Physics is Ultimately Subjective
post by Gordon Seidoh Worley (gworley) · 2023-07-14T22:19:01.151Z · LW · GW · 34 commentsContents
35 comments
N.B. A basic and perhaps obvious point that nonetheless I think people get confused about. Despite the somewhat provocative title, my goal is not to say anything new to the average Less Wrong reader, only to emphasize a point and try to explain it so that it can be clearly seen.
Summary: Physics, by which I mean models of how reality works at the most fundamental level, is a subjective endeavor. Physics seems to be objective, but that's because there's high intersubjective consensus about which models best explain and predict reality. Rounding this off to objective causes confusion, and the point generalizes for all seemingly objective things.
Art, and in particular modern art, is highly subjective. Some people are impressed by the artistry of paint splatters, blank canvases, and signed toilets, while others roll their eyes. People like different things and only sometimes agree, so we generally accept that art is subjective.
Assessing a work of art looks sort of like this:
Physics, by contrast, seems totally objective. The world is how it is, and our models of it are good or bad insofar as they accurately and precisely describe the world. Whether or not you like general relativity, for example, has no bearing on whether it's a good theory. All that matters is how well general relativity explains and predicts what we observe.
The picture for physics might look like this:
But both of these pictures are wrong! The first leaves out the detail that art has to exist in reality—it's not art first, but atoms first, and it only becomes art when it's observed by someone who thinks of it as art. The second leaves out that theories of physics don't exist on their own, they exist in our heads [LW · GW]. To suppose otherwise is to suppose the existence of a paradoxical view from nowhere. So really both pictures are the same picture:
What does this mean for art and physics? Art is just as it was. The picture only makes it clearer that art doesn't exist independently: it exists within someone's experience of the world. For physics, though, it's a bigger deal. Physics feels like it's objective, and physical theories seem like they should be good or bad regardless of who does the assessing. And our experience matches this: people disagree about what art is good, but almost everyone agrees about what physics is good. What gives?
The difference is that art has low intersubjective consensus and physics has high. But that's just a fancy way of saying people strongly agree about physics and only weakly agree about art. Why is that the case?
We ultimately assess art by checking to see if we like it. The whole point of art is to make things we like looking at. We assess physics by checking to see if our theories match our observations. That's the point of physics: to model reality. So they behave differently, because even though in both cases each person independently runs their own evaluative function for art and physics, the art functions are loosely constrained while the physics functions are highly constrained.
At this point you might be frustrated that I've given a convoluted explanation of what we really mean—rather than what philosophers literally mean—when we say that something's objective. And fair, in everyday use you can get away with saying that physics is objective even though that's not literally true. But sometimes the details matter, and even if they don't, it's still good epistemic practice to be precise whenever possible to avoid injecting confusion into your thinking.
But I know some of you will still write that physics, and other seemingly objective things, are objective rather than have high intersubjective agreement. Why? For one, "objective" is shorter. For another, there are people who deny intersubjective consensus, not on grounds of specific evidence, but because they deny the value of evidence all together. That is, there's people who believe in invisible dragons [LW · GW] and they are, in a real sense, the enemies of good epistemology. It's tempting to line up against them [LW · GW], reverse their stupidity [LW · GW], and claim physics, and much else, can be objectively known.
But just because something seems like a good idea if we go up one or more simulacrum levels [? · GW] doesn't mean it's right. Subjectivity gets a bad rap because of its connotations, not because of what the concept actually points to. As bounded agents embedded in the world, we can do nothing other than deal with the world subjectively with our knowledge mediated by our experiences. Any sense of objectivity we have comes from intersubjective agreement that gives us confidence that it's not just our observations that support a belief but everyone's observations.
It's worth remembering that what looks to be objective is actually the result of intersubjective agreement, and that agreement comes from the aggregation of many individual, subjective assessments and not from reality itself…except insofar as we see subjective evidence to suggest a persistent world that exists independent of our experiences.
34 comments
Comments sorted by top scores.
comment by TekhneMakre · 2023-07-15T13:58:12.157Z · LW(p) · GW(p)
So in general, you'd say that nothing is objective? Is that right?
Replies from: gworley↑ comment by Gordon Seidoh Worley (gworley) · 2023-07-16T22:21:56.657Z · LW(p) · GW(p)
I think that's generally correct, although a bit beyond the intended scope of this post. There's no view from nowhere, no actual observer-independent information we can obtain, so any perceived objectiveness is contingent upon the subjective evidence we have about such things and we cannot be certain they are objective. Due to this lack of certainty I think it is better to just taboo the idea of objectivity and think in terms of "things that no observer has yet found sufficiently strong evidence to disagree with".
comment by tailcalled · 2023-07-15T14:30:43.422Z · LW(p) · GW(p)
What would be an example of something that is objective/not subjective, in your view?
comment by Ape in the coat · 2023-07-15T10:08:28.702Z · LW(p) · GW(p)
When we say "art is good" and "physics theory is good" we do not mean the same thing by "good".
