You may be familiar with the idea that temporal disorganization comes from distortions, but not the usual distortions of, say, taxes creating deadweight loss. Rather, some economically literate people have strong opinions on the fact that the tone of Fed communications has measurable effects on asset prices, which they feel should depend on economic fundamentals. They regard the apparent temporal disorganization in the market for complex capital goods such as homes and so-called human capital - which is seemingly caused by central banks shifting priority from stimulating demand to lowering prices - as a dangerous disconnection from underlying economic conditions. That disorganization happened, some of them believe, specifically because of shifts in the interest rate. Perhaps if the interest rate was kept at a correct, steady rate then the intertemporal supply and demand for housing and labor would also be balanced.

But how do all these effects work in the first place? Why isn’t the interest rate just set by the balance of intertemporal supply and demand, independent of central bank jawboning? When mortgage rates reset faster than housing supply can respond, are interest rates operating as part of coordination failure or failing their unique role as a signal of “intertemporal scarcity” (whatever that is)? In short, “In what sense are interest rates causes and in what sense are they effects?”. At their worst, these concerns can drift toward attempts to treat certain financial assets as monetary benchmarks, even when their volatility or illiquidity prevents them from functioning as discounting or hurdle rates.

Though more important to discuss the economy than economics, all these questions turn on the economics question of “what sets interest rates?”. We are back to our original question: what concerns (balance-sheets, optionality, margins, rollover, funding risk, etc.) are operative when things are “disorganized?”

Since we are moving into the abstract, let’s set up the table with a core principle. This is our solid methodological plate on which we can place softer and gooier ideas without making a mess: AN INTEREST RATE IS A RELATIVE PRICE. Whenever you’re confused about capital or interest rates, try this: leave the computer, take a deep breath, relax your eyes, unclench your jaw and remind yourself “An interest rate is a relative price.”

Because interest rates are real market activities, we aren’t able to assume a priori that they perform a particular activity. Maybe they play a role in coordinating market activity in the face of complexity or fundamental uncertainty, maybe not. Maybe they are downstream of deep economic fundamentals like technical progress, capital accumulation and fundamental psychological necessity - or maybe upstream or both or neither. Maybe there is a special interest rate that sets the tone for all the others, or maybe not. Hey, maybe all prices do is conceal exploitation under the flag of profit-seeking and so mystify the relations of production. The principle of “quantifying the quantifiable” only helps us choose what to investigate, not answer what that thing is.

What we can assume is that whatever role interest rates play can be discovered in markets themselves. What matters for market participants is what the rate embedded in a 30-year mortgage, a Secured Overnight Financing Rate (SOFR) swap or a Treasury future is functioning as, not what it could function as.

You see, as relative prices, interest rates are part of the objective economic world, not deep, hidden parameters or deep, hidden residuals. Of course expectations, conventions and uncertainty matter to their determination but the interest rate written into a contract, the outcome, is a quantifiable fact. When going from interest rates to the interest rate, follow the principle of quantifying the quantifiable. Interest rates are quantifiable and observable, don’t give them up for the invisible and unobservable. Use them as guides for the invisible and unobservable.

The principle of “quantifying the quantifiable” is thus not a dismissal of the complex and unobservable. Rather, the goal is to hold onto those few firm bits we do have in a complex, dynamic and uncertain reality. This principle is worth insisting on, as it is deeply important to market analysis that interest rates are market variables, not themselves psychological parameters nor invisible residuals.

The interest rates built into those housing market contracts aren’t parameters or residuals - they are actual relative prices. The actual interest rates embedded in financial contracts and which drive investment decisions can be a ground for analysis in the face of complexity and fundamental uncertainty. Again, a 30-year fixed mortgage, a SOFR-indexed construction loan, and a Treasury future all embed interest rates as tradable relative prices, not inferred parameters. That ground is what is risked in giving them up.

Of course, there are unboundedly many relative prices - like the ratio of the prices of grapes and grape-nuts. They are not all interest rates. Interest rates are distinguished from random price ratios by the fact they can be meaningfully used to discount income flows. This is how interest rates are defined by their roles in economic practice.

Examining a financial contract, we discover that an interest rate is, in the simplest case, the ratio of the difference between the future price of a service flow and the spot price of that service to the spot price of the service. If you switch future and spot then you get a discount factor. Because the discount rate has a simple algebraic relationship with the interest rate, I’ll sometimes use the terms interchangeably.

I am insisting on this objectivity point at length because it can be very difficult for people to differentiate between the fact of an interest rate with their theory of the interest rate. It is all too easy to slide between “interest as a price” and “interest as a conceptual object” without noticing the shift. As Continuous Variation (CVAR) Editor Alex Williams noted, the interest rate is a central and load bearing structure in economics. Given the interest rate, economics has predictive power across wide domains - in micro domains (when will kids mow a lawn today for a pie tomorrow), in growth theory (what technologies will be invested in) and in macro domains (how many millions should be unemployed). Milton Friedman won a Nobel Prize in work that included finding a (fairly weak) effect of the interest rate in the consumption function. But conditional predictions given the rate of interest are much stronger than predictions of the rate of interest itself (because comparative statics require fixed rates). This is a delicate situation in a fundamental part of market analysis.

We therefore can see why there is an impulse to keep such a load-bearing structure safe from empirical revision. But that temptation must be resisted. In fact, we must ask ourselves, is there even really a “the interest rate” in addition to interest rates at all?

The Classical Theory: Knight’s “Capital, Time and the Interest Rate”

We can start with the question of why there is “the interest rate” by answering what the so-called classical (or neo-classical) theories of the interest rate even were. This will land us with Keynes eventually via a roundabout process while demonstrating both that his critique is non-empty and the benefits of his positive theory. The cost is that occasionally there will be long strings of strange names that seem to have nothing to do with market analysis. I believe this cost to be worth getting a feeling of the vast literature that existed even in Keynes’ time and therefore the trade-offs he had to consider in making analytical choices. However, I will try to keep things in real market behavior as much as possible.

Leaving markets just for a moment, Keynes (in “Alternative Theories of the Rate of Interest”) talks a bit about his own intellectual genealogy. He explicitly names Fisher as the origin of his thinking: Fisher -> Hawtrey -> Robertson -> Keynes. We won’t take this route because these thinkers tended to be analytically dense rather than lend themselves to a straightforward exposition. Richard Kahn’s book The Making Of Keynes’ General Theory goes through that genealogical process in detail.

Coming back to pre Keynesian theories of interest, we will find that, like all classical theories, the classical theory of the interest rate is concerned with the causes of price - why does the interest rate exist at all? What classical economists meant by this question is subtle. In market terms, the relative prices of spot and future contracts have a marketwide tendency to correlate. While there are many idiosyncratic forces which shift interest rates in individual financial contracts, the correlation is strong enough that one can talk of a period of interest rates declining or rising altogether. However, that correlation alone does not tell us what anchors the rate. Again, back in the introduction, we saw a money market interest rate drive the housing interest rate rather than shifts in the physical costs of housing production.

Irving Fisher’s genius was to explain this via “intertemporal arbitrage”. If a particular good happened to have a persistently unusually good interest rate, then the actors on the market would be best off if they only bought futures in that good and then bought other goods on the spot market when they needed them. Of course, this simple thought experiment ignores important issues like finance, liquidity and fundamental uncertainty. Fisher was very clear that he considered this an approximation. Frank Knight in Risk, Uncertainty and Profit calls attention to what he calls the “Hegelian contradiction” in all this: intertemporal arbitrage equalizes rates but if rates are equal there are no arbitrage profits to be made. Modern finance theory has deeply analyzed these questions in logical and quantitative terms.

Turning back the clock is justified by clarity rather than depth. And so, I will go with Frank Knight’s exposition in “Capital, Time and Interest” because Knight’s aim of foundational clarity rather than analytic breadth suits our purposes.

Knight’s targets in this paper are the theories of interest descended from that developed by Englishman William Stanley Jevons. These would be picked up by Austrian economist and finance minister Eugen von Bohm-Bawek in an exposition so foundational that they became known as “Austrian” theories (though they were actually developed internationally).

Leaving behind the name issue, the distinction between these lineages is well captured by their answer to the question “What is capital?”. Fisher defined capital as income-producing assets while Jevons defined capital (technically, the “amount of investment of capital”) as the sum of physical inputs to the wage fund multiplied by the length of time of the average period of production. You can probably tell which of these is more fundamental for market analysis.

Moving into said market analysis, the theory of the Jevons side of the family tree, as expounded by Böhm-Bawerk, identified three causes of interest. I will give each of these causes and how they can be tied to the housing market. These mechanisms are not offered as literal descriptions of how market interest rates are set, but as the conceptual foundations Böhm-Bawerk believed must underlie any coherent theory of interest.

• Pure time preference: present income is valued more highly than future income (holding quantities constant) for fundamental psychological reasons.

  • Putting money into an account to create a wage fund for workers on a house over time means less consumption today. Thus profits must be high enough tomorrow to justify that loss.

• Pure technical progress (“roundaboutness”): longer and more capital-intensive production processes yield a greater physical output, which makes waiting economically advantageous. That waiting must be compensated if such processes are to be undertaken.

  • The inputs to housing (like electrical equipment and plumbing) improve over time. What used to be done with a hacksaw can be done with a Sawzall. Those improved equipment are, at least qualitatively, the result of longer-cycle capital which - allegedly - must be maintained with an interest rate.

• Technical-progress-adjusted time preference (“wealth effects on valuation”, “agio”): the marginal utility of income falls over time as people expect to have greater wealth in the future due to the accumulation of capital.

  • As the stock of capital rises, the price of equipment like nails and wire falls. Thus we need a premium on present goods, which manifests as interest.

Theorists in this “Austrian” school vary on the process of how these forces (all of which are very real) work themselves out, as well as the question of which is the most fundamental, etc.. But all “Austrian” theories agree on the existence of an “interest rate” determined by productivities, fundamental psychology, capital structure and the stock of wealth - the natural rate of interest. This idea of a given ‘natural rate’ which actual interest rates can equal or not is the source of the diagnosis of housing supply issues as temporal disorganization from the introduction.

We now turn from exposition to criticism. Knight is correct to point out that the empirical and logical arguments for these sources for interest rates are weak as determinants of an observable market rate rather than as background conditions. Let’s go through them one by one.

First pure time preference: this theory is - to the extent it is not wrong - circular. One cannot meaningfully explain interest by interest by presupposing discounting. These two concepts are simple algebraic transformations of one another. Further, people trade consumption across time depending on many local factors such as wealth, risks peculiar to different income-producing instruments and the opportunities available. Every person and every good should have a constantly shifting discount rate, not one market rate. If there is a market mechanism to harmonize psychological discounting then the mechanism is doing all the work, not time preference.

Next, roundaboutness: as old-growth forests are increasingly harvested, the quality of wood has declined but there is still a positive interest rate in lumber markets. Lumber is an important input to housing markets, showing that this declining roundaboutness is not driving rates. Roundaboutness or technical progress cannot by itself explain why there is a dominant and positive market interest rate. Even if roundaboutness is interpreted structurally rather than qualitatively, it does not explain why one ‘natural’ rate coordinates heterogeneous production processes. And to the extent that coordination happens, then the market coordination process is doing the work, not the roundaboutness itself.

Finally, the agio. On the one hand, wealth effects on valuation are simply the overlap with the previous two. But even if we allow this a non-circular place, why do wealth effects explain the correlation of market rates rather than act as a decorrelating force? Remember it was this correlation which the whole theory was designed to explain. But why can’t a wealthier society tomorrow push the interest rates on copper and gold wire apart rather than together? It seems that the wealth effects are just one market force among many, rather than fundamental. More broadly, the whole notion of technical progress as capital accumulation is highly questionable.

In each case, these mechanisms contribute to individual valuation, but none explains why a single ‘natural’ rate of interest coordinates heterogeneous assets across time. All of them are subject to idiosyncratic reversals at any time: A person who - perhaps by windfall gain - is in a position to consume more today than his psychology prefers will trade away some of their consumption today for income tomorrow. And yet market valuation means discounting that income like anyone else. Any attempt to ground interest rates in an absolute preference for consumption today breaks on this point.

Knight sees all this as sheer error; the word “fallacy” is never far from his lips. But the Böhm-Bawerk approach could be defended as a provisional and approximate place. Still, the methodological sign of what Knight sees as “error” is that in the Böhm-Bawerk system, the interest rate is treated as not an objective relative price found in financial contracts but rather a parameter. This is one of the twin pitfalls mentioned in the opening.

Continuing with Knight, he develops from his critique what he calls his five “main propositions” for the classical (Fisherian) theory of the interest rate. They are, in his order:

1. The quantity of capital is equal to the present value of income generated by that capital discounted by a uniform rate.

2. The quantity of capital is equal to its construction cost = physical (opportunity) cost + carrying charge to entrepreneur for taking on the risk/uncertainty of holding that capital.

3. The value of the carrying (opportunity) cost is discounted at the same rate as physical cost.

4. The rate of discounting construction cost is the same as discounting income.

5. Any creation of an income yielding item is made “under the assumption that the [discounting] rate is the maximum possible under the technical circumstances … of the economic situation in which it is made.”.

Knight would not deny his “main propositions” include important simplifications. But still, his “main propositions” power a transparent theory of interest rate based on the market valuation of income-producing instruments. So far, the interest rate is a relative price. Valuations are primary, psychology and technology matter to the extent they affect valuation.

From a Keynesian perspective, the first four assumptions are acceptable, with similar caveats as Knight or Fisher themselves would make. To see this in terms of Keynes’ first postulate of classical economics (which Keynes provisionally accepts), remember that Knight treats labor as just one form of capital.

But the last assumption is key to making this a (Keynes’ sense) classical theory of interest: it is equivalent to effective-demand full employment. Knight does not allow for the depression condition of shoeless children whose parents have been fired from the shoe factory. Thus Keynes cannot agree with Knight’s “main propositions” in the manner Knight uses them.

Knight’s assumption implies a scarcity in intertemporal exchange under competitive markets - the only way to create more income tomorrow is to “invest” rather than consume today. Income tomorrow is treated as technologically determined rather than allowing for demand constraints.

But this moves us away from market valuation. To paraphrase Keynes, Knight’s fifth assumption means that under free competition “the utility of the price of capital goods when a given volume of capital goods are employed is equal to the marginal disutility of that amount of employment.”. With Knight’s notion of labor as capital, this really does become Keynes’ “second postulate of classical economics”.

This all has gotten very abstract. How does it all shake out for market analysis? Again, by the means of his assumptions, Knight is able to ground interest rates in the equilibrium of competitive intertemporal markets, not the improving technology or the psychology of time preference. Interest rates are prices, not parameters. This is a great insight and Knight is right to insist on it and its consequences.

But how does he explain the empirical fact that interest rates are a spectrum which depend on the peculiarities of particular investments? If spot and future prices are both heterogeneous in an economically meaningful way, then interest rates must differ. Knight could reply: “if different assets truly carry different effective discount rates, then what you are calling ‘interest’ is no longer interest proper but a mixture of pure interest, risk and uncertainty.”.

But the Keynesian has a devastating countercritique: what Knight is calling the interest rate is no longer within the objective market reality of relative prices! Knight defines the interest rate as what remains after abstracting from risk and uncertainty, but that is not what is written in financial contracts. He has fallen into the trap of regarding the interest rate as a residual, not an actual price ratio. This is equally dangerous as treating interest rate as a parameter.

The residual vs. parameter issue is part of the reason that the debate between Knight and Hayek seemed interminable. Maybe Knight’s theory of interest has advantages as an explanation of interest as a market variable. But his residual is defined by a limit concept, and that limit never occurs in real markets. Knight then implicitly issues a vague promise that the rest of the way to observed prices will be covered by some continuity principle.

Hayek, Knight’s direct opponent in this paper, simply prefers to start with time preference as a first approximation. He also implicitly relies on a vague promise that the rest of the way to observed prices will be covered by a similar continuity principle. Hayek’s approach certainly has its own advantages. After all, Hayek wants to work away from the market equilibria Knight focuses on. Who is to say which advantage is decisive?

Coming back to methodology, the principle we call quantifying the quantifiable is key to getting past this impasse. Keynes was only able to hold to his principles and avoid these dual pitfalls by making fundamental changes to the standard tools of market analysis.

Keynes’ Critique of Untheory

In the introduction, we asked the basic question of which operative factors determine interest rates whether or not the market is temporally organized. In the last section, we went over some interest rate concepts which seemed inadequate for market analysis - namely they turned interest rates into parameters or residuals rather than real market variables. We know from history that Keynes needed to change certain fundamental tools of market analysis in order to solve this conceptual issue.

Before we go into details of those changes, I’d like to give an overview of Keynes’ concept of what we now call “the macroeconomy”. Keynes’ chief concerns are with the stock of money (in a given year), the stock of employed labor, the rate of money-wages and the flow of money income. These are the kind of economic realities that you recognize at a macro desk. What we see is that Keynes will not offer a control theory of the economy here (changes in the stock of money do not mechanically change the money wage or money income), but a diagnostic logic for understanding the level to which prices and his fundamental variables tend to stabilize under uncertainty and seeming “discoordination”.

Let’s pause and center ourselves in Keynes’ tests. Keynes discusses interest rates in Chapter 13 (see also) and their connection to capital in Chapter 17 (see also). The central devices which Keynes uses to analyze capital are “liquidity preference” and the “own-rates of interest”. Together with what Keynes calls the “schedule of the marginal efficiency of capital”, a money-denominated schedule governing investment decisions from chapter 11, these elements complete Keynes’ theoretical system.

These arguments have deep logical interdependence. Keynes’ notion of a schedule of marginal efficiency of capital is very different from a typical productivity curve: the marginal efficiency curve is financial first and foremost. We need to establish the deep foundations first so that modeling decisions are properly motivated rather than smuggled in.

I didn’t make the decision to insist on these details lightly. Misunderstanding the modeling decisions has caused people to misunderstand Keynes’ system. We see this historically: in the wake of Keynes’ General Theory, there was a flood of claims that Such And Such Writer/Tradition Knew It All Along. They even did this to Keynes’ face, see his “Alternative Theories of the Interest Rate” for a class of such claims.

To understand Keynes’ modeling decisions, we will run through Keynes’ Chapter 2 (see also). These are about the stock of employment, not interest rates but we can link these to interest rates and capital. It will remind us that “Yes, Virginia, There Was A Keynesian Revolution”, what the revolution was and why it was needed. Then the modeling decisions will be more comprehensible.

In Chapter 2, Keynes separated two claims which play deep, structural roles in what he called ‘classical theory’. Those claims are:

1. Wages (and prices generally) equal marginal product

2. The volume of labor (and inputs generally) is a function of the marginal product.

Today, we would not call these “postulates” as Keynes did. The postulates of our concept of ‘classical theory’ would be more like consistent preference rankings, convex production functions, etc.. Even at the time, Leontief interpreted Keynes’ core claim in a modern way: as the labour supply function not being homogeneous of degree zero in wages & prices (a characterization Keynes accepts in “The General Theory of Employment”).

Despite that, Keynes’ postulates still play an important role in the structure and interpretation of many modern marginalist economic frameworks. Keynes’s postulates no longer enter as behavioral axioms, but as “closure conditions” for market analysis. Thus Keynes’ approach to closure conditions is still very relevant to modern economics.

The first postulate Keynes’ provisionally accepts as a shorthand for sound market analysis. Notice that the first postulate (wages = marginal product) is an attempt to find an observable market anchor for labor, quantifying the quantifiable. Going deeper than this postulate is possible. Joan Robinson showed how Keynes’ system can be employed for a deep analysis of wage determination by investigating the feedback of wages onto profits, general equilibrium and imperfect competition. But in terms of investigating Keynes’ relation to classical theory, it’s best to just follow Keynes and provisionally accept the first postulate ‘with the usual caveats’.

Moving on to the second postulate, which Keynes rejects. There has been evasion on this issue, but the claimed causality from marginal product onto level of employment is indeed a part of classical economics. Even those who saw issues (including Fisher and Hayek above) did not build an explicit, operational replacement for the second postulate’s closure conditions.

That the second postulate was part of classical economics at all can be shown reading the relevant texts and by logical analysis. Starting with a clear textual example, that “purist of marginal theory”, Philip Wicksteed, puts the second postulate and its consequences bluntly in The Common Sense Of Political Economy, Book 3:

We will now turn to the connected problems of unemployment, depression, and commercial crises, which are admittedly amongst the most baffling on the whole field of applied economic science. … Every one knows that persons, not without some dexterity both of mind and hand, may be absolutely unemployable in a given post. Every busy man has had embarrassing offers of “help” from zealous friends who are willing to do anything,—but who can do nothing that does not require more superintendence and correction than the result is worth. …If this fact could be universally recognised, one cause at least of unemployment would be removed or qualified; for it is obvious that the attempt to maintain a standard wage, or to fix a minimum wage, independently of fluctuations in the market of the product, must, so far as it succeeds, throw men out of work when the demand falls, until the marginal value of the reduced product and the marginal significance of the reduced number of workers bring about equilibrium.

I almost prefer Wicksteed’s plain admission of Keynes’ second postulate to the misty evasions that one sometimes encounters. But even without our good old honest Unitarian Minister Wicksteed, one can see through the mist - that is, the postulate can be seen to be a deep part of traditional theory directly.

We will start that direct path with the most solid step in the classical method: prices are set on the market. The solidity makes sense, prices are part of the objective world after all.

The next step is not as solid, but seems to be reasonable: given the prices, managers choose a level of input (say, labor) so that the marginal product of that input is equal to the price (with the usual caveats, of course). This relies on the notion of a “production function” independent of the market activity (Joan Robinson has an excellent critique of the implicit causal order). A typical qualitative example from Frank Knight’s Risk, Uncertainty and Profit is given below.

The x-axis is the number of labor hours per day and the y-axis is the apple output per day at a given plant. Region (1) to (2) is the region of increasing marginal product, after point (2) marginal product is monotonically decreasing. We can ignore region (1) to (2) as unstable in the sense that a firm can have more output (and thus profit) with the same marginal product for input levels between (2) and (4). Wicksteed, to give his point sharp definiteness, discusses the region after (4), negative marginal product but of course no rational manager chooses such a level of input usage. Wicksteed’s, and classical theory’s, real point is that there is exactly one volume of input where the marginal product is stably equal to the wage. In this manner, marginal product determines the volume of input and in particular labor usage - Keynes’s second postulate.

This analysis is, of course, partial equilibrium. The situation can change with general equilibrium. See, for example, IS-LM or more modern “general equilibrium with finance” models. But changing the model’s scope from single market to multimarket (from partial to general equilibrium) does not by itself change the underlying closure logic. Thus it is worth seeing Keynes’ general equilibrium thinking, so that we don’t lose track of sound economic reasoning when models become complex.

Now we’ve gone through Keynes’ Chapter 2. The question is now: How does Keynes’ second postulate tie to interest rate theory?

What is inexplicable in Wicksteed’s theory are the real facts of “depression, and commercial crisis”. On Monday, the shoe factory-employed parents and their children had shoes. Managers happily cited to Wicksteed and pointed to their production functions to explain their plentiful job creation. On Wednesday, the market collapses. On Friday, unemployed parents try to comfort their shoeless children while managers glumly cite Wicksteed and point to their production functions to show they are employing as many as they can. The technical facts of overcrowding (“too many cooks”) hold every day, independent of the level of employment.

This demonstrates the weakness of classical theories with respect to the business cycle. Similar paradoxes exist for problems of growth and monetary theory. All these issues have one thing in common: time. And time is also intimately tied to interest rates. Thus one can hope for a “General Theory” that will solve the problems of interest rates and business cycles. That’s what Keynes was able to do in the General Theory. Not because it is depression economics, but depression was a rare time where different theories of the interest rate act qualitatively differently. As said in the introduction, it was the lesion study which revealed the normal function in that area of the economic brain.

Of course, Keynes did not discover these issues for himself. There was already a vast literature on them, as the previous section’s parade of names demonstrated. To return to the above example, economics seems to work fine on Monday and Friday; it’s the comparison between them that breaks Wicksteed’s closure logic. Wicksteed’s logic holds pointwise in time (with the usual caveats), but fails across time. The standard pre-Keynesian reaction then was, as economist Paul Krugman discusses in his introduction to the General Theory, to move from foundational issues to complex dynamical models.

Some of these models were very insightful, such as Fisher’s concept of debt-deflation cycle (deflation->firms can no longer finance credit stably-> bankruptcy -> credit tightens -> deflation). But Fisher’s powerful intertemporal arbitrage tool becomes undone during debt deflation as interest rates deanchor. Krugman’s point is seen clearly through the questions “How can interest rates ‘deanchor’ in the first place? Aren’t interest rates set by intertemporal scarcity and physical productivities? The factory didn’t change from Monday to Friday, after all. So what sets the price of an asset, say a bond, during a period of debt deflation on Wednesday?”. Fancier models only hid the underlying flawed closure logic.

Krugman’s remarks are related to what economist and historian Gonçalo Fonseca calls Keynes Critique of Untheory. These ideas were rich, they were insightful but they were never part of the foundational structure of market analysis.

This is all becoming abstract again. Let’s look back at the week where we went from many shoes to few shoes. On Monday, there was an incentive to buy and hold shoes to sell for profit on Friday. Then there would never be a period of depression in the first place, just a seasonal production structure where production happens each Monday. So why wasn’t that investment actually made? Why did labor remain purely passive instead of saving? Obviously, the level of saving would be regulated by the interest rate: the future profit of holding shoes is discounted. Something had to raise the effective hurdle rate for carry.

In actual markets an interest rate, a risk premium or a liquidity premium/financing constraint can block intertemporal arbitrage. Classical economics can only bring in liquidity as a “rich insight” disconnected from fundamental theory. Liquidity would block the second postulate by making the level of employment a function of something other than the marginal disutility (i.e. by the financing). What they did was to treat liquidity as a sort of viscosity, effecting timing but somehow optional to fundamental closure logic.

Thus, ignoring liquidity as incidental to fundamental theory, we also see clearly a risk premium will not block our little example, the lack of shoes in the future is an opportunity not a risk. Perhaps if everyone buys shoes to save on Monday profits will be low (a “crowded trade”) but that’s still better than a depression. Thus the only systematic avenue consistent with classical closure to maintain a persistent arbitrage failure is a sudden spike in the interest rate. And now the determinism of classical theory becomes millstone: if the interest rate is set by time preference or productivities in market equilibrium, then there is nothing left to stop intertemporal arbitrage.

Knight has an answer to this: interest rates don’t need to spike to make depressions impossible because the future is uncertain! We didn’t know on Monday it would all collapse. The marginal parent is indifferent between saving another pair of shoes for an uncertain future and holding liquid purchasing power. And yet liquidity was presumed from the start not to be a concern. Knight was not clear on what he thought the alternative to investment is in the face of uncertainty if not liquidity. But, whatever the alternative to investment is a choice between, the marginal parent’s decision is influenced by current interest rates. Yet the interest rates that influence decision making are - necessarily - the objective price ratios, not unmeasurable parameters or residuals.

In this manner, we see how deadly the trap of making the interest rate into a parameter or a residual is. It’s important to see how this is specific to the interest rate. Economic analysis needs parameters and structural variables. But the interest rate cannot just be a parameter like the slope of the schedule of demand. True, demand schedules cannot be observed directly because they involve counterfactual situations (quantity demanded at prices that have not occured). But demand schedules have operational meaning in governing how observable prices respond to changes. They exist as shorthand for consumer based constraints on variation and can thus be inferred from market behavior even if not directly observed. Some (not all) market questions can be translated into subquestions of supply and demand theory. For instance, consider the question of seasonal cycles in the market for fish. The question of the quantity available at a given time as the seasons change the cost of fresh fish can be helpfully translated into statistical questions like “Is the demand schedule for canned saury stable? What about fresh?”. By contrast, Knightian interest has no obvious comparative statics nor does it constrain observable price movements in other ways. Though derived from market equilibrium, there is no market in which the rate itself clears.

To summarize, we have shown that if the interest rate is knowable and determined by productivity conditions, then intertemporal arbitrage destroys the possibility of a commercial crisis. There are twin pitfalls, determining before the market (as a parameter) or after (as a residual). In parameter theories, we have an invisible parameter as the true cause of all business crises - hardly an improvement on just calling the crisis exogenous. Meanwhile, Knight and related theorists’ reduction of interest rate theory to arguing about a residual has parallel issues. Knight consciously allows for leaving behind commercial crises, etc. for the mere analysis of supposed residuals. This amounts to willingly leaving market analysis behind for what Keynes calls “too easy, too useless a task” of telling us “that when the storm is long past the ocean is flat again”.

This is the impasse Keynes guides us through. We must ground capital theory in market analysis while breaking the second postulate, to allow for depressions, growth and monetary theory. That is not easy. Keynes’ was the beneficiary of both of his branches of ancestors - the Fisher side and the Jevons side - even if he wasn’t completely fair to one branch. But this interest rate economics was one of the fundamental tasks that Keynes’ needed to solve. We now turn to his solution.

Keynes’ Positive Theory Of Interest

In the introduction, we looked at some market puzzles to motivate our questions: why can we have inventory gluts amid scarcity? why do some financial instruments act as money and not others? We saw that these can be turned into the question “What is the interest rate?” We saw how classical approaches evade the market questions by making the interest rate into a parameter or a residual. And finally, in the previous section, we saw that the only way to fix this is by making fundamental changes to the process of market analysis. We have finally come to the time for positive theory: what are those changes Keynes had to make?

Read more

• George Elliot, The Mill On The Floss

This section by section analysis of William James’ 1896 talk “Will To Believe” originated in an online reading room. James’ essay is an important analysis of decision making under uncertainty, often caricatured as a license for irrationality. Read carefully, it is something far more powerful: a theory of fixation belief that treats moral risk as epistemically unavoidable and perhaps even desirable.

I want to acknowledge the particular contributions of the users ktb, AmourePrope, Shishi Qivshi, Jason Oakes and Simon for their questions, answers and perspectives. I also want to thank CVAR Editor Alex Williams for his patience and contributions. Statistician Cosma Shalizi inspired some of the basic insights. Others have contributed to this little essay as well. However, I am responsible for any errors in the following.

Contents

1. Introduction

   1. Who was William James

2. Precursors

   1. Charles Peirce’s “The Fixation Of Belief”

   2. William Kingdon Clifford’s “The Ethics Of Belief”

   3. Wilfrid Ward’s “The Wish To Believe”

3. The Will To Believe

   1. Complete essay, section by section

4. Conclusion

   1. Reactions and Aftermath

Subscribe now

Introduction

William James, source

William James was one of the greatest philosophers and psychologists in American history. He had two enormous classics - Principles of Psychology and Varieties of Religious Belief - half of either of which would register their author for immortality. James’ central philosophy of psychology is that the mind or spirit is made of brain. Thus the methods of anatomy (especially neuroanatomy) and physiology are applicable to psychological problems and vice versa. Introspective and behavioristic evidence informs us about neural structure and neural structure underwrites mental and physical behavior.

Today we are reading one of his smaller classics, “The Will To Believe”. This little essay on the process of rigorous belief formation as observed in the human organism illustrates his deep analysis and beautiful style.

But before we start, a bit about James’ writing style. James abounds in paradox, and particularly in long ironic passages where each end balances an over strong conclusion. It makes him a bit windy by modern standards but is enjoyable once you get the rhythm. For example, here is a passage from Varieties Of Religious Experience:

“The next step into mystical states carries us into a realm that public opinion and ethical philosophy have long since branded as pathological, though private practice and certain lyric strains of poetry ​seem still to bear witness to its ideality. I refer to the consciousness produced by intoxicants and anæsthetics, especially by alcohol. The sway of alcohol over mankind is unquestionably due to its power to stimulate the mystical faculties of human nature, usually crushed to earth by the cold facts and dry criticisms of the sober hour. Sobriety diminishes, discriminates, and says no; drunkenness expands, unites, and says yes. It is in fact the great exciter of the Yes function in man. It brings its votary from the chill periphery of things to the radiant core. It makes him for the moment one with truth. Not through mere perversity do men run after it. To the poor and the unlettered it stands in the place of symphony concerts and of literature; and it is part of the deeper mystery and tragedy of life that whiffs and gleams of something that we immediately recognize as excellent should be vouchsafed to so many of us only in the fleeting earlier phases of what in its totality is so degrading a poisoning. The drunken consciousness is one bit of the mystic consciousness, and our total opinion of it must find its place in our opinion of that larger whole.”

Both the phrase “Sobriety diminishes, discriminates, says no; drunkenness expands, unites, says yes.” and the paraphrase “Alcohol is in its totality so degrading a poison.” fail to capture James’ meaning. Both are necessary to understand the balance of his thinking.

James certainly had straightforward and unironic passions and wrote about them plainly. Particularly, he had a deep rooted hatred of racism and trusts. His vicious description of Theodore Roosevelt could fit John Ganz:

“The worst of our imperialists is that they do not themselves know where sincerity ends and insincerity begins. Their state of consciousness is so new, so mixed of primitively human passions and, in political circles, of calculations; so at variance with former mental habits; and so empty of definite data and contents that they face various ways at once.”

But, by and large, James operates in the world of qualification and shading, a world of delicate and descriptive detail. If someone tells you James’ meaning can be distilled to a simple slogan, they are probably trying to sell you something.

Precursors

Now that we have established who William James was, let’s talk about how he wrote this talk. I decided to take a “genetic” approach, examining him in the light of important precursors. After a lot of thought, I pared down the predecessors to discuss to three essays:

1. Charles Peirce’s “The Fixation Of Belief”

2. William Kingdon Clifford’s “The Ethics Of Belief”

3. Wilfrid Ward’s “The Wish To Believe”

Obviously there are many other influences and inspirations, like George Eliot’s The Mill On The Floss quoted above. Further, it must be remembered that James was what they called a “man of science” first and a “man of letters” second. As we shall see, laboratory practice and discovery is as important an influence on his work here as any “predecessor”.

The first paper up is Peirce’s “Fixation of Belief”, published in 1877. Charles Peirce was one of James’ biggest influences as the two men struggled to plant the seeds of empirical psychology in the rough soil of 19th century America. A bit about who Peirce was in relation to James will help. Best to look at Peirce through the lens of his empirical psychology, as this was their original point of contact. Fortunately, the Peirce-Jastrow result is a classic in the field of the psychology of sensation. Methodologically, it could have been written yesterday with its crystal clear statistical reasoning. The basic conclusion that Peirce makes is that perception is probabilistic and graded in terms of intensity. Following this, Peirce and James saw that the problem of psychology is to ground probabilistic and graded judgements (and psychological behavior generally) in the complex and competing activities of the nervous system.

James was familiar with Peirce’s ideas about the fixation of belief from lectures and discussions long before publication. As we shall see, James draws on Peirce extensively for his anti-Cartesianism, psychological concepts and for the analysis of doubt as a problem with a specific structure. But we will also see that James has many differences from Peirce. Most importantly, James is comfortable with psychologistic language (a person feels doubt) whereas Peirce wants more objectivistic language (a person’s evidential situation is doubtful). It is important to be clear that Peirce’s appeal to an unbounded social process was not merely a regulative ideal. For Peirce, the eventual convergence of inquiry reflects the real metaphysical structure of the world, a post-Hegelian commitment James explicitly resisted.For his part, James’ saw Peirce’s process based objective idealism as leaving the laboratory behind. Wherever you land in this tension, keep in mind that if James’ ever seems to be a bit over psychologizing, it is not accidental. James is not replacing epistemic norms with psychology; he is relocating normativity inside psychological constraints.

We move on to Clifford’s essay. Bill Clifford was a very acclaimed mathematician and physicist who suffered a short and tragic life. Clifford’s work on the algebra of spatial analysis is still a fundamental part of modern algebra and his work on the relation of motion to non Euclidean geometries has been praised as a precursor to relativity. But Clifford was an inveterate workaholic: he would teach and experiment all day and write all night, not sleeping for days on end. As a result of overwork, his health collapsed multiple times, and at the age of 34 he caught a case of tuberculosis that his stress weakened immune system could not defeat.

Tragedy aside, we now move on to the writings. Clifford’s essay “The Ethics Of Belief” is half of a two part series published in 1877 with “The Ethics Of Religion”. Now, “The Ethics Of Religion” is mostly dedicated to anti-ecclesiastical thought. This is not without interest, but “The Ethics Of Belief” is much more interesting. This little essay is a foundational work in what has been called the “agnostic movement” of the late 19th century.

The agnostic movement was a complex and many-sided thing, but all of its protagonists shared the conviction that spiritual life had to be reconciled with modern science. There were atheist-leaning agnostics like William Clifford and belief-leaning agnostics like William James. There were even agnostics named something other than William. The movement was controversial even among non-fundamentalists. Nietzsche contended that the English agnostics replaced religion with an ethic more austere without God than the Christian ethic had been even with God (see his unkind comments about George Elliot for more details).

Clifford’s essay starts by raising the stakes: an example meant to convince the reader that if you have the wrong beliefs then people could die. Clifford’s example is a ship with major flaws that is put out to sea because the owner has convinced himself that the ship must be safe because that’s the only way he can make money off of it. Clifford had just survived a sea wreck so he knew what he was talking about. Clifford’s response to the alleged deadliness of false beliefs is to insist that all our beliefs be founded on something he calls “evidence”. Despite the seeming reasonableness of this requirement, Clifford immediately runs into two problems:

1. Surely the future belongs to those who cry ‘Peace’ when there is no peace?

2. Many of our most important beliefs are founded only on the slim evidence of authority. For example, the evidence that people die when shot by guns I believe on the merest authority.

Clifford’s solution to the first problem is suspension of belief - essentially, choosing to avoid moral risk at all costs. There’s been a lot written on this since 1877: Keynes on validation of investment, Merton in self-fulfilling/negating prophecy, fixed point theorems, Schelling on focal points, the list goes on. James has a sharp thing or two to say about this. As we shall see, James would see Gramsci’s famous phrase “Pessimismo dell’intelligenza, ottimismo della volontà.” as closer to a solution than any kind of ‘suspension’.

Clifford’s answer to the second question is what separates him from a naive evidentialist. Clifford argues that though we may not be able to shake off authority all at once, we can always at least begin the process of inquiry. This is directly related to the James/Peirce pragmatist program. Now, nobody can read Clifford’s essay and think he has answered the issues he brings up. A G K Chesterton could have easily reduced him to ridicule. But he writes with honesty and passion far more powerful than jibes can deflate.

Now to finish with Ward’s 1893 essay, “The Wish To Believe”. Wilfrid Ward was an English Catholic apologist, an influence on G K Chesterton and the other apologists of that generation. The essay is a dialog between largely but not exclusively between two thinly sketched characters, a young agnostic and a Catholic priest. It goes pretty much like you think - the priest offers brilliant answers to all the agnostic’s questions & suggestions. The essay is enlivened more through Ward’s impressive erudition than by surprise.

For our purposes, the important bit comes in the first day of dialog when the priest ‘refutes’ the young agnostic’s objections to the Catholic approach to the evidence for miracles. The agnostic says that the traditional religious person’s attitude toward the evidence for religion is untrustworthy. Why? because the religious are intrinsically biased: they wish to believe the evidence has such-and-such conclusions. The priest then says “Sure but there are two kinds of wishes: Wishes like ‘Loving eyes can never see’ and wishes like what Michael Jordan had to defeat the Pistons. The second kind of wish should increase our respect for the reasoner. Only by such wishes can we be convinced to put forth the effort to gather the evidence of a hypothesis as unlikely as miracles.”.

Ward’s second type of wish is an important point for James, related to the first point that tripped up Clifford. What it highlights is a volitional, effort-taxing aspect fundamental to rigorous belief formation. Hard thinking burns calories and strains the nerves. Thus rigorous belief formation is subject to the same kind of Peirce/Jastrow/James empirical psychology as any other strenuous action. From this insight “The Will To Believe” was born.

Subscribe now

The Will To Believe

Introduction

In this section, James is introducing himself to his audience and situating himself in the intellectual world they would know. William James, a good Harvard Man if also a lifelong outsider to all institutions, gently welcomes himself to the “old orthodox College” - i.e. Yale. The orthodoxy that led to the foundation of Yale is Congregationalist Reform Christianity with its sola scriptura and its sola fide.

Moving from the physical location of the lecture, James situated himself in the intellectual world by discussing Leslie Stephen. James had actually met Stephen in 1882 while working on Principles Of Psychology. Leslie Stephen “met” someone else special that year - his daughter Virginia was born that year. Virginia would later cite William James’ concept of “the stream of consciousness” as an important influence on her novels.

But as to why Stephen is quoted here: Leslie Stephen was the social center of the English “agnostic” movement. His essay “An Agnostic’s Apology” is best summarized by this sentence: “We [agnostics] wish for spiritual food, and are to be put off by these ancient mummeries of forgotten dogma.”. Stephen’s name would immediately signal to the audience they are going to hear a work of agnosticism (albeit belief-leaning agnosticism).

James discusses sola fide a bit. This is a central doctrine of large segments of Protestant Christianity. The basic idea is that the Savior sacrificed himself thousands of years ago for you out of sheer personal love (that is, unmerited grace). In some sense and compressing significantly, the point of the imitation of Christ is to recognize oneself as sanctified by the God who always loved you, not to “earn” something. How can you earn an infinite reward anyway? Punning on sola fide, James declares the goal of the essay is justification of faith alone, that in important circumstances the faith in a belief is a sufficient call to hold that belief.

Of course, James’ notion of “belief” is not the same as Martin Luther’s. In James’ concept of “holding a belief”, the belief must be understood in the light of fallibilism. In Quine’s much later phrase, James does not mean believing “hold come what may”.

Now we come to the most important paragraph in the introduction, which guides the whole piece. James says that his analysis will show his iustificatio fidei is “lawful philosophically”. That is: it can be an ordinary part of scientific explanation. In fact, James thinks that all religious experiences, not just the justification of faith, can be examined psychologically (without settling their metaphysical truth). James would bring his idea of the scientific investigation of religious belief to a higher level with his Varieties Of Religious Experience. But here James is only examining this one, admittedly important, part.

James’ idea that faith can be an ordinary part of rigorous belief formation is a provocative thesis. James goes so far to say that all of his students, Christians or atheists, feel that “faith” (i.e. belief that goes beyond justification strictly speaking) is not a part of ordinary scientific thought. But James will have no such nonoverlapping magisteria - he is determined to examine the process of rigorous belief fixation wherever that analysis covers. In the next section, we will start that process.

Section I

In this section, James is introducing a series of conditions to block the method of Cartesian or Hyperbolic Doubt (Peirce’s phrasing). Hyperbolic doubt was, or is held to be by some intellectual historians to have been, the dominant philosophical method of what might be called Idealistic Pre-Critical Modern Philosophy. The basic strategy was to “doubt” all truth until one finds an indubitable non-tautology, then construct all knowledge on top of that. The issue with this form of foundationalism is that doubt is only brought in temporarily, like a vaccine - use a little doubt to extinguish all doubt. Is this the methodologically correct way to use doubt?

