The authorship question begins with an embarrassment of trust. Two names stand on the title pages of some of the most argued over books of the modern age, and for more than a century readers treated the pairing as a single mind with two hands. Marx and Engels wrote together, edited each other, financed each other and defended each other so completely that the composite figure, half jokingly called by later generations a two headed author, seemed a natural unit of intellectual history. Yet the books were produced by distinguishable processes. Some were drafted by one man and finished by the other. Some were written by one and signed by both. One famous pamphlet rests on scaffolding built by the junior partner and a final text written by the senior. Two large volumes of the major work were assembled after the author’s death by his friend, from manuscripts the friend described, honestly but incompletely, in prefaces that few readers ever studied. The question of who wrote what is therefore not gossip about a famous friendship. It is a question about evidence, and it can be answered, case by case, with documents.

What changed the question from a matter of guesswork into a matter of record was the slow publication of the manuscripts themselves. So long as readers possessed only the printed books, Engels’s editing of the later volumes of Capital could be praised or suspected but not measured. Once the critical edition placed Marx’s own draft of the third volume beside the version Engels published, measurement became possible, and the measurements were not trivial. They showed reordering of material, chapter divisions and headings supplied by the editor, and, in the most argued over stretch of the book, a presentation of the falling rate of profit more orderly and more lawlike than the draft underneath it. None of this makes Engels a falsifier, and this article will give the defence of his work its full strength before reaching any verdict. But it does mean that a claim about what Marx argued, when the claim rests on the second or third volume, is incomplete until it says which text it rests on. That requirement has a name in what follows. It is called the edition attribution rule, and building the case for it, document by document, is the work of the pages ahead.

Authorship as a question of evidence

Authorship sounds like the simplest fact about a book, and in this partnership it is among the least simple, because the two men collaborated at every level at which collaboration is possible. They planned works together in conversation and by letter. They divided subjects between them according to talent and deadline. They read each other’s drafts and argued over them. Engels, who earned money in his family’s Manchester firm, subsidised Marx for decades, freeing the hours in which the manuscripts accumulated. He also wrote, at speed and to order, journalism that appeared under Marx’s name, because Marx was the contracted correspondent and Engels was the better military analyst and the faster pen. Any account that imagines two sealed desks, each producing its own signed output, falsifies the record before it begins. The interesting question is narrower and harder: for each text that bears on later argument, whose sentences are these, in what state did they leave the author’s desk, and what happened to them between the desk and the printed page?

Did Engels change Marx’s manuscripts?

Yes, substantially, in editing the later volumes of Capital. Comparison with the original manuscripts, published in critical editions from the 1990s, shows reordering, added chapter divisions and headings, and passages made more systematic than the drafts themselves support, most notably in the material on the falling rate of profit.

That answer, compressed into a single paragraph, conceals three different kinds of editorial act, and keeping them apart is the first discipline the subject demands. The lightest act is correction: fixing slips, completing an unfinished sentence, choosing between two variants of a paragraph where the author left both standing. No serious critic begrudges an editor these acts, and every executor performs them. The middle act is arrangement: deciding the order in which fragments appear, gathering scattered treatments of one theme into a single chapter, supplying the headings under which the reader will meet the material. Arrangement changes nothing in the sentences and a great deal in the book, because order is an argument about what depends on what. The heaviest act is construction: writing the transitions, building a systematic sequence out of exploratory drafts, presenting as a finished doctrine what the manuscript holds as a research programme in motion. The scholarly dispute of the past decades is, at bottom, a dispute about where on this scale Engels’s work on the third volume falls. The defence places it at the lighter end, the critics place parts of it at the heavier end, and the manuscript evidence, which both sides now accept as fact, is what allows the dispute to be conducted precisely.

Three cautions govern everything that follows, and they are worth stating before any case is opened. The first is that editing is not falsification. An editor who reorders fragments has not invented their content, and a charge of bad faith would need evidence of motive and suppression that the record, examined below, does not supply. The second caution is that the two men are not interchangeable, and never were. Marx’s surviving drafts show a mind testing formulations against objections, circling a problem through successive approximations, often leaving the most theoretical material in the least finished state. Engels’s own writings show a different gift: exposition, system, the memorable summary, the willingness to state a doctrine in numbered propositions. When the expositor became the editor of the explorer, some translation between temperaments was inevitable, and the question is what the translation cost and what it preserved. The third caution concerns the standard of proof. Attribution claims in this article rest on documents: manuscripts in identifiable hands, letters in which the two men discuss who is writing what, prefaces in which Engels describes his own procedure, and the critical editions that print the drafts so any reader can check. Where the documents run out, the article says so, and the ledger near the middle of the article keeps a column for what remains disputed precisely so that settled findings and open questions are never blurred.

Readers who want the machinery behind the documents should follow the account of the critical edition project in the mega2 marx manuscripts guide, which explains how the historical critical edition is organised and why its volumes changed what can be claimed. The present article is narrower. It takes the edition’s findings as its evidence base and asks the practical question: given what the manuscripts show, how should the famous texts be described, cited and argued from? The cases come first, because the rule this article defends is worthless unless it is earned by the cases. They begin with the pamphlet nearly everyone has read, proceed through the great unpublished manuscript of the partnership’s youth and the journalism that paid the bills, and arrive at the two posthumous volumes where the stakes are highest.

