We hear this all the time: “Science can’t prove anything to be true, as there’s always a possibility that we’ve made a mistake or that there are other data that we don’t yet know, so all things that science says are “true” are provisional. But science can prove things to be false!”
When thinking about that this weekend, prompted by a talk at the Imagine No Religion Meetings, I felt that that statement is too glib. Bear with me for a minute, as these are just preliminary thoughts and have probably been discussed—and answered—by philosophers of science.
But the statement above seems wrong to me, and for a simple reason: if you can be wrong about finding things true, can’t you also be wrong about finding things untrue? That is, you reject something because it doesn’t seem to fit the facts, and yet that rejection may be wrong because you didn’t know all the facts.
Take creationism. As far as people know, it’s been “proven” wrong. The earth isn’t 6000 years old, animals and plants weren’t created all at once, and there was no great flood. Because it doesn’t fit what we know, we say it’s “proven wrong.” But look at the other side: evolution. It fits every fact we know—as far as we know. But we don’t say that evolution is “proven” right: we say it’s the “best explanation we have” for the data. Why? Because something might crop up to show that evolution is wrong: a passel of Precambrian rabbits, human fossils indubitably dated with dinosaurs, and so on.
But if something might crop up to show evolution to be wrong, why couldn’t something crop up to show that our rejection of creationism is wrong? Maybe there was a Great Worldwide Flood, and we haven’t found the evidence yet. Maybe there’s some flaw in dating that we don’t know about.
Now it’s inconceivable to me that all the evidence we have against creationism is wrong, and that that theory is right. The data are simply overwhelming. But my point is this: if we can’t regard evolution as “proven true”, by what lights can we say that creationism is “proven false”?
And so it goes, it seems to me, for all of science. Nothing can be proven right or wrong in the absolute sense, though, as Anthony Grayling says, things can be proven or disproven in the vernacular sense—in the sense that you’d bet your house that evolution is true or that a molecule of benzene has six carbon atoms and six hydrogen atoms (well, I would). But sticking to the absolute sense, if nothing can be proven absolutely, how can something be disproven absolutely?
Or am I wrong?
In my view, science is just Bayesianism on steroids. So, science can ‘prove’ hypotheses to be true or false only in the sense that it renders them more or less likely given the available evidence. That’s really all we can ask of science, and indeed, it’s the best we can do.
I think we can go a little farther than that. Why? Because the world is interconnected, and certain facts—or lines of evidence, at least—impinge upon others in our vision of it.
This is why compartmentalized thinking is so endemic to religion. Certain facts about the world have to be confined to their own sphere, lest they leak into another one where they don’t fit and wreak havoc.
Of course, the dream of science is to have only one sphere, which accounts for and harmonizes all facts we come across with a few laws of Nature. This is an ancient dream, by the way—the physicist Gerald Holton calls it the Ionian Enchantment, referring to ancient Ionia and the Greek philosophers there who first tried to conceive of a purely natural order.
I think imagining the scientific pursuit as a collection of Bayes estimators, while pragmatically useful, misses this grander vision.
Or in Lord Rutherford’s hoary old chestnut ‘All science is either physics or stamp collecting’
Physics tells me about what a stamp is made of, it tells me nothing of the stamp itself – as a stamp.
But scientists incorporate those interconnections in making Bayesian judgments about hypotheses. If experiments support a hypothesis that doesn’t fit the rest of what we know about the world, that hypothesis is treated met with skepticism (eg, CERN’s superluminal neutrinos or NASA’s bacteria that can use sulfur in place of phosphorus).
That is not good enough, and in fact the usual methods are better. Each area agree on some level of quality that is useful for them.
To wit, astronomy likes to have 7 sigma for observations (lots of stars means lots of patterns), particle physicists 5 sigma (lots of energies means lots of noise bumps), physicists 3 sigma (repetitive signals), biologists 2 sigma (hard to reproduce), medical diagnoses are 80 % correct on average (individuals) and phylogenies have at least 50 % bootstrap consensus (scant data).
What I have learned from bioinformatics is that bayesian methods are a lot more powerful to handle complex problems and find models, but posterior probabilities are *way* to optimistic. By making violence on parameters you can force too much. On the other hand maximum likelihood guarantee just that ML, but are also pessimistic. By not having shortcuts you do not find the optimum models. So we compromise and use both.
Bayesian learning, which is what you end up with moving from prior to posterior, is very good at finding patterns. But I would not expect it to find the general patterns we call laws. For that we need prediction and test against that agreed on standard, to propose the next move outside the comfort zone of available evidence.
YF did refer to science as Bayesianism “on steroids.”
Stair rods?
I took that to mean (s)he is well aware that science involves more than mere Bayesian learning, and was being more metaphorical than literal.
I tend to read literal as I get hurried, but I do not get that impression even on a second read. Testing seems more central in science, millenniums of pre-scientific bayesian reasoning did not achieve much. The metaphorical for that would be that science is testing “on bayesianism”. 😉
I don’t understand why you think there is a dichotomy between “testing” and Bayesian inference. “Testing” (or as we would normally say, “conducting experiments”) is how Bayesians update probabilities of hypotheses.
Recall 2015?
https://whyevolutionistrue.wordpress.com/2015/04/16/why-i-am-not-a-bayesian/
I agree. I think the problem is that people who agree that empirical conclusions are provisional when they’re positive but not when they’re negative are simply being sloppy here. It would have to be the same — if only because claims can be phrased to be either positive OR negative.
It seems to me that the basis of science is that conclusions—either accepting or rejecting—are always provisional pending further evidence, and that is why it works. It’s a system that constantly corrects itself.
I like Sean Carroll’s approach which is that science is not about what can be proven but about what works as an explanation that fits the facts in the real practical world at the present.
We could discover a fact in the future that would be consistent with creationism, but it is pointless to consider this when determining what is a fact, and what works in the present.
I guess the fallacy is attempting to undermine current knowledge based on imaginary future nonexistent knowledge. A typical religious ploy.
I do hope no creationist is quote-mining this thread!
Oh, gosh, that won’t stop them. Antievolutionists have an almost endless ability to parse and contort texts to fit their mythologies. Whole careers have been made on it, from Duane Gish to the redoubtable Jonathan Wells.
I am not a Scientist; but, am married to one and place a lot of confidence in Science; I believe it has the capacity to both prove and disprove; and is also open to being disproved by further Science. It works.
Consider a decidable proposition P. It’s either true or false. If P is true then NOT P is false. If P is false then NOT P is true. That’s obvious. Proving something false (I’m using “proving” in the provisional sense) is no different, logically and scientifically, than proving something true.
Only in a classical logic.
Although I would quibble with Stephen’s “decidable” here (it’s usually more to do with provability than truth—surely a well formed scientific proposition is either true or false), I agree with him that one’s attitude about whether there are propositions where one is 100% certain of their truth should be exactly the same attitude when “truth” is changed to “falsehood”. And for exactly the simple reason he gives.
