as supporting right-wing political parties.
But then when you look at just how the psychological trait was measured, you see that the conclusion is practically built into the instrument used to measure the trait. The authoritarian personality survey effectively asks questions about whether you support right-wing authoritarianism, and isn't designed to pick up support for left-wing authoritarianism. The racial resentment scale treats support for characteristically right-wing positions on race--e.g., colorblind neutrality--as constituting racial resentment. So what looked like empirical support for a substantive psychological hypothesis turns out to be closer to mere tautology.
I think that sort of thing would happen a lot less if, in the course of going from hypothesis, to experimental design, to data collection, to writeup and publication, a paper was more likely to encounter people unsympathetic to its conclusion.
So maybe what matters is, to the extent that the discipline attempts to find evidence for or against claims that--in our current environment--have a kind of partisan political valence, where their truth will be seen as more congenial to some political factions than others, then you want diversity with respect to those factions in the discipline.
I'd want to think more about this, but that's my knee jerk answer.
I wonder: is there a reverse parody of Mill’s trident? Imagine someone who wanted to construct a society where people are wrong and hold their beliefs for bad reasons. He might reason for the need for free speech based on the same trident, but just switching every instance of ‘wrong/unjustified’ with ‘true/justified’ (and vice versa):
People are either right (in which case we need other views to veer them off course), partially wrong (in which case we need other views to get them to full falsehood), or fully wrong (in which case we need other views for them to fully grasp bad reasons for their view).
I guess the point here only bolsters your main thesis: the symmetry breaker between these two tridents must surely be some empirical assertions about why opposing views (of certain kinds, perhaps) will more likely lead to the original trident rather than the reverse trident. Abstract argument alone is insufficient.
I take it to be a background assumption of Mill's trident that debate/inquiry tends to push us towards the truth. I know that can be contested, and it's easy to find examples of falsehoods that have plenty of buyers in the marketplace of ideas, but I still think the thought is basically right. And I think that breaks the symmetry between Mill's original trident and your reverse parody.
I think that’s right — but then I struggle to see what work the trident is doing. Once we buy in to that empirical claim (which I do), that alone seems argument enough for free speech (once we make the empirical claim more specific, perhaps studying which views should be supported and to what extent, etc.)
Fair question. I guess I think the trident is meant to block an argument for censorship that goes as follows: "sure, we needed free speech and open inquiry to find out the truth that P, but now that we've found it, we should institutionalize it, and punish people for saying not-P." I tend to agree we shouldn't be doing stuff like that, even when it comes to claims that I'm not at all uncertain about. But I don't think Mill's trident is the way to make the case.
A very similar point (perhaps just another application of the same point): one thing that has happened in recent years is a massive expansion in the range of views which are regarded as 'beyond the pale' in polite progressive company-- worthy of 'silencing' via scorn or shame or other kinds of social pressure. (I leave aside the important issue of institutional sanctions, like firing, or even prosecution in societies with less liberal free speech regimes than the U.S.). Defenders of free speech, correctly worried about this development, sometimes respond in a way that suggests they think *nothing* should be beyond the pale in this sense. But that isn't very plausible, as the most extreme cases bring out. The real question is which views (and speakers) deserve this kind of treatment. And making that assessment will require a messy weighing up of a host of different considerations, including but not limited to: (1) how likely the view in question is to be true, (2) how likely it is to make a useful contribution to the discourse, even if it isn't true, (3) the expressive value of a firm norm against expressing such views (e.g., for shoring up the status of vulnerable minorities), (4) the risks of a spill-over 'chilling effect' from such a norm, suppressing the expression of views that shouldn't be suppressed, (5) the risks of backlash, resentment, etc., against shaming people for expressing views which are in fact widely held.
I particularly would want to emphasize 5, as I think when a view is widely held in a society, silencing/shaming etc. that targets the view is pretty counterproductive, and mainly serves to reinforce social divisions between the minority who reject the view and the rest of the society in which they live. In some cases, people get this; vegans mostly (not entirely) seem to recognize that they just don't have the numbers they'd need for trying to impose social sanctions on non-vegans to be anything but deeply counterproductive. But in other cases, proponents of minority views act like they have social power that they in fact do not.
Part of the difficulty is when they *do* have social power, but only in a very limited space (e.g. elite cultural institutions). Their social pressure campaign can then seem to be proving effective, within their bubble, while generating immense backlash in broader society.
(Depending on how much they care about broader society vs exercising power within their bubble, this may or may not count as being "counterproductive" to their ends.)
While correct in the abstract, this "weighing up of a host of different considerations" just pushes the problem back one level. That is, the proponents of deeming something "beyond the pale" will have a huge pile of advocacy as to why they are right, and critically, there is a problem that disputing this advocacy itself will also be deemed "beyond the pale". Really, speaking as a free-speecher, people have gone through this in practice.
I don't think this is inconsistent with anything Jake said, but I totally agree that this sort of dynamic exists. E.g., while I admit I haven't deeply thought about it, I think the reaction to Harvey Weinstein was basically right. I don't think it's a scandal that he can't work again in Hollywood. (And I think maybe that's enough to support Jake's point, though perhaps it's not a great example because it's about conduct rather than advocacy.) But I do think it's a scandal that Ronald Sullivan had to leave his faculty dean role at Harvard as a consequence of being part of Weinstein's legal defense team.