Pragmatists think not. The rejection of hyperbolic doubt by Peirce and James is fundamental to the way they see the world. It takes real neural effort to repair what Quine would later call “the web of belief” from doubt, and both reject the idea that such efforts can be avoided by simple words. Thus James introduces three axioms which delimit the actual doubts as experienced by humans from the merely potential doubts examined by Descartes. They are that the doubts must be

1. Live,

2. Forced and

3. Momentous

First, the concept of a “live” doubt. For a doubt to be “live”, the possible beliefs must be contradictory but not decided. James’ axiom of the liveness of a belief is grounded in his psychological research - he knows the attention fixing and reflective association of belief fixation takes energy and effort. Thus he sees a sort of neural “Production Possibilities Frontier” for the production of new ways of thinking. Without a notion like liveness, every belief would have to be treated as perpetually in the process of being doubted, a demand incompatible with finite attention, finite energy and the need for action in real time. James, a fallibalist, thinks every belief can be doubted but not that they always are.

You’ll notice that James defines the “liveness” of an option in a highly subjective way. The liveness of an option is a relation it has with you, not an objective relation the option has with a body of evidence. This was a major source of conflict that James had with Peirce. Peirce wanted to protect objectivity by embedding rigorous belief fixation in an unbounded social process. James felt such infinite processes brought in metaphysics at the wrong level. This is a subtle but real difference. Put another way: James emphasized the subjective sources of knowledge (immediate experience) but Peirce emphasized that sensation becomes knowledge when organized into objectivity (norm-governed inquiry).

The subjectivity of liveness shouldn’t be confused with the liveness of a proposition being a choice. One can no more choose the liveness of a choice than one can choose to not feel the subjective sense experience of pain when touching a hot wire. The liveness of a proposition is directly perceived. Liveness is “felt in the gut”. Liveness is thus subject to the same illusions and faults as any other sensation.

That the choice must be forced has a further target: Captain Kirk and the Kobayashi Maru. The considerations for an analysis of creativity and finding subtle third options are wholly different from the considerations of James’ lecture and thus must be excluded.

Finally, momentousness blocks the attempt to use arbitrary general rules to get out of analyzing the specific proposition. If a decision was not momentous, then one could try to make a rule like “Choose the option which is the least work.” or “Flip a coin.”. These methods are clearly not part of rigorous belief formation.

These are James’ three axioms for a situation in which an organism would bother wasting precious calories changing its mind. It is clear that all three are necessary - negate any one and an easier way of organizing the mind than rigor is possible. Are they sufficient? Later, we will talk about some readers who think they are not. But for now, we will move on to the next section.

Section II

James now moves on to consider what a belief is, grounding in our “volitional and passional nature”. By beliefs which come from our passional nature, James means those beliefs which are cooked up for the stream of consciousness to provide easy flow. Let’s start with an example of a belief grounded in passional nature: imagine coming upon my gazing into the evening sky. You ask me the color of a particular patch of multicolored heaven. You’re shocked that I quickly answered “Red” - that patch was in my blind spot! You ask why I believed it so and I say that I saw it with my own eyes. You direct my attention to the strange green spot in the sky. What happened? In Peirce’s spirit, I in fact (mistakenly) deduced the color of that patch of sky from its presumed continuity with my other sense experience. My answer was swift because I had no conscious feeling of deduction. My brain did that to ease conversation with the illusion of visual continuity. This process is how the greatest mass of our beliefs is formed: our passional nature builds beliefs for the stream of consciousness just in time for their use.

How then does our volitional nature enter? James illustrates the entry place of volition in his magisterial Principles of Psychology with a passage of Leviathan:

“mental discourse, is of two sorts. The first is unguided, without design, and inconstant; wherein there is no passionate thought, to govern and direct those that follow, to itself, as the end and scope of some desire, or other passion.... The second is more constant; as being regulated by some desire and design. For the impression made by such things as we desire, or fear, is strong and permanent, or, if it cease for a time, of quick return: so strong is it, sometimes, as to hinder and break our sleep.”

• Thomas Hobbes, Leviathan, Ch 3 (as quoted in WIlliam James, Principles of Psychology, Chapter XV “Association”, Section 14 “History of the Doctrine of Association”)

James does not agree with all of Hobbes’ ideas, but does take advantage of this insight. What James means by beliefs grounded in volition is those beliefs which have impressed into our patterns of thought through the law of association via repeated experience in voluntary reflection. Consider why a childhood fancy is as clear as day to an old man who cannot remember where he put his keys: the key place was experienced once, the childhood dream experienced over and over again in reflection. So to the belief we once were so critical of is added to our stock of calcified beliefs by repeated internal experience. One can see why James is concerned with the calories the organism spends in impressing a belief via volition!

Because of the expense of volitional reinforcement and the vast quantity of passional belief, one can understand James’ rhetorical question of whether belief formation has any volitional aspect at all. The illusions of blind spot “vision” cannot be willed away. It would be absurd to try to construct the mass of our beliefs through explicit volition, and yet there are cases where volition takes center stage - those covered by James’ axioms.

James starts by discussing Pascal’s Wager. For now, he is highlighting the essentially volitional core. He shows that if such a framework lacks the axioms for belief formation that James outlined earlier (postpone whether it does until later), then Pascal’s argument is unconvincing even if not answered. Such an argument would be subject to symmetry arguments (why be Catholic rather than Protestant, or Muslim or Hindu or Buddhist or …). As we shall see, this is only one phase of James’ dialog with Pascal.

James continues by discussing the agnostic position. Where volition is at the core of Pascal’s model, to the agnostic, “volitional belief” smacks of arbitrariness. He is aware that the nerve of agnosticism cannot be answered by a purely volitional theory. To answer Leslie Stephen’s wish for good spiritual food by telling him he should just enjoy what he has is as crude as it is ineffective.

James lets Pascal stand on its own, but James brings in many agnostics to sharpen their position. He quotes Florence Nightingale associate Arthur Clough. Clough’s poem gives a kind of agnostic creed: “It fortifies my soul to know/that though I perish, truth is so”. This moral fits with Rorty’s analysis that “truth” is a post-Christian substitute for God. James also takes a Thomas Huxley quote from Influence Upon Morality Of A Decline In Religious Belief. Finally, he brings in Clifford’s “The Ethics Of Belief”, which we have already discussed. This shows that James is engaging with agnostics seriously as a group, not just picking on a few poor souls.

Now that James has set up Pascal and the agnostics as his right and left, it remains for him to show that a continuous organism can be built via psychological analysis.

Section III

The basic point of this section is that belief formation is a process. That process involves volition and passion. Indeed “It is only our already dead hypotheses that our willing nature is unable to bring to life again.”. James presages Rorty in saying “Our belief in Truth itself, for instance, that there is a Truth, and that our minds and it are made for each other, what is it but a passionate affirmation of desire, in which our social system backs us up?”.

As a matter of empirical fact, “our non-intellectual nature does influence our convictions.”. James even psychologizes those who would deny it - what is it but Clifford’s passional nature that led to the writing of his wonderful essay? As Schopenhauer put it, man can do what he wills, but he cannot will what he wills.

This part of James’ writing led some to the complaint that his pragmatism is just rehashed utilitarianism. But here James is making an empirical point: that Clifford and Cardinal Newman are both realizations of the belief making process. From James’s perspective, the distinction between evidentialism and the grammar of assent is real but misses the deeper fact that they are both inside the process of belief formation.

James uses telepathy as a deliberately provocative example in this section. He shows that rational agents can refuse to examine evidence although the evidence has not been refuted. Why? Because sometimes accepting the evidence would disrupt their web of beliefs, institutional practices and explanatory norms to a practically intolerable degree.

This can be shown by a less extreme example. Quality engineer W. Edwards Demming asked when we should believe a table is clean. This is not historically arbitrary - Demming was highly influenced by pragmatist C I Lewis. Clearly, the answer to “Is this table clean?” depends on the question “Clean for what purpose?”. Food preparation and chemical compounding might need different notions of “clean”. Clifford can say that the table is not clean enough for a chemical experiment while Cardinal Newman can claim that the table is clean enough to eat off of. It isn’t that the table is volitionally dirty for Clifford and clean for Newman, but rather given the situations - including their respective webs of beliefs as they are - the proper beliefs change: beliefs depend on purposes.

The result is that James says “Pascal’s argument… seems to be a regular clincher” is another example of his balanced style that I brought up in the introduction. It may seem to be a bit odd: if volition has so little to do with belief that even the great rationalists & agnostics are just navigating their web of belief then what does an argument that appeals to volition have to do with anything? Well, remember the criticism that James made in particular: the argument’s insufficiency in the light of symmetry. Pascal only found the argument convincing because he was 99% a believer already, this merely completed an in place process. At the time, this seemed like a critique but now it feels like a necessary part of any believing as such. The phrase “seems to be” is doing real work in this sentence. Seems to who? To Pascal, of course.

And yet this section is not a rehabilitation of Pascal’s wager. James is showing how the issue of belief formation leads a naive approach to difficulties. Belief is a very fraught topic, that’s the concluding sentence of this section. James’ real point is that the understanding - “healthy” as he puts it - of proper belief formation is going to be work.

Section IV

In the previous two sections, we began with the volitional seeming to have little place and ended with it seeming the volitional should have a central place. Now it is time in this section to transition from criticism to positive theory - where does James think is the right place? James puts his theory core thesis as follows:

Our passional nature must, and lawfully may, decide an option between propositions, whenever it is a genuine option that cannot by its nature be decided on intellectual grounds; for to say, under such circumstances, “Do not decide but leave the question open,” is itself a passional decision, just like deciding “yes” or “no,” and is attended with the same risk of losing the truth.

The right place for volitional belief formation is the genuine options - those belief options which are live, forced and momentous.

Having quoted fully half of this section, I will use this time to discuss two related theories: a precursor from Plato and downstream theory from Neyman & Pearson.

Consider Socrates’ speech on psychology in Phaedrus (and related speeches). Like James, Plato assigns the will a small but sneakily vital role. The metaphor Socrates uses for Phaedrus is that of a stout chariot with two winged horses. The rider of the chariot represents reason or the will. One horse is “guided by word and admonition only” but the other is so stubborn that it is “hardly yielding to whip and spur”. Thus the rider only somewhat controls the horses and chariot, resulting in a jagged and stressful path. Every element of the ride is necessary. If the horses were in agreement, the rider wouldn’t have the strength to control the path when they disagree. If the chariot were not strong enough to hold disagreeing horses, then the rider would have no control. Only by means of tension and balance can something as ephemeral as a thought control brute matter. Only by placing the rider (the will) in a chariot (the body) pulled in diverse directions by animal spirits (desires) can the rider have any control. Only that loose control can end up uniting the lovers in the beautiful city.

James, like many of the other agnostics of his generation, is trying to find a way to get the truth of Plato’s vision - the importance of the divided and tensioned nature of man for the efficacy of the will despite the subjective feeling that the tension is resisting the will - in a post Darwin world.

So moving on to Neyman & Pearson. A little history first. In the development of scientific statistics from Pascal to Karl Pearson the emphasis was on the emergence of patterns after many trials. The law of large numbers dominated early statistics. The basic idea was to throw away particular information about members of a sample to get the essentials of a mass of samples.

Things changed with Student (Bill Gosset) and Ronald Fisher. Their idea was to get the maximum amount of information out of a relatively small sample of a given level of reliability. Jerzy Neyman and Egon Pearson took the limited sample approach in a new direction. One of their contributions is the idea of “Type I and Type II Errors”. A maximally powerful statistical test is one that misses as rarely as possible for a given rate of false alarms. A false alarm is a type I error and a miss is a type II error.

James was not a statistician, but he anticipated the Neyman–Pearson insight that different kinds of error must be traded off. He recognized that genuine decisions - the ones covered by his axioms - must be made by powerful methods, and that powerful methods mean stricter tradeoffs rather than looser.

Section V

In this section, James comes to the question “Where is the truth in all this? Isn’t rigorous belief formation about finding the truth?”. He examines these questions by looking at two ways of considering truth, the empiricist (what we would call pragmatist) and the dogmatic.

First, look at the dogmatic. Think back to what James rejected in the method of Hyperbolic Doubt. James rejects that method because of its consequences. If one finds the indubitable belief, then one is done forever: “A system, to be a system at all, must come as a closed system, reversible in this or that detail, perchance, but in its essential features never!”.

This dogmatism is contrasted with the progressive pragmatism (what he calls “empiricism”) James abducts from scientific practice. James’ emphasis on scientific progress prefigures fallibilist philosophers of science such as Karl Popper and Imre Lakatos.

So what is truth without dogmatic standing grounds? For the James-Dewey-Schiller pragmatist, truth is correspondence in relevant aspects. Not an absolute correspondence that works for all people in all situations. The example James gives in Pragmatism is of the time-keeping-ness of a clock. The naive belief is that the clock is simply isomorphic to time as such. But then what is time as such. James argues the clock is isomorphic to (in the first place) our sense of greater and lesser quantities of time passing and (once the process of inquiry has begun) to the web of clocks real and hypothetical, with clocks being regarded as more and less accurate according to their agreements.

Coming back to Peirce, he would add that clocks must become increasingly isomorphic to the ideal possible clock which each actual clock approximates. Only from there do we get the static nature of truth. James never fully replied in print to this, expressing in his letters skepticism about the alleged infinite process and ideal possible clocks. But now we need to return from the truth to belief.

This section ends on a sobering note about belief formation and dogma. It’s all well and good to affirm pragmatism, but the blocking of Hyperbolic Doubt delimits the quantity of belief we can doubt. In fact, for all of pragmatism’s constant experimentation, the great mass of our beliefs will never be subjected to doubt.

Further, even the most demanding critics of belief, Bill Clifford for instance, must lapse into certainty. The point of doubt is belief formation, not preserving uncertainty. Go back and re-read Clifford’s essay, is this the work of a man wracked with doubts? No, it is firm and confident - in a word, dogmatic. This is not accidental, but psychologically necessary - “healthy” as James put it.

To sum up: our arbitrary psychological nature influences our decisions on all genuine options and dogmatism emerges unless an option is genuine. In essence, that consciousness of arbitrariness of belief is an ephemeral shade compared to the calcified belief result of the somewhat arbitrary decision. How do we cross this impasse?

Section VI

Our most important beliefs - the outcomes of our genuine choices - have an element of arbitrariness and yet this arbitrariness leaves no psychological trace but rather the opposite. Once formed, our beliefs lose the feeling of doubt and irony and become calcified: the half conscious and unreflective basis for future decisions.

Unreflective belief is the natural philosophy of man, a gadfly like Socrates is exceptional. Thus Dr James proposes a new role for evidence, quite distinct from the role espoused by the man James calls the “delicious Clifford”. Evidence is not just for doubt extinction, not mainly for ending the suspension of disbelief. Rather, evidence is a powerful tool for doubt creation. Clifford’s ship owner should not just have suspended belief by lack of evidence, he should have gathered evidence to realize he needed to suspend his belief. With evidence, the shipbuilder’s options would become live and the process of rigorous belief formation would start.

Man left to his own devices will tend towards a calcification. Whether from the laboratory to the historical-critical method, evidence and practice is the CLR that removes the calcium. Again, James here prefigures future fallibilist philosophers of science deeply.

James ends this section by rejecting ‘objective certitude’ while explicitly reaffirming what he calls ‘truth itself’. Objective certitude would be truth unconditioned by interpretive act. If it exists, then objective certitude is pure but weakly attached to ordinary reality. How an objective certitude ever got into a mind is a real mystery. Truth Itself is the truth in the context of interpretive action. Truth itself is a strong valuable metal due to its alloys with volitional belief creation and use-value. Think back to W. Edwards Demming’s analysis of the proposition “The table is clean.”. Do you want truth unconditioned by interpretive act or do you want a place to work/eat? Do you want objective certitude or the truth itself?

Section VII

“‘Believe truth!’ ‘Shun error!’—these, we see, are two materially different laws; and by choosing between them we may color differently our whole intellectual life.”

This is how James motivates his innovative idea that it is our volitional attitude toward risk which is decisive when faced with genuine options. It’s difficult to see James’ analysis when faced with purely verbal examples, but fortunately a simple numerical example helps significantly.

In this example, the belief under consideration is whether interest rates will rise, stay the same or fall. This is actually not as different from Clifford’s example of the ship owner as it may appear - the ship owner had a choice between using his capital to launch the ship or do something else. Clifford correctly notes he should have taken into account evidence. Even if by chance everything worked out right, the ship owner was taking unnecessary risks with human lives.

Coming back to the numerical example, the use-value of beliefs about the interest rate comes from the ways they will affect investment decisions. We will assume this belief is live, forced and momentous for the investor who either holds cash or invests in this situation as well.

The table of returns are below. The table is an illustrative table from Wikipedia, not actual returns. Put in words, in this table stocks strongly anticorrelate with interest rates, bonds weakly anticorrelate with interest rates and money markets weakly correlate with interest rates. Further, the level of fluctuation depends on the asset: stocks have large fluctuations, bonds medium and money market small. How do James and Clifford help us understand this situation?

Clifford is writing directly about our attitude towards evidence, but James interprets his maxim as an attitude toward risk: “he who says, “Better go without belief forever than believe a lie!” merely shows his own preponderant private horror of becoming a dupe.”. One can see that James sees Clifford in the light of “maximin” decision making. Maxmin is a rule for making choices: make the choice which makes the best of the worst outcomes. Of course, this is interpretive on James’ part, not literally in Clifford’s essay. But under James’ interpretation, Clifford is strongly recommending the safe but dull money market.

Applying James analysis, we see the maxmin returns rule is not a necessity. The decision rule is a particular attitude toward risk. “I myself find it impossible to go with Clifford.”, he reports. I understand James’ temperament as aligning better with a minmax “regret” approach. The regret of a choice in a situation is the difference between the payoff of that choice and the best choice in that situation. Numerically, stocks have a max regret of 7 (if interest rates rise), bonds have a maximum regret of 5 (if interest rates rise) and money markets have a maximum regret of 11 (if interest rates fall). Thus I see James as preferring Bonds to Money Markets in this artificial situation.

In choosing a different rule, James does not mean to replace Clifford. He is simply reporting his own psychological attitude towards risk: “For my own part, I have also a horror of being duped. But I can believe that worse things than being duped may happen to a man in this world, so Clifford’s exhortation has to my ears a thoroughly fantastic sound.”. But that fantastic sound had a benefit: Clifford’s tune was very definitely in pitch. How does James’ psychologism avoid thorough formlessness? That we shall see in the next section.

Section VIII

James closed the last section by saying that certain attitudes towards risk are “like a general informing his soldiers that it is better to keep out of battle forever than to risk a single wound. Not so are victories either over enemies or over nature gained.”. Let’s look at an example from the statistics of warfare due to Egon Pearson:

“Two types of heavy armour-piercing naval shell of the same calibre are under consideration; they may be of different design or made by different firms. Since the cost of producing and testing a single round of this kind runs into many hundreds of pounds, the investigation is a costly one, yet the issues involved are far reaching. Twelve shells of one kind and eight of the other have been fired; two of the former and five of the latter failed to perforate the plate. In what way can a statistical test contribute to the decision which must be taken on further action?”

The hypothesis to test is “Is the proportion of functioning shells made by company X greater than that of those of company Y?”. The evidence is company X’s shells worked 10 out of 12 times (pi_x=5/6) and company Y’s shells worked 3 out of 8 times (pi_y=3/8). This is practically right on the boundary for statistical testing - it is hard to tell from this evidence whether the null hypothesis (pi_x = pi_y) or the alternative hypothesis (pi_x > pi_y) better fits the data.

Let’s dig deeper into this example. Tests of close sizes give different answers to the hypothesis test. In the Neyman-Pearson way of thinking, the size of the test determines a likelihood ratio threshold for the rejection of a null hypothesis. Cosma Shalizi has shown that the likelihood ratio threshold can be interpreted as the shadow price of the power of the test. Now, to connect this up with James we must remember the nature of a shadow price: shadow price is a marginal concept. The cost of forming a definite belief is the opportunity cost - the possible beliefs one gives up. At what cost a null hypothesis? Similarly, in James, rigorous belief formation is exactly belief formation at the margin: “The most useful investigator, because the most sensitive observer, is always he whose eager interest in one side of the question is balanced by an equally keen nervousness lest he become deceived.”.

Perhaps it should be noted that it is here, at the root insight, James finally cites William Ware. Whatever the source, James has at this point set up all the wires, we are now ready to close the final switch and complete the circuit. James has finally cashed in the promissory note that this is an ordinary part of scientific reasoning, with the example of Röntgen (or X rays) in medicine. Such practices were still novel in 1896. Though not statistical, the linked paper shows exactly the kind of delicate trade off James has discovered: “I have been much impressed by the practical importance of the Rontgen ray process in surgery, but in no instance more than in this, where, in a case in which every hour had become valuable and every effort at exploration dangerous, it substituted accuracy and promptness for otherwise unavoidable uncertainty and delay.”.

Section IX

Before moving into this section proper, let’s look closer at the above surgical case study. Dr J. William White had to form beliefs about a jack stuck in a small child’s throat in order to save the little girl’s life. For the emergency room surgeon, the decision whether or not to knife open a two year old girl’s throat or to try and pry out the jack with forceps is live, forced and momentous - a genuine choice. James has shown that the surgeon is thinking rigorously exactly when he is at the locus where there is a tradeoff between different types of error.

James does not see this as an isolated incident. James argues that what he calls “moral questions” are always what he calls “genuine options”: live, forced and momentous. Thus James finally is able to give a reply to Clifford as a moralist of belief formation, not just as an analyst of such. It may be helpful to paraphrase James’ analysis in a simple maxim. To paraphrase boldly in bloodless modern language, “Thou shalt hold to such belief formation processes as to be sure your beliefs are at the margin of the tradeoff between the two types of error.”.

James’ analysis, even in the maxim paraphrase above, has many advantages. The maxim brings in evidence in the right ways: doubt creating and resolving, as informing but not decisive, etc.. It works with self-fulfilling and self-negating prophecy in the right way: “Where faith in a fact, based on need of the fact, can create the fact, that would be an insane logic which should say that faith based on inner need, and running ahead of scientific evidence, is the ‘lowest kind of immorality’ into which a thinking being can fall.”.

We finally see how James’ approach accommodates without coddling those who cry “Peace!” when there is no peace. James opened the door to a coherent understanding of “Optimism of the will, Pessimism of the intellect.”.

Section X

Wasn’t this talk going to be about the philosophy of faith at some point? James, having his analysis of rigorous belief formation and its connection to moral questions at hand, is now able to loop back to the question of religious belief.

What is, in terms of belief, a religion? James gives two basic poles that define religious belief: “First, she [religion] says that the best things are the more eternal things, … And the second affirmation of religion is that we are better off even now if we believe that first religious truth.”.

In making this definition of religion in terms of belief, James has chosen a perspective. As T S Eliot puts it in his introduction to Pascal’s Pensées:

“The Christian thinker … proceeds by rejection and elimination. He finds the world to be so and so; he finds its character inexplicable by any non-religious theory … [his religion] account[s] most satisfactorily for the world and especially for the moral world within; and thus, by what Newman calls “powerful and concurrent” reasons, he finds himself inexorably committed to the dogma of the Incarnation. To the unbeliever, this method seems disingenuous and perverse; … he would, so to speak, trim his values according to his cloth, …The unbeliever starts from the other end …: Is a case of human parthenogenesis credible? and this he would call going straight to the heart of the matter.”

James would certainly not treat claims like human parthenogenesis as candidates for a genuine tradeoff between types of error. Such claims simply do not arise as live options for anyone. It is on the basis of such issues that, ten years before this lecture, Nietzsche would encapsulate the troubles with belief leaning agnosticism with a pithy maxim: “Aber eine Neugierde meiner Art bleibt nun einmal das angenehmste aller Laster, - Verzeihung! ich wollte sagen: die Liebe zur Wahrheit hat ihren Lohn im Himmel und schon auf Erden.”.

For James, the interesting question of religious belief is never the ontological question of the mechanics of a virgin birth. The religious question is, for James, a moral quandary - and thus, ipso facto, a genuine option. James then checks each of his three axioms to be sure. He finds in accordance with his earlier analysis, that belief and disbelief then are influenced decisively by our attitudes towards risk.

James, of course, did have ontological opinions. In this section, he has early defenses of his later principles of “piecemeal supernaturalism” and the “pluralistic universe”. But this section of James looks forward to his future work. In Varieties Of Religious Experience, James continues to bracket away the question of existence of eternal things and investigate religious questions as they are on earth.

And so, James ends his sermon with a defense of tolerant agnosticisms as a spectrum of religious hypotheses. Like Popper would later, James also sees a connection between fallibilism and tolerance. Whatever way one’s ontological beliefs lean, the Jamesian tolerant agnosticisms give the parishioner a noble investigative task, and James would continue to be one of its great investigators.

Subscribe now

Conclusions

The reception to “The Will To Believe” was enthusiastic and yet it’s difficult to summarize what exactly the reception was. Both Ralph Barton Perry and Jacques Barzun in their books on William James remark how little intelligent commentary for such a popular lecture. The essays anti-Cartesianism, the phenomenological discovery of the structure of belief formation and the extremely important idea of rigor entailing an accuracy/precision trade off were nowhere to be found in the commentary literature.

Part of this reason is historical. After James, the dominant movement in English-language philosophy was so-called “analytic philosophy”, especially its “ordinary-language” strand. G E Moore can be taken as a representative member of the tribe. Moore’s “Defense Of Common Sense” demonstrates his highly original method. Instead of Descartes’ cogito argument of man as a thinking thing, Moore argues that the foundational knowledge is “There exists at present a living human body, which is my body.”. This is not the original part. According to Brian Magee, this insight is also to be found in Schopenhauer. But what was original in Moore is not this foundational sentence, but how he defuses the possibility of a hyperbolic doubt in this case. In essence, Moore argues that we clearly do know this sentence is true and that any definition of “knowledge” which failed to reflect this is a poor definition of knowledge.

In this manner, ordinary language philosophy (at least in its early Moorean form) rejected Cartesian doubt. And yet from what I have seen, Moore remained committed to an ideal of philosophical clarity. It seems to be assumed increasing precision would generally aid accuracy, rather than involve essential trade-offs between precision and accuracy of the sort James emphasizes. I have never seen anything in Moore or Bertrand Russell about such trade-offs. Modern analytic philosophers like Lara Buchak considered this trade off more carefully.

Still, core to the approach of Moore and Russell specifically was the concept of “philosophical problems”, places where language led the user astray. The ordinary language philosopher could then lead the language user back to the truth with the tool of linguistic analysis. There is thus no need for the trade offs of belief formation under uncertainty, as the language user is assumed to be “confused” rather than, say, facing a genuine decision under uncertainty. Wittgenstein ended his Tractatus by noting that the “confusion” concept implied that philosophical problems, so conceived, are not real problems at all.

The above paragraphs admittedly paint Moore and Russell’s position with a broad brush. To put a fine point on it I will post Bertrand Russell’s commentary on the will to believe (from History Of Western Philosophy) in its entirety so that you can see how distant this school of thought was from James.

“The Will to Believe argues that we are often compelled, in practice, to take decisions where no adequate theoretical grounds for a decision exist, for even to do nothing is still a decision. Religious matters, James says, come under this head; we have, he maintains, a right to adopt a believing attitude although “our merely logical intellect may not have been coerced.” This is essentially the attitude of Rousseau’s Savoyard vicar, but James’s development is novel.

The moral duty of veracity, we are told, consists of two coequal precepts: “believe truth,” and “shun error.” The sceptic wrongly attends only to the second, and thus fails to believe various truths which a less cautious man will believe. If believing truth and avoiding error are of equal importance, I may do well, when presented with an alternative, to believe one of the possibilities at will, for then I have an even chance of believing truth, whereas I have none if I suspend judgement.

The ethic that would result if this doctrine were taken seriously is a very odd one. Suppose I meet a stranger on the train, and I ask myself: ‘Is his name Ebenezer Wilkes Smith?’. If I admit I do not know I am certainly not believing truly about his name; whereas, if I decide to believe that that is his name, there is a chance that I may be believing truly. The sceptic, says James, is afraid of being duped, and through his fear may lose important truth; ‘what proof is there,’ he adds, ‘that dupery through hope is so much worse than dupery through fear?’ It would seem to follow that, if I have been hoping for years to meet a man called Ebenezer Wilkes Smith, positive as opposed to negative veracity should prompt me to believe that this is the name of every stranger I meet, until I acquire conclusive evidence to the contrary.”

This certainly illustrates Barzun’s frustrations with finding engaging commentary on “The Will To Believe”. Russell’s example meets absolutely none of James’ axioms:

1. The belief is not live. James explicitly defines “liveness” as a relation between a hypothesis and a believer. Russell states that he thinks there is little possibility this person is named Ebenezer Wilkes Smith. A belief Russell already treats as implausible is dead on arrival.

2. The belief is not forced. Russell’s example allows suspension of judgment, many alternative names and no practical necessity to decide. James’s point is precisely that suspension is sometimes not an option, rather than suspension never being an option.

3. The decision is not momentous. Nothing of value turns on the belief that the man is named Ebenezer Wilkes Smith, no opportunity is lost by waiting and no irreversible consequence follows. James’ whole approach is about balancing the opportunity costs of different beliefs.

Thus Russell, having expounded an example carefully constructed to not exemplify James’s axiomata, notes that James’ constructions do not go through. Remember, James’ analysis is about why organisms faced with genuine options spend the energy to make difficult decisions informed by their attitude toward risk. Not that the types of risk have a trade off generically, but they have a trade off at the margin. Russell’s argument not only does not contradict James’ “The Will To Believe”; Russell’s conclusion is implied by James’ approach!

What this shows is that James is not merely a historical precursor, something that has been tried and discarded. We can recognize his ideas in modern methods like Neyman-Pearson because his thought is still lively and powerful. William James’ ethic of belief is beautiful: structured and yet pluralistic. I will go through some of the schools of ethical thought and what I believe James has affinities with and stands opposed to in each case:

1. Virtue ethics

   1. Accepts: the centrality of practical wisdom (φρόνησις) and healthy-mindedness (similar to εὐδαιμονία); that the good consists of cultivating virtues such as tolerance (ἐπιείκεια) and courage (ἀνδρεία)

   2. Rejects: the detachment of virtues from consequences

2. Deontology

   1. Accepts: the fundamental importance of deep structure (i.e. presenting genuine options) in moral decisions

   2. Rejects: the decisiveness of said deep structure

3. Utilitarianism

   1. Accepts: the evaluation of moral acts in terms of consequences

   2. Rejects: the disjunction between means and ends, the emphasis on individual acts instead of habits (virtues)

With all this, we can at last see that William James was not just a great psychologist and philosopher, he was a brilliant moralist of belief.

• Lichtenberg

Table Of Contents

1. Life Is Just One Nonlinear Problem In Random Theory After Another

   1. Introduction and meeting Herbert Simon

2. “On How to Decide What to Do”

   1. Answer to the question: What is AI?

3. “Trial and Error Search in Solving Difficult Problems”

   1. What was and what is the relation between AI research and consciousness?

PAYWALL

4. An Excerpt From Sciences of the Artificial

   1. Generative Artificial Intelligence as Variety Emitter

5. “Rationality as Process and Product of Thought”

   1. Information and Attention as the scarce resources in AI workflows

6. The Method of Distinguishing the Real from the Imaginary

   1. What has happened in the fields AI has already revolutionized and what is wrong with AI hype

Subscribe now

Life Is Just One Nonlinear Problem In Random Theory After Another

“Il moderno principe, il mito-principe non può essere una persona reale, un individuo concreto, può essere solo un organismo; un elemento di società complesso nel quale già abbia inizio il concretarsi di una volontà collettiva riconosciuta e affermatasi parzialmente nell’azione.”

• Antonio Gramsci, “Notrelle sulla politica del Machiavelli”

We are currently in an AI boom. There are estimates that world capital outlays for computational power in this boom period will exceed $7T by 2030. The high power demands and high power draw volatility of these AI tasks are already shaping the near future of energy investment. Meanwhile, AI related capital cycles rapidly as new chips, new algorithms and new data sources appear. The capital structure changes and the related social impacts are the truly interesting story, but they are not the story I will tell you today - because CVAR Founder and co-editor Alex Williams is doing great work on that front. What we are here to do today is get a grasp on a definition of what AI is and why it can have a specific boom.

By “AI” I don’t mean to limit myself to Generative Artificial Intelligence (GAI). Because of this broad focus, I may use slightly different language than you are used to. Like Polya said: that an important step to solving problems is to attack more difficult and general problems! I will try to stay away from the limiting “AI =GAI” mindset. In order to accomplish this, I will go long before the beginning of the GAI boom to the work of Herbert Simon.

In the old conception of labor as a deliberative change in the state of the world, instruments of labor like AI are of particular interest because they lie on the spectrum close to deliberation rather than close to nature. Thus the end user, the laborer, can benefit from having a strong conception of AI. This is unlike, for example, a lawnmower. A deeper knowledge of the lawnmowing system probably does not make a laborer more effective and transforming the world to a place with more uniform grass heights. This is not to say landscaping doesn’t require thought, but rather that to the extent that the thought is about the lawnmowing system it probably is “lawnmower repair” rather than landscaping proper.

Let’s delimit the arena of Platonic space before we get to a definition. Artificial Intelligence (AI) tools writ broadly are products designed to help make decisions. Thus AI tools are in the same sport as spreadsheets and ouija boards but play different positions, like Junghoo Hu and James Watt. AI tools help us solve what Norbert Weiner called “nonlinear problems in random theory”, expanding the realm of possible actions. An AI tool is being used correctly in the event that it expands its user’s administrative capacity - their ability to “effectively implement policies, manage resources, and deliver services.”. We have previously discussed administrative capacity. The linkages between AI and that discussion will be returned to below.

I want to respect your time as a reader, so let me be as clear about what this post is not as what it is. This post is not a comment on whether any or all AI companies are in a “bubble”. However, I want this post to help you manage AI use, and the AI Bubble topic can derail administrative discussions. To begin with: a market bubble is not “overvaluation” by itself but a complex triadic relationship between prices, values and both credit and other market ratios. Understanding this triad will aid in making strategic decisions. Breaking out that triad out one by one:

1. A price is a ratio at which an actor is willing to part with one good or service for certain quantities of a possibly distinct good or service.

2. Broadly, “values” in the economic sense are attempts to understand price abstracted from peculiar market conditions. There is not one precise meaning to “value”.

   1. For concrete examples of “values”, Substack piece gives two subtly different definitions from Benjamin Graham and Warren Buffett. These value investors base their purchases of stock, bonds etc. on when the market prices and value estimates diverge, making careful definitions important to them.

3. I will give two traditional metrics for assessing different notions of bubbles, one for overvaluation risk and one for liquidity risk.

   1. A high price/earnings to growth (PEG) ratio is sometimes interpreted as an indicator of overvaluation. While not sufficient by itself, I will use it as a back of the envelope indicator.

   2. A conservative credit ratio for liquidity risk is the level of cash held to the value of liabilities (the cash ratio). A lower cash ratio means more liquidity risk.

Loose discussions of bubbles run distinct things like overvaluation and liquidity risks together, but they can actually anti-correlate with certain business strategies. This may seem like a lot of abstraction to take in at once, but as I said: a market bubble is a complex phenomenon.

To illustrate how these complexities work in practice, take controversial AI company Palantir. Palantir’s PEG ratio has been consistently above 2 for more than a year (see current ratios here). Meanwhile, Palantir’s cash ratio is currently more than 5 and has been more than 3 for several years (see here). Both of these are related to Palantir’s aggressive stock based compensation strategy. What this shows is that the question of a “market bubble” will bring us into dense balance sheet questions unrelated to the use of AI.

Given all this, the only responsible, high-level conclusions I can draw are:

1. AI probably has positive long run value, as shown by the level of investment by highly diverse actors.

2. Stock prices of some AI service providers seem high by some traditional metrics. If these metrics indicate market realities, this would make those companies “overvalued”.

3. This is not sufficient evidence for or against a “bubble” in some senses of the word.

This is the limit of what these simple indicators can tell us.

Market matters aside, let’s get back to the main topic. The goal of this piece is to help you use AI tools to expand your administrative capacity better by explaining some of the basic concepts of A. The method of this piece is to use the work of the father of AI, Herbert Simon as guide to conceptual foundations.

Herbert Simon was a fascinating individual. I strongly recommend his autobiography Models Of My Life. Simon recounts formative experiences which led him to realize the difference between his perceptions and the world. His family went strawberry picking. Simons recalled crying because the other children got more strawberries than himself. His family had to explain that the strawberries had a mysterious invisible property - “being red” - that led the other boys to be able to distinguish between strawberry and leaf. That experience of the interplay of hidden information and decision making would affect much of the career.

After achieving majority, Herbert Simon enrolled in the University Of Chicago, drawn to higher education by the Ely-Commons tradition of social science. He remembered Nicholas Rashevsky, Henry Schultz and Rudolf Carnap as particular mentors. From Rashevsky he learned the power of boldly simple models of complex phenomena - and the danger of a cavalier approach to data. From Schultz he learned the difference between a good fit and (what Deborah Mayo calls) a severe test, the core of the Neyman-Pearson philosophy of statistics - in layman’s terms, an idea that probably works vs an idea that’s probably correct (a small subset of what works!). From Carnap, he learned the ideal of logical positivism or logical empiricism - a scientific language where not just some of the important concepts but all content words are defined operationally.

With this educational background, Herbert Simon would become one of the most important computer scientists of all time - of course, his degree was in public administration. How did his horizons expand into the nascent field of computer science? The dark star that guided Simon through his transit towards digital Bethlehem was the attempt “to explain how organizations can expand human rationality, a view quite opposed to popular folklore in our society, which commonly sees them as dehumanizing bureaucracies.” (Simon, Models Of My Life). This optimistic view inspired my use of a quotation from Gramsci at the head of this section: the “modern prince” is more than any one individual could be. In plain terms, effective organizational leadership is when that organization can do more than any sub-coalition of the organization.

So how did this public administration theorist become the father of AI? Here is how the man himself put it: “This sudden and permanent change came about because Al Newell, Cliff Shaw, and I caught a glimpse of a revolutionary use for the electronic computers that were just then making their first public appearance. We seized the opportunity we saw to use the computer as a general processor for symbols (hence for thoughts) rather than just a speedy engine for arithmetic. By the end of 1955 we had invented list-processing languages for programming computers and had used them to create the Logic Theorist, the first computer program that solved non-numerical problems by selective search. It is for these two achievements that we are commonly adjudged to be the parents of artificial intelligence.”. Al Newell had taken classes with George Polya, who dubbed the new science of the psychology of discovery ‘heuristics’. Cliff Shaw provided programming experience. Together, the three men endeavored to make a machine which could, among other things, prove 1+1=2.

As we will see, some of Herbert Simon’s original AI tools are, of course, primitive. His theories evolved along with the tools, starting at points. But one must remember that he was fighting with both arms engaged. His right arm was tied back by Father Time. The rope Father Time used was the ludicrously primitive non-AI aspects of the digital tools available. Newell, Simon and Shaw didn’t store their work as files in folders because the metaphor of the central linked list in a shared computer system as a directory of files is one that had to be created. His other arm was engaged by the practice of empirical psychology. Newell, Simon and Shaw didn’t just want the best AI tool for the job, they wanted a model of how a human did the job! I will discuss this more later.

Even with both arms engaged in work we no longer consider AI, Herbert Simon and his students spent decades refining the theories and developing tools for AI. They kept their collected noses to the grindstone, especially with respect to basic concepts. Thus his definitions and analysis - while definitely idiosyncratic - are in some ways more clear than some modern commentators.

With his unique background and decades of empirical and conceptual work, Herbert Simon really did understand something about modern AI that many today miss. This post uses Simon’s ideas as a jumping off point to cut through the noise and explain what AI is and can do. Before the paywall, I unpack Herbert Simon’s analysis of the meaning of AI, drawing on “On How To Decide What To Do”. This locates AI within the space of decision making tools. Then, as a treat, I use “Trial and Error Search in Solving Difficult Problems” to put strongly the distinction between AI and consciousness research. This closes one source of AI hype. Beyond the paywall, take the fundamental cybernetic concepts of ‘variety’ ‘information’ to map the real decision making niche of GAI. It turns out to be perfectly described by a comment from Herbert Simon in Sciences of the Artificial! Next, I draw on sharp passages from “Rationality As Process And Product Of Thought” to derive necessary conditions for the efficient use of AI.

Finally, the post ends by putting Simon in conversation with other thinkers to provide clear thinking about AI’s future. Frank Knight’s economic analysis of production proves to be an excellent skewer for AI Hype. For the last button on this long article, I combine Simon’s definition with powerful arguments from philosopher Karl Popper to bury AI hype once and for all.

Subscribe now

On How To Decide What To Do

“Proper evaluations of words and letters in their phonetic and associated sense can bring the people of earth to the clear light of pure cosmic wisdom.”

• Sun Ra, Cosmic Tones for Mental Therapy

In this paper, Herbert Simon gives a definition of Artificial Intelligence (AI). Despite being written in 1978, this is the result of nearly 20 years of Simon’s AI research. As such you will find a meaning refined by many years of use and research, and thus something that cuts through popular misconceptions. A true proper evaluation of words, as Sun Ra said.

Specifically, Simon locates Artificial Intelligence (AI) proper as one extreme of a spectrum of decision making techniques which emphasize optimizing on the one end and satisficing on the other. AI is on the satificing end, which makes sense because the notion of satificing was fundamental to his scientific life. Simon labels the other extreme Operations Research (OR).

Understanding the optimizing end of the spectrum makes Simon’s placement of AI much clearer. In OR, the interest is in the existence, uniqueness and computation of an optimum - that is, a special “best point” within the context of the model. All other facts about the model are relative to the interest in that optimum.

A basic example of an OR type approach is the analysis of Jan-Ken-Pon in game theory. This will show in a simple visual way how OR can value an exact optimum within a model at the expense of real life complexities.

The first step in any analysis is to discover the space of possibilities. A “strategy” is a triple of probabilities (p,q,r) such that p>0 is the probability of throwing paper, q>0 the probability of throwing rock and r >0 the probability of throwing scissors and p+q+r=1. A strategy is called “pure” if and only if one of p, q or r is equal to one. Otherwise the strategy is “mixed”. By Viviani’s theorem, each strategy can be identified with a color in the above rainbow colored simplex. The bottom left corner, in violet, represents the strategy of always throwing rock. The orange corner represents a strategy of pure paper. The spectrum between meanwhile corresponds to a certain probability of throwing rock and paper otherwise. The centroid of the triangle is a sort of grey cyan.

What does the OR type approach tell us? The strategies of special interest are the strategies which are the best response to some strategy. There are thus only four points of special interest, the three pure strategies and the centroid. If a strategy is off the centroid, then the best response is the pure strategy that beats the strategy they prefer. Take some examples:

1. (Orange corner) If they throw paper more often than the other two, then throw scissors.

2. (Indigo, bottom center) If they play paper and scissors in equal proportion and rock rarely, then play scissors.

3. (Grey cyan centroid) If they play paper, rock and scissors in equal proportion, then do the same.

