The archive.

A move in chess reshapes the whole board; so, Shackle insists, does a genuinely crucial economic decision, one that cannot be repeated because it alters the very conditions under which any later choice would occur. Replying here to critics of Expectation in Economics, he defends the distinction between unique, isolated, and crucial trials against those who would treat rival imagined futures as additive terms in a single statistical expectation. Mutually exclusive outcomes, he argues, cannot be summed like fractions drawn from an urn; the chooser confronts an act, a moment, and a set of hypotheses. Against Baumol and Graaff he deploys his focus-value method, the φ-function, and potential surprise, rejecting 'degree of belief' as any general rescue of probabilistic ordering.

Now when an experiment, a question about the future, is unique, isolated, or crucial, it does not make sense to add together its rival hypothetical outcomes or answers.

Twentieth-century economics did not merely add topics to an old canon; it replaced the image of a system tending toward stable equilibrium with one shaped by uncertainty, hesitation, and breakdown. Out of that upheaval Shackle draws a map, sorting inherited doctrine by the kind of time and knowledge each theory assumes: perfect adjustment, calculable dynamics, aggregative comparative statics, and the economics of uncertain expectation. The organizing question is temporal—whether a model treats time as timeless adjustment, a dated sequence, or agents' conjectures about futures that cannot be known. Keynes straddles the categories, formally comparative statics yet substantively a theory of imagined futures. The chart doubles as a proposed curriculum and as a warning against teaching incompatible assumptions as though they belonged to one unified doctrine.

Imagined future events still form an entirely distinct category, since they do not constitute a unique series.

Numerical probability divides a fixed unit of belief among rival hypotheses; that additive structure, Shackle contends, is exactly what makes it useless for describing genuine uncertainty, where several incompatible outcomes may each be perfectly possible with nothing known against them. The remedy proposed is potential surprise, a non-additive measure of disbelief that lets rival hypotheses coexist without competing for a common total. Dividing experiments into 'divisible' series, where frequency ratios can render an aggregate outcome knowable in advance, and unique 'non-divisible' acts, where such ratios are meaningless, he weighs an integrative decision rule, drawn via Ralph Turvey from Ingvar Svennilson, against his own focus-values solution and rejects the former as psychologically artificial. Expectation, he concludes, is an act of creative imagination, not rational calculation on incomplete data.

For a non-divisible, unique experiment it is plain that no frequency-ratio can have any meaning or relevance.

May an economist honorably use a theory he does not fully believe? Shackle answers that the alternative would abolish the discipline, since every usable theory remains partial, contestable, and interesting precisely because it is not final knowledge. Sincerity thus becomes disciplined awareness of a theory's limits rather than abstention from theory. The essay runs a sequence of tests—on the arbitrary boundaries that wall economics off from psychology and politics, on the incompatible pictures rival abstractions paint, on equilibrium as a mechanical borrowing, on econometrics and its dangerous phrase 'incomplete information.' Because its subject matter learns, imagines, and invents, economics can never treat fitted equations as eternal truths, and its practitioners, Shackle urges, should form an open craft rather than a guarded mystery.

Economics is not physics, it is psychics, the study of men with all their capacity for learning and experimenting and inventing and imagining.

Orthodox theory said a fall in interest-rates should quicken investment by raising the present value of future returns; businessmen questioned by the Oxford Economists' Research Group flatly denied noticing any such effect. Rather than discard the doctrine, Shackle narrows it. A future receipt must be discounted twice, once for deferment through the pure interest rate and once for doubt, and the two work very differently: interest bites hardest on distant, secure returns, which is why housing and other long-lived, dependable assets remain rate-sensitive. But where invention, fashion, and obsolescence truncate an asset's useful life, a swelling 'marginal rate of risk' absorbs the far future before the pure rate can act, leaving valuations almost unmoved by a one-point change. Entrepreneurs, attending to shifting orders and markets, simply never register interest as the cause of their decisions.

It was until recent years an accepted doctrine that changes of interest-rates powerfully influence the pace at which enterprisers, all taken together, extend or improve their equipment.

Gunnar Myrdal's 1933 essay on monetary equilibrium deserves, Shackle argues, a recognition equal to anything in interwar economics, and he reconstructs it to show why. Myrdal's achievement was to reset Wicksell's problem in time: the natural rate cannot be the observable yield on existing capital, since that yield, once the stream of expected net receipts is discounted at the current interest rate, is tautologically tied to it. The live variable is instead the gap between a projected plant's capital value and its construction cost, a valuation formed before building and therefore a matter of expectations. Equilibrium is recast as compatibility among plans, read alongside Hayek: not balanced aggregates but a state in which realized events force no one to remake their anticipations. Even aggregate equality of investment and 'waiting', Shackle shows, can conceal offsetting individual errors.

Monetary equilibrium in this meaning is a means of classifying the set, considered as a whole, of systems of expectations which are entertained, one system by each individual, at some one point of time.

