The Nature of Causation

What is the cause-and-effect relation like, precisely? What happens, for example, when a brick goes through a window, and the window shatters? Since Hume’s time, most philosophers have taken causation to be a relation between two events. They would say that, strictly speaking, the brick does not cause the window to break. Rather, it is the brick’s moving that causes the window’s breaking. This seems to make sense, but I want to suggest another way of thinking about things.

First, let me be more precise about what events are. An event is:

  • an entity’s coming into existence or going out of existence; or
  • an entity’s gaining some property, losing some property, or simply having some property.

The last alternative on the second line is sometimes called a state in contrast to an event, but here I don’t think the distinction is important. (In this definition, I talk about entities having properties, but I don’t think properties actually exist: they are abstractions that we use to describe the different ways in which things can exist or resemble each other.)

An event is a rather ephemeral thing to have the power to affect reality. Events are not exactly abstract objects, but they are also not things that can exist on their own: events depend for their existence on the things that they happen to. The entities that events happen to are more fundamental than the events themselves. And when I consider it, it seems that it is the things that exist in the world around me, not the events that happen to them, that have the capacities and dispositions to cause changes in reality.

So I would say that it actually is the brick that breaks the window. The brick, rather than the event of its moving, is what causes the window to break. And complementary to that, the brick’s moving is the situation in which its capacity to break the window is realized. So we can say that the window breaks because the brick was moving in such a way (giving an explanation for an event), but what causes the window to break is the brick itself (citing the cause of the event).

My view, then, is that causes are concretely existing entities, and effects are events that happen to those entities. These kind of entities have certain causal powers or liabilities, capacities to affect or be affected by other entities. Powers and liabilities are paired: every causal power entails that something has a causal liability to be affected by that power. And those powers or liabilities may require that certain events happen in order to be realized.

(Note: I don’t have a precise term for the kind of entities that can be causes and that events can happen to. But what I have in mind are things like bricks, windows, photons, people… in other words, the kind of non-abstract things that come to mind most readily when you think about what exists.)

My view about the nature of cause-and-effect is mostly based on intuition, I’ll admit, but it is the view that makes the most sense to me. It coheres with my belief that it is the things in the world that push and pull on one another, not the events they are in. It readily incorporates my powerful intuition that I cause things to happen in reality: I am writing this blog, I am pouring this glass of water. (Writing is thirsty work, you know.) And it grounds cause-and-effect relations in the capacities of things that exist in the world around us.

One argument against this view is that you can’t properly form causal chains if causes and effects are not the same kind of thing. The usual view is that both are events, so that one effect can become the cause of another effect, and so on.

But you can still form causal chains on my view: the entity that an effect happens to can become the cause of another effect, and usually we talk about chains like this when the first effect put that entity into the situation where its capacity to cause the second effect was realized. So I don’t see anything wrong with thinking about causation in this way.

What Kinds of Causation are there?

What kinds of causal powers and liabilities exist in the world around us? I think we can put cause-and-effect relations into three (maybe four) categories.

First, there are deterministic causes. These are the kind that we are most familiar with from the basic physical sciences: a billiard ball strikes another billiard ball and sets it in motion; a sugar cube dissolves upon being stirred into a cup of coffee. In these kind of cases, given the circumstances, no other outcomes were possible. When something is caused deterministically, the effect is completely fixed by the situation that the cause happens to be in.

Second, there are indeterministic causes. Interpretations of quantum mechanics differ on whether this kind of causation is a fundamental part of reality or not; some would say that the laws of physics do completely determine the future state from the initial state, others say that they only determine probabilities that different future states may be realized. But an example of indeterministic causation would be atoms being caused to decay by the interactions of their underlying quantum fields, without the cause determining exactly when it decays.

When something is caused indeterministically, the effect is not fixed, but is a random outcome from a range of possibilities. What may be determined is the range of possible outcomes and the various probabilities for them.

So causes can be deterministic or indeterministic; I appear to have exhausted the options here. Is there a third category? I think we have, or at least seem to have, intimate experience with a third kind of cause: self-deterministic causes. This is the kind of causation that is going on when we act by our own free will and choice. If, for example, I am sitting on a committee and I raise my hand to vote on some motion, I cause myself to raise my hand. But I wasn’t determined to do this by my environment, and it wasn’t just a random effect that happened to me. (At least, that is the way it seems.) Rather, I had certain reasons for raising my hand, and I acted on those reasons.

