Tautological Scientific Realism

This video outlines tautological scientific realism, a version of scientific realism proposed by Carl Hoefer in his article "Scientific Realism Without the Quantum"

0:00 - Scientific realism
6:11 - Tautological scientific realism
12:29 - Viruses
19:20 - Summary
23:10 - Why not fundamental physics?
28:47 - Concerns about conceivability
39:34 - Why can't we conceive that the lore is false?

[bobcousins]

To be precise, I wouldn't say that science tells us "how the world works", but science tells us "how the world appears to us to work". In other words, science creates a model that closely matches how we perceive the universe. This allows us to create technology that works as expected. What the Universe "really is" may be unanswerable.

[channelnamegoeshere]

A thought I had that's similar to the "concerns about conceivability" section:

Yesterday I lost my car keys. I searched every corner of my bedroom for them, and cleared every spot so thoroughly that I concluded that they were not in my room. I figured the only way they could be in my room was if my eyes were not working properly (i.e. radical skepticism). I gave up, and went to get my spare keys from downstairs. As soon as I left my room, I realized that my keys were in my pocket.

In this anecdote, I was mislead by expectations in such a way that my standards for falsification were not adequate. It may seem silly to propose that an entire civilization can not realize the "keys are in there pocket", but there are historical examples of entire countries being mislead due to social influences. Tautological scientific realism says that everything's been adequately falsified beyond a doubt, but falsification is not an exact science. I don't think it's a radically skeptical view to say "there may be a hole in how we falsify what we know".

[michealcherrington]

Science says what stuff does, not what stuff is. i.e. dual-slit experiment results (quantum wave or particle) does not tell us what it is, just what it does. It does not tell us the "wave" literally becomes a "particle", it tells us what it does under experiment. Science does not provide an ultimate what or an ultimate why. Such conclusions are our responsibility.

[GoomySmash]

I think another aspect that is limiting the progression of science is the framing of our assumptions and which assumptions go unquestioned or are simply assumed to not be debatable.

For example, why do we say in chemistry class "the electrons don't ACTUALLY like the protons, they don't ACTUALLY have preferences, it's just an analogy" when such a claim is completely untestable and unverifiable either way? We have no way to understand whether anything is conscious or whether it has an experience. We don't know if certain animals are conscious or whether they experience things or if that experience is at all similar to humans, we don't know if insects are conscious, or plants, or bacteria or even electrons. Yet, it's still so inconceivable for people to say simply "we don't know if the electrons are conscious and do what they do because they have desires or whether they're governed by something else we don't understand at a fundamental level"

Instead we assert very strongly "the electrons DEFINITELY are not conscious entities that have any amount of preference or agency" even though that is essentially a religious assertion based on belief rather than empirical evidence.

That form of radical skepticism, questioning very fundamental foundational beliefs is how i think we progress in science, and the lack of questioning very basic assumptions is why we have seen little progress in fundamental understanding of physics despite knowing it's incorrect for nearly a century

[СергейМакеев-жн]

It is totally conceivable that the interconnection of the "central lore" is not as complete as it appears: as in, there actually is a "seam" between one part of science which is radically false and completely wrong-headed, and another part which is approximately true. This seam might just not be easily noticeable. It might require an impossibly accurate "map of science" to see the discrepancies.

[anthonyspencer]

Great video, Kane. I'm not going to say there is nothing to this. I found myself nodding in agreement. I guess I have three concerns. The first is that the logic of the argument seems to support more than scientific realism: there are plenty of unscientific beliefs we have about reality, which, if they were substanially mistaken would increase the likelihood of broad skepticism. So I wonder if this doesn't come down to a kind of clever way of reframing pragmatism.

The second concern is that we are left with the sore task of deciding what gets included in the sciences. It's one thing to draw up a demarcating line around the vague lines of a high school physical sciences curriculum. But it seems like things would get more difficult for philosophers who attempt to really analyze what inclusion in this set means. For example, what does confirmation look like? And what degree of confirmation is good enough? Are we as confident about the Big Bang as we are about atomic theory? Or, how about the psychosocial sciences? There is a resemblance between methods generally considered scientific, but we also recognize it is varied enough to make science just about impossible to define in a seriously satisfying way. So I am inclined to say that if you can't give that kind of satisfying epistemic account, it isn't a strong case for scientific realism.

Third, it seems like this argument would have worked for the human race at any historical state of its scientific knowledge. Who ever thinks they are so far from the truth that being radically wrong is expected? Hear me out. I'm thinking that if one rejects realism about the prevailing state of knowledge of the day, then one is just accepting the skeptical possibility. You'd have to tolerate the possibility, at any time, of being so wrong that brain-in-a-vat is on the table. If nobody has ever really thought that type of scenario is on the table, then they'e had the right reasons (in terms of conceivability) to believe the prevailing lore of the day. Forget the evidential standard: it is contingent. One always has the evidence good enough for convention at any given time.

