The rationalization of intrinsic policy preferences

 You prefer A to B.  Whatever A is, whatever B is, whatever distinguishes between the two within the common space shared by the propositions, you prefer A to B.  Or perhaps you shall be contrarian, and assert that you actually prefer B to A, even though they are arbitrary labels, so I can reverse them, change nothing, and the statement would have held.  Why?  Not, why were you such a contrarian as to force me to switch the labels, but... why did you prefer A to B?  Were I to assign you a paper defending the choice of A rather than B, I am certain that you could construct a five-paragraph, three-point snoozer following the formula provided to you by high school teachers and perhaps some especially shitty professors.  If you have had some better writing experience, you could write something more cogent, and more enjoyable for the reader.  But would it be the real answer to the question of why you prefer A to B?  The cause?  The thing in your head?  The truth?  The real truth?  Would you even know the real answer?

In that most silly and pointless of disciplines of philosophy, we examine the question of how you know what you know.  Epistemology.  Yet I am asking a different question.  Why you prefer what you prefer, and put slightly differently, believe in the superiority of what you believe.  That is not the same, because built into this framework is more than a truth question, but a normative question.  In writing such a paper, some hack of a teacher once gave you the Greek distinction between logos and pathos, and blah, blah, but that's not really how we study the question.  I am interested in the actual, serious question of why you believe what you believe.  That's a psychology question.  Also, a neurology question.  Even a genetics question.  Yes, really.  All filtered down to whatever the fuck political science is.

Go ahead and make the joke.

So let's just get to that base level.  Genetics?  Seriously?  Yes, seriously.  Ever since Campbell, Hibbing & Funk published their, shall we say, controversial paper, "Are Political Orientations Genetically Transmitted?" in the American Political Science Review in 2005, a subset of political scientists have been examining not only twin studies but even genetic analysis to find strong associations between your political predispositions across a range of measurable outcomes, and that helix-y thing.  No, that is not a claim of genetic determinism, and those who both recoil from and think they have zinger-counterarguments to Campbell et al. tend to do so by mischaracterizing the argument as genetic determinism, which it is not.  Predispositions.  Robust predispositions.  Why?  The short version is that genetic factors influence personality traits, which influence the types of political beliefs one is likely to form.

Of course, that opens up a whole can of brain-eating worms.  It introduces the question of the process by which personality traits lead to the formation of policy preferences, and it introduces the question of what interrupts that process, or at least, what other than genetically-influenced personality traits will influence the formation of policy preferences.

Remember, Campbell et al. were not making claims of genetic determinism.  Stop that shit.

This gets us into the big, wide field of empirical political science, as applied to the study of public opinion.  We tend to study this at the large-N level because small-N analysis tends to be what we call, in technical terms, "bullshit."  We measure a battery of policy preferences, consider a range of potential independent variables, and then assess which independent variables predict which policy preferences, at least at the mass level.  We can ask the extent to which variables like urban/rural, income, race, religion/religiosity and all that kind of stuff predict positions on Issue 1, Issue 2, and so forth.  Income may have more predictive power on Issue 1, race may have more predictive power on Issue 2, and that's the way cookies crumble.  The statistical predictors are not always what you think, but that's why we do the analysis.  The whole point of science is to see how reality works rather than just doing the Aristotle/bullshit thing.

Sorry, but I've been re-reading Aristotle, and fuck that guy.

Anyway, so this gets me to the society-level distinction between who prefers Policy A and who prefers Policy B.  Subgroup 1 prefers Policy A, and subgroup 2 prefers Policy B, on average.  By tendency.  So, if I am making a causal argument, the most important independent variable is whether you are in subgroup 1 or subgroup 2.  Of course, nothing ever comes down to one independent variable, but this is basically how the process works.

Yet I still haven't answered the question of why.  Even if you happen to be in subgroup 1, and prefer Policy A, that doesn't get me why.  It's not just that I don't know what's going on in your head.  There's some meat missing on the bone.

And since all rational people hate vegans, let's tear into that meat.

