A bit of word-dissolving in political discussion

I found Scott Alexander’s steelmanning of the NRx critique to be an interesting, even persuassive critique of modern progressivism, having not been exposed to this movement prior to today. However I am also equally confused at the jump from “modern liberal democracies are flawed” to “restore the devine-right-of-kings!” I’ve always hated the quip “democracy is the worst form of government, except for all the others” (we’ve yet tried), but I think it applies here.

-- Mark Friedenbach

Of course, with the prompting to state my own thoughts, I simply had to go and start typing them out. The following contains obvious traces of my own political leanings and philosophy (in short summary: if “Cthulhu only swims left”, then I AM CTHULHU… at least until someone explains to me what a Great Old One is doing out of R’lyeh and in West Coast-flavored American politics), but those traces should be taken as evidence of what I believe rather than statements about it.

Because what I was actually trying to talk about, is rationality in politics. Because in fact, while it is hard, while it is spiders, all the normal techniques work on it. There is only one real Cardinal Sin of Attempting to be Rational in Politics, and it is the following argument, stated in generic form that I might capture it from the ether and bury it: “You only believe what you believe for political reasons!” It does not matter if those “reasons” are signaling, privilege, hegemony, or having an invisible devil on your shoulder whispering into your bloody ear: to impugn someone else’s epistemology entirely at the meta-level without saying a thing against their object-level claims is anti-epistemology.

Now, on to the ranting! The following are more-or-less a semi-random collection of tips I vomited out for trying to deal with politics rationally. I hope they help. This is a Discussion post because Mark said that might be a good idea.

  1. Dissolve “democracy”, and not just in the philosophical sense, but in the sense that there have been many different kinds of actually existing democracies. There are always multiple object-level implementations of any meta-level idea, and most political ideas are sufficiently abstract to count as meta-level. Even if, for purposes of a thought experiment, you find yourself saying, “I WILL ONLY EVER CONSIDER SYSTEMS THAT COUNT AS DEMOCRACY ACCORDING TO MY INTUITIVE DEMOCRACY-P() PREDICATE!”, one can easily debate whether a mixed-member proportional Parliament performs better than a district-based bicameral Congress, or whether a pure Westminster system beats them both, or whether a Presidential system works better, or whatever. Particular institutional designs yield particular institutional behaviors, and successfully inducing complex generalizations across large categories of institutional designs requires large amounts of evidence—just as it does in any other form of hierarchical probabilistic reasoning.

  2. Dissolve words like “democracy”, “capitalism”, “socialism”, and “government” in the philosophical sense, and ask: what are the terminal goals democracy serves? How much do we support those goals, and how much do current democratic systems suffer approximation error by forcing our terminal goals to fit inside the hypothesis space our actual institutions instantiate? For however much we do support those goals, why do we shape these particular institutions to serve those goals, and not other institutions? For all values of X, mah nishtana ha-X hazeh mikol ha-X-im? is a fundamental question of correct reasoning. (Asking the question of why we instantiate particular institutions in particular places, when one believes in democratic states, is the core issue of democratic socialism, and I would indeed count myself a democratic socialist. But you get different answers and inferences if you ask about schools or churches, don’t you?)

  3. Learn first to explicitly identify yourself with a political “tribe”, and next to consider political ideas individually, as questions of fact and value subject to investigation via epistemology and moral epistemology, rather than treating politics as “tribal”. Tribalism is the mind-killer: keeping your own explicit tribal identification in mind helps you notice when you’re being tribalist, and helps you distinguish your own tribe’s customs from universal truths—both aids to your political rationality. And yes, while politics has always been at least a little tribal, the particular form the tribes take varies through time and space: the division of society into a “blue tribe” and a “red tribe” (as oft-described by Yvain on Slate Star Codex), for example, is peculiar to late-20th-century and early-21st-century USA. Those colors didn’t even come into usage until the 2000 Presidential election, and hadn’t firmly solidified as describing seemingly separate nationalities until 2004! Other countries, and other times, have significantly different arrangements of tribes, so if you don’t learn to distinguish between ideas and tribes, you’ll not only fail at political rationality, you’ll give yourself severe culture shock the first time you go abroad.

    1. General rule: you often think things are general rules of the world not because you have the large amount of evidence necessary to reason that they really are, but because you’ve seen so few alternatives that your subjective distribution over models contains only one or two models, both coarse-grained. Unquestioned assumptions always feel like universal truths from the inside!

  4. Learn to check political ideas by looking at the actually-existing implementations, including the ones you currently oppose—think of yourself as bloody Sauron if you have to! This works, since most political ideas are not particularly original. Commons trusts exist, for example, the “movement” supporting them just wants to scale them up to cover all society’s important common assets rather than just tracts of land donated by philanthropists. Universal health care exists in many countries. Monarchy and dictatorship exist in many countries. Religious rule exists in many countries. Free tertiary education exists in some countries, and has previously existed in more. Non-free but subsidized tertiary education exists in many countries. Running the state off oil revenue has been tried in many countries. Centrally-planned economies have been tried in many countries. And it’s damn well easier to compare “Canadian health-care” to “American health-care” to “Chinese health-care”, all sampled in 2014, using fact-based policy studies, than to argue about the Visions of Human Life represented by each (the welfare state, the Company Man, and the Lone Fox, let’s say) -- which of course assumes consequentialism. In fact, I should issue a much stronger warning here: argumentation is an utterly unreliable guide to truth compared to data, and all these meta-level political conclusions require vast amounts of object-level data to induce correct causal models of the world that allow for proper planning and policy.

