"Real"

Quite a lot of jousting goes back and forth about what is "real." I will claim that the determination of some phenomenon as "real" is often a stand-in for questions like "does x matter," or "should we pay x any mind." Furthermore, I'll claim that the fact of the matter is nuanced enough that often whether something is "real" doesn't really help us answer these questions.

How an algorithm feels from inside and disguised queries

I'll bring in two concepts from Eliezer: how an algorithm feels from inside¹, and disguised queries².

Most concepts are nebulous, but still track real structure in the world. Human intuitions about membership in a category often construct membership as an extra fact, besides all of the facts about an object's properties.

In the proverbial "if a tree falls in the forest, but no one is around to hear it," there are two properties which are normally tightly correlated, namely an event causing acoustic vibration, and it causing aural perception/experience, that have been decorrelated. This question is sort of about asking, "but did it really make a sound?" as though there were a question left to ask, besides the two component phenomena. Yes vibrations, no perception, simple as. To be clear, I don't think anyone is actually vexed by this question, but it's helpful to lay it out very precisely, because indeed I see this exact mistake being made with respect to this class of problem in general.

Another example: are viruses alive, or for that matter prions? Viruses do not have a metabolism and can't reproduce outside of a host, but do contain genetic material, and do evolve. Prions are much simpler, they are misolded forms of ordinary proteins, they don't contain genetic material, and they only reproduce by causing normal proteins to misfold. Again, we might ask, "but are viruses really alive?", as though there were a meaningful answer to this question, beyond the facts of which properties they have in common with most kinds of life, and which they don't.

We can construct a category in such a way that marginal cases are strictly speaking not members, but we haven't really learned anything by this. There is a large diversity of observed biological systems, and gerrymandering "life" to exclude some of them hasn't told us anything new about the ones we excluded, nor would it to arbitrarily choose to include them.

In Eliezer's explanation, his claim is that we might imagine choosing one algorithm (or really, data structure) or another to represent some concept, and that the intuitions that humans have, in which this "but is it really?" question arises, would happen in some kinds of representations and not others. As well, these kinds of ontological intuitions are constructed in his view out of cognitive algorithms, and that the intuitions are "what the algorithm feels like from inside of it."

Eliezer's claim in Disguised Queries is that questions like "is x in category y" are often stand-ins for "is x dangerous," or "is x reasonable," or "is x important." A category is presumably predictive of some other properties, besides those that define the category. Often people will haggle over whether x should be included in a category, in order to be able to determine that we should infer some other property about x. For example, we could (well, at least in principle) imagine someone trying to argue that Gandhi was a criminal—which is absolutely true, strictly speaking! Of course, our imagined speaker here is really trying to argue that Gandhi was dangerous, or immoral, or whatever. This is a silly example, but this is absolutely a move that people make with respect to categories like racism, violence, and terrorism. These kinds of questions are called "disguised queries," since the real question which is being debated is disguised by a question about where to draw a boundary.

Let's assume, though we don't necessarily need to assume this, that the reality described by physics is stably "real" in a straightforward sense. I'll make a few claims:

• There's a fairly large diversity of phenomena which we call "real" (or at least which we regard as real) that are ontologically very different from physical reality.
• Many of the details of physics are quite counterintuitive, and the deepest questions about what's "real" are still contentious in physics.
  • Physical solidity (as in, of wood or stone or bone) is quite a bit subtler and stranger than ordinary human-scale intuitions about "solid matter", though my understanding is that the popsci schtick that matter is "mostly empty" is somewhere between oversimplified and false.
  • Famously, the cosmological interpretation of quantum mechanics is still contentious even in academic physics. Furthermore, both interpretations currently in the ring get openly metaphysical, and in either case are quite contrary to day-to-day intuition.
• I don't believe that there's an account of what's "real" which includes far-flung notions like romantic love, money, status, mathematics, or music, that doesn't also end up including a pretty wide swath of meditative and spiritual experiences.
  • Perhaps, we can hypothesize various grades, or at least a diversity of disjoint types of reality, which is broadly the stance I like, but then we end up with "real" as something either much weaker or much narrower than people usually want it to be. (Also, gosh, wtf, multiple realities? Is that really how far we've stooped?)
    • In any case, this doesn't really help, because there still doesn't end up being space to argue that all this spiritual stuff is just "not real."

I want to be very clear here: I am not claiming that nothing is real, or that reality is completely socially constructed, or that everything which we call "real" has the same ontological status. I am claiming that "real" is both nuanced enough and diverse enough ontologically that it must admit spiritual experiences, one way or another.

And what about this assumption that physics is real?

Look, if we're going to throw that out then you're just really screwed.

More seriously, if we don't assume a basis of reality like "physics," then whatever sense of reality we're working with is yet weaker, and so my broad take is even stronger.

I suppose this obviously leaves out gnostics, or various "mind-only" or "mind-first" metaphysicses. I'm not really writing for those people, but in any case I think they're already going to agree with some claim like "reality is more nuanced and ontologically fussy than modern secular metaphysics would like."