While that's where you draw a line, I'm sure you're aware that there's arguments against it. For example (NOT MY BELIEFS, but describing just one potential response): someone could reply that Sullivan is not being imprisoned or fined, and while defendants have a right to a lawyer, nobody has to be a lawyer for a particular defendant - doing such work represents a moral choice. Further, contend that when there's such an epidemic of sexual violence, to choose to defend sexual predators rather than victims deserves sanction and shunning. Moreover, it can be claimed that as a man, you are completely unqualified to have an opinion here - because you cannot properly judge the climate of violence against women that is being enabled by actions such Sullivan's support of patriarchy, and by YOU by such apologism, etc. etc.
I could go on this is vein, but I presume this isn't a revelation to you. If you enjoy debating this, that's one thing. However, my point about "just pushes the problem back one level" is now everyone has to decided whether the above type of argument is valid itself. And then we're basically back where we started.
I'm not sure what your point is though--is the idea that we shouldn't take arguments like this seriously at all, ever? I see the appeal, but I think it's hard to defend. If I'm running a school or a summer camp, I'm not hiring somebody in NAMBLA as a teacher or counselor. And I think "cancellation" differs only by degrees from that. So I think "there should never be informal, social penalties of any kind for speech of any kind" might sound nice as a principle, but nobody really wants to put it into practice. But I may be reading more into your comments than you intended.
It's more that I urge people thinking/writing about this, to go beyond what might be called the "first level" knocking-down of very weak arguments. I encourage them to engage with the real-world level of the issues, where there are relatively sophisticated ideological argument deployed, though often phrased crudely.
"Herbert Feigl visited Popper in Penn in the summer of 1954 and described, to friends in the United States, his “splendid isolation” in Fallowfield,“an impressive country manor, with a beautiful terrace and parklike gardens.” Popper played a gracious host but complained that members of the circle had repeatedly plagiarized Logik. Feigl found him “hard to bear”: “more autistic than ever,” “paranoic,” and “megalomaniac.”"
Isn't Popper's verisimilitude a tentative to ascribe likelihood to claims?
Also, leveraging deduction rather than induction, we can make precise predictions of outcomes that follow logically a law in a diverse set of settings. Then we can engage in falsification of the law of gravity by running different experiments and varying the settings. Flying a plane should yield the same conclusions as dropping a ball. But we gain something in the process.
Would it be a law about human beings and you could run a meta-analysis, get posterior predictions, compute the likelihood of a theory because different hypothesis derived from the same law have been tested.
In the practice of science, a large set a ideas contain assumptions that are incompatible; modelling yields metrics and evidence in favour of or against; probabilities can be computed - while the notion that we cannot hold an idea for certain is preserved. There's a natural selection process for ideas/theories/paradigms/hypotheses/predictions.
Honest question: in the context of the practice of science (to the extent my above claims are valid given the differences between fields), is it necessary to have the representation of the broadest set of ideas? Also, is it optimal? Don't we reach, especially in quantitative sciences, the same epistemic value by optimizing the exploration (granted we accept the notion of likelihood)?
"Popper’s view implies we never get to be confident about what we’ll see when we rerun old experiments, so we have no reason to do new experiments."
From what I understand, Popper doesn't say we can never be confident about how reruns of old experiments turn out. Confidence is a psychological trait.
"It is therefore all the more important to be clear about the fact that his [Popper] theory is an account of the logic and history of science and not of the psychology of its practitioners." (Philosophy and the Real World, p. 27)
We can be confident that the sun will rise in the east each morning, but at the same time, we know logically that one day it will not. Our confidence level is higher now that the sun will rise, but in a few billion years, that confidence will wane. I think Popper's view is that our feelings should be removed as much as possible from scientific methodology, and that the results shouldn't be validated by feeling, but by surviving refutation.
The following passage doesn't sound like someone who thinks there's no reason to do new experiments, or that we can't have some degree of confidence in what we know. If you can't do new experiments, how would science progress? Popper's view, or is it Deutsch's, is that in our prevailing theories we will eventually find an inconsistency, which spawns a new problem, a new theory, and new experiments. Also, just because we can't be 100% sure that a rerun of an experiment will replicate the original, does it logically follow that we shouldn't design a different experiment?
"As against this, I do not think that we can ever seriously reduce by elimination, the number of the competing theories, since this number remains always infinite. What we do—or should do—is to hold on, for the time being, to the most improbable of the surviving theories or, more precisely, to the one that can be most severely tested. We tentatively 'accept' this theory—but only in the sense that we select it as worthy to be subjected to further criticism, and to the severest tests we can design.
On the positive side, we may be entitled to add that the surviving theory is the best theory—and the best tested theory—of which we know."