Before the cases begin in earnest, the hyphen deserves a moment of suspicion. Marx and Engels, yoked by a punctuation mark, function in the culture as a single authorial brand, and the brand has institutional machinery behind it: collected works in fifty volumes that interleave the two men’s writings in chronological order, presenting the partnership’s output as one river; textbooks that cite the pair as the unit of doctrine; a whole vocabulary, Marxism itself, that names a system neither man lived to see systematised. The brand is not a conspiracy. It is the natural deposit of a real collaboration and of Engels’s own lifelong insistence on subordinating his name to his friend’s. But brands answer questions before they are asked, and the question the brand answers, who wrote this, is the one this article insists on asking again, text by text. The hyphen also conceals an asymmetry that the documents keep exposing. One partner wrote slowly, finished little and left mountains; the other wrote quickly, finished much and left editions. Posterity, which prefers its classics finished, has read the slow man’s mountains through the quick man’s paths across them, and has often mistaken the paths for the terrain. Keeping the two apart is not pedantry about a friendship. It is the precondition of knowing what the terrain actually looks like, paths and all.

It helps, before the cases, to separate three levels at which the single word authorship operates in this subject, because much confusion in the popular literature comes from sliding among them. The first level is composition: who drafted the sentences, in whose hand, from whose notes. The second is construction: who gave the drafted material its published form, its order, divisions, emphases and finish. The third is canonisation: who decided that the resulting object would stand as a classic, be reprinted, translated, taught and cited as the author’s doctrine. For the first volume of Capital all three levels converge on Marx, with minor qualifications. For the Manifesto they divide between the partners in ways the documents specify. For the third volume they divide dramatically: composition overwhelmingly Marx, construction substantially Engels, canonisation the work of Engels first and of a century of publishers, parties and universities after him. Arguments about the authorship question go wrong when a finding at one level is used to settle a question at another: when Engels’s construction is cited to deny Marx’s composition, or Marx’s composition cited to deny Engels’s construction, or the text’s canonicity cited to settle either. The ledger that follows is organised by the levels implicitly, its drafting column addressing composition, its final version column addressing construction, and its citation column prescribing how canonised texts should carry the memory of the first two levels with them.

A partnership conducted on paper

The documentary richness of this subject is not an accident, and understanding its sources explains why some questions in the ledger can be answered to the sheet while others remain matters of judgment. Marx and Engels spent the decisive decades of their collaboration in different cities, Marx in London, Engels in Manchester until 1870, and distance made them correspondents. They wrote to each other constantly, often daily, about money, publishers, politics, the progress of drafts, the incompetence of rivals and the contents of articles each was writing for the other. A partnership conducted by letter leaves an archive of its own working, and theirs survived: thousands of letters in which commissions are given, drafts acknowledged, deadlines negotiated and authorship, again and again, casually assigned in passing. When later scholars needed to know who wrote a Tribune article, the answer was usually sitting in a letter written to arrange its delivery. Face to face collaborators leave memoirs. Absent collaborators leave evidence.

The second source is the manuscripts themselves, preserved with a completeness that was partly Engels’s doing. As executor he treated the papers as a trust, kept them together, had the illegible portions transcribed, and passed the collection on through the family and the movement’s archives until it reached the institutional keeping in which the critical edition could be built. The third source is the printed record of editions: the successive versions of the first volume that Marx revised himself, the prefaces in which Engels described his procedures, the early editions of the manuscripts published in the twentieth century in forms that later editors judged too arranged, and finally the historical critical edition with its apparatus. Each source disciplines the others. A claim in a letter can be checked against a manuscript; a claim in a preface can be checked against both. The authorship question is answerable precisely because the three series survive in parallel, and the places where the ledger records a dispute are instructive: they cluster where one series runs out, a lost letter, a missing sheet, a stretch of work the two men did in the same room in Brussels where no correspondence was needed and the only witness is the paper itself.

Money, the least romantic of the sources, also belongs in the record, because it shaped the division of labour that the documents display. Engels earned a salary and then a partnership income in the Ermen and Engels cotton firm, and from that income he subsidised the Marx household for decades, in regular remittances, loans never called in and emergency rescues when bailiffs threatened. The subsidy bought Marx’s hours in the British Museum reading room, and it bought them at a price that the correspondence makes visible: the subsidised partner owed the other not money but work, and the articles written to order, the drafting taken over at short notice, the whole machinery of mutual service, ran inside an economy of obligation that neither man ever itemised. Recognising the economy matters for attribution because it explains the direction of the ghost writing and the later editorship without recourse to psychology. The partner with the salary had the steadier hand and the freer evenings; the partner with the contract, the readership and the great unfinished book had the name. When the name’s owner died leaving the book unfinished, the other partner’s assumption of the editorship was the final transaction in the same economy, forty years of service ending in a decade of construction.

