What does “existence” mean?

During a recent online discussion I discovered, somewhat to my surprise, that there is no general agreement on what the word “exist” means. Everyone has an intuitive understanding of it but when it comes to an explicit definition of the word there is no consensus, and indeed philosophers have written a vast literature on the topic of ontology, or what exists.

Dictionaries don’t really help; for example Oxford Dictionaries gives a nicely circular set of definitions:

Exist: 1. have objective reality or being
Reality: 1. the state of things as they actually exist
Being: 1. existence.

Of course physicists have a perfectly good operational definition: something exists if it is capable of making a detector go ping. Try arguing that, however, and you’re immediately accused of materialism, physicalism, scientism, being blind to possibilities beyond a very narrow world-view, and a host of similar sins (I plead guilty to at least the first three).

Many want to believe in the existence of things that are non-physical and do not make physicists’ detectors go ping. Such people avoid any explicit definition of what “exists”, preferring to rely on human intuition. I’m often puzzled by the weight given to intuition in philosophy. Yes, it is a guide, but we know that human intuition is horribly suspect and fallible, and unreliable unless corroborated by evidence. As H. Allen Orr says: “a scientific education is, to a considerable extent, an exercise in taming the authority of one’s intuition”.

I thus thought it worthwhile to take the physicists’ “ping” definition and see how it compares with philosophers’ ideas about ontology. Refining the definition a little, I arrived at:

Something “exists” if there is a potential chain of causal links between it and our sense data.

This is an empiricist’s stance, placing our own sense data at the root of our knowledge of the world. Even a pinging detector is useless unless we get to know about it. To adapt Descartes: “I feel therefore I am, and things that cause me to feel are”.

Of course we hardly ever experience anything directly, since even when looking directly at something we are experiencing the photons bouncing off it, not the thing itself. Hence we know about something only because it has affected a thing which affected another thing which affected something else … and ended up affecting sense data. Hence the “chain of causal links” in my definition.

The causal connections need only be between adjacent links in the chain. There might not be a causal connection between the ends of the chain (between the thing itself and our sense data) since, owing to the finite speed of sending a signal along the chain, the signal might not be able to travel that far.

For example, our universe has an “observable horizon” beyond which we can not know about, since even light-speed signals will not have had time to reach us given the finite age of the universe. Presumably, though, stuff still “exists” out there, even if we can’t see it (horrible consequences follow if that isn’t the case, including a violation of the Cosmological Principle, and the fact that material might go out of existence or come into existence as it crossed our observable horizon, as if we mattered).

The word “potential” is included in the definition since many possible causal effects will not, in practice, be realised. Lastly, I should probably remove the word “is” from the definition of “exists” and make the definition iterative: Start with the sense data; Thing 1 potentially affecting it then “exists”; Thing 2 potentially affecting Thing 1 then exists; and so on along the chain. Such a chain is, of course, our mechanism for discovering the world around us.

The above is a definition of whether something “exists”, not of whether we know that it exists. Obviously we can’t know about anything existing beyond the observable horizon, and we can only know about what has produced actual causal effects, not what just has the potential to do so.

The philosophical literature on existent and non-existent entities is extensive, but let’s simply apply the above definition and see what happens.

Do composite objects, such as crystals or fruits or stars “exist”? Yes, these exist, since they are patterns of material, and detectors can be designed to go ping about particular patterns of material. Indeed, this is always the case, since nearly everything is a pattern of lower-level stuff (molecules being patterns of atoms, which are patterns of nuclei and electrons, and the nuclei are patterns of nucleons, which are patterns of quarks and gluons, which … well we’re getting beyond established physics to go further).

What about ideas and thoughts and emotions, do they “exist”? Does pain exist? Yes, since these are manifest as patterns of matter (patterns of electrical and chemical impulses in our brains) and, as above, patterns of matter can quite readily make detectors go ping (an fMRI scanner can read thoughts, though only very crudely as yet).