In the first case we mean "aesthically pleasing", "though provoking", "satisfies a specific the definition of 'art'". In the second we mean "represents the reality".
Taboo the words "subjective" and "objective" - they are the legacy of the times when people didn't understand map-territory relations. You have a much better vocabulary now, you do not need to carry on the confusions of ancient philosophy.
Model of reality is a map. This map can either represent the territory or not. People consensus about a map is also a map, that can either represent a territory or not. Consensus that the map represents the territory isn't the same as map representing the territory.
Replies from: TAG↑ comment by TAG · 2023-07-15T11:17:09.033Z · LW(p) · GW(p)
When we say “art is good” and “physics theory is good” we do not mean the same thing by “good”.
No we dont. The point is what good, as in good physics, isn't. It isn't objective, and it isn't correspondence to reality per se...only a proxy, at best.
Consensus that the map represents the territory isn’t the same as map representing the territory.
What is? We hope that the physics we consider to be good represents the territory, but we have no way of checking directly. You can use the map/territory distinction to define truth, etc, but it doesn't tell you how to achieve it. So it doesn't actually solve the problem.
Replies from: Ape in the coat↑ comment by Ape in the coat · 2023-07-15T11:51:02.960Z · LW(p) · GW(p)
What is?
X=X. Why do you think that something other than X have to be X? We are using whatever imperfect approximations we have. As you've said yourself:
We hope that the physics we consider to be good represents the territory, but we have no way of checking directly.
And it's important to keep in mind that an imperfect approximation isn't the same as the thing it approximates or you can get confused.
You can use the map/territory distinction to define truth, etc, but it doesn't tell you how to achieve it.
It can help you deal with the cases where you confuse map and a territory, or map of a map with the map of a territory, or map of a map of a map with the map of a map of a territory, etc. And if it's this confusion that prevents you from achieving truth, then why - it's a way to achieve truth.
So it doesn't actually solve the problem.
What problem are you talking about? How can any map, in principle, represent a territory?
Replies from: TAG↑ comment by TAG · 2023-07-18T13:57:17.832Z · LW(p) · GW(p)
X=X. Why do you think that something other than X have to be X? We are using whatever imperfect approximations we have.
It's important to understand how imperfect they are.
We hope that the physics we consider to be good represents the territory, but we have no way of checking directly.
And it’s important to keep in mind that an imperfect approximation isn’t the same as the thing it approximates or you can get confused.
Who has that problem? I'm pointing out another problem: we can't quantify how correspondent a theory is. We can quantify how well it predicts, but not how close it is to reality. We don't know how much closeness-to-=reality an extra significant digit of accuracy buys you. Measurement accuracy is objective, closeness-to-reality isn't. We can't even say it's an approximation, because we don't know the error bar.
It can help you deal with the cases where you confuse map and a territory, or map of a map with the map of a territory, or map of a map of a map with the map of a map of a territory, etc. And if it’s this confusion that prevents you from achieving truth, then why—it’s a way to achieve truth.
You're holding a hammer, that doesn't mean I'm holding a nail. M/T can help you avoid some mistakes...but not the one I have mentioned. It's not a universal solvent.
So it doesn’t actually solve the problem.
What problem are you talking about? How can any map, in principle, represent a territory?
No...how we can know it does.
If you have every possible map in some library, then one of them will correspond exactly. But if you don't know which one, the library contains no useful information.
Replies from: Ape in the coat↑ comment by Ape in the coat · 2023-07-23T18:37:41.913Z · LW(p) · GW(p)
Who has that problem?
A lot of people. Topic starter included, apparently.
I'm pointing out another problem
Yes you do, and I'm not sure why bring up this separate issue here.
I'm pointing out another problem: we can't quantify how correspondent a theory is. We can quantify how well it predicts, but not how close it is to reality. We don't know how much closeness-to-=reality an extra significant digit of accuracy buys you.
Predicts what, dare I ask? For the first approximation we can define "reality" by the outcomes that we predict or fail to predict. If the "true reality" is something completely and utterly different that it's not captured by the outcomes we interact with, then either it doesn't exist or doesn't concern us.
You're holding a hammer, that doesn't mean I'm holding a nail. M/T can help you avoid some mistakes...but not the one I have mentioned. It's not a universal solvent.
From my perspective it looks as if I've successfully nailed a thing and now you are telling me out of blue that it's not enough to build a house. Yeah, sure, not that I claimed otherwise. What I actually claim, is that having a hammer and nails is very helpful for construction purposes, much more helpful than trying to hold things together with ropes, even though it's not the only things that are required.
If you have every possible map in some library, then one of them will correspond exactly. But if you don't know which one, the library contains no useful information.
There is a valid question here. It's not likely that a randomly selected map fits the territory. So where does the improbability of a map corresponding to the territory comes from? The answer is that the process wasn't random, that our maps and mapmaking tools are designed by the territory to be able to represent it and are part of the territory itself.