Deterministic causes, and indeterministic causes where the effect is a random outcome from a range of possibilities, may both be thought of as passive kinds of causation. The entities that stand in these kind of cause-and-effect relations are not active participants in the events of nature, but rather are passively subject to the forces that arise from whatever situation they are in.

In self-deterministic causation, on the other hand, the cause is an agent, someone who acts for reasons, goals, or purposes of their own. Self-deterministic causation is inherently teleological, that is, purpose-oriented. And usually, self-determined causes (in the way that I am presenting them here, where the cause is an entity and not an event) are explicitly thought to not be determined to cause the effect that they do by the situation that they find themselves in. These kind of causes could also be called active indeterministic causes, contrasting with the passive indeterministic causes I described earlier.

There is, of course, serious philosophical debate about whether there can be such a thing as a self-determined cause whose effect is also determined by the situation it finds itself in (active deterministic causes, the maybe-fourth category I mentioned above), whether actions that are so determined can be counted as free actions, and just what the relevant sense of being free is that is necessary for assigning moral responsibility. And there is serious philosophical debate about whether self-deterministic causation is possible at all.

I will weigh in on those matters at a later time, as I continue to explore the nature of reality. For now, I will just say that if self-determined causes do exist, reality is very, very different than it would be if they did not exist.

I say this because it is generally thought that physics leaves no room for self-deterministic causation. Electrons and photons do not have goals or purposes of their own, and they do not exercise any active power over each other; the entities of physics on their own are subject only to passive causation. And if apparent instances of active causation, such as our very familiar activities of choosing to walk here or there, or to speak certain words, or to read and think about certain topics, could all be reduced to the motions of fundamental physical particles, then active causation, and agents themselves, never really existed in the first place.

Following that reasoning in the other direction, if self-determined causes do exist, then physical reality is not all there is. So this is an important topic to consider, and I will be coming back to it a few posts down the road.

Cause and Effect

In my last post, in exploring the nature of the physical world around us, I raised the question of whether we are justified in accepting our best scientific theories as true. Weighing the arguments, my conclusion was that, in principle, science is a valid way of obtaining knowledge. In fact, I think it is a very good source of knowledge.

But, we do still need to evaluate our scientific theories to be sure that we are justified in accepting them as true. And in order to understand and clarify what our scientific theories are saying, we need to look a little further at the nature of reality itself.

In this post, I will start doing that by exploring the nature of causation, or, cause-and-effect.

Cause and effect are fundamental concepts in how we explain the world, both scientifically and in the everyday. Try to explain why something is the way it is, and sooner than later you’ll have to talk about what caused it to be that way. (The other kind of explanations that we generally use are conceptual instead of causal; for example, explaining that something is hot because its molecules have high average kinetic energy, and that is what heat is. But then the obvious question to ask is why its molecules have high kinetic energy, and so on.)

There is a sense in which, strictly speaking, the empirical sciences do not actually explain anything. Using science, we discover certain mathematical equations that we call scientific “laws” – but these laws are really just descriptions of the patterns that we see repeated in nature. In order for scientific laws to explain anything, we need to say why nature displays the patterns that it does. And at least to me, it seems to make the most sense to say that nature displays the patterns that it does because the various things in nature stand in certain cause-and-effect relations with each other – or, to put it another way, they have certain causal powers and liabilities to affect and be affected by each other – that are such that their interactions create those patterns.

That is a bit of a round-about way of saying that scientific laws do not explain why anything happens the way it does. Rather, these laws are made true and explained by cause-and-effect relationships existing between the fundamental physical entities that our universe is made of. Cause-and-effect, not mathematical description, is ultimately what explains how our physical universe works. But by making this ultimate explanation, we have left the realm of empirical science, and ventured into metaphysics.

Is Causation Primitive?

The philosopher David Hume held a view of causation that has been hugely influential in the philosophical thought of the last two centuries. Hume challenged the idea that we can have any knowledge of cause =-and-effect at all. He thought that since cause-and-effect is not something we can “see” with any of our five senses, it must just be something we infer from always seeing one thing follow after another (like fire following after a struck match, or the motion of one billiard ball following after its being struck by another). And since this kind of inference is not deductively valid, he thought that our belief in cause-and-effect relationships in the world is not rationally justified.