What about mathematical “objects”, such as an idealised perfect circle, or the number 7, do they exist? Not in their own right, but conceptions of these do exist, as ideas in our brains.

What about properties or attributes, such as the colour red? Does the colour red “exist”? (This is the philosophical Problem of Universals.) Things coloured red certainly do, and the conception of “red” also does in our brains. But “red” doesn’t exist in its own right. This is the philosophical stance of conceptualism, with a long history back to at least William of Ockham.

What about classes of objects, such as the collection of all members of a species, does the class, the species, exist in its own right? I’d say, no, though the concepts certain do, again as patterns in our mind.


Now, suppose one proposed a new type of non-interacting physical particle which interacts with other such particles but has no interaction at all with the known things in the universe. Could that particle be said to “exist”? A similar postulate is that there are faeries which can never be seen or heard or interacted with at all.

One response is to shrug and declare the question uninteresting as unfalsifiable and literally of no consequences. Under the above definition, though, they clearly (ex hypothesi) don’t “exist”.

How about causally disconnected parallel universes in a proposed multiverse? Again, ex hypothesi, causally disconnected parallel universes would not “exist” (though one could invent the concept that they might “para-exist”, along with the unseen faeries and the non-interacting particles). Note, by the way, that some multiverse models propose multiple connected universes, and these universes would exist in the same way that matter beyond the cosmological horizon does.

What about the multiple worlds in Everett’s interpretation of quantum mechanics? Since these go causally disconnected immediately on splitting they would not “exist”.

So that is how a philosophically naive physicist might see “existence”. So far I’m not aware of significant flaws in this way of defining “existence”, nor of any better general definition of the word — but if there are I’d be interested to know about it. Of course those not into scientism would have a fit of the vapours over a definition that is close to: “If we can’t see it then it doesn’t exist”.

Update: Since writing the above I’ve changed my mind somewhat and decided that I need to be more relaxed about composite entities such as species. As above, an animal, such as a deer, “exists” as a composite object. Starting from the “make a detector go ping” definition, one can design a detector with a pattern-recognition device that goes ping when it sees a deer. But it then follows that I could program the device to go ping only if it saw three deer or indeed a whole herd of deer. Further, I could program it to go ping if, and only if, it saw a deer followed by a sparrow followed by mouse. So really I must allow that fairly abstract collections of entities — including “deer then sparrow then mouse” and certainly including “species” — do indeed “exist”. If I don’t then the whole scheme will fall apart.

How about a song? Well, each instantiation of a song would have a physical existence (either as a memory in a brain, or as a recording on some hardware, or whatever), and we do indeed have devices that recognise particular songs. So following the above reasoning the set of all such instantiations, “the song”, does exist. So perhaps I should allow that any set of extant entities, however arbitrarily defined, does “exist”.

20 thoughts on “What does “existence” mean?

  1. Pingback: What does “existence” mean? | As Ric3Hard

  2. fojap

    Reblogged this on fojap and commented:
    I rarely reblog posts, but since I just wrote something in which I declared myself to be a “materialist” and found myself hoping some naive person doesn’t come by and “challenge” me with a silly question like “do thoughts exist”, I have decided to preemptively counter that possibility by reblogging this post.

  3. David Evans

    Does this view imply that Everett’s interpretation cannot be true, or even that it is logically incoherent? By your account, his other universes would not exist even if they did exist.

  4. Pingback: A scientism defence of Logical Positivism | coelsblog

  5. Richard Wein

    Hi Coels. I just got here (to this old-ish post) via your comment at Scientia Salon.

    I would say that the view you express here is based on a somewhat misguided way of thinking about language. Wittgenstein famously said that philosophy is a battle against the bewitchment of our language by our intellect. Unfortunately, I would say that much philosophy remains pretty damn bewitched.

    You gave the following definition: Something “exists” if there is a potential chain of causal links between it and our sense data.