    1. This means that while the Soviet Union is not evidence for the total failure of “socialism” as I use the word, that’s because I define socialism as a larger category of possible economies that strictly contains centralized state planning—centralized state planning really was, by and large, a total fucking failure. But there’s a rationality lesson here: in politics, all opponents of an idea will have their own definition for it, but the supporters will only have one. Learn to identify political terminology with the definitions advanced by supporters: these definitions might contain applause lights, but at least they pick out one single spot in policy-space or society-space (or, hopefully, a reasonably small subset of that space), while opponents don’t generally agree on which precise point in policy-space or society-space they’re actually attacking (because they’re all opposed for their own reasons and thus not coordinating with each-other).

    2. This also means that if someone wants to talk about monarchies that rule by religious right, or even about absolute monarchies in general, they do have to account for the behavior of the Arab monarchies today, for example. Or if they want to talk about religious rule in general (which very few do, to my knowledge, but hey, let’s go with it), they actually do have to account for the behavior of Da3esh/​ISIS. Of course, they might do so by endorsing such regimes, just as some members of Western Communist Parties endorsed the Soviet Union—and this can happen by lack of knowledge, by failure of rationality, or by difference of goals.

    3. And then of course, there are the complications of the real world: in the real world, neither perfect steelman-level central planning nor perfect steelman-level markets have ever been implemented, anywhere, with the result that once upon a time, the Soviet economy was allocatively efficient and prices in capitalist West Germany were just as bad at reflecting relative scarcities as those in centrally-planned East Germany! The real advantage of market systems has ended up being the autonomy of firms, not allocative optimality (and that’s being argued, right there, in the single most left-wing magazine I know of!). Which leads us to repeat the warning: correct conclusions are induced from real-world data, not argued from a priori principles that usually turn out to be wildly mis-emphasized if not entirely wrong.

  5. Learn to notice when otherwise uninformed people are adopting political ideas as attire to gain status by joining a fashionable cause. Keep in mind that what constitutes “fashionable” depends on the joiner’s own place in society, not on your opinions about them. For some people, things you and I find low-status (certain clothes or haircuts) are, in fact, high-status. See Yvain’s “Republicans are Douchebags” post for an example in a Western context: names that the American Red Tribe considers solid and respectable are viewed by the American Blue Tribe as “douchebag names”.

  6. A heuristic that tends to immunize against certain failures of political rationality: if an argument does not base itself at all in facts external to itself or to the listener, but instead concentrates entirely on reinterpreting evidence, then it is probably either an argument about definitions, or sheer nonsense. This is related to my comments on hierarchical reasoning above, and also to the general sense in which trying to refute an object-level claim by meta-level argumentation is not even wrong, but in fact anti-epistemology.

  7. A further heuristic, usable on actual electioneering campaigns the world over: whenever someone says “values”, he is lying, and you should reach for your gun. The word “values” is the single most overused, drained, meaningless word in politics. It is a normative pronoun: it directs the listener to fill in warm fuzzy things here without concentrating the speaker and the listener on the same point in policy-space at all. All over the world, politicians routinely seek power on phrases like “I have values”, or “My opponent has no values”, or “our values” or “our $TRIBE values”, or “$APPLAUSE_LIGHT values”. Just cross those phrases and their entire containing sentences out with a big black marker, and then see what the speaker is actually saying. Sometimes, if you’re lucky (ie: voting for a Democrat), they’re saying absolutely nothing. Often, however, the word “values” means, “Good thing I’m here to tell you that you want this brand new oppressive/​exploitative power elite, since you didn’t even know!”

  8. As mentioned above, be very, very sure about what ethical framework you’re working within before having a political discussion. A consequentialist and a virtue-ethicist will often take completely different policy positions on, say, healthcare, and have absolutely nothing to talk about with each-other. The consequentialist can point out the utilitarian gains of universal single-payer care, and the virtue-ethicist can point out the incentive structure of corporate-sponsored group plans for promoting hard work and loyalty to employers, but they are fundamentally talking past each-other.

    1. Often, the core matter of politics is how to trade off between ethical ideals that are otherwise left talking past each-other, because society has finite material resources, human morals are very complex, and real policies have unintended consequences. For example, if we enact Victorian-style “poor laws” that penalize poverty for virtue-ethical reasons, the proponents of those laws need to be held accountable for accepting the unintended consequences of those laws, including higher crime rates, a less educated workforce, etc. (This is a broad point in favor of consequentialism: a rational consequentialist always considers consequences, intended and unintended, or he fails at consequentialism. A deontologist or virtue-ethicist, on the other hand, has license from his own ethics algorithm to not care about unintended consequences at all, provided the rules get followed or the rules or rulers are virtuous.)

  9. Almost all policies can be enacted more effectively with state power, and almost no policies can “take over the world” by sheer superiority of the idea all by themselves. Demanding that a successful policy should “take over the world” by itself, as everyone naturally turns to the One True Path, is intellectually dishonest, and so is demanding that a policy should be maximally effective in miniature (when tried without the state, or in a small state, or in a weak state) before it is justified for the state to experiment with it. Remember: the overwhelming majority of journals and conferences in professional science still employ frequentist statistics rather than Bayesianism, and this is 20 years after the PC revolution and the World Wide Web, and 40 years after computers became widespread in universities. Human beings are utility-satisficing, adaptation-executing creatures with mostly-unknown utility functions: expecting them to adopt more effective policies quickly by mere effectiveness of the policy is downright unrealistic.

  10. The Appeal to Preconceptions is probably the single Darkest form of Dark Arts, and it’s used everywhere in politics. When someone says something to you that “stands to reason” or “sounds right”, which genuinely seems quite plausible, actually, but without actually providing evidence, you need to interrogate your own beliefs and find the Equivalent Sample Size of the informative prior generating that subjective plausibility before you let yourself get talked into anything. This applies triply in philosophy.