The dominant theories about causation since Hume’s time have been reductionistic, trying to break down the concept of causation and explain it entirely in terms of something else. For example, Hume has been interpreted as saying that cause-and-effect just is nothing more than the constant conjunction of some things following after certain other things. Others have attempted to reduce causation to certain processes or probabilities, often defining causation in terms of counterfactual conditionals involving such concepts (for example, defining “A causes B” to mean “if A had not occurred, B would have been less likely to occur” or something similar).

Some have even questioned whether science supports the notion of cause-and-effect at all. Those in favour of eliminating causation would say that we can just explain the way the universe works by citing the relevant scientific laws. The laws themselves are unexplainable on this view – they are just how nature works. The difficulty of isolating the causes of an effect from the whole operation of the universe is also used as an argument for eliminating causation. However, I don’t think that difficulty is insurmountable. And, as I’ve already said, science is not the only source of knowledge – and it is perfectly valid to interpret scientific theories in terms of cause-and-effect relations.

Contrary to the attempts to reduce or eliminate it, I think we have good reason to believe that causation is primitive. That is, cause and effect are fundamental, irreducible features of reality. Rather than causation being reducible to other features like processes or probabilities, those features ultimately depend on causation.

First, belief in causation is properly basic, just like belief in the existence of the physical world. Causation may not be the kind of thing that can be perceived by our senses, but we still experience it. Our belief in cause-and-effect forms naturally, without any conscious inference – this can be seen by the fact that there are “causal illusions,” situations where it looks like there is a cause-and-effect relation occurring, even when we know that it actually is not.

For example, say you are playing a computer game like Age of Empires. It looks like your cannons are damaging the enemy fortifications, even when you know that both the cannons and the fortifications are just images on a screen. There is no actual relation of cause and effect between the two images – both are just common effects of the program running on the computer. Illusions like this show that our causal beliefs can be wrong, but the principle of critical trust implies that we should not be so skeptical as to think that they are always wrong.

In addition, belief in causation may not be arrived at by deductive reasoning, but it can be arrived at by abductive reasoning. We explain things by citing causes, so an inference to the best explanation is often an inference to the existence of a cause, and a cause-and-effect relation.

The Humean view is that, if causation is primitive, we cannot have any knowledge about it, since all we can observe are sequences of events. But this objection is undercut by the validity of abductive reasoning, and the fact that belief in causation is properly basic. We do not need to reduce causation to something sense-perceptible in order to know that it exists and operates in the world. Hume’s view is false.

(As a side note, Hume believed what he did because he thought that all knowledge had to be derived from our physical sense perceptions in order to be valid. But that belief itself is not one that can be derived from sense perception, so it is self-defeating.)

Second, causation is too central of a concept to eliminate or reduce. Causation seems to be required for understanding how reality works and how we interact with it. Various philosophers have used causation as a central feature in theories about diverse subjects, such as ethics, epistemology, semantics, the philosophy of mind, the metaphysics of identity and persistence through time, and the metaphysics of possibility and necessity. Eliminating causation would undermine much of our understanding and knowledge.

Furthermore, it can be argued that attempts to reduce causation end up being circular, because cause-and-effect relations are implicitly contained in the concepts of process and probability, which is what causation is supposed to reduce to. Most reduction attempts involve counterfactual conditions, but it seems like causation is a key notion in determining how to evaluate such counterfactuals. And there are other problems with these attempts, suggesting that causation is just too basic of a concept to analyze.

Third, some philosophers have suggested that we can conceive of scenarios that are identical in every respect except for the cause-and-effect relations that hold. It is only metaphysically possible for these scenarios to be distinct if causation is primitive. Imagine that:

  • there are three kinds of fundamental particles, A, B, and C;
  • B particles can pass through A particles and each other; and
  • a B particle has a 50% chance of causing an A particle to decay to a C particle upon passing through it.

Now, suppose two B particles simultaneously pass through an A particle, one moving left and the other moving right, and the A particle decays into a C particle. It seems like there are three possibilities:

  • the B particle moving left caused the decay;
  • the B particle moving right caused the decay; or
  • both B particles caused the decay.

Even though our universe does not appear to have A, B, or C particles, it certainly seems like such a universe is metaphysically possible. But reductive accounts of causation cannot distinguish between these scenarios, so their possibility is a problem for the idea that causation is ultimately reducible to something else.

So I think that it makes sense to take causation as a fundamental, irreducible feature of reality. In my next post, I will go on to write about what I think causation is actually like.