The Manifesto: one text, two contributions, unequal shares

The Communist Manifesto is the natural first case because its title page, in the editions that followed the authors’ lifetimes, carries both names, and because the work’s actual production is documented unusually well. The pamphlet was commissioned by the League of the Just, reorganised in 1847 as the Communist League, whose second congress in London late that year charged Marx and Engels with drafting the movement’s programme. Behind the commission lay earlier attempts, and the attempts were Engels’s. In June 1847 he had drafted a statement for the League in the form of a catechism, a sequence of questions and answers known as the Draft of a Communist Confession of Faith. In the autumn he recast the material in a second catechism, the Principles of Communism, twenty five answers long, which worked through the character of the proletariat, the history of class society and the programme of the movement. The catechism form was a standard device of secret societies, and Engels himself judged it inadequate for a public programme. Writing to Marx in November 1847, he proposed dropping the catechism shape in favour of a continuous manifesto, and he supplied the substance such a text would need. Then the partnership’s division of labour asserted itself. Marx took the drafting into his own hands in Brussels, worked through December and January under pressure from the League’s central committee, and produced the published text. Engels was in Paris for part of the decisive weeks. The pamphlet appeared in London in February 1848, anonymously, as the work of the League rather than of named authors.

Who drafted the Manifesto before Marx took it up?

Engels drafted it twice in 1847, first as a short catechism for the League and then as the Principles of Communism, a fuller question and answer version covering the proletariat, class history and the movement’s programme. Marx then set the form aside and wrote the published pamphlet himself, working from that scaffolding.

The evidential situation, unusually for this subject, includes a verdict from one of the participants about the other. In the preface to the 1883 German edition, published months after Marx’s death, Engels wrote that the fundamental proposition of the Manifesto, the idea that economic production and the structure of society arising from it form the basis of political and intellectual history in every epoch, belonged solely to Marx. The statement is remarkable for its precision. Engels did not claim a half share of the pamphlet, nor did he perform a false modesty about his own role, which he described elsewhere without embarrassment. He located the work’s theoretical core in his partner and left himself the scaffolding, the preparatory drafts and the organisational labour. Later scholarship, with the two drafts in print beside the final text, has found no reason to overturn the allocation. Readers can watch the Principles posing the questions and the Manifesto answering many of them in prose of a wholly different temperature. The famous opening, the spectre haunting Europe, the compressed history of bourgeois society, the rhetorical architecture of the whole, are Marx’s writing by every test of style and by the testimony of the only other man in a position to know.

What follows from the case? Two things, one negative and one positive. The negative lesson is that joint signature does not imply joint composition sentence by sentence. The Manifesto was not written in alternating paragraphs, and no passage of it should be described as Engels’s composition on the strength of the shared title page. The positive lesson is that contribution is not exhausted by composition. Engels’s drafts fixed the programme’s agenda, carried over formulations that survive inside the finished text, and embodied the League experience of 1847 on which Marx, who had spent the year in Brussels polemics, drew. The ledger therefore records the pamphlet as Marx’s text on Engels’s scaffolding, which is not a compromise between two claims but a description of a documented process. Anyone teaching or citing the work will find the fuller reading companion in the communist manifesto readers guide, which treats the text chapter by chapter. The point for the authorship question is the pattern this first case establishes: in this partnership, the name on the cover is the beginning of the inquiry, never its conclusion.

A second feature of the Manifesto case deserves emphasis because it recurs throughout the article: the evidence improved with time, and improved in one direction. The pamphlet’s early editions appeared without authors’ names. The attribution to Marx and Engels was established publicly in later editions the two men supervised, Engels’s drafts surfaced in print only after his death, and the letters documenting the drafting winter emerged when the correspondence was published in the twentieth century. At no stage did better evidence make the picture simpler. Each documentary layer separated the contributions more finely, distinguishing commission from draft, draft from text, and text from the doctrine later read back into it. That pattern, evidence refining rather than dissolving distinctions, is the signature of a genuine scholarly problem, as opposed to a legend dissolving under scrutiny. It will be met again in the handwriting evidence of the next case, and again, with far larger consequences, in the manuscripts of Capital.

There is also a warning in the Manifesto case about the afterlife of joint names. Because the pamphlet carried both names into fame, later readers attributed its most quoted sentences to a partnership voice that never existed as a writing instrument, and commentaries multiplied that treated every line as the equal product of two minds. The same habit, applied to the economic work, produced the assumption this article most needs to dismantle: that a text assembled by Engels from Marx’s drafts is simply Marx’s text, full stop, differing from the first volume of Capital only in being later. The first volume is the control case against which that assumption fails, because it is the one large work Marx himself prepared for the press, revised through successive editions, and saw published in his lifetime. Its authority as evidence of its author’s intentions is of a different order, a contrast the capital volume one complete guide sets out in detail. Between a book an author finished and a book his executor constructed stretches the whole territory of the authorship question.

The pamphlet’s own edition history adds a footnote that students of attribution relish. The names of the authors appeared on the work only gradually: the first edition was anonymous, the authorship was an open secret in movement circles, and the title pages of later editions, including the English translation of 1888 that Engels edited and introduced, regularised the joint attribution at precisely the moment when the partnership’s internal division of the text was least visible to readers. Engels’s 1888 introduction, a small masterpiece of movement history, names the circumstances of composition without separating the contributions, reserving the separation for the 1883 statement about the fundamental proposition. Nothing in this is discreditable; title pages are not apparatus. But the sequence illustrates a general law of the subject: public attribution consolidated, into two names on a cover, at the historical moment when private attribution, in letters and prefaces, was quietly recording how unequal the shares had been. Readers who meet only the consolidated form inherit the simplification, which is why the corrective literature always begins by reopening the file the cover closed.