I definitely don't think Popper would have endorsed the idea that scientists have no reason to do novel experiments. I think he had pretty good judgment about what constituted good scientific practice. It's just that it's hard to reconcile that good judgment about scientific practice with the general methodological principles he proposed. He did have a concept of "corroboration", which was a measure of how severely a conjecture has been tested, where he seemed to be happy for scientists to use corroboration to play much the same role that the rest of us would think subjective probabilities play; ie, it's a good idea, when deciding what principles to rely on in designing bridges or airplanes, to rely on well-corroborated physical theories, rather than untested ones. But I think that's very hard to square with the idea that we don't have any reason to expect the well-corroborated theories to continue to withstand tests, which he clearly does say. (Look back at the quote I offered--he doesn't just say we shouldn't be 100% certain in our conjectures, but that we can't even regard them as probable.) Why tentatively accept a theory, why regard it as "best", if doing so doesn't involve forming some expectations about the future? And if it does involve forming expectations about the future, what's wrong with thinking of them as (subjective) probability assessments? After all, there are familiar arguments that if your guesses about the future *can't* be modeled as probability assessments, you'll make predictably bad decisions.
Here's a different example with a similar flavor--one where I don't Popper would've criticized the behavior of scientists, but I do think that behavior is very hard to reconcile with his general principles. He likes the idea of scientists sticking their necks out and making falsifiable claims. E.g., what he liked so much about Einstein, as he saw it, was if the observations hadn't come out Einstein's way in the crucial 1919 eclipse observation, general relativity would've been falsified. But actual scientific practice isn't like that, and for good reason. 19th century astronomers didn't regard Newtonian mechanics as falsified when it got the wrong prediction about the orbit of Uranus. Rather, they tried to figure out how to save the theory--exactly the kind of thing that Popper complains about--and realized that if there was another, as-yet-unobserved planet exerting a gravitational pull on Uranus, then Newtonian mechanics could make sense of the orbit. And that's how Neptune was discovered--in that case, evading falsification was exactly the right thing to do. They tried to do it again with Mercury; Astronomers spent much of the latter part of the 19th century looking for "Vulcan", a hypothetical small planet between Mercury and the sun whose existence would've explained irregularities in Mercury's orbit. In that case, it was the wrong move; general relativity could account for Mercury's orbit without needing to posit any new planets.
I think the proper lesson to draw from all that is that actual scientific practice is much messier than the Popperian picture of scientists proposing conjectures that can be conclusively refuted by a single experiment.
At the heart of the problem seems to be that Popper regarded himself as having 'solved' Hume's problem of induction. While a minority of philosophers think he is the preeminent philosopher of science of his generation, a much, much smaller minority believe this claim.
And this claim leads Popper into all kinds of trouble it seems to me.
On the plus side, Popper’s vision of the only secure knowledge being what was falsified yielded insights about theory choice. The more a theory ruled out, the bigger risks it was taking that it would be falsified by experience and, in the absence of it being falsified, the more information it gave us about how the world might be. On the other hand, especially where all knowledge must remain tentative — for it may always be falsified in some way we had not anticipated — surely for the reasons outlined in the post, science needs some general notion of what is and is not LIKELY?
Popper regarded such statements as dangerously infected with induction. He argued that all scientists knew was defined by the set of propositions that had not been falsified. Yet this leads him down some strange byways that seem to wrench the theorist’s understanding further from actual scientific reasoning, not to mention the practical reasoning by which we navigate the world.
Numerous observations of swans all of which were white might lead the scientist to hypothesise that all swans were white. We now know that such an hypothesis is wrong. But given what was known in the 17th century, it was wasn’t a bad hypothesis. But what of the hypothesis that some swans are born with platinum beaks and ruby eyelashes? Of course it’s possible. But mightn’t it be a scientific assertion to say that, from what we know of swans and nature more generally, the likelihood of their existing is vanishingly small?
However, in his determination to cleave scientific practice from induction in the debates of the 1960s, Popper left no room for the scientist’s judgement of what is probable. And how much can one really get beyond Hume’s problem of induction? If we persuade ourselves that the 1919 measurements of gravitational lensing were not a verification of Einstein’s general relativity, but only its survival in the face of Newton being falsified, induction is still smuggled in when we assume that a natural phenomenon demonstrated today will endure in the future.
This conundrum led John Worrall to compose a dialogue set on the top of the Eiffel Tower with one Popperian trying to persuade another to accompany him down the tower via the lift or the stairs. The other Popperian points out that assuming the laws of physics have not changed is quite a risk to take. Accordingly, mightn’t it be better to just gently ease oneself over the balcony and float down to the ground?
Does anyone actually use Mill's trident style arguments in contexts where you think it's blunt (viz, for supporting rather than permitting the expression of certain opinions)?
The exchange that made me want to write this post occurred on a recent episode of Glen Loury's podcast. He had Cornel West and Robbie George as guests. There was a lot of great discussion and I won't summarize it, but there was a spot where Loury asked them some hard questions about viewpoint diversity in academia, and how to reconcile the idea that viewpoint diversity as such is important with the idea of having substantive disciplinary standards--eg, economists build models. I was pretty unsatisfied with their answers--I thought they didn't squarely address the difficulties Loury was raising--for the kinds of reasons in this post.
My reading of Hume and Popper is that they tend to re-create and build upon Pyrrhonism without ever giving any credit to Pyrrhonism.