Scientific Realism

In my last post I argued that we can know that the physical universe exists. And everyone reading along was probably thinking, “Well, duh. What amazing insight will he share next? That the sky is blue?” But hey. The tagline of my blog does say that I’m building a belief system from the ground up, and I really meant that.

Whether we can believe that the universe exists is one question; what we are justified in believing about what the universe is like is another. Specifically, we use science to investigate the world around us, and our best scientific theories postulate a plethora of entities that are unobservable to our ordinary senses. Electrons and atomic nuclei are far from accessible to our perception – wavefunctions, quantum fields, and curved space-time even more so.

So are we justified in believing that such unobservable entities are real? Or are they merely useful fictions, conceptual devices for predicting the way the world works at the level that is accessible to us? This is the question of scientific realism.

Scientific realism is the philosophical belief that science can and does discover truth, or at least approximately true theories, about objective reality in a rationally justifiable way. Scientific realism is committed to the correspondence theory of truth, and to the claim that the theoretical entities postulated by science – like electrons and protons – are real things that exist, rather than just being conceptual devices for making predictions.

As an example, consider the sentences “this apple is red” and “the average family has 2.5 children.” The first sentence refers to a real thing that exists. You would verify its truth by looking at that apple and seeing if it was red. But you would not try to verify the truth of the second sentence by going out into the world to find “the average family” and seeing if that family had two and a half children. “The average family” is a conceptual device; it signals that the sentence reports the result of a certain mathematical calculation involving families and the number of children they have. Scientific realism says that the statement “electrons have negative charge” is more like the first of the above two sentences than the second.

On the other side is scientific anti-realism, which denies that science discovers truth about the aspects of reality that are not directly accessible to us. Instead, scientific theories are merely useful fictions, devices that let us predict the way the world works at the level that is accessible to us. Some forms of anti-realism even deny that science is a rational pursuit at all.

The Case for Realism

I can only really scratch the surface of the debate about scientific realism in this blog. Look here for something more comprehensive, if you’re interested. I will only try to give a brief overview of the main arguments for and against realism.

Arguments for scientific realism:

  • Realism is the most natural position to hold, and is justified by our common-sense intuitions about rationality and abductive or inductive inferences. Anti-realism is often associated with a degree of skepticism that is unwarranted, at least according to the kind of epistemology that I have described in this blog.
  • The success of science in making empirical predictions, explaining our observations, and enabling us to manipulate the world around us, is best explained by the truth (or approximate truth) of scientific theories.
  • Scientific theories often explain a wide range of phenomena as a product of a single underlying mechanism, and the success of such explanations is unlikely if they are not true (or approximately true).

Arguments against scientific realism:

  • The history of science shows us that theories do not need to be true to make correct predictions, and that even highly successful theories have been replaced and are now thought of as false. The success of scientific theories is explained by their empirical adequacy, not their truth. (Though the realist would probably say that a theory’s empirical adequacy is still best explained by its approximate truth.)
  • The evidence we have in any given field can be explained by multiple different theories, which are all empirically equivalent at the level of things we can observe, but contradict each other at the unobservable level. Since they are empirically equivalent, we have no reason to believe one theory over the other.
  • Since no scientist would claim that our best theories are exactly correct, the realist position relies on the notion of approximate truth. But the meaning of approximate truth is unclear: something is either true or false; truth is not a matter of degrees.

The scientific realist can respond to the first objection by being more selective in considering what kinds of theories, and what aspects of those theories, are deserving of realist attitudes. For example, he could say that only mature, well-developed theories are candidates for realism. Such theories have made novel empirical predictions that have been verified by experiment (e.g., general relativity predicting gravitational lensing), or they have explained phenomena that they were not specifically tailored to explain (e.g., general relativity explaining the anomalous precession of the perihelion of Mercury).

The realist could narrow the field further be saying that we should only have realist attitudes towards certain aspects of a theory, such as just the entities that are indispensable for the theory’s explanatory power, or just the relations encoded in mathematical structure of the theory. Being selective in this way helps explain how discarded theories could be successful without being true, which takes away some of the sting of the anti-realist’s objection. And it can often be pointed out that discarded theories were successful precisely because they had an element of truth to them. For example, the epicycle model of the solar system was false, but it was still true that the planets they knew of back then had nearly circular orbits.

The scientific realist can respond to the second objection by saying that we believe scientific theories not just because they fit the evidence, but because they are the best explanation of the evidence. Explanatory virtues, such as simplicity or plausibility, help to decide between empirically equivalent theories. Against this, the anti-realist may express doubt about whether we are rationally justified in using inference to the best explanation. They would say that at best, explanatory considerations can give us pragmatic reasons for choosing one theory over another, but not rational justification.

