Sunday, 12 June 2016

Estlund Authority & Aumann agreement

This is David Estlund talking to 3A.M magazine-
'I call a government “legitimate” insofar as it is morally permitted to coercively enforce its orders. That’s very rough, but enough for the distinction I want here. I say a government has “authority” insofar as it has the power (a kind of moral power) to put people under moral obligations (defeasible of course) to do things, by commanding them to do so. Again, there are details, but that’s already a clear distinction between two moral questions. It’s possible, of course, to hold that one is never present without the other, and also possible to deny it. By analogy, it might be that all and only the tall flowers in my yard are yellow. But their being tall and yellow are different things about them. A government’s power to generate obligations is different from its being permitted to enforce its commands. I don’t suppose anyone disagrees, but issues are often missed by simply defining “authority” or “legitimate authority” as combining both, and never asking what grounds one part and what grounds the other.'

Suppose there is a Government composed of a single person. Suppose that person feels she is not bound to obey any order given by that Government. Finally, suppose this were common knowledge.

Could such a person give orders, in the name of the Government, which would place 'people under moral obligations to do things'? No, because she is a person and, by our second assumption, we know that she doesn't feel any such obligation. Thus such a Government has no 'authority'. It also has no 'legitimacy' because the person who composes the Government can't coerce herself to enforce her own order to feel a moral obligation to do a certain thing.

Yet, as a matter of fact, there have been plenty of Polities- including ones which are recognisably Democratic- which have enjoyed 'legitimacy' and 'authority' despite maintaining, as a matter of Constitutional doctrine, that there always exists at least one person who is not under a moral obligation to act in accordance with the orders of the Government and, moreover, can't be coerced to do so by any morally permissible means.

It may be argued that 'legitimacy' and 'authority' are not separable but overlapping concepts within a particular type of deliberative discourse a la Habermas. In other words, there is some more or less public process of argumentation and deliberation which everyone would agree was authentic and fair minded, such that at the end of the process each person has a substantively rational acceptation of what is going on in Society such that no scandal arises by reason of any procedural asymmetry concomitant with the existence of a Governing and Governed class.

The problem here is that if such deliberation consumes scarce resources, or if pure strategies can be dominated, or if concurrency type problems arise, or if discourse is not incentive compatible, or if preference revelation is problematic, or if policy space is 'multi-dimensional', or if hysteresis effects predominate, or if Knightian Uncertainty obtains, or if there are increasing returns to scale and scope, or if external effects don't cancel each other out; then there is no a priori way to determine if any deliberative process is 'authentic' though there is an a priori way to show that it must be sub-optimal.

This is not to say we must all sheepishly endorse the Schmittian 'State of Exception'- yet, we can't deny that no State would retain its legitimacy very long if its 'authority' were based on a Universalizable as opposed to an essentially asymmetric, not to say heteronomous, theory of Morality. 
Perhaps there is a non-Kantian theory of Morality such that Elstund's claim to have said something meaningful w.r.t the separability of 'legitimacy' and 'authority' would be true.
If so, this theory would enable the Ruler, though refusing to grant the moral authority of her own commands in at least one case, can still justify her behaviour by some moral, as opposed to expediency based argument to the Governed, even if they do in fact acknowledge that authority. For Utilitarians, there are two possible ways this justification could be made. One would be to argue that she was acting with substantive rationality to achieve a State of the World militated for by Act Utilitarianism. The second would be to justify her actions as procedurally rational in accordance with Rule Utilitarianism.
Even if Utilitarianism is not the 'true Moral Theory', if it is admitted that some people can be swayed by its arguments, then it must be the case that either talk of 'deliberative democracy' or 'epistemic proceduralism' is wholly empty or that the relevant Public Justification we have stipulated for has sufficient alethic force to minimally fulfil an Utilitarian desiderata.

The late Dr. Trejo-Mathys quotes Elstund as attacking Habermas for a sort of nihilism such that Public Justification goes ahead without sufficient alethic force.
Deliberative democratic theory claims to employ only purely procedural standards for the public employment of reason. “The notion of a higher law,” Habermas urges, “belongs to the premodern world.” There are no standards that loom over the political process, policing its decisions, not even any standard of reason itself. “We need not confront reason as an alien authority residing somewhere beyond political communication.” The only normative standards that apply to political decisions are noninstrumental evaluations of the procedures that produced them—in particular, standards of “procedural rationality” based on the power of reason in public political discourse. Any imposition (in theory or practice) of substantive political standards would preempt the ultimately dialogical basis upon which Habermas thinks political normativity must rest. There is an echo of Arendt here: politics is the site of discursive contestation, so politics cannot begin with conclusions. Here is where the price of the nihilist view is evidently too high to pay, or so I will argue. No appeal to good outcomes is permitted on this view, there supposedly being no such thing. … The task for deliberative democratic theory, then, has become the odd one of explaining the central importance of substantive public discussion of the procedure-independent merits of possible political decisions, without ever granting that there actually are any procedure-independent standards.

