Stop Ontology Time

For the past couple of weeks, I’ve been writing about philosophy. This led to the single most dramatic drop in the number of subscribers to this blog, but no matter: I’m doing this in part to repent. In my younger years, I bought into the techno-rationalist dogma of the SF Bay Area; with time, I’ve come to realize that the ideology has fueled many toxic behaviors in our industry. Now, I just want get the word out that philosophy is interesting and goes beyond what’s peddled on LessWrong and Hacker News.

One of the core tenets of tech rationalism is the rejection of metaphysics. In particular, the movement frequently explains subjective selfhood and free will as delusions of biological automatons. As I argued before, the approach seems self-defeating: these notions originate from perceptual experience, just as the laws of physics do. To reject some perceptions and take others as axioms is an aesthetic choice, but not a privileged theory of reality. There’s also something funny about the techno-rationalists’ infatuation with unfalsifiable “Bayesian” prophecies and neo-religious beliefs about AI.

More alarmingly, it’s a straight line from tech rationalism to radical utilitarianism — the belief that normative morality is bogus and that there’s no good or evil beyond the wishes of men. We’re good people, so we pick a noble goal — say, to maximize the overall happiness of sentient beings — but on the way to our utopia, it’s only fair to break a couple million eggs. In fact, to realize peak happiness, maybe we need a galactic-scale computer in which trillions of simulated minds — LLMs? — are kept in a state of perpetual bliss.

Few people in the Bay Area want to build the galactic brain. Far more use this thinking to embrace nihilism: since the universe is a joke, might as well just make money with friends.

Dude, are you done?

Right. I’m doing this out of order: I posted essays about metaphysics before really explaining why I’m blabbing about the stuff in the first place. As a reminder, here are the three main pieces so far:

These articles pertain to epistemology. That’s the branch of philosophy concerned with how we acquire knowledge, what’s knowable and unknowable, and what’s the nature of truth.

Epistemology may appear hopelessly abstract, but it’s a fairly productive field. Formal logic, which underpins computer science and mathematics, can be considered downstream of it. Further, where epistemology doesn’t offer definite answers, it usually gives reasonable choices. For example, on the topic of free will and the sense of subjective self, we arrive at three coherent, well-delineated options:

1. Decide that these perceptions are correct and are a consequence of something outside our understanding of the physical realm. This doesn’t necessarily imply god; it implies a limitation of the scientific theory of reality, perhaps akin to the halting problem or Gödel’s incompleteness.
2. Decide that free will and persistent subjective self don’t exist except as illusions. This nets us a nihilistic universe in which all outcomes are predetermined or governed by chance, and morality is strictly utilitarian (if worth talking about at all).
3. Try to reconcile scientific materialism with metaphysics by tinkering with the definitions of the physical reality, selfhood, and will.

But enough about epistemology. Another major branch of philosophy — and the intended subject of today’s article — is ontology. It’s the study of what exists: the kinds, the properties of, and the relations between entities.

The unbearable lightness of being

In the trades, the word “ontology” is most often used when “taxonomy” would be a better fit. Real ontology operates on a different level: instead of cataloging known entities, it seeks answers to more general questions, such as whether non-existent objects can exist.

This is not as stupid as it sounds. Unicorns don’t exist, but the word has a clear meaning, tapping into a sort of a shared mental state. Almost everyone can imagine an unicorn and describe it, no different than if you asked them about a horse. There’s clearly something to talk about here: a unicorn “exists” more firmly than an ershapfult, a word I just made up.

To analyze the issue more rigorously, Alexius Meinong proposed separating the concept of so-being and existence. His “so-being” is the totality of properties other than physically existing; for a unicorn, it may include characteristics such as “horned”, “horse-shaped”, and “magical”. In this view, unicorns may “so-be” without needing to actually roam the planet. Alas, this is not as elegant as it seems: by severing the link to the physical realm, we can now conjure objects that seem incomplete (e.g., the sole property of “blue”) or paradoxical (“blue” and “not blue”), and we have no good basis to deny them the right to be.