However, the rejection of abductive reasoning is a radically skeptical position, and I don’t think there is any principled way to limit that rejection to scientific inferences alone – if inference to the best explanation is invalid in scientific inferences, then it is invalid in a broader range of cases as well, and this undermines the rational justification for a significant number of the beliefs that we rely on every day. Probably, it would even undermine the reasons that the anti-realist had for rejecting abductive reasoning in the first place! So I don’t think that objection to realism is warranted.

Regarding the third objection, I think it is possible to give a meaningful account of the concept of approximate truth. I haven’t done much reading in this area, but here is my idea of how we can do that:

Let’s say it’s true that (1) my keys are on the kitchen table. Then it is also true that (2) my keys are in the kitchen, and (3) my keys are in the house. Statements (1), (2), and (3) form a sequence of true statements decreasing in specificity.

Now, consider the statements (4) my keys are on the kitchen counter and (5) my keys are on the coffee table in the living room. Both (4) and (5) are false. But (4) entails both of the true statements (2) and (3), while (5) only entails (3). That is, (4) entails a more specific true statement than (5) does. It seems to me that we could consider (4) to be more approximately true than (5) for that reason.

Obviously, more work needs to be done here in clarifying the relevant concepts of entailment and specificity being used here. But if I am right, and we can define the degree of approximate truth of a statement, compared to other statements, in terms of the degree of specificity of the true statements that it entails, then the third objection to realism that I considered can also be overcome.

Evaluating the Case for Realism

I think the case for scientific realism is strong – we have good reason to believe that at least some important aspects of our best scientific theories are true, or approximately true. At the same time, the anti-realist arguments need to be addressed, which is why I qualified the last sentence in the way that I did. We aren’t automatically justified in accepting everything about all of our scientific theories, but we have to sift through them to find out what aspects of what theories are justified by the evidence and by abductive reasoning.

So, if a scientific theory is the best explanation of the relevant data, and especially if it has been successful in making novel predictions or explaining phenomena that it was not specifically tailored to explain, I think we are rationally justified in accepting it as approximately true. But, we should be cautious about theories, or aspects of theories, that suffer from problems that can weaken the rational justification supporting them. These kind of problems include:

  • The theory has conceptual problems, either because it is unclear what it actually means, or it has internal inconsistencies, or it conflicts with other sources of knowledge. (E.g., the different interpretations of quantum mechanics.)
  • The deeper hypothetical aspects of the theory do not actually serve to explain the relevant phenomena, but are used more as mere conceptual devices for making calculations. (E.g., virtual particles.)
  • Non-rational factors account for much of the theory’s acceptance. (See Roger Penrose’s lectures “Fashion, Faith, and Fantasy” for examples of this in modern physics.)

A crucial part of evaluating scientific theories is understanding what they actually mean, so that we can see if they can be fit into a coherent picture of reality. As an example, consider the modern physics of quantum field theory. While it allows scientists to make accurate predictions of the results of high-energy particle collider experiments, scientists don’t really seem to know what this theory tells us about what our physical reality is ultimately made of. Clarifying that is an important task in the philosophy of science.

On the other hand, if we zoom out a little bit, I think we can definitely say that things like electrons and protons and neutrons exist, and that they behave in the way that quantum theories say they do, even if we don’t really know what they are. But to see what we can tell from science about the nature of reality, we are going to need to do – you guessed it – more philosophy. So that is what I’ll be doing for the next few posts.

The Physical Universe

Does the physical universe exist? Can we know that it does?

Given that our sensory experience of the physical world, like other forms of conscious experience, is a ground for properly basic belief, we have very powerful justification for answering yes to both of those questions. Our sensory experience is probably the most well-established of all types of experience. In order to reject all of our sensory experiences as mistaken, we would need very powerful reasons to disbelieve in the existence of the physical world.

So the first and foremost reason to believe that the physical world exists is simply that we experience it.

The second reason is that the existence of the physical world is the best explanation for the fact that we experience it. (This is subtly different from the first reason; it justifies the belief using abductive reasoning, rather than grounding it directly in experience in a properly basic way.) Alternative explanations for our sensory experiences are called skeptical hypotheses, and they include these kind of scenarios:

  • Everything you experience is really just a dream.
  • You are really just a virtual person inside a computer simulation of reality.
  • You are a brain in a vat being given false sensory signals.
  • The universe began five minutes ago with all the appearances of age, including false memories (this scenario specifically targets knowledge about the past, rather than all knowledge of the external world).