The centroid is the unique point which is strategy deviation stable - given that both players are playing centroid neither wants to change their strategy. It is the so-called “Nash Equilibrium”. Does this closely resemble reality? At high levels, Jan Ken Pon is about searching for tells in the opponents’ throw. This part of the Jan Ken Pon is not in this model. They could be incorporated into a higher dimensional model which would have its own set of best responses and Nash Equilibria. This is what I mean by “optimum … within the context of the model”.

The benefits of OR Modeling are obvious. It allowed us to go from a two dimensional continuum of possibilities to four points of special interest, with one point most special among them (the Nash Equilibrium). But the costs are intense. The set at which the best response is the “stable” Nash equilibrium is zero-dimensional. If your opponent has a tendency away from the Nash equilibrium no matter how slight, then you should be playing a pure strategy. The best response is not continuous around the most interesting point! This is exactly the brittleness that Simon argued makes pure optimization unsuitable as a model of intelligence.

So much for the OR approach for the problem. In AI, the interest is moving through the range of a model in a positive direction until one finds an output which works at least to an acceptable extent, what Simon called “satisficing”. That is to say, an AI method is when an outcome which may be of interest for reasons other than having a special claim on being optimum within the model. Good outcomes are not defined relative to that supposed “best” even within the AI’s model of what the world is like.

Coming back to Jan Ken Pon, an AI approach might be to have a computer go through hours and hours of footage of paper rock scissors games until an underlying neural network converges on throwing paper when the hand appears to be throwing rock etc.. The state of the neural network has no relation to the ideal function classifying the real hand (a purely platonic object). The state is judged based on the appearance of convergence when fed more data and success at Jan Ken Pon.

The benefits of AI modeling are obvious. Because we give up the obsession with optima, we can have models of enormous complexity - equal to the task at hand. We can get model outputs better behaved over model inputs. But the costs are intense. The model is now a complex and mysterious thing. Output can depend sensitively on things which are objectively ridiculous. In Jan Ken Pon, maybe due to an accident in the training data, the neural network cares intensely whether the thrower has a freckle over the pisiform bone in the wrist. We are back at the old problem of what works (in the training data) vs what’s correct(in reality).

Simon’s definition of AI is more than just clear - it is an operational definition which allows us to evaluate specific tools. In the next section, I will leverage that operationality to make distinctions between AI research and other branches of the science of decision making, such as consciousness research. Later, we will soon see how it is strong enough to help users of AI (including GAI) develop better processes for production. Finally at the end of this paper, we will see how Simon’s definition, with a bit of help from Poppernian logic, deflates AI hype.

Subscribe now

Trial and Error Search in Solving Difficult Problems

“10. Never hesitate to make a move for fear of losing. Whenever you think a move is good, go ahead and make it. Experience is the best teacher. Bear in mind that you may learn much more from a game you lose than from a game you win. You will have to lose hundreds of games before becoming a good player.”

• José Raúl Capablanca, A Primer of Chess

In the previous section, we saw how Simon’s definition of AI located it within the space of decision making tools. In this section, we go over where AI is located in the space of scientific theories of the mind. This analysis will use the operational nature of Simon’s definition extensively.

Now, Intelligence (artificial or otherwise) by Simon’s definition has only vague and complex relations with Consciousness. Decision making involves conscious and unconscious processes. Consciousness, whatever else it is, is a property that some humans have intermittently, and no one cognitive ability fully defines that property. Likely consciousness is a complex structure of abilities - a “bag of tricks” in Dennett’s phrase. AI tools can have some - perhaps even all! - of these without being conscious.

Despite not equating intelligence and consciousness, Simon’s still hoped that a biomimetic approach would inform successful AI tools, such as automated chess players. See, for example, “Trial and Error Search in Solving Difficult Problems”. In this paper, Simon used a satisficing approach - a now primitive seeming search + heuristics procedure - to analyze chess. He used experimental psychology to try to measure how top chess players reasoned about chess and then make a computer model of human chess playing - artificial intelligence. He included psychological analyses of chess grandmasters to tune his search depth and inspire heuristics.

It’s true that modern chess engines do not use this psychologically grounded approach. But recall that Simon’s definition of AI is a spectrum. On that spectrum modern chess engines are even more AI in Simon’s definition than his were! AI chess systems such as Stockfish use massive neural networks to build a completely opaque model of chess and then satificies itself within that model. One can even tune the degree of satificing! No Stockfish model has no known relation to the ideal chess algorithm, a computationally intractable Platonic object about which little is known by non-constructive existence proofs (Zermelo’s theorem). Rather, Stockfish has taken Capablanca’s advice above into the 21st century by losing billions of games against the best chess player in the world (itself).

The point here is that most modern AI research is not even trying to make conscious tools. Successful modern AI researchers’ goals have not been to make a tool that reasons like a human per se, not in a biomimetic manner. Rather, the main goals have been to develop innovative tools to solve recognizable problems.

Many great minds have attempted to put necessary conditions on what consciousness is. For example, Kant argued that the ability to reason spatially is necessary for consciousness. But humans reason spatially in dreams, showing that this is not sufficient. It’s clear that AI is better now than ever before at performing the tasks these great minds have argued necessary for consciousness. For example, AI can now reason about space well enough that it can create simulacra of objects moving through space (i.e. “animate”). But it would be a mistake (namely, “affirming the consequent”) to think that because AI is getting better at tasks, it is becoming more conscious. This is a confusion of necessary and sufficient conditions.

There is an interesting result of this. With Simon’s definition of AI, Artificial General Intelligence (AGI) can be defined as a system which can output a satisficing model of your input, then find a satisficing response within that model - an AI which outputs AI. There is no need for an AGI to be more conscious than a sleeping man or a large stone. In that peculiarly deflating sense, we already have AGI … but are not even approaching consciousness.

We’ve now seen that Simon’s unusual but deeply clarifying definition of AI doesn’t just tidy up some old debates. His analyses actually sharpen how we think about intelligence, consciousness and modern systems far beyond those he had access to. Now comes a change in tactic to more specifics. Below the paywall, I’ll take Simon’s insights and turn them into practical guidance for GAI use. I’ll start by giving a clean, usable definition of GAI for real-world work. The result of this analysis is put perfectly by a comment of Herbert Simon in Sciences of the Artificial. From there, I’ll draw on Simon’s pioneering work in the economics of information and attention to outline a necessary condition for efficient AI use. Finally, I’ll draw on broader traditions to show why some of today’s loudest “AI hype” claims fall apart. If you want a rational framework for thinking about AI in the present and future - or just a way to separate the useful from the absurd - these next sections will help you.

Read more

Pettis & Klein’s argument begins, as all economic arguments do, with Smith’s pin factory. Smith imagines, or rather read about in a French encyclopedia, a pin factory whose efficiency (output per unit cost - at the time and for all practical purposes, per laborer) rose as more workers were employed. Smith argues that this is due to the “division of labor”. That is, the pin factory, far from being like many pin makers in parallel, is a complex structure of well defined firm-like “branches” in series. Each branch, like a firm, has a well defined input, a well defined output, well defined income (the portion of firm costs, largely wages, that go to said branch), well defined costs (the portion of firm costs, largely wages, that go to the inputs of the branch) and thus well defined profits.

Smith’s theory of the firm as branches in series, at first sight, seems to overcomplicate economics into an infinite theory of massively multilateral semi-monopolies - one for each branch. There are thus at least 18 semi-monopolies in a single pin factory! But Smith has a key: the division of labor is limited by the extent of the market. On the top end, a branch of the firm which takes in cut and pointed proto-pins and outputs proto-pins ground on the top for receiving the head is a semi-monopoly (no other branch in the firm produces ground & pointed proto-pins). But the semi-monopoly branch is only free to act insofar as the upstream branches do not prefer to take their business to outside firms. On the bottom end, just as a firm has certain fixed costs which must be cleared if the firm is to exist, so a branch cannot exist unless its fixed costs can be taken care of.

The superiority, in terms of efficiency, of employing a population in complex branches in series instead of individuals in parallel is Smith’s “system of natural liberty”. The efficiency gains allowed by natural liberty are the manifestation of natural liberty’s superiority over systems of restraint and regulation which limit the extent of the market. Smith’s vision was very beautiful. But, on account of its beauty, the vision is also quite incomplete. Some have found quite little to praise about the process in practice.

Keynes took the vital step beyond this harmonious vision. To do this, Keynes drew on his intellectual background as a logician. Though Keynes was no anti-capitalist, the way he laid bare the logic of markets shocked especially capitalism’s most ardent defenders. For example, common sense and Benjamin Franklin tell us that man is the tool making animal and should be valued as such. But Keynes demonstrated clearly that the logic of the market values man only as a consumer. As a producer, a man is just an unfortunate cost.

For Pettis & Klein, the vital part of Keynes’ logic is his rationalization of income accounting. Keynes’ clearing up of income accounting follows the principle of quantifying the quantifiable. Keynes’ notion of income accounting gives a quantitative description of the current state of the world, measures of the choices of the complex mass of actors. In order that these measures should range over the quantifiable, all aggregate accounts must be functions of the actual and current behaviors of economic actors, i.e. entrepreneurs, households etc.. The measures can not depend on, for instance, what they potentially could do tomorrow.

Look for example at Keynes definition of income: total sales reduced by factor and user cost. Factor cost is those firm’s purchases from other firms and user cost is the loss of value involved in business practice. Income thus is a function of current household decisions (sales) and current entrepreneurial decisions (costs). Prior accounts of income (in Keynes, net income) amounted to total sales reduced by factor, user and supplementary costs. Supplementary costs are those costs which are involuntary but expected. Those costs which are involuntary and unexpected are called “windfall loss” (or gain). Thus, the level of income (in Keynes, net income) depended on how the entrepreneur wishes to constrain his future actions. Wishes are not necessarily quantifiable. Keynes even shows that many of Hayek’s sermons on unmeasurability & aggregability in economics are just a consequence of choosing an income concept that both cannot be measured (and is thus not aggregatable) and is not a function of the choices of economic actors. Excluding supplementary costs with their mysterious border on the land of windfall loss was one step of how Keynes was able to rationalize the world of income accounting.

The particular application of Keynes’ rationalization of income accounting that is core to Pettis & Klein is his analysis of consumption, investment and savings. Consumption is total sales to households, or total sales minus sales from entrepreneur to entrepreneur. Investment is output purchased by entrepreneurs less user cost. Total savings is the excess of total sales over total consumption, which is exactly total output purchased by entrepreneurs less user cost or “investment” for short.

A dialectical way to see the appropriateness of these definitions is to start with the “old view” that an act of saving simply is an act of investment. The only way for a winemaker to fund a film tomorrow is to plant seeds today. Outside of a hypothetical rude state of society, this is incomplete and misleading as an entrepreneur can save by inducing another entrepreneur or household to dissave. The winemaker can also “save” by purchasing a planted vineyard, thus having sales in 12 hours rather than tomorrow. But because accounts are chosen in a manner which is numerically unambiguous and thus aggregatable, all methods of saving by inducing dissaving necessarily cancel out. The result is: aggregate savings level equals the aggregate investment automatically, with all concerns about “genuine” vs “forced” savings taken care of.

Though the level of savings in a time period is always numerically equal to the level of investment in that time period, it is not the case that savings and investment are interchangeable concepts. The old Knightian insight of the “wheel of wealth” taught us that investment and consumption are the active parts of the economy, savings mere residual. True we talk of “saving money” rather than “investing in our checking account” but we also talk of “sun down” rather than “earth turn”.

The consequence of income accounting fundamental to Pettis & Klein is:

domestic savings - domestic investment = net exports

Again, the wheel of wealth insight tells us that investment is the active part of the economy. On the left hand side of the equation, we have one active variable and one passive. The result is the right hand side is, in the main, a function of the one active variable. In economic language, net exports are essentially a function of domestic investment. A debtor country is a country with a high level of domestic investment and nothing else. A lender country is a country with a low level of domestic investment and nothing else.

Finally, the level of domestic investment of an advanced state is, as Abba Lerner painstakingly demonstrated, an essentially political variable. Political variables largely follow class interests. Thus current net exports largely follow class interests.

A subtle distinction is important here. Remember that “savings” and “investment” and all other accounts, unless otherwise stated, are within period savings and investment. Nothing is being held constant, nor are we making dynamic assumptions - we are only discussing within one period. The income accounting and the wheel of wealth relations are functions, not fluents. It might be the case that “yesterday’s investments are today’s savings” or somesuch rot, but that is a dynamic hypothesis not an accounting identity. A debtor country is one with a high level of domestic investment today. The evolution of net exports - unlike its current level - is influenced by many factors other than class interest, especially technological change. The distinction between the current level and evolution of net exports is frequently invoked by Pettis & Klein.

Even with that caveat, the “net exports = f(class interests)” insight can be used to answer questions. Why does big agriculture support Trump when his policies are devastating to their industry (reducing their access to labor and shrinking their market)? They are aware that Trump will simply undo this devastation by increasing their subsidies (as he did last time) as he is the leader of the right wing party that serves their class interests. Thus we see that net agricultural exports are simply a political variable. One can call these interests foolish and benighted, but that is just a roundabout way of calling them class interests rather than universal interests.

A distant but still memorable example may help. In the early 90s, economist Paul Krugman was asked by a trade official what the effect of NAFTA would be on US jobs. Krugman dutifully answered NAFTA would make no difference, as the above simple arguments show that net exports are simply a function of the level of domestic investment landed on by the political process. NAFTA would have political consequences in Mexico. NAFTA protected foreign investment in Mexico, reducing the reliance on domestic investment and driving up net exports. But its consequences in the US were basically the decision of Gingrich and Clinton. Krugman kept talking about finger accounting exercises and his interviewer pressed on the economic, political and philosophical consequences of a massive bilateral trade agreement. Eventually, the trade official exploded,

“It’s remarks like that which explain why people hate economists!”.

Here is where the logic of capitalism as exposed by Keynes interferes with the development of the “system of natural liberty” advocated by Smith. To recapitulate, Smith’s system achieved efficiency by organizing production within branches. As expounded in Marshall’s beautiful chapters on industrial organization show, this system is conducive to growth and class harmony. To quote

“When an industry has thus chosen a locality for itself, it is likely to stay there long: so great are the advantages which people following the same skilled trade get from near neighbourhood to one another. The mysteries of the trade become no mysteries; but are as it were in the air, and children learn many of them unconsciously. Good work is rightly appreciated, inventions and improvements in machinery, in processes and the general organization of the business have their merits promptly discussed: if one man starts a new idea, it is taken up by others and combined with suggestions of their own; and thus it becomes the source of further new ideas.

…

[A] localized industry gains a great advantage from the fact that it offers a constant market for skill. Employers are apt to resort to any place where they are likely to find a good choice of workers with the special skill which they require; while men seeking employment naturally go to places where there are many employers who need such skill as theirs and where therefore it is likely to find a good market. … Social forces here co-operate with economic: there are often strong friendships between employers and employed: but neither side likes to feel that in case of any disagreeable incident happening between them, they must go on rubbing against one another: both sides like to be able easily to break off old associations should they become irksome.”

Harmony, however, is but one possible end point of the system of natural liberty, the other being disharmony. A firm that produces pins can choose to have its pointed pins to be ground for application of the head in house as Smith assumes, or the same firm could purchase semi-pins pointed at one end and ground on the other to apply the head from another firm - even another firm across a national border. In market equilibrium, the costs to the firm would be the same. In income accounting terms, the firm is indifferent between a user cost (producing in house) and a factor cost (purchasing from another entrepreneur). Recall the Keynesian insight that man as a producer is an unfortunate cost and only man as a consumer is valued. The “social forces” which Marshall discussed are sometimes called “implicit contracts”. There is great value to be had in breaking these implicit contracts, which is done by separating ownership and management perhaps by making sure they are in different countries speaking different languages. Smith’s pin making firm was made up of 18 multilateral semi-monopolies which developed to a game theoretical equilibrium with maximal output given the state of technology, largely because they shared the same profit. Other things equal, maximizing the pie to be sliced was everyone’s interest.

Pettis & Klein ask us - is it plausible that 18 firms in 18 countries under at least 18 legal structures with 18 different profits will land on maximal output given the state of technology? South Korea is a famous hub for condensed matter research, China a famous producer of microchips - an economic output dependent on condensed matter research. Even if the market relationships between Chinese and Korean firms are not minimizing efficiency, one cannot say that Korean condensed matter physics lore is “in the air” in China. In market equilibrium, Chinese and Korean firms will be allocatively efficient but may still be x-inefficient.

There is an obvious solution to global x-inefficiency caused by too small industrial districts that are the result of production being spread across industrial districts globally. That solution is: each country domestically invests in its industrial districts, bringing balance to net exports. Balance in domestic investment is actually seen in some countries, Mexico for example, as shown by their switching between positive and negative net exports. But other countries, such as Germany and China consistently have low domestic investment (positive net exports) and the US has had high domestic investment (negative net exports) since 1975.

Now comes the problem that Pettis & Klein confront. Surely class interests are determined by … the class. Why is Germany and China’s ruling upper class so dramatically different from the US’s upper class? Why is Mexico’s upper class served by balanced domestic investment?

Here Pettis & Klein must move from economic laws to economic history. Class interests are shaped not merely by class but by class and history. Further states have a degree of autonomy distinct from the class interests they serve deriving from their limited administrative capacity. A state may serve a class but they can only serve that class insofar as that state can do something possible for a given state in a given situation.

Let’s now summarize the relevant economic history. In 1949, China and Germany were ravaged by war. Their leadership had a mandate from the class they represented to restore equilibrium to the country, i.e. to have high levels of domestic investment. Necessarily, they ran trade deficits during this time. However, as they progressed from low income to middle income states, their upper classes became exposed to volatility. Smith’s insight that the division of labor is limited by the extent of the market worked against them. The branches of its firms must have costs and profits higher than a certain minimum and lower than the global market maximum. The second of these is beyond the capacity of a small state to manage. Over time, the states of Germany & China hit upon the “solution” of guarding class interests within its administrative capacity by moving away from domestic investment, as demonstrated by persistent positive net exports.

Meanwhile, the US’ leadership had a mandate from its ruling class to invest abroad - especially after 1953. This was a confused and confusing policy that balanced Rooseveltian anti-imperialism with fierce anti-communism. During this period, the US necessarily ran a trade surplus. As anti-communism triumphed over anti-imperialism, the US began a policy of courting the upper classes of middle and lower income countries by protecting their investments, i.e. allowing them into the US. As we have seen, in Mexico this led to a modest decline in domestic investment (increase in net exports). But the history of the US has endowed it with a unique institutional feature which causes a paradox. The US state wholly owns a bank called the Federal Reserve. The main products of this bank, US Bonds and US dollars, are the most liquid assets on the market at any given time. Therefore, allowing foreign investors to buy physical assets in the US means that US investors are freed from the burden of having to own physical assets (‘fear of goods’) and flee to liquid financial assets (‘love of money’).

The historical story of the previous paragraphs completes the logical survey of Pettis & Klein’s theory. Pettis & Klein’s theory thus combines the basic insights of industrial organization, income accounting, historical & political development and unique features of the global financial system to explain the persistent features of net exports across countries - positive for China and Germany, negative for the US and zero for Mexico. But there must be a conclusion, what is to be done with these insights?

There are some simple and politically possible “solutions”, such as tariffs and beggar thy neighbor “competitiveness” drives. But these mistake the residual (net exports) for the driver (domestic investment), the symptom for the disease. Indeed, they will only serve to increase x-inefficiency by blocking the division of labor in the first place (tariffs) and breaking the implicit contracts that serve much of the value of localized industry (“competitiveness” drives).

Pettis & Klein’s own solution is “to end the trade war, end the class war”. That is to say,

1. Make it possible for countries that consistently underinvest domestically (China & Germany) to grow the size of their industrial districts to an appropriate level and

2. Definancialize the unique US economy.

Pettis & Klein make specific proposals towards these ends. Item 1 is to be done by allowing Chinese consumption to grow, which validates Chinese domestic investment and puts them on a virtuous cycle back to zero average net exports. For item 2, the US government is to prevent foreign investment from competing with domestic investment into private physical capital by absorbing such investment directly and applying the money absorbed to public physical capital. This would redirect private investment from the financial sector to the private physical capital sector, turning the US into a “normal country” whose upper class interest is served by non-negative net exports.

This has aimed to be a neutral description of Pettis & Klein’s Trade Wars Are Class Wars. As mentioned in the opening, to this we were to have comments from expert contributors to Continuous Variation which describe their reactions, intellectual or otherwise and positive or otherwise, to the theory neutrally described here. Feel free to fill in that gap in the comments below!

• John Maynard Keynes, “Alfred Marshall”

I used to watch a show called Ramsey’s Kitchen Nightmares. In this show, super chef Gordon Ramsey would consult with failing restaurants. Often he would teach basic management lessons, such as demand planning and the importance of courting new customers. One restaurateur was washing her own dishes rather than hire a dishwasher, not realizing as a manager she needed to target costs rather than cash outflow. This story will come back later.

My favorite episode was about a woman who was an extremely talented chef occupying the safe niche of producing authentic American soul food in London. She was told by the bank to cut costs and raise prices to increase profitability in return for a loan. Ramsey advised her to ignore the bank, because they are a bunch of assholes who don’t know shit. And it’s true, no better sign of poor management than leadership cutting costs with no specific vision.

I like to write about management and the economy rather than “economics”. But economics has a real perspective on the simple management issues raised by Mr Ramsey. The perspective is outlined in abstract by the Keynes quote above. This perspective is particularly important when times are hard: depressions, recessions and panics.

When the price of strawberries goes up, the quantity of strawberries purchased will go down. This is obvious whether you are Joe Schmo on the street or berry economist Benjamin Graves. But Graves has a small advantage of Mr Schmo. Schmo believes the price of strawberries goes up then people will buy less strawberries because purchasers will be less able to afford them. He might even think this is a dull truism.

But economic theory largely denies this analysis. The quantity demand effect of a changing price on the purchasing power of a firm is the so-called “income effect”. The income effect is variable. Large direct purchasers of strawberries include retailing firm Wal Mart and agricultural cooperative Ocean Spray. In terms of income effect, these firms could have different reactions to increasing strawberry prices. Wal Mart might see the income effect of an increasing price of strawberries as a reason to cut back on other berry purchases and consolidate in the high profit strawberry market (other considerations could of course overcome this). Ocean Spray however would almost certainly see rising strawberry costs as a reason to consolidate back to their cranberry market home even if just considering the income effect.

Okay, that was a bit much. But in this case we actually can do more than say “It’s complicated.”! The genius of Eugenio Slutsky (not to be confused with CVAR hero Eugenio Colorni) was to show that all changes in quantity demanded can be partitioned into two linear parts: a substitution effect and the above mentioned income effect. The substitution effect says that if the price of strawberries goes up relative to substitute goods, then the purchaser (at a fixed level of wealth) will derive their fruit service from the substitutes to the extent possible. That extent possible is the so-called margin. Thus we finally have all the pieces that Keynes gave us in the paragraph above.

Insofar as they are an instructor of man, the economic theorist must in the first place believe that their students are all but entirely unaware of substitution effects and must be pulled into the light by way of diagram, example and algebra. When in the pulpit, the economist preaches that the economy consists entirely of needed services which can be fulfilled by infinitely various physical means. Thus, the sermon continues, the margin is in the first place psychological: the ‘choice’ (in scare quotes) to make one substitution (strawberries for raspberries) and not another (strawberries for lingonberries). The costs & prices that enter into substitution decisions have money *units* but are not generally a publicly quoted price on a formal market. Going back to the first example, the poor manager that Mr Ramsey was consulting washed her own firm’s dishes because she was making decisions based solely on cash flow rather than on costs. Thus her firm was less profitable than it could have been.

So two cheers for the margin and substitution. But Keynes reminds us that the economic theorist cannot remain solely with the psychological margin. They must build the notion out into a complex system of relations, relations that do intersect cash flow, material goods and other aspects of the quantifiable world. The margin is only expressed by actual substitution, which is not psychological but a change in the state of the actual world. To try to remain with solely psychological margins is like remaining with the question “Where is the sun?” unable to build up to a question with a more definite answer such as “What is the terracentric position of the sun?”. These relations, though shallower than the cost insight, give the actionable content of any economic theory and thus it is by these relations that a theory is judged, not the theory’s depth:

“According to Malthus’s good common-sense notion prices and profits are primarily determined by something which he described, though none too clearly, as ‘effective demand.’ Ricardo favoured a much more rigid approach, went behind ‘effective demand’ to the underlying conditions of money on the one hand and real costs and the real division of the product on the other hand, conceived these fundamental factors as automatically working themselves out in a unique and unequivocal way, and looked on Malthus’s method as very superficial. But Ricardo, in the course of simplifying the many successive stages of his highly abstract argument, departed, necessarily and more than he himself was aware, away from the actual facts; whereas Malthus, by taking up the tale much nearer its conclusion, had a firmer hold on what may be expected to happen in the real world.”

• John Maynard Keynes, “Thomas Robert Malthus”

Yes, often a shallow but stable function serves science more deeply than a tangled web of causes. Causal analysis and depth is good, but comes at certain expenses. A well known economic example is found in the marginal rate of transformation faced by the producers of substitute goods, say frozen concentrated orange juice and fresh orange juice. In a so-called “competitive market” (i.e. where purchasers are indifferent between different producers of orange juice including between in house and out of house production), the marginal rate of transformation between fresh and frozen concentrated orange juice at the currently installed capital base is equal for each producer and equal to the relative price. But this rate does not cause the relative price. Rather, the price is determined by whoever happens to be the marginal purchaser on the order book right now.

In fact, the installed base of capital is itself merely a way of going long the goods & services said capital can supply when mixed with labor. The service of that capital is a substitute for writing a future contract with different risks. If the entrepreneurial opinion shifts, then they will transform that long position to a short. In the case of going long by purchasing capital, they will do so by selling off capital equipment for money or “liquidating”. Often highly specialized equipment will have low or no value on the open market. This is one of the risks involved in investing in plant.

A good real life example happened recently in Texas. The Texas Legislature offered $5B loans at low interest rates to build a gas power plant. The Texas legislature favors gas due to them being rats. They didn’t find any takers. At the end of the day all you get out of taking the loan is a gas power plant. Who wants all that trouble? Capital good risk and uncertainty is basic to the megaproject approach. In fact similar, though more extreme, considerations stymie nuclear plants. No nuclear plant has been delivered on time or on schedule in the known history of the world. This problem is getting worse in the US and Europe. Contrariwise, you can build out wind/solar and battery parks with off the shelf parts. Thus liberalized markets like Texas vastly prefer to build wind/solar and battery parks. I don’t report this as a statement of preferences about energy production but as a set of well known facts.

Once again, we must leave the economy for a bit and come back to “economics”. You see, the pull of the subjective world of profit is exactly that it is an accounting variable: subjective incomes “add” and subjective costs “subtract”. One could, in principle, discover exactly the costs a restaurateur is bearing by washing her own dishes and compare the cost of hiring a dishwasher. This unexpected ability to compare a subjective cost to an expected cash outflow is due to the judicious choice of units, the famous “measuring rod of money”. By judiciously choosing units income and costs (and their difference, profit), however subjective, sum within a firm and across firms. Thus one can speak of the source of profit on a macroscopic scale. The sources of profit are in some sense the “positive sum” parts of the economy. Jerome Levy and Michael Kalecki, following Maynard Keynes’ judicious choices of units, showed that the sum of profits can be meaningfully distinguished into two parts: net investment and net dividends (net over non-business saving). Investment is the “increment in the value of the entrepreneur's equipment beyond the net value which he has inherited from the previous period”, which is to say the negative of the user cost of using capital. Dividends are the distribution of profits paid to households.

The Kalecki-Levy equation may seem unintuitive, since investment and dividends seem like money spent by a company. But consider what happens to the dividend: the receiving household must save or spend. By netting dividends over non-business saving, we must be counting money that is being invested in firms by households. The partition makes sense when one realizes that the receiving of a dividend and the decision to save are separate acts: “To begin with, [individuals] do not know what their incomes are going to be, especially if they arise out of profit. But even if they form some preliminary opinion on the matter, in the first place they are under no necessity to make a definite decision (as the investors have to do), in the second place they do not make it at the same time, and in the third place they most undoubtedly do not, as a rule, deplete their existing cash well ahead of their receiving the incomes out of which they propose to save, so as to oblige the investors with ‘finance’ at the date when the latter require to be arranging it. Finally, even if they were prepared to borrow against their prospective savings, additional cash could not become available in this way except as a result of a change of banking policy.”.

Take the example, Spotify reporting profits for the first time in its 19 year history. The decision to reduce the level of investment that it had last year to record an accounting profit had nothing to do with Spotify’s massive and stable income but reflected the strong growth in wages over the course of the Biden administration and persistently high interest rates. Spotify’s decision to prioritize short term profits (for the first time ever) is the result of expectations of growing wages and that the interest rate will remain above the marginal efficiency of capital (MEC) in the future.

Now, dividends are paid out of profits. Thus the ultimate source of all profit (less net dividends) is net investment: whether the creation of more solar power plants or the creation of more financial assets (i.e. deficit spending). We have previously written much about investment as the point at which the subjective cost/profit world hits the objective liquidity/cash flow world. From the source of profits point of view, one sees that investment is the positive sum part of the economy.

From a macrofounding perspective, we can look at whether Spotify’s recent cost cutting drive has created profits or merely taken profits from other parts of the economy in zero sum. Obviously, cutting costs is not an investment. Thus the only way that Spotify could have created profits is through its dividends. Overall, dividends have increased while non-business savings have decreased, suggesting that this is a major channel for profits currently. Thus Spotify is probably not engaging in a beggar thy neighbor profits strategy.

However, moving from the short run to the long exposes a problem. If Spotify is predicting that interest rates will remain so high that long term investment is no longer worthwhile, then Spotify is eating its seed corn. As Seeking Alpha’s Motti Sapir says: the best case scenario for Spotify in this environment is successful price hikes, a sudden revaluation upwards of podcasts and audiobooks by consumers or an amazing new technology. Spotify’s strategy signals that it does not consider the latter two likely. I think this reflects the general consensus of the business world, thus the general strategy of creating profits via the dividend channel.

But increasing profits by charging customers more for the same product is not sustainable. The only sustainable source of profit is therefore investment. Reducing the interest rate allow for lower MEC plant to be built and thereby investing more (especially in a green electric grid) would shift the creation of profits to a sustainable business environment. The moral of all of this is really quite simple: we need liquidity provision in the short run if we are to avoid liquidation in the long run.

In 2019, Jacobin magazine republished Jo Freeman’s famous 1972 article “The Tyranny of Structurelessness”. That article, which grew out of Freeman’s experiences as an organizer and activist in the Civil Rights and Women’s movements in the late ‘60s had quickly become a cult classic in the social movements, particularly within left-anarchism. It gave a clear account of the limitations of ultra decentralized and informal organizational models popular in that milieu, and connected those limitations to broader patterns of success and failure. It’s a great article and everyone should read it, but it is not the one we are going to look at today. Instead, we are going to look at Freeman’s “The Political Culture of the Democratic and Republican Parties”, on American electoral politics since 1972, rather than up to 1972. 

So why cover this bit of 80s cheese instead of the 70s classic? Simple: because Jo Freeman is a great political mind, and her work deserves attention beyond “The Tyranny of Structurelessness”. Her analysis of the deeper institutional aspects of social movements could be so useful for many Americans in our current political situation, and yet her voice is largely ignored.

“The Political Culture of the Democratic and Republican Parties” is specifically about domestic US politics in the 1970s and 1980s and starts by asserting, straightforwardly, that the Republican and Democratic parties differ primarily in their political culture. Freeman defines political culture from International Encyclopedia of the Social Sciences as

...the set of attitudes, beliefs and sentiments which give order and meaning to a     political process and which provide the underlying assumptions and rules that  govern behavior in the political system…

Freeman is talking about institutions here: the regularized but flexible ways the party does things in a local situation. Looking at institutions in this manner is a methodological view which is very close to my own, so I am sympathetic. Freeman then lays out her basic argument, which has two points:

There are two fundamental differences between the parties in which all others are rooted. The first one is structural: In the Democratic Party power flows upward and in the Republican Party power flows downward. The second is attitudinal: Republicans perceive themselves as insiders even when they are out of power and Democrats perceive themselves as outsiders even when they are in power.

The “attitudinal” point is simple. Democratic operatives here see themselves as permanent outsiders even within their own party, no matter their status. On the other hand, Republican operatives here think of themselves as being rightfully and permanently insiders, regardless of whether they are in power or not. These micro-level differences add up to some significantly different macro-level dynamics. The representative Democratic operator here is an individual forcing their way up the stream of the party, but the representative Republican operator is one who signs on to advance a party consensus. As a consequence of this, power in the Democratic party is the power to say No to a policy or platform, while power in the Republican party is the power to say Yes.

With this essential argument in mind, I’m going to paraphrase and comment. In paraphrasing, my goal is to explicate Freeman’s argument as it relates to the political landscape of the US in the 1970s and ‘80s. The comments are to relate Freeman’s argument to what has been changing in US politics since the 1980s, especially since the first Trump campaign. We will see if Freeman’s evaluation of the centrality of political culture and descriptions of those cultures are still accurate, and if the evaluation can be remediated if things have changed. The first part of this essay, the paraphrase/explication of Freeman’s argument, will follow the structure of her article. Finally, I’ll comment/evaluate the status of the argument in our current situation as best as I can.

Party Structure and Political Power, or Constituencies vs. Big Men

Freeman references Ronald Reagan’s 1984 nomination acceptance speech to draw out a truism that Democrats might want to deny, but which actually reflects something in American political culture: the Democratic party deals with people as members of constituencies (or, disparagingly, “special interest groups”) while the Republicans treat people “as individuals” (or, disparagingly, “cash cows”). This truism describes something essential about how both parties do business. 

As mentioned above, power in the Democratic Party flows upwards from the membership towards the leadership. Therefore membership must organize and use some mechanism for transmitting power. That mechanism is the organized constituency. Patronage gets divvied up by the organized constituency. Complaints and agitation for change get made by and through the organized constituency. The New Deal coalition of labor, African Americans, progressives, intellectuals, and the white ethnic urban machines set the pattern for this kind of constituency management. Since the 1970s some of those constituencies have declined in power, namely labor and urban interests. Meanwhile, new social movements have grown up and joined the scene. The women’s movement, gay rights, environmentalists, immigrants, Asian Americans, and so on. But through it all at its root the post-war Democratic Party is a party of constituencies.

The Republican Party, by contrast, is a party of Big Men. The Mitt Romney Company, the Trump Mafia, the Bush Family, et cetera. These groups tend to cohere around a specific man who can command loyalty to his particular project and who can distribute patronage personally. The Republicans do have caucuses and organized constituencies, but they are much less important than geographical, ideological, and personality-based organizations. Republican constituency groups will also tend to have, from a Democratic perspective, an “astroturf” feel, rather than a feeling of genuine grassroots energy. Compare the Log Cabin Republicans (the organization for Republicans who are gay) with the National Organization of Women (the main liberal feminist organization in the US, officially non-partisan but closely aligned in practice with the Democratic party) or the Congressional Black Caucus (the official caucus for African-American congressman, officially non-partisan but no Black Republican has joined in five years). Where the Democrats are constituted as a coalition of interest groups, the Republicans act as a collection of local notables.

This split in political culture is important because it determines the crucial direction of the flow of knowledge: how the parties get information about the intensity and distribution of problems, how they find innovative solutions, how they connect those solutions to coalition members, and so forth. For example, The National Organization of Women is constantly trying to squeeze information and ideas about its demographic (women) and its solutions (liberal feminism) into the Democratic leadership.

The role of constituency groups and organized caucuses in the Republican Party, however, is not primarily to bring ideas to the attention of the leadership. Rather, the role of constituency groups and organized caucuses in the Republican Party is to distribute information from the leadership to the entourage. The postwar history of the Republican party, from this perspective, is not told in terms of competing constituencies, but rather in grand top down national strategies: Eisenhower’s grudging official alliance with New Deal bureaucracy, Nixonian Hooverism & the Southern Strategy, Reagan’s extremist neoliberalism & alliance with religious fundamentalism - no group really called for any of these. 

This all leads to a question of power and legitimacy within each party. How do you get it and how is it demonstrated? In Freeman's analysis, legitimacy for Democrats derives from the answer to the question of who you represent. For Republicans, it is based on who you know and who you are. The different ways legitimacy is acquired and wielded internal to each party has significant effects on how overall policy platforms get constructed and how political careers progress.

World View

These different ways of interacting with the government and wielding power internal to each party flow from their different institutional cultures, as well as their differing self-conceptions. Democrats feel that a strong central government is “necessary in order to counterbalance private economic domination” but they also feel themselves to be on the periphery of society, and therefore feel more vulnerable to abuse by the government. Republicans also feel ambivalent about state power, but for inverse reasons. The power centers of the party are generally established men who own property, so they are not big fans of redistribution or regulation, for obvious reasons. On the other hand Republican political culture sees itself as being at home in power, as representing the real America, and therefore belonging in charge. They feel like insiders.

The insider/outsider distinction helps explain the different weight the parties put on their political virtues, and how those virtues relate to the parties’ different visions of the national interest. What Freeman means by this is that Republicans strive to promote “individual success” as synonymous with national greatness, but Democrats equate it with “fairness.” 

What makes a country great? Republicans say: when everyone has a chance to succeed as an individual. Democrats say: no, it is when everyone is treated fairly. Finally, Freeman observes that the combination of the importance of group identity for Democrats plus their feeling of perennial outsider status leads to a situation where

Democratic outsiders for whom a group identity is important are quite ambivalent about exactly what they do want. In its own way, each different group experiences a great deal of tension between a desire for self-affirmation and one for assimilation.

This is an important point that we’ll return to in a moment.

Organizational Style & Dissent vs. Disloyalty

Party constituencies organize themselves as separate caucuses within the party and use those caucuses as platforms to get their policy preferences onto the agenda of the larger party. Freeman quotes a description of an internal Democratic Party fight in Massachusetts over who got to be the State Party Chair.

In the Democratic party the affair could best be called a brawl all the way -- at least as the press reported it, no doubt with some gleeful exaggeration. Statements and counter-statements to the press, accusations of falsehood mutually tossed back and forth, gave the dispute most of the elements of an Irish donnybrook, minus only the swinging of fists. There were threats of that too.”

Note well this scene occurred in 1954! The Dems have been like this for a while. It’s not some scary new ideology that makes the inner organizational fights feel so intense! Meanwhile the Republican conventions of this period look and feel like country club meetings: closed, private, quiet, consensus-building exercises.

Freeman then notes in a paragraph so important I’ll just quote it as a block: in the Democratic Party you get on the agenda by picking a fight.

In the Democratic Party, keeping quiet is the cause of atrophy and speaking out is a means of access. Although the Party continues to be one of multiple power centers with multiple access points, both the type and importance of powerful groups within it has changed over time. State and local parties have weakened in the last few decades and the influence of national constituency groups has grown. The process of change has resulted in a great deal of conflict as former participants resist declining influence (e.g. the South, Chicago's Mayor Daley) while newer ones jocky for position (women and blacks). Successfully picking fights is the primary way by which groups acquire clout within the Party.

This is, I reiterate, so important to understanding not just the politics of the Democratic party but also to understanding the perennial tumult and discontent within progressive 501(c)3 social service and cultural organizations. It’s not new, folks!

Furthermore, picking a fight can be a smart move for your career even if you don’t win the actual fight. Willingness to enter into conflict in the Democratic Party can serve as a way of demonstrating power through successful representation of one of the party’s group constituencies. Freeman gives the examples of the fight over “equal division” in 1976, the National Organization of Women’s threats to not endorse Jimmy Carter in 1980, and Jesse Jackson’s Rainbow/PUSH campaign in 1984. None of those challenges were successful within the Democratic party but they did establish a new generation of party activists as contenders inside the party. This finding about conflict and its purpose and results relates to the next point Freeman illustrates: how the parties differ with respect to the display of loyalty and party discipline.

Maybe not surprisingly, the role that open conflict plays in each party also affects how conflicts are ended and their repercussions. Freeman says that while Republican political players do fight, they usually do it in private, and are much more likely to hold grudges than Democrats. Similarly, Democrats may pick fights to get on the agenda, but they don’t expect to be shunned afterwards. Republicans, by contrast, tended to place a greater emphasis on public conformity and tended to punish breaking the public conduct rules much more harshly. Freeman references the famous 11th Commandment: thou shalt not criticize a fellow Republican (at least until recently).

The Future

What sticks with me when I re-read The Political Cultures of the Democratic and Republican Parties is actually not anything about the political parties themselves, but rather something about the political culture of the Democratic Party. I would argue that Freeman’s characterization also characterizes progressive cultural projects as well as the non-profit system, the so-called Non Profit Industrial Complex or NPIC. There’s been a lot of soul-searching and complaining about how badly a lot of these organizations seem to be run, and some of that may actually be correct. 

But the very fact of social conflict and petty fight-picking in these organizations does not, in itself, seem to me dysfunctional in the way it appears to others. It just seems like the political culture of the Democratic party has “won,” so to speak, and its features are now shared very widely in the worlds of the NPIC.

There is a political (or I guess organizational) culture that emphasizes bottom-up flows of power and information. There is a relatively flat organizational structure that mistrusts hierarchy in word if not in deed. There are a number of strong norms about group membership and group representation as the basis for allocating organizational power. There is an ambivalent relationship to the central government that combines an agenda of restraining local and private power while simultaneously fearing government repression due to the vulnerable status of many groups in the organizational coalition. And most importantly we can see the convention of picking fights to get on the agenda and internal campaigns as a means for lower status activists to oppose incumbents and raise their own standing.

The tendency is to see these fractious and sure yes maybe sometimes dysfunctional institutions as some kind of novelty, but one of the things Freeman’s essay shows is that ethnic ward politicking is a fairly old American way of organizing things. “Irish donnybrooks” at the level of the state party and local committee were just understood as normal and the price of doing business.

The existence of this tendency also suggests that the means to amend or improve organizations in the nonprofit sector or the universities could go a couple of different ways. The first would be to just accept the system as-is and accede to the agenda setting demands of the internal organizational challengers, but then be willing to “kiss and make up” (Freeman’s characterization) and get on with it. The other approach would be to initiate a “good government” movement aimed at improving the internal processes and controls of the non-profit sector. That seems like a much heavier lift but also one that might be worth thinking about.

But regarding the major US political parties, what has changed? With the Dems it seems like there has been an exodus of the most conservative members of the party, especially Southerners who previously had strong generational aversion to the Republicans as much as loyalty to the Democrats. But from a structural perspective, this situation is totally consistent with how the party has functioned for a long time. One interest group falls from the center of the party, or even falls out of it entirely. Others rise. It’s business as usual from that perspective.

In some ways there is more grassroots energy in the Republican Party since the election of Barack Obama and the Great Financial Crisis. The Tea Party, Birthers, the whole Trump rally phenomenon, Qanon. Not to mention the extra-parliamentary “Alt Right” fascists. But some things remain the same. The Tea Party was notoriously astro turfed, not that it was unpopular, just that it was not the manifestation of an agenda emerging from the populace. Some of the more hard right wing organizations do seem to have a mind of their own, so maybe one way to put it is the Republicans do have at least one organized interest group now that exerts its own grassroots political power. It’s just that that group is, in a word, nativists.