Replies from: TAG↑ comment by TAG · 2023-07-25T17:54:33.800Z · LW(p) · GW(p)
I’m pointing out another problem
Yes you do, and I’m not sure why bring up this separate issue here.
Because it's relevant to the OP. OP says physics is subjective. You say it's approximate, but still objective. I say that instrumentalism/measurement is objectively approximate, but realism/correspondence isn't .. it can't be calculated or measured.
Predicts what, dare I ask?
Observations.
For the first approximation we can define “reality” by the outcomes that we predict or fail to predict.
No, that's not reality in the realism versus instrumentalism sense.
If the “true reality” is something completely and utterly different that it’s not captured by the outcomes we interact with, then either it doesn’t exist or doesn’t concern us.
Everybody has their own values. If it concerns you personally,it concerns you. If you care whether or not your in a simulation, whether or there is a deistic deity, which interpretation of QM is true, it concerns you.
There is a valid question here. It’s not likely that a randomly selected map fits the territory. So where does the improbability of a map corresponding to the territory comes from? The answer is that the process wasn’t random, that our maps and mapmaking tools are designed by the territory to be able to represent it
Why would the territory care about our ability to represent it? Natural selection kills an organism for failure to predict, not for getting fundamental ontology wrong.
and are part of the territory itself.
How does that help? All epistemology, good bad , or indifferent, is in the territory.
Replies from: Ape in the coat↑ comment by Ape in the coat · 2023-07-26T05:58:32.788Z · LW(p) · GW(p)
You say it's approximate, but still objective.
No I don't. I specifically advise against using objective/subjective distinction in order to describe the phenomena that OP is trying to describe.
Observations.
While it's technically possible to build a purely experience-based epistemology it quickly becomes too complex in the Occamian sense.
No
Yes we can. Try and see for yourself
that's not reality in the realism versus instrumentalism sense.
Too bad. But that's the failure of the framework, not the method. Because realism versus instrumentalism is another confused dichotomy which is systematically leading people astray.
Everybody has their own values. If it concerns you personally,it concerns you. If you care whether or not your in a simulation, whether or there is a deistic deity, which interpretation of QM is true, it concerns you.
Feel free to pursue your values and/or try to answer ontological questions about deities simulations and QM interpretations. This is not relevant to the epistemological question at hand.
Why would the territory care about our ability to represent it? Natural selection kills an organism for failure to predict
You've just answered your question.
not for getting fundamental ontology wrong
As long as "fundamental ontology" is correlated with our observations we were selected for getting it somewhat right. See discovering QM even though our ancestors were not directly selected for particle physics scholarship. And if it's not - it's some other territory then. Maybe it's fundamentally unknowable for us, or maybe we will manage to correlate our observations to it later, if so we can still describe it using the same epistemological framework of map territory relations. Though I hope we will have something even better then.
How does that help? All epistemology, good bad , or indifferent, is in the territory.
Just another example how of M/T framework is superior to subjective/objective one. The latter treats subjectivity and objectivity as separate realms opposing each other: X is either subjective or objective, thus missing the fact of our embedment in the territory.
Replies from: TAG↑ comment by TAG · 2023-07-26T12:52:24.085Z · LW(p) · GW(p)
No I don’t. I specifically advise against using objective/subjective distinction in order to describe the phenomena that OP is trying to describe.
You need to say why it is wrong, or why your approach is better.
Just another example how of M/T framework is superior to subjective/objective one. The latter treats subjectivity and objectivity as separate realms opposing each other: X is either subjective or objective, thus missing the fact of our embedment in the territory.
Subjectivity and objectivity aren't ontological realms at all. They are epistemological terms.
While it’s technically possible to build a purely experience-based epistemology it quickly becomes too complex in the Occamian sense.
Is there some proof of that? Because it would imply that everything a Solomonoff inductor does is "too complex in the Occamian sense".
For the first approximation we can define “reality” by the outcomes that we predict or fail to predict.
No, that’s not reality in the realism versus instrumentalism sense.
Changing the definition of "realism" doesn't allow you to achieve realism, any more than calling a tail a leg proves a dog has five legs.
realism versus instrumentalism is another confused dichotomy which is systematically leading people astray.
You need to prove that.
This is not relevant to the epistemological question at hand.
The question -- according to the OP - is whether physics is objective. A lot of people want physics to tell them what reality is objectively. If you personally don;'t, that has no impact on them.
not for getting fundamental ontology wrong
As long as “fundamental ontology” is correlated with our observations we were selected for getting it somewhat right.
More than one ontology can predict the same observations -- that's part of the problem.
See discovering QM even though our ancestors were not directly selected for particle physics scholarship.
We don't know which fundamental ontology QM indicates. That's the whole problem of interpretation of QM.
And if it’s not—it’s some other territory then. Maybe it’s fundamentally unknowable for us,
So you concede that we don't have some easy, automatic method of figuring out ontology. And you also can't say that it doesn't matter objectively,...only that you don't care subjectively.