A word should also be said about the drafts as physical survivors, because their survival was narrow and their evidence therefore precious. The Principles of Communism circulated in manuscript among League members, was never published in Engels’s lifetime, and reached print only in the twentieth century; the earlier catechism draft survived in a single copy among the papers. Had either been lost, the scaffolding of the Manifesto would be a matter of inference from the 1883 statement and the letters, and the sceptical tradition that periodically claims the pamphlet for Engels, or for some committee of the League, would have more room to work in. The episode is a useful antidote to a lazy assumption that attribution questions are settled by what happens to survive and are therefore arbitrary. Survival is contingent, certainly. But where the documents do survive, they constrain: no reading of the drafts can make the Principles the finished pamphlet, and no reading of the finished pamphlet can erase the drafts standing behind it. Contingency governs how much we can know, not whether what we know is so.

The commission’s mechanics, preserved in the League’s records and the correspondence, add a deadline’s pressure to the picture and explain the pamphlet’s peculiar finish. The congress closed at the start of December 1847. Marx returned to Brussels and wrote through the winter while the League’s central committee, watching the revolutionary temperature rise across Europe, pressed for delivery; by late January the committee’s patience had formalised into an ultimatum threatening to hand the task to others if the manuscript did not arrive by the first of February. The manuscript reached London around that date, went to the printer, a German exile press in Bishopsgate, and appeared within weeks, on the eve of the February revolution in Paris that made its abstractions look like dispatches. The tempo matters for attribution because it forecloses certain romantic possibilities. There was no winter of leisurely joint composition, no manuscript passed back and forth for polishing between Brussels and Paris. There was a scaffolding delivered in the autumn, a single author’s concentrated labour under threat of replacement, and a text whose unity of voice every reader since has remarked. Deadline history is rarely admitted as evidence in attribution disputes, which prefer the testimony of hands and letters. Here it corroborates them, and it also explains the one feature of the case that still puzzles students: how the partnership’s most famous text can be at once a committee’s commission, one man’s draftsmanship and the other man’s unmistakable final prose. It can, because that is what the calendar required.

The pamphlet’s later prefaces add one more attribution datum, of a gentler kind. When the two authors signed a joint preface to the German edition of 1872, a quarter century after composition, they used the occasion to state together that the Manifesto’s general principles remained, in their judgment, correct, while parts of its practical programme had been antiquated by events and by the very development the pamphlet had predicted. The preface is signed by both, was composed in the partnership’s mature manner, and represents their considered joint ownership of the text’s doctrine, maintained without any revision of the internal division of its writing. Ownership, in other words, and authorship part company in the document itself: the men who would not claim equal shares of the sentences claimed equal responsibility for the doctrine. A citation culture attentive to the distinction can honour both facts at once, crediting the doctrine to the partnership where the partnership claimed it, and the text to its actual maker where the documents assign it. Much confusion in the secondary literature is the sound of those two credits being spent as though they were one.

Reading the drafts against the pamphlet

The scaffolding claim can be tested by any reader, because both drafts and the finished text are in print, and performing the test once is worth more than a chapter of assertion. Take the question of the family and of property, which the Principles handles in its final answers as a series of sober policy statements: private property will be abolished as the development of industry makes its abolition possible and necessary, and the consequences are traced through education, the position of women and the organisation of production. In the Manifesto the same material returns transfigured, compressed into polemic, the abolition of the family hurled back at bourgeois critics as a charge they are invited to choke on. The propositions are recognisably continuous; the voice is not. Or take the measures with which the pamphlet’s second section ends, the list of demands addressed to the most advanced countries. Their ancestors stand in the Principles as answers to a catechism question about how the proletariat will use political power, recognisably the same programme, in the same order of business, in the tone of an instructor. The comparison yields the partnership’s characteristic fingerprint: Engels’s expository order, Marx’s rhetorical detonation. Neither contribution is reducible to the other, and the drafts give the lie equally to the legend of equal co authorship and to the counter legend that the junior partner merely fetched the coffee. He fetched the architecture. The other man built the fire inside it.

The drafts also settle, incidentally, a question that recurs in popular discussion: whether the Manifesto’s doctrines were toned down or sharpened in the writing. Set the Principles beside the pamphlet on the state, on religion, on national differences, and the movement is consistently toward compression and edge, the instructor’s qualifications falling away as the pamphleteer takes over. Scholars cite the contrast when they want to illustrate the two men’s prose temperaments, and teachers use it because it can be shown on two facing pages. For the ledger, the point is simpler. A text whose temper can be shown to have changed between draft and publication, in a direction matching the known habits of the man who held the pen last, is a text whose attribution has been done properly, by process rather than by title page. The Manifesto passes that test more completely than almost any other work in the corpus, which is why it anchors the ledger’s first row and why the article opened with it.

The German Ideology and the evidence of hands

If the Manifesto shows the partnership dividing a task in sequence, The German Ideology shows it working in parallel, at the same table, on the same pages. The manuscript was composed in Brussels in 1845 and 1846 as a reckoning with the Young Hegelian philosophy in which both men had been formed: long polemical chapters against Feuerbach, Bruno Bauer and Max Stirner, of which the Stirner chapter ballooned into the bulk of the whole. The two authors could find no publisher willing to take the risk, the manuscript was set aside, and it remained unpublished in their lifetimes, appearing in print only in the twentieth century. That publishing history matters for the authorship question, because an unpublished manuscript preserves the physical traces of its own composition. A printed book hides its drafting behind the compositor’s uniformity. A manuscript displays it, in inks, in hands, in crossings out and insertions, in the fair copy made by one writer and corrected by another. The German Ideology, surviving as a bundle of sheets rather than a bound book, is legible as an event of writing in a way no printed text can be.

