Against Political Chmess

Why too much political philosophy studies a game that nobody plays

Photo by Hassan Pasha on Unsplash

I’m really happy to publish this guest essay by Paul Sagar. Paul is a Reader in Political Theory at King’s College London and writes the ‘Diary of a Punter’ Substack, which I *highly* recommend. In this essay, he argues that much of modern political philosophy is the equivalent of studying 'chmess'—a game like chess, but with different rules that nobody actually plays. He argues that by theorising about this imagined world, many philosophers fail to address the messy, real-world political realities we actually confront. I couldn’t agree more! - Dan.

---

In a typically punchy piece, Daniel Dennett cautioned against the philosophical equivalent of pursuing higher-order truths about chmess. Chmess – as opposed to chess – is a (hypothetical) game in which the king may move two spaces instead of one. Like chess, it is therefore amenable to an (infinite) number of a priori truths. An appropriately programmed computer (or, indeed, a very clever human with too much time on her hands) could come up with them. But this in turn means that it is possible for there to be higher-order truths about chmess:

1. Jones’ (1989) proof that p is a truth of chmess is flawed: he overlooks the following possibility ...

2. Smith’s (2002) claim that Jones’ (1989) proof is flawed presupposes the truth of Brown’s lemma (1975), which has recently been challenged by Garfinkle (2002) ...

The challenge Dennett issues is that pursuing such second-order truths risks being to a significant degree worthless (or so I read him). It is not that doing so is incoherent, or mistaken, and certainly not that it is (too) easy; on the contrary, it might be extremely difficult. It is rather an instantiation of the dictum “if something is not worth doing, it is not worth doing well”. And why is it not worth doing? Because the brute fact is that nobody plays chmess.

Dennett’s primary goal in his paper was to warn philosophers against falling into the trap of pursuing careers in the philosophical equivalents of chmess. It seems to me very good advice, that ought to be widely heeded. But one area in which it has not been widely heeded, at least over the last 50 or so years, is Anglophone political philosophy. A remarkable amount of what now passes for political philosophy is roughly the equivalent of chmess. It is political chmess. But I am against political chmess – and you should be too.

*

To try and make my case, I will first go after the most obvious target. Who else but John Rawls, the most prominent and influential (well, in universities at least) English language political philosopher since WWII?

When stipulating the conditions said to obtain in “ideal theory”, offered as a guide for how to think in a normatively principled and structured way about politics, Rawls states that:

1. Agents are taken to be reasonable, in the technical sense that they endorse a principle of reciprocity such that “they are ready to propose principles and standards as fair terms of cooperation and to abide by them willingly, given the assurance that others will likewise do so”. (This informs Rawls’s crucial understanding of a “reasonable comprehensive doctrine,” a key building block in the construction of the “overlapping consensus” necessary for generating his preferred understanding of “political liberalism”.)

2. Agents exhibit strict compliance, by which it is meant that “(nearly) everyone strictly complies with, and so abides by, the principles of justice”. In other words, all that is in principle required for people to follow the demands of justice is that they come to acknowledge and recognise that these are indeed the demands of justice.

3. Society is conceived of as being closed, with no neighbours either hostile or friendly, and which “we enter only by birth and exit only by death”. This is because “it is natural to conjecture that once we have a sound theory for this case, the remaining problems of justice will prove more tractable in the light of it”.

There is of course a great deal more to Rawls’s immense theoretical apparatus. Nonetheless, focusing even on just these features is instructive.

Recall that the difference between chess and chmess is that in the former the king can move only one space at a time, but in the latter two. Does it not look an awful lot like Rawls is doing something similar, but with regards to politics rather than chess? For consider some things that we know about politics:

• A. There is good reason to believe that significant numbers of people fall short of Rawls’s standards of reasonableness. As the political theorist George Klosko has pointed out, between one fifth and one third of American citizens hold religious views that would be classed as fundamentalist by Rawls, and are hence not “reasonable” by his standards. That is between 60,000,000-100,000,000 people. Klosko is being understated when he writes that “A conception of ‘reasonable’ principles that immediately excludes this number of people requires strong justification.” Whatever the philosophical virtues of Rawls’s conception of reasonableness may be, it is not a conception that corresponds to the world we inhabit.

• B. Not only is it manifestly the case that politics is exemplified by agents not exhibiting “strict compliance” regarding justice, but more fundamentally, the very reason that we have politics is precisely because human beings don’t exhibit strict compliance regarding things like justice. If we assume strict compliance in our ideal theory, we assume away the “circumstances of politics” that we are supposed to be wrestling with in the first place. Such ideal theory by necessity fails to be political theory.

• C. Rather than modern societies being “closed”, they are systematically conditioned by having to share the world with others. This means that issues such as security, trade, immigration, and emigration, will systematically affect the politics of any group of people in myriad, complex, and far-reaching ways, and the institutions through which they try and organise their politics will themselves inevitably have been conditioned by a legacy of inherited, brute-fact, international realities.