Your objection to Popper has a Pyrrhonist answer. From a Pyrrhonist perspective, you are lumping two different things together in what you call "belief." One of those things is a theory - a story of causality. The other is an expectation of future events based on past experience. In Pyrrhonism, this first kind of belief is called a "dogma," and the second kind is common sense based on experience. Pyrrhonism rejects belief in dogmas and endorses common sense based on experience. Even though the Pyrrhonists identified that induction was unjustifiable - long before Hume ever did - they also granted its practicality, explicitly saying it was perfectly reasonable to infer from smoke there is fire, or that scars are formed by wounds. They even had a technical term for this kind of inference: "epilogismos."
Going back to your criticism of Popper, conjectures are theories. If believed in, they're dogmas. However, your example of experiments involving dropping a ball are subject to epilogismos, allowing it to be proper to develop an expectation that repeating these experiments will provide no new information about phenomena.
In general I'm sympathetic to broadly pragmatist, Jamesian take on belief--belief is manifested in behavior, and we should have a default skepticism about putative distinctions in belief that don't show up in behavior; we should suspect that we're dealing with merely verbal disputes.
So on the reading of Pyrrhonists where they say we shouldn't have "beliefs" about the future, but we should make plans for the future based on expectations about what will or won't likely happen--yes get on the airplane, no don't jump out the 10th story window--I find myself worrying that the sense of "beliefs" in which they're saying we shouldn't have them is one that I was never interested in the first place. I worry that Pyrrhonism so understood differs only verbally from commonsense views on which it's fine to have beliefs about the future.
As for Hume, I tend to think his reading of Pyrrhonism was the one where Pyrrhonists really can't even have expectations about the future (not just where they can't have a distinctive sort of belief in dogmas), and he thought that was a bad way to be. Here he is explicitly distinguishing himself from Pyrrhonism in the enquiry:
"But a Pyrrhonian cannot expect, that his philosophy will have any constant influence on the mind: or if it had, that its influence would be beneficial to society. On the contrary, he must acknowledge, if he will acknowledge anything, that all human life must perish, were his principles universally and steadily to prevail. All discourse, all action would immediately cease; and men remain in a total lethargy, till the necessities of nature, unsatisfied, put an end to their miserable existence."
That leaves the question of in what lesson *he* draws from his skeptical considerations. This is a big one, and certainly people differ on it. But my reading of Hume is that he's really not a skeptic. He's interested in showing that our practice of induction lacks a kind of justification it would've been nice to have, but isn't at all recommending that we give it up in light of that. As I read Hume, the big theme in his work is that lots of stuff you might have hoped to be able to give a priori, knock-down arguments for--whether in epistemology or metaphysics or ethics--just need to be treated as assumptions. But he's happy making those assumptions, and isn't advocating we try to do without them.
I don't know whose reading of Pyrrhonism you're getting the idea that Pyrrhonism says you shouldn't have beliefs about the future or that Pyrrhonism differs only verbally from common sense. There are plenty of quotes available from Sextus Empiricus that the Pyrrhonists uphold common sense.
Whether Hume misunderstood Pyrrhonism or intentionally misrepresented Pyrrhonism in order to make his view appear novel and not subject to the historical criticisms of Pyrrhonism, we'll never know. However, that famous statement of his about Pyrrhonism can be rejected as an incorrect understanding.
My concern, though, isn't about Hume. It's about your criticism of Popper.
I know Pyrrhonists were happy to talk about relying on appearances; they certainly didn't recommend the lethargy Hume worries about. But in my reading of Sextus, I find myself genuinely unclear whether he's recommending something different from what we (today--where we rely not just on common sense, but on scientific theorizing for predicting the future) usually do, or not. I see conflicting strands. Here's a kind of dilemma:
Horn 1. The Pyrrhonist is happy with the whole range of techniques that go beyond everyday common sense--statistics, scientific theorizing, etc.--for forming expectations about the future. They just say that we shouldn't believe in dogmas. On this horn of the dilemma, it's not clear that they have much of a target; we can just agree that we don't accept any of that dogmatically, but merely provisionally, we're open to being proven wrong, etc. I don't think many people want to deny this.
Horn 2. Or instead, the Pyrrhonist thinks expectations about the future can legitimately be based on common sense, but not all this systematic theorizing that goes beyond that. Is general relativity a "theory of causality" in the sense that the Pyrrhonist says we shouldn't accept, even provisionally? On this horn of the dilemma, the Pyrrhonist has a real critique of how we ordinarily go about things, but I'm going to prefer sticking with our actual methods for forming expectations about the future, rather than the stripped down ones the Pyrrhonist is OK with.
It's difficult to evaluate the two horns you present on the basis of ancient Pyrrhonism. In antiquity, there was no knowledge of statistics or of the scientific method. - the basis of horn #1. However, I would not conclude from that "it's not clear that they have much of a target; we can just agree that we don't accept any of that dogmatically, but merely provisionally, we're open to being proven wrong, etc. I don't think many people want to deny this." It seems to me that we do have a substantial target and that we disagree about accepting things provisionally. So, that classifies me as one of those few people who want to deny the point.
