Reliable, Therefore Real

[Content warning specific to section III: discussion of gender, sex which remains (I think) in the spirit of social justice culture but does disagree with specific prevalent ways of articulating that spirit]




Source: PhotoJeff at

I have no interest in any definition of the word real which isn’t useful. If I cannot use the word real to distinguish between unicorns and horses, Conan the Barbarian and Conan O’Brien, Iceland and Dinotopia, then it is of no use to me. I want to say that Iceland is real while Dinotopia isn’t—or, more accurately, I want to say that I wish Dinotopia were as real as Iceland is. Therefore as I try to define this word—real—I want to make sure I can use it.

Usefulness may seem like a silly and fringe-case qualification, but it isn’t. An example that a friend recently provided for me: a physicist tried to convince him that the table they were talking over was not real. He did not tell me the details of the argument, but I can imagine them. The table is mostly empty space, with an arrangement of molecules filling out a tiny proportion of the “table”; there is no firm distinction, on the level of fundamental particles, between the table and the air around it. Even those particles are not “real,” being made up of—strings, I guess? I don’t really know.

An example that I take more seriously is that of personal identity. In his I Am a Strange Loop, a book about consciousness and personal identity as a complex, self-referential system of smaller parts, Douglas Hofstadter argues that emergent phenomena or epiphenomena—events or “things” that are the product of smaller things interacting in particular ways: a stampede is an epiphenomena made up of cattle—are not really real. Rather, “table” and “I” and “Canada” and “library collection” are convenient fictions which allow us to reason about the world at our own size; it is silly for librarians to catalogue groups of particles rather than books. He makes much of a marble that does not exist: when he grabs a stack of envelopes, it feels as though there is a marble suspended in the middle of them (because when put together the tiny bits of glue on the point of each envelope’s flap create a less compressible clump which feels like a marble), but when he shuffles through the envelopes there is no marble. He returns to this example often, but I think the example is a bit of a cheat. There is no marble, no; but, when those bits of glue are stacked together, there is a clump which is less compressible than the surroundings. That clump is real. It just isn’t a marble.

When I read I Am a Strange Loop, I wound up agreeing with almost everything Hofstadter wrote—except for his sense that, as a consequence of his claims, “I” do not exist. I’m also sure I would agree with everything the physicist said about the composition of the table whose existence is in question; I would nonetheless disagree that the table did not exist.

So what makes a thing real?

I’ve been working on this problem for about a year now, and I think in conclusion I would say this: a thing is real insofar as it is reliably patterned. Every object we encounter is a pattern, or a pattern of patterns. Arrangements of strings vibrating at different frequencies (according to string theory, anyway) in a certain pattern, fairly reliable over time, create the particles of particle physics; those particles in a certain pattern create atoms; those atoms in a certain pattern create molecules; those molecules in a certain pattern create cells; those cells in a certain pattern create tissues; those tissues in a certain pattern create organs; those organs in a certain pattern create a body; that body operates in such a way that it creates emergent phenomena known as cognition; that cognition in a certain pattern creates me. Is there variation over time, even from moment to moment? Yes, there is; but there is also enough of a pattern that I can make predictions with a fairly high degree of certainty. If that pattern is sufficiently reliable, it is useful for me to acknowledge it.

As far as I am concerned, that is what it means to be real. A thing is real insofar as it is reliably patterned.




Source: [Soren] at

I began considering this issue not from the side of particle physics and personal identity, but through Jorge Luis Borges and his commentary on realism and nominalism. I would love to discuss Borges on this topic at length—indeed, a previous version of this post began with Borges’s discussion of allegories and novels as one instance of a debate, raging through time, between the Platonists and the Aristotelians—but perhaps that can wait for later. I will try a condensed version here.

According to Borges, there are two positions about the nature of reality: the realists believe that categories are the real thing, while the nominalists believe that individuals are the real thing. To a realist (those in Plato’s camp), I am an instance of human; to a nominalist (those in Aristotle’s camp), humanity is the aggregation of all humans.