    First, I would say that this sort of definition does not tell us the meaning (or sense) of the word. It just gives us a necessary and sufficient condition for using the word, which is not quite the same thing. I for one would be wary of even calling it a definition. More importantly, it seems to be circular, since it uses the word “is”, a near-synonym for “exists”. You could equally have said “Something exists if there exists a potential chain…”. To make any sense of this we must assume you are employing two different concepts of existence, in the two places. The second sense must be a broader sense than the first, and you’ve told us nothing about that broader sense.

    Competent speakers understand the word “exists” perfectly well, without any need for a definition, at least when applied to familiar sorts of cases. Problems sometimes arise when philosophers invent weird cases, but in those cases we should question whether the philosophers’ usage is even meaningful.

    Take the case of mathematical concepts, like numbers. Mathematicians use the term “exists” fruitfully. They may say, for example, that there exists an integer between 2 and 4, namely 3. That’s a useful way of speaking, and there’s no reason to tell mathematicians that they’re wrong. There’s no need to try to force mathematical usage–which is talking about pure abstractions–into the same mould as the scientific and everyday usage–which is talking about the real world. We can leave these useful established usages alone. Problems arise when philosophers ask questions like, “Do numbers exist?” In my view this question is meaningless. The mathematical usage makes sense within its normal context, in which we’ve adopted a certain axiomatic system of thought (whether formally or informally). But the philosopher steps outside that system, and tries to ask whether numbers exists in some external, metaphysical sense. In doing so, he deprives the word of any sense, since there is no established metaphysical sense for him to use.

    We should avoid assuming that every question is meaningful and answerable, or that words always have the same meaning. We should ask, “Is this question meaningful in this context?” It’s also useful to ask what the alternatives are, e.g. what difference does it make whether we say numbers exist or don’t exist. Similarly, what is being asked if someone asks whether emotions exist? Surely no one wants to suggest that we never feel any anger, hope, fear, etc. But apart from that absurd possibility I don’t know what is being asked. Part of the problem here, I think, is that we don’t usually use the word “exists” in connection with emotions. We usually talk about “feeling anger” or “being angry”. Introducing the more philosophical-sounding word “exists” may encourage us to think that we have found some meaningful new question to ask: “Yes, you feel angry, but does your anger exist?”.

    Now, suppose one proposed a new type of non-interacting physical particle which interacts with other such particles but has no interaction at all with the known things in the universe. Could that particle be said to “exist”?

    What particle? You seem to be saying, in effect, “If such a particle existed, could it be said to exist?”. I think a more pertinent question would be, “Is there any point in proposing the existence of such a particle?” Perhaps theoretical considerations of fundamental physics might give us a reason for doing so, but I wouldn’t like to speculate.

    1. Coel Post author

      Hi Richard,
      Thanks for your thoughtful comments. Here’s some brief replies: I agree that people have an intuitive understanding of the term “exists” and we could indeed leave it at that without further enquiry, but it seemed to me worthwhile to try for a more rigorous analysis of the concept “exists”.

      I realised the “is” problem when writing the piece, so a few paragraphs on from the main definition I gave a version without using that word. I do agree with much of the rest of what you say. A mathematician saying that the number 3 “exists” is using the word in a different way, and I think most mathematicians would recognise that. On the question of “does anger exist” I’d say yes it does, as a pattern, a particular pattern of firing in the brain, in the same way that a tornado exists as a particular pattern of air.

      Finally, on this point:

      You seem to be saying, in effect, “If such a particle existed, could it be said to exist?”.

      I wouldn’t phrase the question that way, I’d more be asking whether the property of being causally linked to the visible universe is a necessary part of “existence”.

  6. osolev

    I assume that the author of this write-up a male with normal testicles.

    Now, here is the test to come to certainty that he exists, in particular when he has another male with normal testicles.

    First, they each observable to the other squeeze very hard with their hands their each one’s testicles, or use a pair of pliers in each hand to grip hard on the testicles, when they want to arrive at greater certainty of their testicles’ existence.