Returning to the chess-chmess comparison, we might say that points A-C have the following kinds of equivalents in chess:

A.* - the king can only move one space

B.* - knights move two spaces forwards and one space sideways, and are the only pieces that can “jump” over other pieces

C.* - each player starts with 16 pieces, and the board is made up of 64 squares

By contrast, Rawlsian ideal theory looks more like the following sorts of innovations that we might add to chmess:

1* - the king can move two spaces instead of one

2* - knights can move any number of spaces, in any direction, and if they choose, they can hover above the board (but not above the king)

3* - each player starts with five pieces, and the board is made up of 20 squares

Now to be clear, there may be nothing (necessarily) incoherent about my expanded version of chmess. And it would certainly not be easy to come up with philosophically rigorous truths (a priori or otherwise) about expanded chmess (the hovering knights alone guarantee that). If someone so wished, they could spend a lifetime working out what follows about expanded chmess, enthusiastically building up a body of published work in which they reply to, and contend with, other leading figures. Whether or not that is the best way to spend one’s life is absolutely beside the point; each to their own, depending on whatever makes them happy (under the usual liberal constraints about fists stopping short of other people’s noses, etc). Still, some of us might point out that nobody actually plays chmess. But they do play chess. And we are interested in chess. So what, exactly, is all this scholarship on chmess teaching us about what we want to know about?

It should be obvious where I am going with this. To say that Rawlsian ideal theory is political chmess is not to say that it is incoherent. (Or rather, it is not to say that it is necessarily incoherent: it might be, but that is subject to further investigation.) And it is not to deny that great intellectual power is at work there. Nobody who has read Rawls can seriously deny that his is a towering intellectual achievement taken on its own terms. But why should we take those terms, if what we want to know about is political philosophy? What he gives us conspicuously leaves out crucial fixed points of politics, and as a result, offers an analysis of something else. Indeed, the situation is worse here for Rawls than for the proponent of chmess. After all, chess is just a game, one of many that we play. At least some of us could swap it for chmess, if we really wanted to (although more on this shortly). By contrast, it is not an option for us to “play” ideal theory in real life, so what is achieved by our discussing how we should play it if we could, and yet knowing full well that we can’t?

It is rather like turning up at the house of a chess tutor, wanting to learn more about chess, and being told that one must spend years studying the intricacies of chmess. Being a good student, one does what one is told. Chmess might in fact be very interesting, and indeed some extremely intelligent people may have carved out distinguished careers expounding on its details (and if you work hard, you might too!). But when handing over the money for another lesson, you start to feel that something isn’t quite right here. That you came to learn about chess, not chmess…

*

Of course, the tutor has a reply ready. Don’t worry, he tells you, learning about chmess is an indispensable prerequisite for correctly playing chess. This is because without the “ideal theory” that chmess provides, one will have no principled basis upon which to critique and improve the practices of chess. Chmess provides this; that is its ultimate job.

But doubts having been kindled, you find this hard to take seriously. As far as you can tell, no actual chess players in the world have more than a passing awareness of chmess, and none of them ever make serious reference to it when planning or playing their games. This applies just as much to the amateur club down the road as it does to Magnus Carlsen. OK, but perhaps they should; perhaps until they do, all their chess is proceeding on an unprincipled basis, lacking rational grounding. But why? Or rather: why and how could chmess show this, whilst also delivering chess from such flaws, through the lessons it purports to offer? After all, by changing the rules in the way it does, chmess is talking about something which is not chess.

The analogy may seem overstrained, so evidently ridiculous that it cannot be fair to use it to critique Rawls. But I do not think that it is unfair. Rawls insists that without his ideal theory (or something like it) “there is no rational basis for continually adjusting the social process so as to preserve background justice, nor for eliminating existing injustice”. But this is not plausible, and not only for the reason that it is a bare assertion, one we should no more accept than if somebody told us that until we have an ideal theory of the best ship then we have no rational basis for repairing our leaky boat. More than this, it is as if the ideal theory we were being told to consult turns out, upon closer inspection, not to be about water-borne vessels at all, but flying carpets. Or if you like, about chmess rather than chess.

*

Let us return to the classroom. Somewhat exasperated, the tutor calls in an eminent colleague to help set you straight. The eminent colleague concedes that it is indeed a (sad!) fact that in the real world no chess player, from amateur through to grandmaster, seriously consults the wisdom available through studying chmess. But so much the worse for the chess players. You see, the eminent colleague has consulted high-profile academic works showing (or so he contends) that if the players of the world switched the rules of chess for the rules of chmess, the resulting game that everybody then started playing would be better. Getting into his stride, the eminent colleague anticipates that you might be troubled by this notion of “better”. Better how? Never fear! He has a battery of arguments that any sceptical student will have to go along with. Given what chess students already believe to be true about the value of chess, they have to conclude that a world in which everybody played chmess would be better than a world in which everybody plays chess.