I think that a scientific proposition which, for example, has no universal quantifier when formulated fussily (and so not likely interesting) might very well be accepted as 100% true (or false). I think Keith gave one.
But to answer Keith on this, do you know of any serious physicist whose work employs anything but standard (classical, if you prefer) logic? So-called quantum logic never came to anything and is long dead as an attempt to understand microphysics, in particular the measurement problem.
Sorry, it was Gregory Kusnick who gave one.
I agree that quantum logic is useless in physics.
There are other nonclassical logics, though, which deny the equivalence between p and not not p.
Intuitionistic logic is likely the best known and likely also has the most supporters.
As YF pointed out, science is Bayesian inference on steroids. And the way people usually frame hypotheses, “P” typically has a much lower prior probability than “not P”. Take P=creationism, for example. It’s a very complex hypothesis, with many features that all have to be exemplified to count the hypothesis as true. Solomonoff induction would rate creationism as much less likely than its denial, prior to the collection of geologic and biological evidence.
So: Jerry is right. BUT, you can usually get a few more orders of magnitude odds ratio in favor of “Hypothesis 1 was wrong” than “Hypothesis 2 was right”, in a situation where evidence favors 2 over 1.
You are NOT wrong.
We are just less concerned with false negatives than false positives in most (not all) domains of science. If creationism turned out to be true, our rejection of it would have been a false negative. But if by declaring evolution true, we are wrong, we’ve made a type 1 error (we have a false positive). We don’t want that. It’s embarrassing. So are false negatives when the stakes are high. They would be if Hell were real. This is certainly the case in newborn screening for treatable congenital disorders. In this setting, a false negative means incorrectly saying a baby is healthy when they are sick.
So, the degree to which scientists fuss about false negatives depends on the stakes involved. We don’t want to say we found something that’s not real (false positive), but false negatives get less attention in the literature, usually.
Someone else mentioned Bayes. Yes. Keep going with that somebody.
It’s a bias.
Ah, I don’t mean Bayes is a bias. The bias statement (“It’s a bias”) was meant to be further up in my draft.
I mean it is a bias that we don’t focus as much on false negatives, but their importance depends on context.
If a theory, say General Relativity, fits all known observations and is integral to certain commonly used technologies, GPS, I believe, but we cannot say it is absolutely correct, for the general reason that something new may come up to upend it. and the specific reason that it is not quantized and most physicists suspect that is vital. On the other hand, we can reject as false any theory that does not agree with observations can we not. How can any theory of gravity or space-time/mass-energy be correct if it does not agree with know, verified observations? So I am not in agreement.
Yes but when you have no fact to prove or disprove, like “god,” then how does it even figure into the problem?
How absolutely true or positively false something is has more to do with the subject than with the truth or false of it. We know that the polio vaccine was successful in preventing the disease. We know it is false that vaccines cause autism. The likely-hood that the first statement will change is nonexistent and the second statement is slightly less so.
Are you looking at the wrong level? I don’t think you can conflate hypothesis and theory. And it feels like that might be what you are doing. You can disprove a specific defined hypothesis, or find data that supports, but does not prove the concept.
However a theory is a lot bigger, and at a practical level there may be too many hypotheses to knock down – or two new ones might spring up every time you add data. Something is “accepted” once it seems beyond any reasonable doubt, not beyond any possible unfounded objection.
I rather agree with Alric, that at some level it’s what works. How you take that “what works” idea can, of course, vary too. For example, most of the medics that I know are more concerned that a pill cures the patient than knowing the details of how the pill works. Some however want to design new pills, and for that they need to understand the underlying mechanism to whatever degree of detail is needed to make their own point. That’s the level of detail that interests me, however biology is a subset of chemistry, which is a subset of physics – so the level at which one is interested in “mechanism” is also variable.
“Maybe there was a Great Worldwide Flood, and we haven’t found the evidence yet.”
What do you mean by “worldwide”? One that covers the entire planet, so to what depth? And where did all the water (rain?) come from?
If you are talking about the flood in the Old Testament, how do you explain the continued existence of animal species in Australasia and at the Poles which would never have gone two-by-two into Noah’s Ark near the eastern end of the Mediterranean Sea? And how did the African lions and Asian tigers get on with the various herbivores when on the ark?
I was for a time working as a field geologist in Africa, looking for kimberlites. Did our survey miss some? Quite likely, because – at least in that work – one could never prove a negative.
You’ve hit the factual nail on the methodological head. False notions tend to not actually explain much of the data, or even be aware of the range of what they needed to account for (as a whole endeavor, antievolutionists, ID & YECers combined) barely bump into 10% of the available science data stream).
Two symptoms of the Flood problem: (1) creationists don’t typically look down at their feet where they are to ponder where the “Flood deposit” is. Instead they invoke (usually secondarily) iconic spots like the Grand Canyon. (2) how come so much interbedded volcanic ash in the fossil record, did the spigot flip on and off spasmodically? Andrew Snelling bumped the closest to this problem, suggesting there was more than one Flood, but has backed off mentioning it since, avoiding the dogmatic ire of the YEC movement.
You can tell who’s doing the science and who’s not by who does the work, who pays attention to as much of the data as possible, and who ends up moving forward based on the successful utility of the hypothesis. Creationism (and all beliefs that are really wrong) fail miserably there.
So in order for disproven hypotheses like Flood Geology be given fresh life, it would be necessary for the body of data that shouldn’t have existed at all if their view were true to not be there. That is not a recipe for revolution, but a requirement for erasure.
[ gif of the guy in glasses and ]
[ turtleneck with fingertips ]
[ indicating his mind is blown ]
… aaand sub.
I think you are exactly right and that many people have been led astray by Karl Popper’s writing (I’m not familiar enough with it to say whether or not they’re interpreting it correctly). Anyone who actually does science knows that experiments can increase the probability of a hypothesis being true, as well as decrease that probability, depending on the outcome. Now that we have Bayesian statistics, this understanding has a formal mathematical basis.
Popper’s writing has led to this popular notion that we should constantly be trying to disprove our hypotheses rather than support them. I don’t think that makes sense, but the Bayesian baby in the bathwater is that an experiment must have the ability to undermine a hypothesis for it to also have the ability to support it. So while the word “falsify” is much too strong of a term, an experiment must be capable of reducing the probability of the hypothesis being true, or else it isn’t informative.
Also note that one can have a confirmationist approach and *not* be a Bayesian – Haack and Bunge, for example, are such.
“But my point is this: if we can’t regard evolution as “proven true”, by what lights can we say that creationism is “proven false”?”
We can’t and I’m ok with that.
“Creationism” as described here comes from one book written by a tribe in a small country in a very small part of the world. Why should Jewish creationism be any more true than other creation stories from as far apart as Australia and North America? Because it was written down and became the forerunner of Christianity that then spread the story worldwide?
