The Wage and the Worth

Imprint manuscript · July 2026 · revised after external review (Paper No. 008, Thread I — Motivation). Four internal adversarial rounds converged at v5; a coverage review then surfaced the retirement natural experiment, the goods-of-work and contributive-justice literatures, and the status-politics bridge, all added here. Response letters, corpus index, and the five prior manuscript versions are filed with the archive.

· · ·

“I must study Politicks and War that my sons may have liberty to study Mathematicks and Philosophy. My sons ought to study Mathematicks and Philosophy, Geography, natural History, Naval Architecture, navigation, Commerce and Agriculture, in order to give their Children a right to study Painting, Poetry, Musick, Architecture, Statuary, Tapestry and Porcelaine.” — John Adams to Abigail Adams, post 12 May 1780

· · ·

Abstract

Paid work has always delivered two goods at once, and because it delivered them together, we have argued about automation as though it threatened only one. The first good is the wage — subsistence, and a claim on the social product. The second is the worth — the standing that comes with a recognized contribution, the answer to what one does and therefore where one belongs. The worth is not a single thing but a bundle: structured time, a sphere of company, a shared purpose, the dignity of being counted, and the activity itself — the facets a century of research on unemployment keeps recovering. Of these, recognized standing is the facet the wage cannot buy and the policy debate cannot see, and it is the one this paper tracks most closely, while conceding that the decomposition — which facet carries most of the weight — is not yet settled and is itself a measurement task. Only paid work delivered the whole bundle at scale, which is why its components have gone unnamed; and the bundle is precisely what automation acts on. This paper argues that the social outcome of automation turns not on how much labor is displaced — the variable both the optimists and the pessimists track — but on whether institutions re-route both goods or only the wage. Displacement sets the magnitude of the shock; the re-routing architecture sets which basin a society falls into. A settlement that re-routes the wage and not the worth — of which unconditional cash alone is the paradigm case — is predicted to land in a distinct failure mode with two faces that fall on different people at once: for the long-term displaced, an anomic harm — provision without recognized contribution, a categorical loss an architecture can address, predicted conditional on a decomposition of the worth bundle this paper concedes is unsettled; for the newly displaced, who still see a working reference class, a separate positional harm — relative standing lost even where material need is met — that an architecture of recognition does not reach, and whose evidence base transfers relative-pay findings to relative standing on a proxy the paper concedes is unverified. Because displacement is gradual, the two are concurrent sub-populations rather than sequential stages, so both are live at once and neither waits on an equilibrium that may never fully arrive. The argument runs in three registers kept strictly apart — what is owed, what is measured, what can be expected — and commits to three falsifiers, the first of which is a measurement the literature has not yet made and whose instrument this draft can specify but not yet validate. The contribution is integrative and methodological: a reframe of the automation debate from the quantity of displacement to the architecture of re-routing; the worth distinguished from the wage and tied to the instruments that could measure it; and a discipline for arguing about a transition without forecasting its arrival. The claim is conditional on displacement, silent about its timing, and specific enough to be wrong about — and this revision marks, more sharply than the first draft did, the places it is most likely to be wrong.

· · ·

§ IThe Question and the Two Goods

There is an old line, usually misattributed to Jefferson, that frames the optimistic case for every labor-saving revolution. John Adams wrote it to his wife from Paris in 1780: he studied politics and war so that his sons might study mathematics and philosophy, so that their sons might study painting and poetry. It is a theory of progress as a generational escalator — each cohort’s labor purchasing the next cohort’s liberation from labor, until the descendants arrive at the activities a free person would choose for their own sake. Read as a prediction about automation, it is one pole of a debate that has not moved in two centuries. The optimist says the machines are the escalator: they take the drudgery, free the hours, and the freed hours convert into higher pursuits. The pessimist says the machines are something new — that this time the cognition itself is automable, the displacement structural rather than frictional, and the terminus not an ascent to poetry but a descent into mass unemployment, immiseration, and the disorder that follows when a society can no longer find a use for most of its members (Susskind, 2020).

The two camps appear to disagree about everything. They share one premise so completely that neither states it: that the outcome — escalator or descent — is set by how much labor the machines displace. The optimist’s case is that net displacement is small or self-correcting, so the escalator runs; the pessimist’s is that it is large and permanent, so the floor gives way. Both read the social outcome off the displacement curve. Both treat the technology’s reach as the variable that decides whether the grandchildren study art or fight over scraps.

Here is the assumption stated so that it can be doubted. The displacement rate sets the magnitude of the shock; it does not set the basin a society falls into. The same degree of automation can produce Adams’s escalator or its opposite, depending on a different variable — whether the institutions that catch the displaced supply both of the goods that paid work supplied, or only one. To see that this is a different variable and not a rephrasing of the first, the paper has to name what work delivers, and to insist on a distinction the bundled history of work has kept invisible.

Work delivers two goods. The first is the wage: subsistence, and with it a claim on the common product — the material side, the side the policy debate already sees. The second is the worth: the standing that attaches to a recognized contribution — the answer every social order asks its members to give, what do you do, and so where do you belong. The worth is not the wage and is not reducible to it; and it is not, this revision is careful to say, a single thing. The research on what unemployment destroys keeps recovering a bundle of non-material goods — a structure for time, a sphere of company beyond the household, a purpose larger than oneself, the standing of being counted a contributor, and the activity itself (Jahoda, 1982). This paper calls that bundle the worth, names recognized standing as its distinctive facet — the one a transfer cannot supply and the accounts cannot see — and tracks it most closely, while holding open, as an empirical question rather than a settled premise, how the weight is distributed across the facets. The standing facet is the one Smith placed at the center of his moral psychology when he observed that “man naturally desires, not only to be loved, but to be lovely” — to deserve regard, not merely receive it (TMS III.ii.1); and it is the one Rawls, locating the social bases of self-respect among the primary goods, called “perhaps the most important” (Rawls, 1971, §67). For almost the whole history of the wage, only paid work delivered the whole bundle at scale: the job was the socially legible certificate of contribution, the thing one could name when asked what one did, and the site that supplied the company and the structure and the purpose along with the pay. Because the goods arrived in one envelope, no one had to ask whether they were separable, or which of them did the work. Automation is the question made unavoidable, because it acts on the one institution — paid work — that bundled them.

The claim, stated so it can be wrong. Paid work delivers two goods, a wage and a worth, the second a bundle of non-material goods of which recognized standing is the distinctive facet; only paid work delivered both at scale; automation can sever them. The social outcome of automation is therefore not determined by the displacement rate but by whether institutions re-route both goods. A society that re-routes the wage and not the worth — income without a recognized channel of contribution, of which unconditional cash alone is the paradigm case — is predicted to land in a distinct failure mode: materially provided for and contribution-starved. That failure has two faces, which §6 and §7 keep apart and assign to concurrent sub-populations rather than to sequential stages: for the long-term displaced the deprivation is mainly anomic and categorical — the loss of recognized contribution itself — which an architecture can address; for the newly displaced, who still see a working reference class, it is also positional — relative standing lost net of material provision — and that positional, political face is the one no architecture of recognition reaches. The failure will be partly invisible to instruments that read income, employment, and output, because by those instruments a provided-for population reads as a served one. Three observations would disconfirm this claim, named in §7 and §8 and collected in §9. The first — that income replacement, properly measured, reproduces what work supplied — is the one the paper would most cleanly die of, and the paper concedes at the outset that the measurement which would settle it has not yet been made.

A word on method, because the method is half the contribution. Every non-trivial sentence belongs to one of three registers. What is owed: claims about the social contract, the two goods, and the philosophers who named the worth before economics dropped it — Smith, Rawls, Polanyi, Arendt — cited at passage level, with the paper’s own constructions labeled as constructions. What is measured: claims about what work supplies beyond income and what its loss does net of income, each carrying its magnitude, its venue, and its contest, with the laboratory-to-field and pilot-to-permanent discounts stated wherever the record is partial — which, for the contemporary cash evidence, is nearly everywhere. What can be expected: projections, each carrying the trend it extrapolates, the assumption it requires, and the observation that would disconfirm it. Two standing rules keep the registers apart. The paper asserts no timeline — no date for mass displacement, no probability of artificial general intelligence; its structure is conditional throughout, if displacement is large and durable then the deciding variable is the worth channel. And the paper indulges no nostalgia: it concedes in full that paid work has been alienating, coerced, degrading, and rationed by class and race and sex — that Smith himself, in the Wealth of Nations, watched the division of labor render the worker “as stupid and ignorant as it is possible for a human creature to become” (WN V.i.f.50) — and that for a large share of the workforce the worth a job conferred was thin. The claim is about a channel and its severance, not a golden age, and §4 will follow that concession to its uncomfortable consequence rather than retreating from it.

§ IITwo Caricatures to Clear

Two pictures dominate the conversation this paper enters, and each must be cleared before the argument can run.

The first is the cornucopian: automation has destroyed jobs for two hundred years and employment has always recovered, therefore it always will, and the fear of technological unemployment is the oldest fallacy in the discipline. The concession is real and the paper makes it in full. The lump-of-labor assumption — a fixed quantity of work, so that a machine doing some permanently subtracts that much human employment — is indeed a fallacy; total work is not fixed, and two centuries of mechanization did not produce mass permanent unemployment (Autor, 2015). American agricultural employment fell from roughly forty percent of the workforce to about two percent across the twentieth century, and the displaced did not become a permanent idle class; new tasks and new sectors reinstated the demand for labor (Acemoglu and Restrepo, 2019). But the cornucopian hides a substitution inside the word “jobs.” That employment recovers is a claim about the wage channel. It is silent on the worth: re-employment at degraded contribution satisfies the count and still fails this paper’s claim, and the record of the American deindustrial belt and the China-shock counties is precisely the record of places where the jobs number partly recovered and the standing did not (Autor, Dorn and Hanson, 2019). The lump of labor is a fallacy about the wage. There is no equivalent theorem guaranteeing the worth.