At this point, we return to the distinction between positional issues and valence issues, based on the distinction made by Donald Stokes in his 1963 APSR article, "Spatial Models of Party Competition."  Stokes was writing in response to the emergence of the Downsian model, as we sometimes call it, although Downs was borrowing heavily from Hotelling (1929).  Put voters along a line, representing the left-right spectrum, and parties strategically place themselves along that line in order to compete for votes in an election.  Simple, intuitive, and by this point, deeply ingrained in how we think about all things political.

But... what the hell is it that makes you prefer -2.71, or 3.14, or whatever?  What is it about that location?  That policy?  Because that number has to mean something, besides the obvious.

Downs didn't give a fuck what that number meant.  Stokes did.  Stokes says there are two kinds of preferences.  There is a positional preference, and a valence preference.  A positional issue is an issue over which you prefer a policy 'cuz you want that policy, because you think that policy is intrinsically good and right.  Others will disagree on which policy is intrinsically good and right.

Then there are valence issues.  Valence issues are issues on which we collectively agree on what we want to achieve-- the goal, the outcome-- but disagree either on the method best suited to get there, or which candidate/party has the secret sauce that gets us there.

In that classical Stokes framework, and how we generally talk about elections, the standard valence issue is something like "a good economy," with the catch being that this is a multidimensional thing.  Unemployment and inflation are inversely related, the stock market is kind of its own thing, and there's a lot wrapped up in "the economy," much of which people do not understand, but there's some ideal, Goldilocks economy that people want, even if they don't necessarily know what that is.  Nor Thing Fucking 1 about economics, and hence, what economic analysis says about how to produce said unspecified Goldilocks economy.  But gimme gimme gimme!

The complications of "a good economy" make it a conspicuously odd "valence issue," even if it is a canonical one, yet you can list many others.  There are plenty of outcomes on which Americans collectively agree.  There's lots of stuff we'd all like.

Start writin'.  Have fun.  Be as topical as you'd like.

...

...

How do you get that stuff?

Well, there's the rub.

So we have a contrast.  There are preferences you have, instrumentally, because you think that they will produce outcomes that people generally want (valence outcomes), and you have positional preferences, which are things you want, just because you want 'em, 'cuz that's the stuff you want, intrinsically.

And here's where Stokes makes this even harder.  A valence issue, like "a good economy," may lead to positional disagreements over, say, fiscal policy.  Macroeconomic analysis can lead to disputed conclusions about which policies produce which economic outcomes, so a person might favor left or right policies, thinking that such policies will produce valence outcomes.  The policy dimension and the valence dimension (as conceived by Stokes*) are not even truly orthogonal.

But it's worse than that.  What if you form a positional preference, for reasons other than an empirically-motivated assessment of the outcomes that the policy would produce?  You then have strong cognitive-psychological incentives to conclude that your policy will produce valence outcomes.  You want Policy A just 'cuz you want it, and that's the real truth, but once you form that preference, your brain offers you candy to conclude that Policy A will make everything sunshine & rainbows from now until eternity, and don't bother me about the farmers' need for rain, it'll fix that too!

And then you tell me that the reason you prefer Policy A to Policy B is because of the outcomes it will achieve.

Which would be a lie.  Your belief in those outcomes would be motivated reasoning, to produce a valence-based conclusion in support of a positional preference, which existed prior to your assessment of the valence-based effects.  You're just rationalizing.

Stepping back from Stokes, into a more abstract realm, why do you want what you want?  Because there is a hole in you.  An emptiness.  It's called, "your navel."  Gaze into it, and feel yourself filling up with the completeness of knowing the truth of your navel.

I hate philosophy.

What I mean to say is that you can have preferences that are instrumental, or preferences that are just... intrinsic.  You can want a thing because it gets you something else, or you can want a thing, just fuckin' 'cuz.  Why do I want money?  That's an instrumental preference.  Money is a medium of exchange.  That's it.  It does me no good, except in that I can buy shit I want, or invest it so that I can buy even more shit later.  That is a purely instrumental preference.  Why do I want coffee?  A little from column A, a little from column B.  There is some utilitarianism to caffeine-- not gonna lie-- but I also just love coffee.  It tastes good.

