Digressions&Impressions

On Information and the State: from Justin Smith to George Stigler

[Over time I plan to phase out D&I at typepad. This post was first published at  digressions.impressions.substack here. To receive new posts and support my work, consider becoming a free or paid subscriber at <digressionsimpressions.substack.com>]

A recent wide-ranging, erudite, and politically salient Substack post by Justin Smith defends the thesis that "we may identify a vast process currently underway whereby we are moving, in our efforts to understand reality and to harness it to our ends, from its to bits. It seems to me the real story of the past few decades, a story that easily explains the rise in popularity of the simulation argument...is of a fairly total regime change, from physics to informatics —here I am adapting the French informatique, which names the research domain in question far more appropriately than “computer science”— as the Prima Scientia or ultimate science of reality." (The post may be gated so you will have to trust me on its interest or take a subscription to Justin's excellent substack, Hinternet.) Presumably, by 'regime' Smith is suggesting that the wide world of learning is fundamentally ordered and (he makes an explicit nod to Hacking here) ordering and authoritative in some sense. 

In these digressions, I have been joking over the years that economics will soon be displaced by computer science. But as is usual while I have been focusing on a change in one tree, Justin is noticing the shifts in the wider forest.

Unfortunately, Smith never defines 'information' (and its non-trivial cognate,) 'units of information processing' despite it playing a key role in his overall argument. That's probably a feature and not a bug of his argument because part of his point is that in virtue of the significance and fluidity (within constraints) of symbolic representation information is not (ahh) ontically stable, eternally out there. 

Inform has Latinate roots (to give form to, to shape, educate, etc.), and, perhaps, there are echoes of the scholastic doctrine of hylomorphism lurking here. (Not so long ago, I was pleased to learn that computer scientists use 'hylomorphism' to refer to a certain recursive function which struck me as very funny and profound at once.) As I was barely conscious contemplating the possible significance of this, through the mystery of the associative process and the rapidity of modern networks (and Google), I looked up and downloaded George Stigler's (1961) "The Economics of Information." This paper has over 12000 citations, and alongside his work on economic regulation, crucial to his importance to the development of economics (for which he got a Nobel).

Somewhat amusingly, Stigler also fails to give a definition of information. He tends to use 'knowledge,' 'advertising,' subjective beliefs, and the comparison of prices by consumers as ways of understanding and describing information. Information can be 'pooled' and has a 'flow.' Most famously, information is costly (and can be bought and sold).

That information is costly is both part of the cause and part of the effect of the dispersion of prices. (In an older tradition of economics -- as David M. Levy taught me, one can think of Wicksteed -- prices were thought to be uniform in a market at a given time; although earlier, during the 18th century, Adam Smith and his contemporary, James Steuart, thought that prices dispersed, too although their analysis is not identical.) I don't mean to suggest that this exhausts the role of information in Stigler's analysis of economic life because it connects means and ends, but about that some other time more.

In Stigler's paper, I encountered a fascinating passage that is worth quoting and discussing (and this is the real point of this digression):

The maintenance of appreciable dispersion of prices arises chiefly out of the fact that knowledge becomes obsolete. The conditions of supply and demand, and therefore the distribution of asking prices, change over time. There is no method by which buyers or sellers can ascertain the new average price in the market appropriate to the new conditions except by search. Sellers cannot maintain perfect correlation of successive prices, even if they wish to do so, because of the costs of search. Buyers accordingly cannot make the amount of investment in search that perfect correlation of prices would justify. The greater the instability of supply and/or demand conditions, therefore, the greater the dispersion of prices will be.
In addition, there is a component of ignorance due to the changing identity of buyers and sellers. There is a flow of new buyers and sellers in every market, and they are at least initially uninformed on prices and by their presence make the information of experienced buyers and sellers somewhat obsolete.
The amount of dispersion will also vary with one other characteristic which is of special interest: the size (in terms of both dollars and number of traders) of the market. As the market grows in these dimensions, there will appear a set of firms which specialize in collecting and selling information. They may take the form of trade journals or specialized brokers. Since the cost of collection of information is (approximately) independent of its use (although the cost of dissemination is not), there is a strong tendency toward monopoly in the provision of information: in general, there will be a "standard" source for trade information. (Stigler 1961, 220)

Chicago economists are not exactly known for assuming that any market will tend toward a monopoly. As I noted last week, in the 1950s Stigler's work was, in fact, prominent for undermining empirically the Marxist (and Marx-inspired) thesis that markets tend naturally tend to monopoly. 

Now, later when (in part inspired by the work of the sociologist Morton and in part because of his interest in which economics Phds would be hired where) Stigler turned to the sociology of scientific knowledge, he tackled this tendency toward monopoly in knowledge production head on. And so I will return to it some other time. But in the 1961 paper he leaves the point in the quoted passage alone.