This might be bad for Republican performance going forwards, but on the other hand the patrimonial structure of the GOP means their tendency to drive away other voter blocs may be constrained by the siloed patronage streams. On the other other hand, the old “11th Commandment” (thou shalt not criticize a fellow Republican) has clearly been violated by the Trump campaign and administration. Trump wrecked the country club aura of the Republicans and the party, despite being one of the most country clubby of Republicans. It remains to be seen if Trump as a phenomenon has totally reworked how the Republican Party does business, or if he was just another in a series of Big Men who was owed loyalty and deference. In other words: more of the same, but more so. 

John Ganz’s recent book When the Clock Broke on conservative political culture in the 1990s is good on this. This is probably also bad news for national political culture going forward, but you didn’t need me to tell you that.

But Hayek vs. Keynes is a boring framing because of how much more Keynes can be used for, and how little of Hayek is worth absorbing. Since it is cheap heat, and slop that has been run a million times over, and we don’t want to publicly reward that kind of thing, I am going to be putting the discussion of “The Use of Information In Society” below the paywall.

When interest rates were low, and demand was slack, there was essentially no reason to read Austrian economics. The only stuff that’s useful in there, to my eyes, is the questions about the relevance of the level of interest rates to the optimal “roundaboutness” of production in a capital-using economy. So when interest rates are structurally low, and with little prospect for moving, there’s not really a good reason to engage, so I didn’t.

Now that rates are moving around again in a five year window, let’s go through the good parts of Austrian econ, and then, if you really want to hear it, we can go through the bad parts.

The Good: Worrying About The Roundaboutness of Production

Contrary to what you might expect, there actually are places where Austrian Theory is worth remembering, and it is on a narrow technical question that is essentially downstream of Marx. In Capital II, we hear a lot about the circulation of commodities and money. One idea is that one way capitalist firms can do better on their M-C-M’ (money -> commodities -> more money) responsibilities is by shortening the length of time moving from one step to the next takes.

This might mean figuring out ways to speed up production, to improve marketing, to expand market access (sell faster by selling to a wider audience), or, crucially, to shrink the number of discrete steps involved in production. There’s even a special name for when this stuff gets too out of wack — a “circulation crisis”.

Now, there are a lot of ways to do each of those things to speed up the pace of circulation. You can put the speedup on line workers, and make them work faster at their existing tasks. You can create “ideological state apparatuses” and “mass media” and use it to increase the public’s desire to buy specific things, or even just encourage their somatic impulse to buy stuff in general. You can “Do Imperialism” in the broadly Marxian sense of colonizing other places in order to open them up as markets for your domestic production. You can shrink the number of steps that a good passes through on its journey from being raw materials to being finished goods, whether through technical innovation, or vertical consolidation, or any other approach to the reduction of the “roundaboutness” of production. 

It is this last one that we are going to talk about, because it is the one interesting piece of Austrian economics, to my mind.

Production can be very roundabout, the more the division of labor increases, in terms of the proliferation of firms and dispersion across geographies. People get really upset when they hear about how far this process can go. Remember that diagram of a cup of pears taking three or four transatlantic journeys before being sold? This is one of the most intuitively understandable of Austrian frustrations about low interest rates: the lower the cost of credit (so, the lower the rate of interest), the lower the financial penalty for “excessively roundabout techniques of production”. If you don’t pay interest on your holding costs for the value of inventory, then in theory there is less rush to move product or tighten up the “roundaboutness” of production than if interest rates were higher. 

At worst, the goal in this situation becomes maximization of profit through minimization of production cost through dispersal of supply chain, which means every player chases down and chooses only contracts with the precisely lowest cost producer. As we have seen over the pandemic, doing this for years and years leads to fragile supply chains which have been excessively optimized to one specific state of the world, and which break down rather than adapt when hit with shocks.

Instead of developing vertically integrated techniques in-house, this regime would push more firms towards more brittle but concentrated supply chains. This “excessive roundaboutness” leads to a production system with a lot of potential breakage points, because it will have longer sequences of connected actions to provide the same product, whether in geographical or firm-boundary space.

This notion is well-formalized in the “Hayekian Triangles” that Roger Garrison explains so well. The idea is that, essentially, for a constant interest rate, the more steps in production, the higher the final cost of production for a given final good. Different kinds of goods have different minimal network complexity (and so their triangles are different sizes), but for each good, there is also a range of choices available to each firm as to how many steps more than this necessary minimum should be undertaken. Specialization is good until it is excessive, and overspecialization is bad until it becomes necessary.

On the logic of the Austrian idea here, one of the goals of monetary policy is to act as a forcing mechanism for production networks to target some optimal degree of roundaboutness. Too low, and the system is presumed to provide insufficient incentive towards concentration and “cutting out the middleman”. Too high, and the range of goods that it is financially viable to produce is unduly constrained by the interest rate’s impact on the network of production. Granted, all this is in the most ideal-typical of terms, and the actual impact of this is undoubtedly completely swamped by the other impacts of raising or lowering interest rates, but there is a cute little logic to this account of this system.

But even without this logic, the triangles are an interesting conceptual tool for theory, because they make it easier to talk about how different classes of goods really do have different minimal roundaboutnesses. This is important for thinking about things like all the corner cutting forced on engineers by MBAs, the example of Boeing comes to mind.

But it’s also important for thinking through what has to happen for decarbonization. We want to ramp up the level of production of end goods, which means ramping up the production of everything earlier on in the process of their production. Ramping up the level of output of things on the far right side of the triangle, when the triangles are large, will require some significant time-coordination in terms of standing up upstream and downstream capacity simultaneously. Without this, regardless of interest rates, the baton cannot be smoothly passed between production steps without hitting bottlenecks.

The IRA very explicitly is trying to address this issue, but because people broadly lack the theory background to see that in it, they don’t “get it”.

There are folks out there who do “get” this theory stuff though – especially Hyun Song Shin building on the work of Pol Antras building on the work of Hyun Song Shin. In 2023, Antras published a working paper outlining an “Austrian” model of the impact of the changed rates environment on the “optimal roundaboutness of supply chains” thru the channel of the compounding cost of trade credit. This was broadly based on the work that Shin has done to develop formal representations of models of the dynamics of global value chains relevant to the work of the BIS. 

As foreign exchange and interest rates are both involved in the global transmission of this balance-sheet side impulse, the actual implementation of the model can get messy quick. Despite this, the dynamics are relatively intuitive, and have interesting implications:

I won’t gloss the papers – you should read them – but they leave two points worth noting for policy. First, incremental increases in interest rates within these models seem to broadly swim in the same direction as the promotion of near-/on- shoring, because of the dynamics implied in the HS Shin paper. Coupled with subsidies to firms in targeted sectors to pay these added costs, there is a certain logic to the use of rates to intervene in industrial organization (tho, as mentioned, this effect liable to be swamped by the comovement of other factors/forces with interest rates).

Second, the way that the Fed took interest rates from 0% to 5% is emphatically not the way to use higher interest rates to achieve this goal, which would seem to require a smoothly telegraphed hike trajectory at a speed that businesses can productively react to by rearranging their supply chain topology. What yanking rates straight up and holding them there will do is generally reinforce uncertainty about changes in upstream costs and broadly leave firms to take a defensive rather than exploratory or experimental posture, with respect to their supply chains.

In conclusion: the parts of Austrian theory that you have heard of are bad, and the useful parts are rarely talked about because they are sort of abstruse and too close to Marx for the tastes of the theory’s boosters. Despite this, it’s worth thinking seriously about these other parts, especially as industrial policy looks increasingly towards understanding and reinforcing specific supply chains. It can be hard to understand things without a clear theory.

Now that you have had your vegetables, if you have paid for a ticket, there is red meat below the paywall.

The Bad: Markets as Computational Institutions

Read more

• John Dewey, “Influence of Darwinism In Philosophy”

I suppose it best to start at the beginning: the title. This little piece is titled after John Dewey’s “Influence Of Darwinism In Philosophy”, which conveniently also begins its analysis with the title of Darwin’s book.

The very word “species”, as Dewey notes, comes from the Latin verb “specio” (to look) and the stem “ies” which converts a verb into an abstract noun. The word is a Romanization of the central Greek concept of “ἰδέα”. “Ιδέα” is not just a word, it’s a phoneme that encompasses an entire world view built into all of western thought at its very foundation. How long before Darwin it had been the goal of philosophers to escape the world of ἰδέα is shown by this: Aristotle himself declares that escape is his goal in both Logic and De Anima. Careful philologist/philosophers such as Paul Shorey (Roosevelt Lectures) and D’Arcy Wentworth Thompson (Excess And Defect) have amply demonstrated Aristotle’s failures. Darwin himself gives another example of Aristotle’s failure in his “Historical Introduction” to Origin Of Species.

The goal of this piece is indicated by the title: I submit to see how quantum mechanics has influenced philosophy, as Dewey submitted to see Darwinism. But there the title play stops, for reasoning from the words in the title “Quantum Mechanics” bears poor fruit. The word “quantum” is also Latin, meaning “how much”. Cicero says that the good and the wicked are never close friends for “nisi quod tanta est inter eos, quanta maxima potest esse, morum studiorumque distantia.”. Mechanics comes from the Greek μηχανή, meaning device or method. There is a history to these words. The term “mechanics” for the abstract dynamics comes from the Victorian belief that abstract dynamics would provide a way of describing apparently qualitative change as the quantitative change of state of what were ultimately very small machines. But this is not a philosophical history - or at least the argument that it is would take up an entire post in its own right. Dewey’s use of the title is illustrative rather than essential. The point is that philosophy and physics share concepts:

“Wo sich ein Körper bewegt, da ist Raum und Zeit, das simpelste empfindende Geschöpf in dieser Welt wäre also das Winkel und Zeiten messende. Unser Hören und vielleicht auch unser Sehen besteht schon in einem Zählen von Schwingungen.

• Lichtenberg, Sudelbücher: D

Dewey illustrates that science and philosophy share not only concepts but also aesthetics – senses of beauty. Traditional philosophy and science had held that the most beautiful and complete explanation was the explanation from ultimate origin to ultimate end. Even the motion of a pendulum was, as Kuhn put it, an “infinitely complicated” problem. The explanation of the pendulum started with its conception and design, passed through an infinite series of faster and smaller beats and ended with its conceptual transition from a moving object to a stationary object. Galileo’s revolution was to reconceptualize the pendulum as a system described in three numbers - radius, amplitude and frequency. As the pendulum users refined Galileo’s theory, the debate shifted, but the revolution remained. What the users want is for the system to be described simply by the theory, not to be given a conceptual path from ultimate conception to ultimate end. This change in aesthetic has become more permanent as classical pendulum theories have been left behind.

Most people agree that what is now called classical mechanics has had enormous philosophical impact, but many would deny this role for quantum mechanics. Perhaps this is for historical reasons: the scientific issues which confronted western philosophy (ignoring the Athenian Greek period) just before the industrial revolution put the scientific issues at the core of human society were issues within the world of classical mechanics. Philosophical conversations about “determinism” and mechanism are often conducted in oddly classical registers even today. Quantum processes are more vital to our lives than classical mechanics was to the industrial revolution, but scientific issues can be put at the core of a philosophy but once.

The example I often return to as an illustration of the doing of quantum mechanics is ammonia synthesis, the industrial chemical process which produces the fixed nitrogen necessary to continue human life on earth. Quantum mechanics links the chemical composition of the chemicals under examination to the spectra produced when heated by x-rays. Examining these spectra allows one to develop an extremely detailed analysis of the dynamics involved. 

Despite the extreme importance of quantum mechanics, I don’t pretend it is uniquely important. One could write a great essay about the influence of, say, fluid mechanics on philosophy. Hopefully CVAR will have such an essay one day soon. The main benefit of choosing quantum mechanics is pedagogical: enormously many people (everyone who has taken an intro to chemistry course for instance) have used quantum mechanics but not nearly as many people do quantum mechanics. Many people have this nagging feeling that they know something important about atoms but can’t quite explain anything about it. They know that quantum mechanics should influence ideas of space, time, measure, etc. but are not quite sure how it does so.

This use/do distinction is universal. Even Niels Bohr was careful to title his books like “The Theory of Spectra and Atomic Constitution”, “Atomic Physics And The Description Of Nature”, and (my favorite) “The Penetration Of Charged Particles Through Matter”. These titles emphasize the using of quantum mechanics.

Perhaps, as a society, we are in a similar situation to Socrates’ contemporaries or Leibniz’s contemporaries in our relationship with our scientific visions. To make the meaning of this implausible assertion concrete, I will look at three areas in which quantum mechanics has (or should have) fundamentally altered our philosophical conceptions. As a prophylactic against mysticism, I will not go into traditional “philosophy of quantum mechanics” subjects like measurement. Instead our subjects are simpler and more experimentally grounded concepts like the identity between creation and motion and the pluralization of determinism.

Identity Between Creation And Motion

“Ce principe [de raison suffisante] posé, la première question qu’on a droit de faire sera : Pourquoi il y a plutôt quelque chose que rien ? Car le rien est plus simple et plus facile que quelque chose. De plus, supposé que des choses doivent exister, il faut qu’on puisse rendre raison pourquoi elles doivent exister ainsi, et non autrement. ”

• Leibniz, “Principes de la nature et de la grâce fondés en raison”

Why is there something rather than nothing - surely few questions can be as annoyingly philosophical than this. Leibniz moves too quickly, but his next sentence contains the only solution actually available to ordinary physics: everything that exists is a little miracle. Nothing new can come into being, there is only transformation - as Leibniz puts it “Il n’y a donc point de métempsycose, mais il y a métamorphose”.

To talk a little shop, we can talk about this in classical mechanics terms. Physicists are used to relating the “dimension” (number of unknowns) of a classical mechanics problem to the number of particles. For example, the number of dimensions of a one component system where each molecule has no internal degrees of freedom (rotations, vibrations, etc) should be 6 times the number of molecules - three for each position, and three for each momentum. Now, we mathematicians have elaborately proven that dimension is an invariant under continuous transformation. In practice, this means no continuous motion can ever create or eliminate a molecule, including the actual physical motion as a special case.

The idea of the preservation of matter has been very powerful. For Arthur Eddington, the first to propose that stars are powered by nuclear fusion, the matter was so settled that the trick was not to show how each miraculous little proton was necessary but to count even count them to a unit:

“I believe there are 15, 747, 724, 136, 275, 002, 577, 605, 653, 961, 181, 555, 468, 044, 717, 914, 527, 116, 709, 366, 231, 425, 076, 185, 631, 031, 296 protons in the universe and the same number of electrons.”

• Arthur Eddington

Yet it was Eddington’s own scientific field - the quantum mechanics of the atomic nucleus - that fundamentally altered the question.

Perhaps a little historical narrative and a focus on experiments over theory will clear things up. All radioactivity experiments from Marie Curie to Ernest Rutherford had found exactly four components of the atomic nucleus, each of which turned out to be known physical objects. In order from least to most penetrative of matter, these components are: 

1. A larger positively charged “alpha” particle 

   1. Shown to be a helium nucleus

2. A much, much smaller and negatively charged “beta” particle 

   1. Shown to be an electron

3. and the uncharged and electrically neutral “gamma” particle 

   1. Shown to be light.

These particles had a range of apparently mysterious properties. The alpha particle has ~4 times the mass of the hydrogen ion and only twice the charge. The sole possible explanation in terms of already observed particles: the atomic nucleus was made of bound hydrogen ions and beta particles. The positive charge on the nucleus is the excess of hydrogen ions over beta particles. If one didn’t think about it too hard, it made sense: there was a mysterious nuclear force binding a hydrogen ion to a beta particle more tightly than the electrostatic force holding together neutral hydrogen. Even the conditions under which alpha and beta coalesce into a neutral particle can be guessed at: the other Charles Darwin had shown under certain relativistic conditions, these particles must coalesce. These conditions are hard to attain, in line with the rarity of elemental change.

The resulting picture of the nucleus is a bit disturbing. Even the helium nucleus consists of two hydrogen ions and an H2 molecule confined into a distance comparable with the scattering length of one free electron. In other words, there seemed to be a proper part of most atomic nuclei which had the parts of four hydrogen atoms squeezed into less than 100,000th the size of a hydrogen atom. 

Further experimental facts only make the picture more disturbing. As noted above, beta radiation was distinguished from alpha by its highly penetrative character. Yet the bound electrons refuse to escape the tiny radius of the nucleus!

So far I have been careful not to actually raise the issue of quantum mechanics, only experimental facts. As these experimental facts piled up, quantum mechanics brought the issues of the proton-electron picture to the level of a contradiction. The so-called Klein Paradox means that an extremely steep potential cannot trap an electron. The Klein Paradox means that if the energy for an electron to escape a well (modeled as a step function for simplicity's sake) is greater than the sum of the kinetic and rest energy of the electron then there must be a positive transmission probability. This is a result of the same relativistic symmetries that make antimatter necessary: in order for a reflection to be possible - that is, a solution where the electron reverses its velocity at the barrier - there must also be a refraction solution - that is, a solution where the electron keeps its velocity but reverses its momentum at the barrier. An electron would be freed, not bound by a strong enough potential. 

But then, if all this is the case, whence all the electrons? The simplest conclusion is that the electrons are not in there at all - but then why do they fly out during beta decay. Bohr proposed that the electrons are created by the process of beta decay. The creation of a particle is thus a part of the motion of the system no more privileged than any other motion. This is a proposal that makes sense only in the light of quantum mechanics, where the quantum numbers, such as electron number, are the most fundamental considerations.

The existence of an electron - the number - is simply quantized motion, rather than being a little miracle like the existence of a classical particle. But from whence electrons at all? And how did they get there?

There’s a traditional quantum mechanics approach to understanding the existence of objects that goes back to Bohr called “building up” or “aufbauen”. If an aufbauen approach, one imagines that the object of interest is made up one piece at a time, and proceeds by asking why that subobject is stable. Since quantum particles do not have a history (no hair as some say), an atom built using this method is fully general.

The aufbauen approach is the modern interpretation of Leibniz’s method, outlined above, and is a great approach for answering why there are atoms and other compound objects. Thus it explains everything physical that Leibniz would want to explain. But what about simples such as electrons? Can we go beyond even Leibniz’s wildest goals?

Quantum theory rephrases the question of creation of simples from the total miracle that happens once and for all time – as it was in classical theory – to a mini miracle that may or may not be happening at any time, all the time. Because the particle creation operator in quantum mechanics is not an observable we cannot go beyond “may or may not”. The unobservability of particle creation is a necessary feature of the correspondence principle that is formalized by modern quantum mechanics.

The nonobservability of particle creation is the quantum correspondent to the classical mystery of creation. Perhaps there is more to say, but once we try to get into “explaining non observable phenomena” we are well outside the experimentally verified quantum mechanics that is the purview of this column.

The Pluralization Of Determinism

Quantum mechanics teaches that creation is a form of motion, but what can be said of motion itself? From Plato and Democritus we have often been told that all change is motion, so the science of motion becomes science tout court. Similarly, determinism has been a requirement of science for a long time: Aristotle criticized the Atomists for their indeterminism. Contrariwise, many modern philosophers - Bertrand Russell, John D Norton - have criticized the concept of a single all encompassing causal relation as a pre scientific idea. If you like, you can dismiss the whole “free will vs determinism” argument as one kind of folk ontology (vague, inchoate causes) vs another (vague, inchoate mindstuff). Now, Norton & Russell’s arguments are more general than what we need, so I will just describe one of Norton’s examples to get the flavor. I will call this example Norton’s Hill, though the idea is much older (there is a generic variant on Wikipedia).

Figure 1a from “Causation As Folk Science” by John Norton. I’m calling this “Norton’s Hill”.

Imagine that you’re placing a very small hot wheels toy car on the hill and flicking the car with your finger. Consider the relation between a pair of points on the hill and the possible times it can take to get between those points. Generally the range of times to get from point a to point b will be a large set, because if you have one nice path from a to b that doesn’t go over the peak, then you can try to move the initial angle a little closer to the peak and raise the initial kinetic energy slightly to hit the same point on the downslide.

If points a and b and the peak of Norton’s Hill are in line with point a closer to the peak than point b, then that correspondence has a range with no minimum. One can always get from point a to point b faster by pointing your hot wheel at point b and flicking it harder.

But is there a maximum time to get from point a to point b? If one points your hot wheel at the peak of the hill, then one gets a significantly longer time of transit. Consider a sequence of paths getting closer and closer to the peak before turning around. In order to slide back down to point b, the energy has to decrease as one gets closer to the peak.  There are two competing considerations: the infinite slow down as one approaches the peak and the smaller and smaller distance traveled one gains in each adjustment. Since the time is the ratio of distance to velocity, the outcome depends on the limit of this ratio. Norton chose the slope of the hill so that one can reach the peak at zero velocity at finite time from any point. By symmetry, the amount of time to reach point b should be twice the amount of time to roll to the peak at zero final velocity from point minus the time spent between b & a.

This value seems to be the maximum… but the solution makes no sense. At the peak of the hill one has no momentum and no force (the peak of the hill is flat). Why would the hot wheel start rolling down hill? Much more reasonable to say that the hot wheel rolls uphill and we sit at point b with our stopwatch iterating seconds. Morally, it should take infinite time for one to go from the peak to point b if one has zero velocity at the peak.

These two solutions, one finite and the other infinite, form the basis for the two solutions which Norton gives analytically. To connect these paragraphs to his exposition: I started with what he calls the time reversal symmetry solution and built up to the simple analytic solution. 

Okay, enough of that. The point of Norton’s hill as an example (not to be confused with its peak as a location!)  is to show that determinism is not a rigorous consequence of Newton’s laws. The issue is that criticizing the concept of determinism is easier than abandoning it. Bertrand Russell himself had trouble: after demolishing the vague concept of a causal relation, he invoked against free will causes far vaguer than any supposed causal relation.

Perhaps going back to the history of the problem will help. Aristotle criticized the atomists for being unable to explain the origin of motion as if he had an explanation himself. Perhaps an impulse to outdo the Aristotleans led Newton to so strongly emphasize the determinism of his system of the world (a system that included more than just his laws of motion!). Newton is careful to demonstrate that each possible solution to his orbital theories is deterministic (for instance Book I, Section VI, Proposition XXXI has the proof for elliptic orbits). 

But Newton’s boasts betray his doubts.  As Norton’s hill demonstrates, classical mechanics is not in general deterministic. Even in Newton’s lifetime, Leibniz had made terrific hay of the unacceptable indeterminism of the laws of impact. In modern language, Leibniz was complaining about the effects of a delta potential. Even with all this, following Newton it became common, especially in France, to assert that the “laws of physics” are themselves “deterministic” or “mechanical” - that determinism was something that did not need to be proven for each problem as Newton did but instead somehow simply followed from the fact of the structure of Newtonian mechanics.

This kind of thinking is typical of 18th/19th century French mathematics. The dimension - “number of unknowns” - of a typical problem of point particle motion is 6 times the number of molecules. That’s three unknown position coordinates and three unknown velocity coordinates at time t. The initial conditions provide 6 times the number of molecules separate equations. The dimension of the solution is the dimension of the background space minus the number of equations one is required to satisfy. Assuming everything is in general position, equation counting tells that at each moment one has a discrete collection of where to go next moment.

But this is not enough for determinism. For determinism, you need a unique next step rather than just a finite number of possible next steps. Pace the French, determinism requires analytic assumptions beyond just equation counting. You can reinterpret Norton’s hill as a disk with very small arrows pointing in the direction gravity would pull you, a so-called “vector field”. The vector field has deterministic dynamics if the arrows never cross, which happens if the force function has a derivative almost everywhere which is bounded above by a constant. The derivative of Norton’s force function blows up at the origin, which breaks determinism.

Equation counting is a siren song in the true sense: the Homeric sirens were not birds but goddesses who sang incomprehensible truths. For physics and philosophy of physics, equation counting always has a force far greater than its rational criticism. Thus is the origin of the famous Problem Of Free Will. So strong is the song of equation counting that before quantum mechanics the only notion of determinism that mattered was the notion implicit in classical mechanics. This put the philosopher at a disadvantage: good philosophy would show that mechanics is not mechanistic, but if one said in so many words that mechanics is not mechanistic one was accused of misunderstanding the concept. 

There were a couple of people, of whom Charles Peirce & Henri Bergson are the most important, who did deny the notion of determinism (supposedly) implicit in classical mechanics. Russell in the piece linked above duly accused Bergson of not understanding scientific concepts. 

But these men were thus indeterminists. This simple classification is no longer necessary nor possible, for now there are at least two fundamental notions of determinism: that of classical and that of quantum mechanics.

First of all, why is the concept of causality different in quantum rather than classical mechanics? The difference in causal concepts is, as Bohr showed, a necessary consequence of the famous wave-particle duality. A simple - say, spherical - particle is described by center and radius occupied at a particular time. A simple wave - say, a sine wave - is described by wavenumber, frequency and amplitude. Planck’s constant times the frequency (wavenumber) is the energy (momentum). Confinement in spacetime is natural to particles but unnatural to waves and vice versa. What classical mechanics hides is that having a particular energy/momentum is unnatural to particles.  A conspiracy of waves with varying frequency/wavenumber (i.e. energy/momentum) to make a wavepacket confined in space, the spread of this conspiracy is the famous “uncertainty principle”.

To recapitulate, tying motion to spacetime location requires such a host of extra equations - the 6N equations above - constraining the possible functions from clock time to spatial location that the solution is not knife edge (1D) but knife point (0D). What we just established is that the knifepoint relation between kinematics and dynamics is Newtonian mechanics. We earlier cast doubt whether this was the intuitive ideal of “determinism”. In the era after Newton, this latter demonstration didn’t matter: determinism was the notion within the structure of Newtonian mechanics and if that wasn’t some “intuitive” idea, then so much the worse for intuition. But waving our hands at Newton’s laws and saying “determinism” is not appropriate for quantum mechanics. But then what notion is appropriate? We can look at an example related to Norton’s hill: a central potential with the form V=(2/3)(r^(3/2)). 

A few facts about central potentials generally, because often much of the facts about the solution to a quantum problem can be read off the symmetries. Spherical symmetry means that Schrödinger equation for a central potential separates into a radial and an angular part. The angular part is the same for all central potentials with our boundary conditions gives us two quantum numbers: the azimuthal quantum number ℓ and the orbital quantum number m. The meaning of these numbers is: 1) the total angular momentum of the particle is a function of ℓ and 2) the angular momentum in a particular direction is proportional to m. 

Perfect spherical symmetry is obviously a strong constraint on the angular aspects of a problem, but it doesn’t seem like it tells us anything about the radial aspects, which is where all the classical problem is! Okay yes, the classical problem is independent of angular momentum but we will see the quantum problem is not independent of ℓ in particular. I had to introduce those numbers. But what of the radial dependence? Here we must add boundary assumptions. Our assumptions are

1. That there is any localization at all (the wave function decays as r goes to infinity)

2. That the wave function is finite at the origin.

Independent of the shape of the potential, these already imply that

1. The wave function near the top of the hill has the form has the form r^ℓ.

2. The wave function exponentially decays far from the origin

There’s one more bit that comes from pure wave mechanics: with a localized wavefunction we expect a finite number of nodes (zeroes). There are ℓ at the origin and one at infinity, the remainder being the nodal quantum number n (not to be confused with the principal quantum number). The energy is a function of the count of nodes, n and ℓ.

Putting all this together, we expect a form something like: 

r^ℓ F(r) exp(-r) 

Where F is a sequence of functions that depend on n and ℓ, each with finitely many zeros. (Çiftci, Ateser and Koru, 2003) suggest approximate solutions of this form based on interpolation between two exact solutions. 

So what is the philosophical result of all this? I hope that I have convinced you that the outcome of the quantum algorithm is perfectly well behaved wave functions and thus perfectly well behaved probability density functions. That is, for each triple n, ℓ and m there is a differential P(r) dr which gives the probability of being found at a radius of about r from the top of Norton’s Hill. This probability can be interpreted as an expected waiting time to detect a particle a distance r from the origin.

Come back to classical mechanics for contrast. Classical mechanics cannot answer the question of how long should I expect to wait to roll down Norton’s Hill. As soon as one reaches the peak with zero velocity, all bets are off. Maybe you’ll fall down, maybe not. This is the sense in which quantum mechanics is more deterministic than classical mechanics. It’s not the case that classical mechanics answers every question but sometimes wrongly and quantum mechanics answers few questions but always rightly, but rather that quantum mechanics and classical mechanics answer overlapping but distinct sets of questions, with the classical answer standing as an approximation to the quantum answer where the explanations overlap.

There are thus at least two fundamentally different notions of determinism in physics: the relation between dynamics and kinematics in classical physics and the well posedness of wave function in quantum mechanics. 

This is our result, but again what is the philosophical implication? What one observes is that successes of the classical mechanics notion did not depend on their metaphysical “correctness”  but rather that they inspired researchers to ask questions that were appropriate to classical mechanics. In the long and short, we return to Dewey. To paraphrase Quine: the considerations of which guides to choose from one’s scientific heritage to solve one’s actual problems are, where rational, pragmatic.

Things are different now, and since we’re all good Marxists, we know that that is because of changes in the macroeconomy. But since we’re all also human beings, we know that while statistics might suggest how to operate in situations, they can never really tell you what they mean, either the statistics or the situations. Summarizing Mr. Pound The Learnèd:

“Artists are the antennae of the [I am going specifically against the wishes of that old antique jackass Ezra here when I change the sentiment to] human race; any who neglect the perceptions of their artists decline.”

If you’re looking for meaning, look for meaning where it is. Economic statistics might make a decent lamppost in the markets, but there’s no guarantee that light reaches all the way to meaning. For meaning, I usually turn to poetry. This is mostly chauvinism on my part; verse makes more sense to me more quickly than film or novel or image, and can simply do stranger things.

The goal of this short piece of literary analysis is to make one idea legible in two different ways. Since we’re talking about two different generations of folks — specifically as generations of folks — we should explain things in terms of two generations of poetry, one recognizable to someone about 30 in Cold War America, and one for someone about 30 in Post-Cold-War-America-At-30. For consistency’s sake, I’m also going to try to pick poetry representing a roughly analogous social position across the two time periods, and a roughly analogous artist-audience relationship and rhetorical occasion. Again, we’re trapped by my chauvinism here, but, sadly, as the hammerer would have said to the door were the language different at the time: “Here I stand, I just can’t do no other”.

Let’s take a look at a couple Bruce Springsteen songs, and a song from the Menzingers. Since one of the main themes of the volta under literary analysis is how we have to go back and do things right this time, forgive me if this all points to an even earlier era of literary criticism than was current at the time of the songs themselves, much less now.

Bruce Springsteen

There are two interestingly linked couplets from the songs “Atlantic City” and “The Promised Land”, where the meter and rhyme are such that one could land two different punchlines on the same setup. More simply, you can swap the bolded lines without changing meaning or singability — try it — but when you swap the italicized lines, the songs remain singable but change meaning.

On the left, we see the Koo Problem in full swing. Debts that no honest man can pay — extend and pretend — mortgages or greek sovvies — monetary accommodation insufficient without fiscal air support — everyone who’s already got assets is gonna love specifically the combination of inflation below target and zero rates, not just zero or low rates. And it’s the same for people too: can’t work your way up to a change in asset class without enough demand moving through the system — to say nothing of what happened to the homebuilders — wonder how many folks who built homes in the 2000s got to keep theirs through the 2010s. Just got to gamble — hey, some of the era’s most important growth sectors (like Silicon Valley and Oil & Gas) can look a lot like gambling sometimes, so smile, you’re in good company, It’s Time To Build (tm) — because maybe you can get over that way; who knows what happens to the hindmost but you better believe they’re gonna be stuck with it so they best not be you. A devil to take them is too lavish to even believe in, it’ll probably just be some medical biller in an ashen polo shirt through the phone lines. You can feel it in the ambiguity implied in the relationship too: getting a job and putting money away is for suckers — it’s never gonna work and you just gotta flee while you still can because maybe you can trick the people there this time, not like last time. Everybody knows it’s not gonna work, and if you can make it somehow, you’re safe to just assume every one will stick with you…not go away — and you don’t need to do anything for them because who else are they gonna find who got it working and will deal with their complaining.

On the right we see that things feel, psychically and somatically, worse, but that’s the price of real recovery. It feels bad and you want to explode — everything’s locked up at the damn Walgreens now; what do you mean we have to pay these stupid teenagers sixteen dollars an hour; when did takeout get so expensive what do you mean people have better jobs than delivery boy now — and the dogs on Main Street are howling because they absolutely agree. But they’re only howling; they never bite. Exploding is from weakness — same as biting — if you have to show force there’s no power. Climate change is a real problem, but exploding a pipeline barely produces a halfway decent movie, much less a global solution. It might be that you want to explode and tear this whole town apart, but you know that even if you exploded that wouldn’t happen. There’ll still be things to do afterwards; towns — even previously torn-apart ones like Halifax or Lac Megantic — are resilient because there’s things that need done. Building enough stuff to replace everything built out by folks with big aspirations who didn’t understand how half the systems worked even at the time is hard, and it sucks, but it has to get done.

“It wasn’t easy but we grown men.
I really mean it like the song said.”
Earl Sweatshirt, 100 High Street

One thing to bear in mind for the sake of this analysis though, is that these songs all show up in pretty different contexts. The Menzingers are able to dramatize the tension more clearly by explicitly thematizing it; the construction on Bruce above is totally artificial and even flips the historical timeline. The events that separate optimism and pessimism in the couplets above — rather than “waking up to the last ten years” as we will see in the Menzingers’ case — can be neatly summarized in the figure and actions of one Paul Volcker — another Jersey Boy, but one of distinctly different interests.

The Menzingers

What was unconscious in the Springsteen example by virtue of its artificiality is clear in the Menzingers’ song “Tellin’ Lies”. These are Pennsylvania boys after all, not Jersey — of about the same age as “Post Cold War America” — working out what to do about being how, where and what they are. They are in earnest and they deserve the same treatment as the octave and sestet of sonnets past, with some judicious edits around the level of repetition native to the form of pop punk lyric. Plus, they just make it so easy to do:

I hope you all can understand, because I would find it garish to press the point further.

Aaron Benanav and Seth Ackerman have been having an extended debate about capitalism across a variety of media properties recently. This dialog is the latest phase in a longer exchange between a number of different folks, but largely centered around the work of Robert Brenner, a major American academic figure.

The broader exchange itself is a part of the exploration of the possible political implications of a vague yet determined commitment to a particular version of the “tendency of the rate of profit to fall” (TRPTF) within the history of Marx interpretation. The immediate inciting incident of the latest round of this debate is a piece by Ackerman in Jacobin (which, full disclosure, I said very positive things about when it came out) on the rate of profit and the political implications of its dynamics. Ackerman links those implications to an account of a social-political worldview associated with a post-Bretton Woods “long crisis” of overaccumulation and under-investment alongside pervasive mercantilism.

The dialog’s one smoothing feature is that everyone involved, myself included, seems to basically agree at the level of broad political goals. Our main differences seem to me to be in the theoretical means by which each participant arrives at a picture of the “objective” macroeconomic situation and its plausible dynamics. The main point of contention outside these “objective” questions are few, vague, and largely over who ought to direct the state-directed investment that we would all agree – were we to be in actual conversation – is needed. We would all likely say that we support the UAW strike. We each would all probably say we find the technique of hitting particular supply chain bottlenecks to amplify strike impact to be both impressive and innovative. We were (I assume) proud to see an American President walking a picket line.

The biggest point of disagreement over objective matters is on the question of the general Tenor Of Things over time. This is a debate rooted in ancient history: the 2010s, a decade of global stagnation. Was that stagnation a failure of policy or was it an inevitable denouement to capitalist growth? For my part, I believe the stagnation was a policy failure and one which says terrible things about society. 

Despite this, there are a lot of folks of different political and theoretical stripe who all agree that things could not have been different during the 2010s, and that that says something about society as well. Neoliberals take inevitability to mean that the government response was “good enough”. Other, more radical voices take the inevitability of crisis to mean we have to scrap the whole global system without a thought to what might come after: “You pull the whole tooth.”.

The inevitability vs failure framing ended up differentiating oodles of different strands of “2010s Leftism”. I think the reason that this debate winds up centering on Brenner is not so much about his analysis in particular as it is about the mood his analysis is often presented in, its affective valence, how it feels. For all the soi-disant objectivity of this debate, there is something aesthetic at the core.

This is not to diminish the debate. As a Keynes person, it’s natural to me to see vibe and aesthetic as of critical importance to the “animal spirits” of entrepreneurs and the broader economy. Right now, there are real stakes to getting the mood right. The short run has to be managed by tea leaves and gut feelings. The situation has changed relative to the 2010s on a number of fronts: 

1. governments are beginning to demonstrate broad-based commitment to (at least starting) to make the investments necessary for global decarbonization; 

2. household and firm balance sheets are in a radically better place relative to the early innings of the Great Recession; 

3. labor markets have less slack across the board, contributing to higher wages at the bottom end of the income distribution and better outcomes for labor actions (in addition to the genuinely better Biden NLRB). 

The ‘inevitability’ framing – combined with an agreement that things were indeed worse in the 2010s – leads to an oppositional mood organized around the impossibility of things ever getting better. As a pure idea, pure oppositionalism might be fun, and feel very interpersonally radical. But team ‘inevitability’ – even on their own view – have only described reality, when the point is to change it.

Austerity, neoliberalism, etc: all these things are not over. Neither is crisis. The pandemic experience has one clear message though: we know how to use fiscal firepower to kill a recession dead in its tracks. If the soi-disant left plays a role in using pandemic-era inflation to emphasize the inevitability of the 2010s crisis and the impossibility of achieving full employment and sustained growth through a sufficient fiscal response, they should not feel embarrassed to stand alongside that position’s other advocates (Libertarians, Republicans, Generalized Liquidationists, etc). I would hope they might be, however.

As a student and interpreter of the work of John Maynard Keynes, I also don’t really like how Keynes has been portrayed in this fight tout court. I get that Keynes is a helpful stand-in for a particular historiography of the American Twentieth Century. In the context of something like the “Social Structures of Accumulation” literature that emphasizes the importance of the move from “Fordism” to “Post-Fordism” as representing a holistic movement of a relation of wages-prices-and-production, Keynes’ notion of full employment looms large. Even with that caveat, it’s hard to get away from the fact that the theoretical content in Keynes’ actual work presents a lot of problems for how the Benanav/Ackerman Brenner debate is being carried on. For a conversation about theory, the discussion of Keynes avoids theory in favor of (dubiously-)ascribed history of “practice” of Keynesianism. I don’t think this clarifies anything or strengthens anyone’s argument.

Since we have the same goals, we need to be reading from the same book. In the interest of that goal, I’m going to try to recap Benanav and Ackerman’s arguments — in, admittedly, fairly tendentious form. I welcome comment if important detail or accent is missing in any of these recaps. After these critical recaps, an unsurprising guest and I will walk us out with a new analysis for these same issues.

Benanav’s Story As I Understand It

My understanding of the core of Benanav’s argument is:

1.  Productivity growth is both cyclically and structurally weak;

2.  Longue durée demographic factors in developed countries are also structurally weak; 

3.  These two factors together constitute the core “drivers” of economic growth.

The result of the twin weaknesses means weaker economic growth, and more intense and more zero-sum class conflict. A number of different reasonably incompatible accounts of productivity slowdown are all given as examples: Gordon, Baumol, an oddly-summarized Keynes. Benanav deduces that everyone is definitely identifying a productivity and growth slowdown, but can’t quite explain why.

We can look at the background theory in terms of a single causal chain:

Technological Innovation (e.g. Automation) → Productivity Increase (e.g. more Widgets per labor-hour) → Economic (nominal GDP deflated by inflation index) Growth → Class Conflict Over Distribution Of The Social Surplus → Consumption and Savings Decisions → Investment

This chain is understood in a broadly Marxian cast. That Marxist cast rests on some assertions about firm behavior which I don’t think are too important to get into here, though others have hashed them out in some detail.

I have two basic problems with Benanav’s story as I understand it: 

1. The structure of possible empirical identification of this story is weaker than it appears on the surface; 

2. The extreme Ricardianism embedded in its heart.

Empirical Identification

Most of Benanav’s story — and all of its empirical confirmation — rests on movements in productivity data. Let’s just do a quick refresher on how bourgeois economists measure productivity, since the story is using the bourgeois economists’ measures. The first instinct is to think that someone is going around measuring and adding together how many things all the different factories make, but that won’t work, because what units are you going to use to add them up? Accounting for GDP this way would literally require adding apples and oranges.

So, what bourgeois economists have agreed to do is to add up the prices paid for all the things that are sold in the economy, even the implicit payments (implicit prices and quantities are “imputed” by various methods). This is the idea of nominal GDP. Then the bourgeois economists divide the sum of all prices real and imagined by the movement in a weighted basket of goods prices. The prices of that basket of goods is called that GDP Deflator. Of the infinitely many possible GDP Deflators, PCE has become standard.  Finally, the bourgeois economists divide the GDP:Deflator ratio by the number of hours worked in the economy. 

In sum: changes in measured labor productivity have to be driven by changes in one of the three following measurements:

1. Nominal GDP; 

2. The GDP deflator;

3. And the total number of hours worked.

Benanav’s Baumol argument is, within this frame, an argument that 

1. Nominal GDP will grow slower than:

2. The GDP Deflator (inflation) due to a shifting composition of the economy towards services and away from manufacturing while:

3. labor hours will hold constant or grow.

In Baumol’s toy model, these 3 outcomes are society’s inability to improve productivity in high-labor-intensity sectors (the usual examples are barbers and symphonies).

In comparison, Benanav’s version of Brenner’s particularly 70s inflected argument seems to be an unexplored skepticism of the validity of government investment. Benanav implicitly and sometimes explicitly is skeptical of investments' ability to bring faster nominal GDP growth back without stoking inflation, as a consequence of the limits to productivity improvement. Thus we cannot have a broadly capitalist mixed economy. Benanav takes as given the experience of inflation in the 70s and relatively high industrial wages then, by global standards. Sure, centralized government investment worked for Japan, and then China, but only at the cost of industrial growth elsewhere.

Benanav’s broadly Schumpeterian (at least, the Schumpeter of Capitalism, Socialism and Democracy) argument is that increasing automation and capital investment leaves less and less opportunity for further investment by reducing the number of workers in manufacturing and undercutting its traditional developmental role. Others have pointed out this path recently. Productivity statistics seem then to be almost like a bell curve over time. The rise in the bell curve is due to more nominal output per unit labor for constant inflation rate over time as improvements in technique lead to labor productivity. The fall in the bell curve is due to the relative fruits (producer surplus in aggregate) of those improvements competed away, after which firms turn the screws on labor. OR, if combined with the Baumol argument, then marginal restraint in service sector wage pressures the curve down as inflation underperforms due to something like a price war in manufactured goods.