The second is the terminator: the machines are coming for all of it, the displacement total and imminent, the result collapse. This fails the paper’s own no-prophecy rule, and the paper declines it as firmly as the cornucopian. What is measured — §5 will be precise — is exposure, not destruction: the share of tasks a technology could in principle touch, which is a ceiling on impact and a poor predictor of it. The single largest swing in the literature is not between optimists and pessimists about the future but between two ways of measuring the present: occupation by occupation, Frey and Osborne’s much-quoted forty-seven percent of American jobs “at risk” (Frey and Osborne, 2017); task by task — because most occupations are bundles of tasks, only some automatable — about nine percent (Arntz, Gregory and Zierahn, 2016). A five-fold disagreement produced by a change of method is a warning against any argument resting on a displacement headcount.

A discipline follows from clearing the second caricature, and the paper binds itself to it here because a later section will be tempted to break it. Exposure measures are informative about composition — which tasks and occupations a technology reaches — and uninformative about magnitude and timing — whether, when, and how much displacement actually occurs. The terminator’s error is to read magnitude off a composition measure. The paper will, in §8, use the composition of exposure to argue something about where displaced labor could go; it may not, and will not, use it to assert that displacement of a given size arrives. The line between composition and magnitude is the line the no-prophecy rule draws, and §8 is written to stay on the legitimate side of it.

What the two caricatures share is the assumption of §1: that the technology sets the outcome. The cornucopian reads a benign outcome off the recovery of the jobs number; the terminator reads catastrophe off the size of the exposure. Neither has named the second good, and so neither can ask the question that decides between escalator and descent — not how much work will the machines take, but what happens to contribution when they take it. Clearing the caricatures is the first statement of the thesis.

§ IIIThe Bundle Named, I: What Is Owed

What did paid work supply beyond a paycheck? This section names the second good in the register of what is owed, and draws on four witnesses who each named a part of it before economics decided it could manage without one. A note on method governs the section: the witnesses are read for what their texts will bear, and where this paper turns a witness against the use it is usually put to — as it must, twice — it says so.

Smith supplies the sharpest distinction, and a warning the paper takes rather than evades. Against the modern habit of folding all regard for others’ opinion into “reputation,” Smith split the desire for approval in two: “man naturally desires, not only to be loved, but to be lovely; or to be that thing which is the natural and proper object of love” (TMS III.ii.1). The desire for praise is satisfied by approval, deserved or not; the desire for praiseworthiness only by deserving it, observed or not. The distinction matters to this paper twice over. First, it identifies the standing facet of the worth: not applause but the regard one has earned, the answer to the contribution question that would survive an honest accounting. Second — and this is the use the paper makes of it later, against an objection the first draft did not face — praiseworthiness is not a positional good. To be praiseworthy is to deserve regard; it does not require out-ranking anyone, and everyone in a society can be praiseworthy at once, where not everyone can be eminent. Smith was equally clear that this standing is among the strongest motives a person has — what drives the pursuit of betterment is less the comforts of wealth than the wish “to be observed, to be attended to, to be taken notice of with sympathy, complacency, and approbation,” that “it is the vanity, not the ease, or the pleasure, which interests us” (TMS I.iii.2). Worth is load-bearing in motivation — the substrate this thread studies — which is why the paper belongs here.

Rawls supplies a normative upgrade and, read honestly, a correction the first draft needed. The social bases of self-respect are a primary good, and Rawls ranks them first among equals — “perhaps the most important primary good” (Rawls, 1971, §67) — on the ground that without the conviction that one’s life and contributions are worth carrying out, nothing else in a life of plans gets purchase. But Rawls does not locate those bases in work or contribution. He locates them in the equal basic liberties, the public status of equal citizenship, and the plurality of associations in which a person finds others who affirm what they do. This cuts against any glib claim that the job is the irreplaceable source of self-respect — and the paper accepts the correction, because it cuts in the paper’s favor. If the social bases of self-respect can rest on citizenship and association rather than on the wage, then worth can in principle be carried by a settlement that is not a labor market — which is not an embarrassment to the thesis but the strongest available warrant for R3, the legitimacy narrative of §6, and for the claim that the worth can be deliberately re-routed rather than merely mourned. Rawls is evidence that the decoupling this paper contemplates is conceptually possible. What he does not supply, and what §4 must, is evidence about whether, in the society we actually have, work is currently doing the work.

Arendt supplies the distinction that keeps the goods from collapsing into one. Labor is the cyclical toil that meets the body’s needs and leaves nothing behind; work fabricates the durable artifice; action is the disclosure of who one is among others, in the space where standing is won (Arendt, 1958). The hinge the paper takes is narrow: the part of work that confers worth is not the part that meets the body’s needs. Automating labor — relieving toil — need not automate worth, and replacing the wage need not replace the standing. The goods are conceptually independent; a technology or a policy can move one without moving the other. That independence is what makes the central claim possible and what §4 measures.

Polanyi supplies the mechanism of reaction. Labor is a “fictitious commodity” — human activity is not a thing produced for sale, and organizing society as though it were inflicts a dislocation society will not tolerate (Polanyi, 1944, ch. 6) — and the result is the “double movement,” the protective counter-movement that the drive to extend the self-regulating market provokes (Polanyi, 1944, ch. 11). The counter-movement can be humane, as the welfare state was, or ugly, as Polanyi watched it become in the 1930s. A transition that severs the worth and re-supplies only the wage is a fresh attempt to treat the human relation to contribution as a pure commodity question, settleable with money; Polanyi’s prediction is that the attempt summons a reaction whose character is not guaranteed. §7 puts that prediction in the measured register, and concedes there how much weight the prediction can bear.

Two disciplines govern the section. Nothing is attributed to a witness that the text will not bear; and the central construction — that the wage and the worth are separable goods that production bundled contingently, and that the worth is itself a bundle of facets — is labeled the paper’s own. The witnesses establish that the worth is real, that it is owed, that it is distinct from subsistence, that severing it provokes a reaction, and — the corrections from Smith and Rawls — that the standing facet is non-positional and that its social bases need not run through a wage. Whether work actually supplies the bundle, which facet carries the weight, and whether money alone restores it, are not questions a text can settle. They are the questions of §4.

§ IVThe Bundle Named, II: What Is Measured

A paper that intends to say the worth can be severed from the wage must establish that there are two goods rather than one, and must be honest that the evidence identifies a bundle while the paper’s distinctive interest is in one facet of it. This is the sobriety chapter, and the revision has made it more sober in three places the referee was right to press: what the evidence can and cannot decompose, where its keystone case comes from, and what the quantitative record actually weighs.

The separability evidence first. When the single factory of Marienthal closed in the early 1930s and the Austrian town fell into unemployment while continuing to draw some relief, Jahoda, Lazarsfeld and Zeisel went to watch what happened to people who had lost work but not, entirely, income (Jahoda, Lazarsfeld and Zeisel, 1933/1971). What they found was not the leisured self-cultivation the escalator promises but collapse: apathy and resignation, a measurable breakdown of the structure of time, withdrawal from the town’s associational life, a narrowing of horizon. Out of that tradition Jahoda named the components of the lost good — time structure, social contact, collective purpose, status and identity, enforced activity (Jahoda, 1982). And here the first draft overreached, in a way worth correcting in plain view: it took evidence about the loss of the whole bundle and attributed it to the one component it had chosen to elevate. The honest statement is that the separability literature identifies a large non-pecuniary cost of joblessness and does not, on its own, decompose that cost across Jahoda’s five functions. Fryer’s critique — that material and agency deprivation, the loss of control over one’s own future, do much of the damage — is not walked past in this draft but conceded: it is a live and well-supported rival to the institutional-functions account (Fryer, 1986). What the paper claims, and no more, is that recognized standing is one facet of the bundle, the facet a transfer most clearly cannot supply, and a plausibly important one; the decomposition is an open empirical question, and §6 folds it into the paper’s first falsifier rather than assuming its answer. The reframe costs the paper a clean slogan and buys it a defensible construct — and it does not weaken the re-routing argument, because the facets are not five unrelated goods that happened to co-occur. They co-occurred because the job was a site of recognized collective contribution, and an institution that restored that — R2, below — would restore the structure and the company and the purpose along with the standing, which is the point.

The cleanest demonstration that some non-pecuniary good exists comes from holding income fixed and watching wellbeing anyway. Clark and Oswald found the unemployed dramatically less happy than the employed controlling for income — joblessness depressing wellbeing by more than almost any other circumstance studied, divorce included, with the gap not closing when the income gap closes (Clark and Oswald, 1994). The decisive follow-on is adaptation: people return toward baseline wellbeing after marriage, divorce, widowhood, after most of what life does to them — but do not adapt back to unemployment, which scars for as long as it lasts (Clark et al., 2008). Non-adaptation net of income is the fact hardest to explain if the wage were the whole story. The magnitudes are not small: the meta-analysis of several hundred studies puts the distress difference between unemployed and employed at about half a standard deviation, with roughly a third of the unemployed showing psychological problems against roughly a sixth of the employed, and the longitudinal studies — distress worsening on job loss, recovering on re-employment — supporting a causal reading (Paul and Moser, 2009). The revision adds the cleanest causal designs the first draft omitted, because they strengthen the individual-level claim: mass-layoff and plant-closure studies, where workers lose jobs for reasons unrelated to their own health, find large and lasting losses — in lifetime earnings (Jacobson, LaLonde and Sullivan, 1993), and in life itself, with displaced workers suffering elevated mortality for years after (Sullivan and von Wachter, 2009; Eliason and Storrie, 2009). These designs answer the selection worry — that poor mental health causes joblessness rather than the reverse — that contaminates the cross-sectional record. They do not, and the paper says so, rescue the decomposition: clean causal evidence that job loss has non-pecuniary costs still does not identify which non-pecuniary good is lost.