Value theory, Shackle charges, quietly presumes perfect knowledge — that the buyer can see every satisfaction in advance — a fiction exposed by the very existence of information, a good worth having only because its contents are not yet known. The essay accordingly shifts the object of economic choice from satisfactions to actions whose consequences remain hypothetical. Each rival hypothesis carries a face-value, the gain or loss if it proves true, and a second variable measuring its claim on the imagination; numerical probability, he demonstrates through five separate objections, cannot serve as that second variable for unique, non-seriable decisions. His replacement is potential surprise: not a lesser degree of certainty but a positive recognition of some disabling incongruity, a scale on which any number of mutually exclusive hypotheses may all sit at zero.

But the theory of consumer’s behaviour assumes that we always know what we are going to get.

The entrepreneur who sinks his fortune into a single plant makes a choice no lottery can model — and it is mathematical expectation, the workhorse of investment appraisal, that Shackle attacks in this sequel. Multiplying outcomes by probabilities and summing them, he argues, is legitimate only where an experiment is divisible or seriable, so that a spread of results can be possessed as a statistical aggregate; one business commitment has no such structure. The textbook urn and the game of chance are closed worlds that bar by rule the very unknowns constituting reality. In their place stand focus-values — the strongest gain one can plausibly hope for and the gravest loss one must plausibly fear — standardized on a gambler's indifference map, where a steeper feared loss demands a larger promised gain. The framework recasts Kalecki's principle of increasing risk without objective probability.

When the course of action is a non-divisible non-seriable experiment, such an additive procedure loses entirely the relevance it has for a divisible experiment, and has only one claim to fall back on: that of being a compromise.

A Chinese sentry, Kwong Hui, weighs loyalty against treachery in a choice he can make only once — and for whom, as Shackle drily observes, a severed head is rather final. From Keith West's parable Shackle draws his standing objection to orthodox decision theory: frequency-ratio probability describes a series of repeatable trials, but says nothing about the single occasion whose outcome absorbs a person's whole future. Such crucial experiments may destroy the very conditions under which they were run, so they cannot be rerun. In their place he offers not calculation but imaginative appraisal, where rival hypotheses are ranked by their power to stir hope or fear and by their degree of potential surprise, and choice settles on a representative focus-gain and focus-loss. It is a founding statement of the Knight–Keynes–Shackle line dividing calculable risk from genuine uncertainty.

For a non-divisible non-seriable experiment the concept of frequency-ratios is wholly irrelevant.

Once the Exchequer acquires a duty to stabilize aggregate demand, the old arithmetic of matching revenue to authorized expenditure no longer suffices. Written in 1947 in the wake of Keynes, this essay treats every fiscal stream as a force acting on monetary demand relative to the supply of goods, and builds a pair of indices—deflative P and inflative Q—to measure the initial thrust of a specified receipt or disbursement before secondary reactions unfold. Shackle's taxonomy of pensioners, policemen, postmen, palace-builders, and paper-makers shows why a payment that adds no saleable output pushes prices up while a purchase for resale may prove deflative. A tax label alone, he insists, never fixes the direction of pressure; only the composition of spending does.

The Exchequer, in deciding the size, method and timing of its levies and disbursements, must nowadays be guided by two quite distinct sets of considerations.

'Profit,' Shackle observes, names two quite distinct things: the realized figure an accountant records and the forward-looking conjecture that induces an enterpriser to commit resources at all. Because production takes time, those resources must be specialized before the future market is known, and it is this unavoidable uncertainty that creates the enterpriser's double role as decision-maker and uncertainty-bearer. Written for accountants but aimed at economic theory, the essay dismisses both the rough 'best guess' and mathematical expectation, whose frequency ratios describe repeatable series but say nothing about founding a firm or building a factory. In their place stands potential surprise: ventures compared through focus-gain and focus-loss rather than a single maximized number. Timeless Walrasian equilibrium, he charges, excludes the very time, novelty, and monopoly from which profit springs.

In all production, because it takes time, there is an ineradicable uncertainty.

How can a person coherently expect to be surprised? Shackle's Economica note turns the paradox into a piece of logic. To expect an event, in his vocabulary, is to attach zero potential surprise to it — so a counter-expected outcome, one already imagined and rejected, cannot be the source of true surprise; only the genuinely unexpected, a possibility never before entertained, can. His resolution is the residual hypothesis: a deliberate heading for the possibilities one cannot yet specify, to which zero potential surprise may attach even as its detailed content is bound to astonish, like a nineteenth-century physicist confronted with a digital computer. The economics follows. When residual ignorance threatens widely separated gains and losses on a gambler's indifference map, doing nothing and holding cash may outrank any active policy — a theory of enterprise inhibited by news too obscure to interpret.

In contrast with this I define an unexpected event as one which has never been formulated in the individual’s imagination, which has never entered his mind or been in any way envisaged.