[silverharloe]

3:21 ah, if you said "Homeric myths" instead of "Homeric gods" you could have asked, "what's the difference between electrons and Electra?" which would have earned quite a few points for style.

[NikolajKuntner]

Very interesting, I haven't heard those arguments being packaged into a framework like that.
Now as someone who both has a physics PhD and is also generally trying to tackle math constructively (indeed I combine the two even on my channel), 22:30 in (1) strikes me as just one big usage of double-negation elimination, just being put into nice prose. I'm going to make a longer mathematical tangent to drive my point home. [So I don't know how hands on you're into that corner of logic, but here's a example of the treatment. Say we have the two-element set X:={0, 1} and some proposition P choose so as to be inaccessible to us, as far as our proof calculus is concerned (say P equals the Riemann hypothesis. Or say P equals the Riemann hypothesis in conjunction with the consistency claim of our theory. I.e. P is a conjunction of 1. a statement not yet proven or rejected and 2. a statement independent of the proof calculus for a consistent and sound theory, by Gödel 2).
If our theory got a comprehension principle, we may form the subset Y:={x in X | (x=1) or P}. This Y is a subset of X and by definition contains the number 1. Moreover, X contains 0 if and only if P holds, and vice versa. Because P per design cannot be proven or rejected, the same holds for the claim that 0 is a member of X. Now for any subset S of the naturals, the mathematical induction principle can be shown to imply that S being empty is equivalent to S not having a least element. The contrapositive of this say that any non-empty set N cannot be ruled out to have a least element. In particular, since our X contains 1 and is thus non-empty, it cannot be ruled out to have a least element. A bit stronger yet, we have shown that it cannot be ruled out that either 0 or 1 is the least element of X.
Now assuming the principle of excluded middle (PEM) here, one may show that "either 0 or 1 is the least element of X". However, constructively, we cannot make the claim that "either 0 or 1 is the least element of X". As noted, constructively, we can only show that "We cannot rule out that either 0 or 1 is the least element of X".
Indeed even classically, by design we certainly cannot point to 0 and say it's the least element of X, and we can also not point at 1 and say it's the least element of 1. Due to this (due to us not being able to point to an individual), the intuitionist mathematician will argue (and I sort of agree) that there is little gained in assuming the PEM or the double-negation elimination principle (DNE). All that we really want to say, we can say constructively. And all we cannot say constructively, is because it would be elusive existence claims like the above.
A more casual example is being in a care accident and waking up in a hospital without windows. Say we have no evidence for it being day outside right now, not do we have evidence for it being night outside right now. Granting that night and day are two dual parts, we derive that "finding evidence that it's neither day nor night outside", that would be inconsistent". As this is the statement we can constructively derive and as it captures our meaning, there is not much to gain from postulating PEM. What PEM would enable us here is to derive "it's either day outside or night outside", but yeah that's not really of value. We can classically make this claim without also providing evidence that it's day, or that it's night, and we already constructively derive the inconsistency of rejecting this disjunction.
The issue with classical mechanics is that if you don't keep track of uses of PEM, then just citing a theorem of classical mathematics does not tell the one you recite it to whether the mathematical object can be exhibited. If the object can be exhibited indeed, then you can produce the strong theorem also constructively anyway. On the other hand, the Gödel-Gentzen translation of the theorem (basically the double-negation of it) can always be proven also constructively.]
So anyway, sorry for the long side comment and back to topic. Let's say we grant (2) in 22:12 in its "cannot conceive" phrasing, for all of realism, i.e. that "there is no alternative" and so we reject the idea of non-realism. I'm not finding myself compelled to go further than that - it seems that "our meaning" is captured, similar to my mathematical example above. Having ruled out the non-existence of a thing, but also not being able to witness/demonstrate/provably witness the thing, why apply DNE and make the claim it exists? I feel like adopting a double-negated form of Tautological Scientific seems compelling. Conceiving an alternative is hard, and we adopt the view that this means we won't adopt an alternative to realism. The axiomatic jump from "cannot conceive alternative" to "won't permit any alternative" is already strong and captures our meaning - end of story, we don't have to also accept realism on the nose, as this framing would always prompt us to back it up with evidence against conclusive evidence for the reality of things (some existence claims) that we cannot actually provide.
Now we can take (1) and debate the notion of "strong evidence" as opposed to mathematical "conclusive evidence." Again, I'd say given the practical nature of physical theories - its means and use to make predictions - we don't need to treat the former as if it's the latter. Strong evidence compels us to use the theory and that's what we want to get out of it. Not having to doubt virus theory is false is what captures our meaning. Virus theory has practical applications. No real gain in now adopting DNP and making the claim that virus theory holds true and they exist.