Replies from: Ape in the coat↑ comment by Ape in the coat · 2023-07-27T08:19:21.083Z · LW(p) · GW(p)
You need to say why it is wrong, or why your approach is better.
Being doing it for quite some time. Your next citation is literally about that.
Subjectivity and objectivity aren't ontological realms at all. They are epistemological terms.
I didn't claim that they are ontological realms. The claim was that they are defined as opposing each other. As if something is subjective it's not objective and vice versa and that this framework harms out understanding.
Is there some proof of that? Because it would imply that everything a Solomonoff inductor does is "too complex in the Occamian sense".
Not too complex to exist. To complex to be considered the best hypothesis by the Solomonoff inductor. At some moment it's just simpler to postulate some "shared reality" than disjointed experiences. And yes most of the hypothesis that Solomonoff inductor consider would naturally be too complex in this sense.
Changing the definition of "realism" doesn't allow you to achieve realism, any more than calling a tail a leg proves a dog has five legs.
Unless, of course the tail was originally called a leg by mistake and I'm now using a better fitting term.
You need to prove that.
Yeah, I'm going to write a whole sequence on these matters eventually.
The question -- according to the OP - is whether physics is objective. A lot of people want physics to tell them what reality is objectively. If you personally don;'t, that has no impact on them.
As always people have different definitions of "reality". I'm pointing the author towards definitions that I find less confused. You claim that they are not the same definitions you like. Which is true, and not to their credit. But more importantly what definitions you prefer to use isn't really relevant to the point I'm trying to make.
More than one ontology can predict the same observations -- that's part of the problem.
And yet only one of the hypothesis is the most accurate and the least complex considering all evidence. Come on, you brought up Solomonoff inductor yourself, you have to understand the underlying principle.
We don't know which fundamental ontology QM indicates. That's the whole problem of interpretation of QM.
Yes, that's the next question in chain. This doesn't contradict my point that we've already answered so many.
So you concede that we don't have some easy, automatic method of figuring out ontology. And you also can't say that it doesn't matter objectively,...only that you don't care subjectively.
Only in the narrow edge case of "metaphysical reality" being completely uncorrelated with our observations. Do you concede that in every other case we do? Not sure what you mean by "matter objectively". I'm expecting that the best map describing the territory it's actually correlated to, isn't confused that it doesn't describe some other territory that it's not correlated to.
Replies from: TAG, TAG↑ comment by TAG · 2023-07-29T17:21:18.596Z · LW(p) · GW(p)
I didn’t claim that they are *ontological *realms. The claim was that they are defined as opposing each other.
Lots of things are. Why would tha matter?
As if something is subjective it’s not objective and vice versa and that this framework harms out understanding
At some moment it’s just simpler to postulate some “shared reality” than disjointed experiences.
So what you mean by experience based explanation isn't empiricism, it's something like solipsism or idealism...?
Unless, of course the tail was originally called a leg by mistake and I’m now using a better fitting term.
That would be an extraordinary claim in need of supporr.
The question—according to the OP—is whether physics is objective. A lot of people want physics to tell them what reality is objectively. If you personally don;’t, that has no impact on them.
As always people have different definitions of “reality”. I’m pointing the author towards definitions that I find less confused.
Without explaining why.
More than one ontology can predict the same observations—that’s part of the problem.
And yet only one of the hypothesis is the most accurate and the least complex considering all evidence.
There's multiple simplicity criteria, too.
Come on, you brought up Solomonoff inductor yourself, you have to understand the underlying principle.
Yes. In fact, unlike a lot of people here, I can see its drawbacks.
SIs try to process an infinite list of programmes in a finite time. To acheve this, the process cess shorter candidate rgramnes first. In other words, the way they work has a justification that has nothing to do with their probability of successfully representing the territory , assuming they can be unproblematically translated into hypotheses. It might conveniently be the case that the same simplicity criterion does the job, but it doesn't have to be
So you concede that we don’t have some easy, automatic method of figuring out ontology. And you also can’t say that it doesn’t matter objectively,...only that you don’t care subjectively.
Only in the narrow edge case of “metaphysical reality” being completely uncorrelated with our observations.
No, in the cases where metaphysical reality is not completely correlated. That's why you need simplicity criteria.
Do you concede that in every other case we do? Not sure what you mean by “matter objectively”.
According to some universal values.
I’m expecting that the best map describing the territory it’s actually correlated to, isn’t confused that it doesn’t describe some other territory that it’s not correlated to.
Again , the problem is *knowing that you have axhieved correspondence.
comment by redlizard · 2023-07-15T17:15:17.093Z · LW(p) · GW(p)
This seems very confused.
What makes good art a subjective quality is that its acceptance criterion is one that refers to the viewer as one of its terms. The is-good-art() predicate, or the art-quality() real-valued function, has a viewer parameter in it. What makes good physics-theory an objective quality is that its acceptance criterion doesn't refer to the viewer; the is-good-physics-theory() predicate, or the physics-theory-accuracy() real-valued function, is one that compares the theory to reality, without the viewer playing a role as a term inside the function.