    Do that until they are already feeling excruciating pain in their testicles.

    That should make each of them certain that their testicles exist.

    Now to make each one sure that the testicles of the other also exist: they now grip hard the testicles of the other, until both are feeling excruciating pain in their testicles caused by the other, on each other.

    Now, that is one test for the certainty of existence, and what is existence, in regard to testicles.

    Now do the same exercise on anything and everything that you doubt to exist, like your nose, which you can even cut a small slit in one of the nostrils until it bleeds.

    Now, with things which are not parts of your body, like the bullet in a gun, shoot with the loaded gun’s barrel pointed at and even in contact with your laptop and fire.

    There, that proves to you the existence of your erstwhile good laptop.

    I am sure you now know what is existence, and that you exist and things exist even when you are not around.

    I like you to think on this sentence from me:

    “The default status of things in the totality of reality or being is existence.”

    Go to: http://forumserver.twoplustwo.com/137/religion-god-theology/how-come-god-existing-thinking-truths-facts-logic-1623068/

  7. thennicke

    Hi Coel,

    Interesting article, and a very important topic. I have a question:
    Isn’t causality a perceived phenomenon? Wouldn’t it make a lot more sense to take it out of the definition and say, simply, “something exists if you can perceive it”? If perception comes first, then it doesn’t make sense to talk about causality until you’ve seen a billiard ball hit another and cause it to move or whatever. That explains why, as somebody pointed out above, there lies a circular “is” in your phrase, “Something “exists” if there is a potential chain of causal links between it and our sense data.” I can see why you wanted it in your definition but it seems to break with strict empiricism. Let me know if I’m misunderstanding things of course,


    1. Coel Post author

      … and say, simply, “something exists if you can perceive it”?

      The problem is that most things we cannot perceive directly, we only perceive them indirectly. Even if you’re looking at a tree in your garden, what you’re really doing is perceiving photons that have bounced off the tree, not the tree itself. But, if you allow for indirect perception, then that’s more or less the same as me talking about chains of causal links. The “photons bounced off the tree” is a one link in the causal chain.

    2. thennicke

      Ahh, so this is a disagreement about what constitutes perception. The question is, do we perceive the stream of photons itself, or do we perceive patterns/forms within that stream of photons (e.g. a tree)?
      Because we are talking about sight specifically, we might as well just rephrase this: do we “see” photons, or do we “see” a tree?
      To me, the answer seems to be *both*; it’s a false dichotomy. This is because visually, a tree is just a pattern of photons.
      The point is that you don’t have to talk about this in terms of causality; just in terms of basic abstraction. And if you take the causality out of the definition, you can avoid circular reasoning.

      By the way, I really think you are on to something when you “adopt Descartes”; I honestly think that, in terms of radical scepticism, “I perceive therefore I am” is far more useful than “I think therefore I am”, because you have to exist to be able to think, and “existence” is defined with respect to perception – if you don’t have perception, you can not exist, or even talk of existence. Nor can you think. As I see it, Descartes was frustratingly close, but people like Hume got it right and triggered the scientific revolution.

      The fundamental difference between the “time” before you were born, and now, is that of perception. Perception is the only reason you can be sure that you are conscious; in fact, every word in this sentence is defined with respect to the stream of information that constitutes perception. It’s also why science is so powerful: it taps into the very fabric of existence.

      This is about the fundamental difference between empiricism (e.g. Hume) and rationalism (e.g. Descartes). Or, to put it another way, between foundationalist knowledge, and the “web of knowledge” that you describe elsewhere on this blog. I think the latter is more accurate (kudos to you for promoting it), and that if we take it to its logical extension, we can’t talk about causation until we’ve perceived.

      Again, let me know if you spot any errors here. Thanks.

    3. Coel Post author

      The question is, do we perceive the stream of photons itself, or do we perceive patterns/forms within that stream of photons (e.g. a tree)?