Not defending any supernatural creation stories. Indeed, I believe all to be false. But I believe that because all the evidence contradicts them, not because there is some ultimate scientific way of determining they are in fact false.
What I meant is that all knowledge and conclusions about the natural world are provisional. Some are better supported than others (some, like the various creation stories, have no support at all) but just as none can be proved true, none can be proved false.
I rather like Gould’s admonition; “In science, ‘fact’ can only mean ‘confirmed to such a degree that it would be perverse to withhold provisional assent.'”
I think humanity has proven to itself that we never really know anything to be absolute.
I would amend that to “should have by now proven to itself.” The majority of people alive today do not understand, or even care to understand that the accuracy of knowledge is a spectrum of probabilities rather than absolute true or false.
I like the “should have”
We can’t, with absolute certainty, but the same applies to the ‘theory’ that we are merely brains in bottles tormented by mischievous imps, and I’d rank creationism closer to that ‘theory’ than evolution.
Yes, if the “tentative ” nature of the claim involves a science fiction or fantasy scenario which would hypothetically invoke brain-in-a-vat, the Matrix, Q, or “it’s all been a dream and you wake up to a radically different reality” — then those scare quotes aren’t in pencil, they’re in permanent marker.
On a scale of 1 to 10…
😊
Are there not two different issues being discussed here? One is whether a new observation might prove a theory wrong e.g. the discovery of Precambrian rabbits; the other is whether an observation might be wrong e.g. a flaw in the dating.
As Woody Allen once wrote as a humorous description of a college epistemology course, “Is knowledge knowable? If not, how do we know this?”
The Buddhist philosopher Nagarjuna (around 100 CE) had a similar thought:
The nature of things is to have no nature; it is their non-nature that is their nature. For they have only one nature: no-nature.
Regarding our old friend creationism, it isn’t necessary for anybody to know it is false since its proponents still need to agree on what actually suggests that divine creation is even possible.
To chase a tail why try to “prove” something false when nobody has yet “proved” it is true.
Perhaps we could use the Stephen Jay Gould quote you shared at INR7 and say that “we consider something to be false if it is unsupported by evidence to the extent that it would be perverse to withhold provisional rejection”.
I challenge the claim that all findings of science (positive or negative) are necessarily provisional.
Consider the hypothesis that there’s an exact full-scale replica of the Taj Mahal sitting out in plain view under the stars at the north pole of the asteroid Juno. That’s a question we can answer definitively one way or the other simply by sending a spacecraft to look. After thoroughly mapping the asteroid and concluding that the hypothesis is false and there’s no such artifact there, what conceivable new evidence could change our minds, without also overturning the entire scientific worldview (including the notion that evidence matters)?
This is not absolute proof in the sense that it remains logically possible that we’re insane, or that a trickster demon is systematically deceiving our instruments, or that reality itself is an illusion. But we’re no longer doing science when we entertain such possibilities.
+1
Axiomatic truth values and empirical facts are different things. We can only test or take consensus in some statistical sense.
But when the uncertainty becomes really small I would not worry about the theoretical difference. We can be safely right in claiming that a water molecule is one oxygen and two hydrogen atoms, and safely wrong in claiming that is is two oxygen and one hydrogen atoms.
Given that hydrogen and oxygen are a highly combustible combination of you get the proportions wrong I’d disagree.
Reasoning By Explosions?
http://www.sluggy.com/comics/archives/daily/20090731
I’d actually say we are *more* right when we say that water is dihyrogen monoxide rather than (pace Dalton) hydrogen monoxide or hydrogen dioxide. (Why? In part, the dissociation – 10^-7 mol / L at 25 C and 1 atm pressure.)
I think science can prove the Sun goes round the Earth is wrong. And that the Moon is not made of cheese.
In fact what Einstein showed is that there is no right or wrong answer regarding whether the Sun goes around the Earth or vice versa. All frames of reference are equally valid, and the theory yields correct predictions no matter which frame you choose.
I see that one a lot. The Sun going around the earth IS wrong. It is not in any way shape or form equivalent to the Earth going around the Sun, as a very simple analysis of the energies involved in both systems will show. “All frames of reference are equally valid” is a word-mangling of what either special or general relativity shows through its mathematics. To say they are equivalent violates the more general principle of mass-energy equivalence, upon which all relativistic theories are predicated.
I didn’t use the word “equivalent”, and I’m not sure what you mean by it. What I mean is what I said, that you can derive correct predictions in any frame. Are you saying that’s not true?
Any inertial frame is equivalent (equally valid, etc.) to any other. The key is “inertial”. The earth orbiting the sun is less “inertial” (not accelerating) than the Sun orbiting the centre of the Galaxy and neither are completely free of acceleration, so they are not equivalent in the sense that Einstein defined his reference frames.
But, wait. If the Sun goes around the Earth wouldn’t that be inconsistent with other physics? If we drop an apple does the Earth jump up to meet it? Wouldn’t we all get whiplash?
Why would we get whiplash? Whatever geometry moves the Earth in the apple-centric frame would move us along with it.
Did you have your coffee this morning?
To quote Sean Carroll:
“In particular, gravity in the Solar System is extremely well described as “flat spacetime (as in special relativity) plus a small perturbation.” From this perspective, we can very well define inertial frames in the flat background spacetime on top of which gravity is a tiny perturbation. And in those frames, it’s the Sun that is basically stationary and the Earth that is truly moving. So even the most highly sensitive general-relativists would not complain if you said that the Earth moved around the Sun, unless they hadn’t yet had their coffee that morning and were feeling especially confrontational.”
😁
OK, I’ll go get some coffee.
🙂
This is actually wrong from what I’ve been told by philosophers of general relativity and physicists – I’ll defer to any here for details.
I think it depends on what you mean by proof. If you mean absolute mathematical proof, the answer is that it is not proven.
On the other hand, the probability that the Earth goes round the Sun (or more accurately the common centre of mass of the Earth-Sun system to a reasonable approximation) is so high that there’s not really any point in assuming the contrary.
Science produces models based on assumptions. Like Newton’s three laws. You then explain natural phenomena with it. If your model does the most with the least you win. The heliocentric model made an assumption about the sun being the center of the solar system. It simplified things a great deal. Far better than any competing model.
When you start looking at systems with very large masses and high velocities the Newtonian model starts to fall apart. Then a new model is needed.
And so on.
Is there truth in this? Don’t think so. Just better models based on better assumptions.
Right. The business of science is to find better and better approximations to some ever-receding ultimate truth which may or may not actually exist. The late, great Douglas Adams made a funny but ultimately pretty philosophically sound observation regarding this conundrum: “There is a theory which states that if ever anyone discovers exactly what the Universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarre and inexplicable.