But you have been growing restless, and interject. You grant for now that it is true (even though you are not sure that it is) that a world in which we replaced chess with chmess would be better (however defined). You even grant that all fully coherent and rational people who understand the eminent colleague’s arguments will agree that it would be better if everybody started playing chmess instead of chess. Here’s the thing, though: they aren’t going to, are they? People have invested years in learning the rules of chess; some of them have built prestigious careers in it, and their income and status depends on it; chess is embedded in our culture, has deep roots in our history, is in many ways bound up with the psychology of what it means to be a gameplaying animal at this point in the development of the species. Plus, most people will never even hear about chmess, and of those who do, a large number will lack the intellectual chops to handle its intricacies (it rather looks as though one needs a PhD to even start to understand it properly). And that’s before considering all the people who have vested interests – some perfectly respectable, others nefarious, a few outright insidious – in preventing a transition to chmess, and who can therefore be expected to keep chess going. You can see that although it might indeed be better if everyone gave all this up and started playing chmess…they just aren’t going to, are they?

But now the eminent colleague raises a wry smile. You have walked into his trap. He replies: but they could, couldn’t they, if they really wanted to, and tried really hard? Sure, it might be difficult to start playing chmess instead of chess. Sure, it is very unlikely that enough people will start doing it such that the state of the chess world changes. But it is not impossible, is it? After all, people saying that they (or others) won’t adopt chmess cannot serve as a reason for rejecting the correctness of chmess (which, we all agree, would be better than chess). And this is so even if the people saying that they won’t itself serves as a declaration of intent to put insurmountable barriers in the way of chmess-realisation. The superiority of chmess stands independent of won’t, in all its guises! In turn the validity of chmess, and its attendant superiority to chess, stands vindicated. The eminent colleague leans back in his armchair.

And yet you find yourself unmoved. The reply seems obvious. Right now, around the world, millions of people are playing chess. They have been playing it for thousands of years (long before theorists of chmess came on the scene). They will be playing it tomorrow. And the day after. And the day after that. And you have an intellectual desire to understand it better, perhaps even (although you are less optimistic about this) one day suggesting some modest ways in which it might perhaps go better. You plan to use the philosophical tools of conceptual analysis, but are open to others when appropriate. And so, sure, it is not impossible that human beings could all play chmess, instead of chess. But they don’t, and we can be very sure (based on literally all the facts that we have) that they won’t, not now and not in the future. Hence, if the aim is to better understand chess, well then chmess just looks…beside the point.

The eminent colleague is beginning to look a little flustered. You can’t say what I’m doing is incoherent though, can you? It is plainly coherent! What you are is a chmessophobe! But I can be a chmessophile, if I wish!

It is getting uncomfortable. You have clearly overstayed your welcome. You start to back out of the room. Aiming to lower the tension, you reply in complete sincerity: of course you can research whatever you want! I am not for a moment accusing you of any kind of rational failing. If you want to study the philosophical depths of chmess, that is your prerogative! Hell, you are more successful and eminent than I will ever be! But as you turn to leave, you just can’t resist: the thing is, I’m interested in chess, and whilst you’re welcome to do your chmess, I personally just don’t see why I should care…A shoe hits the door as you close it behind you.

(I will not belabour the point any more than I already have: David Estlund is as much a practitioner of political chmess as Rawls. Although I must certainly apologise to him for my literary flight of fancy. The one time I met him he was generous, good natured, and supportive. I doubt that the real Estlund throws shoes.)

*

How might one proceed from here? I suggest that the best option is not to double down on political chmess – although some certainly have. For example G.A. Cohen, the former Chichele Chair of Social and Political Theory at Oxford, attacked Rawls by substituting a recognisable conception of justice with an appeal to fraternal benevolence, but which he labelled “justice” to retain the normative oomph of a term now stripped of its conditions of meaning. Cohen replaces political chmess with political chmess².

By contrast, Cohen’s successor, Jeremy Waldron, has rightly emphasised that political philosophy should be the philosophy of politics, which means that what we are theorising must in some sense be politics. It can certainly be a simplified form of politics, perhaps one that leaves things out if doing so helps us to produce better theory. (We can think of theories as being rather like maps: if they do their job properly they help us to navigate the world – but they do not, and must not be allowed to, take priority over the world.) But political philosophy cannot be about something completely different if it is to remain political philosophy. Ideal theory of the kind practiced by Rawls (and many, many others) does not involve simplifying assumptions; it involves counterfactual assumptions that mean it is theorising something altogether different. It is playing a different game – chmess, not chess.

Is there a better option? Yes, and it goes under the heading of what Rob Jubb has recently labelled sober political realism. It is a burgeoning approach, growing in strength (see for example here, here, here, here, here and here). It is properly identified and distinguished not via unhelpful debates about method, or “normativity”, but by what such work shares in common: taking seriously the fixed points of politics as a pre-condition for doing political political theory. It offers investigations of political chess, rather than political chmess. I recommend it accordingly.