One design does bear on the decomposition, and the first five drafts of this paper missed it — a coverage failure the external review named and this revision repairs, because the missing study is the closest thing in the published record to the mechanism this paper claims. Hetschko, Knabe and Schöb followed German panel respondents across the threshold into retirement, comparing those who entered it from unemployment with those who entered it from work (Hetschko, Knabe and Schöb, 2014; SOEP, 711 and 1,603 transitions respectively). For the formerly unemployed, almost nothing observable changes at that threshold: they were not working before and are not working after, their leisure is identical on both sides (23.94 hours either way), and their income barely moves. What changes is the category — and the socially legible answer to what they do. Life satisfaction rises from 6.32 to 6.59 on the eleven-point scale (p < .01), against 7.30 to 7.33 for those retiring from employment, whose income falls and whose leisure jumps by six and a half hours a day. In first differences, controlling for income and leisure change, the effect of having been unemployed is 0.33 points (SE 0.11); among the 110 respondents legally compelled to retire once pension-eligible — the subsample with exogenous timing — it is 0.75 (SE 0.18). The authors’ reading is explicit and is not this paper’s imposition: retirement “is associated with a switch in the relevant social norms,” and what the unemployed lose is not only income but conformity with a norm whose violation costs “a loss in purpose, a low recognition from others and a failure to meet one’s own self-concept.”

This is the categorical-standing mechanism, measured — non-employment held constant, income controlled, only the social category moved — and it is the nearest existing analogue to the study F1 asks for. Four honesties travel with it. The effect is male-specific: 0.53 for men, −0.01 and null for women, consistent with the same literature’s finding that the work norm binds asymmetrically by gender, and a real limit on how far the mechanism generalizes. No specification controls the respondent’s own health, so the paper does not claim it does. An independent reanalysis of the same panel finds a short-term satisfaction gain for both transition groups, more pronounced for the formerly unemployed — which qualifies the sharp “the employed gain nothing” reading without touching the differential (Wetzel, Huxhold and Tesch-Römer, 2016). And it is one country, one norm regime, with no non-German replication of the design located.

Retirement complicates the thesis in two productive ways the paper now states rather than absorbs. It is an existing, society-scale instance of R3: a legitimacy narrative that already absorbs exit from paid work for something approaching a third of the adult lifespan, and absorbs it well enough that crossing into it raises the standing of people whose material circumstances did not change. That is the best evidence available that deliberate re-routing of the worth channel is possible — and simultaneously a qualification of this paper’s “only paid work delivered both goods at scale,” since retirement has been delivering the categorical good, at scale, to non-workers, for as long as the institution has existed. And it suggests that standing may be partly a stock — earned by past contribution, retained after exit — rather than purely a flow. If so, §6's stock/flow split of the displaced misses a third sub-population, and §6 now names it.

One counterweight belongs beside this result rather than after it, because it cuts the other way on the same instrument. The cleanest evidence on unearned money finds that a large lottery win raises overall life satisfaction — 0.037 SD per $100,000, undecayed over five to twenty-two years — while leaving happiness and mental health statistically indistinguishable from zero (Lindqvist, Östling and Cesarini, 2020). Unearned income does move the same eleven-point instrument the retirement study uses. The honest statement is therefore narrower than “money cannot buy standing”: money moves evaluative life satisfaction, and the retirement result shows that a category change moves it too, with income held still. Both are real; neither subsumes the other; and the decomposition of §4 remains open.

Now the deflation, at full strength, because the paper’s credibility depends on §4 surviving it — and the revision concedes more here than the first draft did. First: money helps, really and measurably. The meta-analytic effect of cash transfers on wellbeing and mental health is positive, scales with the size of the transfer, and fades once the payments stop (McGuire, Kaiser and Bach-Mortensen, 2022; Wollburg et al., 2023). The worth channel does not make the wage channel optional. Second: the keystone case is from the wrong conditions, and the paper now says so. Marienthal was catastrophic-Depression unemployment with minimal relief and heavy stigma — the opposite, on nearly every axis, of the scenario the paper worries about, a permanent income floor with, in the better constructions, a legitimacy narrative that honors non-waged contribution. Marienthal is therefore an existence proof that worth-loss can devastate, and that stigma — itself a standing phenomenon — is part of how; it is not a forecast of life under a secure, de-stigmatized floor, and Projection 1 is re-scoped accordingly. Third, and most serious, the strongest version of the worth-loss thesis meets a powerful rebuttal at the population scale. The deaths-of-despair literature ties rising mortality among less-educated Americans to the long erosion of the life a high-school education once secured (Case and Deaton, 2015; 2020); but Ruhm finds economic conditions explain “less than one-tenth — quite possibly zero” of the rise in drug deaths, the dominant driver being the drug environment (Ruhm, 2019; concurring, Currie and Schwandt, 2021). The honest statement is that both sides credit the opioid supply shock for the drug deaths and dispute how much economic and status decline adds, and that the contest bites hardest exactly where the paper would most like to point.

The referee pressed a fourth deflation harder than the first draft allowed, and it is conceded here in the body rather than defended away. The quantitative record is, at best, neutral on the worth-specific claim. Finland’s basic income raised life satisfaction and reduced strain with an essentially nil employment effect (Kangas et al., 2020) — money improving the wellbeing of the still-unemployed, which shows the income channel is real even where work is not restored. The paper’s first draft parried this with “Finland did not measure standing either,” and the referee is right that the parry proves too much: a move that neutralizes the strongest evidence against the thesis would equally neutralize evidence for it. The defensible position is narrower and is the one the paper now takes. Finland speaks to a different margin — cash among the unemployed, not the gap between employed and unemployed net of income — so it does not directly refute the separability evidence; but the deeper reason the quantitative record cannot yet adjudicate the worth claim is the decomposition problem above. Even the evidence for a non-pecuniary good does not isolate standing, and the cash trials were not built to. The quantitative record establishes that a non-pecuniary cost of joblessness exists and that money substantially mitigates distress; it does not establish that the residual is worth rather than security, company, structure, or control. The paper’s confidence in Projection 1 is reduced to match.

One consequence of the no-nostalgia rule, conceded as a paragraph and not a clause because the referee was right that it changes the welfare sign. The wellbeing and non-adaptation results are population averages, and they average over enormous variance in how much worth a job confers. For the substantial share of the workforce in precarious, alienating, or stigmatized work — the jobs Graeber’s polemic called “bullshit jobs” and that Smith’s own Wealth of Nations anticipated when it described the division of labor stupefying the worker (Graeber, 2018; WN V.i.f.50) — severance of a low-pay, low-worth job in exchange for security and time may be welfare-improving. If many jobs confer little worth, the aggregate effect of the well-funded vacuum is not uniformly negative; it is indeterminate, and for a non-trivial population positive. This does not defeat the thesis — the claim is about the worth channel’s existence and its absence from the policy menu, not about a uniform welfare loss — but it bounds it, and the bound belongs in the body.

The chapter’s honest summary. The worth is real — visible wherever the wage was replaced and the standing was not, from Marienthal’s relief rolls to the scarring that income does not heal and the displaced-worker mortality that selection cannot explain. It is a bundle, of which recognized standing is the distinctive and least-measured facet, and the decomposition is unsettled. It is unequally distributed — for some jobs its loss is a wound and for others a release. And at the population scale the strongest version of the thesis is contested, with a powerful supply-side rival the paper concedes. With that sizing fixed, §5 asks where the severance stands and what is being built to catch it.

§ VThe Documented Present

Everything in this section is measured or dated. It comes in two halves — what the machines are doing to the wage’s source, and what is being built to replace the wage — and both contain deflation.

The displacement record. Exposure is broad and rising; realized displacement is, in the aggregate, not yet visible; and the paper holds those facts apart. The most-cited estimate finds about eighty percent of American workers with at least ten percent of their tasks exposed to large language models, and about nineteen percent with at least half exposed — the authors stating, in terms the headlines dropped, that exposure “does not distinguish between labor-augmenting and labor-displacing effects” and making no prediction about timing (Eloundou et al., 2024). Exposure is a ceiling, and — per the discipline fixed in §2 — it is informative about composition, not magnitude. What is realized so far points more to augmentation than destruction: the cleanest field study, of customer-support agents given an AI assistant, found a fifteen-percent average productivity gain concentrated among the least experienced, who gained about thirty percent (Brynjolfsson, Li and Raymond, 2025). Macroeconomic estimates of realized effect are modest — on one accounting, a total-factor-productivity gain well under a percent over ten years (Acemoglu, 2025) — against far larger industry projections, a gap the paper notes and, under the no-prophecy rule, declines to adjudicate. The early displacement signals are narrow, recent, and contested: a relative decline in automation-prone freelance writing and coding posts after capable models shipped (Demirci, Hannane and Zhu, 2025), and a disputed finding of relative employment decline among the youngest workers in the most exposed occupations (Brynjolfsson, Chandar and Chen, 2025, a working paper challenged by Federal Reserve and labor-survey analyses, cited as a signal under contest). The paper takes no position on when, or whether, the aggregate picture changes.

It records one compositional fact, and — having bound itself in §2 — uses it only as composition. Every prior automation wave hit routine, codifiable, middle-skill work, and the escape ran toward tasks machines could not do. The polarization literature is precise about the shape: automation hollowed the middle and pushed labor toward both tails, including a large expansion of lower-skill in-person services, not a clean march upward (Autor, Levy and Murnane, 2003; Autor and Dorn, 2013). This wave’s exposure has a different composition: it is high for higher-income, higher-education, cognitive and creative occupations (Eloundou et al., 2024). The paper claims from this only what composition licenses — that if displacement of significant size occurs, its incidence would fall differently than before, on the very tasks the optimist’s escalator points toward — and it claims nothing about whether displacement of that size arrives. §8 builds the re-bundling question on that conditional, and concedes there how weakly the conditional can be tested in real time.