For the most part, the only thing these hypotheses have going for them is that they are possible – we may have reasons to think that they could be true (though even that is questionable for some of them), but we don’t actually have any reasons to think that they are true. Because of that, they all seem significantly less plausible than the common-sense hypothesis that the physical world actually exists.

In addition to that, these skeptical hypotheses often lack the power to explain why we have the experiences that we have, without being ad-hoc relative to the common-sense hypothesis. The dream hypothesis, for example, just gives no reason for why we would have a consistent and coherent dream of a reality that did not exist, while the common-sense hypothesis can say that our experiences reflect an objective reality outside ourselves.

Some skeptics would argue that there are skeptical hypotheses that are better explanations for our sensory experiences than the existence of the physical world itself. For example, take the dream hypothesis again. It is in some sense a vastly more simple theory than the common-sense hypothesis. Rather than requiring the whole universe and the various entities in it, the dream hypothesis only needs the dreamer.

But simplicity is not the only explanatory virtue: the lack of plausibility and explanatory power puts the dream hypothesis below the common-sense hypothesis, and things are similar with other skeptical scenarios. So in addition to accepting the existence of the physical world as a properly basic belief, I think we are also justified in believing it as the best explanation of our sensory experiences.

As a side note, I am, perhaps, slightly misrepresenting these skeptical hypotheses. Their point is not to show that we should believe any of them. Rather, their point is to show that it is possible that all our experiences could be deceived, so we can’t know that the physical world around us is real. But I think that is based on a misconception of what knowledge requires. We do not need to prove that we can’t possibly be wrong about a belief in order to know it.

Philosophical Idealism

Positions which hold that there is no physical reality, or that we cannot know that there is a physical reality, are known as various forms of philosophical idealism. (This has nothing to do with the more common sense of the word, the practice of forming or pursuing high ideals.)

Epistemological idealism is the belief that we cannot know that anything is real aside from our subjective perceptions. It holds that we can know about the internal mental world, and maybe the world of abstract objects, but we cannot know that the physical world even exists. Aside from using skeptical hypotheses like the ones above to undermine our confidence in the physical world, epistemological idealists often argue that since our subjective experiences are the only reality that we can directly access, they are the only reality we can meaningfully talk about.

I think that my reasons to believe in the physical world, based on the principle of critical trust, are sufficient to deny epistemological idealism. And the argument that we can only meaningfully talk about our subjective experiences simply does not follow. Here’s an analogy. In one sense, you could say the only things we can really see are the photons that hit our retinas; but that does not make it meaningless to talk about where those photons came from. Even if it’s true that we only have direct access to our subjective perceptions, we can still know about the physical reality that we perceive through them.

And further to that, epistemological idealism especially (though some forms of metaphysical idealism have this problem as well) is susceptible to lead to an absurd conclusion: solipsism, the belief that you are the only person that exists, or at least, the only one you have ever met. The reason for this is that if you are unable to access any external reality, there is no way you could ever interact with anyone aside from yourself. There has to be some kind of objective, external reality for other people to exist, and it has to be accessible in some way for there to be a medium of interaction and communication between different persons.

To put this another way, I think interpersonal interactions could count as a well-established type of experiences, different from sensory experiences of the physical world. Even though it overlaps with the latter, it is a different type because there is a different feature that unifies such experiences: they are about people rather than the physical environment. Face-to-face, phone, and email interactions with another person are unified in a way that merely looking at something in front of you, listening to a recording of a sound, and reading some text are not. There are different introspective states associated with interpersonal interactions as well, such as recognizing or predicting the mental states of others as different from your own.

So via the principle of critical trust, interpersonal interactions give us good reason to think that other people exist and that we can interact with them, against solipsism. But since it seems intuitively true that there must be an objective, external reality in order for such interaction to be possible, that gives us a good independent reason to believe that such an external reality exists, against epistemological idealism.

Metaphysical idealism is the belief that the physical world does not exist, and that reality is exhausted by the mental world, or maybe the mental world together with the abstract world. A metaphysical idealist might hold that all that really exists are minds and their perceptions, or that minds somehow create the objects of their perceptions through the act of perceiving them. A metaphysical idealist does not have to be an epistemological one: he could believe, for example, that minds communicate telepathically, and can access an objective, external reality (other minds) in that way.

