The Meta Level Doesn’t Justify Itself

The Meta Level Doesn’t Justify Itself

Scott Alexander distinguishes between thinking on the object level versus the meta level.

You are an Object-Level Thinker: You decide difficult cases by trying to find the solution that makes the side you like win and the side you dislike lose in that particular situation.

You are a Meta-Level Thinker: You decide difficult cases by trying to find general principles that can be applied evenhandedly regardless of which side you like or dislike.

The general trajectory of Western thought since at least the Enlightenment might be encapsulated as something like “Meta level good”. The institutions of modernism – liberalism, capitalism, and rationalism – are all very “meta” in the sense of dealing with particulars on the basis of more or less rigid pre-announced rules. Liberalism as a political institution regards everyone as equal before the law. Capitalism as an economic institution divorces the employee from the occupation, so you can conduct transactions anonymously. You don’t have to know your grocer to buy groceries – you just have to know he’s a grocer. And rationalism as an epistemic institution defines explicitly the kinds of valid connections that can be made between propositions.

With a few hiccups here and there, these institutions did a pretty good job of dealing with the conflicts and difficult cases that arose in the Western world for a few centuries. But what if the “sides” in some difficult case involve the validity of the meta level in the first place? After a few centuries experience as colonists of non-European societies, Europe found that liberalism, capitalism, and rationalism are rather impotent in settling difficult cases when one side doesn’t accept the “meta level good” dictum at all.

The transition from modernism to postmodernism can be thought of as a way to grapple with this question; an attempt to apply the meta logic to the meta level itself.1 In the limit, two different cultures are completely incommensurable. If the institutions of modernism are intended to settle disputes, there’s not much they can do when they themselves are in dispute. Someone who doesn’t accept the norms of good-faith rational discourse can’t be persuaded rationally to do so. A liberal order can’t engender coexistence if one culture is implacably dedicated to object-level advancement.

All this is to say, the institutions of modernism don’t justify themselves. The meta level is itself an object-level imperative. It’s culturally bound; not all cultures accept it. And so one of the most trenchant critiques of modernism is to deny the validity of the meta level at all – ironically, because this position is more meta than to actually affirm the meta level.


This was once a critique associated with Catholics in defense of the old order. Today it’s made mostly by the descendants of Marx’s sociology, both on the right and the left. Liberalism, capitalism, and rationalism – far from being neutral with respect to interests – just advance the interests of the liberals, the capitalists, and the rationalists! There’s no such thing as a neutral ideology, since any ideology will have its own partisans. You can always out-meta any ideology or institution that thinks of itself as operating on the meta level by pointing out that it’s no “meta” level at all; just another object level ideology/mythology trying to displace the old ones.

This position is called relativism by its detractors. And every Evangelical is trained from birth with the gotcha: “Ok, bud, isn’t ‘there are no absolute truths’ itself an absolute?” In other words, isn’t “meta is good” itself an object level principle? – exactly the same as the postmodern critique of modernism!

Ultimately this is a bullet that has to be bitten. Meta-level institutions are worth defending for their own sake. The more we can reduce the specific value commitments necessary for coexistence, the better – and it was liberalism’s great virtue that it embodied a sort of “minimal set”. But these value commitments can never be completely eliminated. Against certain assaults – specifically from those who don’t accept the validity of meta-level dispute resolution – they can only be defended on the object level – that is, on the basis of liberal interests. In the last resort, liberalism isn’t just good because it’s more meta; it’s good because judging some set of things on form rather than content advances certain interests and not others. But one can’t judge everything based on form. At some point, the specific content must be evaluated. At some point you have to plant a flag and say: these interests are defensible because they’re ours, and these are not.

Ha ha, but seriously

Unfortunately there seems to be a piety contest surrounding the meta level. It’s become something of a sport, for example, to fault Christopher Hitchens for failing to be meta enough with respect to Islam – specifically for taking the position that those whose goal is peace and coexistence must at some point resort to violence against those whose goal is violence and domination. Whatever else his faults, this position cannot be disposed of by invalidating its form. The content must at least be engaged with on its own terms. Aesthetically ugly as it is, it may yet be right.2

In the end, “Relativism is itself an absolute principle” is not a gotcha. “You’re not very tolerant of the intolerant” is not a gotcha. “You’re gonna force people to live peacefully?” is not a gotcha. The answer to each is simply, “Yes, and?” – hopefully followed by a discussion of the actual merits of the position. There is no meta-level principle that doesn’t need practical grounding in the object level in this way. Any principle that tries to get along without it will find itself stumped by a similar gotcha question, and ultimately swept away by institutions less scrupulous about advancing their own object-level goals.