Sure, both of these functions are in the end computed by human brains, which adds a level of subjectivity to the imperfect physical act of actually evaluating these functions. But that doesn't mean that the ideal things-that-these-brains-are-supposedly-evaluating are themselves subjective. A human brain evaluating how accurate a certain physics theory is results in a subjective assessment of an objective truth; a human brain evaluating whether a certain painting is art results in a subjective assessment of a subjective property. The act of assessment by an imperfect brain adds a layer of subjectivity over something that may or may not be objective in the abstract; but that abstract ideal of that which these brains are supposedly evaluating has a real difference in kind that is well beyond mere consensus.
Modifiers and subjectivity-affecting operations can be applied to both objective and subjective criteria, of course. The degree to which a theory of physics reflects reality is an objective measure; the degree to which a reader likes a theory is a subjective measure. The degree to which a viewer considers a painting to be art is a subjective measure; the degree to which the average human viewer considers a painting to be art is an objective measure, because the viewer parameter has been aggregated out. But these complications only obscure the basic distinction, they do not fundamentally challenge it.
↑ comment by Gordon Seidoh Worley (gworley) · 2023-07-16T22:18:48.834Z · LW(p) · GW(p)
But that doesn't mean that the ideal things-that-these-brains-are-supposedly-evaluating are themselves subjective. A human brain evaluating how accurate a certain physics theory is results in a subjective assessment of an objective truth; a human brain evaluating whether a certain painting is art results in a subjective assessment of a subjective property. The act of assessment by an imperfect brain adds a layer of subjectivity over something that may or may not be objective in the abstract; but that abstract ideal of that which these brains are supposedly evaluating has a real difference in kind that is well beyond mere consensus.
But how would you know if anything is actually objective? Have you experienced objectivity not through a subjective lens?
This is my point: however strongly we might believe something to be a particular way, it is still something we believe. That doesn't diminish the quality of the belief or its potential truthfulness, but it is important not to miss this.
But these complications only obscure the basic distinction, they do not fundamentally challenge it.
Again, how do you know that this basic distinction exists? Can you point to it without relying on subjective evidence (evidence that passed through some observer)? If you cannot, then anything we might claim to be "objective" still rests upon subjective assessment of evidence. Objectivity is a constructed idea created by subjective agents.
comment by Hastings (hastings-greer) · 2023-07-26T16:53:34.002Z · LW(p) · GW(p)
Sometimes the number of stones in the bucket really does match the number of sheep in the field [LW · GW]. If we can’t call that objectivity, we still have to call it something. Intersubjective agreement doesn’t seem like a better name to me, but of course naming is nothing if not subjective ;)
Replies from: TAGcomment by Shmi (shminux) · 2023-07-15T00:45:43.174Z · LW(p) · GW(p)
That's the point of physics: to model reality.
One would be a lot less confused if replacing the above with "That's the point of physics: to accurately model a specific type of observations" In art "accuracy" is not the main goal, and that's the only difference. The "intersubjective" part is in how the models are built.
Replies from: TAG↑ comment by TAG · 2023-07-15T11:23:57.307Z · LW(p) · GW(p)
This a statement of the realistic philosophy:
That’s the point of physics: to model reality.
This is a statement of the idealist philosophy:
“That’s the point of physics: to accurately model a specific type of observations”
Neither is a fact. The realism/instrumentalism debate has been going on for a long time. Calling the realist position "confused" isn't going to settle it.
Replies from: shminux↑ comment by Shmi (shminux) · 2023-07-16T04:40:27.345Z · LW(p) · GW(p)
I mean, instrumentally, physics about making and testing models, whether your philosophy is idealism, realism or anything else. Philosophy is a different kettle of fish, and yeah, the debate there has been going on forever, since there is no experimental way to resolve it.
Replies from: TAG↑ comment by TAG · 2023-07-16T12:21:41.490Z · LW(p) · GW(p)
I mean, instrumentally, physics about making and testing models
It's about that in the sense that that is the method, but it is not about that as a point. You don't just build models, and put them in a filing cabinet, you build models and use them to make predictions, and also to make claims about the nature of reality.
Replies from: shminux↑ comment by Shmi (shminux) · 2023-07-16T13:43:42.946Z · LW(p) · GW(p)
Right, it is, but the second part is basically "philosophy", though most physicists are naive realists and "make claims about the nature of reality".
comment by Noosphere89 (sharmake-farah) · 2023-07-15T13:55:49.744Z · LW(p) · GW(p)
I think the key here is whatever reality really is, it's laws and parameters aren't controlled by anyone, and thus it's useful to reliably say that this reality is objective. I suspect all cases of subjectivity are essentially cases where someone can control the rules and parameters of something, and given different preferences, subjectivity arises in concepts.
Replies from: gworley↑ comment by Gordon Seidoh Worley (gworley) · 2023-07-16T22:28:09.619Z · LW(p) · GW(p)
I think the key here is whatever reality really is, it's laws and parameters aren't controlled by anyone, and thus it's useful to reliably say that this reality is objective.