Finally, Benanav’s demographic argument is that the supply of “more labor hours” will dry up as developed world populations age, leading, within this frame, to an upper bound on the growth of labor hours which has to translate into lower nominal GDP or higher inflation in order to constrain measured productivity.

Rank Ricardianism

There is little more normal or boring than arguments over the dynamics of the GDP Deflator in contemporary macroeconomics. The thing that really rubs me the wrong way about this whole exchange is the more or less explicit Ricardianism lurking within the overaccumulation/underinvestment story. I talk about Ricardianism in depth in a number of places in my explication of Keynes, but I will just link the one, and provide a simple explanation here.

The story of a basically fixed set of possible investments which can run out is fundamentally pre-Keynesian. Keynes’ point is that eventually the Quasi-Rents that capital goods are able to command eventually will fall to zero at a point of non-scarcity. This is radically different from saying that we are going to run out of things to invest in. The disappearance of scarcity is when things get interesting, not boring. 

Keynes' story is diametrically opposed to the Ricardian explanation. Ricardians hold that eventually both the extensive and intensive margins to cultivating land in a given area fall to zero. The intensive margin is the reward for investing into the same land - improving crop, ordering the latest seed drills, etc. The extensive margin is just buying more land. The original Ricardians were fascinated by the zero sum nature of the extensive margin and not confident the intensive margin would save us. Further, the margins seem to be a permanent feature of reality that could potentially fix values without recourse to the contingent social structure of price making. Those values have yet to be convincingly fixed.

If the margins on investment in given land were to disappear then logically enough the investment would cease. Everything has reached a steady state. The classic story goes like this: First, you use up all the good farmland; then, in order to get more good farmland, it’s worth it to irrigate the bad farmland; then, when you use up all the bad farmland, you intensify investment into the good farmland; then, when there’s nothing more you can do to make it more productive, you intensify investment into the bad farmland. Then, once you’ve done this, the things that can change how many farm products you get are: enough (disciplined enough) labor, the weather, the cost of inputs, and whether you own the farm or just work it. From there, the economy is a zero-sum contest of wills. Implicitly this is all happening at permanent full employment, because everyone whom the margins allow is working on the farm. Anyone not working on the farm simply doesn’t contribute enough to be worth it to the margins.

This story is eminently criticizable and was much criticized. Keynes’ critique is far more fundamental than most of the criticisms in the literature though. The simple fact of fundamental uncertainty means that not only will the economy not necessarily reach a steady state given by fixed physical margins, the economy is not even pointed towards one ever. Instead, what happens is folks will make use of the cash system to build things they think will be helpful in securing more cash. Their thinking will be right sometimes and wrong other times, and those past purchases will structure the future things that can be coherently incorporated into the existing physical capital structure of the economy. In sum, the messy process of price making is more permanent than the vaunted margins. 

And really, the intensive/extensive development metaphor is fine so far as it goes, but the possibilities of new things to invest in does not seem quite so limited by the farm these days. In fact, the dearth of investment opportunities has far less to do with the characteristics of the existing physical capital structure than it is is a function of the expected demand picture. Keynes was a trader, and so used a definition of the word “investment” that would be given by an actual person making an actual investment decision, and not a Marxian one. Whether or not that was a principled decision, Keynes used it to make sense of the fact that new and likely-profitable investments can become EXTREMELY limited when one is in an environment of structurally low demand.

The Ricardian story is one of secular stagnation: the tendency of the margin to fall. Secular stagnation — as many midcentury Marxists pointed out (Kalecki, Baran, Sweezy, Steindl are the first to come to mind) — is a question of insufficient demand-side intervention on the part of the government to reduce inequality and restore conditions of wage-led growth. To my thinking, this viewpoint culminates satisfyingly in (Bhaduri & Marglin 1990 (a cousin of (Taylor and Krugman, 1979)!).

From a Keynesian perspective, the Ricardian story gets the macroeconomic dynamics completely backwards. For Keynes, “technological innovation” was a question of the change in the technique of production employed. Problem is, “technique of production” is neither abstract, nor a simple: production is a social arrangement of the use of machines and other objects to make things people are willing to trade cash for. For Keynes, the width of production means the central question of productivity is a question of “which actual machines are being used to produce which actual things”. Economic analysis of any stripe which relies on nominal measurements in order to successfully aggregate quantities of the production of diverse things cannot speak to the question of actual production.

Firm-level productivity is then primarily a question of which machines entrepreneurs own, and which machines they use, and which ones they would like to replace when. Labor discipline and surplus extraction are more like failures than inevitabilities. The relevance of the functional dependence of firm-level productivity on technique has obvious implications for decarbonization and the climate transition. To transit, we will have to replace tens or hundreds of millions of machines and techniques. 

What all this economic jargon means is just this: if there is technological innovation which drives a change in productivity, the productivity change has to do its work through a decision by some specific entrepreneur to invest in some specific new capital good for use in production.

The question is now: what leads entrepreneurs to invest in new equipment? The answer is simple: effective (which is to say expected, and expected to be actualized) demand.

Only if consumers are going to want what the equipment makes, and have the cash to act on that want at a price point that makes sense for the machine and technique of production, will productivity change happen. When firms feel like they’re going to need more capacity badly enough in order to meet future demand that it’s worth spending money on, they invest in new equipment.

Keynes’ problem with trusting the market and its firms — far from being any kind of strict technocrat — is that investment is an incredibly finicky and bewilderingly social mechanism for the development of the productive structure of the economy. Entrepreneurs get freaked out at the slightest sign of trouble, especially after the 2010s. When Entrepreneurs get scared they drop the load and bolt. The next Great Recession might be hiding around every corner, in the same way the next “Great Inflation” is lurking around every corner for economists of my parents’ generation.

There’s also a complicated feedback cycle involved here: when investment falls, spending and incomes fall, which makes investment look less worthwhile; run that enough times, and it’s hard to imagine the case for building a factory in the US ever again (at least at this level of abstraction). 

With that critique in mind, we come to the government. The US government in particular has a mammoth hand in deciding what the level of effective demand for the economy as a whole is. The example of having the government pay people to dig holes and fill them in with fifty-pound notes in glass bottles was specifically used as a joke about the OPERATION OF THE GOLD STANDARD about the kinds of things which would still work even if useful things like building more housing and public goods and hospitals and elder homes were out of reach for POLITICAL reasons. Not economic ones.

For Keynes, it ultimately gets too irritating to continue leading private capital around by the nose past a certain point. The government itself is going to have to undertake the investment necessary to make sure full employment is maintained and stagnation doesn’t set in due to an economy-wide investment strategy that focuses solely on profits and not on needs.  

Keynes' ideas might involve some transvaluation of values in order to really grok, but that’s usually the price of progress. However, grokking Keynes does entail that if you want to do any of these things, you have to actually get ahold of the mechanism for administratively doing them. Lack of mechanism is one of the major problems in macroeconomics: there are a lot of opinions, and not a lot of commonly-used-tools.

To put this story in the terms used for Benanav’s arguments, Keynes would argue that 

1. nominal GDP is something that governments can positively affect by reversing recessions;

2. Once that’s taken care of, governments can use policy to look to reduce the forces that push up GDP Deflator, which is to say: lower inflation by enhancing supply-side capacity and gearing market governance towards inflation prevention. 

3. Hours worked is also well within the US government’s capacity to affect, especially given the history of labor forcing the governmental establishment of various wage and working hour regulations.

In the General Theory, Keynes argues that the best way to get to this point is by simply letting a boom run: businesses have little incentive to economize on labor — and thereby upgrade to more productive machines — without some friendly competition for topline revenue. Ruthlessly forcing “exit” on firms that are not at the immediate productivity frontier is a way to raise productivity, but one traditionally advocated by the Austrian economists. Keeping healthy demand in the market when recession threatens gives firms breathing room and space to compete, rather than being forced together in a room whose walls are shrinking.

When businesses raise productivity by forcing fewer workers to do more with less…more doesn’t actually get done most times. This is why Keynes’ advice is simply to let booms run, while governments take on a gradually but permanently expanding role in deciding what ought to be invested in. In this way, they can contribute positively to the building-up of the supply side, enhancing capacity and reducing inflation risk in the long run.

Ackerman’s Story As I Understand It

Ackerman’s story has three different parts. 

The first is a lineage of a specific tradition of Marxian thought and ecclesiology – the social organization of the body of the church. The dynamics of tradition and its ecclesiastical work is what I think Ackerman does a good job at.

I have a personal investment in this race too. The lineage I learned Marx within flows through students of Shaikh. I’m not partial to that lineage, but rather that understanding is load-bearing for me in the way that I think about the economy. Shaikh’s tradition has structured my thinking in some pretty idiosyncratic ways, given that his thought was presented to me sub rosa as A Standard Interpretation, so I rebelled against it like a normal teenager does. It’s nice to see it laid out as a Tradition And Ecclesial Structure, especially in relation to the broader arcs of Ackerman’s piece.

The second is an account of the political strategy and affective valence attached to the idea that “things will not get better” that Ackerman argues are an important part of Robert Brenner’s work and underlying thesis. This is a real 2010s type question, and feels like something approaching culture war at this point: is it possible for things to get better in the near term if we work at it, or no? The 2010s — and the whole broad Mark Fisher aesthetic that came alone with it — answered “no”. Without rehashing the argument of the previous section, one can see how that “no”came from a misperception of the limiting factor on the 2010s economy. The limiting factor was effective demand, not objective margins or productivity or demographics.

The third and last part is an explanation of how easy it is to jump between these two things – the TRPTF and the Brenner Story – on the basis of affective similarity. But, as Ackerman correctly points out, this jump can’t really be done successfully without sacrificing a pretty significant amount of overall theoretical coherence. This isn’t really a critique of Brenner, and much more a critique of the use of Brenner, or even “critique of The Brenner Guy as a Type Of Guy”.

Ackerman uses three parts to explain a possible worldview that might have social or political implications for left strategy. I don’t really do politics, so I won’t be commenting on that, but I have certainly talked with folks on the long end of connecting all these stories together in pursuit of an explanation of their bad vibes. The connective tissue I would offer is, you guessed it, Maynard Keynes.

The Ghost Of Keynes’ Story As I Would Tell It

Well, you knew I had to invite him. The ghost of Maynard Keynes and I would start with a broad observation. From 1945 - 1973, we had a historically novel governmental contribution to aggregate demand coupled with a growing ability to meet that demand through investment. His ghost, always tasteful, would likely point to Minsky on this point. Then the combination of: 1) changes in the structure of the global economy consequent on the investment undertaken in that period, 2) an energy shock which produced a shock to the whole system of 3) exchange rates and international trade upset the apple cart.

Throughout the 1970s and 1980s, the US built out a ton of energy (esp. fossil fuel) infrastructure to deal with inflation from the energy shock. At the same time, Japan had been aggressively pursuing mercantilist strategies (which can be good for the home country, and also are allowed, but are definitely not optimal) combined with domestic industrial policy. The legacy of MITI shows this clearly.

The US response to inflation, energy shocks, and external mercantilism was done in a manner more or less specifically designed to punish labor as a means of getting out of the problem of destabilizing inflation. This did a lot of very bad things over the short, medium, and long run.

In the short run, it stabilized supply and demand by making working people relatively poorer and thus less able to buy energy, rather than by bringing more energy supply online. In the medium run, the punishing response made bringing that necessary supply online cost more than it otherwise would have, stretching out the time taken to build out capacity to meet newfound monetary demand from, essentially, women entering the workforce much more strongly. In the long run, it led to a structural bias in the accumulation of income and wealth towards lower-MPC individuals (the very rich), as a consequence of the method of stimulus used to offset the strong short-run negative demand impact of rate hikes (this is the Reagan tax cuts). These lower-MPC (richer) individuals spend a smaller proportion of their marginal dollar of income, providing fewer jobs and less economic stimulus than if that same dollar had gone to a higher-MPC (less rich) person.

These long run issues are still actually very big. In this version of the Keynesian story, programs of economic stimulus from the 1980s onward were being directed towards their least-effective ends: cutting the taxes of the already wealthy. Not only do the wealthy have a lower MPC, Engel’s Law means the wealthy also tend to purchase things that have a very low elasticity of investment: positional goods. This whole redirection of income within the circular flow of the economy contributed to an environment of structurally lower demand, which led to lower growth, and thus slower investment over that period. Even worse, the scale of the stimulus continued to shrink alongside ambitions as a growing body of theory which argued “there is nothing that can productively be done about recessions by the government, whether fiscal or monetary” grew in academic influence. At the same time, bigger and bigger pools of dollars were exiting the circular flow to sit in a sinking fund as ballast for a range of financial activities. These have an associated MPC of 0.

In 2008, that process hits its absolute nadir. TARP and QE might have sounded like a lot of stimulus money if you were just hearing about the economy for the first time, but in terms of the actual engine of economic growth — dollars in the pockets of high-MPC folks — they did not get very far, because the program just swapped one kind of asset for another. 

One way to understand this decoupling of apparent stimulus from actual outcome is through the basics of effective demand. If everyone can see that demand is in the toilet (the vibes are off and NO ONE wants to buy anything), then no matter how low interest rates go, you won’t actually get very much economic stimulus out of those low rates because banks won’t want to lend to folks interested in investing in new capacity. Everyone is saying “show me where your demand is coming from” and there is a smoking crater. The government declines to fill that crater in any way. Things like the pandemic-era Temporary Superdole, or more aggressive financial strategies that centered the wealth of homeowners and the employment prospects of non-homeowners as targets were off the menu. The economy just stays in a slack position, seemingly, forever after that. The cratering has the apparently unintuitive effect of decoupling the GDP Deflator from “money printing” as well. Why would the price of tea rise if nobody is buying it at the current price? And yet folks will still post charts with M2 on them. When the Fed started raising rates later in the 2010s, most involved in monetary and fiscal policy at the time were perfectly satisfied with an unemployment rate of 5%.

That is, until we finally get a large burst of actual fiscal stimulus when the pandemic happens.

The fact that real fiscal stimulus happened during a period of moderate-to-severe supply disruption was, all things considered, not optimal. In fact, the fact that it still worked under these hindered conditions – far from the largely financial problems of the Great Financial Crisis –  is further evidence of its power. Were we to have responded to the 2008 recession similarly, the ensuing decade would have been dramatically different. 

On the one hand, the legacy of the past 40 years of running the economy “cold” on purpose combined with the optimizations made to a low demand economic environment made it so supply chains had a hard time opening up the throttle in the immediate moment when demand returned. On the other, the pandemic itself made it generally extremely difficult to open that throttle up anyway.

That avalanche of money was going to agents that had been cash-constrained for the past 10+ years: firms, households and state + local governments. Households finally cleared up some long-running balance sheet problems that were still hungover from the 2010s. Most importantly, the tight labor markets compressed the nominal wage distribution at the bottom of the income ladder. Just think, if you owe some payday lender $450, it’s a lot easier to pay that off at a nominal wage of $15/hour than $7.25/hour.

Rehearsing these facts is not to say that everyone has it great now and should quit whining. Rather I am saying that certain numbers have moved in positive ways over time, and that should be the basis for demanding more. Most particularly, it should be the basis for demanding more recession responses like the most recent one, and stronger and more powerful systems of unemployment and social support. 

The hope now, in the aftermath of the most successful recession response in the history of the US government, is that it actually is possible to put people back to work quickly by giving more money to high-MPC (low income) folks. Genuine commitment to full employment would be a massive reversal in economic policy. Achieving this reversal, given the degree of recent empirical success. However, dogmas die hard. Maybe everyone will just accept that the next time there’s a recession that pain is actually good for workers and we shouldn’t fight for a similarly overwhelming federal fiscal response as that which ALREADY WORKED during the pandemic.

At the risk of retreating into litotes, that’s the conclusion of this piece. But me, personally, I think everyone is really overlooking how important Mercantilism is to this whole thing, which Keynes also provides a way to think through. But this post is already too long, so I’ll just link to another, longer, post here.

What’s Actually In The Chapter

So far, Keynes has basically been elucidating how to understand the dynamics of an economic system considered more or less in isolation from the whole rest of society. Isolation is frustrating for some folks. Economics is a means for understanding markets, which are a means for understanding The Economy, which is a means of surviving a capitalist society. This isolation is, however, a way of getting to clarity on what parts of the economy are inward facing: those that are really just about the economy (namely money, interest and - unfortunately- employment). Today, in the last chapter (!!!), Keynes is ruminating on What The Fact That Employment Is An Inward Facing Part Of The Economy Means For Culture and Politics.

A lot of folks understand that “The Economy” is what drives most directed human effort nowadays. Even projects to meet basic needs like domestic water use are thrown into the same field - The Economy - as massive bottling concerns. Setting everything up viz a viz The Economy this way also has the secondary effect that the vast majority of folks’ cultural experiences are now given color by the often-bizarro logics that crop up out of the habits of people participating in markets, rather than people participating in world-making for its own sake. On the one hand, the often bizarro logic of markets can make the world extremely confusing: things aren’t always what they say they are. But the fact of a logic, however bizarro, also, makes getting the organization of the economy right seem like One Weird Trick for changing the world as a whole: change the economy, change the world.

Economics has thus become sort of like education itself in some ways. There are always spirited political delegations showing up claiming that they have figured out how things really ought to be. Or worse: they claim to have figured out how things AcTuAlLy ReAlLy ArE. These folks then try to mount an opposition to an ill-formed mass of habits of reading situations, such as using nominal dollar accounting to measure value, and then the whole thing falls flat half the time.

On the right, the political economists who claim to know essentially consist of folks asserting that, no, actually, there is no “economy” and no emergent dynamics and all power should devolve to small businesses and heads of household, because they are Responsible and know how to have Skin In The Game. On this view, the main problem in the economy is that people other than oneself are lazy or don’t know their place.

In the center, there are always people claiming to know that what the economy is doing now is good enough. As long as we don’t make any sudden moves at all — even ones unambiguously for the better — we might avoid scaring The Economy. When the economy is very scared it will start acting out: that could mean a recession or worse. If we scare The Economy off, things will be very bad, so it’s best we don’t change anything and let it be itself: Laissez-faire is probably best translated as “let it be.” On this view, the main problem in the economy is that people come along and try to use it to accomplish their own ends, whether private, or worse, public. Their own public ends, whatever that means, now that’s a problem.

On the left, there are always people claiming that what the economy is doing now couldn’t possibly be seen as good, or good enough. Not just “is not” but “could not be.” Thus, unless we radically change who is responsible for “what the economy is doing now,” we are collectively morally responsible for every bad thing that happens. Specifically every bad thing, simply because every bad thing can, somewhere in the causal chain, be traced back to The Economy. In the main variant of this view, the major problem in The Economy is that the bizarro logic of “production for profit and not for use” prevents everyone from using their presumed natural abilities to do things in a way that would — without producing any irresolvable conflicts — sync up everyone’s wants and needs and abilities and communities, whether at the level of use of money or ownership of the means of production.

Each of these groups marshals one form or another of social-scientific theory and brings out results to show why the things that it would like to do are correct. It’s not really a great form of discourse, but it’s what we have right now.

Okay, but what about Keynes?

As far as what’s actually in the chapter, well, Keynes talks about how, as a consequence of how we understand the economic system to work, we ought to do a lot more public investment and leveling out of incomes and wealth in order to run the economy faster and better, so we can stop worrying about economic problems and get back to more interesting questions like “How does the nucleus of the atom work?”, “Is this end rhyme too obvious?” or “Was this anime any good?”.

The way Keynes phrases it is more poetical than I can do justice to here I think, so I’m just going to do some Tallying Up. This is substantially the end of this whole Keynes project. It’s been super fun and I’m really proud of the words here. Thank you all for supporting this project. Analyzing Keynes from the tail end of America’s lost decade through today has been a wild ride. The world and I have really changed more than I could have imagined.

Why I Think Keynes and Deleuze Are Interoperable

I am going to talk my personal mental book here for a second. I want to explain “why” I focus in on the things I have within the General Theory. But also I’m not going to cite anything in a serious way right now because I am very lazy.

I am a Fan of both enlightenment and the enlightenment. I think light is cool as hell. There is nothing better than drinking unlimited coffee and trying to figure out what’s going on, in the broadest possible sense. I think the political ideals of The Enlightenment are broadly good. Everyone deserves dignity and a living because we transparently have enough things to provide that; we’re just not well-enough organized to yet. I also think feudalism is stupid, and have a real problem treating as “valid” the distinctions given by birth or merely assigned by authority alone. The tricky part is figuring out how to set the system up so that the enlightenment can survive and thrive by making things better for everyone.

I got into Marx because he seemed like an enlightenment thinker who also took the fact that — oh shit, living in The Economy fucking Sucks for a lot of people. The Economy does weird confusing stuff that its soi-disant fans and explainers have a really hard time accounting for — seriously. But the thing is, there are later enlightenment thinkers than Marx as well. Some are Marxo but also some who are not Marxians. I don’t want to try to present an exhaustive list of who I would place there, but two later enlightenment thinkers I find particularly useful are, in historical order, John Maynard Keynes and Gilles Deleuze.

Both were major public figures in their time and context: 

1. Keynes had a hand in most of the important government decisions of the mid-20th century; 

2. Deleuze wrote bestsellers and lectured to mobbed auditoriums. 

Both Keynes and Deleuzehad “methodologies” more than “theories”: they were like mapmakers who looked for recurring patterns and explained their recurrence and sources, rather than total descriptions of the dynamics of the space of possibilia for, respectively, economy and metaphysic. Both of them get bad raps because their followers had a really hard time communicating the full majesty of the methodologies, as well as why all the different parts are necessary, in the sense of implying one another. This led to a kind of botched uptake in American economics (for Keynes) and American letters (for Deleuze), which was not helped by the confused intrusions of a long-running red scare on scholarly processes.

This led to proponents who put forward far flimsier assertions than are really justified on an honest reading of the text. These flimsy assertions were then easily shot down by the political opponents of the policy recommendations the flimsier asserters were used to making. This happened within both economics and philosophy departments here. The philosophy backlash pushed Deleuze studies towards comparative literature, where students often lacked mathematical and scientific background. In economics the backlash sidelined study of Keynes into a handful of historical “Post-Keynesian” departments (Rutgers, East Tennessee, UMKC, etc) with minimal funding, influence, or ability to command public attention.

Perhaps two concrete examples from both would help.

1. From the Deleuze point of view, Mill’s definition of matter as virtual stimulation - “the permanent possibility of sensation” - makes perfect sense. The material objects referred to in a sentence are those sensations such that in the conversationally relevant ordering of possible worlds, the subjects of the sentence are having the stimulus in some possible world of every neighborhood of the possible world centered by the sentence. Mill’s matter is the core of Deleuze’s analysis of film: the objects in the frame are those you could be experiencing. But there was a time not so long ago where phrasing this as Deleuze was laughed out of American philosophy departments, as the style at the time required one to credit JK Lewis.

2. From the Keynes point of view, it is obvious that “Money matters.”. Money is the name we give to unusually permanently liquid goods, such as cash or bank deposits. Liquidity is the manner in which markets deal with uncertainty. Dealing with uncertainty matters. But Milton Fredman for one would often brag that he both proved money mattered and that he didn’t know what money is!

These all too easy victories should have spelt the doom of Keynes and Deleuze. But the World Spirit is not so predictable as the American academy. Instead each spontaneously cropped back up in disciplines where each thinker had made real advancements that were discovered to be useful: Keynes (The Philosopher Of Liquidity) in Finance, and Deleuze (The Philosopher Of The Virtual) in Tech. So the parts of their theories that get fleshed out relatively seriously are the parts that are most relevant to the subdomain they were fleshed out in.

Keynes is useful for finance, because his work can offer a better way to make practical macro judgement calls, which are a resource that can be made to be worth money. Deleuze is useful for tech, because he develops a (semi-mystical) language for talking about the use of emergent dynamical systems as alterable social-scientific metaphors. This is very fun when it clicks, and “exclusive and difficult but fun once it clicks” is something that tech folks love!

But these resurgences upset the narratives created by the all to easy victories. Had not Samuelson left Keynes in the past? American academic narrative space is usually very simplified. These two weird multidimensional thinkers. 

1. Deleuze, who cares about freedom above all else, gets jammed into “accelerationism” because some folks find the style of his translated writings fun, and want to riff on it. This eventually becomes right-wing, because all purely aesthetic movements do, eventually. 

2. Keynes, who spent his whole time trying to make it so more folks could live how his Bloomsbury friends did with their champagne and casually failing magazines and lovable conniving schemes, winds up being most useful to traders trying to snipe each other back and forth betting over whether a number goes up or down.

Everyone also got very upset about something called “postmodernism” for a while. But so far as I can tell, the jury is still out on what exactly that all was about. Deleuze is often grouped in with The Postmodernists, especially as a respondent to Foucault and Lacan. But though the projects are related and they were all frenemies, they remain distinct.

I think both Keynes and Deleuze are thinkers who take seriously the anti-foundationalist “next step” in the enlightenment project that parallels in a broad sense the “next steps” taken in, say, physics by general relativity and quantum mechanics. I think it is strange that they have not been recognized as such yet, but they also had florid and complex lives that are not quite so far in the rear view yet.

What makes it important that they are natural allies, is that Deleuze offers some really handy ways to think through some of the more dynamical-system like implications of Keynesian economics, without introducing the parochialisms of Arrow or Samuelson. Keynes provides a rigorous way to really pose the problem at the economy at multiple levels of emergence, and across a range of assemblages. Frankly, Deleuze’s ideas about assemblages are really statements of the problem, rather than solutions. I think they recursively strengthen and locate one another.

How I Personally Reconcile Keynes and Marx

Another question that has been lurking all throughout this substack is how to ultimately reconcile Keynes and Marx. If we want to reconcile them, and reconcile them seriously, we should look to pivot around a node that both thinkers share a Very Important Relation to.

That thinker in this case — I would argue — is David Ricardo.

Marx presents his work as a negation of Ricardo. I read it — and especially the mathematical approach — as a sublation of Ricardo which specifically accepts his description of the economy as “true”. Marx’s advance comes in using Hegel to argue that the implications that Ricardo draws from his “correct” description of the working of the economy are philosophically flawed. 

Even in Ricardo’s time, there were Ricardian Socialists. There is, however, no way to see beyond Ricardo in Marx: the fact that Ricardo is Flawed by not being Hegelian enough asks us to take as read that we will be following the implications of a dialectical engagement with Hegelian dynamics for…our….future….economic….designs…..which is suspicious at best and boneheaded at worst. Usually people resolve this tension by saying “well, the economy will do what we tell it to” in one tone of voice or another.

Keynes also presents his work as a negation of Ricardo. I read it — and especially the methodological approach — as a negation of Ricardo which specifically rejects his description of the economy. Keynes' narrowest goal is to see beyond Ricardo! The program is to find a “path from the present to Ricardo” — a way to make the dynamics of Ricardian long-run models actually bind, in the limit of solving all of the problems of underutilization Keynes identifies. But getting from the present to Ricardo requires going far beyond the existing institutional arrangements (that was also why Ricardo decided to get into politics after all) — and, pointedly — the existing moral justifications for capitalism’s working. To actually get to Ricardo though, you also have to use the state to do a ton of stuff, contra to the permanent demand of “Laissez-faire.” Keynes’ goal was to chart a path for the beginning of an actual economic science based on observing the economy and working to meliorate its constant problems but we have not always stayed on this crooked and narrow path. At the base of this path is a rejection of the idea that there exists an ultimate and final Fix for the economy out there. What there is, for Keynes at least, is a target destination of abundance for all.

Where followers of Marx and followers of Keynes clash, that clash often can be found on the territory of whether Ricardo was right or wrong – something which is essentially understood as a foundational assumption of the underlying thinker for either party – and not something that could plausibly be at stake in a policy debate. That’s a hard conclusion to actually practically get to though, so it is easier for Marxos to assume that Keynesians don’t care about the class struggle and for Keynesians to assume the Marxos aren’t interested enough in how the world currently works to make implementable suggestions for changes to achieve their goals. This doesn’t have to be this way but it is right now.

I think this is why Keynesians and Marxos have historically talked past one another, and then independently made discoveries already long-settled in the other camp at different times throughout the twentieth century. With Kalecki and Keynes, they basically sorted out the same thing at the same time even. However, looking at the situation this way, we might be able to see some commonalities that make the map clearer. 

Keynes was definitely not a Marxist socialist. He may have been some kind of socialist — the history of the UK is full of fun goofy intentional-living type socialist economic thinkers — but it’s gotten quite difficult to tell what that word means these days.

But it’s true: what Keynes advocates for in this chapter is decidedly less radical than a Marxist-Leninist or Trotskyite programme. In a way, his argument throughout has been: “in order to make the limits of the Classical Surplus Approach world of Ricardo binding – and durably binding in the Long Term – you have to first do a ton of different things that completely violate Ricardo’s suggestions for economic management simply in order to get to Full Employment.” Maybe once you actually get to the world of Ricardo, then Marxism will work out as a kind of inverse picture of that Ricardianism, but at that point one has reached a “steady state” in the capital stock to such a point that the onrush of uncertainty must have already been decisively defeated.

Despite this, the program implied by the General Theory is recognizably fairly radical along a number of dimensions that seem reasonably important. Let’s think about this as a kind of eightfold path of noble actions to take with respect to the economy:

1. Ending the “Disciplinary Power of the Layoff” → basically, employers — those who own the means of the production — should be able to decide who works on their projects, but not to threaten workers’ livelihood and ability to make rent and buy food. When you get fired, you should have the resources to find something that you’re better at guaranteed, so you don’t have to waste the economy’s time figuring out how to be broke or doing something less useful. It’s just not worth it.

2. Making the economy “play until you win” for everyone → basically, the degree of shortfall from full employment and full capacity utilization sets the “difficulty level” for each firm out there in the market, and also for everyone out there looking to get a job. It might be that making it so that only the best of the best survive produces more “best of the bests,” but the social and cultural benefits from making it possible for anyone with a half-decent idea to do ok business at it for a while are absolutely immense (and in my opinion, much more fun than the austere regardability of efficiency). Imagine how much less boring it would be if startup costs for a business were like $15,000. We’d get so many more fun businesses.

3. Using the state as a publicly-directed swing producer or dealer for key goods → The state doesn’t own all the firms, but it owns enough of the means of production to take up labor slack when the capitalists are Fucking Up. It also probably ends up owning extra of the means of production in certain sectors that are way upstream, or politically sensitive (whether domestically as in the case of housing or internationally as in the case of semiconductors) in order to make sure all the incentives stay aligned and everyone stays coordinating.

4. Setting up an international payments system that is designed to end the mercantilist race to the bottom while providing the same specific benefits as successful mercantilist policy to all countries → this is a huge step forward in “development” thinking, because it produces agency for developing countries and makes sure they cannot lose the game. We explored how exactly this works in depth last month, with chapter 23. Maybe the internal politics in any particular country are all messed up, but the job of international organizations in a Keynesian schema is to create systems where good things can happen eventually on each domestic level without severely destabilizing anything else.

5. Seeing colonialism and authoritarianism as larger, more proximate, and more tractable problems than capitalism for strange-seeming but true technical reasons → the problem for Keynes, is ultimately “folks using the economy to terrorize over other folks.” He doesn’t assert that hierarchies or leadership are irrelevant or unimportant: the role of the entrepreneur as “one who decides to Do Something” is carved out in order to praise it in specific distinction from the rentier, or “one who Owns Something That Makes Money.”

6. Making certain public capital goods superabundant, so that folks can do stuff they want to do → this is like, investment in public spaces and hanging out and things like this. It’s easy to imagine all kinds of urbanist daydreams fitting snugly under this heading, at the same time as large-scale nature conservation and, say, public ownership of 100% renewable power. The idea is to make it so that organizations are almost “ideal” things, by virtue of having just so much equipment lying around.

7. Letting people still play economy for fun, even if they are safe and have enough to eat at the end of the day → Keynes is right to note that people do love playing economy and making money and being on the top of the leaderboard. He also acknowledges that the things that people do to get there can often be unacceptable, but in an economy where people can’t exercise power over one another, in the same way, it would be fine to let people run up the score against their friends. No harm no foul there, plus who doesn’t love a good game of rollercoaster tycoon.

8. Elimination of poverty by directing money at the bottom of the income distribution as a means of boosting demand, and thus consumption and investment → This is much closer to a technical explanation of how to reduce what is currently understood to be poverty with the tools that we currently have than the Marxian account of exploitation is. It is also much better than the account of poverty given by some economists (implicitly or explicitly): “some folks just ain’t worth nothin.” Keynes would see low-end wage increases and cash payments to poor folks as very good ways to achieve control over the macroeconomy. Disagreement is a mere unsupportable aesthetic preference which also happens to be tasteless.

The idea here is that this kind of economy would essentially destroy the last remnants of Feudalism. The achievement of “full employment” — or, say, the obliteration of the reserve army of labor in more floridly Marxian terms — effectively ends the bad part of the deal of the “double freedom” imposed on pre-capitalist peasants. This famous double freedom is most often explained as: “peasants became free from the limitations of being peasants, but also free from the benefits of being peasants.” We can read Keynes as looking to restore a version of the safety of benefits whose loss has been long lamented, without restoring serfdom.

Something to think about with Keynes is that the program is designed as a way to end the threats posed by the “natural” world of the kinds of emergent structures that a monetary production economy gives rise to without making everything Static, Boring and Permanent in the way that the economics of Ricardo drive towards.

Keynes is an enlightenment, open society, “new things will always be happening no matter what, and that’s great to see because otherwise I’d be bored as hell” kind of guy, at least in this book. The threat to this is, at the end of the day, nature: the accidental, un-meliorated, un-adjusted order of things. But that nature runs all throughout human society. There are myriad of places where things simply aren’t adjusted to one another because they haven’t needed to be yet.

Another thing that’s important about this, is that the goal is to create a flexible and resilient system with respect for the individual by essentially designing the state such that it’s possible – and eventually – easy for everyone to Win.

This is the mindset of someone who Doesn’t Want To Have To Bother Thinking About The Organization Of Production, so that they could think about partying or more delicate pursuits instead. It’s a world where “don’t worry about it, we figured out how to make it good times all over, as long as we all love and trust each other into the future and find exciting things to do” is the desired outlook to employment or financial worries. Everyone provides, everything provided, and organized based on what people think to organize — not based only on who owns what’s there now (though knowing how the existing system is organized will always be an advantage).

What the Marxian mindset often ends up pointing towards on the horizon is something like: “we will finally solve all of our problems by making all of the decisions correctly and executing on them, including all the ones that contradict the other ones, for the good of the Workers.

A Last Detour Through Evangelion

Maybe the easiest way to think about the difference between them is through two different ways to think about the ending of the show Neon Genesis Evangelion.

The original ending, which was the one Anno liked, and wanted, and which really was based on the really-existing material constraints of the show’s making (it had to be cheap as they ran out of money), lends itself to a broadly Keynesian reading of the whole show. Everyone figures out how to love and trust each other after a period of really struggling with what that would entail, and congratulations are justified, because this way, all the things that contradict the other things can be reconciled through Love rather than have to do battle.

However, this upset so many people so violently, that they mailed him death threats, and said that the drama of the show was so obviously leading up to something different, and that this was a ridiculous cop out for a show that they (the audience) had cathected so much emotional energy into. Obviously things were heading for a Grand Confrontation, and obviously that had to happen, and obviously the audience had to see it.

Let me just make the economic stakes of this goofy metaphor clear.

In Keynes, the point is to treat the economy like an excitable animal that keeps getting caught in nets. It wants to run around, you want it to run around, but it just keeps getting tangled up and hurting itself and freaking out. Where there are contradictions discovered — strikes, inflationary bottlenecks, political upset over a rising standard of living for people who “aren’t supposed to” have one — the Keynesian response is to untangle them and send the economy on its way.

In Marx, the point is to treat the economy as the site of contradiction, specifically in class relations — the distinction between owners of capital goods (the Means of Production) and non-owners of capital goods (to the extent that the latter constitutes a “class,” I usually find it easier to think “capital” “labor” and “other” when I want to use Marxs’ terminology in a strict manner).

This leads to a worldview where the contradictions need to be heightened, the two dialectical sides need to be Smashed Into Each Other to create Some Third Thing. It also tends to assume that the achievement of growth in a capitalist system is something which can be safely assumed, and which the system itself is constantly militating towards, such that opposition to that growth can never be “wasted” because at least it’s moving things in the right direction. So you smash all the contradictions into each other, and then end up with what you wind up with after. Theory says it will be better. Anyway, back to the metaphor.

Anno obliged — suddenly there was a bunch of money coming in — and made End of Evangelion, the only movie to try to address what it would be like if Kojevean “Real Stalinism” were ever to be tried. In it, everyone gets everything they want — even the things that contradict the other things — but they get the Business End of them in Confrontation. As a result, the contradictions all smash into each other, and annihilate the whole scene — the whole planet, really, as the scale of the narrative expands outwards — and everyone gets everything they want, at the expense of any of it existing. “Here,” Anno seems to be saying, “this is what it would look like if you ran the solution where all dialectical oppositions are directly sublated and crashed into each other in the material realm. Looks like shit and Tang, doesn’t it. You all wanted this, not me.”

And people went bananas for the movie. It gave them what they were looking for in the grand confrontation. I think this is something of an understandable but conservative result: given the choice between material tragedy and love, folks prefer the tragedy. Maybe they’ve been hurt too many times, maybe it just makes for a better story that way, maybe people just want to watch a fun giant robot show, I don’t know. I know that Anna Carabelli shares this tragic reading as well.

Keynes also takes great pains to point out that nothing in the parts of capitalism that we like require the parts that we don’t like — and that that specifically is what makes those parts so deplorable. It’s not really a question of the child from Omelas here. Not only are people being exploited, they don’t even have to be! Simple-minded libertarian arguments about how people have to be paid more than others might work in the extremes, Keynes allows, but for a normal person living a normal life, there’s no need to force them to compete to survive. We’ve advanced far enough to not need that anymore, but the challenge is in figuring out how to convince people of that while succeeding in the world-historically difficult task of successfully governing a modern nation-state. The problem of figuring out how to guide the macroeconomy is probably harder than the problem of figuring out how to justify the existing distribution of income and wealth, for good or for ill.

Keynes isn’t really trying to “hold back the revolution” — he’s just fighting Ricardo. The twentieth century does make it seem that some people who earnestly would like a revolution got Ricardo’s economic strategies tangled up in that desire though. Hopefully we can all find some path towards a better life together despite that.

BY ORDER OF THE AUTHOR

Per G. G., CHIEF OF ORDNANCE.

What Happens In The Chapter

This is one of the more complicated chapters to follow. Not because the logic is difficult, but rather because Keynes is placing his theory in the history of economic thinking. We are discussing “economics” rather than “the economy”. That necessarily involves a lot of context that we have mostly shied away from. Paul Krugman called this chapter “a delightful desert” rather than the “tough meat” of the earlier chapters. But really it’s more like an aperitif, in the sense that it’s the time when backroom deals get made.

Keynes disclaims the priority of identification of any of the parts of the system he presented. Keynes claims rather he was just the first to tie all the parts together. By tying them together as a logical whole, he could make sure that the causal linkages close and the dimensional analysis checks out. Keynes predecessors who stumbled on bits of this system also usually happened to be of two kinds: 

1. extremely practical statesmen and 

2. baffling cranks. 

This is a new point of view for a long established fact. So-called “deviant” or “heretical” economic viewpoints throughout history have happened to line up with the practical advice of the folks running the state.

‘Cranks’ are the far side of another famous Keynes quote: “the hard part is getting out from under old ideas” not coming up with new ideas. The Crank is, to put it mildly, someone who is already out from under old ideas. They are in fact so far out from under so many old ideas that they become either wildly wrong, stubbornly difficult to communicate with, or both at once.

Despite this, Keynes fucking loves Cranks. That’s an extremely good thing about him! Keynes likes Cranks because they are a source of Random Variation in intellectual evolution. And if used right, that Random Variation can help a body of thought trying to advance rapidly. Through Random Variation, large jumps of logic get produced, and then either whittled down by selection, or identified as features mis-perceived or mis-argued in the existing framework.

But who are the Cranks? A Crank is just the “organic intellectual” of any kind of person sufficiently practical or impractical that they find themselves at odds with the “mainline scholarship” of their era. “Organic intellectual” is Gramsci’s term, a way to talk honestly and fairly about the scribbling portion of every social class and stratum– even as a peasant, “One becomes a priest or a bandit.”. When the mainline scholarship gets too obviously wrong — or recommends too impracticable policy — Cranks have their chance to shine. They start being able to score points off of the mainliners, and eventually there’s a crackdown or a social crisis within or about the discipline.

There is, though, a parallel between the process of Cranks and the best truth discovery process that predates modern science: dialectic. For Socrates, figures like Euthyphro and Thrasmychus are Cranks, social science cranks. They argue that evil is good and everything is evil, respectively. But it is exactly because Socrates was forced into an extreme position by tangling with Thrasmychus that Πολιτεία [the original Greek at Trombley’s insistence] has bite. In Due nuove scienze, Galileo uses both engineering folklore and Aristotlean reasoning as his Cranks. Today we see who Keynes uses.

The ability to distinguish the wildly wrong Crank from the insightful but stubbornly difficult to communicate with Crank is the highest mark of taste in a scientific — and especially social scientific — context. To figure out if there’s anything to a Crank’s point, you have to 1.actually dig in and follow their system and 2. have a sufficiently strong grasp on one’s own system that you can identify what is living and what is dead in the other’s.

In my opinion, bothering to give credit to cranks as someone at the forefront of your field is the absolute height of good taste. Whatever issues their systems had, the best cranks really do put in the hours. If one is swinging for the fences and playing for the ages, and if your gambit works!, then those cranks will get to hang a harmless jersey up in the rafters of the literature too, forever. Too many people fail to give cranks their due. It is true that Cranks do, often, inspire irritating and monomaniacal followers, but so do most things worth paying attention to. Keynes in this chapter is here to celebrate reason, not social forms.

But Keynes also loves to reconcile the less-than-fully-formed doctrines of “practical folks” with the economic literature, in order to show the obvious faults in the latter. This violates a traditional code among historians of economic thought: Never Admit That The Practically Right Could Also Be The Theoretically Right. Ricardo’s theory that replacement cost of gold is the sole and invariable stabilizer of currency value is not practically right, but you’re not supposed to admit this entails that his critics had a theoretical point.

In today’s chapter, Keynes takes on the question of why, historically, maintaining a surplus in the balance of payments — for technical reasons relating to international settlement, that meant accumulating gold bullion — was seen as critical to effective statecraft, despite economists historically warning against it in favor of free trade.

Since Keynes’ theory claims to explain all of the underlying dynamics that each of the more crank theories identify that are at variance with Ricardo and them, he needs to be able to explain where he agrees and where he disagrees with all of the relevant cranks.

Keynes’ system is simple once identified, but complicated to demonstrate: since employment comes from demand and demand comes from consumption and investment, the main risks to employment and growth are insufficient consumption or insufficient investment. Getting to sufficient consumption is a question of gross labor income weighted by propensity to consume that income. Getting to sufficient investment is a question of keeping the Schedule of Marginal Efficiency of Capital (the expected yield on new capital goods) above the Rate of Interest (the expected yield on debts).