There is another theory which states that this has already happened.” 🙂
Proof is very paradoxical. Experimental or laboratory proof in science obeys physical laws of nature so what’s to say you are wrong when you mess up the experiment and get the desired results. Absolute proof or disproof is a two-way street, every time you walk on it you are never certain which way you’ll take. How do we know all this evolution evidence isn’t anecdotal and supportive of another theory of our existence? Its like Schrödinger’s cat. Its one way or the other. True or false. I would say its quadratic, having more than one solution
Are there at least a few creationists/religiosos out there who with oceanic hypocrisy presume to accuse scientists of wishful thinking? Where is the scientist who claims the truth of something based on merely and solely thinking or believing it, or who says, “I KNOW,” without offering a scintilla of evidence?
I belong to the if-it’s-useful school of thought. It doesn’t matter which is true. Evolution makes testable predictions. Creation, while it may provide after-the-fact explanations, doesn’t make testable predictions. Therefore the former is useful while the latter isn’t.
So, are religions useful?
Some people seem to find them so. I have no problem with them as long as they are personal beliefs that don’t infringe on anyone else’s rights.
This is not true for any given creationist hypothesis. For example, the YEC hypothesis as espoused in the Bible predicts we will never find any object older than about 6,000 years. It also predicts evidence of a massive flood about 4,000 years ago.
It’s the observation of phenomena that contradict these predictions that led the first real geologists to reject the idea of a young Earth and embrace geological time.
Some things are surely falsifiable? The hypothesis – all swans are white, is proven to be false by the existence of black swans.
Why think if something is true or false?, just like a jigsaw it either fits or it doesn’t.
I think existance is like this, there is only one way to connect the pieces.
Of course it is possible to accept a false theory/reject a true theory (Errors type I and II, respectively https://en.wikipedia.org/wiki/Type_I_and_type_II_errors ). There’s a big historical example of this sort in the history of the cure for scurvy – ignorance in placing citrus juice in copper/reactive vessels led people to disbelieve its effectiveness. http://idlewords.com/2010/03/scott_and_scurvy.htm
That said, these sorts of reversals tend to happen out near the edges of science, not in the core of science that has been settled for hundreds of years. Even relativity and quantum mechanics, revolutionary as they were to hundreds of years old Newtonian mechanics, have to reduce to ordinary Newtonian mechanics in the realm where Newtonian mechanics had already been shown to work.
What I think the statement about an inability to prove things right, but having the ability to prove things wrong, is trying to get at is the reality of fallibility in a world that has, or seems to have, infinite resolution. Our vision has fuzziness to it, so we cannot tell the difference between the One Truth, and all of the infinite Not Quite Truths that crowd around it. We can, however, tell when something is way off of the mark.
Granted, we could always discover a conscious conspiracy of deception, but that gets into brain in a vat solipsism, and isn’t particularly useful to speculate about.
Unfortunately, it is very easy for science to prove that this is false: Gal Gadot is currently a guest in my home.
It would be a wonder if that woman were.
/@
oh no you di-int.
I predict that she did not have you tied up with the lariat of truth when you made that statement.
If she had, it would have been true!
/@
So, which theory of truth (there are lots, unfortunately) does it use? 😉
The one …
😍
Sorry. What was the question again?
/@
Theories of truth often differ on how they handle “liar sentences”. So: does the lariat require you to state the truth, or does it require you to state something that is not false? Etc. (See above about non-classical logics.)
Sorry, Keith; I was just being facetious.
(♥ Gal Gadot.)
To state the truth, I think, only flimsy data from the comics.
/@
^ based on ^
Yes, I realize – but it might be a teachable moment for some enterprising logic instructor if they want to illustrate non-classical logics of various sorts.
Here we go.
“Proof” narrowly understood is something someone does formally in mathematics or logic.
Scientists don’t generally “prove” things, they measure things and construct hypotheses for what they measure. Sure, mathematical physicists “prove” things in mathematics, but the proofs don’t “prove” anything about the real world, at best, they provide decent models that allow us to predict empirical phenomenon (sometimes). Such as Einstein’s General Relativity predicting the precession of the perihelion of Mercury.
However, to the extent in the vernacular we speak of scientists “proving” things, it necessarily follows that establishing any negative fact is logically connected to the establishment of a positive fact.
If we prove that there was not two or more shooters involved in the JFK assassination, then the only logical possibility is that there was one shooter. If we know it is precipitating and we rule out snow, then we know it is raining.
Apparently you’ve never been to Seattle, where there are more kinds of precipitation in heaven and earth than are dreamt of in your philosophy.
Proof is for mathematicians and alcoholics.
The kind of evidence which a jury takes as proof of the story of how a crime was committed and by whom is a kind of proof which science exceeds in those theories which have gained the provisional assent of the consensus of scientists in a given field. Mathematical, completely incontrovertible proof only applies to the abstractions which comprise mathematics. IMHO, exceeding the rigor of legal “proof” is good enough to convince most rational people to give provisional assent to a proposition or to reject a hypothesis. We’ve come a very long way in the 400 years or so during which this mindset prevailed among the practitioners of the sciences. The (non-mathematical) proof of the value of this approach is in the real world results it has wrought. From hand-held devices granting access to all the accumulated wisdom (and sadly all the bullshit) of mankind to anyone for the equivalent of a couple of hours of paid work to space probes, and medical advances which have raised the average human lifespan from < 30 years to over 80 in those countries who've made the best use of those advances.
In the UK the level of “beyond reasonable doubt” is the norm for any jury’s decision to convict a defendant. But that is based on the evidence presented to the court. And some evidence may not be presented, because it does not suit the defence or the prosecution, because it was not available at the time, or because the court can ban its presentation. Thus conviction is not the same thing as guilt, leading to wrong convictions and occasional wrong executions. Equally, failing to convict is not the same thing as innocence.
As a layman I always considered that proofs are for mathematicians and science deals with levels of confidence.
If a scientific theory (like evolution) has enough (by some measure of enough) evidence to support the claims than you can declare that the theory explains the phenomenon, or put another way the theory is true (e.g. WEIT).
Is this naive?
Next edition of WEIT–“Why Evolution Is Almost Certainly True.”
😉
“Or am I wrong?”
No, you’re absolutely right. Another reason we might not be able to disprove something is that the phenomenon may not amenable to the tools we have at our disposal, including human reason.
This issue was at the core of the famous feud between Galileo and his old friend, Maffeo Barberini, aka Urban VIII (they really were friends; Galileo wrote a poem in praise of Barberini before he became pope). Contrary to legend, Urban VIII, educated by the Jesuits, was untypically open-minded for a pope. Like many Catholic scholars of the time, he was inclined to agree with Galileo’s heliocentric model as opposed to the Ptolemaic, Aristotelian, or Tychonic models. Hence, he gave Galileo permission to disseminate his views publicly, but only under the condition that he present his model as being ex hypothesi—as a theory rather than as fact. The Pope’s position was that we can’t “necessitate God”—that is, just as God might have created the universe so that the sun only appears to revolve around the earth, he could have created it so that the earth only appears to revolve around the sun. In short, we mere mortals have no way to prove or disprove anything with absolute certainty.
