In a fascinating recent paper, the philosopher Craig Warmke has offered an analysis of Bitcoin that purports to show (i) how it works and thereby (ii) what it is. And he concludes that, contrary to popular belief, Bitcoin isn’t code, but is rather a fictional substance. I agree with Warmke concerning (i), about how the Bitcoin protocol works. In fact, I can say something nicer than that. His model for how the Bitcoin algorithm works is a thing of beauty, and it should be used in courses on blockchain technology to show how consensus is achieved in the Bitcoin blockchain. However, I definitely want to take issue with (ii), his conclusion about what Bitcoin therefore is. It isn’t a fictional substance. And more to the point, the model that he uses to accomplish (i), doesn’t really get him to his conclusion in (ii). This having been said, I agree that “Bitcoin isn’t code.” But it isn’t fiction, either. It is as real as you or me, or at least it is going to be.
Before we get into the heart of the matter, I need to make a couple of preliminary observations. Preliminary observation one: Warmke never said that “Bitcoin is Batman,” or at least I don’t think he did. That slogan comes from my friend Brady Dale at Coindesk. And I take it that what Brady means is that Warmke’s thesis is that, like Batman, Bitcoin is fictional — albeit a work of collaborative fiction. And who knows, if we include Batman fanfic, maybe Batman is a work of collaborative fiction too. As I said, I intend to argue that this is just not the case — that Bitcoin is not a fictional substance — but before I get there, a second observation is in order.
Preliminary observation two: One problem with saying that Bitcoin is a fictional substance is that we really have no clear understanding of what fictional objects and substances are. Warmke discusses this, but I don’t think he puts enough ashes on his forehead when he does so. That is, I don’t think he comes to grips with how big an issue this is. Put 10 philosophers in a room, and they will give you 100 theories of what fictional objects are. Maybe fictions are nonexistent objects, or maybe they are possibilia, or maybe they are pretenses, or maybe there are no fictional objects. Maybe fictional names are just definite descriptions in disguise or maybe there is a big operator IT-IS-A-FICTION-THAT taking scope over all the fictional discourse to follow. My point is that even if you successfully show that Bitcoin is like Batman, you haven’t shown that much, since we have no idea what Batman (or any other fiction) is like.
I also want to point out, in fairness to Warmke, that he is not out there on an island with this sort of analysis. So-called “fictionalism” has become quite popular in philosophy. So popular, that virtually everything has been labelled fiction by someone or other. Numbers? Fictions. Ethical claims? Fictions. Truth? That’s a fiction too. I should probably say that I am not a fan of this strategy in general, not least because of considerations raised in the preceding paragraph. No one knows what fictions are. But I have other reasons to doubt that Bitcoins are fictional objects.
So, preliminaries out of the way, IS Bitcoin a fictional substance? Warmke has an argument that it is, and the argument runs something like this: I have a wonderful model for how bitcoin works (he’s right about that part) and in that model, fictional substance plays the role of Bitcoin, therefore Bitcoin is a fictional substance (he’s wrong to draw that conclusion).
I’m oversimplifying Warmke’s argumentative thread, obviously. One way in which I’m oversimplifying it is that I am handing him the idea that he has a model in which fictions play an important role. But it is more correct to say that he has a thought experiment in which he imagines (and asks us to imagine) constructing a model that has fictional substances as components. We are in weird territory here; we are in the land of models that don’t get built. They are merely imagined models. Are imagined models real models? One can expect objections here, but I’m going to accept the idea for now. Let’s suppose that imagined models are real models, and thus that we have a real model of Bitcoin, using fictional substances (whatever they are) as components of our model for Bitcoin.
Granting all this, what is my objection? I believe that Warmke is confusing a property of the model for a property of the thing being modelled. Let me illustrate the objection with the example of models for atoms and how they combine into molecules. You could model atoms with wooden tinker toys, assuming you had the right number of holes in the connectors. So, for example the hydrogen atom would have one connector hole, the Oxygen atom two, etc. You could get at a lot of truths about chemistry with that model, but you would be mistaken in supposing that atoms are made of wood, or that they are therefore flammable etc. The problem of course, is that it is merely an artifact of this model that it is made of wood. We wouldn’t have to use wooden tinker toys. We could build our model using steel connectors or using Swiss cheese and celery sticks. There are all sorts of possibilities. So, when using models in this way, we have to distinguish the artifactual properties of the model from the more important mathematical properties. We could even abstract away from all these physical models and end up with a purely formal mathematical model that would do everything our physical models can do.