Regarding the target, nothing about the scientific method or about statistics substantively affects the Pyrrhonist issues with ethical dogmatizing, and it is the issue of ethics that these days has people interested in modern revivals of ancient Hellenistic philosophies of life. As for accepting things provisionally, that sounds a lot like the Academic Skeptic way of thinking about the matter. I think one can get a functionally indistinguishable Pyrrhonist approach via a different way of thinking, following Popper. If a conjecture fits with the phenomena and has fended off a lot of refutation attempts, one can hold the conjecture as open to further inquiry. It may be the only conjecture that is still standing at the time, and as such, for practical purposes, seem to be provisionally accepted.
On horn #2, again, what's normal now did not exist in antiquity. The way I see it, the scientific method - at least as described by Strevens in "The Knowledge Machine" - is an attempt to overcome the Pyrrhonist/Epicurean objections via the primacy of empirical data over any other form of reasoning. (The ancients often disregarded empirical evidence in favor of other methods of coming to a conclusion). And although Sextus unfortunately didn't come out and say so, in practice at several points outright rejects the plausibility of some theories because they are contradicted by evidence. The best example of this is in his "Against the Astrologers," but other examples include his rejections of Stoic dogmas on providence and virtue.
I'm not sure I understand what you mean to say about horn 1. At the end of the first paragraph you seem to be saying you want to deny that we should accept things, even provisionally. But at the end of the second paragraph you seem to say that one can be functionally indistinguishable from a Pyrrhonist while, at least for practical purposes, provisionally accepting conjectures that have survived many attempts at refutation. Apologies if I misread you in either case.
If both of those paragraphs can stand, as I've interpreted them, then the distinction between "accepting" and "accepting for practical purposes" must be bearing a lot of weight. That's the kind of distinction I meant to be expressing something some skepticism about by mentioning James. Roughly, the Jamesian idea is that regular old "belief" already is "acceptance for practical purposes", so there's no real ground to be occupied in saying that we shouldn't believe any claims, but should merely accept them for practical purposes.
In my second paragraph, I am carefully trying to hairsplit. Note the "seem" in the final sentence: "...as such, for practical purposes, seem to be provisionally accepted." It's not the same way of thinking that the Academic Skeptics use, but produces the same actions. If there's only one theory standing, one is going to act like it is accepted, just like one treats other things that are commonly accepted. It's like saying non-dogmatically that there are gods and that they have foreknowledge (Outlines of Pyrrhonism III.2).
Again, there's a big difference - to a Pyrrhonist - between accepting that wounds result in scars and accepting a theory about how wounds result in scars. In antiquity, there were all sorts of crazy theories that were firmly believed in because there was no agreed-upon method for refuting them. To the best of my knowledge, the ancient Pyrrhonists were not exposed to a situation in which a single theory was the lone survivor of efforts to refute it and all competing theories empirically.
I haven't read the whole exchange between you and Dan, but I'm unclear about how the distinction between ghe mere application of commonsense and going beyond it applies here. Dan's example of dropping a ball for the millionth time is an experiment done to confirm commonsense, but others are not. Nonetheless, Dan is right that Scientists eventually move on from even the most theory-laden experiments which presently mark the frontier of knowledge. Scientists were at one time super busy smashing atomic nuclei together trying to form element 113, in order to verify the going theories about atomic stability. But once they showed that it was possible to create such an element, as our theories of atomic stability predicted, we moved on. The Periodic Table now features more than 118 elements.
So Dan's point seems quite unharmed by the Pyrrhonist distinction you draw.
Nicely presented!
On Popper-hatred, there are at least two factors worth naming and emphasizing over "science envy": (1) his personal reputation for irascibility and hypocritical tyranny, to the point local students used to joke about "The Open Society by One of Its Enemies" and (2) his vociferous antagonisms to Hegel and Plato, on the basis of what quite a few philosophers thought were indefensibly shallow grounds. He developed a reputation as an impolite, blowhard hack among many philosophers, for better and for worse, deserved or undeserved.
"The Open Society by One if Its Enemies" is hilarious--thank you for that story I'd never heard.
Happy to share! Iirc I heard it via philosopher David Corfield, who had been there to see it in person, at a conference a couple years back.
Loved reading this. The clarity of your writing is a pleasure. Thanks!
As a lay reader trying to reason about this:
1. What is a characteristic of disciplines where political viewpoint diversity is beneficial? I can't image, say Mathematics or Computer Science, or Chemistry research and training to be much affected by political viewpoint diversity.
2. It seems like this viewpoint diversity is transient. Should you have fair representation of federalists and anti-federalists? Of Whigs and Jacksonian Democrats? Or theists and non-theists? These seem like fashions of a time rather than enduring facts about the world.
Good questions.
I think it's easier to point to examples than to find some non-trivial feature they all have in common. So here's a thing that I think happens in psychology more than you might imagine. Psychologists try to show that some psychological trait that we'll all view negatively--an authoritarian personality, or racial resentment, say--is associated with some political behavior, such as supporting right-wing political parties.