The replacement record. What is being built to catch the displaced is, almost entirely, a wage instrument, and the most rigorous evidence shows it leaving the worth channel unmeasured. The flagship is the OpenResearch unconditional-cash study, the best-powered and longest American trial of its kind: a thousand dollars a month for three years against an active control (Vivalt et al., 2024). The revision states its findings more carefully than the first draft did, because the referee was right that the prose ran ahead of them. The effects on work were small: a roughly two-percentage-point fall in labor-force participation and a little over an hour less worked per week, with the great majority of recipients continuing to work — not a picture of a population stepping off the wage into a vacuum. The early improvements in stress and food security were “large but short-lived”; there was no detectable effect on physical health; and gains in subjective wellbeing decayed across three years toward insignificance. The study did not measure purpose, standing, or recognized contribution, and its authors, reaching to explain their own null, reached toward the variable this paper names — that unearned income may not supply the “accomplishment and self-esteem” that earned income does. That authorial gloss is a discussion-section hypothesis about a null, not a measured result, and the paper now treats it as exactly that: an unmeasured variable can fail to be refuted but cannot be confirmed. The honest reading is that the flagship neither confirms nor refutes the worth claim — it leaves it unmeasured, and notices the gap on its way past.

The rest of the cash record tells the same story from other angles. Where standing was approached at all it was approached qualitatively: the Stockton demonstration and the Baby’s First Years trial report, in interviews rather than primary quantitative outcomes, gains in agency, dignity, and meaning that recipients tied to the unconditional character of the money (West and Castro, 2021; Castro and West, 2023) — even as Baby’s First Years’ headline quantitative result, an early finding on infant brain activity, failed its preregistered four-year follow-up (Noble et al., 2025). The 1970s negative-income-tax experiments, the deepest well of guaranteed-income evidence, measured labor supply and family outcomes and never measured standing at all. Across the literature — flagship to pilot to meta-analysis — the instruments are cash, hours, employment, health, and generic wellbeing. The worth is the variable the replacement debate does not have a column for, which is the condition §6 turns into a framework and §10 into an agenda.

The section’s structural observation: the severance is legible on both sides of the ledger, asymmetrically. On the loss side, exposure’s composition points at the contribution-rich upper rungs, though its magnitude is unknown. On the repair side, the entire policy apparatus is a wage apparatus, silent on the worth. The second good is unmeasured where it might be lost and unaddressed where the loss would be caught.

§ VIThe Hinge: The Severed Bundle

This section is the framework. It is labeled as such: what follows is not a finding but the structure that organizes the findings and generates the falsifiers. The revision rebuilds it in two places the referee was right to press — what the worth is, and whether it can be supplied at all.

Define the two goods as variables. The wage, W, is subsistence plus a claim on the social product — re-routable by transfer, the channel the debate already instruments. The worth, Θ, is the non-material good of recognized contribution — and Θ is a vector, not a scalar: its components are the facets §4 named (structured time, company, shared purpose, recognized standing, activity), with recognized standing the facet a transfer cannot supply and the accounts cannot see. The paper’s central construction — labeled construction — is that paid work delivered the pair (W, Θ) jointly and contingently: jointly, which is why the components were never named apart; contingently, because the joint delivery is an artifact of how production has been organized, not a law of nature. Automation acts on the organization of production. Therefore it can sever W from Θ.

The standing facet requires one further decomposition, because without it the constructive program is incoherent — this is the positionality problem the first draft did not face. Recognized standing has two components that must be held apart. A categorical component: whether one has a recognized, culturally legible answer to the contribution question — whether one is counted a contributor at all. And a positional component: whether one’s answer is high-ranked relative to others. The categorical component is non-positional; it is Smith’s praiseworthiness (§3), which everyone can possess at once, and it is the social bases of self-respect that Rawls grounds in equal citizenship, securable for all by construction. The positional component is the zero-sum good of the long tradition the first draft ignored — Veblen’s invidious distinction, Hirsch’s positional goods, Frank’s rank (Veblen, 1899; Hirsch, 1976; Frank, 1985) — and it cannot be raised for everyone, because rank is relative by definition. The distinction disciplines the whole constructive program: an architecture can supply the categorical component and cannot supply the positional one, and a paper that promised to give everyone high standing would be asking for everyone to be above average. The recognition tradition in social philosophy (Taylor, 1994; Honneth, 1995) is precisely an argument that the categorical component — being recognized as a contributing member — is real, is owed, and is not inherently scarce; that is the component R2 targets. Note that broadening Θ from “standing” to the full bundle (above) already shrinks the positional exposure, because most of the facets — structured time, company, purpose, activity — are absolute goods, not ranks. Positionality bites only on the standing facet, and within it only on the positional component. What survives as a genuine constraint is real and the paper concedes it: the positional component is irreducible, no architecture removes it, and it is one reason the dark substitute of §8 — status competition routed through tribe and crime — never fully closes.

That concession would be a loose end if the post-work scenario were a single state; it becomes a structural distinction once one notices the positional component is pulled in opposite directions for two groups who exist at the same time. Displacement is gradual, so a long transition always contains both a stock of the long-term displaced and a flow of the newly displaced — and the paper treats these as concurrent sub-populations, not as sequential stages, which removes any dependence on a universal equilibrium ever arriving. For the newly displaced, a working reference class is vivid and positional comparison is live; for the long-term displaced, the reference class has psychologically faded — the finding that joblessness hurts less where more of one’s reference group shares it (Clark, 2003) runs, for them, in the architecture’s favor — leaving the categorical deprivation, the absence of recognized contribution as such, as the residual harm. A third sub-population belongs beside those two, and it follows from the retirement result of §4 rather than from the stock/flow logic itself. If standing is partly a stock — accumulated by past recognized contribution and retained across exit, which is how retirement can confer it on non-workers — then the people with no stock to draw on are the never-employed young: those displaced before entering the labor market at all, who hold no accumulated contribution, receive none of the fading-reference-class relief the long-term displaced eventually get (their reference class is working and stays working), and have no legitimacy narrative available to them, since retirement’s narrative is explicitly one of contribution already rendered. This population is the one the paper’s own §5 exposure record makes concrete, since the contested early-career signal is precisely about entry-level cognitive work. It is also the population for which R2 would have to create standing rather than re-route it — the hardest case in the architecture, and the one an index would detect last, because the never-counted register as absent rather than as declining. The paper names the sub-population and claims no magnitude for it; the projections of §7 are not extended to cover it.

The split sorts the paper’s two dangers by who rather than when. The long-term-displaced danger is anomic and categorical, and it is the one R2 is built to address. The newly-displaced danger is positional and political, and R2 does not reach it: it is a problem of distribution and reference, not of recognition, and the paper says so rather than letting R2 pose as its solution. Both are live at once, and both tools are needed at once. §7 assigns each projection to its sub-population accordingly.

Θ admits measurement, and the revision discharges the claim the first draft only asserted. The components map onto instruments that already exist: occupational-prestige scales index the positional component (Treiman, 1977; the ISEI), which is exactly why they are the wrong tool for the categorical one — a scale that ranks occupations cannot register a good that is not a rank; subjective social status instruments such as the MacArthur ladder (Adler et al., 2000) capture perceived position; the meaningful-work literature operationalizes experienced contribution (Hackman and Oldham, 1980); eudaimonic-wellbeing scales (Ryff, 1989) and the “mattering” construct in social psychology (Rosenberg and McCullough, 1981; Elliott et al., 2004) come closest to the categorical component — the sense of being counted, of mattering to others, independent of rank. The proposed operationalization of Θ for a population is therefore not a restatement of the concept but a constructible index: the distribution of a mattering-and-recognized-contribution measure, reported with its inequality, alongside but distinct from the prestige distribution that the labor market already produces. The missing column the abstract promised is now nameable as a column: not occupational prestige, which exists and is positional, but population mattering, which is categorical. One honest limit travels with the proposal, because it cuts at the construct: the mattering and subjective-status instruments measure a felt, received state — whether a person believes they matter — which is closer to Smith’s praise (perceived approval) than to his praiseworthiness (deserved standing, real whether observed or not), and a self-report scale cannot reach the social-fact dimension of recognized contribution, whether the surrounding society actually counts one. The index as specified therefore captures the subjective facet and not, on its own, the social-recognition facet the theory is built on. The claim is accordingly specifiable now, validation pending — pending evidence that felt mattering proxies deserved standing, and a social-recognition component the self-report scales omit — which is a real advance over “nameable” and an honest retreat from “buildable now.” §10 states the agenda; F1 (§7) stakes the claim on it.

The claim of §1 can now be stated as structure. The automation debate tracks D, the displacement rate. The paper’s variable is R, the re-routing architecture — whether the institutions that catch the displaced re-supply both W and Θ, or only W. D sets the magnitude of the shock; R sets the basin. This is, deliberately, the variable the parent paper isolated: Beyond the Binary argued that the reward architecture is the load-bearing thing an economy must design rather than inherit, and engineered one for a market that still runs on paid work. This paper takes that architecture to the case that breaks it — where the wage-bearing job, the channel that delivered both goods, is what automation withdraws. The parent is cited, not re-argued.

A humane re-routing has three requirements; the debate is loud about the first, silent about the second, naive about the third.

R1 — the wage channel. Decouple subsistence from labor. The requirement the debate sees, and technically the most solved: cash works as income (§4). R1 is necessary, and the paper says so without hedging, because the argument is constantly mistaken for an argument against unconditional income and is the opposite of one. R1 is necessary. It is not sufficient.

R2 — the worth channel. Supply a recognized, non-waged way to contribute that the culture counts, restoring the categorical component of standing — and, because a genuine contribution institution is a site and not a badge, the structured time, company, purpose, and activity that came bundled with it. This is not a requirement the debate lacks a word for — the debate has several, and §6's ancestry paragraph now names them: contributive justice, recognition, the goods of work, the purpose problem. It is the requirement the debate has no instrument for, and no falsifier: the good cash does not deliver (§4's flagship null), only qualitatively grazes, and no trial has yet been built to detect. It is Beyond the Binary's machinery ported one step over, from how output is rewarded to how contribution is recognized when it is not sold for a wage. Its ambition is bounded by the positionality result: R2 can make care, maintenance, learning, and the arts count; it cannot make everyone eminent, and it does not try. The clearest existing test bed for whether non-waged-but-structured contribution confers standing is the job-guarantee proposal (Tcherneva, 2020; Paul, Darity, Hamilton and Zaw, 2018), which the paper engages in §8 not as a policy endorsement but as the available experiment.

