# Learning the word "ontic"

Gold Member
The meaning of most words is not learned through their definitions. When I say "Jack is walking", no one will ask me to define "Jack", to define "is" or to define "walking". Even in pure math, if I say "The set exists", no mathematician will ask me to define "The", to define "set" or to define "exists". The meaning of most words is learned gradually, through examples. (Artificial neural networks in machine learning also learn through examples, but humans usually need much smaller number of examples to learn it efficiently.)

The word ontic is one such word. The meaning of "ontic" is somewhat similar to "real", but "ontic" has a narrower meaning. There is no precise definition, but it can be learned through examples. Here are some examples:

1. In classical mechanics, the particle position as a function of time ##x(t)## is ontic. Its Fourier transform ##\tilde{x}(\omega)## is not ontic.

2. In classical mechanics, anything that can directly be derived from ##x(t)## is ontic. The meaning of "directly" also has to be learned through examples. For instance, the velocity ##\dot{x}(t)## and acceleration ##\ddot{x}(t)## are directly derived from ##x(t)##. The momentum ##p(t)=m\dot{x}(t)## and the force ##F(x)## are not directly derived from ##x(t)##.

3. In classical mechanics, quantities that cannot be directly derived from ##x(t)## are not ontic. For example, mass ##m##, momentum ##p##, force ##F## and Lagrangian ##L(x,\dot{x})## are not ontic.

4. A classical wave ##\phi(x,t)## is ontic. Its spatial Fourier transform ##\tilde{\phi}(k,t)## is not ontic.

From those examples, one can use intelligent extrapolation to determine whether many other concepts in classical physics are ontic or not. (But in some cases it may not be obvious, so we may have have different interpretations of classical physics. That's particularly true in the theory of relativity.)

When one grasped the meaning of "ontic" in classical physics, one can start to think and talk about "ontic" in quantum physics.

Last edited:

## Answers and Replies

martinbn
Position is ontic needs some motivation. Otherwise it looks arbitrary. Why not ##p(t)## is ontic, and ##x(t)## is not, and go from there? There are so many possibilities, that it requires some justification for the choice of ##x(t)##.

Gold Member
Position is ontic needs some motivation. Otherwise it looks arbitrary. Why not ##p(t)## is ontic, and ##x(t)## is not, and go from there? There are so many possibilities, that it requires some justification for the choice of ##x(t)##.
There are many motivations. It's compatible with common sense and basic intuition. I see position, I don't see momentum. Position is something that you learn in physics before you learn about momentum. I can compute ##p(t)## from known ##m## and ##x(t)##, but I can't compute ##x(t)## from known ##m## and ##p(t)##. It's only in the Hamiltonian formalism that momentum appears on an equal footing with position, in other formalisms (e.g. Newtonian or Lagrangian formalism) position is more elementary than momentum. In higher derivative theories momentum loses its meaning, position doesn't. All this suggests that position is more basic than momentum.

Jimster41 and dextercioby
1. In classical mechanics, the particle position as a function of time ##x(t)## is ontic. Its Fourier transform ##\tilde{x}(\omega)## is not ontic.
...
From those examples, one can use intelligent extrapolation to determine whether many other concepts in classical physics are ontic or not. (But in some cases it may not be obvious, so we may have have different interpretations of classical physics. That's particularly true in the theory of relativity.)
One reason why the notion of ontic seems non-trvial to translate to for example my agent interpretation is that in classical mechanics, one is allowed to entertain the idea that things exists, wether observed or not, as they are assumed to just exist. This kind of thinking is not used in agent centered picture (it's forbidden).

So i have to twist things quite a bit to find the correspondence. What I entertain to likely be the corresponding "real" things, when I associate your solipsist HV with my own view, is:

It's what is inituitively "real" to the actual inside agent, regardless of wether it's fellow observers in the environment are in agreement. [this relates to the observer democracy vs equivalence condition mentioned before]

What i mean is, what I see, observe and touch, is not less real just because I am the only witness! or that all my attempts to communicate this to the environment is compromised. It only means that it's an information war, without any external jury. The winner just eats the looser, there is no right or wrong, no true or false.

In this sense, the agents reality is as real as anything can be, but it's hidden (or screened behind a communication channel).

The reasons why this reality was rejected in Bell's theorem, is because it's assumed that what is "real" can be communicated without beeing compromised, and that every agent, and thus every pary of the universe should be in agreement with this hidden real structures (ie reality is some sort of truth that does not need to be communicated or inferred, it just is), and there follows the ignorance assumption beening the basis of bells theorem. But this sort of extrinsic, non-inferrable truth violates the constructing principles in the agent view.