Galieo complied with this condition in his Dialogue Concerning the Two Chief World Systems, published in 1632. But, confident that he was right, he put Urban’s caveat in the mouth of the character Simplicio, who, as the name suggests, was an idiot. Predictably, the Pope was not amused, and lowered the boom. Moral: you don’t mess with the Pope.
Two ironies here: 1) by allowing for human error, the Pope was being more “scientific” than Galileo; 2) the historical antagonism between science and religion began as a personal feud between two strong-willed—not to say “pig-headed”—men, one of whom just happened to have more power than the other. Go figure.
Gary
Have you read the Dialogue? Simplicio is not portrayed as an idiot at all, just a dogmatist, which is something else. (And does shade, sometimes, in some people, as an idiot.) And it *is* also a reference to Simplicius, the great commentator on Aristotle.
As for the pope’s position, I’m not sure, but Galileo’s direct opponents were what would today be called some variety of “anti-realist”, and like today’s (e.g., van Fraassen) they are selectively such. (To the point of hypocrisy, but that’s another story.) There’s a good analysis of this in the Cambridge Companion to Galileo and the Wooton biography, which I was surprised to see was good on the philosophy of science. (It also agrees with my conclusion that Galileo was not a Christian.)
Yes, I’ve read the Dialogue, though not in the original Italian. I’ve read Wooten’s biography as well, and found it interesting though not as interesting as Redondi’s Galileo: Heretic. Wooten took a lot of flak for his portrayal of Galileo as a non-Christian, but I agree with you. Like nearly all Renaissance natural philosophers before Descartes, Galileo believed in an anima mundi, but was certainly not the devout Catholic that the Church has tried to make him out.
As for Simplicio being an “idiot,” I was using shorthand. That he harkens back to “Simplicius, the great commentator on Aristotle” is hardly homage on Galileo’s part. Galileo had no use for Aristotle except as a whipping boy. (I’ve read Arisotle as well, and in the original Greek—not by choice: the Jesuits made me do it!)
You’ve got me on Aristotle in the original – I’d love to do so but I have so many other projects.
I’m not sure you’re right about Galileo and Aristotle – I think in the letters on sunspots or perhaps The Assayer he says that he has no trouble with Aristotle, just the overzealousness of his contemporary followers. He says basically that you’re insulting the memory of a genius by saying that he couldn’t possibly have changed his mind on anything, and what human (especially a pagan) could possibly have gotten everything right to begin with?
I’ve read The Assayer but don’t recall any defense of Aristotle, so will have to take your word for that. It surprises me, however, that Galileo would acknowledge that anyone other than himself was a “genius” or that he hadn’t got everything right. I’ll try to find the passage you refer to.
Those works (not the Dialogue and Discourse) are much more “humble”, oddly. I am not sure why. I also wonder sometimes if it is correct to assume that Salviati really does “speak for Galileo” completely.
I agree with your main point, Jerry. It’s why I ultimately call myself an “agnostic atheist.” I am 99.999% sure that god does not exist (nor any other higher power). I don’t believe in any god or higher power because I think this is the explanation that best fits the data, but I leave that .0001% of “doubt” as a conditional statement that I could be wrong for various reasons: I do not have all the data, I have misinterpreted the data, I and all other human beings do not have sufficient brainpower/senses/whatever to know of the data that proves a higher power, etc.
“agnostic atheist.”? At least there is some positivity in there. Better than an apathetic agnostic, I guess. Their motto; “don’t know, don’t care”.
It’s easier to say than “atheist who practices epistemic humility.” I have to admit that I can’t be 100% certain about the nonexistence of a higher power, as I feel it’s my duty to intellectual honesty to do so.
You’re in good company;
“I can’t be sure God does not exist… On a scale of seven, where one means I know he exists, and seven I know he doesn’t, I call myself a six… That doesn’t mean I’m absolutely confident, that I absolutely know, because I don’t.”
-Richard Dawkins
Huh, I’ve never seen that quote. OK, I’m as smart as Richard Dawkins, everyone! Bow to me.
6 on a scale of 7 seems quite low for me.
I think he later amended his position to “6.9”.
/@
It seems to me that one can be less than 100% certain about, say, the existence of the Higgs boson without thereby assuming a duty to self-identify as an “agnostic Higgsian”.
Epistemic humility is fine, but I don’t see why the nonexistence of a higher power should be more deserving of such humility than other propositions about the natural world.
Well, let’s take your example for…um, example. We have what appears to be conclusive data about the Higgs Boson and its existence. We don’t actually have any conclusive data about the nonexistence (or, on the other hand, existence) of a higher power. That’s a decision you have to make on your own: does the idea of a higher power fit the data I have available to me, or doesn’t it? But since none of that data is *directly* associated with the existence or nonexistence of a higher power, you can’t really treat is the same as something measurable or detectable, like the Higgs Boson, or something that forms a framework to take into account all data points, like evolution.
How about Klingons then? We have no conclusive evidence of their nonexistence. Do we have a duty then to declare ourselves “agnostic aKlingonists”?
Or perhaps the clearly fictional nature of Klingons counts as conclusive evidence of their nonexistence. But if so, why doesn’t the same consideration apply to notions of gods, higher powers, and supreme beings, which have their origins in myth and legend? Why does ancient fiction call for more epistemic humility than modern fiction?
“Why does ancient fiction call for more epistemic humility than modern fiction?”
Because babies and young children are not taught from birth that klingons exist and must be worshipped. Because no one turns to klingons in time of disaster and war for mental support. Because no one expects the klingons to provide answers to the many unknowns. Because the klingons have yet to establish power networks which require lifelong obedience and vast amounts of money to maintain.
Eric, those are all reasons why entrenched myths often get the benefit of the doubt even when they don’t deserve it. I’m trying to understand why BJ apparently thinks they deserve it.
Ask BJ?
And in addition to everything Eric Hayman said: because we have *conclusive evidence* that Klingons are modern-day fictional characters *created by a single person.*
Moreover, *Klingons wouldn’t be a sufficient answer to multiple questions about life and why and how the universe exists.* The existence of god is, in fact, a sufficient answer, regardless of whether you or I personally believe it.
You’re just throwing out random examples of fictions at me without answering any of the points I’m making now.
“The existence of god is, in fact, a sufficient answer”
Not at all; it’s an evasion. It explains nothing about ultimate origins, but just pushes the questions back a level, and then erects a barrier to further inquiry.
‘Nothing can be proven right or wrong in the absolute sense…’
What does ‘absolute sense’ mean? To what does it refer? I think the problem is that we use language in, as you have said, a rather glib way.
I think it means exactly what it says, being 100% certainty, or “absolute” certainty.
I’m not a scientist. I’m a scientist wannabe; but at my age (77), I’m not very likely to ever be one. Just call me an informed science enthusiast.