Let's take some specific problematic cases:

• Math: There have been a variety of theories with respect to the ontological status of mathematics, over millenia. Math seems both more inescapable than physics, but also completely intangible. We can imagine (and, it's now often hypothesized) worlds with different laws of physics, but we largely can't imagine "worlds" with different arithmetic, or calculus, or topology.
• Music: Is music real?—sorry, wait what? Since when is music even the kind of thing that can be real?—well, we care about music, we devote quite a lot of economic productivity to it, we argue about what music is right, and some people dedicate their lives to the practice of it. For most people it's often much more salient and powerful than math or history or certainly theories of physics. On the other hand, it's not quite like most things we would ordinarily call "real." It's just sounds, often without words, in some pattern, how could that matter? How could that possible mean anything?
• Love: Yes yes, "true love" and all that; but in any case, once again, is love real? What would it mean for it to be "real"? Sometimes people feel bamboozled or manipulated by love, and there's certainly deception and fraud presented as though it were love. On the other hand, love, whether romantic, familial, love of one's country, etc., seems like one of the most important forces in world. People will give their lives for it, they'll work themselves to the bone for it, etc. etc. etc. Love is also richly imaginal: tied into fantasy, imagery, and elaboration that are both more salient, powerful, and motivating than what's actually in front of our faces, but also never quite present "in the here and now."
• Money: The value of money is certainly "rational" in a clear sense: macroeconomic shenanigans aside, you can generally expect that you will find a counterparty willing to accept your dollars for most kinds of transactions. Still, money isn't quite just "made out of physics," or at least not just physics. There's a large, only-somewhat-centralized cooperation in the construction of a coordination mechanism, which relies on expectation across the network that the rest of the network will continue to exist, and continue to behave largely the same. Of course, money is quite real! Money is incredibly salient and powerful, almost anywhere in the world. So we have both a very high degree of indirection, and something to the effect of "social construction," which is meanwhile quite "real," in the sense of "important, and worth worrying about."

So, again, if a tree falls in the forest, but there's no one around to hear it, is it real?—sorry, again, there's a variety of properties, some in common with other things which we take as real, some not, and no extra fact about whether it's actually real.

Ontological and spiritual commitments

An ontology, among other things, serves as the framework through which we understand what exists, and what can matter at all. Descriptively, most people have commitments to particular ontologies: moving from one ontology to another often inspires a fear that what one now understands to be real and will become meaningless, wrong, or harmful, and vice versa. This is often correct! One metaphysics will often, in fact, reify or valorize notions which are fake or irrelevant in another.

These kinds of ontological commitments might be made by way of organized religions, ideologies and philosophies, or may be cultural and implicit, idiosyncratic, and even completely unspoken. Some examples would be things like reductionist materialism, Karma³, Heaven and Hell, souls and spirits, and even persistent personal identity.

To be clear: this list is not to say "these are utter nonsense, and everything that flows from them is wrong." Rather, mostly I want to inspire some sort of ontological flexibility and a skepticism of over-reified systems of metaphysics. (Still, as an empirical claim about cosmology, I'm pretty confident that Heaven and Hell are not real.)

Most of the point here is to say, a. that the ontologies that people use are largely unjustified (though in the next section I'll get to why this is so hard), and b. that in general, these domains are nebulous enough that there isn't going to be a singular, stable ontology on which to build all of our theories of what matters or is worthwhile⁴.

The naive view which says "physics!" or "science!" in response to questions of ontological nebulosity simply can't make these problems go away. But, still, there are things which aren't real! Certainly, there are lies, and there are dreams, and there are hopes that do not come true.

Often, scientism and atheism maintain a commitment that entertaining many ways of seeing is "wrong", since these are "not real." Except, again, a. what's real, but more importantly, b. how have you determined that the collection of ways of seeing that you use are correct, or the only ones that should be used?

Now, my real question is: what should we make of both claims and experiences of prayer, meditation, psychedelics, mysticism, etc.? I think these are real questions, I mostly don't have a clear answer. One of the main claims of this book is something like, "actually, I still don't have a clear answer, but it's definitely not any of these."

Note that I'm not claiming that eg. those who report talking to God are definitely talking to a persistant physical being etc. etc. Mostly, I think these experiences often have very substantial existential value, and can't be excluded. However, determining which are baloney and which are actually worthwhile is much harder. I discuss this more later.

(Though, I do have a very small probability on some spiritual reports having a straightforwardly literal ontological/cosmological interpretation. Again, I discuss this more later.)

I'll note, that if I say, "well, there's something interesting going on metaphysically with respect to spiritual stuff," this isn't yet to say that you should regard it as important or valuable, but just to say that it can't be flatly determined to be not real, just by writing it out of your map.

I'll discuss this all in much more detail in the section on spirituality and spiritual practice.

"Real" and "True"

One last bit here: I'll claim that "true" is nuanced in basically the same sorts of ways as "real" above. What's true? Just mathematical theorems? What about very high confidence predictions, from scientific theories? What about just high confidence predictions? And, which theories, or is it mostly about their being socially acceptable, and endorsed by the current priesthood? What methods are even acceptable for determining what's true? What does it mean, if anything, for introspective or spiritual claims to be true?

Footnotes

1. Eliezer Yudkowsky, 2008. How An Algorithm Feels From Inside ↩

2. Eliezer Yudkowsky, 2008. Disguised Queries ↩

3. This is usually left out of bowdlerized western Buddhisms, but Buddhist Karma traditionally amounts to something to the effect that the goodness of actions is ontologically/cosmologically basic. I think it's a bit hard for us to appreciate both how weird and how intense such a perspective would be, if fully taken on. ↩

4. Chat, is this real? ↩