The thing is, the Keynesian system works so well that it’s actually empirically relatively easy to stumble on certain successful descriptions of the economy’s working. The thing is though, it’s hard to stumble on all of them at once. In the absence of a network of good-faith scholars to build a real research program around, these cranks then fill in the gaps in their own knowledge different ways, but rarely with the best arguments or equipment.

So who are the Cranks here who have seen through to parts of Keynes’ system throughout history? And what parts of it did they get?

1. The Ancient Statesmen got that when interest rates were too high, it made everyone poor and slowed the economy, so they enacted usury laws.

2. The Mercantilists got two weird parts: first, lower interest rates lead to higher investment and second that businesses have “liquidity preference” type issues around producing and holding inventories of goods. They then knitted these two facts together — one problem of the inducement to invest and another of the business cycle — into a common policy prescription: sell domestic wares for foreign gold so that financial conditions will loosen. We will have a lot more to say about them below though.

3. Silvio Gesell got that a tax on liquid wealth was essentially the same as a negative interest rate, and could be employed with the same effect as lowering interest rates, either in stimulating household consumption or dissuading firms from the maintenance of large sinking funds.

4. Mandeville got that if elites spent more money throwing better parties, then the jobs and demand involved would produce enough effective demand keep the domestic economy moving and all its workers employed, much to the chagrin of bourgeois or religious notions of personal morality.

5. Malthus got that a major cause of depression was jobs disappearing because people were too broke to buy things, which then made more people too broke to buy things, which then led to more jobs disappearing.

6. Hobson got that “excessive” domestic saving could easily hamstring consumption without creating domestic investment, thereby leading to apparent “overproduction” which is really under-consumption. He also got that this got worse as the distribution of income gets more skewed. He also got the sort-of unrelated point that if you do too much outbound FDI relative to domestic investment, you can effectively artificially tighten domestic financial conditions.

The problem is, relying on these prior theories involves relying on a lot of assertions that either aren’t true, or are neither desirable nor necessary for the limits to explanatory fidelity the ill-formed parts of crank theories enforce. So Keynes’ goal is to show how his theory subsumes, extends, situates, and corrects all of these approaches.

This is a hugely important thing to do for philosophy of science reasons: a new General Theory needs to show how it can plausibly re-derive successful descriptions from previous paradigms while showing how the limits of those past paradigms are sublated by the description from higher generality. This is what Keynes does.

Read more

There are three chapters in this last book. This first post-theoretical chapter is on the “business cycle”, which is the tendency of markets as a whole to boom and bust, over and above the fates of the particular firms and families within it. The next chapter is on mercantilism and the strategy and consequences of international trade, while the last is on social philosophy downstream from the Economic Base. 

It’s nice to be doing the Business Cycle chapter now. The broader Pandemic Business Cycle is proving incredibly confusing to be for folks approaching it as a mere shock. For the last eight or nine months, nearly everyone has been saying that a recession was merely six months away. What I want to use Keynes to say is that, simply because nothing has happened to bring a recession about, no recession is imminent. For a recession to become imminent something deep has to happen to the state of the economy.

Hopefully, as a result of this chapter and thinking through how the mechanics Keynes describes works, the business cycle might be a little easier to get a handle on. Ideally, we’ll have some time to sneak in a contrast with Marxian Crisis Theories and talk about some of the goofier cycles as well. At the end of the day though, cycles rule. They are so fun that there is a whole industry of guys who liked Fight Club who all read newsletters and pretend to one another that there is some absolutely critical cycle going on that everyone else is missing. You can just watch cycles go up and down all day long, or listen, if you’re more musically inclined.

Read more

Above is an excerpt from Keynes’ A Treatise On Probability. In this short passage, Keynes proves Bayes’ Theorem from an equational structure on multiplying probabilities for the first time in history. Keynes motivated those equations with a theory of still logical but merely probable inference. Keynes method of proving of Bayes’ theorem is what gives the idea its modern importance. But for what blasted reason did Keynes develop a whole theory of probable inference in the first place?

“But, of course, the real test of any new principle in science is not its ability to re-derive known results, but its ability to give new results, which could not be (or at least, had not been) derived without it.”

• Edwin Jaynes, Probability: The Logic Of Science, Chapter 30: ‘Maximum Entropy: Matrix Formulation’.

This question is related to one of the recurring questions we have had: “Is philosophy about forming a Socratic dialectic with one’s cultural endowment or about sharply defined intuition gadgets?”. We have frequently used probability theory as a place to investigate this question. In addition to Keynes, one might wonder where Kolmogorov, Cox and de Finetti fit into all this. Indeed, how can our dialectic be precise without axiomata and the resulting deductive structure?

This question represents a certain view on the foundations of science. The name “foundations” suggests a hard place on which structures can be constructed, like Kirkcaldy. The foundations of a building are laid first and those foundations are never disturbed lest the structure whole and entire must fall down. This instability is rather unlike Kirkcaldy, as far as I know.

In certain old ways of thinking (I have never been able to find exactly who holds this oft refuted doctrine), all that is necessary in the foundations of science is the true meanings of the technical words: from the nature of a triangle, conclude that the sum of the angles must sum to two right angles. Devastating criticisms of this point of view are contained in Kritik der reinen Vernunft, Πυῤῥώνειοι ὑποτυπώσεις, वज्रच्छेदिकाप्रज्ञापारमितासूत्र, Πολιτεία and 莊子, to name an arbitrary handful of old books.

This is not to say the dialectical method is necessarily historical. I recall some classicist arguing that the original Euclid had no postulates himself, and one of the points of the book was to reconstruct what the postulates would have to be (much like Book XI, which has 25 definitions but no postulates for solid geometry). That would be an example of a deductive dialectic which is not historical.

As those critics and classicists suggest, Kirckaldy manages to get things done in a more stable manner. So, I believe this view of ‘foundations as foundation’ is incorrect. Instead, I hold that the foundations of a science are the part of that science where that field itself is the positive domain, though the foundational theories may be positive or normative. In the foundations of mathematics, the subject is mathematics itself: can mathematics be proven or at least suggested? The aim of the foundations of statistics, then, is to see if statistics can be proven or at least suggested.

As Jaynes said, success is when that proof or suggestion is and provides something beyond previously observed statistical methods into new statistical methods. As we shall see, the questions our approach raises apply to machine learning as well as traditional statistics.

Thus I see, for example, Kolmogorov’s pamphlet more like a lamp than a pyramid. Kolmogorov’s axioms of probability are tested immediately by whether they can successfully hold up the important relations of probability - Bayes Theorem, Independence and Conditional Independence - and are tested again in a mediated way by the fact their interpretation in a space with finite Lebesgue measure allowed for the discovery of new statistical methods. The research program of Kolmogorov’s axioms are, in this way, justified by their success.

Let’s take one example. Kolmogorov had a structural theory of the notion of independent experiments. By structural, I mean that the theory is developed using a recursive system of equations without regard to their real world interpretation. The structure itself was abstracted from existing statistical practice. The idea of independent experiments did not burst into the world virginally from the skull of Kolmogorov!

To name just one example of how independence was used in statistics before Kolmogorov, Airy had treated independent and entangled (i.e. nonindependent) observations in his book Errors Of Observations. Airy gives the example of two reports on the transit of the moon, one from observatories a and b and the second from observatories b and c. The probable error in each report is (related to) the sum of the probable errors from each observatory’s data, but the total probable error is not the sum of four probable errors. Thus he sees entanglement (and, by implication, independence) not as structural, but substantive. 

Still, Kolmogorov’s structural theory of independence has been successful in Jaynes’ sense. Its formalizations have been seen to be useful for information theory (see Kolmogorov’s “The Theory Of Transmission Of Information” ), population genetics (see Kolmogorov’s “Solution Of A Biological Problem”), statistical mechanics, etc.. But, damn it all, can a structural approach really capture the idea of independent experiments?

Let’s go a little further. Kolmogorov’s structural approach aimed to sail between the Scylla and Charbadis of interpretation. The Scylla was the open danger of philosophizing in public, especially if one cited officially deprecated reactionary philosophies such as positivism. Unfortunately, Kolmogorov was highly influenced by Richard von Mises: an empirio-critic and brother of an economic adviser of Engelbert Dollfuss. The Charbadis was the danger of simple failure: Borel’s speculations into probability theory failed to gain scientific followers (though they had a scientific audience) largely due to Borel’s lack of statistical feeling.

This point can be illustrated by a criticism of Kolmogorov by the aforementioned Richard von Mises. From his recursive system of equations, Kolmogorov derives a necessary and sufficient condition for the independence of two events A and B, simply P(AB)= P(A) P(B). But von Mises objects that the above equation could hold “by coincidence”, as demonstrated by the fair four sided die S = {1,2,3,4}, the event of rolling a number less than the mean A={1,2} and the event of rolling a prime B={2,3}. The probabilities of events A and B are both ½. The event of rolling a prime less than the mean is the joint probability AB = {2}, which has a probability of ¼. Thus the independence equation is satisfied. Yet, the probability of rolling a small number and a prime are intuitively related because the primes are “front loaded” in the natural numbers in the sense that the average gap between primes between 2 and n is proportional to the count of digits in n.

Richard von Mises argued that this and similar considerations suggest the structural approach is merely necessary, not sufficient. Further conditions, ones subject to empirio-criticism, must be also implemented. In the case considered by Airy, the condition for independence would be something like physical separation in space. 

The goal of this essay is by no means an attempt at criticism of Kolmogorov’s theory of probability, not its structural core nor its measure theoretical interpretation. Even less is the goal to propagate empirio-criticism as an approach to the philosophy of science. Rather, this brief exploration of independence is of interest because independence is something between a goal - we want independent corroborations of our theories - and a fact. We shall soon see that if we can take the distinction between goals & facts as our guide, then we can come to a deeper understanding of statistical practice.

A Typical And An Atypical Backstory

Before we can use those distinctions as our guide, I must look at a navigation method you might have seen before. On Wikipedia, that weighty authority of all knowledge, there is a fun chart of the main schools of probability. These schools have given us weighty considerations on the question of how we can and should use statistics.

This chart is original work, which can be seen from the fact that “reference class problem” is put under “subjective” but the alleged source, which seems to be mostly following Donald GIllies’s Philosophical Theories Of Probability, puts it under “frequency”. Of course the chart is far too clear and symmetrical to represent the positions of actual theorists. The alleged distinction between “inductive behavior” and “inductive inference” which caused such a row between RA Fisher and Jerzy Neyman isn’t present. But the chart is still a nice place to start, as it is a typical and thus easily shared backstory to statistical philosophizing.

“It seemed to me then, as it does now, that something new in statistics most often comes about as an offshoot from work on a scientific problem.”

• George E P Box, An Accidental Statistician, “Chapter Five: An Invitation To The United States”, emphasis in original

There are other methods of organizing than this. As Box said, the problem that one is working on is something that must be accounted for. This is what is meant by being guided by fact, the facts of one’s problem.

Therefore, what must be done for the achievement of the goal of this essay is to propose another little oversimplifying view on probability. This one will be sensitive to goals and facts rather than basic philosophy. Further, rather than a list, this one will be structured as a good ol’ consulting firm 2 x 2 matrix. This matrix will be complete at the end, with the remainder of this post being a dialectic which develops the reasons why the entries are placed where they are.

Before that dialectic can start, let us map out the space we are exploring. The basic axes of the matrix are:

1. Method: Is the underlying action driven by compensation or inundation?

2. Goal: Do I want to predict something or describe something?

The dialog has thus started where Hegel recommended: mere abstract existence. The next step is the concrete process of seeing how the table will - or should have - come to be. That step will take us to the goal of a determinate table, which we can then use for ourselves.

X-Axis: Compensation vs Inundation

Any answer to the question of what we can and should do with statistics is already related to the system we are examining. We give the two rows to “compensation” and “inundation,” terms which come from Richard A. Epstein’s Theory Of Gambling And Statistical Logic. We will start at the top left and head towards the origin in defining and illustrating these terms.

X-Axis, Part 1: Compensation

Let’s start with an illustrative definition: a system with an equilibrium is driven by “compensation” if deviation from the equilibrium causes a change in the state of the system. An example of a compensatory system is a tensegrity structure where each piece has disposition to resist tension or compression, so that if one piece is disturbed the force is distributed over the whole structure so that the dispositions of each piece to resist tension/compression optimally reinforces the disposition of the whole structure to remain solid. Changes of state are capable, essentially, of compensating for the change in equilibrium.

Philosophers Karl Popper and Ian Hacking have advocated interpreting the probabilistic dispositions of a system as objective properties of the system whatever the particular outcome of the experiment may be, similar to how fragility is a disposition of glass whether the glass ever breaks or not. I will write this way as well, ignoring whether it is a façon de parler ou pas.

The simplest example of a compensative system is a system whose equilibrium is a stationary state. The stability of that state can be found by the Lypunov function method. If the state of your system is exactly described by a vector x and the dynamics ẋ are of the form ẋ=f(x), then we can define a stationary state x* as any vector such that f(x*)=0. Then a Lypunov function V is a differentiable function of the state of the system x which is positive definite except at its root at x* and V’ is negative definite. If such an arbitrary function V seems like an odd thing to chose, then just recall that by the chain rule V̇=V’ ẋ so that dx = dV/V’. The last equation says that the differential of the system’s state is proportional to the differential of any Lypunov function, with the proportionality coming from the derivative of the specific Lypunov function. Since V is positive definite except at x* its differential dV are pointing away from x*. Because V’ is negative definite, the differential dx must be pointing towards x*. Thus the dynamics must be moving towards x* and x* is stable.

Though a crude visualization, this still is a bit abstract for a blogpost. If we recall that thermodynamic stability occurs if a system has a minimum energy given its entropy, we see that the function V above is really just a generalization of the idea of energy. The system tries to minimize the generalized energy V and the minimum of the generalized energy V is at x*, thus the system is trying to get to x*. This is helpful because generalized energy has a structural rather than substantive meaning: it just is some function that satisfies some inequalities, not anything to do with the specific properties of “energy” as such.

Before moving on, I do want to mention one fun historical tidbit about Lypunov functions. Lypunov functions are only useful for stable stationary states, so they actually did not accomplish what he wanted to do (examine the stability of the periodic orbits of the solar system). As a result, the method was little used until Nicholas Minorsky, who popularized the concept of PID control, advocated their use in stability analysis for large ships. After this, use of the approach grew like wildfire, especially in the field of aircraft control. 

The merely structural definition of the Lypunov function gives us something to consider when thinking about goals. Do we care about learning the substantial energy of the system as an accounting or as a function? Or does one only care about the stability of equilibrium? In the later case something structurally defined, like the generalized energy function V, will work. This is a good question, but we can put off wondering about those goals until a later section.

X-Axis, Part 2: Inundation

Compensation is relatively easy to understand, and inundation is its obverse:  a system with a stable equilibrium is driven by “inundation” if deviations from an equilibrium do not cause a change in the state of the system. The well known notion of the gambler’s fallacy is based on a confusion of compensation and inundation. The equilibrium of a sequence of fair coin flips has the number of heads arbitrarily close to the number of tails, but a run of heads does not make a tail any more likely the next time the coin is flipped. Return to statistical equilibrium happens, in this case, by inundative trivialization of even long runs of heads. Essentially, without a change of state you “run out of” any given amount of deviation from equilibrium as the system adds more coin flips to its history. In fact, this inundation can be quite slow as seen by the fact that the most probable number of origin crossings of 1D random walk is zero. This is the well-known “beginner’s luck” that can take time to overcome.

On a more abstract level, consider a chaotic deterministic dynamical system with decay of correlations, such as Sinai billiards. Sinai billiards present a model where a perfect ball bounces around a frictionless billiard table with an obstacle with a smooth boundary in the center. The system is deterministic: the path of the ball is a function of its initial condition. However, it can be shown geometrically that some obstacles, including a circular one, are dispersing and cause decay of observable correlations. A good example of an observable for Sinai billiards is the “free path”, the distance between the ball’s current position and its next collision. We can use a camera to observe the free path at different moments of time and compare the observed average free path to the mean free path. The observed average free path will be normally distributed around the mean free path. But even a significant deviation from the mean free path - if by chance we happen to photograph the ball near the obstacle several times - does not cause a change in the path of the ball. The path of the ball remains a function of the ball’s initial condition. Many observations will cause the observed average free path to converge to the mean free path by inundative trivialization of chance photographs near the obstacle. 

Perhaps this is all a bit abstract. A pure gambling illustration of compensation vs dispensation can be seen in the well known 18th century card game La Her. In the first round, the Dealer shuffles the deck and deals one card to themself and one for the Receiver. Epstein calls this the random play round. After this the Receiver may exchange cards with the Dealer. Next, the Dealer may draw a card off the deck. However, there is one extra rule: the position of Sir (a king) cannot change after the randomization round. If the Receiver tries to exchange for a king, then the Dealer declares “La Her” and the exchange is negated. If the Dealer tries to draw a king, then the Receiver declares “La Her” and the draw is negated. Whatever happens in the two strategic rounds, in the end the two cards of the Dealer and Receiver are compared and the winner is whoever has the higher card. The cards are ordered ace low/king high and Dealer wins ties. One might expect that because the Dealer wins ties, the game is in the Dealer’s favor. However, in fact during minimax play, this game is actually in the Receiver’s favor. By always trading low cards and trading middling cards often enough, the Receiver can make it difficult for the Dealer to have a high card.  The Receiver's strategy is an inundative strategy which aims to simply outwait any streaks of kings dealt to the dealer. Because the game is simply “draw the highest card” with extra steps, it can be played very rapidly. I had my students play this game in a statistics class and record their results. It was noticeably difficult for the students to compensate for initial runs of luck, as expected by inundative strategies. Contrariwise, if the dealer aims to win by trick shuffles and/or dealing (this would be very simple for a skilled card mechanic as only the top three cards are relevant to the game), that would be a compensatory strategy as he is changing the state of the system to affect the probability of winning.

The importance of inundation vs compensation to philosophy of statistics is shown by the Ramsey/de Finetti theory of Dutch Books. The underlying assumption to this method of justifying probability axioms is that any inundative strategy that violates the probability axioms has a perfect response, which guarantees loss. Thus the presence of arbitrageurs provides a Darwinian compensatory force which flushes out inconsistent actors. For de Finetti and Jaynes, compensation is allowable structure but inundation is not.

By now it is clear that whether a system works by compensation or inundation obviously has something to do with what we can and should do with statistics. A good engineering metaphor would be the difference between measuring forces which can change an outcome and measuring viscosities which merely slow an outcome down. But what we can and should do with statistics is not a function of inundation and compensation. This is why there is a second dimension to deal with: the dimension of goals.

Y-Axis: Description vs Prediction

To reiterate: the distinction between compensation and inundation just defined and illustrated is a difference in the action of a system. The time has come to define and illustrate the x-axis distinction between the goals of the experimenter: prediction and description. These will be vital in examining what we can and should do with statistics. Having just arrived at the floor, we will carriage return back to the top and go from left to right in defining and illustrating the concepts.

Y-Axis, Part 1: Description

“The data of an experiment consist of much more than a mean and its standard deviation. In fact, not even the original observations constitute all the data. The user of the results, in order to understand them, may require also a description or reference to the method of investigation, the date, place, the duration of the test, a record of the faults discovered by the statistical controls, the amount of nonresponse, and in some cases, even the name of the observer.”

• W. Edwards Deming, “On Probability As A Basis For Action”

If the goal is description, then what to do is to maximize the information content in the interpretation of data restricted by the need for the interpretation to be true. The information content of a hypothesis is what an inference tells you not only about how a system is, but also how it could be, thus Deming’s dictum above that the measurements do not exhaust the information content of the experiment. This goal is often called an “explanation” framework because the information content sought is information about the causes of the system. However, I prefer the phrase “description” because the goal - translated into a propensity framework - is to accurately describe the dispositions of a system, including the probabilistic ones.

“The old doctrine of modals is made to give place to the numerical theory of probability.”

• Augustus De Morgan, Formal Logic

Now, cause is a modal concept. If the experimenter has a descriptive goal, then their use of probability and statistics are substitutes for the uses of modal logic in medieval philosophy. Where Thomas Of Aquino might use modality to judge whether or not a former prostitute can be saved, a modern crisis counselor uses statistics.

This connection between modality and statistics is as old as probability theory itself. The very term “probability” was introduced by Antoine Arnauld - or, more probably, a ghostwriting Blaise Pascal - in the last chapter of Ars Cogitandi. Arnauld - or Pascal - adopted the term from an infrascholastic debate. The so-called “probabilists” had claimed that if the application of divine law was unclear on a subject, one could follow conscience, even to the least probable possibility. One must realize that at this time “probability” means “acceptability” or “social acceptability”. Thus Gibbon can say that “Such a fact [that the Persians faithfully discharged their agreement to supply provisions to the Roman army] is probable, but undoubtedly false.”. Similarly Hume discusses not probability, but the probability of chances. We would now say “probabilities accepted by the masses”.

Catholic probabilism means that if Church doctrine can be interpreted in a way than that way is acceptable, limited only by coherence with doctrine generally. This evolved out of the Roman concept of precedent and into the modern concept of precedent. Now, probabilism is a vital part of Catholic philosophy. As Adrian Fortescue put it, Catholic “unity consists not in a mechanical uniformity of all [the Church’s] parts, but on the contrary, in their variety”. The doctrine of probabilism allows this, and cybernetics gives us good reason to understand the need of a social system to encompass variety. 

However, this kind of probabilism easily leads to mere casuistry. In the 17th century, the Jansenists attempted to halt this decline with a vigorous attack in the opposite direction. The greatest of the Janesists, Pascal (the analysis is Pascal’s even if the writing is Arnauld’s), offered a complete replacement of the Jesuetical logic of “possible interpretation = permissible interpretation” with a logic of dominance in expectation.

“Those who come to this conclusion [the logic of dominance in expectation], and who follow it out in the conduct of their life, are wise and prudent, though they reason ill in all the matters of science ; and those who do not come to it, however accurate they may be in everything beside, are treated of in the Scripture as foolish and infatuated, and make a bad use of logic, of reason, and of life.”

• Antoine Arnauld, Pierre Nicole (and probably Blaise Pascal), The Logic, Or The Art Of Thinking, Chapter XVI “Of the Judgment which we should make touching Future Events”

At the apex of this mode of thought, Hume defined a causal link between two objects as the presence of the second object being incompossible with the absence of the first. In other words, a counterfactual absence of the first object strictly implies the counterfactual absence of the second object. In more readable terms, the presence of the second object strictly implies the presence of the first. Hume holds this definition of to be immediate with a frequency notion of experiment: constant conjugation. Interestingly, when Hume actually defines probability, he is a strong subjectivist. As Richard Jeffrey put it: “David Hume’s skeptical answer to those questions [of the seeing the existence of objective chance] says that chances are simply projections of robust features of judgmental probabilities from our minds out into the world, whence we hear them clamoring to be let back in.”.

Ontology aside, this Hume/Pascal view does not truly capture how the probability-based information content of a hypothesis replaced the old approaches to modality in reasoning between choices. The first person to distinguish the information content of a theory and its probability was Karl Popper, who distinguished between degree of confirmation and probability in Logik Der Forschung. The dog that did not bark which shows that Popper originated this distinction is that Poincare is unaware of it, as he talks instead of maximizing the probability of a theory. One can see Gillies’ Objective Theory Of Probability, Introduction, Part vi “The Two Concept View” for more details on the history. However, the distinction between probability of a theory and the confirmation of a hypothesis is not peculiar to Popper or objectivists about chance, as an excellent Bayesian analysis can be found, for instance, in Chapter 6 of Betting On Theories by Patrick Maher.

To go deeply into the history of how descriptive statistics evolved out of modality would take up a whole book, namely Hacking’s The Emergence Of Probability. I don’t mean here to reproduce any of our illustrious predecessors' particular analyses but only to illustrate the distinction between information content of a hypothesis and probability of a theory. Maher quotes Cavendish as a typical example of scientific description: “We may therefore conclude that the electric attraction and repulsion must be inversely as some power of the distance between that of the 2 + 1/50th and that of the 2 - 1/50th, and there is no reason to think that it differs at all from the inverse duplicate ratio.”. The point estimate of 2 is not merely the center point of the interval estimate, but an assertion of conjectural (in the sense of Popper) confirmation of the hypothesis that the electrostatic force varies as the inverse square: extremely informative but with an a priori probability of zero. The pair of statements thus balance information content and probability sufficiently for Cavendish to write them.

Y-Axis, Part 2: Prediction

“Whenever I wake up with a headache, my wife tries to figure out why I got a headache: did I drink too much, did I stay up too late… I just take an aspirin.”.

• George E P Box, “Rethinking Statistics For Quality Control”

So much for probability as descriptive modality. Let’s now move on to “prediction” as a goal. Returning to Arnauld/Nicole/Pascal’s we see the logic of dominance has something to do with behavior. Though the logic is a replacement for non-numerical modality, the motivation has nothing to do with the description of modal properties such as cause. In fact, it is generally true that when one says one wants “prediction” in statistics, what one really means is a guide to best behavior. The goal of an archer is to aim where the enemy is going to be, not to wonder why the enemy is running there.

Typically, one develops a parametric model of the situation. We shall return to the question of what a parameter is later, but what matters now is that parametric models reduce the space of description to the space of parameters. For Jaynes, the hallmark parametric model is a physics model which describes all relevant interactions, like the example of Cavendish above. The parameters which in the archer’s case describe the motion of the enemy, might in Box’s case be numerical descriptors of the level and distribution of pain in the head. The agent decides on a function - perhaps even a random one - of the parameters that then dictates their own behavior. The parameter values chosen then are a function - perhaps a random one - of the data observed. This allows the behavior of the parametric model to be a function of the observed data, which is desirable in certain instances.

We can now see von Mises’ critique of Kolmogorov’s structural theory of independence as implicitly predictivist, thus failing to achieve Kolmogorov’s desired neutrality. Though the fair die makes primality and smallness independent in Kolmogorov’s sense, primality and smallness could have been dependent (if for instance the die had a different number of faces). Richard von Mises’ commitment to empirio-criticism highlights descriptive rather than predictive statistics. 

Kolmogorov’s predictivism, such as it is, doesn’t seem to have been intentional. But there are in fact pure predictivists in statistical philosophy: Bruno de Finetti and Jerzy Neyman for two. Probability and statistics are vital for a predictivist. As shown in the article on the game of La Her above, generally the best behavioral function of data is a random function. Further, the random function of behavior can be a random function of a random variable. This random variable can have more information than a single number. For a simple example of behavior as a random function of a random variable type use of statistics, we can check de Finetti:

“[Brier’s Rule] is used, for instance, by American meteorologists to estimate the probability of rainfall. And in the morning — so I am told — those estimates are published in newspapers and broadcasted by radio and television stations so that everybody can act accordingly. To say that the probability of rain is 60% is neither to say that ‘it will rain’ nor that ‘it will not rain’: however, in this way, everyone is in a better position to act than they would be had meteorologists sharply answered either ‘yes’ or ‘no’.”

• Bruno de Finetti, Philosophical Lectures On Probability, Chapter 1: “Introductory Lectures”, Section 4: “Proper Scoring Rules”

The goal of linking behavior to data functionally differentiates predictive from descriptive statistics. Counterfactual considerations, necessary to consider cause, are ruled out. For example, it is perfectly good experimental design for a predictivist to continue an experiment until the p-value is below a certain level:

“Suppose I tell you that I tossed a coin 12 times and in the process observed 3 heads. You might make some inference about the probability of heads and whether the coin was fair. Suppose now I tell that I tossed the coin until I observed 3 heads, and I tossed it 12 times. Will you now make some different inference?”

• Anonymous, “The Likelihood Principle”, Wikipedia

If behavior is a function of the data, then it is a function of the data. It is of no matter what the intention behind the experiment was. One should not make a different decision in the second case. If the readings of a probe do not go off the scale, then the statistician does not have to worry about censoring, even if they could have gone off the scale. But if one wants to describe causes, such as whether the coin is physically deformed, then the second experiment is different as the number of heads is predetermined. Cause is a modal concept after all. The second experiment is not an experiment designed to maximize the information content of the number of heads thrown. 

At this point we have both distinguished the types of system one encounters and the goals that one has when encountering a system. The principle of continuity complicates all distinctions: a system may have compensative and inundative aspects and one’s goals may be a mixture of prediction and description. Of course, these distinctions are only a step on the process of figuring out how our table will come to be. The distinctions must only be sharp enough that we can arrive at a determinate table which can help us see how we can understand and use statistics.

A Matrix And Its Morals

We have defined and illustrated our axes so that the content of the matrix promised in the first section can be cashed in. Perhaps you have already been able to intuit what the contents of the matrix would be. If so, that would be a credit to the importance of the axes. But regardless of intuitive powers, it is time to examine the content of the matrix. We will start with the point (Description, Inundation) and work clockwise.

(Description, Inundation)

When a process is inundative and the goal is to describe the process, one should randomize experiments to estimate constants. There are two parts to this process: Randomization and Constants. First it must be shown how randomization as a part of design delimits our ability to make firm descriptions of inundative processes. Then parameters will be distinguished from causes and forces in order to understand why they are the correct descriptions of inundative processes. 

To start with randomization, as stated in the previous section, the goal of predictive experiments is to the extent possible make behavior a function of data. Thus randomization of experiment is sheer nonsense. If one is designing an experiment to distinguish between Proper Victorian Ladies (PVL) and Time Travelers In Disguise (TTID), one could use whether the subject can truly distinguish a proper cup of tea. The subject is fed tea from an arrangement of cups. Fisher would recommend the cups be arranged randomly, perhaps drawing from a deck of well-shuffled cards and serving a hot cup on a red card and a cold cup on a black card. Let’s say one drew Red, Black, Black, Red, Red, Black and she correctly labeled each cup. You have successfully found a PVL.

Jaynes would argue that Fisher and you are merely pretending to not understand how shuffles work. A shuffle is a deterministic physical process that determines the state of the deck. The outcome is a (possibly random) function of the state of the deck and the state of the lady (PVL or TTID). Thus behavior is a (possibly random) function of the state of the deck and the state of the lady. If the goal is to determine whether to let the lady go or send her for further interviews and Red, Black, Black, Red, Red, Black gave a particularly successful behavior function, then why not keep to that pattern every time? What is the point of pretending to be bad at cards?

Of course, the answer is obvious. Fisher’s goal is not to make behavior a function of data, but to truly describe the state of the woman, PVL or TTID. The design of experiment, therefore, must be such that considerations beyond what is observed are accounted for. The description of the propensities of the system are causal and thus modal. By design, the state of the woman is not varied through the modal space of possible experiments, but the state of the deck is. The experiment, and its repetitions, is thus designed to capture that unvarying part of the experiment. The experimenter must care not only about the present data, but the counterfactual data as well.

Randomization is one side of Deming’s Dictum that observations from an experiment do not constitute all of the data. The flip of that modal coin is that one cannot make firm descriptive statements of inundative processes unless the experiment is designed before the data is collected. What is allowed to wander and what is fixed across possible worlds is a part of the description of dispositions of a system. If one simply comes across a vat of data without considering the propensities that are not expressed, then the description cannot be successful. Thus Jaynes’ perhaps successful description of Rudolf Wolf’s dice based on data is, in this view, due to his implicit modal assumptions that the dice are not varying over the 20,000 rolls but only the initial conditions of the dice upon release. This can be seen in his more detailed earlier analysis, where he considers the various physical features of the dice as “constraints”. Jaynes states boldly that his entropy approximation to chi-square tests is “combinatorial, expressing only a counting of the possibilities;” (emphasis in original) which is to say modal. This also helps show that these considerations are not dependent on frequency vs Bayesian preferences but really are motivated by goal and system.

Such is randomization, but what about constants? Partly we have already covered this: a constant is what is held still through the possible worlds described by the interpretation of an experiment. They are the propensities that characterize a system as an object in modal space, the replacement of what used to be called the essence. Constants are not necessarily a parameter or a statistic which can come from an experiment. Nonparametric methods, like the jackknife, allow us to describe a constant of a system without assuming much about the shape of the probability curve, so a constant cannot be a parameter. Similarly, a statistic like a p-value, might tell us a lot about a constant of a system, but is rarely the constant of interest itself.

It will help to see how inundation relates to constants rather than causes or forces by a famous attempt to do without inundation. In the early 19th century, it was an item of faith among scientists that highly sensitive, very controlled mircoexperiments would be virtually free of noise. In Laplace and Airy, the law of errors is motivated by the model of an enormous number of independent sources of error. This led ultimately to Maxwell’s Demon: a measuring probe smaller in scale thermal fluctuation and yet so sensitive that it could filter incoming molecules by their speed. 

The point can be made by returning to the earlier example of Sinai billiards – a ball bouncing about and reflecting off the sides of an idealized billiard table with an appropriate central obstacle. As stated, the dynamics of the ball are chaotic. Let’s say that you want to get the measured average free path (i.e. distance to the next wall) as small as possible. You set up a film camera with a variable frame rate. A computer will attempt to raise the frame rate when the ball is near an obstacle and reduce it when it is far from an obstacle, which it can learn from the photographs. The first difficulty comes from the fact that if the rate changes continuously, then you have to start lowering the rate before or after striking the wall, either of which limits your ability to lower the sample average. But once your frame rate is low then you have little information about where the ball is right now, meaning you have to take a risk when raising the rate. Over time, the graph of the frame rate will look more and more like white noise. The accumulation of errors and the limited power of each success will drag the observed average free path towards the mean free path. This is the well known phenomenon of decay of correlation.

What statistics can reveal is the rate of decay of correlation, which is a function of the geometry of the system. That is, the rate of decay of correlation is constant over almost all possible initial conditions. The statistics are not for revealing the state of the system, as Maxwell assumed. That is to say, the goal is not which possible world (initial condition) you are in but what is constant through the possible worlds. Maxwell’s mistake was mistaking an inundative process for a compensatory process.

(Description, Compensation)

But what about compensatory processes? If our goal remains description, then we should still use experimental designs that take into account modality. But now the form of modal and counterfactual reasoning that we care about isn’t just stability across possible worlds but identifying causes: our measurements make what objects “necessary”?

In terms of the Sinai billiards above, we know that the obstacle causes the decay of correlations,  in the sense that the presence of decay of correlations is incompossible with the absence of the obstacle. But I think analogizing the process of estimating constants with the independence of initial conditions and cause as dependence on geometry is simultaneously too unintuitive to be enlightening and too precise to be useful.

So I will switch to the place of Fisherian statistical practice and its chief descendents, biometrics and scientific quality control. The lady-tasting-tea example described above had the goal of describing causes, that is characterizing the functional dependence of dispositions of processes on conditions. The point of examining treatments of crops in factorial designs was not to find which of the treatments was the most successful in order to fast-track its adoption into agricultural practice. Instead, the goal was to figure out the causes of successful treatment which could then be improved on as treatments approached agricultural practice. We care about counterfactuals because practices are meant to be iterated, thus we care about conditions outside the experiment.

Fisher illustrated his concept of descriptive biometrics, in essence the generalization of crop experiments to nature, with the example of Johannes Schmidt and the European eel. Since ancient times, the life cycle of eels has been shrouded in mystery. Nobody had ever seen a young eel (immature eels had been misidentified as a different fish), nobody had ever seen an eel die of old age (eels have had observed life spans >150 years). A generation before Schmidt, Giovanni Grassi had shown by gross anatomy that the so-called ‘glass eel’ was in fact eel larvae and proposed that eels bred in Mediterranean bays. Schmidt took decades of vertebra and myomere measurements and found the histogram of measurements was near independent of where the eels were caught. This is unusual: fish born on opposite sides of the same lake often have distinct myomere measurements. Convinced that European eels have a single breeding ground, Schmidt began tracking that location by tracing the distribution of glass eels. As he went up the currents of the Atlantic, the average observed length of the glass eel declined. The place of reproduction - that is, the place where observed length of glass eels approached a minimum - is in the Sargasso Sea.

Schmidt and the eels are a perfect example of how statistics empower the description of the dispositions of systems that allow scientists to trace down causes. The regularity of glass eel myomere measurements is incompossible with glass eels having scattered origins, thus the single origin causes the glass eel myomere lengths (see Hume’s definition of causation above). If a glass eel wanted different myomere lengths, then it would have to have been born in a different location. And indeed this is so: American eels are born southwest of European eels.

This example also happily demonstrates the assertions that inundative and compensative processes are not perfectly distinct. Darwin taught us that the compensatory processes underlying species stability occur within a larger scale inundative process which is now called evolution. I will not speculate whether the speciation of European and American eels happened by natural selection (the accumulation of small genetic advantages) or by genetic drift of eel families in the two breeding centers. I do tend to be a follower of Sewall Wright in these matters, emphasizing multiple local maxima in the genetic fitness landscape which are reached by drift as much as hill climbing.

Before moving on, I want to draw one last moral from the biological example about the iterative nature of descriptive experiments. Fisher had a very distorted vision when describing the dispositions of one particular species he was fond of. For example, he had scientifically indefensible opinions about their propensity to form cancers late in life (his son-in-law and daughters did not share this opinion). We don’t have to listen to those wrong opinions, as we have iterated past them. Caring about iteration means worrying about counterfactuals. One of the chief advantages of the descriptive mode, the coin that rewards worrying about counterfactuals, is that a description can be wrong.

(Prediction, Compensation)

Prediction means choosing a function that relates one’s behavior to one’s data. If a process is compensative, then the data is a function of underlying forces. Unlike a description, behavior cannot be wrong=incorrect. Behavior can only be wrong=immoral. Before we come back to this point, I want to talk a bit about a group of people who thought this quadrant was the only place to be, the Bayesian revivalists. This will give us a firm footing to see the attractions and dangers of this location.

When one talks about the Bayesian revival - especially de Finetti and Jaynes - one often notes the importance of ‘determinism’ in their thinking. Jaynes and de Finetti present statistics as precise thinking in a world where all processes are compensative and all thinking is determining action. One might phrase a de Finetti flavor of determinism as saying all behavior is a function of underlying forces, but usually a random function.

The behavior is a (possibly random) function of the data, a (possibly random) function of the underlying forces. I chose to characterize the underlying facts as forces rather than constants or causes to emphasize the non-modal nature of this mode of thinking. The definition of force in modern physics is the rate of change of the quantity of motion. Newton originally defined the total force on an object as the curvature of the path of the object, with rectilinear forces treated by a limiting argument. This is a function of the actual motion of the object. Then he employed physical arguments to deduce the forces emitted by objects follow rules now called vector addition. Thus the individual forces are a function of the actual kind and actual arrangement of force emitters. But it was Euler who first developed the venomous part of Newton’s concept: forces are not merely functions of the actual process but deterministic functions. Other maximum principles in dynamics - such as the stationary action principle or the principle of least constraint - suggested modal readings, the principle of maximum effect of force suggests a modal skepticism. Counterfactuals aren’t merely alternate but also wrong.

One can make a loose link between this view and de Finetti’s Paretan fascism. For Pareto, the measured extreme inequality in post-war Italy was a function of market forces, which is to say efficiency itself. Competition at the top level of economic inequality merely engendered slow circulation of those near the same economic scale. Later, de Finetti would try to mute the extreme implications of Pareto’s theory by introducing the idea that a corporatist government could act as a mild countervailing force by considering symmetry in addition to efficiency. These considerations would lead de Finetti to join the left-libertarian Radical Party after the collapse of Fascism.

The incoherent politics of professorial weirdos aside, the goal of making one’s behavior a function of underlying forces is one we see increasingly prevalent. Social media, text and image archives, etc. are constantly generating stuff which may be euphemistically called the Data. Companies and governments want to use the Data. All the stuff Deming mentioned as useful to the quality engineer - “the date, place, the duration of the test, a record of the faults discovered by the statistical controls, the amount of nonresponse, and in some cases, even the name of the observer” - may be well outside the Data. Further, the conclusions drawn from the Data must be functionally related to the data for the conclusions to be drawn by a computer. Statisticians used to rail against cookbook statistics, now they must develop techniques far more rigid than any cookbook.

In my opinion, the rise of Bayesian Ideology is not because it has better arguments now than in the 20th century. If anything, the arguments are worse: the rise of Markov Chain Monte Carlo and related methods have made the beautiful functional structure which attracted Jaynes and de Finetti in the first place rather ugly. Instead, I believe the rise is due to the fact that Bayesian Ideology, deliberately or not, gives license to the practice of putting functions on vats of data and calling it inference.

I don’t say all this to bemoan the decadence of  post-postmodern society. The issue is that behavior functions can only be wrong=immoral not wrong=incorrect. The classical theory of errors in Airy etc. conceived of uncertainty as caused by uncorrectable mistakes in measurement. But a bankrupt firm did not make mistakes, they revealed their risk preference was not to miss profits over passing their survival constraint. Nor is this a stupid inference as the recent bankruptcy of Pyrex shows. Their corporate owners clearly preferred inflated paper profits over long term survival.

(Prediction, Inundation)

“When the capital development of a country becomes a by-product of the activities of a casino, the job is likely to be ill-done.”

• John Maynard Keynes. The General Theory of Employment, Interest and Money. Chapter 12, “The State of Long-Term Expectation”

Before getting into this location, let's consider the success of the simple analysis of the bankruptcy of Pyrex. The success is due to the fact that the bankruptcy of Purex reflects clear market forces. It wasn’t an inundative process of poorly chosen behavior functions. Cornell Capital bought Pyrex’s parent company, Instant Brands, and loaded it with volumes of debt that quickly became unsustainable as interest rates rose.

But what about behavior functions on inundative processes, that is to say this location? We’ve actually already discussed one, the frame rate of a camera being a function of the position of a particle in Sinai Billiards. We saw that the behavior of this function would inevitably become just white noise. Earlier, we called James Clerk Maxwell’s assumption that he could make a function of an inundative process reflect his preferences a “mistake”. The pawl of a brownian ratchet simply becomes more white noise in the system without organizing the behavior of the gear in accordance with its preferences.

The success of applying algorithms to the Data depends on the nature of the process generating the Data. Yet often in machine learning and other applications of modern statistical methods, we are not improving our knowledge of the nature of those processes. This choice is affected by market incentives. As Keynes said long ago, the predictive and inundative nature of stock market speculation means that “[t]he measure of success attained by Wall Street, regarded as an institution of which the proper social purpose is to direct new investment into the most profitable channels in terms of future yield, cannot be claimed as one of the outstanding triumphs of laissez-faire capitalism…”.

Management guru Nassim Taleb further dramatized the folly of predictive methods in an inundative process by highlighting cases where a few large outcomes dominate any sample. This is related to the case of the Sinai Billiards/Brownian Ratchet convergence of the control to white noise, but with 1/f noise. The point made here is that white noise management is already bad, even without worrying about 1/f noise.

Here is an example of white noise management in quality engineering, taken from a lecture by George E P Box. Unfortunately because this is proprietary data owned by Ford motors, it was left out of the printed version. The x-axis is time from 1972-1982 and the y-axis is the count of truly pathetic lemons, cars returned in the first six months of purchase. Each unit height represents a car with severe quality issues, a part with complete mechanical failure. You can see the frequency is pretty stable over the decade. Six month returns were rare even in the pre-consumer protection era. Some bad parts are gonna escape the factory. This is an issue for insurance, not a target for quality engineering.