Borges would say that the majority of people today, when pressed, are nominalists; we do not believe in an Ideal Human of which all humans are mere imperfect instances. I agree, for the most part, but I would add that the majority of people today still carry with them certain realist assumptions, probably due to pervasive cognitive biases. We are, for instance, what psychologists call naïve essentialists. Imagine that a wizard changes a person into a horse; that person is now, in all respects—physical and mental—a horse; later, a good fairy changes the person back into themselves. This story probably makes sense to you. And yet, why should it? How can you change the person back? “The person” doesn’t really exist anymore, right? If you change a person in all respects into a horse, in what way can the person still be said to be the same person? And yet we easily, naturally assume that the person carries with her an indelible identity (in Platonism, that indelible identity is called an essence) regardless of her particular physical features (in most Greek philosophy, those features are called accidents). And, while the vast majority of people are nominalists except in these unconscious assumptions, there are still some avowed realists remaining.

The short version, as I said above, is that individuals are real but categories are constructions. And as a rule of thumb, I think that maxim is a very good thing to assume; most of the work people have done to break down social injustices—against racism, most especially—has been anti-essentialist, which in itself should be a good enough indication of which side to err on. And yet, on closer inspection, nominalism doesn’t work.

If categories aren’t really real, why are they so reliable? If there is no category humanity that exists independently of the various organisms we have grouped together and called humans, why do so many organisms so reliably fit into that category’s general contours? I knew the answer, of course, is that a human’s shape and behaviour are formed from much smaller things which act reliably: DNA determines my features, protein behaviour determines my DNA, molecular physics determines protein behaviour, and so on. At a certain point we hope to find a truly fundamental thing—strings, maybe, or perhaps numbers?—which are actually Platonic objects, relevant only as they are a member of a category and not as an individual. But at all other levels nominalism is truer than realism.

Except… what is an individual? If I want to say that I am an individual and thus am somehow more real than or more relevant than the category humanity, I have to figure out what an individual is. This is where physics and emergence come into play: at all levels but the most fundamental one, what we ordinarily think of as “things” are groups of many smaller things organized in certain ways. What we normally think of as “things” are really groups of patterns arranged so they act in reliable ways, altogether producing an effect which, on our level of perception, has coherence. And if “things” are really groups of smaller things that behave in roughly similar and predictable ways, how is that any different from a category?

The solution to the realism v. nominalism problem, it seems to me, includes acknowledging that individuals are categories. The individual I know as “Christian Hendriks” is a category which contains various instances, across time, of tissues and thought patterns and information-encoding proteins arranged in roughly reliable ways. This is still a specifically nominalist vision, but articulated in different terms: I know longer say, “The individual is ontologically prior to the category”; rather, I say that categories are human constructs, not objects independent of human minds, but also that those categories nonetheless approximate patterns in the world which are independent of human minds, and insofar as the category reliably describes the pattern—and insofar as the pattern is reliable enough to be described—the category is real. And, moreover, as with the nominalists, I would still say that a human is more relevant than humanity, but that is just because the first is a much more reliable pattern than the latter. (Or, maybe more accurate, that a human is a more useful and information-rich pattern than humanity.)

What are the consequences of all this thinking?

I will probably be addressing that question for some time to come, but I can get started with two thoughts, below.



As the friend mentioned at the start of this post pointed out to me, the pattern’s reliability is not the only important thing: the pattern’s relevance also matters. His favourite example is that you could gather together all people with blond hair, blue eyes, and an underbite, and call that a category. It is a reliable category, so it is real. It isn’t useful category, though, the way that men and women have been useful categories.

I would add, though, that it isn’t a useful category in our context; I can easily imagine a context in which it would be useful.

Imagine three people, each with a different type in the romantic-sexual sense. The first person is attracted to thin leggy redheads; the second person is attracted to short curvy brunettes; the third person is attracted to blue-eyed blonds with underbites. In that context, those categories do matter and we could give them helpful names: Lucy’s type, John’s type, and Hal’s type.

The point of that example is that while a thing might be real if it is a reliable pattern, I would only care about its reliability if it were useful for me to notice that pattern. But patterns that I don’t care about might still matter to another person; a useful category for me may be useless to another, and vice versa. “Lucy’s type” might be mostly invisible to me because I do not know Lucy; that does not mean that Lucy does not have very reliable tastes or that the people Lucy finds attractive do not reliably act in a certain way. If I were to get into an argument with Lucy’s friends over these categories, both reliability and usefulness are relevant; in that argument, the distinction between reliability and usefulness is also relevant.