But then when you look at just how the psychological trait was measured, you see that the conclusion is practically built into the instrument used to measure the trait. The authoritarian personality survey effectively asks questions about whether you support right-wing authoritarianism, and isn't designed to pick up support for left-wing authoritarianism. The racial resentment scale treats support for characteristically right-wing positions on race--e.g., colorblind neutrality--as constituting racial resentment. So what looked like empirical support for a substantive psychological hypothesis turns out to be closer to mere tautology.
I think that sort of thing would happen a lot less if, in the course of going from hypothesis, to experimental design, to data collection, to writeup and publication, a paper was more likely to encounter people unsympathetic to its conclusion.
So maybe what matters is, to the extent that the discipline attempts to find evidence for or against claims that--in our current environment--have a kind of partisan political valence, where their truth will be seen as more congenial to some political factions than others, then you want diversity with respect to those factions in the discipline.
I'd want to think more about this, but that's my knee jerk answer.
I wonder: is there a reverse parody of Mill’s trident? Imagine someone who wanted to construct a society where people are wrong and hold their beliefs for bad reasons. He might reason for the need for free speech based on the same trident, but just switching every instance of ‘wrong/unjustified’ with ‘true/justified’ (and vice versa):
People are either right (in which case we need other views to veer them off course), partially wrong (in which case we need other views to get them to full falsehood), or fully wrong (in which case we need other views for them to fully grasp bad reasons for their view).
I guess the point here only bolsters your main thesis: the symmetry breaker between these two tridents must surely be some empirical assertions about why opposing views (of certain kinds, perhaps) will more likely lead to the original trident rather than the reverse trident. Abstract argument alone is insufficient.
I take it to be a background assumption of Mill's trident that debate/inquiry tends to push us towards the truth. I know that can be contested, and it's easy to find examples of falsehoods that have plenty of buyers in the marketplace of ideas, but I still think the thought is basically right. And I think that breaks the symmetry between Mill's original trident and your reverse parody.
I think that’s right — but then I struggle to see what work the trident is doing. Once we buy in to that empirical claim (which I do), that alone seems argument enough for free speech (once we make the empirical claim more specific, perhaps studying which views should be supported and to what extent, etc.)
Fair question. I guess I think the trident is meant to block an argument for censorship that goes as follows: "sure, we needed free speech and open inquiry to find out the truth that P, but now that we've found it, we should institutionalize it, and punish people for saying not-P." I tend to agree we shouldn't be doing stuff like that, even when it comes to claims that I'm not at all uncertain about. But I don't think Mill's trident is the way to make the case.
A very similar point (perhaps just another application of the same point): one thing that has happened in recent years is a massive expansion in the range of views which are regarded as 'beyond the pale' in polite progressive company-- worthy of 'silencing' via scorn or shame or other kinds of social pressure. (I leave aside the important issue of institutional sanctions, like firing, or even prosecution in societies with less liberal free speech regimes than the U.S.). Defenders of free speech, correctly worried about this development, sometimes respond in a way that suggests they think *nothing* should be beyond the pale in this sense. But that isn't very plausible, as the most extreme cases bring out. The real question is which views (and speakers) deserve this kind of treatment. And making that assessment will require a messy weighing up of a host of different considerations, including but not limited to: (1) how likely the view in question is to be true, (2) how likely it is to make a useful contribution to the discourse, even if it isn't true, (3) the expressive value of a firm norm against expressing such views (e.g., for shoring up the status of vulnerable minorities), (4) the risks of a spill-over 'chilling effect' from such a norm, suppressing the expression of views that shouldn't be suppressed, (5) the risks of backlash, resentment, etc., against shaming people for expressing views which are in fact widely held.
Yeah this is very much how I see things.
I particularly would want to emphasize 5, as I think when a view is widely held in a society, silencing/shaming etc. that targets the view is pretty counterproductive, and mainly serves to reinforce social divisions between the minority who reject the view and the rest of the society in which they live. In some cases, people get this; vegans mostly (not entirely) seem to recognize that they just don't have the numbers they'd need for trying to impose social sanctions on non-vegans to be anything but deeply counterproductive. But in other cases, proponents of minority views act like they have social power that they in fact do not.
Part of the difficulty is when they *do* have social power, but only in a very limited space (e.g. elite cultural institutions). Their social pressure campaign can then seem to be proving effective, within their bubble, while generating immense backlash in broader society.
(Depending on how much they care about broader society vs exercising power within their bubble, this may or may not count as being "counterproductive" to their ends.)
While correct in the abstract, this "weighing up of a host of different considerations" just pushes the problem back one level. That is, the proponents of deeming something "beyond the pale" will have a huge pile of advocacy as to why they are right, and critically, there is a problem that disputing this advocacy itself will also be deemed "beyond the pale". Really, speaking as a free-speecher, people have gone through this in practice.
I don't think this is inconsistent with anything Jake said, but I totally agree that this sort of dynamic exists. E.g., while I admit I haven't deeply thought about it, I think the reaction to Harvey Weinstein was basically right. I don't think it's a scandal that he can't work again in Hollywood. (And I think maybe that's enough to support Jake's point, though perhaps it's not a great example because it's about conduct rather than advocacy.) But I do think it's a scandal that Ronald Sullivan had to leave his faculty dean role at Harvard as a consequence of being part of Weinstein's legal defense team.