Now we see the same data with Japanese manufacturing overlaid. One can see both the low initial level and the sharp drop in cars returned in the first six months of purchase. The drop, at least, is because the Japanese manufacturers took an descriptive & iterative approach to quality engineering. Interestingly, the Japanese did not try to upscale their output, but rather continued to pitch their product as practical economy cars. As consumers learned higher quality was available at lower prices, the Japanese automobiles began dominating the global market. While 1/f noise management is spectacular in failure, it was white noise management that brought American automobile manufacturers to the malaise era.

I do not mean this to be a criticism of Taleb’s beliefs, as the concept of “quality Japanese automobile” was as absurd in the 60s (in the US and Japan) as a “black swan” was to Aristotle. Rather, I agree that use of the appellation “mistake” for the application of predictive methods to inundative processes is justified.

And this time, I do say this to bemoan the decadence of post-postmodern culture.

thing is, the tech (driven by demands from VR) wants to drive forward towards “spatial audio”, where what’s created is a sculptural audio “environment” where particular sound sources are given Locations. Video game players like this, for sensical reasons: the higher fidelity the aural location of an event of interest (enemy, animal, river, etc), the better a skilled player can rely on it as a signal, whether in terms of an increase in spatial fidelity in inner representation, or in terms of a lower error rate for signals on that channel. 

in terms of personal sound, we are thinking of the spatial audio on the airpod maxes, and in terms of room sound, we are thinking of the kind of wavefield synthesis that recent developments in so-called AI have made more plausibly solvable (im thinking here of a wavefield synthesis rig that can take a LIDAR reading of the space it is in, and then calculate dynamically how the frequency response of the room sound will embed itself within the room, and introduce cancelling standing waves, doing a kind of “noise cancelling” for a closed room of arbitrary contents. would be just a fucking lot of convex optimization in the frequency domain but that’s what NVIDIA’s for now right??)

thing is, this creates an interesting challenge for people who make audio-based art. the stereo field is two channels that represent the hardware environment into which many (hundreds, sometimes) of microphone channels are fed down into. right now, state of the art is figuring out fun ways to do “trompe le œil” while projecting those n channels onto a two-channel space. when you do this, you’re summing a bunch of the channels, so the trick is to kindof convert the space of possible variation into one that encodes the most meaningful aspects of the original group of signals. The problem though, is that you’re playing a game against the tensions of constructive and destructive interference. so you have to have technical know-how of how signals combine, how they can be cut to fit one another like building a brick wall, BUT ALSO social know-how to execute on questions of genre appropriateness (what kinds of bands are there, and how do these different kinds of bands answer the characteristic questions of recording differently, and in what way do audiences local to them understand that difference) BUT ALSO art know-how to see how the song relates to itself as projected into stereo recording. and so you use frequency-domain tools to most artfully reduce the number of channels going to the folks at home. which requires technical and social knowledge to execute on well. 

NOW. what spatial audio does is actually provide a way of consistently representing channel-size signals at psychoacoustically identifiable locations, due to the construction of specific speaker systems combined with tuned filters to move signals in three dimensions in the stereo field, rather than two.

so, at the level of the ontology of recording, you can now have the construction of a 3-d space taken care of for you through pre-made technological aspects that the composer can count on the standardization of, as well as a system that can be used to “place” signals which retain their uniformity as single channels by having identifiably unique locations within the stereo field.

this gets rid of the “trompe le œil” but it also makes things Way more about the quality of the initial signals, rather than the fitting-together of signals to achieve a unity of purpose. or rather, the unity it achieves now is one whose goal has to be its creation in the mind of the viewer, rather than on the opaque 2D of a stereo recording.

THE THING IS: this means that a lot of genre techniques that have to do with achieving specific and socially-identifiable genre-effects through the use of a trompe le œil in mixdown can’t plausibly take advantage of the new technology. what it would do, to retain sonic genre, would be to figure out what was meant to be achieved by the social recognizability of those aspects and then work out how to achieve those things within the new technical space.

I would like to see someone like Dan Lopatin actually take on the challenge of composing specifically within this medium, but in a way meant to communicate something external to existing questions of music “genre” that the prevention of widespread sampling thru IP laws prevents. 

I think Magic OPN kindof abdicates on it, because lopatin just takes us on a genre walkthrough of things he’s anxious about or things he enjoys remembering as a way of showcasing what the tech can do. that’s helpful, but it’s more of a demonstration record than a real artwork. He mostly sounds lonely and confused and looking for some reminder of what makes him who he is. Bums me out.

HOWEVER, I also particularly believe in his ability to do that, given that Garden of Delete executed on the whole 90s khaki computer peripherals as techno-desert storm II death as suburban backyards where mp3 players were just invented. the genre work accomplished the identificatory thrust of the music’s meaning. i don’t think anyone else “got” what 2004 was like for a certain kind of boys in the way that Lady Bird “got” 2004 for a certain kind of girl.

like, nothing in culture is straining against its inability to become sound-sculpture, so the operation of genre won’t produce the form, which is why we would need someone like Dan Lopatin to take the anticipatory step of announcing and explaining the goal.

the opportunity now exists to take a sculptural approach to audio, but we don’t have anything to point to as examples of even attempted sound-sculpture, simply bc the tech has never existed, so there’s no sense of how to understand something produced as one unless it’s announced as one, because no genre will grow towards it: the sculptures they produce will be sculptures of paintings with the traditional kind of painting painted on it. this is something that is like the problem of the fourth wall, but weirder.

the core idea is that the details of specific channel-points need to be managed as the point where “meaning” is generated in order to maximize the compositional opportunities presented by the form. a rock band is about producing an effect as a whole that recordings then represent. the thing is, it’s a small number of pieces. at the bit depth we are able to do now, and with some decent ear training, you really can hear all the way into this stuff. the problem is, we just don’t really have anything obviously ready to be composed in such a way as to take advantage of new tech like this.

basically what i’m worried about is that musicians are not clever enough to know how to use this compositionality and video game people are too artless to make something worth experiencing this way, even if they know the tech better.

we see a similar problem in 4k so far. people mostly use it to make smoother, sharper versions of the thing they were already going to make. in places where nature fills in all the details —so any 4k nature documentary —the tech itself is able to stand out bc the source material in some sense is the same way as it’s depicted, and is as detailed on as minuscule a level.

the problem comes in when this stuff is used for narrative: what happens is the details all get sharper than any of the detail-makers care to pay attention to. the fidelity of the medium is higher than the imagination of the filmmaker, and you get the inclusion of endless meaningless details that undermine the work by not presenting detail for its own sake or detail for the sake of the film, but rather Accidental Detail. this is ugly and dispiriting—even when people get what the problem is, they end up resolving it by “having nature take care of it” as in the hobbesian theodicy of Triangle of Sadness. it’s a cheap cop out, but at least it’s the characteristic cop out taken by incipiently-monarchistic critics of liberalism’s failures to have already reached its goals, so it’s funny in context, if icky.

the way we know this is true is because we have an example of this fate being avoided by someone whose imagination tries actively to adjust itself to the fidelity of the medium he is working in: David Lynch. he says all over the place that with each film, he’s looking at the state of the art, and then he proceeds to engage the state of the art tech in a painterly manner, by asking what of its subject it can represent uniquely that past media couldn’t.

this was expected to be a huge problem with Twin Peaks s3, since the soft-focus VHS of the first two ceilings is not only beloved, distinctive, and well-understood, but was also in the heights of a supercycle of fashionability when season three came out.

thing is, season 3 makes completely clear from the outset that Lynch is aware of how much information goes into the amount of space made available by 4k. i mean just look at it, im sure i’ll have to give an actual argument for it, but every blade of grass gives the sense of “having been paid attention to” in terms of how it’s represented.

this is even better, because the time distance allows the distance in media of representation to express themselves in terms of the distance in time between seasons one and two and season three. in the first two seasons, one was working in analogue, filming fields of light of different color, with the intention of projecting those fields of colors to some different size in a home television. everything is floating and hazy, and made more really so by looking like what representations of that era looked like to people in them, no less than the decor in the houses. in the third season, by contrast, you can tell lynch knows that he is filming 8,294,400 individual pixels, specific determinate spots with specific determinate details. this is even truer with an OLED t.v., where each individual point of color also produces it’s own light—earlier LED tvs relied on a common light field refracted in different colors. and the things he films that way are the things folks are now filming poorly with them, but in a way that is composed, and also reflected in the sort of transition from analog to digital that is happening over the course of the series as a whole.

the apotheosis of failing to do this ought to have been some kind of Outer Space B Movie, if we were still keeping up with tradition, but instead it’s the marvel films. I can’t think of a firmer example of not paying attention to how the details and the whole integrate into a narratively satisfying unity as the Marvel Studios practice of filming actors and actresses delivering lines on green screen while a team unrelated to the team selecting lighting for the shoot meticulously draws out in whatever virtual AutoCAD they use all of the details of the explosions and distant sky and mountains.

the only rock and roll artists i can think to have plausibly addressed this issue would have been steely dan, who also made sure to operate at the forefront of how different ways of using the existing technology could “mean” differently. they say as much about Gaucho, and the icky 2000s brickwall L2 gloss of Two Against Nature really solves it. it’s basically the audio equivalent of Inland Empire.

At the heart of it, there is some cool new tech and as a critic, I’m worried our auteurs aren’t up to the task.

• Walter Bagehot, “Adam Smith As A Person”, 1876

According to friend and disciple Duguld Stewart, the philosopher and economist Adam Smith had, on his deathbed, regretted that he had written so little. This might surprise anyone who is aware of the heft of Wealth Of Nations (WON). For Smith WON was merely the pendant of a vast project which would range over all subjects in their historical and analytic aspects: “the immense design of showing the origin and development of cultivation and law; or, as we may perhaps put it, not inappropriately, of saying how, from being a savage, man rose to be a Scotchman.” as Bagehot phrased it with his inimitable wit. 

Fortunately for the devotees of Baghot’s “historical method”, Adam Smith was a great celebrity in his time. We can make use of John Rae’s classic biography Life Of Adam Smith as well as Bagehot’s witty article and some other sources lightly sprinkled in. Rae’s book is certainly the best way to absorb Smith’s complex personality. From Rae, we learn that - unlike Hume who adopted French cuisine - Adam Smith never abandoned strawberries, haggis and cock-a-leekie soup. We can also learn that though Smith was absent minded and had no mechanical dexterity, he was an expert card player. Smith, whose religious beliefs are still a matter of debate, regularly sung self-penned Latin hymns to the Divine Geometer but was unable to rhyme in English. “When asked about Shakespeare Smith quoted with apparent approval Voltaire's remarks that Hamlet was the dream of a drunken savage, and that Shakespeare had good scenes but not a good play; but [his unnamed interviewer] gathered that he would not permit anybody else to pass such a verdict with impunity, for when he himself once ventured to say something derogatory of Hamlet, Smith replied, ‘Yes, but still Hamlet is full of fine passages.’.”. 

Wherever we look, we can find plenty of material for today’s post. Today, we will look through the sparse but beautiful field of his smaller works organized by the major centers of his life. We will not here enter into casuistry over The Wealth of Nations, there being enough of that everywhere else. Hopefully you will find some freshness yet in Smith on this, his 300th birthday.

The Scottish Years & Background

Adam Smith was born to Scottish parents in Kirkcaldy, Scotland on June 16th, 1723. But we also learn of the intellectual world of Glasgow in the 1750s, now famous the world over. Smith befriended both Whig bishop Frances Hutchinson and Tory deist David Hume. What united the odd trio was their love of scholarship. Hume famously kept Virgil under his pillow, where Smith kept Hume. Enlightenment Scotland was Hume’s home and the influences of that world lasted his entire life. Rae notes in his biography that the last sentences Smith published were a tribute to his math teacher (which were added to Theory Of Moral Sentiments). He further asserts that Smith “is sometimes considered a disciple of Hume and sometimes considered a disciple of Quesnay; if he was any man's disciple, he was Hutcheson's.”. This evaluation is quite just, if one sees the matter from Smith’s point of view.

Yes, Smith had a great, even exaggerated love for the intellectual world he came from. Rae records Smith rating fellow Scots Stewart and Simson above his friend d’Alembert. Smith considered Hume’s History Of England the equal of Livy, this speaks more to his respect for Hume than his agreement with Hume’s famously Tory biased book.

In my opinion, the work of Smith which best displays his affections is his letter to William Strayhan on the death of Hume. Indeed, this may be Smith’s best writing in general. Hume’s cheer and good nature moved many readers. David Dalrymple was stirred to translate the letter into latin verse. In this passage, David Hume sings to Adam Smith of the ailment which will soon end his life:

“...Pol me moribundum, Adame, vocarea

Nomina si rebus dare convenientia quelles:

Altera jam morbo implicitum me detinet aeftas,

Ille quidem, vegeti quanquam quis corporis essem,

Difficilis, senium gravat, haud sanabilis ulla

Arte, Diarrhream medìci dixere Pelasgi.
Membrorum virtutem horae praedantur euntes,

Sensi equidem, et sedes inter vitalia fixit
Dira valetudo, nec te spes lactet inanis,
Adstat summa dies…”

The most noble part of this letter - Hume’s discussion with Charon - and its final sentence letter provoked quite a vicious reaction in late 18th century England. The idea that Hume could have died peacefully offended the principle that an atheist - which Everyone knew Hume was, that far more important thing than whether he actually was - must be in chronic fear of death. Critic, poet and lexicographer Dr Samuel Johnson scoffed at it. Dr Johnson confronted Dr Smith. People began to leave the party in droves. Eventually Johnson called Smith a liar to his face. Hearing about this shocked even Sir Walter Scott, no friend to Smith’s theories. Scott asked Smith 'And what did you reply?'.

“'I said, You are a son of a ——!' On such terms did these two great moralists meet and part, and such was the classical dialogue between two great teachers of philosophy."

• Sir Walter Scott, in Boswell’s Life Of Johnson, quoted in John Rae’s Life Of Adam Smith

Johnson quickly made amends for his outburst by sponsoring Smith for membership in The Club. Almost equally uncool in their response was a strongly worded letter by an unknown writer. 

While the logic of these reactions is non-existent, their strength is founded quite honestly in the depth of feeling of the letter. Bagehot warns us not to take the statement that Hume was “approaching as nearly to the idea of a perfectly wise and virtuous man, as perhaps the nature of human frailty will permit.” literally. What should be taken for this is that Smith’s praise is not just for Hume the individual but for The Scottish Enlightenment as an age.

Oxford & The Historical Essays

After being baptized in this Scottish intellectual world, Smith left for Oxford as a postgraduate student. It was an infamous catastrophe. In “Adam Smith As A Person”, Bagehot’s attempts to apologize for the state of English learning are accidentally hilarious in their infra-isle bigotry. The fact is that Smith went from hobnobbing with Hume to a university system without a single influential scholar. More modern attempts at a defense are even more pitiful, one writer - perhaps attempting to illustrate the paradoxes of quantification over an empty set - attributed to 18th century Oxford “the most significant traditions and innovations in the history of English music”. In a more honest moment, when ethnic antagonism was not an obstacle, Bagehot admitted that the Oxford of Smith’s time “was at the nadir of her history and efficiency.”. He allows one of Smith’s contemporaries to describe an unusually thorough doctoral examination: “I was asked who founded University College; and I said, though the fact is now doubted, that King Alfred founded it; and that was the examination.”.

As unlikely as the theory is descriptively, Bagehot’s theory provides an excellent excuse to fold in Smith’s historical trilogy. As printed in The Essays Of Adam Smith, the trilogy is

1. “History Of Astronomy”

2. “History Of Ancient Physics”

3. “History Of Ancient Logics And Metaphysics”

The three essays form a trilogy with History Of Astronomy, with each essay highlighting its connections to the previous. There are five other short essays in The Essays Of Adam Smith, of which the most interesting is “Considerations Concerning The Formation Of Languages”. In that essay Smith proposes that verbs were the original linguistic units, all others forming by a congealing process. Though necessarily completely speculative, it is in fact quite modern. David K Lewis’ notion of signaling in Convention proposes the same theory. The signals are acted on by the actors. “One if by land” in this interpretation means “Charge!” and two if by sea means “Fire!”. But there has to be some structure if this is ever going to be finished, so I will say no more.

Instead I will start with “History Of Astronomy”. This is the longest, most researched and best of the historical trilogy. There is no Enlightenment sneer in this history. Rather, Smith’s approach to philosophy of science is highly modern, as can be seen in this appreciation. Adam Smith begins with an analysis of the subjective reactions to unknowledge, which he divides into three categories: Wonder, Surprise and Admiration. Wonder is the most important emotional reaction to surprise as it is the most painful. In Ethica, Spinoza has the following to say about Wonder (“Admiratio”):

“Hæc mentis affectio sive rei singularis imaginatio quatenus sola in mente versatur, vocatur admiratio, quæ si ab objecto quod timemus moveatur, consternatio dicitur quia mali admiratio hominem suspensum in sola sui contemplatione ita tenet ut de aliis cogitare non valeat quibus illud malum vitare posset.”

• Spinoza, Ethica, Pars Tertia - “De Origine et Natura Affectuum”, Propositio LII - Scholium

Wonder is caused by the object of fear within a surprise, just as Adam Smith says. Admiration is caused by that which is beautiful in a surprise. Spinoza sees wonder as an impediment rather than a spark to science. The passio of wonder is apt to be answered by Peirce’s “method of tenacity”:

“If the settlement of opinion is the sole object of inquiry, and if belief is of the nature of a habit, why should we not attain the desired end, by taking any answer to a question which we may fancy, and constantly reiterating it to ourselves, dwelling on all which may conduce to that belief, and learning to turn with contempt and hatred from anything which might disturb it?”

• Charles S. Pierce, “Illustrations In The Logic Of Science I: The Fixation Of Belief”, Section V.

This analysis leads Spinoza to define Wonder as an emotional viscosity:

“Admiratio est rei alicujus imaginatio in qua mens defixa propterea manet quia hæc singularis imaginatio nullam cum reliquis habet connexionem.”

• Spinoza, Ethica, Pars Tertia - ‘De Origine et Natura Affectuum’, Affectum Defenitiones - IV

Smith, by contrast, prefers the traditional theory that links the passio of wonder to the passio of active inquiry. “Philosophy is the science of the connecting principles of nature.” says Smith. Wonder is the affective or inner half of a seeming disconnection. Thus it is for the relief if this inner tension connecting principles are tested. This theory can be compared to that of Plato’s Theaetetus. Socrates has just gone through a lightning fast elenchus. Theaetetus responds as follows:

“Θεαίτητος: καὶ νὴ τοὺς θεούς γε, ὦ Σώκρατες, ὑπερφυῶς ὡς θαυμάζω τί ποτ᾽ ἐστὶ ταῦτα, καὶ ἐνίοτε ὡς ἀληθῶς βλέπων εἰς αὐτὰ σκοτοδινιῶ.

Σωκράτης: Θεόδωρος γάρ, ὦ φίλε, φαίνεται οὐ κακῶς τοπάζειν περὶ τῆς φύσεώς σου. μάλα γὰρ φιλοσόφου τοῦτο τὸ πάθος, τὸ θαυμάζειν: οὐ γὰρ ἄλλη ἀρχὴ φιλοσοφίας ἢ αὕτη, καὶ ἔοικεν ὁ τὴν Ἶριν Θαύμαντος ἔκγονον φήσας οὐ κακῶς γενεαλογεῖν.”

• Plato, Theatetus, 155c-d 

The word for wonder here is θαῦμα, a word for emotions which covers both miracles and magic tricks. Which one Socrates did - and which one Plato believed Socrates did - is a matter of considerable debate. Still, Socrates sees Theaetetus as passing this test of the passio of wonder and thus continues to educate him. Theaetetus - a mathematical genius and perfect gentleman - can be contrasted with the arrogant cruelty of Dionysius II:

“ἡ μὲν δὴ πεῖρα αὕτη γίγνεται ἡ σαφής τε καὶ ἀσφαλεστάτη πρὸς τοὺς τρυφῶντάς τε καὶ ἀδυνάτους διαπονεῖν, ὡς μηδέποτε βαλεῖν ἐν αἰτίᾳ τὸν δεικνύντα ἀλλ᾽ αὐτὸν αὑτόν, μὴ δυνάμενον πάντα τὰ πρόσφορα ἐπιτηδεύειν τῷ πράγματι. οὕτω δὴ καὶ Διονυσίῳ τότ᾽ ἐρρήθη τὰ ῥηθέντα. πάντα μὲν οὖν οὔτ᾽ ἐγὼ διεξῆλθον οὔτε Διονύσιος ἐδεῖτο: πολλὰ γὰρ αὐτὸς καὶ τὰ μέγιστα εἰδέναι τε καὶ ἱκανῶς ἔχειν προσεποιεῖτο διὰ τὰς ὑπὸ τῶν ἄλλων παρακοάς. ὕστερον δὲ καὶ ἀκούω γεγραφέναι αὐτὸν περὶ ὧν τότε ἤκουσε, συνθέντα ὡς αὑτοῦ τέχνην, οὐδὲν τῶν αὐτῶν ὧν ἀκούοι: οἶδα δὲ οὐδὲν τούτων. ἄλλους μέν τινας οἶδα γεγραφότας περὶ τῶν αὐτῶν τούτων, οἵτινες δέ, οὐδ᾽ αὐτοὶ αὑτούς.”

• Attributed to Plato, Ἐπιστολαί, 341a-b

Dionysius II needs not be taught, he already knows... I am not certain there can be a better illustration of the difference between truth and certainty.

Coming back from Plato’s Greece, Smith advises that we look not to Plato but the presocratics: “the philosophy of Leucippus, Democritus, and Protagoras … successfully revived by Epicurus.”. Like most attempts to go back to the presocratics, Smith’s account is a projection of all the virtues that a philosopher should have, which he then easily shows to be far superior to those philosophers - chiefly Plato - who have the temerity to not be projections of ourselves.

Astronomical accounts that predate the presocratics are not remarked upon. In particular, Smith leaves out mythological and astrological sources such as the shield of Achilles and Odysseus navigating away from the isle of Calypso. Smith is of the school that observation leads to science, rather than “myths and the criticism of myths”.

Thus Smith builds a story where observation of the πλανήτης directly leads to the Eudoxan system, the theoreticians of the Pythagorean and Platonic schools providing - at best - moral support. The dismissal of the Pythagorean and Platonic schools reveals Smith’s source exactly: Aristotle. Aristotle criticized these schools so the role of our historian is to figure out why they are wrong, not what they were doing. Eudoxus was Aristotle’s beloved mentor, and so receives no criticism. Smith can be saved here only by noting that he was likely relying on older histories of astronomy unaware of their biases. In addition there is no original source of Eudoxus or any pre-Ptolemaic Greek astronomy to turn to anyway.

A much simpler story can be told if one avoid’s Smith’s Aristotelean fundamentalism. Archaic Greek writers were scarcely aware of the πλανήτης perhaps because they posed no problem for them. Orphic writers already conceived of the universe as a sphere surrounding a round earth (perhaps round like a table, perhaps round like a ball). Pythagoras proposed the theory of Music Of The Spheres not as a “wild and romantic idea” but as a serious proposal to study planetary motion by ratios. This new approach to star charts lead to the discovery that Φωσφόρος and Ἕσπερος are the same object for they have the same ratios to the other πλανήτης. Only with this bold claim did the πλανήτης become problematic and demand theorizing, which climaxed in the work of Eudoxus. This story also has the advantage of flowing from wonder - the shocking identity of Φωσφόρος and Ἕσπερος - to science, as in Smith’s theory.

Moving from Eudoxus to Ptolemy, Smith remarks “Nothing can more evidently show how much the repose and tranquillity of the imagination is the ultimate end of philosophy, than the invention of this Equalizing Circle.”. Truth as an end of philosophy is pointedly left out. But the actual dialectic he follows is more complex. The comedy of science is that we have ways of finding the repose and tranquility of the imagination, but we want ways of finding the truth.

Passing over Smith’s comments on Arabic and Medieval astronomy - which are brief and simple - we come to the Copernican Revolution. Some window dressing is needed before we start. Aristotle considered Eudoxus’s spheres to be physical objects: they can be meaningfully counted and the fact that they don’t make noise when they grind on each other has astronomical implications. Though the point is debated, I think it likely that Eudoxus thought of them the same way. Hipparchus & Ptolemy no longer considered astronomy from a predictive point of view. Their goal was to predict when and where a shiny point of light would show up in the near future. The result, as Smith discusses, is that astronomy, once a vital part of Platonic education, dropped out of the philosophical conversation altogether. Smith argues that the Roman philosophes were so concerned with mores they lost all interest in mere accuracy. Why this should be astonishing if “repose and tranquillity of the imagination is the ultimate end of philosophy” is left unclear. What Copernicus did was, for the first time since Aristotle, propose a theory of the heavens which aimed to be descriptive rather than predictive.

Adam Smith’s discussion of the Copernican system is excellent. He owned a first edition of Copernicus’s De Revolutionibus Orbium Coelestium, which you can read right here. Copernicus was convinced his system was true because it was “simpler” than its contemporary competitors in the sense that it lacked equants entirely (it still had other Ptolemaic hacks however). In truth, the Copernican system had no real competitors. This perhaps lulled Smith into a false confidence, as he follows Copernicus in talking boldly of the relative simplicity of theories as if that had a simple and uncontroversial meaning. I recommend this workshop as a good indication of how complicated it is to say what simplicity means.

Smith’s judgements were obviously pitched to consolidate his narrative, because here they make no difference. The churches that condemned Copernicus did not have a rival theory of the structure of the heavens but rather denied that such a theory was possible or desirable. That is not to say it had no issues with internal coherence or external accuracy. The only objection of any force was that, as Smith says, “A ball, it was said, dropped from the mast of a ship under sail, does not fall precisely at the foot of the mast, but behind it; and in the same manner, a stone dropped from a high tower would not, upon the supposition of the Earth’s motion, fall precisely at the bottom of the tower, but west of it, the Earth being, in the meantime, carried away eastward from below it..”, a matter of internal coherence with Aristotelean physics. This, if true, would have been absolutely fatal, for it threatened not Copernicus’s observations - observations are always fallible anyway - but his entire philosophy of a descriptive understanding of astronomy. Smith explains the Copernican reaction perfectly:

“It is amusing to observe, by what, subtile and metaphysical evasions the followers of Copernicus endeavoured to elude this objection, which before the doctrine of the Composition of Motion had been explained by Galileo, was altogether unanswerable. They allowed, that a ball dropped from the mast of a ship under sail would not fall at the foot of the mast, but behind it; because the ball, they said, was no part of the ship, and because the motion of the ship was natural neither to itself nor to the ball. But the stone was a part of the earth, and the diurnal and annual revolutions of the Earth were natural to the whole, and to every part of it, and therefore to the stone. The stone, therefore, having naturally the same motion with the Earth, fell precisely at the bottom of the tower.”

• Adam Smith, “The principles which lead and direct philosophical enquiries, as illustrated by the history of astronomy”, Section IV: “The History Of Astronomy”

This false attack and its phony evasion aside, Tycho Brahe made himself the first true critic of the Copernican system. He was distinguished from his predecessors both by one golden fact. Tycho Brahe took the descriptive aspect of Copernicanism seriously. For Brahe, discovering that the data showed “Venus and Mercury were sometimes above, and sometimes below the Sun; and that, consequently, the Sun, and not the Earth, was the centre of their periodical revolutions” was evidence of something, not just a series of numbers he logged in a table. 

The next phase of astronomy was founded by Galileo Galilei, the first to use the telescope in astronomy. Perhaps the revolutionary findings of Galileo can be exaggerated, but it would take a great deal of effort. Pythagoras is said to have discovered that Φωσφόρος is Ἕσπερος and thereby introduce scientific astronomy to Greece. Galileo seemed to make even more shocking discoveries daily. Before Galileo the crescent moon of man’s imagination - a symbol of virginity dating back to Ἄρτεμις - looked like this. Completely pure and silver. After Galileo, the crescent moon looked like this. The Church was so furious that they refused to accept the second painting as an allegory for the immaculate conception. Before Galileo, all heavenly bodies were supposed to glow. After Galileo, only the stars glow - the moon and the πλανήτης (notice the two πλανήτης at the bottom corners) merely reflect light.

After Galielo it could no longer be a mark of sophistication to consider astronomy as a purely predictive science. The heavens - the very place philosophy and religion should be the most secure - was now a place of investigation.

Kepler is the name that follows Galileo most fluidly, neglecting his own students Castelli, Viviani, Guiducci, Borelli, Magiotti, Ricci and Torrecelli who continued and contributed to their leader’s vision and work. Kepler’s discoveries - the elliptical nature of orbits, the conservation of angular momentum, the relation between the orbital period and the mean distance  - were of a far more technical nature than Galileo’s bluntly descriptive discoveries. This does not diminish their enormous importance. Smith describes how Cassini confirmed Kepler’s laws for the moons of Jupiter and Saturn, probably a deliberate simplification on Smith’s part, but he is no doubt right that the fact that one could in principle go and see shiny points of light in the sky act according to post-Copernican rules seemed to deflate serious anti-Copernicism. 

“If the Earth and the Five Planets were supposed to revolve round the Sun, these laws, it was said, would take place universally. But if, according to the system of Ptolemy, the Sun, Moon, and Five Planets were supposed to revolve round the Earth, the periodical motions of the Sun and Moon, would, indeed, observe the first of these laws, would each of them describe equal areas in equal times; but they would not observe the second, the squares of their periodic times would not be as the cubes of their distances: and the revolutions of the Five Planets would observe neither the one law nor the other. Or if, according to the system of Tycho Brahe, the Five Planets were supposed to revolve round the Sun, while the Sun and Moon revolved round the Earth, the revolutions of the Five Planets round the Sun, would, indeed, observe both these laws; but those of the Sun and Moon round the Earth would observe only the first of them. The analogy of nature, therefore, could be preserved completely, according to no other system but that of Copernicus, which, upon that account, must be the true one. This argument is regarded by Voltaire, and the Cardinal of Polignac, as an irrefragable demonstration; even M‘Laurin, who was more capable of judging, nay, Newton himself, seems to mention it as one of the principal evidences for the truth of that hypothesis. Yet, an analogy of this kind, it would seem, far from a demonstration, could afford, at most, but the shadow of a probability.”

• Adam Smith, Adam Smith, “The principles which lead and direct philosophical enquiries, as illustrated by the history of astronomy”, Section IV: “The History Of Astronomy”

Smith does his best to do justice to Descartes’ system of vortices. This speaks to both his francophilia and generosity, as recently deceased systems are far harder to give their due than those long dead. Essentially, the vortex picture encouraged people to think of the heavens as an immense ocean. It was then easy to picture the planets as globally moving but locally stationary, even if the vortex picture actually does not do justice to that picture.

Finally, the essay concludes with a discussion of the Newtonian system. Smith considered this portion of the essay imperfect, but its imperfections lie in the direction of being standard rather than being deeply flawed.

The next essay is “History Of Ancient Physics” and it is. I put it this way because, as one can see from the table of contents, this essay is 10 pages - scarcely a sixth of the length of “History Of Astronomy”. The essay begins with an apparent connection to the previous, but this connection is only to the title. “If the imagination, therefore, when it considered the appearances in the Heavens, was often perplexed, and driven out of its natural career, it would be much more exposed to the same embarrassment, when it directed its attention to the objects which the Earth presented to it, and when it endeavoured to trace their progress and successive revolutions.” says Smith, but is this perplexion wonder? The appearance of a rainbow - color with no pigment! - is often surprising, but surely the natural reaction is admiration rather than wonder. One can tell from this that this essay does not refer to its noble predecessor but to the idea of it, a draft that lacks all that makes the “History Of Astronomy” unique.

Smith says, and it is natural enough, that a simplicity behind all the myriad phenomena of nature must be supposed in order for physics to begin. But then he starts not with the monists, who offered the simplest of all possible theories but with Empedocles and the four elements. The only excuse for this is that the four elements is the physical theory Aristotle prefers, so obviously it is the only one worth taking seriously. We have already seen the previous essay also has this defect, but since this essay is so much shorter the result is much more powerful here.

After a bit of description of the Four Elements theory, Smith goes on to discuss the origin of the world. Archaic Greece had no notion of a creator god. Gods were masters of this or that, even if they were supposed eternal and omnipotent in hymn. Anaxagoras and the Pythagoreans are credited with originating the notion of archaic mind/order which reaches its complete form in Plato.

Aristotle, contrariwise, is credited with proposing the eternity of the world. “[W]hat could hinder…” Aristotle may have asked “the First Cause from exerting his divine energy from all eternity[?]”. Certainly there could be have been no cause to arrest him! The cool reason of Aristotle so completely harmonized the human vision of the world that it, Smith says, “saps the foundations of human worship”, which probably would have connected to the analysis of surprise had this truly been a completed essay. Despite this - as Smith notes - it is Aristotle’s system which gave birth to the Scholastic system. The world would have to wait until Hegel until someone took up the thread of this analysis.

Finally, there is a brief note on the Stoics “the most religious of all the ancient sects of philosophers.”. The Stoics, Smith claims, smoothed and symmetrized all that was difficult and cutting in Plato’s analysis. The Stoic’s model world is both eternal and created, these opposites joined by the simple supposition of a cyclic cosmos. This ordering speaks to Smith’s proto-Hegelian logic rather than to historical fact: it’s very possible (though the point is much debated and I in fact doubt it) that Heraclitus proposed a cosmic cycle from conflagration to conflagration and Plato’s theory is a reaction against it. This is not a true flaw since the point of the essay is to derive what principles we can from the practice of physics not present a complete and independent history of physics. The flaw is the incompleteness of the essay: it simply stops here.

The last essay, “History Of Ancient Logics And Metaphysics”, is even more imperfect than its predecessor. The essay relies very little on the research and speculations of the previous essays. Honestly, it reads rather like someone quite smart but over opinionated going over summary articles and pontificating.

Anyway, to get into the text, the majority of “History Of Ancient Logics And Metaphysics”’s length is a defense of the Aristotlean system which takes the form of downplaying the pertinence and existence of deviations from The Philosopher. The essay starts off reasonably enough, though following Aristotle’s version of the history of philosophy to a fault. The principle of primeval philosophy announced in the title comes easily in the first sentence “In every transmutation … there was something that was the same and something that was different.”. This leads naturally enough to monistic presocratic philosophy, which laid a solid metaphysic for science. Opinion relates facts about a particular arrangement of the one substance - say, waters. Science looks to that which is common to all waters.

This was a reasonable starting place even by Smith’s lights. However, it obviously both could be and actually was exposed to great criticism. Smith then moves on to discuss Plato and swiftly loses all patience and good sense. “...Plato …imagined [he] could still further confirm, by the fallacious experiment, which showed, that a person might be led to discover himself, without any information, any general truth, of which he was before ignorant, merely by being asked a number of properly arranged and connected questions concerning it.”. The fallacy comes in the undefendable addition‘, of which he was before ignorant,’. The fallacy thus was both created and demolished by the critic, as it so often is. The invented flaw is then magnified beyond all proportion by taking one passage of Meno as Plato’s final and complete theory of education, ignoring successes such as Theatetus and failures such as Dionysius II, as discussed above.

Seemingly aware of the weakness of his own argument, Smith abandons all pretense of finding principles which lead and direct philosophical enquiries for mere verbal criticism: “What seems to have misled those early philosophers [i.e. Plato], was, the notion, which appears, at first, natural enough, that those things, out of which any object is composed, must exist antecedent to that object.”. This would immediately obliterate rather than support the theory of education which Smith attributed to Plato, as it would necessarily contradict the ignorance existing before the knowledge. That this paragraph does not at all follow from the previous discussion of how the notion of All Triangles arises when every experienced triangle is obtusangular, or rectangular or acutangular except in being is another serious demonstration of the imperfect state of the essay.

The essay ends with a discussion of the Stoic philosophy of fine grained distinctions. The essay ends with no particular conclusion with a list of these distinctions. In other words, it follows the exact pattern as “History Of Ancient Physics”, yet another clue as to its undigested fate. All that I can say is that it is sad that the author died before this essay could be brought to the level of its grand first entry. Had he lived longer, perhaps Smith could have disproven Sydney Smith’s sneer (quoted with approval by Bagehot) “Greek has never crossed the Tweed with any force.”.

Edinburgh & Theory Of Moral Sentiments

That ended on a rather disappointing note. We can see that Oxford provided little material for Smith’s life’s work and then saw his life end before he could transcend it in his historical work. But Smith did not live a sad life. Smith left Oxford for the more intellectual environment of Edinburgh. There he found again the stable base to write his first masterpiece, Theory Of Moral Sentiments (TOMS). Bagehot accurately but unfairly reviews TOMS in the already mentioned article as follows:

“In Adam Smith’s mind, as I have said before, [moral sense] was part of a whole; he wanted to begin with the origin of the faculties of each man, and then build up that man—just as he wished to arrive at the origin of human society, and then build up society. His Theory of Moral Sentiments builds them all out of one source, sympathy, and in this way he has obtained praise from friends and enemies. … One party says the book is good to gain authority for the conclusion, and the other that you may gain credit by refuting its arguments. For unquestionably its arguments are very weak, and attractive to refutation. If the intuitive school had had no better grounds than these, the Utilitarians would have vanquished them ages since. There is a fundamental difficulty in founding morals on sympathy; an obvious confusion of two familiar sentiments. We often sympathise where we cannot approve, and approve where we cannot sympathise. … Adam Smith could not help being aware of this obvious objection; he was far too able a reasoner to elaborate a theory without foreseeing what would be said against it. But the way in which he tries to meet the objection only shows that the objection is invincible. He sets up a supplementary theory—a little epicycle—that the sympathy which is to test good morals must be the sympathy of an “impartial spectator”. But, then, who is to watch the watchman? Who is to say when the spectator is impartial, and when he is not? If he sympathises with one side, the other will always say that he is partial. As a moralist, the supposed spectator must warmly approve good actions and warmly disapprove bad actions; as an impartial person, he must never do either the one or the other. He is a fiction of inconsistent halves; if he sympathises he is not impartial, and if he is impartial he does not sympathise. …”

Of course, Bagehot is as ever sublimely hubristic in his confidence in Imperialistic Utilitarianism. But a quite opposed thinker, Noam Chomsky, says something similar. “Adam Smith… felt that it shouldn’t be too difficult to institute humane policies.”, he wrote with obvious irony. This flaw left “[c]lassical liberalism shipwrecked on the shoals of capitalism,...”.

For myself, I agree with Brad DeLong that Smith’s theory has aged far better than the simplistic unidimensional theories of utilitarianism that Bagehot defended. Further, Bagehot misinterprets TOMS as a theory of moral acts, rather than as a theory of the foundation of virtues. It is a theory that aims to do the work of Mill’s variation of pleasures into sorts. The impartial spectator is merely the Socrates who knows both sides rather than the pig who knows but one. The matter which Smith’s impartial spectator judges is not the relative merits of the possible actions by multiple parties, but of the relative weights of conflicting virtues. The bulk of TOMS shows by construction that this theory sublated the best of our variegated moral feelings - Burke enumerated “the just, the fit, the proper[ and] the decent” in his review - into an internally ordered unity. Only in the last part of the book, Section IV of Part VI, does Smith turn to the problem of applying his constructions to the engineering of actions. It is rather later, in WON, that Smith constructs a partial theory of moral acts (by a state) out of his theory of moral sentiments. 

Conclusion 

The last paragraph of TOMS promises a philosophical history of jurisprudence. Smith makes a distinction between old methods of  ordering society - mere police - and the new method of justice. Smith abandoned this plan, which has been fulfilled as Hegel’s Philosophie des Rechts. That plan became WON rather than what was advertised. In WON, Smith would develop his concept of justice in an economic rather than historical order. I wrote at the top that I wouldn’t discuss WON in detail, there being enough of that here, there and everywhere. However, I would like to end on some kind of summative notion so we can see how TOMS leads into WON.

The method of justice is, we would now say, is that  of freedom. Freedom is when, within wide limits, one’s conduct is guided by reason: symmetry, efficiency and equity. It is the brutish man who is the slave, not those who fear him. The slave to his passions fears justice and knows, or will soon will know, that time and chance brings equality.

Smith’s Justice is not freedom from government, no, it is freedom of government. Whig history - Smith would appreciate this - is to try to interpret him in the light of 20th Century totalitarianism or libertarianism. Nowhere in Smith do we see anything like Montesquieu’s notion of the eternal tax break: once a petty aristocrat is promised tax relief all his descendants are so sheltered for the law is the aggregation of history. Instead, the government is free to adjust taxes. Free is the government guided by reason: symmetry, efficiency and equity. The injudicious tax creates the smuggler who must be punished in proportion not to the crime but to the tax. The despot always prefers caste and hierarchy, but it is caste and hierarchy which creates racial and other strife.

The necessary addition to Smith’s notion of free government, in my view, is the principle of continuity that says the free government is continuous with its free people: government of the people, by the people, for the people. As another old reader of Adam Smith said: Nur Dann –

aus dem Kelche dieses Geisterreiches

schäumt ihm seine Unendlichkeit.

“As [the imagination’s] ideas move more rapidly than external objects, it is continually running before them, and therefore anticipates, before it happens, every event which falls out according to this ordinary course of things. … But if this customary connection be interrupted, if one or more objects appear in an order quite different from that to which the imagination has been accustomed, and for which it is prepared, the contrary of all this happens.”

Adam Smith, “The Principles Which Lead And Direct Philosophical Enquiry, As Illustrated By The History Of Philosophy”, Section 2: Of Wonder, or of the Effects of Novelty

A Table of Contents for a Long Post

1. What Do We Read Minsky For and What Should We Read Minsky For?

   1. Minsky and Capital Theory in the Designer Economy

   2. Reading Minsky Properly

2. Minsky’s Intellectual Environment

   1. Part 1: The Paradoxes Of Rationality

      1. What Exactly Do We Mean By Rationality

      2. Paradoxes of Rationality in the Work of Minsky’s Contemporaries

   2. Part 2: The Contradictions Of Growth

      1. Quantitative Models of the Qualitative Transitions of Growth

      2. Growth, Demand, Inventory and the Revaluation Impulse

3. Minsky’s Economics

   1. Part 1: Demand Rules Employment

      1. Neoclassical Growth and the Fifth Channel From Demand to Employment

   2. Part 2: Investment Rules Demand

      1. Interlude: The Pure Theory Of Strategy

         1. It’s Time For Some Game Theory

   3. Part 3: Liquidity Opens Strategy

      1. Liquidity Opens Strategy; Unknowledge Closes It

4. Conclusions

What Do We Read Minsky For and What Should We Read Minsky For? 

Maybe you remember 2006 well. A young fresh face named T-Pain topped the music charts, film bro classic In Bruges dominated theaters and the international banking system violently collapsed causing an international financial crisis which would dominate global politics for the next ten years.

