I’ve been meaning to make a few comments with respect to Jim Antle’s recent TCS piece on the right’s internal schisms. Except that instead of a thoughtful, well-structured consideration, I fear that at this hour I’m only up for a core-dump of free associations on the topic. I think I first wrote about it here, but the 10-second version of it if you don’t feel like following the link is: the Vietnam War occasioned the last conservative/libertarian split, and that as the rationale for the Cold War marriage of convenience fades, issues like technology, domestic civil liberties concerns, and, especially, American foreign policy would inevitably bring about another one. It’s only a matter of time before the mainstream media stop running obtuse “odd-bedfellows” stories and realize that the largely contingent polar categories in their metaphor no longer map political reality.

If the “right,” or at least its moderate incarnation, is defined by the policies of George W. Bush, then a couple of things are clear. First, the term no longer means remotely what it once did: smaller government, fiscal responsibility, free trade. Second, the rough categorization of libertarians as part of the political right no longer makes very much sense. It was for a long time, though always with caveats, an accurate enough functional description under an electoral system that forces two main coalitions. But those coalitions are never set in stone, even when they last long enough to be given a name. Again, as an empirical matter, it seems libertarians are now more often found arguing in concert with (elements of what previously passed for) the left. Whether that would continue to be the case under a Democratic administration is unclear, but the hypothetical also seems unlikely to be tested terribly soon.

I think Jim makes one misstep in his thinking about the current fragmentation. That is, he sees libertarians as a kind of free-floating Sandra Day O’Connor of politics, allying with stable groups called “left” and “right” as whim and circumstance dictate. From that perspective, the idea of a libertarian “alliance with the left,” by which he seems to mean voting for Howard Dean or some other candidate of the left, strikes him as unappealing.

But left and right, as I’ve probably said five times too many in this post alone, are not permanent categories. They’ve changed in meaning before and will change again. The question is not whether libertarians will support the “left” or “right” if that means clones of Bush and Dean stretching on into the indefinite future. The question is what “left” or “right” will mean in ten years.

Imagine political clusters as striated rocks, with bands of different colors. The right might break down into a few main stripes, say “conservatives” and “libertarians.” Within each, there might be gradations in shading to indicate differences in practical positions—just how much government? But each stripe would also itself be composed of bands running perpendicular to the grain of the larger bands, to represent, perhaps, the different comprehensive ethical theories that make up the set of persons at any given point along the policy-opinion dimension. There would be patterns and clumping here—more deontologists among the anarchists, more consequentialists in the minarchist camp, say—but still a good deal of variety. And, of course, there are any number of further ways to segment any given band—including by one’s stance on some potentially novel question. The same fractal subdivision could be carried out for either of the major coalitions.

All of these are potential faultlines, and so each coalition can be broken up into a large number of different shapes. Constraints on the new coalitions that form from the fragments produced by cleavage along a given faultline are provided by the electoral system and the median voter theorem. I’m not going to replicate a bunch of poli-sci theory here, but there’s plenty of interesting work on the dynamics of coalitional formation in this sort of system.

Imagine the standard Nolan Chart, but pretend at first that only economic liberties are at issue, that only that left-right spectrum counts. You can use the standard Downsian analysis to predict (under first-past-the-post, winner-take=all) a pair of dominant coalitions (parties) situated just far enough from the median to deter third party entrants. Call those points at which the parties form the consensus points. Now, drop that second axis, personal liberties, into play. Depending on the distribution of voter preferences in this new 2-D opinion-space, the consensus points shift, not only up and down, but also along the original left-right axis. You can drop a third line representing some other issue (foreign policy, abortion, etc.) and repeat the shift in 3-D space, and in principle on into N-dimensional space for N issues. (Each of “economic” and “personal” freedom are conflations of positions on a range of sub-dimensions, which could be represented as separate dimensions as they grew salient.)

Well, goodness, I got off on a ramble there… the upshot is just that realignment to a new axis (the line defined by the new consensus points) is very likely to cause a shift, not only into new space, but along the old axis. When the realignment represents the introduction (or new salience) of multiple dimensions simultaneously, the jump is potentially highly discontinuous, yielding shifts along multiple previous dimensions. There’s little reason to expect the new consensus points after such a shift to bear terribly close resemblance to the previous ones.

I decided to call these “consensus points” for a reason. The standard polisci model takes voter preferences as given and party alignment as a straightforward function of those preferences. But we know it doesn’t actually work quite that way: there’s feedback between the preferences of individual voters (or subgroups) and the dominant coalition at a given time. Consensus points exert a gravitational pull on public preferences, thereby reinforcing themselves as equilibria. Coalition realignment is not just the rearranging of rigid puzzle pieces to form a new image; it’s a rearrangement that changes the shapes of the pieces.

If the new left and new right prove stable in the medium term, we can anticipate that, for instance, some progressives and libertarians would begin to construct a new, semi-shared “left” identity. (I cede “right,” whatever that now means, to the party in power, as spoils.) But that also can’t help but modify each group’s reflexive notion of what it means to be a progressive or a libertarian.

Anyway, late-night sketchy thoughts… I may try to turn this into something less babbling; suggestions welcome.