This is a reasonable supposition to make, but as I point out in another comment, we only know that its laws and parameters aren't controlled by anyone insofar as we've only thus far seen evidence to suggest that. And that fact that tomorrow we could obtain evidence that suggests otherwise means there's a subjective layer between us and making a claim like "reality is objective", and thus any claims to objectivity are necessarily subjective claims, and because of what objectivity means, this disqualifies a strong claim to objectivity and permits only a subjective, contingent claim to it.
comment by Rami Rustom (rami-rustom) · 2023-07-15T11:45:05.336Z · LW(p) · GW(p)
I don't understand the point of this post except to try to clarify the meanings of the words objective and subjective.
But if that's the case, what's the point of that? What does it help us do better?
What happens if somebody does not hold your view and sticks with their current view?
My guess is nothing. The act of doing physics (i.e. the scientific approach) does not get affected by how we understand the terms objective and subjective. Do you agree?
Side note: I wonder if it would have been better to start your post with a clarification of the terms objective and subjective, then explain what physicists and artists do, then explain how physics or art are objective or subjective.
comment by juliette (julietteculver) · 2023-07-18T16:01:26.024Z · LW(p) · GW(p)
I'm reminded of the book What is Mathematics, Really? by Reuben Hersh. I read it too long ago to be able to summarise it well or debate the arguments, but it takes a fairly similar stance to mathematics and is very readable from what I recall.
comment by gjm · 2023-07-15T02:53:54.677Z · LW(p) · GW(p)
Imagine some variety of art for which it happens that all humans agree on the merits of any given instance. Is this picture more beautiful than that one? Everyone gives the same answer. No one knows exactly what it is about the art that they're assessing, though.
Would you consider judgements about these artworks to be just the same as ones about physics? I don't think I would. For instance, suppose we learned that we were soon to be visited by intelligent aliens. I would be much more willing to bet that the aliens would agree with me about the answer to "what will happen when I put these magnets and little iron balls in such-and-such a configuration?" than about the answer to "how beautiful is this picture?". Even if literally all human beings agree about the latter.
I don't know whether Gordon would take the same view. But for me this seems to point to a sense in which physics is "more objective" than artistic judgement, in a way that doesn't just come down to how reliable the intersubjective agreement is.
Gordon does say
That's the point of physics: to model reality. [...] the art functions are loosely constrained while the physics functions are highly constrained.
and it seems to me that this (and not the degree of intersubjective agreement) is what makes physics "objective" where art is "subjective". Whereas Gordon seems to be claiming that actually "objective" just means "very high intersubjective agreement" and the only way in which "the point of physics" and the degree of constraint matter is by affecting how much intersubjective agreement there is. And I think this is just plain wrong.
(A family of counterexamples in the other direction, less clear-cut but not requiring a counterfactual about an actually-nonexistent art form: pick any scientific topic that has become highly politicized. How effective are masks, or vaccines, or ivermectin, against COVID-19? How safe is it to give puberty-blocking drugs to gender-dysphoric preadolescents? What should we expect to happen to termperatures across the world over the next 50 years, for any given pattern of industrial activity? I claim that these questions are more like physics than art, in terms of how "objective" their answers are; more like art than physics, in terms of intersubjective agreement. Again, the point is that my notion of "objectivity", which may or may not match Gordon's but I am pretty confident is a reasonable one, plainly diverges in these cases from intersubjective agreement, and I think this is good reason to think that objectivity and intersubjective agreement are different things.)
Replies from: gworley, TAG↑ comment by Gordon Seidoh Worley (gworley) · 2023-07-15T06:14:34.943Z · LW(p) · GW(p)
Imagine some variety of art for which it happens that all humans agree on the merits of any given instance. Is this picture more beautiful than that one? Everyone gives the same answer. No one knows exactly what it is about the art that they're assessing, though.
Would you consider judgements about these artworks to be just the same as ones about physics?
In what sense? Clearly they aren't the same because they're judgements about different things grounded in different assessment functions. In this scenario everyone thinking some art is good is still grounded in what people like. That's still different from being grounded in a function based on what best describes what we observe. That your hypothetical work of art produces high intersubjective agreement doesn't suddenly make it the same as the situation with physics, just makes it more like it in one respect.
Whereas Gordon seems to be claiming that actually "objective" just means "very high intersubjective agreement" and the only way in which "the point of physics" and the degree of constraint matter is by affecting how much intersubjective agreement there is. And I think this is just plain wrong.
You've misunderstood me. The point of physics, as I say, is to have a model of the world that explains and predicts reality. High intersubjective agreement is the phenomenon that gives the illusion of objectivity.
Replies from: gjm↑ comment by gjm · 2023-07-15T15:01:21.476Z · LW(p) · GW(p)
I agree that my hypothetical artform is not the same as physics. That was rather my point: it seems (maybe I have misunderstood?) that you're saying either that when we say "X is objective" what we really mean is "X has extremely high intersubjective agreement", or else that what causes us to say "X is objective" is extremely high intersubjective agreement; and I think both of those are just wrong, and so I gave (what seems to me to be) a hypothetical example in which there is extremely high intersubjective agreement but I would not be at all inclined to call it objective.