So sum up, my association of hte solipsist HV if yours is that they are part of the agents own microstate, they are very real. But hidden to other agents. But it's available for inquiry. This "picture" IMO explains why the ansatz in the theorem does not hold.

The suggested problem one arrives seems however very difficult, and certainly confusing. So the medicine will appear worse than the disease at first. This has been my view for years, but I am still struggling with the new puzzle.

/Fredrik

3. In classical mechanics, quantities that cannot be directly derived from ##x(t)## are not ontic. For example, mass ##m##, momentum ##p##, force ##F## and Lagrangian ##L(x,\dot{x})## are not ontic.
Do you have a reference for those assertions?

Even in pure math, if I say "The set exists", no mathematician will ask me to define "The", to define "set" or to define "exists".
Even if they won't ask you, there are definitions by Tarski from 1933 and 1956 for "exists". They might appear trivial, but they are not. The notion of "set" on the other hand is taken as fundamental, and does not "need" a definition.

russ_watters and Demystifier
martinbn
There are many motivations. It's compatible with common sense and basic intuition. I see position, I don't see momentum. Position is something that you learn in physics before you learn about momentum. I can compute ##p(t)## from known ##m## and ##x(t)##, but I can't compute ##x(t)## from known ##m## and ##p(t)##. It's only in the Hamiltonian formalism that momentum appears on an equal footing with position, in other formalisms (e.g. Newtonian or Lagrangian formalism) position is more elementary than momentum. In higher derivative theories momentum loses its meaning, position doesn't. All this suggests that position is more basic than momentum.
All these suggests that position is fundamental or primitive, but why is it ontic? If I understood you correctly the whole point was to get an understanding of the use of "ontic". To understand what "ontic" means. So far it seems that the definition of ontic must include position. Why? I am not convinced yet. Also you motivation seems to be centered about the human "I see position, I have the basic intuition, I have the common sense, I can compute,.." Shouldn't it be something that is intrinsic to the actual system not the observer, especially if you are going towards QM?

Here is the question. Why call position ontic and not fundamental?

Gold Member
Do you have a reference for those assertions?
See e.g. the books by Durr on Bohmian mechanics and quantum foundations, or the book by Norsen on quantum foundations. Durr says that ontology of a theory is what the theory is really about. In #8 I explain that classical mechanics is really only about positions.

I probably don't understand it very well, but I think Tarski eliminated (not defined) quantifiers ("exist" and "for all") to define the truth.

Last edited:
gentzen
Gold Member
Why call position ontic and not fundamental?
It's both. Ontic because we think of partice position in classical mechanics as something which is there even if we don't see it.

We also think of position as something which is there even if we don't have a specific theory about it. We can formulate a classical theory of particles without mass ##m##, but we can't formulate a classical theory of particles without particle positions. So it's really about particle positions. Everything else (mass, momentum, Lagrangian, ...) are just auxiliary concepts to describe the behavior of positions. Ontology of a theory is what the theory is really about.

Last edited:
Ontic because we think of partice position in classical mechanics as something which is there even if we don't see it.
If we can make this mapping

"we" ~ the set of all observers (implying that an agreement among agents => equivalence)

I think this is abstractly in agreement with my definition?

The differnece between QM and classical mechanics is then that in classical mechanics, the equivalence can always (at least in principle) be manifest. But in QM, this is not so. Some part of reality is bound to remain relative to the agent. And, ignoring this, has implications for interactions. This is why, the consensus in QM is pushed out to "ensembles" or statistics.

But I do not label this as "incompleteness" as the incompleteness has intrinsic, reasons, not out of "ignorance".

/Fredrik

Gold Member
So sum up, my association of hte solipsist HV if yours is that they are part of the agents own microstate, they are very real. But hidden to other agents. But it's available for inquiry. This "picture" IMO explains why the ansatz in the theorem does not hold.
Let us try to make in concise.

What's ontic in my solipsistic HV's? Short answer: particle positions responsible for consciousness.

What's ontic in your view? Short answer: ... ?

What's ontic in your view? Short answer: ... ?
= The agents microstructure, whose state is responsible for encooding its infered expectation of the future.