It seems to me that proof is always too strong a word, for two reasons:
1. Something may appear to you to be absolute proof, but I don’t see it. Then is it proven or not? Answer: It depends.
2. Even in the same mind, what one see as absolute proof today may e seen as error tomorrow. Either in the face of additional evidence, or just rethinking the old evidence.
Nevertheless, certain things are backed by such overwhelming evidence that they are almost certainly true. A water molecule will always contain an atom of oxygen and two atoms of hydrogen. We will never again think he sun orbits the earth. Even in the face of no real evidence, all sane people expect the sun to ‘come up’ tomorrow morning.
Could we be wrong about any of these things? I don’t think so. On the other hand, I can’t imagine anything that would prove to me the god of the Bile exists. But my brother can’t think of anything that might prove to him His God is a myth.
Proof is a subjective thing. All we can ever have is evidence, no matter how much evidence we might have. But that’ OK.
Even if someone could convince me that there is a god, it’s very hard to accept that he’s a dictator.
Fortunately, the Bible was not written so cleverly that you need a biology PhD to be a skeptic. One would think that common sense is enough (though it’s not so common at all).
The concept of falsifiability is a narrow one, but useful nonetheless. Hypotheses lead to predictions. If Hypothesis A must logically lead to prediction B, and prediction B fails in experiment, then, logically,there must be something wrong with, either Hyp. A, or the logical underpinning of Pred. B. If the logic is sound, then the hypothesis, as stated, must be false.
Any expansion of the concept of falsifiability is treading on dangerous ground.
Hi jackwildeadventures. The problem is that scientific hypotheses are never tested in isolation. They are tested in conjunction with what are called auxiliary hypotheses and background assumptions. These include environmental conditions and the theories underpinning the use of test instruments. So when a scientific theory is “falsified”, it could be the case that the theory is true while one or more of the auxiliary hypotheses/background assumptions require revision. A classic case in the history of science was how the recalcitrant orbit of Neptune initially seemed to “disprove” Newton’s law of gravitation. For more on testing theories, see my comment below on the Duhem-Quine underdeterminacy thesis.
I have been thinking the same thing. Even when an experiment seems to disprove a hypothesis, it COULD still be the case that the hypothesis is correct, but some confounding factor just happened to make it seem incorrect. I agree that both proof and disproof are all about probabilities.
perhaps the only absolute objective truth is that there is no absolute objective truth in which case all we can hope to achieve is a narrowing of the margins of error
My thoughts – could this be overthink? For simplicity reduce “science” to a hypothesis you want to test using data you collect. The notion “science can proves something to be false” just means that if the data don’t support your hypothesis, you can rule it out, end of story. If the data do support your hypothesis, it *may* be true. But on that side, there may be other hypotheses also consistent with the data, and it becomes a judgement call, perhaps requiring more data, to break the “tie.” Sean Carroll frames this in a more nuanced way in the form of Baysian “credences,” which allow us to assign more or less confidence in a given hypothesis as more information is obtained. None of this addresses the problem that the *data* (or their interpretation) might be flawed somehow. In that sense you could still claim science can’t prove an idea wrong. In some sciences, data ambiguities may be a systemic issue; in others, not so much. But as a philosophical statement about science it seems to have more in common with the brain the vat objection to there being a “real” world – maybe not deserving of too much of your energy.
I think that assuming that science can’t prove anything to be true is to say that there aren’t any absolute truths in the Universe. There are, and likely will continue to be, situations where science gets it exactly right. That’s not to say that we shouldn’t continue to question and scrutinize the results of scientific inquiry. But there must surely be times when the absolute correct answer has been found. Especially if repeated testing bears that out.
….This also applies to disproving certain things. As long as sufficient data exist to do so, i’d say it could be done. I don’t suppose we’ll ever disprove the existence of a supreme being/force in the Universe. But who knows…given the time and desire to do so we may someday know all the secrets of the Universe!
Yes, Jerry, you are right. This has been looked at in vast amounts of detail in the philosophy of science literature. The quandary you pose is best encapsulated in what is known as the Duhem-Quine underdeterminacy thesis. Here is a good acount of the issues in the SEP Encyclopedia >
https://plato.stanford.edu/entries/scientific-underdetermination/
Thanks for this comment. Thanks PCC(E) for the question!
In my view sometimes one has underdetermination, but what D-Q fail to notice is precisely *how* background knowledge works. Quine almost gets it, with the “web of belief”, though.
thanks!
It seems to me that “proof” of a theory or its “truth” still lies on the spectrum of probability. As has been noted many times, there is only one true or correct answer (perhaps here and at this point in time as Carroll argues) but there are so many more wrong ones. This sets up an asymmetry that it’s far, far more probable that a wrong answer to a proposition is being proferred than the right answer. Thus, one is far more likely to find any single explanation to be untrue than to be true. Example: Lincoln was born in 1809: true.
Lincoln was born in 1835: untrue.
Lincoln was not born in 1809: untrue
Lincoln was not born in every year that isn’t 1809 (e.g. 1975): all true.
I think I just restated Jerry’s thesis.
Ugh, sleepy time!
In principle, yes. In practice, I don’t think there is parity between the two, and I think scientific research will often lead to a higher confidence that some alternate theory is false than the prevailing theory being true.
The disparity comes from the “facts on the table.” For a wrong theory to be somewhat accurate is conceptually easy – that is Hume’s classic problem of induction. And we see that all the time in science, whether it’s the ideal gas law or Newtonian mechanics or what have you. However for a right theory to be inconsistent with the facts…that’s much harder to explain in a convincing manner. How could that be the case? Well, the humans who collected the data might be biased. Or they might be making a systemic error. Theists could also claim Omphalos/last-Thursdayism as a reason, or claim Satan made things look like some wrong theory is right. But none of those are very appealing to most people, and bias and systemic errors are overcome by repeating the measurements using different people and different (orthogonal) data collection methods. So it seems to me that as more and more repeat measurements of fact X are collected, by more different people, our confidence that a theory that predicts not-X is wrong will rationally go up, and up, and up, ad infinitum. But our confidence that a theory that predicts X is right does not necessarily go up ad infinitum.
In short, IMO it’s possible to be wrong both ways, but once you have a lot of decent quality data its much easier to be wrong about a theory being right than it is to be wrong about a theory being wrong. 🙂
Well said.
The Wikipedia pages on falsification are probably quite close to how I would answer this question.
I think I would summarize my view as saying that the thing that separates science from a hobby (such as stamp collecting, perhaps?) is that it makes testable predictions.
Of course, it’s hard to positively confirm a single theory, because multiple other theories (including unknown ones) will make the same prediction. I think that’s why falsification is needed, as you can rule out specific options. The loopholes are usually in the assumptions you put into your test.
You can sort-of squeeze religion into this category: If there were a god, perhaps you’d expect prayer to be effective, or “miracles” to happen, a booming voice from the sky explaining the error of your ways, etc.?