      But the patterns within the photons are not the tree, they are an *image* of the tree. The tree itself is made of wood and leaves and is not just a pattern of photons. That’s why one has to add in the concept that the pattern of photons is caused by the tree.

    4. thennicke

      What if you close your eyes and put your hands on your head? It’s easily possible to infer its round shape (a form of pattern), despite the absence of your most useful sense.

      I was only talking in terms of pure sight to simplify things. Wood and leaves are also patterns of sense data: not just patterns of reflected photons, but patterns of all our senses put together.

      A tree is not something that can only be seen. It is something that can be touched, that can be heard when something is struck against it (a crude form of scientific experiment), etc.

      So yes, patterns of photons will form a *visual* image of a tree, but similarly, patterns of atoms will form a *tactile* image of a tree (assuming you put the right sensory organs in the right place obviously). And so on for the other images (sonic, olfactory, gustatory, etc.) When you put all those images together, you get a tree.

      So I would maintain that a tree is an abstraction, and its inference has nothing to do with causation.

      Especially when you consider that causation itself is an abstraction (albeit one that we perform with incredible ease): how can causation exist if the criterion for existence requires causation?

    5. Coel Post author

      Yes, a tree can be perceived in other ways also such as touching. So let’s consider something a bit more distant such as the (former) planet Pluto and its moon Charon. You cannot touch them and you cannot “see” them directly.

      A feature on Pluto would be perceived by: (1) a Solar photon bounces off it, (2) it gets detected by a detector in a spacecraft, (3) that information gets stored in an on-board computer, (4) it gets passed to a transmitter, (5) the transmitter emits photons, (6) those photons are then detected by a radio antenna, (7) that gets stored in a computer, (8) it gets passed to another computer, (9) it gets processed into an image, and (10) only then does it get looked at and “perceived” by a human.

      Presuming that you accept that features on Pluto exist, and that we can’t perceive them directly, but only via the above chain (which, by the way, is shortened and simplified), how can we get away without including chains of causal links in our notion of “perception”?

    6. thennicke

      I disagree that “You cannot touch them and you cannot “see” them directly.” It might be difficult to do, but it’s possible; you just have to get yourself close enough without dying in the vacuum of space. I guess what you meant is that for these objects, sophisticated devices must be built in order to infer their existence – that we can’t see or touch them directly from Earth. I agree with that.

      But this is still just a case of “putting the right sense organs in the right place”, in the same spirit as I described in my last reply. Except that these “sense organs” are inorganic (cmos sensor, hard drive, transmitter etc.) Think of the satellite, the transmitter and so on as extensions of your own nervous system, with which you can reach out and touch Pluto, in the same way that you can put your hands on your head.

      Yes, your nervous system is a causal mechanism, but we don’t need to know *anything* about how it works (i.e. causal explanations) to receive and process the sensory information it gives us. We simply get the live feed from our senses (and from the technological extensions of our senses) and that’s what we get to work with.

      I would claim that fundamentally, we are pattern recognition machines. This is supported by the fact that you can provide the brain with any kind of information stimulus and it will unconsciously do its best to “make sense” of it (for some amazing examples of this see https://www.ted.com/talks/david_eagleman_can_we_create_new_senses_for_humans).

      The concept that we hold in our minds of reality itself (our “worldview”) is one such abstraction, made sense of out of the information available to us. When we find evidence (sense data) that supports the existence of Pluto, we adjust our worldview to incorporate it – unless we are too emotionally invested in a different worldview, and would like to ignore our own eyes (bias).

      It seems to me that everything you see, taste, feel, etc. is fundamentally just a sea of information. We then take that information and “make sense” of it. The scientists among us go out and actively look for new sensory information (think of this as “extending the nervous system”), so as to help us formulate a more accurate picture of the world we inhabit. Nothing in that process involves causation (well, apart from the engineering of the technology, for which we obviously need causal explanations – how things work – of the world). Causation is fundamentally just another abstracted concept with which we can make sense of (interpret; process) the sea of information..