R3 — the legitimacy narrative. Supply a public account of who belongs and why — the answer a person without a job can give and a society will accept. Rawls’s location of self-respect in citizenship rather than work (§3) is the warrant that R3 is possible; Polanyi’s double movement (§3) is the warning of what happens without it. Without R3, R1 and R2 can be technically met and socially rejected, experienced as management rather than membership.

A settlement that meets R1 only — income without a recognized contribution channel, of which unconditional cash alone is the paradigm case — lands in the well-funded vacuum: provided for, contribution-starved. Not the pessimist’s starvation, which R1 prevents, but the more specific failure §4 describes — needs met and the second good unbuilt — sized, now, to what §4 actually supports: a predicted loss in the categorical-standing facet and the bundle around it, not a uniform collapse in wellbeing, and contested by the supply-side and security-channel rivals the paper has conceded.

Two of the structures just built lean, quietly, on the decomposition §4 left open, and the paper flags the dependence rather than hiding it. R2 is presented as a recognition problem — the hard thing no one is working on — which presupposes that recognized contribution is the generative core of the bundle. The paper believes it is, for the reason given in §4 (the facets co-occurred because the job was a site of recognized collective contribution), but concedes this is a claim about the bundle’s causal structure no better established than the decomposition itself: if the core turns out to be structured activity among others, then community institutions plus an income floor could restore much of the bundle without solving the harder recognition problem, and R2's bar is lower than the alarm implies. And Projection 2 needs the standing facet specifically to be the one whose deprivation drives the dynamics; if the decomposition returns “mostly time-structure or control,” P2 loses its mechanism on top of the mobilization step it already concedes. Both structures are therefore conditional on F1, and the paper marks them so.

The idea that markets consume social goods they do not produce is not this paper’s invention, and neither — the external review was right to press this, and the correction improves the paper — is the second good itself. The claim that work delivers non-monetary goods, recognition among them, has a developed literature this draft owes acknowledgment. Gheaus and Herzog enumerate four goods of work obtainable only through work — excellence, social contribution, community, and social recognition — and argue that distributive justice must ask who has access to which of them, not only who receives what wage (Gheaus and Herzog, 2016). Fraser and Honneth supply the grammar beneath that claim at political-theory scale, and their disagreement is this paper’s own two-variable structure argued at the level of justice: whether recognition subsumes redistribution or the two are irreducible (Fraser and Honneth, 2003). Sandel names the corresponding principle contributive justice, and argues directly against transfer-only settlements on the ground that what is at issue “is not just inequality of wealth, but inequality of esteem” (Sandel, 2020). The post-work school runs the opposite valence and is engaged here rather than ignored: Weeks, Srnicek and Williams, and Danaher argue that standing and meaning should be deliberately decoupled from paid work rather than re-routed through a substitute for it (Weeks, 2011; Srnicek and Williams, 2015; Danaher, 2019) — a position this paper’s R2 is a direct alternative to, and which would be vindicated if F2 held. One qualification the paper owes its own citation: Weeks’s central historical resource, the Wages for Housework tradition, is a recognition-of-contribution demand rather than a decoupling one, so the post-work camp is less unified on this point than the label suggests. And Susskind, cited elsewhere in this paper for displacement, occupies the middle: his conditional basic income conditions payment on recognized contribution, economic or otherwise, precisely to preserve the meaning function his own closing chapter identifies (Susskind, 2020) — the nearest existing policy proposal to R2, and uncredited on that point in the first five drafts.

The consequence for the paper’s novelty claim is a resizing, and it is the resizing that makes the claim true. What the debate lacks is not the concept — it has several — but a measurement and a falsifier. No one in the contributive-justice or goods-of-work literatures specifies an instrument for the distribution of recognized contribution, states in advance what would disconfirm the account, or asks a running cash trial to carry the column. That is this paper’s contribution, and it is sufficient. Two further moves distinguish the present claim from its ancestry. A mechanism at the right grain: the ancestors named the meaning problem at the level of a culture and supplied no measurement; this paper names Θ as a vector with instruments and a test site, the cash pilots, which already let us ask whether re-routing W restores Θ and already answer that they were not built to. And a stance: the ancestors treated post-work as prophecy; this paper treats R2 as an engineering requirement, conditional on displacement and silent about its timing. The meaning problem, relocated from prophecy to design, becomes a specification — and, the revision adds in candor, a framework whose near-term empirical payoff is deferred to a measurement not yet made, not a bundle of live near-term wagers.

§ VIIWhat We Can Expect, I: The Well-Funded Vacuum

Everything in this section and the next is projection. Each carries its trend, its assumption, and its falsifier, and none borrows the authority of the measured record it extrapolates. The revision states, more plainly than the first draft, that these are projections whose confidence the §4 concessions have lowered.

Projection 1 — wage-only re-routing leaves the worth bundle unbuilt, and its categorical-standing facet most of all. Trend: income replacement raises wellbeing transiently and then fades (§5); the net-of-income and non-adaptation record says some persistent loss is not income (§4); the cash trials do not measure whether the residual is standing. Assumptions: that the residual non-pecuniary good is substantially the worth bundle rather than wholly security or control (the §4 decomposition, conceded open); that the fade-out generalizes from time-limited pilots; that the standing facet is not spontaneously regenerated by leisure and consumption. Expectation: among the long-term displaced — the sub-population in which Projection 2's positional dynamics have faded — a society that meets R1 and not R2 shows, in the facets §4 can see, the deprivation Marienthal documented — provision without purpose — most among those for whom paid work was the only recognized contribution; sized down, per §4, from “its own kind of ruin” to “a predicted and measurable loss the accounts do not capture,” anomic rather than explosive, and bounded by the heterogeneity concession that for low-worth work the effect may run the other way. Falsifier (F1, the substitution falsifier): a well-identified study of unconditional income at scale and duration that measures the worth bundle — the mattering-and-recognized-contribution index of §6, not only income and employment and generic wellbeing — and finds recipients converging on the employed net of income. The first draft framed F1 as a measurement commitment; the revision keeps the candor and adds the instrument: F1 is now a study that can be designed now, with a specified column, and run to an interpretable result once that column is validated (§6) — and the paper predicts its sign while conceding the quantitative record to date is neutral.

Projection 2 — the newly displaced carry a positional, political risk net of material provision. This is the leg the Round-2 revision relocated and the Round-3 revision re-grounded. It concerns the flow of newly displaced (§6), for whom a still-working reference class is vivid; the danger is not a property of any end-state but of being displaced against others who are not. The relocation, in the prior draft, leaned this leg on the China-shock record — and that was the wrong support, in the paper’s own terms. Those counties had no R1: their workers suffered material loss, status loss, and community collapse together, and the documented political reaction cannot be pinned to relative position specifically, any more than §4's individual evidence could isolate standing from security. The China-shock literature is cited here, accordingly, only for what its regional designs can show — that uneven displacement produces political backlash (Autor, Dorn, Hanson and Majlesi, 2020) — and not for the positional mechanism, which it cannot separate from material distress. Trend: the positional-specific mechanism is instead grounded in evidence that holds own material position fixed — that relative standing moves wellbeing net of absolute income: neighbors’ higher earnings lower one’s own satisfaction at a given income (Luttmer, 2005), peer salaries move job satisfaction with pay held constant (Card, Mas, Moretti and Saez, 2012), and the wellbeing cost of joblessness shrinks as one’s reference group shares it (Clark, 2003). That literature isolates the positional channel China-shock cannot. And one study, added at external review, closes much of the distance between that wellbeing evidence and the political claim P2 actually makes — the gap the first five drafts conceded and left open. Gidron and Hall find, across ISSP data covering fifteen countries for vote choice and twelve across six waves from 1987 to 2014 for trends, that lower subjective social status predicts populist-right support: moving from medium-high to medium-low status roughly doubles the predicted probability, from about six to eleven percent, and the association survives conditioning on occupation, income, education, and employment status (Gidron and Hall, 2017). The detail that matters most to this paper is one line of their table: the unemployment dummy is −0.001 and null in the same specification where subjective status remains significant. Employment status per se carries nothing there; perceived standing carries the effect — which is P2's mechanism, measured at the level of politics rather than wellbeing, and measured net of the material variables. Three limits travel with it. The design is cross-sectional and the authors explicitly disclaim causal identification; the relationship is curvilinear, peaking at medium-low rather than lowest status, which is weaker than a pure deprivation account predicts; and the general methodological objection that status threat and material interest cannot be cleanly separated with cross-sectional data (Morgan, 2018) applies to it in full. It narrows the gap between wellbeing and rupture. It does not close it, and the aggregation step P2 concedes below remains unbuilt. The classical revolution mechanisms — relative deprivation, the J-curve, Hirschman’s tunnel effect, Merton’s strain (Gurr, 1970; Davies, 1962; Hirschman and Rothschild, 1973; Merton, 1938) — supply the path from positional loss to disorder, and the newly displaced, unlike the long-term displaced, still have the reference class these mechanisms require. Assumptions, with the priors conceded: that relative standing net of material provision is a live driver (supported by the net-of-income evidence above); and — the load-bearing and weakest assumption, inherited from Round 1 and untouched by the re-grounding — that distributed positional loss aggregates into collective rupture, which it does not do without mobilizing structure (Tilly, 1978; and the conflict econometrics that found opportunity dominant over grievance, Fearon and Laitin, 2003; Collier and Hoeffler, 2004). The honest gap is triple, and the third is the twin of a concession the paper already makes for the categorical instrument (§6): the clean positional evidence is about wellbeing, not rupture; the bridge to collective action is unbuilt; and the relative-income studies measure relative pay, drawn from wage economies in which pay tracks standing, so carrying them into a post-wage world — where pay is decoupled from standing by construction — is itself an unverified transfer. The paper flags this proxy gap rather than conceding it for the categorical measure (felt mattering for praiseworthiness) and passing over its exact analogue for the positional one (relative pay for relative standing). Expectation: the newly displaced are the sub-population in which the Polanyian reaction can form, where mobilizing structures convert positional loss into collective action; absent them, the same loss runs to the quieter, anomic pathologies of §8. Falsifier (F3, the inflection falsifier): evidence that instability tracks absolute deprivation net of relative position — that the displaced-against-a-working-reference-class, once their material needs are met, stay quiet while only the absolutely poor rebel. The paper states its prior plainly: even re-grounded, F3 is the leg it is likeliest to lose, because the aggregation step runs against the mobilization literature; what the re-grounding buys is that the positional claim now rests on evidence that isolates the positional channel, rather than on a combined-shock backlash wearing a worth label.