You say "You've misunderstood me" but if there is a specific thing where you think I think you said X but actually you meant Y, I haven't been able to work out what that thing is, what X is, and what Y is.
... Oh, maybe you think I'm accusing you of saying that physics being about trying to model the world matters literally only because that leads to high intersubjective agreement; of course I don't think you think that and I apologize if I gave that impression. What I think (perhaps wrongly?) is that the only way that feature of physics matters for determining how "objective" physics is is by making there be strong intersubjective agreement.
It would help me (and maybe others? I don't know) if you could clarify a few specific things.
- There's (at least potentially) a distinction between what something means and what actually makes people say it. I think you are saying that what makes people call things objective is the presence of good intersubjective agreement, and that actually e.g. physics is not more "objective" than art but merely seems so because it has good intersubjective agreement. Is that right?
- If so: what exactly do you mean by "objective"? Like some other commenters here (tailcalled, TekhneMakre) I am concerned that you're defining "objective" in a way that makes it (fairly uncontroversially) not apply to anything, and it seems to me that there are plausible ways to understand "objective" that make it apply more to things commonly thought of as objective and less to things commonly thought of as subjective, in which case I think that might be a better way to use the word. But I'm not sure, because I don't know quite what you mean by "objective". (It seems like you mean something with the property that "theories of physics are in our heads" implies "physics is not objective", for instance. But that doesn't really nail it down.)
- When I say "physics is objective" (actually I would generally not use those words, but they'll do for now) what I think I mean is something to do with physics being grounded in the external world, and something to do with my opinion that if aliens with very different mental architecture turned up they would none the less have quite similar physics, at least to the extent that it would make similar predictions and quite likely in its actual conceptual structure, and really not very much to do with intersubjective agreement. Do you think I am just deluding myself about what's going on in my head when I say that physics is more objective than art, and that actually all I'm doing is comparing levels of intersubjective agreement? Or what?
- (I do think that intersubjective agreement is relevant. The way it's relevant is that what-I'm-calling-objectivity is one possible explanation for intersubjective agreement, so strong intersubjective agreement is evidence of what-I'm-calling-objectivity. But it's not the only possible explanation, and it's far from being proof of objectivity, and it certainly isn't what "objectivity" means.)
↑ comment by Gordon Seidoh Worley (gworley) · 2023-07-18T16:51:24.278Z · LW(p) · GW(p)
1. There's (at least potentially) a distinction between what something means and what actually makes people say it. I think you are saying that what makes people call things objective is the presence of good intersubjective agreement, and that actually e.g. physics is not more "objective" than art but merely seems so because it has good intersubjective agreement. Is that right?
Yes, for reasons that might seem obvious after I answer the next question.
2. If so: what exactly do you mean by "objective"? Like some other commenters here (tailcalled, TekhneMakre) I am concerned that you're defining "objective" in a way that makes it (fairly uncontroversially) not apply to anything, and it seems to me that there are plausible ways to understand "objective" that make it apply more to things commonly thought of as objective and less to things commonly thought of as subjective, in which case I think that might be a better way to use the word. But I'm not sure, because I don't know quite what you mean by "objective". (It seems like you mean something with the property that "theories of physics are in our heads" implies "physics is not objective", for instance. But that doesn't really nail it down.)
I generally think we should taboo objective because I don't think there's agreement on the definition. I have two definitions in mind, and I think there's a motte and bailey situation going on with them.
Definition 1: not dependent on a mind/observer for existence
Definition 2: stuff that seems to be the same for all known observers
Definition 1 is something like the strong version of "objective". Definition 2 is a weak version that's equivalent to a definition for "intersubjective consensus".
Definition 2 is the thing that's defensible, but Definition 1 is what some people want to mean by "objective", yet nothing exists independent of minds because existence is a property of ontology (the map) not reality (the territory). I say more about this fine distinction between existence and being here [LW · GW].
3. When I say "physics is objective" (actually I would generally not use those words, but they'll do for now) what I think I mean is something to do with physics being grounded in the external world, and something to do with my opinion that if aliens with very different mental architecture turned up they would none the less have quite similar physics, at least to the extent that it would make similar predictions and quite likely in its actual conceptual structure, and really not very much to do with intersubjective agreement. Do you think I am just deluding myself about what's going on in my head when I say that physics is more objective than art, and that actually all I'm doing is comparing levels of intersubjective agreement? Or what?
- (I do think that intersubjective agreement is relevant. The way it's relevant is that what-I'm-calling-objectivity is one possible explanation for intersubjective agreement, so strong intersubjective agreement is evidence of what-I'm-calling-objectivity. But it's not the only possible explanation, and it's far from being proof of objectivity, and it certainly isn't what "objectivity" means.)