/Fredrik

Demystifier
martinbn
It's both. Ontic because we think of partice position in classical mechanics as something which is there even if we don't see it.
Ok, but it seems that it is completely unnecessary. Why give it a name! Which in addition is not strictly defined. We know what positions are, we know what forces are and so on. What's the motivation to label one of them ontic!
We also think of position as something which is there even if we don't have a specific theory about it. We can formulate a classical theory of particles without mass ##m##, but we can't formulate a classical theory of particles without particle positions. So it's really about particle positions. Everything else (mass, momentum, Lagrangian, ...) are just auxiliary concepts to describe the behavior of positions. Ontology of a theory is what the theory is really about.
Not sure about this. Also positions need not be the most fundamental in the theory. Configurations could be what you need. For example a free particle in the plane can be descibed by its ##x## and ##y## at any given time. These are the positions. But if the particle is constrained to move along a unit circle you can use instead the angle ##\theta## to completely discribe the dynamics of the particle. Also positions are not invariant, they depend on some choices. In some cases it may be better to discibe the particle by its worldline.

This whole ontic business seems like you are tying your hands for no good reason. At least, I am not convinced yet what the point is. Of course the cynic in me says that this is just part of the BM propaganda , but let's not go there.

russ_watters
I probably don't understand it very good, but I think Tarski eliminated (not defined) quantifiers ("exist" and "for all") to define the truth.
Even if Tarski had succeeded to define truth in arbitrary object languages by quantifier elimination, it would still be a definition. (From my perspective, the important part of first order semantics is that a model can be specified first, independent of the sentences of a theory that it may satisfy, and that this fixes the meaning of "exists" for such a given model.) Note that the truth definition by quantifier elimination only works in special cases:
... a much less direct route, which Tarski describes as a ‘purely accidental’ possibility that relies on the ‘specific peculiarities’ of the particular object language.
But when it works, then it even gives an algorithm for determining which sentences are true (or false):
Tarski gave a truth definition by quantifier elimination for the first-order language of the field of real numbers. ... Later he gave a fuller account, emphasising that his method provided not just a truth definition but an algorithm for determining which sentences about the real numbers are true and which are false.

Gold Member
What's the motivation to label one of them ontic!
To differ real physical stuff from its mathematical representation.

For example a free particle in the plane can be descibed by its ##x## and ##y## at any given time. These are the positions. But if the particle is constrained to move along a unit circle you can use instead the angle ##\theta## to completely discribe the dynamics of the particle.
Have you ever seen a real particle constrained to move along a circle? I did, in a roulette. It had well defined all three position coordinates ##x,y,z##. That was the ontology. A 2-dimensional description with ##x,y## and a 1-dimensional description with ##\theta## was just convenient descriptions, not the ontology. The particle was really there in 3 dimensions, not in 2 or 1 dimensions.

Minnesota Joe
Lynch101
Gold Member
The meaning of most words is not learned through their definitions. When I say "Jack is walking", no one will ask me to define "Jack", to define "is" or to define "walking". Even in pure math, if I say "The set exists", no mathematician will ask me to define "The", to define "set" or to define "exists". The meaning of most words is learned gradually, through examples. (Artificial neural networks in machine learning also learn through examples, but humans usually need much smaller number of examples to learn it efficiently.)

The word ontic is one such word. The meaning of "ontic" is somewhat similar to "real", but "ontic" has a narrower meaning. There is no precise definition, but it can be learned through examples. Here are some examples:
To define terms such as 'exist' one would nearly need to undertake a cartesianesque radical deconstruction and build form the 'ground' up. Thankfully Descartes has already done something like that, so we can stand on his shoulders, so to speak.

One thing that we know for certain, an undeniable fundamental truth, is that there is existence. We don't need to define existence here since 'existence' is simply a label that we apply to the undeniable fact that we are experiencing [what appears to be] a physical world. The same goes for the term 'experience'. These are all just labels which have their terminus in the physical world.

So, existence is undeniable. What is up for question, however, is what the nature of that existence is i.e. what is it that exists? In trying to answer this question we are trying to find out what is 'ontic'. Again, 'ontic' is just a label which attempts to point in the direction of the physical world. It is the 'absolute terminus'.

As humans we started thinking that the contents of our visual perception were 'ontic', but with scientific investigation we found that our interpretations of our sensory input were misleading. This lead us to the atomic picture. [Some] ancient Greek philosophers thought that atoms were 'ontic' but further scientific inquiry demonstrated that the atomic picture was not the 'absolute terminus'. It might be that quantum fields are 'ontic'.

physika
jim mcnamara
Mentor
Ontic comes from philosophical ontology. The meaning in that context: ontic is physical, real, or factual existence. So it's use in the Physics literature had to have started from that point.