While that's where you draw a line, I'm sure you're aware that there's arguments against it. For example (NOT MY BELIEFS, but describing just one potential response): someone could reply that Sullivan is not being imprisoned or fined, and while defendants have a right to a lawyer, nobody has to be a lawyer for a particular defendant - doing such work represents a moral choice. Further, contend that when there's such an epidemic of sexual violence, to choose to defend sexual predators rather than victims deserves sanction and shunning. Moreover, it can be claimed that as a man, you are completely unqualified to have an opinion here - because you cannot properly judge the climate of violence against women that is being enabled by actions such Sullivan's support of patriarchy, and by YOU by such apologism, etc. etc.
I could go on this is vein, but I presume this isn't a revelation to you. If you enjoy debating this, that's one thing. However, my point about "just pushes the problem back one level" is now everyone has to decided whether the above type of argument is valid itself. And then we're basically back where we started.
I'm not sure what your point is though--is the idea that we shouldn't take arguments like this seriously at all, ever? I see the appeal, but I think it's hard to defend. If I'm running a school or a summer camp, I'm not hiring somebody in NAMBLA as a teacher or counselor. And I think "cancellation" differs only by degrees from that. So I think "there should never be informal, social penalties of any kind for speech of any kind" might sound nice as a principle, but nobody really wants to put it into practice. But I may be reading more into your comments than you intended.
It's more that I urge people thinking/writing about this, to go beyond what might be called the "first level" knocking-down of very weak arguments. I encourage them to engage with the real-world level of the issues, where there are relatively sophisticated ideological argument deployed, though often phrased crudely.
From Hacohen's bio of Popper
"Herbert Feigl visited Popper in Penn in the summer of 1954 and described, to friends in the United States, his “splendid isolation” in Fallowfield,“an impressive country manor, with a beautiful terrace and parklike gardens.” Popper played a gracious host but complained that members of the circle had repeatedly plagiarized Logik. Feigl found him “hard to bear”: “more autistic than ever,” “paranoic,” and “megalomaniac.”"
Isn't Popper's verisimilitude a tentative to ascribe likelihood to claims?
Also, leveraging deduction rather than induction, we can make precise predictions of outcomes that follow logically a law in a diverse set of settings. Then we can engage in falsification of the law of gravity by running different experiments and varying the settings. Flying a plane should yield the same conclusions as dropping a ball. But we gain something in the process.
Would it be a law about human beings and you could run a meta-analysis, get posterior predictions, compute the likelihood of a theory because different hypothesis derived from the same law have been tested.
In the practice of science, a large set a ideas contain assumptions that are incompatible; modelling yields metrics and evidence in favour of or against; probabilities can be computed - while the notion that we cannot hold an idea for certain is preserved. There's a natural selection process for ideas/theories/paradigms/hypotheses/predictions.
Honest question: in the context of the practice of science (to the extent my above claims are valid given the differences between fields), is it necessary to have the representation of the broadest set of ideas? Also, is it optimal? Don't we reach, especially in quantitative sciences, the same epistemic value by optimizing the exploration (granted we accept the notion of likelihood)?
Is this really what Popper thought?
"Popper’s view implies we never get to be confident about what we’ll see when we rerun old experiments, so we have no reason to do new experiments."
From what I understand, Popper doesn't say we can never be confident about how reruns of old experiments turn out. Confidence is a psychological trait.
"It is therefore all the more important to be clear about the fact that his [Popper] theory is an account of the logic and history of science and not of the psychology of its practitioners." (Philosophy and the Real World, p. 27)
We can be confident that the sun will rise in the east each morning, but at the same time, we know logically that one day it will not. Our confidence level is higher now that the sun will rise, but in a few billion years, that confidence will wane. I think Popper's view is that our feelings should be removed as much as possible from scientific methodology, and that the results shouldn't be validated by feeling, but by surviving refutation.
The following passage doesn't sound like someone who thinks there's no reason to do new experiments, or that we can't have some degree of confidence in what we know. If you can't do new experiments, how would science progress? Popper's view, or is it Deutsch's, is that in our prevailing theories we will eventually find an inconsistency, which spawns a new problem, a new theory, and new experiments. Also, just because we can't be 100% sure that a rerun of an experiment will replicate the original, does it logically follow that we shouldn't design a different experiment?
"As against this, I do not think that we can ever seriously reduce by elimination, the number of the competing theories, since this number remains always infinite. What we do—or should do—is to hold on, for the time being, to the most improbable of the surviving theories or, more precisely, to the one that can be most severely tested. We tentatively 'accept' this theory—but only in the sense that we select it as worthy to be subjected to further criticism, and to the severest tests we can design.
On the positive side, we may be entitled to add that the surviving theory is the best theory—and the best tested theory—of which we know."