Trejo-Mathys defends Habermas as best he can but gives the game away when he says - 'In Habermas’ view, all normativity rests on an “ultimately dialogical basis”, and science, morality and art criticism are just as much “sites of discursive contestation” as politics is.'
The problem here is that people who allow themselves to be Governed, rather than seeking to run away from any such stationary bandit, only do so because this is an Evolutionarily Stable Strategy. For them 'normativity' is linked to those preferences or tropisms which ensure survival and reproduction upon an uncertain fitness landscape. Nobody, not a gobshite by profession, actually gives two shits for 'discursive contestation' and nothing has an ultimately 'dialogical basis' because only professional gobshites indulge in dialogic which is why we relegate them to the role of Child Minders to the more cretinous portion of our Credentialized offspring. Meanwhile, cheap 'out of control' mimetic effects shape Revealed Preference though, no doubt, Preference Falsification based Availability Cascades regulate what people say they want as opposed to what they vote with their feet for. There is a good reason why Public Discourse is essentially hypocritical and misleading. The fact is, Newcomb type problems lie at the heart of Language. We all swear to quaff Kafka's toxin at Dijkstra's banquet because- since substantive, that is ergodic, rationality and procedural, that is hysteresis afflicted, rationality never sync up on an uncertain fitness landscape, hence breeding 'capacitance diversity' which is why these types of rationality evolved in the first place- it serves only to starve philosophers- whereas we have quietly stolen as much of the silver as we decently could and are already plotting our escape to where the grass is greener if only for having fed on others' bones.

Thus, it is perfectly adaptive, at least if one is by nature a worthless gobshite rather than someone whose Loyalty might be valuable, to indulge in the sort of 'cheap talk' given below (my comments are in bold) while reserving 'costly signals' for the purpose of Exit, rather than Voice.