It would be tempting here to discuss the role of google or wikipedia or the stanford encyclopedia of philosophy as standard sources of information in our ecologies. (And to what degree ChatGPT can displace these or become one itself.) But I want to close with an important broader context in which i situate Stigler's passage.

In the liberal tradition — with which Stigler was very familiar — which understands, as a a core function of the state, it (the state) as a machinery of record, a collector and disseminator of accurate public data. A lot of our social practices, inside and outside the market, presuppose a social infrastructure in which the machinery of record is reliable, allowing public authorities and private actors to plan their activities. For that to happen, the public must be well-informed, and the only way citizens can possibly be well-informed on complex matters of policy is for state experts to organize and process information. Elsewhere, also in joint (forthcoming) work with Nick Cowen I have argued that Walter Lippmann's (1922) Public Opinion puts this vision forward.[1]

In fact, the germ of Lippmann’s idea is expressed in the closing paragraph of John Stuart Mill’s and Harriet Taylor’s On Liberty (1869, Ch. 5), where they claim that “the greatest dissemination of power consistent with efficiency” should be allied with “the greatest possible centralisation of information, and diffusion of it from the centre.” In context, it is clear (recall) that they are responding to Tocqueville’s fears that in order to prevent democratic despotism through local self-government, incompetent art of government is inevitable.

So, lurking in Stigler (1961) is a causal account that both explains the origin of the state and justifies the state's continued existence as a (perhaps even natural) authoritative monopoly provider of the machinery of record. (In fact, this is how Smith explains the epistemic origin of the state: as guarantor of measures and weights and the soundness of currency. ) And that even in a bit-rich world this will remain so.

Of course, Marxists and Libertarians may object that the account hinted at here effaces the violence of the state. This is undoubtedly so, but I consider this a feature not a bug of the argument.

• This first appeared at: <On Information and the State: from Justin Smith to George Stigler (substack.com)> a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

On the Strange Afterlife of Kant (the Antinomies) in Analytical Philosophy/metaphysics; some reflections on our methods

There are two mutually reinforcing founding myths in analytical philosophy: the one centered on Cambridge, England, tells the story of how G.E. Moore and Russell liberated us from the unintelligibility and inflated metaphysical pretensions of British Idealism by the application of common sense and conceptual decomposition aided by Frege's new mathematical logic that the young were using to dismantle the cowwebs of intellectual dinosaurs (like Bradley). The other is centered on Vienna, where in the aftermath of Einstein's overthrowing of Newtonian physics, it was understood that Kantian philosophy (with outdated commitments on space, time, causation, etc.) was untenable and (with help from Wittgenstein) forged a powerful program that mixed logical acumen, again indebted to Frege, and respect for the sciences, even Enlightenment values. From the vantage point of these two narratives, the ongoing attractiveness of transcendental philosophy and species of idealism within Continental philosophy is a kind of intellectual embarrassment--hadn't they heard the news -- one might ask paraphrasing Nietzsche -- of the death of  Kantianism?*

Yes, such myth(s) is no history, but during the last few months, in reflecting on the inexorable march of (knee-jerk realist) metaphysics within analytical philosophy, and the hardening of attitudes toward anyone willing to doubt the solidity of Moorean facts (despite the existence of widespread semantic and pschological illusions discovered by way of the sciences), I have started to wonder about the following question:

What happened to the antinomies in the rise of analytical philosophy?

For, the antinomies, diagnosed by Kant, do not go away just because Kant got physics wrong. In fact, while the new logic is a fantastic tool, it also gave rise to new logical antinomies (starting with Russell's paradox and rediscovery of very old ones, e.g., The Liar).

Of course, the logical antinomies stimulated extraordinary fertile, ongoing attempts to develop logic(s) in ways that circumvent even clarify these antinomies. So, rather than ignoring the antinomies they can present research opportunities either by acting as known constraints or as obstacles to be overcome. Reflecting on this history made me see analytical philosophy in a different light.**

For, one can understand the methods of analytical as responsive to the threat the antinomies pose. Here's a rough and ready taxonomy:

The first, and dominant, approach is to espouse a systematic caution that expresses itself in various ways within the analytic tradition: by this I do not just mean the obsession with rigor and clarity, nor a certain ponderousness about the mundane (e.g., ordinary language philosophy) but also a certain self-limitation on what can be expressed.+ In Carnap, this is exhibited in a fondness for precise, engineered languages, or frameworks, and a suspicion of questions that are external (or optative) to these frameworks. In Quine, it entails deference to a suitably regimented language that collects the findings of science. In both there is, despite the best efforts of Stebbing, also a mistrust of metaphysics as such.