It’s also fun to apply to some areas of science: What are the testable predictions? (Post-hoc rationalizations aside).
The volvox leaves no fossil
but maybe solves the puzzle
why we seem to die physically
and have a determined, sexed,
Mortality, trading chromosomes
as investments in Life & Love for
Sunlight & Moonshine & Star-bright &
Death-right, with an open eye for opulent,
Brilliance, leading inward and onward
as if time and gravity, were secondary
to celestial, cyclical mobility! How it feeds
with crudely elegant efficiency and did,
Lead us to the apple, and civilized society.
Now our tamed sapian has grown so dignified
so consciously self-deified & satisfied with
his modern meal of bigotry, he so forgets what
Edified us all, Yes! God holds his breath in envy,
is covetous of our transient being, thus, ponder
how the timeless, boundless cannot improve but
wee small mere corporeal mortals evolve & persist
into that radiant 4th dimension of growth beyond breadth.
Jerry, you are correct. Because the universe is infinite, nothing can be completely proven or disproven. However, we can assume that the preponderance of evidence favors one or the other. For instance, we can assume that there are material causes for all effects, but we can never prove that completely until we discover the causes for all effects, which is impossible. The opposing assumption that there are no material causes for effects cannot be completely proven (or disproven) either. As scientists we simply choose which ever assumption works for us. Once we choose determinism as our guiding assumption, we do not need to be agnostic about it.
This is the same question as proving the existence of a god. I can think of a god but it would defy the laws of physics. It would seeem that nothing could completely falsified.
I can’t help thinking that using terms like ‘true’ and ‘false’ and ‘proof’ are misleading. They are philosophical terms but science is about the *descriptions* we assign to observed regularities. If there were no humans there would be no Law of Gravity but masses would continue to attract each other.
Better I think to say that science is about observations of regularities being coherent with other observations. For example the Theory of Evolution is an explanation that is strongly coherent with many other observations. Biblical Creationism is increasingly incoherent as we observe more regularities and discount more Biblical explanations.
Bunge (usually) distinguishes between laws, which are objective patterns in the universe, and law statements, which are our attempts to reconstruct these in thought. (He’s a constructivist in a weak sense – not a silly pomo sense or even a Kantian sense.)
Im not a scientist but it seems to me that science deals in extreme likelihoods – based on empirical evidence that ultimately can be confirmed some way by the senses – even maths is ultimately based on perception of space and quantity.
Rejecting true hypotheses happens all the time, so It is indeed logically possible that ID has been wrongfully rejected. But should we believe this? To paraphrase Dennett; it is not so interesting what is logically possible but what is possible in the real world.
This is a reversed version of Hempel’s dilemma expanded to all scientific knowledge: How can we know science is true when science constantly adjusts the facts? We assume that science tell us the truth, but we also know that science is a constant process of finding new facts – thereby implicitly assuming the facts we know now are wrong. This seems like a paradox?
There are many solutions to this problem. It (mostly) all comes down to what one considers the meaning of “true”.
It is is a semantic trick to compare “prove” in the strict logical sense and scientific “prove”. A rule of thumb is the logic version deals with truth values 1 and 0 whereas scientific “truths” are probabilities – everything between 1 and 0. Just because something is not true in the logical sense, does not mean that it is the best plausible explanation. It is logically possible that ID can be true, but the real world probability is so low we might as well call it non-existent – like planning your life based on expecting to win the lottery every day for the rest of your life. There are so many plausible facts in science (i.e. the scientific “truths”) that need to be contested for ID to be considered facts.
Proving one hypothesis false does not prove a competing theory true. If Darwinian evolution suddenly turned out to be wrong, it does not prove that ID is true. AKA the “just because Hillary was a bad candidate, does not mean Donald Trump is a good president”-fallacy.
I would agree that in mathematical terms, falsifying a hypothesis is just as provisional as proving a hypothesis. However, in practical terms, proving (in the sense of finding overwhelming evidence for) the negation of a scientific hypothesis is often much easier than proving the hypothesis. I think this imbalance is one of the things that makes scientific hypotheses of interest.
As an example, take the following hypothesis (note that I use proof in the sense of “finding overwhelming evidence for”, not the mathematical sense)
A: The speed of light is constant
and its negation
B: The speed of light is not constant
If you want to prove A, it’s very hard. You have to take lots of measurements of the speed of light in different scenarios and even then, there’s always the suspicion that you might have missed a situation in which you would have measured a different speed of light.
On the other hand, proving B, is easy. If you measure the speed of light, in two scenarios and come up with two speeds, B is true.
So you say, “I’ve measured the speed of light in all sorts of places all over the world and it always comes out the same. The speed of light is constant (probably). I say, “let me measure how fast it goes through this block of glass – oh, it’s slower”. A is disproved by one measurement and B is proved.
However, this does not mean B cannot be overturned. New information comes along about quantum electrodynamics and photons and it is discovered that light is apparently slowed down in glass because the photons keep bumping into electrons in the atoms of the glass and interacting with them. Careful measurement shows that between these interactions, the photons do travel at a constant speed.
So, no falsifying a hypothesis does not settle the question in an absolute mathematical sense. There’s always the possibility that we misunderstood something or even were just unable to measure things accurately enough.
The claim is correct but turns of the logicians meaning of “true” and “false”. In other words:
IF (hypothesis) THEN prediction.
If the prediction doesn’t occur then we can *logically* conclude that the hypothesis is false. However, if the prediction does occur then nothing in logic allows us to conclude that the hypothesis is true (the prediction might also follow from other premises).
This is why it is logically possible to falsify an hypothesis but not logically possible to prove an hypothesis. As soon as we move away from the logicians world of binary logic, and enter the empirical Bayesian world, then we can never “prove” of “disprove” empirical claims, but only assign probabilities of them being correct or false.
Scientific proof is inductive and depends crucially on Occam’s Razor. The Ptolemic system made good predictions, but was replaced by the Newtonian system because of its needless complexity. The same principle, I would argue, applies to all (or nearly all) scientific hypotheses, notably including evolution vs. creationism. Induction cannot conclusively prove a positive or a negative, but can only support or refute a proposition with greater or lesser confidence. In many cases, such as the examples above, the degree of confidence is so great as to approach certainty.
In our world, I think we can’t wait to make decisions sometimes. We have to go with the information in our pocket. Not deciding is a decision. Sometimes what we decided was based on bad information. Sometimes we have unintended consequences. 66 years old here and I have watched scientists go back and forth for years, changing their minds. I don’t think I see many absolutes. I thought I did before, but I was wrong. We need decisions sometimes and we base them on information, good, bad or partial. I may make a bad decision because I didn’t research, another person with much more intelligence than I because although he did tremendous research, the answers he used were wrong. The outcome is about the same. So maybe the question about proof, after consideration, tells us that “Common Sense” may be all we have sometimes. But who’s “Common Sense?” Ah, we get up every day and try to find cures, make sense of our origin and have something that gives us stability that may not really exists. When I see a child starving, I give them food if I can. All you folks in here make me think I missed out on a lot of education somewhere. I enjoyed reading this stuff.