Imagine two men, each with a different type in the romantic-sexual sense. The first man is attracted to people with male primary and secondary sexual characteristics. The second man is attracted to people with female secondary sexual characteristics; primary sexual characters are of no importance to him. Because of the high correlation between primary and secondary sexual characteristics, it is useful for both men to call themselves heterosexual and say that they are attracted to women, but they are using slightly different definitions of the word woman when they say this. It seems to me that if they were to get into an argument about the category woman, what matters in that argument is not whether the patterns they are point out are reliable—both patterns are somewhat reliable, and so both categories are real—but rather whether those categories are useful.1

Of course, the categories man and woman are implicated in many more contexts than just the romantic-sexual one, so we can expect even more conflicts between different definitions, with different stakes.

This is one of the things that frustrates me when people argue that, for instance, man and woman are human constructs and so that biological sex isn’t real, or insist that trans men are (or aren’t) really men while trans women are (or aren’t) really women; or, conversely, that the high correlation between female primary and secondary characteristics means that we should disregard the slightly larger category of people with female secondary characteristics irrespective of their primary sexual characteristics (ie. that only people with uteri are woman). I am frustrated because all competing definitions of man and woman are human constructs, of course, and all of them point to a pattern in the world that is at least somewhat reliable (and therefore real). The issue is not the reality of these categories, or who really falls into which category; the issue is the usefulness of these definitions. What does each definition do?

In my view, as described above, the question is not, What is the truest way to define biological sex?, but rather, What is the most useful way to define biological sex?

And of course the next and more important question is, Useful in what context, and to whom?

I’m not going to answer that here—that is an enormous and fraught conversation which I cannot hope to engage all on my own—but I think this example indicates how defining and understand what real means matters because it changes the way we enter and conduct conversations.



Second, it is embarrassing how long I have been trying to write some version of this post. It’s important to me that I do, however, because for the last year or so it’s been impossible to explain myself to most people; so much of my understanding of the world is now so rarified that I cannot even begin to talk about politics and ethics. At the very least, if I do the foundation work here, I can refer back to it if anything I want to say relies on it. Also, it is helpful to practice explaining yourself, trying to find more efficient ways to convey a worldview.

What is interesting to me is how many dividends this way of looking at the world has started to pay out. At first, as I was trying to figure out individuals and categories and realizing that the problem was in a definition of real, it seemed an esoteric intellectual exercise that would have only a few specific consequences (sex and gender being one of those places today where the conflict between realism and nominalism is especially raw). In the last few months, though, I have been surprised by the number of applications that cropped up.

One of the biggest of these is in the realm of religion. I have not written much about my own religious life here, preferring to write more anthropologically. However, it turns out that “a thing is real if it is reliably patterned” has some overlap with religion. I intend to discuss this in a separate post but for now I will say that intuitions I had trouble articulating became clear in this system while at the same time I have had to change my mind in certain places, becoming somewhat more religiously orthodox than I had been—a surprise, I’m sure, to everyone involved.

Expect further writing to build off of this post.


1. It is not my intention here to represent trans people entirely or particularly with this distinction; I know, and I hope my readers know, that there are various different ways to be trans and furthermore that people with anatomies like those I described may not be trans.

8 thoughts on “Reliable, Therefore Real

  1. Pingback: On Santa Claus and Hofstadter’s Souls | Accidental Shelf-Browsing

  2. Pingback: Clouds, Daffodils, and Jam That Will Not Come Together Again: A Series Index | Accidental Shelf-Browsing

  3. Pingback: Moral Error Theist | Accidental Shelf-Browsing

  4. Pingback: Kegan-Chapman Developmental Model | Accidental Shelf-Browsing

  5. Pingback: Guilt and Shame in the Colossus of Rhodes | Accidental Shelf-Browsing

  6. Pingback: What Would It Mean to Be Autistic? | Accidental Shelf-Browsing

  7. Pingback: A Glut of Tradition | Accidental Shelf-Browsing

  8. Pingback: On Fallibilism, Protestantism, and Woo: 25 Fundamentals | Accidental Shelf-Browsing

Leave a Reply

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

You are commenting using your 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