Of course, in Lewis the caution is methodological in that any philosophical system has to find some optimal cost-benefit equilibrium between common sense intuitions and theoretical desiderata and is to be expressed in some final-physics language (for a more careful presentation of the conservativism in Lewis's program see here). To this day, some analytical metaphysicians discuss their projects in terms of the ontological/ideological features of a hypothetical, ideal language at the end of the progress of science.

These examples of systematic caution all respect the challenge the antinomies pose without having to address them and even without having to be aware of them. (If it turns out that my historical speculations cannot be substantiated in the record, I can always say  that I am putting my finger on -- to borrow a phrase of Dennett's --free-floating rationales without representation.) Such systematic caution connects with another familiar fact of analytical philosophy: its focus on puzzle-solving rather than systematizing. (Analytical philosophy has always rewarded specialization rather than breadth [recall].) The advantage of this kind specialization and precision is not just enhanced focus and skill at puzzle-solving, but one also avoids bumping up against antinomies that would reveal themselves if one were forced to systematize. (As an aside, this may help explain some of the status-hierarchies within philosophy: a flatter hierarchy among areas of specialization runs the risk of accidentally  generating the kind of systematicity that would generate awareness of antinomies.)

Of course, while the Kantianism in Carnap, even Kuhn, is now well-established, a second approach is to be found in other analytical philosophers, who explicitly reverted back to structural features of Kant’s philosophy most notably Strawson at Oxford. Strawson's ‘descriptive metaphysics’ is a metaphysics of what (that other important Kantian), Sellars at Pittsburgh, would call the ‘manifest image’ of ordinary life not fundamental reality. The widespread interest in normativity reminds us of the enduring vitality of this approach.

The third method is to treat philosophical systems as (revisable) inferences to the best explanation in which one appeals to theory-world fit in light of a variety of scientific virtues. In principle, this approach need not respect the threat of the antinomies (despite its origin in Pierce's philosophy), if it remains focused on relatively isolated topics. Those that advance it also advocate the disciplining (Williamson's term) of philosophy by constraining it in light of the demands of, say, some formal system. This piggy-backs on the ways in which the logical antinomies have been deflected and tamed. The downside of this approach is that one's philosophical commitments get blackboxed in one's formal apparatus.

So, here's a prediction: as analytical philosophers are emboldened to get ever more systematic and to bring the problems of cosmology, consciousness, and axiology, including politics, into their collective orbits, the antinomies will reappear in modern guise, especially if they are using inference to the best explanation in each individual area or topic.++

 

On Being Longwinded

Don't let your question (or your answer) run on forever.--David Chalmers "Guidelines for respectful, constructive, and inclusive philosophical discussion."

Most of Chalmers's "guidelines" are sensible suggestions for the smooth runing of public philosophical discussion. It's sobering to look at them and recognize that I violate a number of them regularly without even thinking about it. Even so I have some qualms about some of the guidelines. (In what follows I leave aside the risks associated with the heavy handed emphasis on civility that is growing in popularity among university administrators and some of my peers. For discussion, see Johnson and Kazarian and Leiter.)* In particular, throughout the guidelines, time is treated as a scarce resource and the norms appear to be designed to guarantee a fair allocation of it (see especially the section on "procedural norms"). But more important than temporal equity is quality. And it is not obvious that temporal equity serves quality. More important, it is by no means obvious that brief (or brief-ish) questions really serve inclusivesess. Let me explain.

If one shares a language and there is considerable mutual trust, then one can be brief, even use short-hand, and still expect to be understood. Brevity or to-the-pointness is, then, a skill that can be developed in the context of a lot of background agreement, philosophical as well as social. I call it a 'skill' because it requires considerable expertise and practice. It is easy to forget once one has mastered it, but it is very difficult to ask a focused philosophical question in real time. So my first criticism is that this norm may well be exclusionary in practice, especially toward novices and those that are not quick on their feet. (Obviously, a lot turns on what it means to 'run on forever.') This criticism can be accomodated by a wise and informed moderator (who can recognize the inexperienced and slow-thinking in our midst).

My second, more important, criticism builds on a feature of the first, but is really distinct. Sometimes one's position on an issue, or more general philosophical sensibility, is very different from a speaker's. But, even then, it may still be mutually illuminating, especially if there a students presents, to have a dialogue between outlooks that are at odd with each other. In such a circumstance, a question may well require considerable 'stage-setting,' in order to make a precise and even focused question possible. Sometimes the 'stage-setting' alone can be already very illuminating to others in the audience. Not all philosophy needs to be about challenging each other's foundational premises or the exploration of alternative frameworks, but it strikes me as a shame if we codify a set of norms that unintentionally, but predictably, makes such an event, which would be a paradigm of respectful, constructive, and inclusive discusssion, even rarer than it already is.