Keep it up.

[jolssoni]

I think Hoefer's argument pairs nicely with Vickers' book "Identifying future-proof science": Looking at the track record, any finding that enjoyed a ~95% consensus in the wide scientific community was never overturned. Basic physics may get rewritten every 50 years or so and our monkey brains may never grasp the final ontology assuming there is one, but the table of elements, materials engineering, basic laws of thermodynamics etc is settled science.

[hernanmurua]

It is a complex problem in the sense it is emergent so not reduced to the best, the smallest, the coherence. All is just a part, and hipereal (in Edgard.Morin's sense) can only be understood in a whole.

[aniksamiurrahman]

I guess one of the roots of the problem is the concept of "truth" or more appropriately - the concept that truth lives seperately from observation.
Probably connected to this is another thing: Philosophy of science almost always considers science detached from the the physical constrins of doing it, as well as underlying society that performs it. All the observation and testing take time, energy, effort etc. Since all of them are finite, this puts constrain on the extent on scientific endevour, and contributes to what will be seriously persued and what will not. For example, success in science gave us technological success that caused both: 1. Serious climate change 2. Severe social disparity. which in turn is visibly contributing to rise anti-scientific mentality, or even making doing science very difficult. Shouldn't those too be part of global picture of the philosophy of science?

[plasmaballin]

This arguments makes a lot of sense to me, since I've always thought that scientific anti-realists focus way too much on physics. The go-to example is Newtonian mechanics being overturned in favor of general relativity and quantum mechanics. This makes scientific anti-realism seem very plausible because we can always imagine that, whatever we think the fundamental constituents and laws of nature are, they can actually be reduced to something even more fundamental. But if we try to apply scientific anti-realism to any other domain, it suddenly looks a lot less plausible. If you try to tell me that our theories of biology are all wrong and that our bodies really work in a completely different way that just looks the same to all of our experiments, I don't know what to call you other than a radical skeptic.

It may always be impossible to say with high confidence that our theories of fundamental physics are exactly true, that is, that they really have an ontology and metaphysics that matches objective reality, but this seems to be more of a case of not being able to know the exact nature of a thing-in-itself, rather than a problem for all of science.

[GeorgWilde]

11:10 "There is no middle path between realism and radical skepticism." (aka there is no anti-realism) - Said the law of excluded middle (A or non A)


The law of excluded doesn't hold in anti-realism.

[fontenbleau]

the great irony is that this topic video are on Youtube, service which very strongly filters any slight details of harsh reality to be published on platform and in current times censoring comments almost to mindless reddit level. Youtube is like 2D very narrow bubble in 3D world.

[GeorgWilde]

Yeah. I'm agnostic even about common sense beliefs. Definitely anti-realism. I don't see the advantage of this tautological realism. It's like "i want to have strong a-priori global beliefs". And i'd rather be epistemologically humble.

[mundoheavy]

I never heard of Tautological Scientific Realism before, but based on your explanation I feel like a better name would be:
"Lack of imagination" realism or "Lack of intelectual humility" realism.

[QuantumSeanyGlass]

HMMM. I think that some information theory would be very helpful for making assertions about this topic. Scientific theory can be seen as compression on the space of known correlations. From this, for scientific theory to be wrong, 1 of 2 things must be true:
a. There is a theory that compresses known correlations better than existing scientific theory while being substantially different
b. There are a large number of correlations that have not been detected or recorded erroneously which, taken together, cause science to compress poorly
I'm fairly certain you could have a good shot at showing that a. is unlikely by measuring how well existing scientific theory compresses known correlations vs. a computed lower entropy bound. We could also have a shot at calculating how likely it is that enough correlations would be found to be incorrect to cause Science's compression to increase unacceptably above the lower entropy bound, given that correlations are always reported with a probability vs. the null hypothesis. Integrating new correlations into your calculus there seems significantly more challenging.

[garygallagher]

The 'many varieties of evidence' that put the central lore beyond doubt resonate with Hasok Chang's Operational Coherence. But I don't think Chang would endorse the 'realism' of the lore.

[ErinDonnelly-yt]

I think changes in scientific theory are just the theories becoming more precise, not more accurate.

Miasma theory accurately described the way disease can spread between people and be caused by "gross" conditions. These things continue to be true, even though now we understand the underlying microscopic factors.

[Eta_Carinae__]

Wrt foundational physics: if all the objects of physics are just treated as free variables that satisfy some number of empirical constraints, then couldn't those be considered tautologically real?