      I need to also mention that I was misusing the word “perception” in prior posts, as though it were a synonym for “sensation”. It’s not; it refers having to the ability to sense, rather than having the conscious awareness of sensing. And the phenomenon of blindsight shows us that the brain can perceive without sensing, which is all the more bizarre.

    7. Coel Post author

      Except that these “sense organs” are inorganic (cmos sensor, hard drive, transmitter etc.) Think of the satellite, the transmitter and so on as extensions of your own nervous system, with which you can reach out and touch Pluto, in the same way that you can put your hands on your head.

      But in order to create these technological extensions of our senses, and in order to use them properly, we need to know *lots* about how they work and how one thing causes another (such as how a photon causes a change in the cmos detector, and how every other step works).

    8. thennicke

      Yes, that’s exactly right, as I already said above: engineering requires causal (how/why) explanations. We need to know that stuff in order to construct, maintain and use the equipment that brings us new data.

      But that’s a separate issue. I don’t need to know how or why the information gets to my mind to be able to experience and interpret that information.

      We don’t need to know that the big bang caused the expansion of space and created matter that formed stars, which fused hydrogen, throwing off photons which travelled through space, reflected off Pluto and hit a sensor, where they were converted to an electrical signal and transmitted to Earth via more photons, where they were turned back into an electrical signal and printed, where different photons from a desk lamp bounced off that photograph and hit our retina, where they were converted into an electrical signal, etc. etc., in order to *see* an image of Pluto in our visual field. Yes, there’s a causal chain there, but no, it’s not relevant to how we observe and label the world.

      We don’t need explanations to have observations. It’s actually the other way around: you need observations to have explanations. So you can’t invoke causality (explanation) when asking whether something exists (abstracted observation).

    9. Coel Post author

      I don’t need to know how or why the information gets to my mind to be able to experience and interpret that information.

      But if we want to say that the information is “about Pluto”, and thus say that we perceive Pluto — as oppose to just perceiving a piece of paper held in our hand with different shadings on, and not knowing how to interpret it — then we do need to know about and understand the whole chain of causal explanations that link the information back to Pluto.

  8. YF

    Hi Coel,

    I was wondering what are your thoughts on whether or not a particle, such as an electron, exists prior to its measurement. According to some interpretations of quantum mechanics, particles do not exist until they are measured- and even then, some interpretations dispute the existence of particles altogether, suggesting that fields are the primary bedrock of reality. I believe that according to the Copenhagen interpretation of the double-slit experiment it is inappropriate to even ask what is really happening on the way to the detector as there is, in principle, nothing there to describe. Admittedly, I find this interpretation to be a bit unsatisfying, but I was just wondering what’s your take on this issue.

    1. Coel Post author

      Hi Yon,
      First, if fields are indeed the bedrock of reality, then a particle is an excitation of the field, essentially a ripple or wave in the field. So the particle would then be “real”, it would be a real pattern in the field.

      The other question, about the effect of a “measurement”, is harder to answer. Any measurement must interact with the field, and thus must disturb it, and thus the state after measurement is different from the state before. The problem is that any calculation of the outcome of the measurement seems to involve non-locality (“spooky action at a distance”) in order to make the calculations come out right (Bell’s inequalities).

      The Copenhagen interpretation just agrees not to ask about how this happens. Which, as you say, is not very satisfying, and no-one really holds to that nowadays. So there is something about how quantum systems interact to produce “measurements” (aka “collapse of the wavefunction”) that is not currently understood (though it could involve non-local effects, or instantaneous communication, or many worlds, or something else outlandish).

      Having said all that, most physicists would maintain that nature gets on with proceedings regardless of whether there are physicists around doing measurements, so, yes, everything does “exist” prior to a measurement (and Copenhagen is wrong to suggest otherwise), but that there is something we don’t currently understand about quantum mechanics.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s