Variations of this philosophy have argued that there is something contradictory or incoherent about the existence of physical reality, in order to rule it out as impossible. Alternatively, it may just stem from idea that we have no good reason to believe in the physical world. You can probably guess at this point that I don’t find any such arguments to have much force.

A popular version of idealism (which can be either epistemological, or metaphysical, or both) that has been around in the past few decades is motivated by a certain interpretation of quantum mechanics. This position holds that, since quantum objects exist in superposition and the wavefunction collapses into a definite state only when quantum objects are observed, it is actually observation that creates reality. Quantum mechanics reveals there is no objective, perspective-independent reality, only subjective perceptions. Or so the idealist would say.

The only problem is that quantum mechanics does not tell us that observation creates reality; one interpretation of it does. There are other interpretations that say nothing of the sort. For instance, it could be interaction with a macroscopic system (that is, one with a very large number of entangled degrees of freedom) that causes wavefunction collapse, rather than conscious observation.

So the argument can just as easily go the other way. Through our experience we know that the objective physical world exists; this particular interpretation of quantum mechanics implies that it does not; therefore this interpretation is false.

So in summary, I think we can be confident in our common-sense belief that the physical world around us exists in objective reality, and that there are not comparably good reasons to believe any of the various forms of idealism. Thus, I believe that the physical universe exists.


In the previous post, I talked about the debate between platonism, the belief that abstract objects really exist, and nominalism, the belief that they do not. To defeat platonism, the nominalist needs to explain how we can talk about abstract objects, such as numbers, without committing ourselves to their existence. I believe the answer is that, most of the time, when we talk about abstract objects, we are not really intending our statements to be taken completely literally.

NOTE: See the notes on my previous post for how my view has shifted on this issue; since writing this part of my blog I have come to think that I must explore the basic structure of reality much more deeply than I did originally, and the result (when I try to formulate what I mean when I say things like “abstract objects like properties are means of speaking about the ways that things exist,” as I originally did below) actually ends up in some ways looking like platonism more than nominalism. I still think figuralism has some important insights, insights related to the true nature of abstract objects. More to come.

The theory that we are speaking figuratively when we talk about abstract objects is known as figuralism, and it has been recently developed by the philosopher Stephen Yablo. Yablo distinguishes between the literal content of a statement, and the real content:

  • Literal content: the conditions which make the statement literally true.
  • Real content: the conditions which make the statement true according to the way that reality is to be imagined in interpreting the statement.

Yablo refers to “the way that reality is to be imagined in interpreting a given statement” by calling it a game. The game being used is an implicitly understood background of any statement. (For literal statements, the game is just “interpret this statement by imagining reality as it actually is.”)

For example, when we say that someone has a heart of gold, the game we are using is something like “interpret this statement by imagining that the quality of a person’s character is reflected in the material composition of their heart.” Essentially, part of understanding figurative language is understanding, often subconsciously, the game that is being implied by the statement and its surrounding context.

Different games, different ways of imagining reality, can lend different interpretations to a statement. And in conceptualizing reality in a figurative way, abstract objects can be used both as representational aids and also as things to be represented. For example, we can use numbers as representational aids to count objects, but we can also treat numbers as objects that can be counted.

A nominalist can say and mean all of the following four statements, and figuralism gives a reasonable way to interpret all of them so that there is no contradiction:

“There are no numbers.”

  • Numbers are things represented.
  • Reality is conceptualized as the nominalist believes it actually is (with no abstract objects, and therefore no numbers, actually existing).

“The number of numbers is 0.”

  • The occurrence of “number” (singular), and the number 0, are representational aids.
  • The occurrence of “numbers” (plural) has numbers as things represented.
  • Reality is conceptualized as the nominalist believes, but the representational aids being used to talk about it are imagined according to our concept of numbers.

“There are infinitely many numbers.”

  • Numbers are things represented.
  • Reality is figuratively conceptualized as if numbers exist the way we imagine them to be when using them as representational aids

“The number of numbers is infinity.”

  • The occurrence of “number” (singular), and the term “infinity,” are representational aids.
  • The occurrence of “numbers” (plural) has numbers as things represented.
  • Reality is again conceptualized as if numbers exist.