Another wonderfully wacky ontological question is the status of holes: is a hole an object of its own? If so, it’s a weird one: it’s a parasite that can’t exist without a host (and can’t nest in another hole). Perhaps it’s more useful to view it as a property of the object in which it appears — but then, every golf hole becomes the property of planet Earth. The third option is to look at holes not as material structures, but as qualified (distinguished) slices of space. Again, not as elegant as it seems: in that last model, we need some mechanism by which space can earn this designation or be stripped of it.

Of course, it’s fair to take a step back: epistemology asks questions that matter. In contrast, not much hinges on the ontological status of holes. Heck, never mind that: is there any chance of a thought experiment that helps us clarify the choice?

The time traveler’s wife

One way to breathe new life into ontology is to consider the question of time. Perhaps the most intuitive understanding of it is that it’s a physical dimension we travel through together. We can’t control our travel, but it’s otherwise no different than moving through space: we used to be there and now we’re here.

This reasoning lends itself to presentism: the belief that all that exists is a thin slice of reality hurtling through time. Yesterday, your cat was in what’s now the past; right now, it’s with you in the present. There are no past-cats left behind: what surrounds the present is just nothingness.

Although presentism makes sense, it also invites two interesting objections. First, it seems like a very inefficient use of spacetime: is it possible that other time-slices exist and are inhabited by sentient beings too? Second, the theory appears to clash with Einstein’s relativity, which tells us that depending on your frame of reference, different observers have different notions of simultaneity. This makes it difficult to describe the geometry of the now-slice we supposedly inhabit.

Another philosophical approach is known as eternalism. It’s the belief that the past and possibly the future already exist; it’s just our subjective self that’s being carried through time. Think of presentism as a stage play and eternalism as a film reel. The latter doesn’t necessarily imply determinism — the film can be an interactive feature, or it might be that the future is still ours to film (also known as the growing block universe). But the key point is that the time dimension isn’t empty: it’s occupied by us.

Yet another way to look at this problem is to posit that time is not a proper physical dimension at all; it exists only as a way to make sense of the perception of causality that’s internal to the observer. In this view, there’s no present or past, except as a mental model that lets us separate causes from effects. That’s a pretty big wrench in the works of scientific materialism, but not a position we can easily discount.

Other than pondering cosmology, one “practical” use of the philosophy of time is to realize that most time-travel sci-fi is somewhat incoherent. It posits that the past and the future already exist — you can travel back and forth — but changing the past somehow changes the present, a ripple-like effect that propagates faster than light. This requires some sort of inherent connectedness between the frames of the film reel. But what necessitates this effect, other than the fact that it’s a convenient plot device?…

The final possibility is that these thought experiments are just gibberish, running into the semantic wall I talked about in one of the earlier articles:

“We start by asserting that concepts have meaning only as they relate to a specific, semantically closed reality. For example, it’s meaningless to ponder if bunnies can exist outside the fabric of space and time. The concept of a bunny — or indeed, the concept of existence — is inseparably tied to the substrate of the universe … In other words, the question makes no sense.”

This possibility led some philosophers, most notably Rudolf Carnap, to divide ontological questions into two categories: internal ontology, which seeks answers within the bounds of well-formalized systems; and external ontology, which is utter nonsense. To illustrate, internal ontology may ask about the properties of prime numbers in the context of the Zermelo–Fraenkel set theory, which formalizes modern math. In contrast, external ontology may ask “do prime numbers exist on the Moon independent of mathematics”, which sounds pretty cuckoo.

That said, many object to Carnap’s quip. For one, it takes away the ability to ask questions we’d like to know answers to. More fundamentally, without asking mushy questions about the nature of space and time and then examining possible answers, it’s hard to arrive at any sort of a formal theory.