Example urrent ontology efforts:
P. Magnus "Scientific Enquiry and Natural Kinds: From Planets to Mallards (New Directions in the Philosophy of Science) "

Why do I mention that? Philosophy of Science used to be a required part of the undergrad hard sciences curriculum in my experience. Not any longer it seems.

For example, I assume most professionals get the concept of empirical falsification and have heard of Karl Popper. I hope. Which you would have gotten as an undergrad course, or reading.

But word borrowing like this means, I think, that the originator of the usage had academic exposure to Philosophy. So, the use of ontic does not seem too very different from what I learned long ago. What is different is @Demystifier 's "osmotic" method of learning the meaning of ontic. New to me.

Whatever works well for you, keep on....

russ_watters
martinbn
To differ real physical stuff from its mathematical representation.
The coordinates are not real physical stuff, the particle is. The coordinates are part of the mathematical representation. They are not made of atoms. It doesn't matter if they have exact values at any given time. They are still part of the map, not the territory.
Have you ever seen a real particle constrained to move along a circle? I did, in a roulette. It had well defined all three position coordinates ##x,y,z##.
The ##x,y,z## are well define only if you have chosen a coordinate system. Otherwise there are not. And what is so special about Cartesian coordinates? The momentum of the ball in the roulette was also well define, right? But you make a very big distinction between position and momentum.
That was the ontology. A 2-dimensional description with ##x,y## and a 1-dimensional description with ##\theta## was just convenient descriptions, not the ontology. The particle was really there in 3 dimensions, not in 2 or 1 dimensions.
I honestly don't see how these are different. You simply declare it to be so. Somehow for you the Cartesian coordinates ##x,y,z## are real, but any other equivalent description is not. It's just convenient maths.

I don't see how any of this is supposed to be about learning what ontic is. You just say this here is ontic and those over there are not.

Gold Member
The coordinates are not real physical stuff, the particle is. The coordinates are part of the mathematical representation. They are not made of atoms. It doesn't matter if they have exact values at any given time. They are still part of the map, not the territory.
I knew you will say that. And you are right. But that's a part of learning of the concept "ontic", through gradual refinement. The particle is real and its coordinates are not, as you said. But the point is that, among all mathematical objects in the theory called classical mechanics, the object ##x(t)## is closest to the real particle. This is the object that best represents the ontology, even though it's not ontology by itself. Or, if you know a mathematical object in classical mechanics that even better represents the real physical particle, I'm ready to listen and to refine my understanding. Any idea?

For instance, you may propose that an even better representation is a coordinate-free differential geometry where the particle is a point on a 3-dimensional manifold representing physical space. I'm fine with that. But the idea is that it is a point on a three dimensional manifold, not on a lower dimensional manifold. If that would satisfy you, I'm fine with that. But I just don't feel that the difference between a coordinate-free definition of a point and its coordinate representation is a big difference. When I say ##x,y,z##, I always have in mind a point in coordinate free 3-dimensional space.

Last edited:
martinbn
I knew you will say that. And you are right. But that's a part of learning of the concept "ontic", through gradual refinement. The particle is real and its coordinates are not, as you said. But the point is that, among all mathematical objects in the theory called classical mechanics, the object ##x(t)## is closest to the real particle. This is the object that best represents the ontology, even though it's not ontology by itself. Or, if you know a mathematical object in classical mechanics that even better represents the real physical particle, I'm ready to listen and to refine my understanding. Any idea?

For instance, you may propose that an even better representation is a coordinate-free differential geometry where the particle is a point on a 3-dimensional manifold representing physical space. I'm fine with that. But the idea is that it is a point on a three dimensional manifold, not on a lower dimensional manifold. If that would satisfy you, I'm fine with that. But I just don't feel that the difference between a coordinate-free definition of a point and its coordinate representation is a big difference. When I say ##x,y,z##, I always have in mind a point in coordinate free 3-dimensional space.
Ok, this I understand and I think that we don't have a different views, just different phrasing of them. But it gets me back to why give it a name. The point is to be able to describe the behavior of the particle. What does it matter if I use one mathematical model, or another even if one is more a faithful representation of the existence of the particle?

But it gets me back to why give it a name. The point is to be able to describe the behavior of the particle. What does it matter if I use one mathematical model, or another even if one is more a faithful representation of the existence of the particle?
Regardless of what interpretation we are into, I expect a more faithful that may yield more insight and add explanatory value.