The first two statements communicate what the nominalist actually believes about reality, while the second two statements communicate what the nominalist believes about the concept of numbers, speaking as if they actually exist. The platonist disagrees with the first two statements, but the nominalist, the platonist, and the mathematician (who may be a platonist, a nominalist, or neither) can all agree on the second two statements, in the same sense that the nominalist intends.

On figuralism, the real content of the statement “1+1=2” is not that there really exist these things called numbers, and that they add together and equal each other in such a way. It is just a much less wordy way of expressing that if you have one object, and you have another object, then you have two objects.

Arguments for Figuralism

I think figuralism explains and unifies several facts about abstract objects:

  • We often use abstract objects in speech without noticing it; we see through statements that are literally about abstract objects, to what the speaker is really expressing. For example, if I said “the number of zebras has doubled in the past year,” you probably wouldn’t think about the fact that literally speaking, this sentence is about a certain number, rather than being about zebras.
  • Even if a trusted, infallible source of knowledge revealed that abstract objects did not really exist, we would still all speak as if they did. If you said “the derivative of this function f(x) is…” and someone claimed that your statement was false, because there are no functions, you might even say “stop being so literal – you know what I mean when I talk about mathematics.”
  • Abstract objects are insubstantial – there is nothing more to their nature than what is required by our concept of them. It would be ridiculous (unless you have synesthesia) to ask what colour numbers are, or how much they weigh, or whether they are happy or sad. In other words, abstract objects always seem to be exactly what we need them to be to function as representational aids, and nothing more.
  • The fact that our abstract object talk is figurative (or something similar) explains why statements about mathematical objects, for example, seem to be necessary and knowable a priori – because the real content of those statements is logical, derived purely from deductive reasoning.

The main objection to figuralism is that if we have been talking figuratively about abstract objects all this time, we should have known that. How can we be unaware of when we are talking figuratively instead of literally?

The answer is that figuralism does not require our abstract object talk to be completely analogous to the more noticeable, deliberate forms of figurative speech. All it requires is that we can talk about abstract objects as if we believed that they exist, without it being necessarily the case that we find we actually believe they exist, upon reflection about it. That is, all figuralism requires is that we simulate belief in abstract objects, not that we deliberately speak figuratively about them.

There are other scenarios where this kind of thing happens. For example, we go about our everyday lives acting and even thinking as if the earth is stationary, even when we know that it is revolving on its axis. When we think about the sun’s changing position in the sky, we don’t always feel the need to remind ourselves that this is a result of the earth’s rotation. Usually, we don’t even think about it. We simulate belief that the earth is stationary, without actually believing that it is stationary.

And the fact that conceptual metaphors pervade the way that we think and talk shows that completely literal language is not the norm. So it seems entirely plausible to me that figuralism is correct, and moreover, it is a useful way of understanding why our talk about abstract objects is the way it is.

The Abstract Realm

So does the abstract realm exist? Are there any reasons for thinking that abstract objects are real, mind-independent things? If figuralism is correct (and I believe it is), the arguments I presented in favour of platonism are no longer justified. We don’t need to believe in abstract objects for any explanatory advantage, and in that case, Occam’s razor suggests we should not believe in them.

Furthermore, since abstract objects have no causal powers to affect reality, nothing about their existence could lead us to have knowledge about them. We could come to know about them in other ways: through intuition (though, given figuralism, such intuition could be reinterpreted as being about the real content of statements about abstract objects, instead of the literal content), or maybe through divine revelation. Or we could have knowledge about abstract objects because reality is such that every possible abstract structure exists, so whatever concept we have of an abstract object, it refers to something.

But without some other reasons to believe in abstract objects, it seems simpler to take the nominalist position: abstract objects, and the abstract realm that they inhabit, do not actually exist. Rather, they are useful ways of thinking and speaking that expand our ability to understand reality.

A Side Note

Since I am now making the claim that I don’t see any good reason to believe in the existence of abstract objects, it is only fair that I remark on the fact that I have already made some use of certain kinds of abstract objects in discussing what I believe about reality. Specifically, in explaining the correspondence theory of truth, I used propositions and states of affairs, both of which can be construed as abstractions.

However, I don’t consider the correspondence theory to require the actual existence of propositions or states of affairs as abstract objects, any more than I think that doing mathematics requires the actual existence of numbers somewhere out there. Truth “fits” reality while falsehood “misfits” reality; propositions and states of affairs are just ways of speaking to clarify what we mean when we say that.

And, for the reasons I have outlined in this post, I believe something similar can be said for every other kind of abstract object there is.