I think somehow you've come to believe there is evidence to suggest there's an external reality and you're drawing conclusions about other things based on having assumed there's an external reality independent of you as an observer.
For comparison I would use reality/"the world" to point directly to experience. Anything else we think we know is known only through that experience, and that includes any claims we might make to the existence of external reality. But in an important sense external reality, however real it seems, is not real because we only know about it indirectly as mediated by our experience and thus its existence is a claim not an assumption.
Replies from: gjm↑ comment by gjm · 2023-07-19T01:34:11.172Z · LW(p) · GW(p)
Your distinction between "existence" and "being" seems ... idiosyncratic, and it seems to me that you should probably split your "definition 1" into definitions 1a (not dependent on a mind/observer for existence) and 1b (not dependent on a mind/observer for being). In so far as I understand the distinction you are making (which may not be far enough) it seems to me that nothing is objective-1a by definition (because you take "existence" to be a property of people's ideas) but some things might be objective-1b. I don't think anyone means objective-2 when they say "objective", and I think your insistence that they do is just a mistake.
I think the way the mistake arises is that, given other assumptions you make, what other people mean by "objective" is crazy, and so it feels to you as if saying they mean objective-2 is being charitable, replacing a crazy notion with one that's wrong but at least makes sense. Whereas, to those who make different assumptions, what you're doing looks highly uncharitable, replacing a perfectly reasonable notion with one that's wrong. (This is a very common phenomenon.)
Specifically, of course the assumption you make and they don't is something along the lines of "all talk of 'reality' and 'the external world' as something separate from our experience is nonsense". I'm sure it's true that nothing is objective given solipsism, but I find it difficult to care because I find solipsism unconvincing, your version (sorry!) as much so as any other.
(I expect you don't like having your position called solipsism. I'm going to call it that anyway. Sorry.)
Also, even given solipsism, I think your definition of "objective" makes it a useless word: by definition, nothing is "objective". I -- like you -- generally prefer to avoid the word, precisely because different people mean different things by it and it causes confusion; but I do think there's a useful concept somewhere in its vicinity; there really is a useful distinction between physics and (hypothetical) highly-intersubjectively-consistent art, and that distinction has something to do with what people commonly mean by "objective". An analogy (borrowing on a nice little essay by David Chalmers from when "The Matrix" first came out): suppose it turns out that we are all brains in vats, living in a painstakingly constructed simulated world; then it is still true, in a useful sense, that black swans are real and unicorns aren't, even though in another sense "nothing inside the simulation is real". If we-in-the-Matrix explore our world very thoroughly, we will find black swans but we will never find unicorns. Similarly: it's not just that physics is agreed on by (in some sense) everyone; it's also that it seems clear that it would be agreed on by aliens, AIs, archangels, etc., whereas we should expect those beings to have quite different taste in art from ours, and this points to an important difference between physics and art, and the word "objective" isn't such a bad word for it, even if in some sense "nothing is objective".
(Relatedly: even though we only know about "reality" via our experiences, I claim that there is a useful distinction between the things I am seeing right now and the things I might see if I were on a large dose of hallucinogens, and words like "real" and "objective" are useful ways to point at that distinction. This isn't really any different from e.g. saying "the monster is behind the building" when talking about a computer game, even though "really" the monster and building are both being displayed on the same flat surface.)
Replies from: gworley↑ comment by Gordon Seidoh Worley (gworley) · 2023-07-19T02:18:37.535Z · LW(p) · GW(p)
I think, based on this reply, you basically get my point, we're just quibbling about some details.
I take this sort of hard line stance on "objective" because surprisingly many people, when pressed, turn out to be naive realists, including a whole bunch of rationalists I've interacted with over the years. So if I seem maximally uncharitable it's because there's a bunch of folks out there who are failing to grasp the point I make in this point under any terms.
↑ comment by TAG · 2023-07-15T11:35:11.533Z · LW(p) · GW(p)
and it seems to me that this (and not the degree of intersubjective agreement) is what makes physics “objective” where art is “subjective
If "objective" means " less subjective than art", that argument works, but if it means "not subjective at all", it doesn't.
comment by David Gross (David_Gross) · 2023-07-15T01:45:23.794Z · LW(p) · GW(p)
This reminds me of a POV that I find perennially tempting, a sort of Buddhism verging on solipsism:
"Reality Itself" is already completely here and completely available to you. That indeed is the definition of Reality Itself: the actual subjective contents of the present moment, your (pre-"assessed") subjective experience. To discover Reality Itself, you don't have to assess your subjective experience as though it were merely evidence left behind by the real reality that you have to examine for clues. It's already the real deal. Our ideas about "objective reality" (atoms and quarks and fields, but even chairs and tables and people) are models and linguistic conventions we find helpful for discussing, finding patterns in, and predicting Reality Itself, but they are not themselves Reality Itself hiding behind our subjective experience. You can learn interesting things about Reality Itself by examining it and subjecting it to scientific analysis, but you can't get any realer that way: any closer to reality. You're as close as you can get already -- it's right here; you can't miss it.