If a putative entity is described as having certain characteristics and these characteristics are actually self-contradictory, then the entity does not exist.
Disclaimer: there are a few weird logicians (including the brilliant, but IMO crazy, G. Priest) who assert that there are or might be contradictions in the world. This to me is a mistake, but it is important to realize in my remark above I am assuming it.
A fine line between “brilliant” and ‘academic charlatanistic’, and I’m on the other side. I think John Burgess’s book on non-standard logics (“Philosophical Logic”) deals with that genre of logics and logicians rather well!
I wondered at first if Priest was a charlatan myself. What convinced me was _Beyond the Limits of Thought_. In the textbooks of non-classical logics a lot of the “philosophical motivation” is *bad*, I agree. As for Burgess’ book, I found it too cursory as well, though a good counterpoint.
I’m not wedded to any non-classical logic – Bunge convinced me of how much work one would have to do to *introduce* a new logic (though at the time there was almost no work on “hybrid logics” which might help her). However, once one starts talking about very fundamental matters in logic itself, epistemology, etc. one has to address the question, I think.
I bring this up with (lay) followers of Plantinga, who often have no idea that there are many systems of modal logic and his ontological arguments only “work” in S5.
Oh, and one more thing on Priest – I do think he’s horribly wrong, but I think most of the arguments against his views don’t work and Lewis was right to say the dialectical situation is awful.
How can we be 100% sure that a shoe-box does contain a $100 bill?
How can we 100% sure it doesn’t?
We can only be 100% sure if we have the ability to test that something is true or false.
If someone claims something that’s not testable and it contradicts multiple other well tested claims and we have good explanations why someone would make such a false claim we can be confident it’s a false claim.
Isn’t something, which implies a testable other thing which tests ‘false’, itself therefore testable (and of course false)?
IOW, is testability closed under logical consequence? or other “nice properties”?
One needs a theory of testability first – I suspect it won’t be, for the reasons espoused by the (mathematical) Intuitionists, or perhaps for computability theoretic reasons. The latter would only apply if computability is a desiratum, presumably. If it is, I think the work on computational learning theory shows “no” to be the answer, now that I think about it.
Two major differences I think:
Rather than seeming to say testability is transitive under logical consequence, I’ve said the obvious, that testing false is transitive under the reverse of logical consequence. That is, if theory A implies theory B, and if theory B tests false by experiment/observation, then theory A has clearly tested false in that same test.
What I said is clearly a succinct statement of exactly that.
However it seems clear that testability is not transitive under implication:
‘A: Nothing exists’ implies ‘B: gods do not exist’. But clearly A is testable, yielding False, whereas B seems not to be really testable. Perhaps it is under some kind of definition of “testable”, but that’s the whole trouble here.
My 2 cents: There are things we can know to be true with 100% certainty outside the realm of math and deductive logic. For instance, if an object is red all over, then we know with certainty that it is not simultaneously green all over, unless we futz with the definition of ‘red’ and ‘green’. We can know that this is true even of objects we haven’t seen, so probabilities don’t apply. This is not a logical consequence; it derives from facts about the world. Euclid’s axioms are another example of necessary synthetic truths; things that are necessarily true based on facts about the world, not on facts about meaning and derivation. I suspect there are many other examples if one were to go looking.
“For instance, if an object is red all over, then we know with certainty that it is not simultaneously green all over. . . .”
Even aside from the possibility of color-blindness, your choice of an example isn’t a good one. For one thing, it overlooks Galileo’s distinction, in The Assayer(1624), between the primary and secondary qualities of objects: “To excite in us tastes, odors, and sounds I believe that nothing is required in external bodies except shapes, numbers, and slow or rapid movements. I think that if ears, tongues, and noses were removed, shapes and numbers and motions would remain, but not odors or tastes or sounds.”
What Galileo is proposing here is a new set of rules for talking about the physical world. No longer will it do to say “the fire is hot” or “that object is red.” “Hot” and “red” refer to unique, subjective sensations, not to an objective property of the material world. If you wish to speak scientifically you need to say, “The fire transmits the sensation of heat” or “that object transmits the sensation of red.” Not that I recommend talking this way.
There is a physical basis for the perception of an object being red or green having to do with wavelengths along the visible light spectrum. The qualia of redness is not an intrinsic property of the object, but it does arise (in part) from some fact about the red object that is not true of a green object. The wavelength cannot simultaneously be 510 nm (or thereabouts) and 660 nm (or thereabouts), and this is, if not the main reason certainly part of the reason, that an object can’t be simultaneously red and green all over. My point is that there is a physical, not a semantic or logical basis for asserting that if the predicate ‘red’ is true of an object, it is necessarily the case that ‘not green’ is also true. And this doesn’t arise from either deductive reasoning or Bayesian reasoning– it is a necessary synthetic truth, as philosophers use the term.
My 2 cents just went up to 3 cents… 🙂
Unless it’s a sinople object …
/@
“My 2 cents just went up to 3 cents…”
Copy that. If you send me a bill, I’ll be happy to remit the 3 cents.
“The wavelength cannot simultaneously be 510 nm (or thereabouts) and 660 nm (or thereabouts)”
There certainly can be objects that reflect both of those wavelengths. They’d appear yellow under white light, but I’m guessing we could contrive a Land Effect experiment in which such an object could appear either red or green depending on ambient light — and maybe both simultaneously, to different observers. (E.g. one observer sits in a red-lit room and views the object through a window against a red background; the other sits in a green-lit room and views it against a green background.)
“Euclid’s axioms are another example of necessary synthetic truths; things that are necessarily true based on facts about the world, not on facts about meaning and derivation.”
That’s a quite surprising claim, given other (non-euclidean) geometries in mathematics together with physical considerations about what is the geometry of the cosmos, at least in a non-quantum approximation! Sounds a bit like one of Immanuel Kant’s famous errors.
Do you know of the paper which has title something like “Seeing Reddish Green and Yellowish Blue”? I have heard about this since Dennett’s _Consciousness Explained_. My philosophy of perception instructor a few years later did not know the state of the art on this and I keep forgetting to follow up. Does anyone know?
Also, if anyone in the Ottawa area can set up the lab experiment so I can see it myself, I’d love to try.
Supposedly some people report that they do indeed see a colour that they have never seen anywhere else before.
Crane & Piantanida’s 1983 paper in Science did the reddish green paper, there’s been some follow-up science since that suggested the perception was more complex than that. https://en.wikipedia.org/wiki/Impossible_color lists quite a few sources on that, including link on the Crane paper.
All I know is that Lee put in 3 cents, so I think he wins.