Is Wermke making this sort of mistake? I think so. We are asked to imagine a state of affairs, and with that imagined state of affairs come imagined properties (like being fictional). In this case we are asked to imagine a bunch of people engaged in a collaborative work of fiction called BREAD. And given the rules by which this collaboration must proceed, it really models how Bitcoin works, and quite elegantly so. The problem is that you can imagine models, isomorphic in structure to Wermke’s model, that don’t have anything to do with fiction.
So, for example, instead of being engaged in a collaborative work of fiction following certain rules, one might be engaged in a game, called BREAD, in which bread tokens are acquired following certain rules. The goal? The degen with the most BREAD wins the game. OK, here we can get into an argument about ludic versus narrative theories of games, but if you think games are different than works of fiction, then the conclusion you would extract is not that Bitcoin is a fictional substance but rather that it is a game or game component.
Alternatively, an emperor, call him Trajan, might proclaim a new alimentary scheme using the rules of BREAD. In which case Bitcoin would not be a fictional substance or a game, but a method for distributing actual bread, which would make it an alimentary scheme or a component thereof. There is nothing fictional in that once it is in prod. People holding BREAD get to trade it for actual bread. Not only is this version of BREAD possible, but I think it is perhaps the closest to what is actually going on.
Let’s circle back to the part I agree with in Warmke’s paper — the idea that Bitcoin isn’t just code. There are a lot of reasons to think this. One reason is, I think, the one that Warmke is getting at here. If you can instantiate the Bitcoin algorithm without actually executing computer code, well then how could Bitcoin be identical to that code — the execution of which is not necessary? And that is a good point. You could paper-game Bitcoin. It would be a very slow game, but, in theory, enough angels with enough paper and enough centuries to play the game could do it.
Even apart from paper-gaming Bitcoin, it is not like there is a single way to instantiate the Bitcoin algorithm using computers. Different kinds of hardware can instantiate it in different ways. Maybe we could instantiate the mining portion of it using organic computers. Who knows? But what this possibility tells us is that there can be great variation in the code that Bitcoin runs on — clearly so at lower levels of abstraction, such as machine language. This is so, because the basic Bitcoin protocol can be executed — indeed must be executed — in many different ways when we start talking about the level of hardware and code.
This then takes us to the issue of what Bitcoin really is. If it isn’t a fictional substance and it isn’t code, then what is there left to say about it?
To say more, we need to make a distinction. There are two kinds of questions one might be interested in when one asks, “what is Bitcoin.” There is the narrow reductive question we have been discussing, and a broader question. I can illustrate the difference via another question: What is money? The narrow answer to the question “what is money” might look at the physical properties of money. So, you might talk about the metallic properties of coins, or the physical properties of paper money. But usually, those aren’t what one wants to know about when one asks, “what is money?” One wants to know about the functional role that money plays in human affairs. So too, when people ask “what is Bitcoin?,” they perhaps are not interested in the technical details of Bitcoin’s algorithm in isolation. That is one reason that “it’s code” is not a great answer. It is also a reason why “it’s like Batman” is not a great answer. If we really want to get deep into what bitcoin is, we need to get into an extensive discussion of its role, and indeed its future role, in the conduct of human affairs.
And honestly, when we get into questions of what there is, most of the furnishings of our world have this character. They are social constructions, or at least there is a large social component to their constitution. This is true even for tables and chairs and dinner plates and knives and forks, chopsticks and sushi. These are all canonical examples of things that are “real,” but if you think about it, their existence is not grounded in straight physics or materials science but rather in the social role that they play for us in family dinners and whatnot. So too, if we want to know what Bitcoin is, we will want to understand its role in the social fabric of our lives, a role that is even now just starting to be understood.
So maybe the question is not what Bitcoin is, but rather, what it is going to be. And because whatever it is going to be depends upon its future role in human affairs, maybe the real question is what Bitcoin ought to be. Sometimes metaphysical grounding is yet to be determined, and sometimes it is our responsibility to determine it.