Projection 3 — the measurement blind spot certifies the vacuum as success. Trend: the replacement debate measures income, employment, health, and generic wellbeing (§5); the national accounts read output and joblessness; none reads the distribution of recognized contribution. Assumption: official statistics keep instrumenting material and employment outcomes and not the worth bundle, because the instrument did not exist — the assumption the revision can now relax, since §6 supplies one. Expectation: a wage-only settlement passes its own audit — output intact, poverty down, unemployment addressed by transfer — while the categorical-standing facet erodes unrecorded, the instruments reporting success because nothing official records the thing being lost. Falsifier: deployment of the §6 mattering-and-contribution index showing a wage-only settlement healthy on the worth dimension — which would be F1 arriving through national statistics, and would be welcome.

§ VIIIWhat We Can Expect, II: Can the Worth Be Re-Routed?

The constructive question deserves its own section, because the paper’s posture is that the basin is chosen, not caused — that the escalator is achievable and the vacuum avoidable, and that the difference is R2. This section runs the question in both directions, converts the optimist’s best objection into the paper’s second falsifier, and is careful, where the first draft was not, to obey the composition/magnitude discipline of §2.

The objection, at full strength: the worth has always re-bundled on its own. Farmhands became factory workers, factory workers became clerks, and each transition ended with new kinds of recognized contribution no one designed. Markets regenerate worth channels faster than they destroy them; they will again; R2 solves itself. The historical evidence is real (§4): the escalator mostly ran. An argument that it will now stall owes a reason the past two centuries do not simply recur — and the first draft’s reason violated the paper’s own discipline, which the revision corrects. The first draft argued that exposure’s skew toward cognitive work shows “the frontier is closing.” But exposure is a composition measure, and to read the closing of an escape route off it is to do with composition what the terminator does with magnitude. The corrected, weaker, and defensible claim: if displacement of significant size occurs, its composition (§5) would fall on the cognitive and creative tasks toward which displaced labor has historically climbed — so the direction of the historical escape is the direction now exposed. Whether displacement of that size occurs, the paper does not assert; and it concedes two further limits the referee named. The “escape ran upward” history is too clean: automation polarized, hollowing the middle and expanding both tails, including low-skill in-person services (Autor, Levy and Murnane, 2003; Autor and Dorn, 2013), so the escape sometimes ran down into work that was secure-ish but low-worth — which is the vacuum arriving as employment, not the escalator. And the reinstatement of labor demand through genuinely new tasks has historically been unforeseeable in advance (Acemoglu and Restrepo, 2019), which means “this time the frontier is closing” is structurally the same claim made, and falsified, in prior waves. Falsifier (F2, the re-bundling falsifier): evidence that automation transitions regenerate standing-conferring work without deliberate design, faster than they destroy it, and independent of a human-advantaged frontier that cognition-automation reaches — with the paper conceding that F2 is weakly decidable in real time, since by its own logic the new tasks would be unforeseeable until they appeared, and that this weakness makes F2 a wager about structure rather than a crisp near-term test.

If the worth does not re-bundle on its own, it must be engineered, and the engineering is R2's content — bounded, now, by the positionality result of §6: R2 targets the categorical component of standing, not the positional one.

R2's design problem has a dilemma at its center that the first five drafts labeled and did not face, and facing it is the difference between a box in a diagram and a specification. Categorical recognition that everyone receives risks deflation: a certificate all hold carries no information, which is the “participation award” this paper disclaims but never designed against — and the disclaimer is not a design. Recognition that can be failed, on the other hand, reintroduces assessment, and assessment reintroduces rank at the margin, which is the positional good §6 concedes no architecture can supply to everyone. The dilemma is real and this paper does not resolve it; what it can do is state the horns precisely enough that a designer knows which they are choosing, and note that the space between them is where every existing recognition institution actually sits — credentials, professional licensure, civic honors, and, in the paper’s own §4 evidence, the retirement category, which is categorical, universal at threshold, unfailable, and still confers standing, apparently because the contribution it certifies is understood as already rendered. That the unfailable case works at all is the most useful datum available to R2's designers, and it suggests the deflation horn is softer than the symmetry implies.

The precedent the paper should have carried from the start is the largest historical population of contribution-without-wage: unwaged care work. The Wages for Housework current argued precisely that unrecognized contribution should be counted as work (Dalla Costa and James, 1972; Federici, 1975), and the economics of care has spent decades on the valuation problem beneath it (Folbre, 2001). The precedent is simultaneously R2's natural case study and a warning about its timescale: the recognition of care as contribution has been contested for two generations and is not settled, which is what R3-narrative failure looks like from inside. Any estimate of how quickly a worth channel can be built should be calibrated against that record rather than against the speed of a transfer program. The shape of the problem is statable even where its solution is not. It requires recognition the culture actually registers — not a participation award but a standing the surrounding society treats as real, which is why R3 sits beside R2; it requires channels with structure, because the facets Jahoda named do not arrive with a check and a contribution institution must supply the time and company and purpose a transfer cannot; and it requires a reward architecture that points motivation at the worth-bearing activities the wage economy under-counts — care, maintenance, learning, the commons — which is the parent paper’s machinery applied to recognition. The job guarantee is the clearest existing test of whether structured non-waged contribution confers the categorical good (Tcherneva, 2020; Paul, Darity, Hamilton and Zaw, 2018): the paper engages it not as endorsement — it declines a policy menu (§10) — but as the experiment whose results would most directly inform R2, and notes that its standing question (does guaranteed public work confer recognized contribution, or only income-with-tasks?) is the same question F1 poses to cash, asked of a different instrument.

And a quieter projection the section should state in the cooler register the rest of the paper now keeps, because the worth motive does not wait for R2 and because this is where the positional residue of §6 actually lives. When a legitimate worth channel is missing, standing-seeking does not vanish; it routes, at the margin, to whatever supplies recognition more cheaply. The gang and the black market have long been described as worth economies of a kind — they answer the contribution question, confer rank, and structure time for people the wage economy has stopped answering, and Wilson’s account of what joblessness did to such communities, net of poverty, is the nearest record (Wilson, 1996). Identity-based belonging — the tribe, the online faction — is a worth economy of a newer kind: standing conferred for loyalty rather than contribution, available instantly and free to those whose legitimate standing has thinned, the demand-side complement to a supply-side account of responsive, individually-adaptive persuasion developed elsewhere in this archive (No. 005, The Broadcast and the Model). The claim is sized to the conceded priors: this is the channel through which the positional component — the part R2 cannot supply and the transition does not dissolve — exerts itself when legitimate outlets are absent, and it is offered as a mechanism worth watching, not a prediction of scale. Falsifier: status-starved populations whose material needs are met show no elevated turn to illegitimate standing — if R1 alone absorbs the worth motive into private consumption — which would also be F1 arriving from the side; the paper expects otherwise on the strain record and marks the expectation as a projection, not a finding.

§ IXThe Ledger: What We Claim and What We Do Not

Before the register audit, a disclosure the first five drafts never made, because it is visible only when the three falsifiers are read together rather than one at a time. Each was conceded locally; the conjunction was not. F1 is gated on validating an index whose validation criteria this paper itself would set. F2 the paper concedes is weakly decidable in real time, since the new standing-conferring work would be unforeseeable until it appeared. F3 the paper expects to lose, and has pre-accommodated the loss. The joint consequence is that no observation available in the next several years can disconfirm the framework — which sits badly beside a rhetoric of “three ways to prove it wrong.” The paper states the position plainly: as of this draft it is a framework with deferred falsifiability, and the deferral is a real cost, not a stylistic one. Two things reduce it, and both are commitments rather than claims.

First, the index gate is handed to third parties, so the paper does not control its own tribunal. A worth index counts as validated if, on a general-population sample: its items load on a factor distinct from generic life satisfaction and from occupational prestige (discriminant validity, latent correlation below .85 against each); it correlates at least .40 with an independently collected other-report measure of the same person’s recognized contribution; it shows metric invariance across employed and non-employed strata; and it distinguishes known non-waged contributors — sustained caregivers, sustained volunteers — from the long-term idle at d ≥ 0.5 under that invariance. These thresholds are stated here, in advance, so that a research group with no stake in this paper’s thesis can run the validation and rule on it. If a validated index by those criteria shows income recipients converging on the employed net of income, F1 has fired, and this paper’s first leg is wrong.

Second — and this is the escape hatch the review was right to close — the paper forecloses the deflection it would otherwise be tempted into. The conceded praise/praiseworthiness gap (§6) could be used to dismiss any disconfirming index result as “felt mattering, not social-fact standing,” which would rebuild inside the instrument the same proves-too-much parry the Round-2 revision disallowed on Finland. The paper renounces it here: a validated index result counts against this paper’s thesis on the criteria above, and the felt-versus-conferred distinction may not be invoked after the fact to discount a result the paper predicted the sign of. If the distinction is to bear weight, it must do so in the instrument’s design, before the data exist — which is the successor paper’s business, not a retrospective defense of this one.

Third, short of strict falsification, three near-term observations would move this paper’s confidence without waiting on the index: cash-trial waves that add any standing or purpose measure and find recipients converging on the employed (confidence down); a demonstrated re-bundling of standing-conferring work in an occupation whose cognitive core was automated (down, on F2); and evidence that populations meeting R1 but not R2 show no elevated status-related distress relative to matched employed populations (down, on P1). None is a falsifier. All are available before the index is.