Manuscript scholarship has exploited that legibility with great care. The hands of the two men are distinguishable: Marx’s writing is famously difficult, a compressed and heavily corrected scrawl, while Engels wrote a clear, regular hand that made him the partnership’s natural copyist. Working sheet by sheet, editors of the critical edition have assigned the drafting of chapters between the two, and the resulting map is instructive. Much of the fair copy stands in Engels’s hand, in his capacity as scribe, carrying Marx’s marginal additions and corrections in the margins and between the lines. Other sections were drafted by Marx and copied or left uncopied. A third contributor complicates the title page further: Moses Hess, then close to both men, supplied material that entered the manuscript, so that the work conventionally shelved under two names contains the labour of three. The Feuerbach chapter, the portion on which the manuscript’s fame as the birthplace of the materialist conception of history chiefly rests, is itself a layered object, assembled from drafts in both hands, never brought by its authors to a finished state.

What does the handwriting in The German Ideology actually prove?

It proves who physically drafted and copied particular sheets, and where the second author revised the first, which lets editors assign chapters with reasonable confidence. It does not by itself prove whose ideas a passage contains, because Engels often served as scribe for dictation, copying and joint drafting sessions in Brussels.

That distinction, between the hand and the mind, is the central methodological lesson of the case, and it generalises across the whole authorship question. Handwriting evidence is genuine evidence, the hardest kind this subject possesses, but it answers a physical question. In a partnership where one man routinely wrote out what the two had discussed, where dictation and copying were ordinary labour, the presence of Engels’s hand on a sheet is compatible with several different accounts of the ideas on it. Conversely, a sheet in Marx’s hand may record conclusions the two had reached jointly in conversation the previous evening. The mature scholarship therefore uses the hands as a skeleton and clothes it with the correspondence, with the men’s other writings of the same months, and with the internal evidence of style and preoccupation. The result is neither the naive picture of a dual authored book, every sentence jointly owned, nor a cynical picture in which attribution is a lottery. It is a graded map: this chapter chiefly drafted by one, that chapter by the other, this famous passage surviving in a copyist’s hand with the partner’s corrections thickening in the margin.

The German Ideology also teaches the article’s second methodological lesson, about the difference between a text and a work. Generations of readers met the manuscript through editions that arranged its fragments into a smooth doctrinal treatise, with the Feuerbach material presented as a systematic statement of a finished philosophy of history. The critical edition, printing the sheets in their documentary state, revealed something less tidy: drafts at different stages, polemics abandoned mid argument, a conception being worked out against opponents rather than expounded. The distance between the arranged treatise and the documentary bundle is an editorial distance, created by well meaning twentieth century editors facing the same problem Engels faced with Capital: how to make a readable book out of unfinished paper. The parallel is not an accusation against either set of editors. It is a demonstration that the problem of posthumous construction arises wherever this partnership’s manuscripts are published, and that the only safeguard is the reader’s awareness of which object, manuscript or construction, a given citation actually touches.

One further point from this case will matter when the verdicts are weighed at the end. Nobody, in the long argument over The German Ideology, has accused its twentieth century editors of falsifying Marx and Engels, although their arrangements were substantial and although the arranged version shaped decades of interpretation. The scholarly community treated arrangement as a service to be assessed, corrected and eventually superseded by better editions, not as a scandal to be exposed. Remembering that calm judgement is useful ballast when the same community, or its popularisers, turn to Engels’s far more difficult editorial task and reach for the vocabulary of betrayal. The standards applied to editors should not change because one of the editors was the author’s closest friend.

The afterlife of The German Ideology in print offers the comparison case that proves the partnership’s problem was structural rather than personal. When editors of the 1920s and 1930s first brought the manuscript to press, they faced sheets without a settled order, in two hands and a third contributor’s, for a book the authors had abandoned eighty years before. Their solutions were the solutions of their trade: they selected, ordered and titled, presenting the Feuerbach material as a coherent doctrinal chapter and giving generations of readers a treatise the Brussels table never saw. Mid century editions in several languages reproduced the arrangement, and a whole interpretation of the young Marx, the theory of history as a finished system of 1845, grew up on the arranged text. The critical edition’s documentary presentation then did to those editors what it did to Engels: it did not accuse them, it superseded them, printing the sheets in their manuscript order and letting readers watch the conception being assembled rather than receive it assembled. The scholarly moral travelled beyond this one manuscript. A field that had learned, expensively, to ask which object it was citing in the case of the third volume learned to ask it of the 1845 manuscripts as well, and the question turned out to be the same question wearing different dates.