Another difference between representations is that they may have various requirements in representation and computation, and thus be more or less fit. This would add insight in the agent model, as an agent while doing inference, is still contrained by limited resources. For this reasons it seems reasonable that nature would choose (evolve) a compact and efficient interactions and way to code information. This is also at least in my view, the line of reasoning that will help guide "scaling" of physical law, from the low energy to the higest possible energy. At any scale, one would expect some sort of "optimum" representation, such a representaion would hopefully exhibit som isomorphism to the agent (which mean regular matter).

This would be in contrast to various ad hoc renormalisations, where the regulators are just mathematical tools, having no correspondence. In what I think about, it's the agents inference capacity, that is the natual regulator, if we can give that a physical interpreration, that would really add insight and explanatory value I think.

This is in the sense, in which I see that the choice of mathematics differs. Some models also contain an obvious redundancy, that is not enlightning as you can to deal with it. The contiuuum model is a good example of such haze, that adds no explanatory value at all.

/Fredrik

See e.g. the books by Durr on Bohmian mechanics and quantum foundations, or the book by Norsen on quantum foundations. Durr says that ontology of a theory is what the theory is really about. In #8 I explain that classical mechanics is really only about positions.
Thanks for those great references. In "Bohmian Mechanics" by Dürr and Teufel (2009) on page 38, I found something similar to your explanation in #8:
Here is another point one may think about from time to time. Although all variables needed to specify the physical theory are “real”, there is nevertheless a difference. In a particle theory, the particle positions are primitive or primary variables, representing what may be called the primitive ontology. They must be there: a particle theory without particle positions is inconceivable. Particle positions are what the theory is about. The role of all other variables is to say how the positions change. They are secondary variables, needed to spell out the law. We could also say that the particle positions are a priori and the other variables a posteriori.
That explanation (and its context) is slightly more in line with investigations about "ontological commitments" of theories that I encountered before. The goal there is more on arguing which parts are "necessarily ontic" than on identifying which parts are "not ontic" (or rather "not necessarily ontic").

Of course there can be parts that would better be "not ontic", for example the set of all sets in set theory, or the wavefunction in Bohmian mechanics. But momentum doesn't belong to that category, despite being only a secondary variable in classical mechanics, and despite being even less ontic and slightly annoying in Bohmian mechanics.

But when it works, then it even gives an algorithm for determining which sentences are true (or false):
Here I risk to create a similar type of confusion, because that algorithm determines the sentences which are "necessarily true" (or the sentences which are "necessarily false"). It doesn't change the fact that there can be sentences which are true for one model and false for another, i.e. neither necessarily true, nor necessarily false.

Last edited:
Demystifier and Minnesota Joe
Gold Member
Ok, this I understand and I think that we don't have a different views, just different phrasing of them. But it gets me back to why give it a name. The point is to be able to describe the behavior of the particle. What does it matter if I use one mathematical model, or another even if one is more a faithful representation of the existence of the particle?
The point is to develop some intuition which is not purely mathematical. Perhaps someone who finds easy to think very abstractly as in modern algebraic geometry does not see it that way, but most physicists like to have physical intuition behind the equations in physics. The idea - that some mathematical objects describe the real physical stuff more directly than the others - helps to develop such intuition. Perhaps it's not helpful to you, but it's helpful to many. So it's nice to have a special word for that intuitive notion of real physical stuff.

If it's not helpful to you, fine, then you don't have to use it. But even then, understanding its meaning may help you to understand others when they use it. Just because it's not useful for you doesn't mean that those who use it are talking nonsense. If the notion of ontology helps someone to make a computation and measurable prediction which all people in the community can understand, then the notion of ontology is a useful tool. And for me, that's just a tool. The purpose of a tool is not to be true, its purpose is to be useful. For me it's useful. For those who grasp its meaning immediately is probably useful too. For those who struggle to understand it it's probably not very useful, but even they can benefit from understanding it because then they can more easily understand those who use it.

You may compare it with learning foreign languages. You don't learn a foreign language because it's better than your language. When you learn a strange phrase in the foreign language, you don't complain that it doesn't make sense. You try to absorb it anyway, because then you can more easily communicate with those who use that phrase naturally. The same should be with learning the word "ontic". It may be a part of a strange foreign language for you, but you can learn it anyway.

Last edited:
Lynch101, Fra and gentzen
Gold Member
What is different is @Demystifier 's "osmotic" method of learning the meaning of ontic. New to me.
Yes, osmotic, I like the way you phrased it.

Life is ontic and is applied biology. Biology is applied chemistry. Chemistry is applied physics. Physics is applied mathematics. Mathematics is a model.

Is mathematics just a model?

physika
TeethWhitener