In the house manner, the register audit, so that no reader need reverse-engineer the paper’s accounting.

The registers, named for a reader who has not seen the companion index: Register C is what is owed — conceptual and normative claims, carried at passage-level citation; Register E is what is measured — empirical claims, each with magnitude, venue, and contest; Register P is what can be expected — projections, each with trend, assumption, and falsifier.

What is owed (Register C, §3): the two-goods distinction, the worth as a bundle with recognized standing its distinctive facet, the praiseworthiness account of standing as non-positional, the Arendtian independence of worth from subsistence-labor, the Polanyian reaction — each cited at passage level, with the central construction (work bundled the goods contingently; the worth is a vector; standing splits into categorical and positional) labeled the paper’s own, and with two witnesses turned against the usual use: Rawls, who grounds self-respect in citizenship rather than work and so warrants R3, and Smith of the Wealth of Nations, who watched the division of labor degrade the worker and so disciplines the no-nostalgia rule. What is measured (Register E, §§4–5): the non-pecuniary cost of joblessness and its magnitudes, including the mass-layoff causal designs; the explicit concession that the evidence identifies a bundle and does not decompose it; the deflation in full — money’s real effect, Marienthal’s wrong-conditions provenance, Ruhm’s supply-side rebuttal, the quantitative record’s neutrality on the worth-specific claim, and the heterogeneity that makes the welfare sign indeterminate for low-worth work; the displacement record as exposure-composition-not-magnitude; the replacement record as a wage apparatus blind to the worth. What is projected (§§7–8): the projections, each with its falsifier, sorted by concurrent sub-population — Projection 1 the anomic, categorical harm of the long-term displaced, which R2 addresses; Projection 2 the positional, political risk of the newly displaced, grounded in net-of-material relative-position evidence (Luttmer; Card et al.; Clark) rather than in the China-shock record, which is cited only for the existence of transitional backlash and cannot isolate the positional channel; gated by mobilization, and explicitly not something R2 is claimed to avert; with the §8 substitute cooled to match and Projection 2's prior still stated against it. What is framework (§6): the two variables, Θ as a vector, the categorical/positional split, the operationalization against existing instruments, the deciding variable R, the three re-routing requirements.

What the paper does not claim. No displacement timeline and no probability of artificial general intelligence; the argument is conditional on displacement and silent about its arrival. No nostalgia for the wage economy and no claim that work was good; the worth was a channel that ran through drudgery and exclusion, and for many jobs its loss is a release. No claim that the worth is a single good or that recognized standing is established as its dominant component; the decomposition is open and routed to F1. No claim that recognized standing can be made universal in its positional sense; R2 targets the categorical component, and the positional residue is conceded irreducible. No claim that unconditional income is a mistake; R1 is necessary, and the argument is that it is insufficient alone. No claim that immiseration is fated; the basin is chosen. No claim that the political rupture of Projection 2 is likely rather than possible; the mobilization literature is conceded to lean the other way, and the rupture is located among the newly displaced, not the long-term displaced. No claim that the China-shock record isolates the positional mechanism; it is cited only for the existence of transitional backlash, and the positional-net-of-material claim rests on the relative-income literature instead. No claim that R2 averts the newly-displaced positional politics; R2 addresses the long-term-displaced anomic harm, and the positional, transitional risk is a problem of distribution and reference the paper assigns elsewhere. No claim that the worth index is validated; it is specifiable from adjacent instruments, measures felt rather than deserved standing, and omits a social-recognition component still to be built. No claim that R2's necessity as a recognition problem is established independent of the open decomposition; if structured activity is the bundle’s core, R2's bar is lower. No villain; the vacuum is an architecture’s default, and defaults have no author. No claim that the social contract was ever honored equally. The paper claims a distinction, a deciding variable, a predicted and now-measurable failure mode, and three ways to prove it wrong — one of which, it concedes, it is more likely to lose than win.

§ XImplications: Changing the Diagnostic Vocabulary Before the Policy Menu

The paper ends where the house tradition ends its arguments: with what to measure, not what to mandate. The deepest practical consequence of the two-goods distinction is that the instruments by which a society will judge its automation transition are blind to the variable that decides it. Income, employment, output, and poverty are wage-channel measures. A settlement securing every one of them while the worth channel emptied would report, by every official number, a success — and the failure would surface later and elsewhere, by which point the diagnosis is an autopsy.

So the agenda is measurement before menu, and the revision can now state it as a build rather than a wish. The instruments for the positional component of standing already exist — occupational-prestige scales rank what the labor market already ranks — and they are the wrong tool, because the good the vacuum destroys is the categorical one: being counted a contributor at all. The buildable column is the distribution, across a population, of a mattering-and-recognized-contribution measure, drawing on the subjective-social-status, meaningful-work, eudaimonic, and mattering instruments (Adler et al., 2000; Hackman and Oldham, 1980; Ryff, 1989; Elliott et al., 2004) and reported, like income, with its inequality — a Θ-distribution beside the income distribution. Evaluate proposed transitions on their worth content, not only their wage content: a jobs guarantee, an unconditional income, and a shorter week are not interchangeable income instruments but different bets about R2, and the question to ask each is whether it confers recognized contribution or only delivers cash — which is, precisely, F1 asked of each instrument. And treat the categorical-standing distribution as a leading indicator, watched the way inflation is watched, so the politics of the vacuum can be seen forming. The paper does not know whether the worth channel should be built through public work, a transformed civil society, a revaluation of care, or institutions not yet named, and declines to pretend the choice is settled. It claims that the choice is the one that matters, that no one is measuring whether we are making it, and that the instrument which would tell us is, on the evidence of §6, specifiable now and buildable with validation — close enough to existing scales to start, far enough from them that the social-recognition column still has to be earned.

§ XIA Research Program

This paper is one move in a longer argument, conducted in the open, and the revision has made plainer than the first draft what kind of move it is: a framework whose near-term empirical payoff is a measurement, deferred until the measurement is made. Its three falsifiers are offered as invitations, and the revision has also made plainer which the paper expects to win. The substitution falsifier (F1) is now a study that can be run — an unconditional-income trial instrumented for the worth bundle, with the mattering-and-contribution column §6 specifies and the net-of-income identification that would tell us whether the residual is worth or security; the paper predicts its sign and concedes the present record is neutral. The re-bundling falsifier (F2) is a historical and structural wager the paper concedes is weakly decidable in real time. The inflection falsifier (F3) is the political economists’ contest, relocated to the newly displaced and re-grounded on relative-position evidence net of material loss (Luttmer; Card et al.; Clark) rather than on the China-shock record, which shows only that transitional backlash occurs — still the leg the paper is likeliest to lose, because the aggregation-into-rupture step runs against the mobilization literature, and retained because honesty requires naming a favored disconfirmer. The paper invites the experimentalists whose pilots it leans on to add the missing column; the unemployment-and-wellbeing tradition to test the decomposition its construct now turns on; the scholars of mobilization, not only of grievance, to supply or deny the mechanism Projection 2 needs; and the designers of job guarantees to treat their programs as the standing experiment they already are.

The paper belongs to the motivation thread and takes the thread’s question to its limit. The Conditional Frontier asked what an institution drafted around motivation would have to specify; Beyond the Binary engineered the reward architecture such an institution needs, for an economy that runs on paid work. This paper takes that architecture to the case that breaks it — an economy in which the wage-bearing job, the channel that delivered both subsistence and standing, is withdrawn — and finds the architecture’s hardest task is the one no one is working on: not replacing the wage, which we know how to do, but re-routing the worth, which we have only now specified how to measure. John Adams imagined an escalator carrying his grandchildren up to the things a free person does for their own sake. He was right that the destination is the point. He did not have to ask who would still have a place on the stairs. We do — and we will not know whether we have built one until we agree to measure the thing the stairs were carrying.

Notes

On the registers. Register C is what is owed — conceptual and normative claims, carried at passage-level citation. Register E is what is measured — empirical claims, each with magnitude, venue, and contest. Register P is what can be expected — projections, each with trend, assumption, and falsifier. The definitions appear in §9 for readers without the companion index. [§§3–9]

On what this revision added, and why. Four adversarial rounds audited this paper inward — consistency, candour, proxy hygiene — and none audited it outward. A coverage review then surfaced three absences that materially bear on the argument: the retirement natural experiment (§4), which is the closest thing in the published record to the mechanism claimed here; the goods-of-work and contributive-justice literatures (§6), which mean the paper's construct was never as unprecedented as earlier drafts implied; and the status-politics bridge (§7), which closes much of the distance between the wellbeing evidence and the political projection. A coverage pass is now a standing gate for this archive. [§§4, 6, 7, 9]

On the retirement result's limits. The Hetschko, Knabe and Schöb effect is male-specific — 0.53 for men, −0.01 and null for women — which is a real limit on generality and is stated in the body. No specification in that paper controls the respondent's own health, and this paper does not claim it does. An independent reanalysis of the same panel finds a short-term satisfaction gain for both transition groups, more pronounced for the formerly unemployed. The design has no non-German replication. [§4]

On the falsifiers. Read jointly rather than one at a time, the three falsifiers cannot fire in the near term: F1 is gated on an index validation, F2 is weakly decidable in real time, and F3 the paper expects to lose. §9 states this as a framework with deferred falsifiability, hands the validation criteria to third parties in advance and in numbers, and renounces the praise/praiseworthiness deflection so that it cannot be used after the fact to discount a disconfirming result. [§9]

On figures carrying verification flags. Two are open at imprint and are to be discharged against primary sources before external submission: the panel years in Hetschko, Knabe and Schöb, where the accepted manuscript reads SOEP 1984–2010 and the circulating abstract reads 1984–2009; and the column assignment of the quadratic term in Gidron and Hall's vote model. [§§4, 7]

On the epigraph. The escalator line is John Adams to Abigail Adams, post 12 May 1780, transcribed from the Adams Papers with original spelling preserved; a manuscript cancellation shows Adams struck "Painting and Poetry" and replaced it with "Mathematicks and Philosophy," sharpening the three-generation sequence. It is widely misattributed to Jefferson. [epigraph, §1]

References

Acemoglu, D. (2025). The simple macroeconomics of AI. Economic Policy, 40(121).