Policy.  Here's where things get really tricky.  You have strong cognitive incentives to formulate a policy preference on something other than scientific and data-based analysis of the valence-associated outcomes that policies achieve.  You have all this other background stuff, beginning honestly at the genetic level.  You seriously have genetic predispositions.  Which is not to say it is necessarily immoral to have intrinsic policy preferences divorced from the valence outcomes that policies would achieve, but you have pressures working on you that start at the genetic level, work neurologically, culturally from the moment you can process information as a mere tot, you've got your family, friends, extended network and everyone around you, all operating in a way that makes you likely to fail the Asch line experiment because most people fail the Asch test, and the basic point is that you are not a neutral, objective, scientific data analyst, stepping back and asking, what valence outcomes do I wish to achieve?  What data must I gather?  Upon gathering these data, which policy best achieves the valence outcomes that are most important?

That ain't how your brain works.

And yet, if I ask you why you prefer A to B, you're gonna tell me, it's because A creates some valence outcome, and B doesn't.

Is that really why you prefer A to B?

No.

That's a rationalization of an intrinsic policy preference.  Which is not to say that it is immoral to have an intrinsic policy preference, but cognitively, one should be aware that this is happening.

And upon making one's self aware that this is happening, what does one do?  One interrogates that policy preference.  And one interrogates the logic and data behind the conclusion that it produces a valence outcome.  One becomes one's own interrogator.

Remember how science works.  Formulate a hypothesis.  Specify your hypothesis in contrast with a "null hypothesis," such that the two are mutually exclusive and exhaustive.  The null hypothesis should be, generally speaking, that you are wrong, the world is random, nothing to see here, folks, move along.  Default to the null hypothesis.  Do not let yourself believe your own claim until you have tried to debunk it, tried to adhere to the null hypothesis, and found that adhering to the null hypothesis requires such absurdity in the context of observable data that one must "reject the null hypothesis" in the context of observable data.  And, since you have defined the null hypothesis and your own hypothesis (the "alternative hypothesis") to be mutually exclusive and exhaustive, rejecting the null becomes equivalent to accepting your own hypothesis, although for epistemological/terminological reasons, we don't put it that way.

The point is, you don't let yourself believe yourself until you have tried to convince yourself otherwise, and failed.  It's like The Hitchhiker's Guide to the Galaxy, and flying.  It's the art of throwing yourself at the ground, and missing.  Scientifically, you don't get to believe your own claim until you try to convince yourself that you are full of shit, and you fail.  You find some small part of yourself, however small, that is devoid of shit.

But you gotta try.  You gotta look.

If you didn't try, you cheated.

And hell, who are you?  Nobody reads this damned blog, so it ain't like I'm gonna dock your grade.  There is no "you," not because I have turned solipsist, but because nobody is reading this post, or at least, so few people as to be negligible.  You're not scoring any points on me by cheating.  You aren't getting an undeserved A by shirking.  You're just fooling yourself.

I'd rather be right, but that requires doing the hard work of becoming one's own interrogator, in order to converge towards right answers over time.  Not "being right" at any particular point in time, but having a process to get there, which is how science works.  It doesn't work if you cheat.

So here's the thing.  Or at least, a thing, there being many things out there.  You may have some intrinsic preferences over policy, on any particular issue of the day, whatever day "today" is, which may very well be different from the day on which I am posting this rambling, blathering bit of pseudo-scholarly, self-indulgent nonsense.  But if you then rationalize that your intrinsically preferred policy must create a valence outcome, and furthermore, assert that this rationalized effect is why you prefer that policy...

You're just lying to yourself.

Carolyn Wonderland, "Victory of Flying."  The studio version is on Peace Meal.  This woman is BAD-ASS.

*Later scholars developed the concept of valence characteristics, which are candidate traits that voters just supposedly like.  According to modern valence models, voters intrinsically want candidates who are competent and honest, just 'cuz.  You know.  Obviously, such models have tons o' predictive power.