The Stirner chapter, the manuscript’s sprawling bulk, adds a final refinement about hands. The polemic against Max Stirner’s The Ego and Its Own runs to hundreds of sheets, and the division of labour within it has been studied as closely as any stretch of the corpus. The findings present a characteristic Brussels tableau: long runs of fair copy in Engels’s hand, passages of draft in Marx’s, corrections layered over both, and stretches where the pace and the pressure of the jokes suggest dictation, two men amusing each other at a third man’s expense across a table. Scholars assign probable authorship of particular sections with appropriate humility, and the honest summary for the ledger is that the chapter is jointly produced in the fullest sense the partnership ever achieved, conversation made visible as text. Set that beside the Manifesto, drafted apart, beside the journalism, ghosted, and beside the third volume, constructed, and the partnership’s repertoire is complete. Four modes of dual production, four different citation practices, one pair of names on the covers. That is why a ledger, and not a slogan, is the right instrument for this subject.

The small print of the partnership: editions, updates and a famous correction

Between the great cases stand smaller ones that repay attention because their scale lets the evidence be seen whole. The most charming is Engels’s edition of Marx’s Wage Labour and Capital, the series of articles from 1849 that Engels republished in 1891 as a popular pamphlet. In preparing it he altered the text in one systematic respect: where the young Marx, following the economists of his day, had written of the worker selling his labour, Engels substituted the mature terminology of Capital, the sale of labour power, throughout the pamphlet. He then did the thing that distinguishes an editor from a ventriloquist. In his introduction he announced the alterations, explained that Marx would certainly have made them himself in a new edition, argued that the original wording could mislead the pamphlet’s intended readers about the theory’s central distinction, and left the decision visible for anyone to challenge. Party critics did grumble, and later editions have recorded the changes. The episode is a perfect miniature of the whole problem: a real intervention, theoretically motivated, openly disclosed, small in extent, and arguable in principle. Every quality that critics seek in the third volume case, at pamphlet scale, with the disclosure done right. Set beside the third volume, where the disclosure was general and the interventions architectural, the miniature proves that the partnership’s editor understood passage level candour perfectly well. What he did not do, in the great case, was supply it, and the difference in scale between a pamphlet’s terminology and a system’s architecture is the usual explanation offered for the difference in candour. The explanation is plausible. It is not quite an excuse, since the candour mattered more, not less, where the construction was greatest.

Earlier works supply the contrast class that keeps the partnership’s categories honest. The Poverty of Philosophy, Marx’s 1847 answer to Proudhon, is wholly his, written in French, in his own polemical voice, with no Engels in its production beyond the background of their discussions. The Eighteenth Brumaire of Louis Bonaparte, the Class Struggles in France, the Civil War in France with its address drafted for the International, are Marx’s solo performances, prepared by him for publication and printed under his direction, and they show what Marx’s finished prose looks like when no editor and no collaborator intervenes: periodic, allusive, historically saturated, structurally wayward by any textbook standard, and unmistakably the same mind that fills the manuscripts with crossings out and fresh starts. Reading the solo works against the posthumous volumes is an education in the editor’s effect that requires no apparatus. The solo works sprawl and gleam. The constructed volumes march. A reader who prefers the march is entitled to the preference, and Engels, who built the marching order, deserves the credit for it. What the reader is not entitled to do is call the march the author’s own step. That sentence, in the end, is the whole of the edition attribution rule before it puts on its scholarly dress.

The Holy Family itself, the ledger’s control case for joint work, rewards one more paragraph, because its internal division illustrates the difference between collaboration by assignment and collaboration by osmosis. The title page of 1845 announced a critique of critical criticism, against Bruno Bauer and his circle, and the preface assigned the work to both authors. The correspondence and the text together show the shares: Engels contributed a few short chapters, written during a brief stay in Paris, while Marx, who remained with the project after Engels left for Barmen, expanded his own assignment into the bulk of the book. Even the partnership’s most genuinely joint production, in other words, divides unevenly once the documents are consulted, and divides in the same direction as the Manifesto a few years later: Engels quick, programmatic and brief; Marx slow, expansive and unable to leave a subject until it had been turned inside out. Temperament, visible in the juvenilia, is still visible in the executor’s study forty years on, where the quick man’s vice, the love of a clean system, and the slow man’s vice, the unfinished mountain of paper, met in a single editorial act.

Journalism under borrowed names

The third case is the least dignified and, for working scholars, among the most useful, because it concerns the texts through which the partnership earned its living and reached its public. From 1851 Marx was the London correspondent of the New York Daily Tribune, then among the largest circulation newspapers in the world, and the arrangement lasted a decade. The correspondence was his financial lifeline and his drudgery. His English in the early years was unequal to the speed required, his knowledge of military affairs was thin, and his health and his economic research competed for the same hours. Engels, in Manchester, supplied the deficiency. He wrote, in whole or in large part, many of the articles that crossed the Atlantic under Marx’s name, above all the military correspondence, where his study of warfare and his experience of the 1849 campaign made him the partnership’s authority. The letters between London and Manchester document the traffic in detail: requests for articles on named topics by named deadlines, drafts sent down for Marx to translate, revise or forward, complaints about editors, calculations of payment. The byline was a commercial fact, the name of the contracted correspondent. It was not an attribution in the scholarly sense, and the two men treated it with a pragmatism that later collectors of their works found deeply inconvenient.