Acemoglu, D., & Restrepo, P. (2019). Automation and new tasks. Journal of Economic Perspectives, 33(2); (2020). Robots and jobs. Journal of Political Economy, 128(6).

Adler, N. E., Epel, E. S., Castellazzo, G., & Ickovics, J. R. (2000). Relationship of subjective and objective social status. Health Psychology, 19(6).

Akerlof, G. A., & Kranton, R. E. (2000). Economics and identity. Quarterly Journal of Economics, 115(3).

Arendt, H. (1958). The Human Condition. University of Chicago Press.

Arntz, M., Gregory, T., & Zierahn, U. (2016). The risk of automation for jobs in OECD countries. OECD Working Paper 189.

Autor, D. (2015). Why are there still so many jobs? Journal of Economic Perspectives, 29(3).

Autor, D., & Dorn, D. (2013). The growth of low-skill service jobs and the polarization of the US labor market. American Economic Review, 103(5).

Autor, D., Dorn, D., & Hanson, G. (2019). When work disappears: Manufacturing decline and the falling marriage market value of young men. AER: Insights, 1(2).

Autor, D., Dorn, D., Hanson, G., & Majlesi, K. (2020). Importing political polarization? American Economic Review, 110(10).

Autor, D., Levy, F., & Murnane, R. J. (2003). The skill content of recent technological change. Quarterly Journal of Economics, 118(4).

Brynjolfsson, E., Li, D., & Raymond, L. (2025). Generative AI at work. Quarterly Journal of Economics, 140(2).

Card, D., Mas, A., Moretti, E., & Saez, E. (2012). Inequality at work: The effect of peer salaries on job satisfaction. American Economic Review, 102(6).

Case, A., & Deaton, A. (2015). PNAS, 112(49); (2020). Deaths of Despair and the Future of Capitalism. Princeton University Press.

Clark, A. E. (2003). Unemployment as a social norm. Journal of Labor Economics, 21(2).

Clark, A. E., Diener, E., Georgellis, Y., & Lucas, R. E. (2008). Lags and leads in life satisfaction. Economic Journal, 118.

Clark, A. E., & Oswald, A. J. (1994). Unhappiness and unemployment. Economic Journal, 104.

Dalla Costa, M., & James, S. (1972). The Power of Women and the Subversion of the Community. Falling Wall Press.

Danaher, J. (2019). Automation and Utopia. Harvard University Press.

Davies, J. C. (1962). Toward a theory of revolution. American Sociological Review, 27(1).

Demirci, O., Hannane, J., & Zhu, X. (2025). Who is AI replacing? Management Science, 71(10).

Durkheim, É. (1893). The Division of Labor in Society; (1897). Suicide.

Eloundou, T., Manning, S., Mishkin, P., & Rock, D. (2024). GPTs are GPTs. Science, 384.

Federici, S. (1975). Wages Against Housework. Falling Wall Press. [reprinted in Revolution at Point Zero, PM Press, 2012, 15–22]

Folbre, N. (2001). The Invisible Heart: Economics and Family Values. The New Press.

Fraser, N., & Honneth, A. (2003). Redistribution or Recognition? A Political-Philosophical Exchange. Verso.

Frey, C. B., & Osborne, M. A. (2017). The future of employment. Technological Forecasting and Social Change, 114.

Fryer, D. (1986). Employment deprivation and personal agency during unemployment. Social Behaviour, 1.

Gheaus, A., & Herzog, L. (2016). The goods of work (other than money!). Journal of Social Philosophy, 47(1), 70–89.

Gidron, N., & Hall, P. A. (2017). The politics of social status. British Journal of Sociology, 68(S1), S57–S84. [table column flagged]

Graeber, D. (2018). Bullshit Jobs. Simon & Schuster.

Gurr, T. R. (1970). Why Men Rebel. Princeton University Press.

Hetschko, C., Knabe, A., & Schöb, R. (2014). Changing identity: Retiring from unemployment. Economic Journal, 124(575), 149–166. [panel years flagged]

Hetschko, C., Knabe, A., & Schöb, R. (2019). Looking back in anger? Retirement and unemployment scarring. Demography, 56(3).

Hirsch, F. (1976). Social Limits to Growth. Harvard University Press.

Hirschman, A. O., & Rothschild, M. (1973). The changing tolerance for income inequality in the course of economic development. Quarterly Journal of Economics, 87(4).

Honneth, A. (1995). The Struggle for Recognition. Polity/MIT Press.

Jacobson, L. S., LaLonde, R. J., & Sullivan, D. G. (1993). Earnings losses of displaced workers. American Economic Review, 83(4).

Jahoda, M. (1982). Employment and Unemployment: A Social-Psychological Analysis. Cambridge University Press.

Jahoda, M., Lazarsfeld, P. F., & Zeisel, H. (1933/1971). Marienthal: The Sociography of an Unemployed Community.

Kangas, O., et al. (2020). Evaluation of the Finnish Basic Income Experiment. Ministry of Social Affairs and Health, 2020:15.

Keynes, J. M. (1930). Economic possibilities for our grandchildren. In Essays in Persuasion.

Lindqvist, E., Östling, R., & Cesarini, D. (2020). Long-run effects of lottery wealth on psychological well-being. Review of Economic Studies, 87(6), 2703–2726.

Luttmer, E. F. P. (2005). Neighbors as negatives. Quarterly Journal of Economics, 120(3).

McGuire, J., Kaiser, C., & Bach-Mortensen, A. M. (2022). A systematic review and meta-analysis of the impact of cash transfers on subjective well-being. Nature Human Behaviour, 6(3).

Merton, R. K. (1938). Social structure and anomie. American Sociological Review, 3(5).

Morgan, S. L. (2018). Status threat, material interests, and the 2016 presidential vote. Socius, 4.

Paul, M., Darity, W., Hamilton, D., & Zaw, K. (2018). A path to ending poverty by way of ending unemployment: A federal job guarantee. RSF: The Russell Sage Foundation Journal of the Social Sciences, 4(3), 44–63.

Paul, K. I., & Moser, K. (2009). Unemployment impairs mental health. Journal of Vocational Behavior, 74.

Polanyi, K. (1944). The Great Transformation. Farrar & Rinehart.

Rawls, J. (1971; rev. 1999). A Theory of Justice, §67. Belknap/Harvard.

Ruhm, C. J. (2019). Drivers of the fatal drug epidemic. Journal of Health Economics, 64.

Sandel, M. J. (2020). The Tyranny of Merit. Farrar, Straus and Giroux.

Smith, A. (1759). The Theory of Moral Sentiments; (1776). An Inquiry into the Nature and Causes of the Wealth of Nations. Glasgow Edition.

Srnicek, N., & Williams, A. (2015). Inventing the Future: Postcapitalism and a World Without Work. Verso.

Sullivan, D., & von Wachter, T. (2009). Job displacement and mortality. Quarterly Journal of Economics, 124(3).

Susskind, D. (2020). A World Without Work. Metropolitan Books.

Tcherneva, P. R. (2020). The Case for a Job Guarantee. Polity.

Veblen, T. (1899). The Theory of the Leisure Class.

Vivalt, E., Bartik, A., Miller, S., Broockman, D., & Rhodes, E. (2024). The employment effects of a guaranteed income. NBER 32719; and companion papers 32711, 32784.

Weeks, K. (2011). The Problem with Work. Duke University Press.

West, S., & Castro, A. (2021). Stockton Economic Empowerment Demonstration, first-year report.

Wetzel, M., Huxhold, O., & Tesch-Römer, C. (2016). Transition into retirement affects life satisfaction. Social Indicators Research, 125(3).

Wilson, W. J. (1996). When Work Disappears. Knopf.

· · ·

Companion papers in the archive: No. 001, The Conditional Frontier, and No. 002, Beyond the Binary (the thread's first two papers; No. 002 is this paper's parent, and its reward architecture is ported here from production to recognition); No. 009, The Praise and the Praiseworthiness (the successor, which builds the instrument this paper's first falsifier waits on).

The companion deck — The Wage and the Worth — sits beside this paper; the corpus index, claim outline, framework and projection set, four internal referee rounds with their response letters, the external review, and the six manuscript versions are filed under cowork. The successor paper specifies the worth index in full and hands its validation criteria to third parties. Comments — and corrections to the flagged figures — welcome at contactme@marshallcahill.com.

The Wage and the Worth.

Abstract

§ IThe Question and the Two Goods

§ IITwo Caricatures to Clear

§ IIIThe Bundle Named, I: What Is Owed

§ IVThe Bundle Named, II: What Is Measured

§ VThe Documented Present

§ VIThe Hinge: The Severed Bundle

§ VIIWhat We Can Expect, I: The Well-Funded Vacuum

§ VIIIWhat We Can Expect, II: Can the Worth Be Re-Routed?

§ IXThe Ledger: What We Claim and What We Do Not

§ XImplications: Changing the Diagnostic Vocabulary Before the Policy Menu

§ XIA Research Program

Notes

References

Slides · accompanying lecture

References

Abstract

§ IThe Question and the Two Goods

§ IITwo Caricatures to Clear

§ IIIThe Bundle Named, I: What Is Owed

§ IVThe Bundle Named, II: What Is Measured

§ VThe Documented Present

§ VIThe Hinge: The Severed Bundle

§ VIIWhat We Can Expect, I: The Well-Funded Vacuum

§ VIIIWhat We Can Expect, II: Can the Worth Be Re-Routed?

§ IXThe Ledger: What We Claim and What We Do Not

§ XImplications: Changing the Diagnostic Vocabulary Before the Policy Menu

§ XIA Research Program

Notes

References

Slides · accompanying lecture

References

Related work