(Logic of Scientific Discovery, p. 419)
I definitely don't think Popper would have endorsed the idea that scientists have no reason to do novel experiments. I think he had pretty good judgment about what constituted good scientific practice. It's just that it's hard to reconcile that good judgment about scientific practice with the general methodological principles he proposed. He did have a concept of "corroboration", which was a measure of how severely a conjecture has been tested, where he seemed to be happy for scientists to use corroboration to play much the same role that the rest of us would think subjective probabilities play; ie, it's a good idea, when deciding what principles to rely on in designing bridges or airplanes, to rely on well-corroborated physical theories, rather than untested ones. But I think that's very hard to square with the idea that we don't have any reason to expect the well-corroborated theories to continue to withstand tests, which he clearly does say. (Look back at the quote I offered--he doesn't just say we shouldn't be 100% certain in our conjectures, but that we can't even regard them as probable.) Why tentatively accept a theory, why regard it as "best", if doing so doesn't involve forming some expectations about the future? And if it does involve forming expectations about the future, what's wrong with thinking of them as (subjective) probability assessments? After all, there are familiar arguments that if your guesses about the future *can't* be modeled as probability assessments, you'll make predictably bad decisions.
Here's a different example with a similar flavor--one where I don't Popper would've criticized the behavior of scientists, but I do think that behavior is very hard to reconcile with his general principles. He likes the idea of scientists sticking their necks out and making falsifiable claims. E.g., what he liked so much about Einstein, as he saw it, was if the observations hadn't come out Einstein's way in the crucial 1919 eclipse observation, general relativity would've been falsified. But actual scientific practice isn't like that, and for good reason. 19th century astronomers didn't regard Newtonian mechanics as falsified when it got the wrong prediction about the orbit of Uranus. Rather, they tried to figure out how to save the theory--exactly the kind of thing that Popper complains about--and realized that if there was another, as-yet-unobserved planet exerting a gravitational pull on Uranus, then Newtonian mechanics could make sense of the orbit. And that's how Neptune was discovered--in that case, evading falsification was exactly the right thing to do. They tried to do it again with Mercury; Astronomers spent much of the latter part of the 19th century looking for "Vulcan", a hypothetical small planet between Mercury and the sun whose existence would've explained irregularities in Mercury's orbit. In that case, it was the wrong move; general relativity could account for Mercury's orbit without needing to posit any new planets.
I think the proper lesson to draw from all that is that actual scientific practice is much messier than the Popperian picture of scientists proposing conjectures that can be conclusively refuted by a single experiment.
Thanks for the thoughtful reply!
At the heart of the problem seems to be that Popper regarded himself as having 'solved' Hume's problem of induction. While a minority of philosophers think he is the preeminent philosopher of science of his generation, a much, much smaller minority believe this claim.
And this claim leads Popper into all kinds of trouble it seems to me.
On the plus side, Popper’s vision of the only secure knowledge being what was falsified yielded insights about theory choice. The more a theory ruled out, the bigger risks it was taking that it would be falsified by experience and, in the absence of it being falsified, the more information it gave us about how the world might be. On the other hand, especially where all knowledge must remain tentative — for it may always be falsified in some way we had not anticipated — surely for the reasons outlined in the post, science needs some general notion of what is and is not LIKELY?
Popper regarded such statements as dangerously infected with induction. He argued that all scientists knew was defined by the set of propositions that had not been falsified. Yet this leads him down some strange byways that seem to wrench the theorist’s understanding further from actual scientific reasoning, not to mention the practical reasoning by which we navigate the world.
Numerous observations of swans all of which were white might lead the scientist to hypothesise that all swans were white. We now know that such an hypothesis is wrong. But given what was known in the 17th century, it was wasn’t a bad hypothesis. But what of the hypothesis that some swans are born with platinum beaks and ruby eyelashes? Of course it’s possible. But mightn’t it be a scientific assertion to say that, from what we know of swans and nature more generally, the likelihood of their existing is vanishingly small?
However, in his determination to cleave scientific practice from induction in the debates of the 1960s, Popper left no room for the scientist’s judgement of what is probable. And how much can one really get beyond Hume’s problem of induction? If we persuade ourselves that the 1919 measurements of gravitational lensing were not a verification of Einstein’s general relativity, but only its survival in the face of Newton being falsified, induction is still smuggled in when we assume that a natural phenomenon demonstrated today will endure in the future.
This conundrum led John Worrall to compose a dialogue set on the top of the Eiffel Tower with one Popperian trying to persuade another to accompany him down the tower via the lift or the stairs. The other Popperian points out that assuming the laws of physics have not changed is quite a risk to take. Accordingly, mightn’t it be better to just gently ease oneself over the balcony and float down to the ground?
https://link.springer.com/chapter/10.1007/978-94-009-2380-5_16
Does anyone actually use Mill's trident style arguments in contexts where you think it's blunt (viz, for supporting rather than permitting the expression of certain opinions)?
The exchange that made me want to write this post occurred on a recent episode of Glen Loury's podcast. He had Cornel West and Robbie George as guests. There was a lot of great discussion and I won't summarize it, but there was a spot where Loury asked them some hard questions about viewpoint diversity in academia, and how to reconcile the idea that viewpoint diversity as such is important with the idea of having substantive disciplinary standards--eg, economists build models. I was pretty unsatisfied with their answers--I thought they didn't squarely address the difficulties Loury was raising--for the kinds of reasons in this post.