The scale of the inconvenience became clear with the series on the German revolutions of 1848 and 1849, published in the Tribune in 1851 and 1852 as the work of its correspondent. The series, later reprinted as Revolution and Counter-Revolution in Germany, was attributed to Marx for sixty years and entered the canon as his account of the revolution he had lived through as a newspaper editor in Cologne. When the correspondence between the two men was published in 1913, the letters showed that Engels had written the series, at Marx’s request, while Marx was absorbed in economic studies. The reattribution was accepted without serious resistance, because the documentary evidence was complete, and the episode stands as the cleanest demonstration in the whole authorship dossier that the signature on a piece of this partnership’s journalism proves almost nothing. What proved authorship was the archive: drafts, letters, the internal evidence of who knew what.

Why did articles by one partner appear under the other’s name?

Because the Tribune contracted with Marx as its correspondent while Engels, uncontracted and employed in Manchester, was frequently the faster and better equipped writer, especially on military subjects. Deadlines, income and Marx’s slow English made the substitution practical, and their letters record it as routine business between them.

The same correspondence disciplines the reverse question, which is asked often enough to deserve a plain answer. Did Marx write articles that Engels signed? The traffic the letters document runs overwhelmingly in one direction, from Engels’s desk into Marx’s column. Engels was the partner with a salary, a fluent journalistic hand and time bought by a regular working day; Marx was the partner with the contract, the reputation with editors and the chronic shortage of hours. Occasional pieces moved the other way in the partnership’s earlier newspaper ventures, and each man at times drafted material the other revised and published, but no body of journalism signed by Engels has been shown to be Marx’s composition in anything like the Tribune proportions. The honest summary is asymmetry. As ghost writer, Engels served Marx’s byline on an industrial scale; the reverse service was episodic. Any collected edition that prints the journalism must therefore attribute article by article, from the letters and the manuscripts, and the good editions do exactly that, reassigning dozens of pieces between the two names as the evidence dictates.

Three consequences follow for readers. First, arguments built on a single Tribune article attributed by signature alone are built on sand, and the history of the subject contains embarrassing examples of scholars explicating Marx’s view of a campaign he never analysed, in prose his friend wrote between office hours. Second, the episode recalibrates what the two names mean on any title page from the partnership’s working years: they denote a firm as much as a pair of individuals, a going concern with a division of labour, internal subsidies and house styles. Third, and most important for what follows, the journalism established the working relationship that the editing of Capital would later extend. Engels had been writing as Marx, with Marx’s knowledge and for Marx’s purposes, since 1851. When, after 1883, he came to construct books as Marx from Marx’s papers, he was not assuming an unfamiliar role. He was continuing, in a graver register and under a self imposed discipline he described in his prefaces, the ventriloquism of thirty years. Whether the continuation was faithful is the question the remaining cases decide, but its roots in this documented history of borrowed names should be kept in view. The man had been the partnership’s second voice, in print, for most of his adult life.

The journalism case has an earlier chapter that belongs here, in the revolutionary years themselves. At the Neue Rheinische Zeitung in Cologne in 1848 and 1849, the division of labour was institutionalised in a masthead: Marx as editor in chief, Engels among the editors, writing leaders, reporting the Frankfurt Assembly, and eventually chronicling the Hungarian war and the defence of the Palatinate and Baden, in which he served as a combatant and adjutant. The paper’s suppression in May 1849 ended the experiment, but not before it had established the pattern the Tribune years would industrialise: Marx’s name fronting the enterprise, Engels’s pen supplying a share of its contents out of proportion to his billing, and the two men treating the arrangement as the natural economy of a firm. When the paper was revived as a review in London in 1850, the pattern persisted under exile conditions, with Engels contributing the studies of the German campaign that drew on his soldiering. Scholars assembling either man’s journalism from these years must work as the Tribune’s editors learned to work, from the correspondence and the manuscripts outward, and the collected editions print the two men’s articles in interleaved volumes that are monuments to the impossibility of separating the firm’s output by signature.

One further journalistic case has done disproportionate work in the popular literature and should be returned to its proper size. Marx’s articles on Lord Palmerston and on British rule in India, written for the Tribune in 1853, are his own compositions, however startling their judgments, and they are sometimes quoted in arguments that attempt to redistribute the partnership’s opinions between the two men. The attribution of the India articles is not in doubt, and their presence in the record is a useful reminder that the authorship question cuts no ideological figure: the same documentary method that takes Revolution and Counter-Revolution away from Marx leaves him in sole possession of texts that other readers would prefer to reassign. A method that only ever moved credit in one direction would be a suspicion machine. This one is a scale, and it weighs the famous pamphlet, the military journalism and the Indian correspondence with the same weights.

The working detail of the Tribune arrangement, recovered from the letters, deserves preservation because it shows ghost writing as the partnership’s ordinary metabolism rather than its occasional expedient. In the first years Marx’s written English lagged behind his reading of it, and the workflow ran accordingly: Engels composed articles in German or English as time allowed, Marx revised, translated where necessary, and dispatched the result over his own signature, sometimes telling his friend exactly which paragraphs he had altered and why. The military correspondence of the Crimean War years, when the Tribune’s readers demanded operational analysis at a pace set by the telegraph, fell almost entirely to Manchester, and Engels followed the campaigns with maps and the professional literature, complaining in the letters about the necessity of predicting the movements of generals he despised. Charles Dana, the paper’s managing editor, dealt with the correspondent in London and for years had no reason to know that a second hand in Manchester was supplying a share of the copy, a share that can be estimated for some years at close to half the output. When scholars speak of the Marx of the Tribune, the great journalist interpreting America, India and the European war, they are speaking, in a proportion no honest citation should conceal, of a two man bureau. The bureau’s existence is not a scandal and was not, within the conventions of the trade and the friendship, a deception practised on anyone in particular. It is simply the largest single demonstration that in this firm the byline and the pen had parted company early, by design, and never entirely rejoined.