  1. Culler, for example, described an exercise in textual deconstruction thus: “To deconstruct causality one must operate with the notion of cause and apply it to causation itself.”
  2. Though it is a pity that the people who insist upon this point most forcefully are, like Hitchens, otherwise terrible people who hit upon it in a broken-clock manner. When you make an opinion taboo, those left holding it are probably objectionable norm-defectors in general – which, of course, further justifies the taboo.




Facebook Twitter Reddit Threads


  • 1

    Jim Caton

    Apr 23, 2016 at 14:47 | Reply

    The meta-level refers to the hierarchical ordering of objects. The only way out of this problem seems to be competition amongst meta-frameworks. The competition will itself involve the “us vs. them”, object-level norm.

    The discussion seems to be an iteration of arguments concerning the choice of priors. Choose your priors, let the logical structure form around them… Of course another means of construction would be to critically observe the world and continually adjust and expand one’s logical construct. Here we let the logical structure emerge from observation, then identify priors ex post.

    • 1.1

      Cameron Harwick

      Apr 23, 2016 at 15:01

      We’ve identified a few meta-frameworks within which competition selects for good results – i.e. liberalism, capitalism, rationalism. But competition among frameworks I’m less sanguine about – especially given the demographic problems that seem to follow from liberalism. There’s no guarantee that “good” institutions will be selected for. If competition among frameworks is going to select for beneficial ones, we might need a return to Western object-level norms like colonialism, or at least a less bashful “Westernism”. Otherwise I don’t see selective pressure in any desirable direction.

  • 2

    Brad Kells

    Apr 24, 2016 at 11:33 | Reply

    Okay, so I’ll try to not make this into a reducto gotcha argument, but object-level ranking is an absurdity.

    Take the impetus for Scott’s quiz:

    A ha! says Scott, look at these plebs with their lack of principle, they are only focused on having their Team win!

    Well, no. Its just a competing Meta. Scott is on Team Federalism/States Are The Unit Of Analysis. He wants us to notice that his opponents are willing to change their view of federalism/states if it suits their outcome. But the other Team also has a Meta: they are on Team Gay RIghts / Or Nah Are The Unit Of Analysis. They point out that Scott is completely undecided on the morality of gay marriage.

    Both sides are consistent with respect to one thing, and inconsistent with respect to another. They are asking different questions.

    You can’t say Process and Result, Means and Ends, because the process is a result, and the ends can be means.

    The two are necessarily both variables. Capitalism is a process, but you can’t say it is objectively good until God writes it in the clouds. You can say it produces good things (the process produces results) but now you’re back to the start: what is a “good thing?”

    You can say that the Meta level is morally correct. I.E. capitalism is a Good Process.
    The Post-Modernists have dismantled this pretty thoroughly, I’d argue.

    But saying that “the Meta level is good because it produces good Object level results” is a deception, a trick of the light. Capitalism is a Good Process because it makes people rich. Obfuscation. It just hides the fact that you still have ranked outcomes, in this case *rich better than poor*. You haven’t actually justified anything.

    An Object level defense of the Meta is just an obfuscated Meta.

    • 2.1

      Cameron Harwick

      Apr 24, 2016 at 20:47

      Interesting reversal – assimilating everything into the meta level rather than the object level. But I think we’re basically agreed that the postmodernist critique is pretty fatal to any grounding in the nature of things.

      You’re right that this isn’t a justification of Capitalism, etc – but it is a defense. Actually that’s a good way of summing up what I was trying to argue, that a defense doesn’t have to be a justification. You can’t justify cooperation in a finitely repeated prisoner’s dilemma, but you see people cooperate anyway. Why? Is cooperation an end in itself? Do they think there’s a justification when there really isn’t? If so, am I making things worse by pointing out the postmodernist argument? Or, given the fact that there are people pointing out the argument in order to egg people on into defection, is it better to shore up cooperation as an end in itself by embracing an identity as partisans of cooperation? We believe in liberalism, capitalism, and rationalism – not because God wrote it in the clouds, but because that’s who we are. An identity as people who (among other things) reject identity politics.

Leave a Reply

More Content