3:AM: .. you’re defending the idea that we can’t do without an epistemic or truth-seeking dimension to the democractic ideal and as we’ve seen, you think this is the way to go given that the commoner defences don’t work. How do you not fall into the despotic character that Hannah Arendt observed when she said ‘ from the viewpoint of politics, truth has a despotic character.’
DE: I’m glad you asked that, and I have a long answer. The reason I’m glad is that Arendt’s view about “Truth and Politics” (as the title of her central essay on that topic puts it) is understandably, but wrongly I think, thought to ground a deep objection to thinking of democratic authority in even partly epistemic terms (as I do). If normative truth has no place in political discourse, then the question whether democratic arrangements might tend to get things right can’t ever arise…it might seem. So let me try to explain why I think that is not her view at all. (The key: she and I agree there are procedure-independent standards of political judgment, which is all I need, but for certain reasons she doesn’t want to call them “truths.”)
Arendt does not mount a critique of normative truth or standards that would show that they have no place in political thought or discourse. And we should note that there are two ways in which they can hardly be avoided. One is the point at which a person participates in politics. A person might simply amorally pursue her aims, I suppose, but this is not how almost anyone understands their participation. But only because such misunderstanding is strategic and also has a lower computational cost. Rather, a person will normally (so we know from piles of empirical research) have a view of how her interests, competing interests of others, and other considerations, ought to be somehow brought to bear in action. Yes, in the short run, it is cheap to have an 'attitudinal' or 'trait based' view and even, at the margin, to send 'costly signals' on that basis; however there is a mimetic tipping point after which 'situational rationality' prevails and 'Identity' proves increasingly spectral. The participant won’t necessarily think that her own opinions about these things ought to be imposed even as others disagree, but she does have opinions, and she does disagree with others. She can hardly add that there’s nothing to her view except that she happens to hold it, as if any view is as good as the next. She will take herself (however uncertain she might be) to hold it for reasons that warrant her holding it. She has a view about what people ought to support doing collectively, and she thinks (maybe tentatively, maybe not) others to be mistaken, at least partly. So she appeals to what she will see as certain controversial normative—I think broadly moral—judgments whose validity (let’s call it) is not up to her or anyone. I'm afraid this isn't so at all. Where there is an 'uncorrelated asymmetry' something like the 'bourgeois strategy' is eusocial in the short run- i.e. before the fitness landscape has been explored. That's what fools philosophers into thinking there are universalizable moral laws. However, this 'bourgeois strategy' disappears longer term( because resources can be invested in acquiring or losing a particular signal) so ergodicity is restored. The second point at which even a follower of Arendt must not think that everything normative is up to us is this: The very idea that what gets done ought to be determined politically (rather than despotically, say) is a moral or normative principle or standard that her view cannot avoid, and surely its own validity is, again, not subject to what gets decided in actual politics. If a Society knows it is Democratic or Despotic then an uncorrleated asymmetry arises. Assuming Societies compete under scarcity, it follows that resources will be diverted to changing signals till the informational asymmetry disappears. The comedy of 'People's Democracies' competing with genuine ones ended, at least in Germany, once 'Exit' became possible. Parts of the former G.D.R are demographically worse off, and face a grimmer future, now than 20 years ago because Exit, especially among better educated young women, has continued even under Frau Merkel. Wolves, however, are flourishing, as are Neo Nazis.
I think Arendt could easily agree with all this. When Arendt says that “truth has a despotic character,” she explains that she means that as a single person investigates questions in, say, physics or math (even aspects of history, though that’s more complicated), the correct answers are not in any way affected by the fact that there are other minds with other perspectives. It’s simply between me and the non-negotiable world. By contrast, in the domain of the moral and political, there is no subject matter at all if not for multiple minds and perspectives. The right answer is not only hard to know without help from others—that’s also true about lots of math, physics, and history (she emphasizes the need for witnesses, for example). More importantly, in normative matters the right answer is itself some accommodation of multiple perspectives, such as competing aims and convictions. But her point here is easily exaggerated. She is not arguing that there are no moral or genuinely political judgments that have any validity or correctness apart from multiple people’s actual opinions or statements or arguments. She rejects that conventionalist view, and holds that people’s actual statements or views might be mistaken even about moral or political matters. Indeed, a common way for a person’s moral or political opinion to go wrong, on her view, is through a failure of (solitary or socially assisted) imagination, an insufficient “representation” of the positions of others. Yet, what is happening in East Germany has nothing to do with 'the domain of the moral or political' or the problems of preference aggregation or mechanism design or the pseudo-problem of alterity. The same thing is happening in other European countries with wholly different moral and political availability cascades.
For purposes of an epistemic dimension of democratic authority, this is enough, and it’s highly consonant with how I see things. Arendt doesn’t doubt for a moment that even democratic political decisions can go wrong, be invalid, lack “impartial generality,” in her term. That impartially general perspective is not up to us, as a matter of decision on her view (or Kant’s, on which she deeply draws, or Habermas’s which draws on them), though it does essentially involve the existence of multiple points of view, which is her point. I think it’s misleading to say, as she does, that in the political (a part of the realm of “opinion” rather than “truth”) “our thinking is truly discursive,” since that might suggest that it depends on what actually gets said. That is not the view as she explains it. On the other hand, she clearly thinks that trying to understand or reach that impartially general stance requires, as an epistemological matter, hearing from others—not because their opinions are somehow automatically valid or determinative (they aren’t) but because our imaginations are limited and need this aid. But it is crucial to this whole view that the impartially general (and so valid) stance is not constituted by what actually gets said in discourse, or even by whatever people happen to think, since even together, even democratically, we could always get it wrong. The impartially general stance sets a standard that is independent of any actual human (even democratic) procedures, even if it isn’t independent of the existence of multiple minds and agents. In my account of democratic authority, that’s the aspect of the vernacular idea of “truth” that is operative: procedure-independent standard for judgment or belief. (This leaves entirely open whether the valid impartial stance might be explicated by some imaginary idealized procedure, as Habermas argues. The point here remains: its independence of any actual procedures. The implications of the hypothetical procedures, on such a view, are still not up to us. Real politics could always get it wrong. Habermas is easy to misunderstand on this.)
In that case, where Aumann agreement obtains, so must Elstund type procedure-independent 'truth'. However, in real time, if membership in the Government is stochastic in at least one respect, an uncorrelated asymmetry can arise during an Aumann agreement process. This means that a legitimate Government may lack Elstund 'authority'. Weakening Aumann's conditions won't get us to robustness because we know a priori that this entails so exponential an increase in concurrency and race hazard type problems- which militates for Hotelling type convergence or even Coasian internalization- compounded by congestibility and signal degradation problems with the result that Braess's paradox gains salience and only closing at least some autobahns and boulevards of Public discourse can rescue Communication itself from logjam. 
One workaround is to make policy space 'multi-dimensional'- but that just gets us to McKelvey chaos.

No comments: