Billmon, my favorite ex-blogger, makes a series of interesting observations about the state of parties.
Billmon finds it puzzling that parties have given up the role of being intermediaries, but, somewhat ironically, that is entirely consistent with the way political science theories about parties have been evolving last few decades. Somewhere along the line, the idea that parties are supposed to play a brokering role among contending factions within the party has been essentially dropped and forgotten–especially in the US context. Theories, wrapped in neat formalism, that purported to show that parties are simply tools for the party medians to rig the institutions and force the outcomes of their choosing on the rest of their parties became the norm. (“purported” because this logic runs into a serious logical problem: the same logic that would allow the parties medians to govern, i.e. that the “majority” rules should imply that the majority party’s median should not be able to prevail ultimately over the popular median…but the logic of allegedly powerful parties persisted mainly by huffing and puffing.) Of course, if indeed the role of the parties is for the alleged majorites within each to force their will on the rest, this semi-authoritarian role of the party institutions is precisely what we should expect.
In a sense, the evolution of the political science to focus on the “control” rather than “mediation” is natural and logical, even if ultimately self-subverting. Control is easy to theorize about and measure. What exactly “mediation” means is not easy to theorize and nearly impossible to measure in a consistent sense. But if so, this is typical of the anti-scientific “data science” type mindset: we will just hang our brains out to dry and just let data tell us whatever where we have data, because, well, we have no brain.
There is even more disturbing aspect to this: how do we know who/what the medians are? The only way one could tell is that the medians are “supposed” to win therefore whoever wins must have the median on their side. But if the role of the incumbent party leaders is to rig the party institutions that they win, even if the numbers may be against them, can the results of the “elections” that led to their victories be trusted as reflecting the popular sentiments? The practitioners of techniques like DW-Nominate are somewhat aware of this, but they are seemingly willfully refusing to see it as a fundamental problem–merely a methodological problem that will somehow sort itself out with some fancy formulas, not a fundamental problem–although, to be fair, this would threaten the validity of a lot of their measurements (and worse, how they are being used). In the actual practice of politics, this becomes even more perverse, to the extent of Orwellian proportions: we know we have the support of the (mythical) median because we won even though we won because we count the votes through the rules we have concocted to our advantage.
The interesting implication of Arrow’s theorem that has been persistently ignored is that majorities are ephemeral like shifting sand dunes: they are inherently unstable. Hayek, with his obsession with the epistemological problem in politics and economics, would have counseled that the “majority” that rules at any one time should be advised to insure themselves against losing their status. Of course, this is precisely what Mayhew argued as the the basis of his “universalist” theory of legislative/coalitional politics, even if not in so many words. Instead, the actual practitioners of politics of taken up the path he advised against, as inducing too much instability: of building a narrow coalition that forces its will over unwilling or even hostile masses. Some amount of institutional rigging would help prolong this, but only at cost of increasing instability. Beyond some threshold, the whole lashup can only come crashing down.
PS. I’ve been reminded that this “political science” perspective applies rather narrowly to American politics worldview. This is definitely true.