In the wake of the rapidly developing global financial crisis, many people turned to the economic sage Hyman Minsky for guidance. In some sense, this literature was a boon. After all, Hyman Minsky was a great economist, and it’s always good when they see a comeback. But in another, more accurate sense, this pretense was bad, and the literature it brought about confused and reduced Minsky’s economics to such a degree that they could easily fit in a few lines of newsprint.

First of all, the phase change in Minsky citations seems to have been triggered by the Lehman Brothers collapse. While this collapse was certainly important, one hardly needs a sage to analyze it. Equally annoyingly, in the trendiness of collapse finance, Minsky was absorbed as a bit player in the then-current acronym debates - MMT, QE, TARP, etc.. But most importantly, the reputation of Minsky as a business cycle doomsayer has severely reduced the audience of his theories, and broadly reduced the comprehension of his non-audience.

This started long before 2006. See, for example, Paul Samuelson’s ‘A Personal View on Crises and Economic Cycles’ (paper 497 in volume 7 of his collected papers). Samuelson is certainly justified in asking if the world needs yet another voice constantly shouting “A qualitative credit crisis is in the intermediate-term cards. Wolf! Wolf!”. Is Minsky’s role merely to assign a determinate name and career to the fact of that voice?

Assuredly not:  Minsky’s work is even more relevant now than it was in the months and years following the financial crisis. The way the COVID19 Mother Of All Demand Shocks has revealed that, for example, our JIT production infrastructure was more fragile than anticipated, is far more Minskyian than the collapse of Lehman Brothers was.

Minsky and Capital Theory in the Designer Economy

So how does Minsky help us solve this problem? Asking companies to go back to 1950s production methodologies is not a plan. Creating the conditions for the production base to gravitate to more stable methods requires an understanding of inventory as a part of the portfolio of a company. Revolutionary, evolutionary or even liberal: you still have to do capital theory.

Taking a Keynesian approach to capital theory means identifying what kind of part inventory is in the portfolio of a company: it is an illiquid one. Holding goods means being unable to hedge the risks involved in not-holding cash. Because of this fact, a capitalist economy is structurally stuck with a deep paranoia. The ‘fear of goods’ is built into the heart of the manufacturing processes because inventories, as a kind, intrinsically move slower than merely financial processes. Inventories are also creatures of time and timing: of changing demands, changing capacities, and of the need to put them somewhere for the duration of their time as inventories.

This is at the heart of the “financial theory of investment” which powers Minsky’s model. The case I wish to make in this review is that rather than the mere stormcrow he is often made out to be, Minsky was instead a prophet of the designer economy.

At the most general level a Designer Economy is an economy which aims to equilibrate through the creation of possibilities rather than by the elimination of possibilities. Our solution to the oft-justified paranoia of the capitalist is not to play into it and allow him to bunker down, but instead to help him develop habits that allow him to open up: a kind of administrative therapy. As Spinoza put it:

PROPOSITIO LXXIII : Homo qui ratione ducitur magis in civitate ubi ex communi decreto vivit quam in solitudine ubi sibi soli obtemperat, liber est.

Spinoza, Ethica, Pars quarta: ‘De servitute humana seu de affectuum viribus’

The ideas Minsky developed will be even more important to the success of a Designer Economy than they were to explaining the global financial crisis.

Reading Minsky Properly

But one truth must be admitted. Minsky is even harder to read now than he was at the turn of the century. The writings themselves are difficult, but in 2008, the world’s problems were financial, Minsky’s home turf. A financial collapse fed forward into an investment desert which fed forward into a decade-long demand shortfall. Contrariwise, in 2023, our problems are pandemics, wars and climate change which Minsky’s writings are not directly about. Despite this, the deductive chain that links his premises to the topics that make his work vitally important to today’s economic issues is strong but long.

Why then has Minsky’s Ngram risen? I fear that discussions of Minsky often turn into

1. a list of reasons to read/not to read Minsky that make it clear the list’s author has not read him 

2. inscrutable insider baseball,

3. Political/Philosophical/Personal jeremiads,

4. jargon unintelligible to anyone who would need to hear a discussion,

5. or, worst of all, high quality economic analysis that tells truths many don’t want to hear. 

Like Chedorlaomer, I wish to destroy four of these five cities. I will start Minsky’s case by reading him into the intellectual environment he thrived in, which will give us the right choice of tools to understand him. Once we understand the broad sweep of his intellectual environment, we can plow through the land Minsky covers. That is to say, the Finance -> Investment -> Demand -> Output channel mentioned above. I find it easiest to go backwards:

1. Demand -> Output 

2. Investment -> Demand 

3. Finance -> Investment 

After this groundwork is laid, we can close out with a few finer points.

As a result of this approach, some of the most important and well known of Minsky’s theses won’t be mentioned at all and others will be restated in my own peculiar argot. This therefore is more a prolegomenon than a review. My only hope for the resulting review is that you are able to read Minsky with fresh eyes.

Minsky’s Intellectual Environment

Part 1: The Paradoxes Of Rationality

Minsky’s theory may seem strange. Who asked for a financial theory of recessions? Do we think sophisticated investors structure portfolios that crash? What arrogance to think sophisticated investors with decades of experience just work their minds toward their own doom, but mere academics can do better.

Some have even asserted Minsky violates some “assumption of rational behavior”. There is an analogy in this accusation to the economics of machinery, as highlighted by one of Minsky’s closest friends. Naively, one might believe that the introduction of a more efficient machine must be non-decreasing in output, since society can always choose to simply not adopt a technique that reduces output (and to the extent it cannot, that is a problem of monopoly rather than machinery). This misses that the effect of a labor saving invention can be to raise rents and thereby reduce production in the long run. In a mythical Ireland where the nobility eats cake and the peasants potatoes, improving wheat land means the nobility has less need to suffer the existence of the peasantry. 

Samuelson above and Kindleberger elsewhere see these intuitive mistakes as examples of the infamous ‘mathiness’ of current economic theory - one can often get away with an unsupported assertion by dropping a technical sounding buzzword.

The buzzword is “Pareto optimality” for machinery and “rationality” for finance. If you are versed in the jargon, you might notice these two buzzwords - when used properly - denote the same closely analyzed concept. But finding the buzzworder’s mistake is only half the battle. Actually developing a theory that replaces buzzwords with good sense is the other.

What Exactly Do We Mean By Rationality

Let’s focus on rationality. Rationality is a grand subject: rationality per se has been a goal and object of investigation ever since The Enlightenment. We shall soon see that Minsky's analysis of rationality in particular highlights what is often called a “paradox of rationality”. A paradox of rationality is a particular case of the famous “fallacy of composition”: the rationality of each member of a group doesn’t entail group rationality. These are well known enough that I can name three examples off the the top of my head:

1. The Paradox Of Thrift, in which a group of rational savers cause a reduction in total saving,

2. Tendency of the profit rate to fall, in which a group of rational capitalists invest until their original investments can’t be validated, and

3. Minsky’s Example: a group of investors who invest in the status quo destabilize the status quo. (See, for instance, a crowded trade)

Partly under the influence of Keynes and partly because of the game theory revolution, economists and philosophers became interested in a particular paradox coming from the organic nature of judgment and action just as Minsky was maturing as an economist. The basic economic paradox is simple: a person who may judge rationally often must act “irrationally”.

An extreme example can be given, if I am allowed some license. In 1945, the Japanese high command was aware that it had lost the war. They planned to draw out the conflict while offering conditional surrenders in order to go to the negotiating table in a strong position. The Japanese high command posited that they were willing to see the home islands flattened in order to avoid seeing the home islands flattened. This posit became a pose which did not work. The Japanese high command eventually went to the negotiation table with an unconditional surrender.

Paradoxes of Rationality in the Work of Minsky’s Contemporaries

We can see how broad this fascination with paradoxes of rationality became by looking at three contemporaries of Minsky, at least one of which was influenced by the example of the Japanese high command. Those three are Thomas Schelling, Robert K Merton and David K Lewis.

Thomas Schelling is well known for fascination with paradoxes of rationality, which he sees as a distinction between micromotives & macrobehavior (i.e. the fallacy of composition). But Schelling didn’t start off as a strategy theorist, but as a Keynesian economist. Indeed, Schelling first got interested in paradoxes of rationality by the famous “paradox of costs” in Keynesian economics!

That Schelling moved from economics strictly speaking to general strategy theory shows that economics was not the only field discovering paradoxes of rationality. Around the same time, Robert K Merton was writing influentially on self-fulfilling/negating prophecies. Like the economists Minsky and Schelling, Merton was fascinated by how actors in a capitalist economy deal with the unyielding uncertainty of life. The center of his theories was on the unending tension between imitation and adaptation. He created a model in which an actor within an organization imitates his “role model” until they predict that imitation will not fulfill their desires. In that case, they attempt to adapt and the organization then must adapt to this adaptation. We see then that there is a fundamental paradox of rationality. In this almost cybernetic vision, a self-defeating organization is much like a self-defeating prophecy, snatching defeat from the jaws of victory.

Speaking of almost cybernetic visions, we come to the last of our trio: David K Lewis. Lewis was a so-called analytic philosopher in a tradition anchored by his mentor Quine. This variety of philosophy emphasized mere logical analysis, but made up for this by vaunting classical first order to almost religious levels. This technique oriented philosophy unsurprisingly reached a polish and sheen rarely seen in any academic discipline. Lewis showed that a group of rational organisms could use arbitrary symbols to coordinate non-arbitrarily. If each organism acts according to a rule “if I see signal s, then I do action a.”, then in a very real sense signal s ‘means’ “We are in situation n, which makes a the best course of action.”. This is an example of a paradox of rationality in reverse: rules which are arational to the individual beget group rationality. There is actually a quite direct contact between Lewis’s conventional signaling analysis and uncertainty, as a more general statement of rules of behavior is more like “If the signal is s with probability p, s’ with p’, s’’ with p’’, etc, then I do action a with probability q, a’ with q’, a’’ with q’’, etc.”.

This has started to take us away from our goal, and besides the point has been established. In the introduction, we noted that it’s not difficult to see rational irrationality as a property seen in economic failures. We have now gone over briefly how the rationality which structures the micromotives of agents do not necessarily strongly constrain the macrobehavior of society. It’s sheer mathiness to go from rational agent to rational system. What it now comes time to do is establish how this ‘paradoxes of rationality’ applies to economic growth.

Part 2: The Contradictions Of Growth

The concept of economic growth developed from the same enlightenment theorists that promoted the concept of rationality. In fact, the hard right wing of the enlightenment, such as Swift & de Maistre, claimed there was a paradox of rationality in Leibniz & Spinoza. They proposed that a group of rationalists would be unencumbered by social restraints and thus immediately overthrow all progress. This is A O Hirschman’s “Perversity” from Rhetoric Of Reaction.

A more interesting theory developed from the moderate right of the Enlightenment: Mandeville, Hume and Adam Smith. Growth is a paradox of rationality, but the paradox is that a group of rule-followers can form revolutionary theories. We might recall Adam Smith on the history of  astronomy. Out of the sense of wonder, the ancient astronomers created a structural theory of the cosmos based on the four part classification of celestial objects into Sun, moon, planet and fixed stars. By a process much like a paradox of rationality, the dynamism of wonder was replaced with the husk of belief. Eudoxus was then, by Smith’s lights anyway, led by his unshakable belief in this system to overthrow it: making the theory geometrically precise required more spheres than the obvious classification makes it seem. “Work-to-Rule” strikes are an earthier variety of a similar process.

But this is to talk about growth in general, including qualitative improvement. Ever since Adam Smith, economists tend to think of economic growth strictly in terms of quantitative growth without qualitative change. 

Quantitative Models of the Qualitative Transitions of Growth

Let’s be precise for a change, although it requires a detour from Minsky per se. Adam Smith defines growth as an increase in the intensive margin. Compare two identical pin factories vs one pin factory twice as large. The second one has more output per capita due to more division of labor. 

The modern version of this quantitative growth theory is the theory of exogenous growth which started with Wicksell and was perfected by another one of Hyman Minsky’s close friends. I will go through a particularly comprehensible case because my goal is exposition not criticism. Besides, the right goal for exogenous growth theory is precision rather than strict accuracy or generality. As an approximation, assume all n+1 capital funds are rights to physical capital goods. Output per capita will be further assumed to be a fixed and well behaved function of all the capital funds in the economy in real value terms. This can be best comprehended in a couple equations:

where k_i is the value of capital good i,  p_i is the unit price of capital good i, q_i is the quantity of capital good i, P is a price deflator, N is the quantity of labor in the economy, y is real output per capita, b is a factor of proportionality and a_i is the weight of good i. Dimensional considerations tell us that the a_i sum to 1. The power of the model comes from assumptions like assuming the weights constant (or similar assumptions for different functional forms). The additional term b is usually allowed to be a function of time, interpreted as “technical change”.

In 1956, Robert Solow used this simplified model to argue for the stylized fact that technical change contributed more to growth than capital deepening did. This was bound to be controversial, but strangely enough neither this result nor its negation became a part of economics. Rather the style of model became de rigeur for so-called mainstream economics. This mystifies everyone, including Solow.

This approach ignores, quite deliberately, crucial qualitative aspects of economic growth. Returning to Adam Smith, look at the example of technological change (his comment is deleted from Wealth Of Nations, but present in the Canan edition). Implementing Jethro Tull’s steam powered seed drill improves the intensive margin of the same labor hour on the same land producing the same wheat. But this causes a revaluation of all the hand powered seed mills. So we see that holding consumption goods - wheat - fairly constant so that y has a stable meaning, means that production processes are constantly changing. Then in order to compare outputs at different times, the k_i cannot be the market value of actual capital goods, as each capital good is constantly being revalued. Instead, it has to be some sort of index. This leads us into the muck of the ‘measurement of capital’. As Robert Solow put it:

“If you think of growth as an uninterrupted, smooth process, and if the economy is in some kind of a natural equilibrium all the time, then you can give a fairly complete analysis of that. But when a surprise happens, a shock of some kind, a war, maybe a big invention, any kind of a disturbance, then the existing real capital – machinery and buildings, computers, telephones, and all that – has to be revalued in smooth equilibrium conditions. This is what capital theory is about. You can talk in pretty precise terms about what the value of a capital good is. The reason why, when a disturbance occurs, you cannot do that any more, is because the value of the capital good depends on the present discounted value of its future earnings. The essence of the surprise is that you thought you knew what your earnings would be but it turns out that you don’t. And then the question is to understand in exact detail what happens then and how the economy does or does not get back to a different equilibrium state. You have to be able to get a grip on how capital goods are valued under such circumstances of real uncertainty, that is, uncertainty that cannot be described by probabilities. That’s very difficult. I have never had a very good idea. Nobody else has ever had one either. That’s a very important question, it’s not esoteric.”

• Robert Solow in Karen Ilse Horn’s Roads to Wisdom, Conversations with Ten Nobel Laureates in Economics, emphasis mine

We will return to the time structure detailed by Solow later. For now, the main point is economic growth often comes not from the same goods just more, but rather, from different goods entirely. This results in two related paradoxes of growth: ‘creative destruction’ and ‘fear of goods’.

Growth, Demand, Inventory and the Revaluation Impulse

Because growth involves qualitative change, a firm cannot guarantee that there will be the same demand for their output after the change as before. Simply from an uncertainty management point of view, a firm is usually better off holding cash rather than a good of near equivalent value in this situation. As Shigeo Shingo said, inventory is evil. This is the ‘fear of goods’. This uncertainty can shadow you even if your wares are superior to the new substitute goods, which is the body of the process of ‘creative destruction’.

Having now hit two famous buzzwords, it will do to make sure we really have done the work and not just dropped them because it feels good. In Part 1 of this section, we re-aired some famous pitfalls in the application of rationality. Now we have claimed that economic growth intrinsically involves a paradox of rationality: firms, simply by trying to grow in value, subject themselves and each other to downward value revisions. To paraphrase Popper in the introduction of The Poverty Of Historicism, it is in qualitative technical change where we expect our predictions to be the least successful - if we can predict which capital goods would bring the highest return, then why haven’t we already bought them? 

Paradoxical valuation is a very important question, it’s not esoteric. If you drive a barely functional $4,000 clunker to work every morning then you can lose food stamps and thereby lose your family’s access to a good dinner because of the lurching pricing dynamics of the used car market. Minsky’s economics tries to come to terms with an economy characterized by pervasive growth-driven valuational paradoxes of rationality by following the chain of causality from finance to output. As stated in the introduction, we will follow that chain backwards: Demand -> Output, Investment -> Demand and Finance -> Investment. Like Ancient Greek plays, we will break up the economics with a Satyr play: a brief interlude into the pure theory of strategy. The strategies created by such a paradox ridden valuation system are the stuff of Minsky’s economics itself. 

Minsky’s Economics

Part 1: Demand Rules Employment

The above is more or less Keynes’ model which he developed within his background logic in the first four books of The General Theory (GT). Feedforward from dynamics in financial portfolios consisting of physical and financial capital is stabilized, with lots of leakage, by feedback from dynamics of employment vs liquidity preference. 

Right now, we will concentrate on the connection of demand to employment. Classical theory allowed for four sources of unemployment: friction, leisure preference, productivity in wage-good industries and price of wage goods. This creates a specific dynamic: “That which we call dearness is the only remedy of dearness: dearness causes plenty.” - this is how Bagehot summarizes what he believed valuable in the Physiocrats. Say’s law was “There is no real[, i.e. long lasting,] dearness but that which arises from the cost of production.” (David Ricardo, On The Principles Of Political Economy And Taxation, Chapter 20: Value and Riches, their Distinctive properties). This is the reverse of Keynes’s position: rather than demand pulling resources into employment, the fact of employment creates demand for resources. This is not a foolish position to take, but there is no end to examples of inverted price/quantity spirals that fall outside of its purview. What can happen in general happens to the specific: there can be inverted wage/employment spirals of collapsing wages and employment.

Neoclassical Growth and the Fifth Channel From Demand to Employment

In order to understand this situation we must return to fundamentals. The issue for now is under what conditions is pulling an employee out of the pool of unemployed labor a profitable investment by the hiring firm, given its portfolio of physical and financial capital. A capitalist wants to hold a contract in his portfolio when he can soundly expect to thereby obtain the fruits of the contracted labor. The four sources of unemployment can be easily understood from this perspective. 

Most firms are small and illiquid, and so survival by itself demands the firm not take on labor even at a near zero money wage rate in times of high uncertainty. The fifth channel of unemployment is the uncertainty channel. Keynes goes so far as to show that expectations in the above sense determine output and employment. When this fifth channel is the operational one, it is demand which creates employment. The firm hires a laborer when it expects the demand for the increment of output to be strong and stable enough to justify the expense. 

One extremely minor result which has dominated the Minsky discourse to the point I must include it is his ‘right sizing’ of the so-called neoclassical theory of growth, that is the exogenous theory of growth expounded above and its endogenous descendants. He comes to the same conclusions as his friends Samuelson and Solow: the exogenous theory of growth has stringent assumptions and can be used in the hope that some principle of continuity will preserve some of its qualitative results. Exogenous growth is, at best, a “trust but verify” fuzzy big picture view. The neoclassical theories of endogenous growth have tried to maintain the basic structure of the exogenous model while allowing some bells and whistles. The nicest thing that can be said about these models is that many of them collapse to exogenous growth theory. It’s not unusual for an endogenous growth model to exhibit endogeneity on a parameter set of measure zero.

The results of decades of endogenous neoclassical growth theory has convinced me of the truth of what Minsky predicted long ago: greater insight will mean changing not just the form of neoclassical growth theory but also the structure. We will have to open Keynes’s fifth channel of unemployment, the tension between employment and liquidity preference.

But let’s return now from the tedium of discussing economics to discussing the economy. Summarizing: the basic paradox of rationality Minsky examined is that qualitative changes in capital markets cause revaluation, switching goods from cheap to dear unpredictably. Given this, if we require all resources be in use, then the demand for resources should be unstable. Liquidity preference ‘stabilizes’ the market system by enfeebling the force of dearness so that demand is the primary actor. But, in what amounts to other words for the same thing, liquidity preference also frustrates the axiom that dearness causes plenty. 

We must pick our poison. Do we let dearness exert its feeble impulse bearing dearness for a long time? Or do we take the risks involved in acting to change costs of production? Let’s go over what brought us to this impasse. In the introduction there was highminded talk about creating possibilities. The possibility of acting to lower costs of production directly in the face of long lasting dearness is how Minsky paves the way for this freedom. The last two sections went over the difficulties in achieving that freedom. Now we are set up for the practical. What we want to do with Minsky is find what can replace the feeble force of dearness to induce investment.

Part 2: Investment Rules Demand

Yes, it is finally time to actually discuss the book under review. There are two qualitative facts about the theory of investment that must be emphasized even in a brief review like this: 1) yields are quasi-rents and 2) except for the supply schedule for capital assets and the consumption function, the investment function fluctuates wildly.

Let’s start with the first. The yield of an investment is, to some degree, a rent on the underlying asset. In other words, the yield of a capital good is not a measure of its marginal productivity as a capital good. This quasi-rent yield is why the investor might choose to hold the capital good rather than cash, a reward for ignoring his fear of goods. But still, yield is partly a reward merely for owning something. While working on this problem, Adam Smith was wise to quote Hobbes: “Wealth is power.”.

Now we turn to the second point. Starting with definitions: the supply schedule for a capital asset is its replacement cost and the consumption function which maps income(s) to consumption. The replacement cost the price of making a new capital asset as an alternative to buying one off the market. As noted in the previous section, classical economists such as Ricardo and Say concentrated on replacement costs as a part of the supply schedule to the point of exclusion. The supply schedule for a capital good is usually, both by economists and by economic actors, assumed to be stable in the short run. Recently we had an astounding exception, as the replacement cost for microchips and a range of industrial and capital equipment exploded during the COVID19 pandemic. However, this assumption isn’t theoretically all that important, for theorists or participants, and few people were surprised by the lack of an endogenous market change in chips because of it.

The stability of the consumption function is much more important and is Keynes’s contribution to economics praised even by the most anti-Keynesian. Now, filleting out the consumption function to do investment theory is the reverse of econometric practice. One normally takes investment as an input (measured by phone surveys) and tries to fit a consumption function. Milton Friedman’s most important contribution to economics by far was in estimating a consumption function in a way that found a slight effect of interest rates. Since interest rates are a relative price (the price of a forward contract divided by the price of a spot contract), this shows that consumer behavior is to some extent directed by a relative price. This was, and still is, the sole evidence for the empirical relevance of economic rationality.

The importance of this result for microeconomics only highlights the irrelevance of the form of the consumption function for macroeconomics. In a closed, balanced budget economy, aggregate demand is simply the sum of investment and consumption. Given the stability of replacement costs and consumption functions (and thereby the supply schedule for capital goods), for such an economy, to the extent demand is unpredictable it is because investment is unpredictable. 

In other words, as phrased in the title of this section, it is investment that rules demand. Let us again summarize what we have seen so far, so as to illustrate the importance of this result. We want to use Minsky to create possibilities. In Minsky’s intellectual environment, there was a strong interest in the paradoxes of rationality. Qualitative economic growth itself involves such a paradox, as a company grows that company causes revaluations on its own assets. In the last section we saw how experience and theory both show we cannot rely on self-generated price fluctuation sensitive demand shifts to keep resources in employment. In this section we showed that it is investment which must make up this shortfall. This is a core part of the merely analytical side of Minsky’s economics. Now we will leave the text of Minsky for a moment to see whether Minsky and Keynes’s concerns are truly as general as they claim - can the problem of fear induced investment shortfall be simply regulated away? Or is it a fundamental part of strategic action?

Interlude: The Pure Theory Of Strategy

With the pure theory of strategy, we can start to cash in on the promises made in the background section. The point of going into paradoxes of rationality then and game theory now is to show that investment shortfalls are not accidental happenstances which can be swept away with one weird trick. The reason investment shortfalls are fundamental is that they are part of the uncertain strategy landscape of investment itself. One way that Minsky can be understood is by exploring the idea that market decisions are closer to so-called ‘Normal Form Trembling Hand Perfect Equilibrium’ (NFTHPE) rather than a necessarily ‘Subgame Perfect Equilibrium’ (SPE). 

There are two ways to get at what I mean by this. One is to follow the literature. Hyman Minsky, Robert Solow and others have pointed out that the ‘rational expectations’ school was misnamed: RatEx economists actually often explicitly broke assumptions of rationality and besides that, assumptions like market clearing actually did more explanatory and justificatory work than expectations. But I promised in the introduction not to get into insider baseball. This is just a small note in this direction: SPE also implies much more than mere rational expectations.

It’s Time For Some Game Theory

Let’s go into what we mean by our strategy concepts. We will start with NFTHPE. has three parts: Normal Form (NF), Trembling Hand (TH) and Perfect Equilibrium (PE).

1. Normal Form means that we list all possible strategies for each agent and the payoffs that are the product of each combination of possible strategies.

   1. This normal form tends to hide the sequential structure of strategy. In this case, we aim to leverage this rather than bemoan it. 

2. Trembling Hand means that agents act as if there is a non-zero chance they may use any feasible strategy in the future.

   1. Thus to the extent an agent can control themselves they put as much weight on playing any best response strategy they can. As their level of control rises, they put more effort into the bests of the best response strategies.

3. Perfect Equilibrium means that the strategies played are in all trembling hand strategies.

   1. To put it informally, they are the actions of those who do have control who are aware they might not have perfect control.

Before moving on, we can build a little intuition. If a best response gives a non-zero chance to all strategies  - a la Rock, Paper Scissors - then it is automatically a NFTHPE. If the best choice is to tremble, then just do it. The basic intuition of the NFTHPE can be put something like this: “More control should never make you worse off - if randomness works better, then choose randomness.”.

Contrariwise, an SPE is more like an extension of Tic-Tac-Toe than Rock, Paper, Scissors. The main point of Tic-Tac-Toe is to place your symbol in a place such that whatever your opponent does, you don’t regret placing it there.  For an SPE in general, control is assumed and avoiding regret is the only consideration. As a result, an SPE oriented agent often cannot take advantage of an opportunity. To paraphrase EEC Van Damme’s Refinements Of The Nash Equilibrium Concept, an SPE oriented agent is restricted to act as if his a priori expected payoff was true, even if the world and other agents act completely differently than that suggests. In other words, an SPE oriented agent never expects qualitative change.

What are the implications for financial strategies? One is that because finance is more like an NFTHPE is not a SPE, so-called ‘non-credible’ threats and even strictly dominated strategies can matter. The concept of ‘non-credibility’ assumes that the agent’s a priori expectations of credibility are stationary. Financial agents do not act as if their a priori expectation will hold true come what may. They can revise whether they think a threat is credible or not on the fly. This makes the job of the financier - whether a central banker or an ordinary individual - harder. They have to care not only about how their fellow actors revise their ex post expectations, but also have to plan for the whole market ecology to shift as actors revise their revisions. This finally brings us to the last link in the Minsky’s chain, the method by which actors can plan for their a priori evaluations to be wrong.

Part 3: Liquidity Opens Strategy

In order to have access to a wide variety of strategies, economic agents hold so-called “liquid” assets. “Liquidity is”, as Ken Boulding says in “Implications For General Economics Of More Realistic Theories Of The Firm”, “a measure of the degree of perfection of the market for an asset; i.e., the extent to which it can be exchanged for any other asset in indefinite quantities without loss due to worsening terms of trade.”. By holding assets of relative liquidity, a firm or household enables itself to abandon course when revising beliefs.

Now, the liquidity of an asset is not an a priori fact about that asset, it is anterior to market behavior. And yet, it is access to liquidity which is the sole justification for the ancient social practice of marketing. A market gives some actors (not enough) some access (not enough) to a form of certainty which opens possibilities. Markets are just places where liquidity happens. 

The ‘openness’ of liquidity leads to Minsky’s famous “two price levels”:

1. The price level of financial assets whose prices move fluidly

2. The price level of physical outputs whose prices slowly

Starting with the second price level, one can see that short run strategies around markets for physical outputs are practically limited to threats involving moving/creating financial assets. Source: every page of the FT every day. Many of these threats are not credible, but become credible.

For instance, a monopolist can hold liquid assets and threaten to lower prices against a potential competitor. If the monopolist doesn’t hold enough liquid assets to bankrupt a potential entrant, the entrant may regard the threat as noncredible. As the entrant sees the monopolist  take on debt to raise the funds to survive a period of low prices, the threat becomes credible. The current monopolist has an advantage over the potential entrant via a principle of increasing risk which was not part of the a priori evaluation. Monopolies supporting themselves through manipulating liquidity instead of producing product is, in essence, the financialization crisis thesis.

Liquidity Opens Strategy; Unknowledge Closes It

There is a dual to the title of this section: unknowledge closes strategy. In Samuel Bowles’ Microeconomics, a “risk dominant equilibrium” is defined to be the best strategies assuming other agents act randomly. Risk dominant strategies are rarely payoff dominant, which Bowles uses as a model for development traps. It is hard for laborers in Palanpur, Bowles claims, to trust their neighbors to plant earlier in the year. That strategy is thus closed to them (although Hirschman might, and probably more truly, see this as an opportunity for growth than as a growth-preventing trap).

Let’s return to the example of financialization, Minsky would note that a monopolist taking on debt to fend off entrants is actually a losing strategy over time. Every time the monopolist has to fend off a potential entrant they trade their old position for the same position but with more debt. In fact, in the long run, only debt that changes the position of the debtor is stable. Manipulating liquid assets to fend off competition risks rather than investing is what Minsky calls hedge borrowing. Eventually, the monopolist’s position may worsen to the point that its illiquid assets are in hock to continue the competition risk games. This is what Minsky calls Ponzi finance.

Now, laugh, friends, the comedy is over! Let us take the time to integrate over our progress. In our strange interlude I propounded the idea that financial agents are like trembling hand actors in a game - they like control but not because they assume their a priori expectations will be stable. This fed into the idea of liquidity as a source of financial agency. Once we see that it is liquidity that opens up the field to radical belief revision, we check liquidity’s effect on investment. Playing games with liquidity to keep one’s position is unstable because it leads to Ponzi finance. Liquidity’s essence is freedom and she will not be mocked long. This leads to instability in investment as qualitative changes force market wide revaluations. Instability of investment is itself instability in demand. Without stable demand, the ability of dearness to defeat dearness is enfeebled. There is a systemic and essential weakness in the traditional price oriented program. 

What is the solution? One is to reduce uncertainty by shifting to a design framework. Once this framework is outlined, the process of achieving that framework is through creating and using liquidity to strengthen the power of competition and the adaptability of firms and households. Minsky is, as argued in the introduction, the prophet of the designer economy:

Conclusions

All of the above only scratches the surface of Minsky’s JMK, which is only one part of his system of thought. I haven’t covered his detailed algebraic models or his historical work. Having reviewed at least some small part which is actually in Minsky, I will be more speculative and less pedagogical in this conclusion.

In my opinion, Minsky’s ‘Two Price’ system is a rational generalization of ‘money’ in the neoclassical system. Recall Yap, the well known island of stone money. Their economic culture was admired by neoclassical economist Milton Friedman. What is it that Friedman is admiring? It is that the Yap people have a payment system. 

“Acheter, c’est vendre et vendre, c’est acheter” Bagehot quoted Quesnay with approval. This is still taken as axiomatic by economic models today. But most trade is not barter. Most trade takes place through the plumbing of a payment system. How did that happen?

To put it another way, Minskyians often say anyone can invent a currency, the trick is to get someone to accept it. To go back to David K Lewis - who seemed so far out at first - how can an actor introduce a new signal which actually shifts the behavior of those receiving it? Can you take a phrase that’s rarely heard, flip it and make it a daily word?

Coming back to economics, Fisher’s circular flow with a meaningful velocity is a simplified payment system without checks, letters of credit, commercial bills or anything but gold coins really. A price then becomes a simple signal s in a Lewisian game.

Minsky’s generalization - and, as he says, Keynes’ generalization before him -  is to replace Fisher’s “velocity of money” with a more realistic portfolio theory of finance. This is an extensive change - the Keynesian Revolution - because for Minsky the portfolio is the active part of the economy but for Fisher the velocity of money is a passive part of the economy. In essence, the duality between signal and action is blurred.

The majority of post-Minsky research has been towards continuing this rational generalization towards denser theories of the plumbing of the financial system. This isn’t bad per se but this research misses the important growth aspects of Minsky. In particular, Minsky’s system is a rational generalization - a ‘general theory’, as it were - of neoclassical economics of both growth and the business cycle without the crutch of so-called ‘natural values’, such as the infamous ‘natural interest rate’ of Wicksellian monetary theory.

A natural value in general is a relative price which satisfies additional relations to the typical market relations seen in economics. For instance, an actual interest rate - being the relative price of a spot vs forward - satisfies certain market relations. Contrariwise, the ‘natural interest rate’ is an almost mystical relative price which would bring an end to net intermarket money flows, even those into the money market or those meant to bankroll production (“real bills”).

These natural rates dominate narratives of economic dynamics. Natural rates are so popular because they seem to give meaning to the non-market idea of a relative price being “too low” or “too high”. However natural rates are unnecessary if, like Minsky and Keynes, we use intelligently chosen national accounts.

For instance, if the Federal Reserve thinks - as it says - millions of people need to be unemployed in order for the economy to stabilize, then it should be questioned of what sectors should this unemployment come from?

This question is obfuscated by changing it to which sectors are being paid above their ‘natural wage’. The obvious difficulty is that the natural values are subject to change during qualitative growth periods. Thus they fail to guide us in exactly the case when we need them.

Even this relatively simple point demonstrates how Minsky’s Keynes gives us a simple and rigorous market oriented economics. 

To start out with praise for famous men, we can note that Hacking’s stature in the philosophy of science can be summarized by association. Hacking’s introductions bless the opening of the most popular version of both Kuhn’s The Structure Of Scientific Revolutions and Feyerabend’s Against Method - both of the two most controversial books in the field. I also say stature rather than structure: Hacking’s interests were quite different from Kuhn’s historicism or Feyerabend’s tedious recapitulation of Carlyle. It is that structure that we will attempt to draw here.

Hacking’s interest was in the human use of human beings. For Hacking, the rise of science was a rise in humanity’s self-understanding in a decidedly non-Hegelian sense: a rise in what humanity understood can be done. What ties the alchemists with the quantum chemists is that both participated in the growth of knowledge of what substance transformations can actually be done. This stance made Hacking one of the leaders of what came to be called the “New Experimentalism” and what Hacking called his Back To Bacon project. For Hacking, it was the laboratory which measured the incommensurabilities of Kuhn and Feyerabend. To do this, the  laboratory could not be the ‘reified theorems’ of Bachelard, but rather a self-justifying field of considerable stability. Hacking created a new Historical Ontology, saying “What we accumulate  are experimental techniques and styles of reasoning. Anglophone philosophy of science has too much debated the question of whether theoretical knowledge accumulates. Maybe it does not. So what? Phenomena and reasons accumulate.” (Historical Ontology, Hacking, 2002 emphasis in the original). New ways of doing and ways of seeing aren’t always knowledge, but neither are they always not-knowledge.

In the 18th century, Benjamin Franklin overturned traditional ontology by discovering the world was not largely non-electrical, but actually teeming with charge. In the 20th century, Anglo-Philosophers of science were convinced that they were little Franklins, showing that the world was not non-theoretical, but rather teeming with concepts. We have already talked about how Hacking was the Lavoisier who used experiments to further break down science into its elements. But there was another aspect of Franklin’s science which fascinated Hacking. That was the science of statistics and demography. When applied to the nation, suddenly marriage, life and death became scientific objects when applied to large sections of humanity. “In fine”, wrote Franklin, “a Nation well regulated is like a Polypus; take away a Limb, its Place is soon supply'd; cut it in two, and each deficient Part shall speedily grow out of the Part remaining.”. This human use of human beings is close to Hacking’s work in the pure philosophy of science - especially since experimentalists spend as much time dealing with the theory of errors as they do with the special theories of their subject - but it also connects to the thorniest human use of human beings: politics.

In the philosophy of politics, Hacking was also an innovator. Along with his older contemporary John Rawls, Hacking made politics a part of so-called ‘analytic philosophy’. But where Rawlsians (and most non-’analytic’ philosophers) had politics as part of the normative part of his philosophy, Hacking had politics as part of the positive side. In The Emergence Of Probability and The Taming Of Chance, Hacking chronicled the rise of the kind of rule by numbers that characterizes the modern state.

Writing on the cusp of that rise, Hegel declared that he couldn’t believe the planets were dragged around the cosmos like a bunch of dumb children as the French materialist determinists had it. On the other side of the 19th century, Hacking quotes George Knapp making much the same complaint with two differences. First, for Knapp’s nip at the French schools, the boys being dragged around are not Mars and Jupiter but actual boys. Second, the laws pulling the boys around are not the certain laws of gravity but the stable laws of statistics. Franklin’s image of a self-healing society which overcomes the deadliest obstacles became what Hacking called ‘statistical fatalism’: if you reform one thief, another will just come after.

Just as the laboratory provides continuity through incommensurability, Hacking shows that it is the actual practice of biopower that creates continuity. As one example of how Hacking uses the practice of biopower as data for doing philosophy, we can look at his work on human kinds. 

A human kind is, its proponents claim, a kind of natural kind. A kind is a set of individuals who all fall under some other predicate (which I will call the synthetic predicate) who also fall under at least one other predicate which is not analytically entailed by the defining predicate. Almost like a “type of guy,” where the icons by which you can identify one mean something about their behavior, or vice versa. A natural kind is a kind which is stable: the individuals don’t spontaneously fail to fall under the synthetic predicate. The notion of kind was introduced by Mill for a deflationary account of scientific method. The snub-nosed animals do not form a kind, by Mill’s lights, because they share no peculiarities save those analytic to being a snub-nosed animal. But being a mammal - at the time, meaning an animal which has mammary glands - is a kind because all mammals have four chambered hearts, a fact not analytic to having mammary glands.

But are there such a thing as human kinds in this sense? Mill is ambivalent, though he says that most conventional classifications - such as Christian or Mathematician - are not kinds. Hacking argued that further human kinds - Christian, Mathematician, Pervert - are intrinsically different from natural kinds. A human member of a kind can decide whether he wants to be part of that kind - or not! Hacking looks at the examples of Multiple Personality Disorder and Autism. People with these disorders can read the narratives he analyzes, and Hacking’s meta-analyses, and react to them. In principle, the members can adopt enough different behaviors that all synthetic predicates of the human kind disappear, demoting the kind to mere predicative set.

Anyone familiar with the LockPickingLawyer is aware of a simpler example. The LockPickingLawyer is a YouTuber who picks locks, usually locks produced en masse and usually locks with severe design flaws. One of his most common victims is MasterLock, a large lock making concern known for its inexpensive and low quality locks. In principle, MasterLock could watch this channel and change their behavior, which would fundamentally alter the inexpensive lock market. After watching a few videos, “MasterLock makes bad locks.” feels like a law of nature but it is not. And yet, this is how we treat much of mankind.

In his mature philosophy, Ian Hacking would take a different argument against human kinds. Hacking would move away from the importance of ‘natural kinds’ and the underlying ‘analytic’ notion of “knowledge as justified true belief” towards a Poppernian/Platonic theory of conjectural knowledge. I happen to agree with this way of evolving the new experimentalism. But more than that, I believe that Hacking’s work represents a model of how philosophers should engage with science and history. He has provided me with inspiration for some of the essays in this very online magazine. I know Alex can say the same as well. I hope his work will continue to flower by inspiring philosophers all over the world.

“The object of our analysis is, not to provide a machine, or a method of blind manipulation, which will furnish an infallible answer, but to provide ourselves with an organized and orderly method of thinking out particular problems; and, after we have reached a provisional conclusion by isolating the complicating factors one by one, we then have to go back on ourselves and allow, as well as we can, for the probable interactions of the factors amongst themselves. This is the nature of economic thinking. Any other way of applying our formal principles of thought (without which, however, we shall be lost in the wood) will lead us into error.”
JM Keynes, Chapter 21, General Theory

Alright! This is the last chapter of the book that is any “work” at all. The last three chapters are all meditations on the broader social, philosophical, and political issues that come up as a consequence of following this description of the economy out to its end. The quote up top is pretty much the best summary of the approach to studying the economy that I try to take. Hopefully the math gets weirder and the description gets thicker over time as people look closer at it, but the basic idea should still, essentially, be the above. There’s just too much going on otherwise.

Prices

Keynes opens this chapter by pointing out that most previous economics (and, a hundred years on, plenty of economics today) would start by trying to work out a theory of value, and then figure out how to map that theory onto price. The goal of this chapter — and the reason it comes all the way at the end — is that Keynes sees how prices are much more often determined by the economic system, rather than determining the economic system. And as measures determined by the system, they constitute a way for it and its constituent parts to essentially check in with themselves.

In my opinion, “price” isn’t actually a natural category, which causes problems for the idea that a globally-valid bundle of “relative prices” would actually be globally meaningful. Instead, relative prices tend to describe something like “local curvature”: markets, production and labor all have internal structures that become external structures by being projected onto price. However, this is like applying the same transformation to a bunch of different things in different spaces. For our purposes here, it doesn’t matter if that transformation is thought of in a formal mathematical way, or in a simpler intuitive way, like how you can “yassify” any image by feeding it through a function that alters certain aspects of it in common ways regardless of the appropriateness of the input material. Different market structures weight different prices (and thus different information bundles about different relative prices) differently, and have different structures of response to changes in price.

What is so cool about “money” and “moneyness” is that it is simultaneously the means by which one can access any market regardless of internal structure and the means by which every market structures the relationship between its past, its present, and its future.

Since this is the last “serious” chapter, I am going to give what I think of as a potted history of the intertwined conceptual genealogies of Price Theory. Price Theory has always bugged me, because it seemed like exactly the wrong place to start asking questions about the economy. Yet, because buying and selling are the places every individual person is most likely to first see the economy in their daily life, it’s usually folks’ assumption that it must be where things are the most economic. I don’t think this is right.

Read more

On its face, this chapter gives an account of “The Employment Function,” which sounds at first like something so vague that there’s no way it can still be useful. It actually is though! For the whole book, we’ve been hearing that employment is a function of effective demand, as translated by the investment decisions of firms. This is the first place where he explains a little bit about how that function might plausibly work–not enough to write math papers about it, but enough to use it as an aid to thought. In this chapter, Keynes develops the idea of an employment function to show how it is in-principle possible to map the level and distribution of effective demand onto the physical capital goods and social and market-institutional arrangements of firms in a way that tells you something about how many workers will have jobs, and which sectors those jobs will be in. It’s a little bit IO (Industrial Organization) and a little bit I-O (Input-Output Economics). 

So, although Keynes has been telling us for the whole book that the level of effective demand drives the level of employment, it never really explained how it envisioned the causal structure of that functional relation. Here, Keynes makes a fairly simple point: the amount of employment that a given level of effective demand translates to is dependent on the physical and institutional structure of the economy as well as the allocation of that demand on the consumer side. However, at this level of conceptual complexity, it is difficult to give closed-form structural equations that describe how two kinds of things will “always” interact or think through general equilibrium. This is something that has historically made it relatively difficult to capture Keynes’ ideas in a handful of equations in a textbook model. It’s more like a method for making Sankey diagrams than anything else.

Read more