Boundary cases: the will, the notebooks and the other editors

Four neighbouring cases sharpen the ledger’s distinctions by showing the partnership’s attribution problems in their other forms, and a serious treatment owes them a section because each has generated its own confusions.

The first is The Origin of the Family, Private Property and the State, published under Engels’s name alone in 1884, the year after Marx’s death. Its preface performs an act of attribution rare in any literature: Engels explains that the book represents the fulfilment of a bequest, since Marx had intended to present the results of Lewis Henry Morgan’s researches himself, had filled notebooks with excerpts and critical comments on Morgan’s Ancient Society, and had left those notebooks as the material from which the study was to be built. Engels worked from them, supplementing Morgan with his own reading and with the later literature, and said plainly that Marx’s notes stood behind his exposition. The book is therefore Engels’s composition erected on Marx’s research dossier, a third configuration beside the Manifesto’s scaffolding and the third volume’s construction: research by one, writing by the other, declared on the title page’s behalf by the writer himself. Citations that attribute its arguments simply to Marx invert the actual relation; citations that attribute them simply to Engels erase the dossier. The ledger’s lesson generalises: in this corpus, the honest citation often needs a sentence, not a name.

The second case is Marx’s Critique of the Gotha Programme, a set of marginal notes on the draft programme of the unified German workers’ party, written in 1875 and circulated privately among the party leadership. Engels published it in 1891, against the wishes of the party’s leading figures, on the eve of a congress about to adopt a new programme, judging that the movement needed its founder’s criticisms in print. The case adds a category to the editor’s repertoire: not selection, not arrangement, but the decision to publish at all, taken against living custodians of the author’s reputation, on the executor’s judgment of the author’s deeper intent. Whatever one thinks of the decision, and the party leadership of the day thought poorly of it, it displays the same trait as the third volume’s construction at a smaller scale: Engels exercising sovereign judgment over the posthumous text, openly, and accepting the controversy as the price. Attribution here is uncomplicated, since the notes are wholly Marx’s. What the case complicates is any picture of the executor as a passive instrument. He was an actor with a programme of his own for the papers, and the programme, publish, complete, make usable, is the context in which his heavier interventions should be read.

The third case extends the problem beyond Engels altogether. The manuscripts Marx had drafted in 1861 to 1863, containing the great historical and critical survey later known as Theories of Surplus Value, were edited for publication between 1905 and 1910 by Karl Kautsky, then the movement’s leading theorist. Kautsky, facing the same sprawling, digressive draft that any editor of Marx must face, rearranged its parts into a thematic order he judged more logical, and presented the result as the fourth volume of Capital. Twentieth century scholarship, once the manuscripts could be consulted, judged the rearrangement a distortion of Marx’s design, and the critical edition eventually restored the draft’s own sequence. The episode matters here because it breaks the personal frame. Given Marx’s unfinished paper, editors of very different temperament and generation reached for arrangement as their instrument, because arrangement is what such paper demands. Engels’s interventions belong to a class of editorial acts that his successors repeated, and the comparison cuts both ways: it normalises his construction, and it warns that even sympathetic, expert, well intentioned construction, Kautsky’s no less than Engels’s, can quietly replace the author’s design with the editor’s sense of order.

The fourth case is a single manuscript with an instructive physical history: the theses on Feuerbach, eleven brief propositions Marx jotted in a notebook in the spring of 1845, which Engels found among the papers, recognised as the germ of the new world outlook, and published in 1888 as an appendix to his own book on Feuerbach. In publishing them he made small alterations, smoothing the notes toward print. The critical edition prints the notebook text and the alterations side by side, and they are the slightest of touches beside the third volume’s architecture. Yet the case is a favourite of the seminar room precisely because it is small enough to hold in the hand: a canonical text, eleven sentences that include the most quoted sentence Marx ever wrote about philosophy and practice, reaching its readers through an editor’s polishing, with the unpolished original recoverable and the differences itemised. Students who grasp why the theses can be both Marx’s beyond doubt and Engels’s edition beyond doubt have grasped the logical space in which the whole authorship question lives.

The authorship ledger

The cases surveyed so far, and the two great cases still to come, can be gathered into a single findable object. The authorship ledger sets out, for each major text of the partnership, who drafted the material, who produced the final version, what the manuscript evidence shows, what remains disputed, and how the text should be cited by anyone who wants an attribution to survive scrutiny. Read down the citation column and the pattern of the whole article is visible at a glance: the closer a text stands to a version its author prepared for print, the simpler its citation; the more editorial construction intervened, the more the citation owes the reader a specification of which object is being quoted.

Text	Who drafted it	Who produced the final version	What the manuscript evidence shows	What remains disputed	How it should be cited
The Communist Manifesto	Engels, in two catechism drafts	Marx, writing the pamphlet itself	Drafts and letters assign the scaffolding to Engels and the text to Marx	How much phrasing carried over from the drafts	As Marx’s text, with Engels’s drafts noted
The German Ideology	Both men, with material from Moses Hess	No final version was prepared in their lifetimes	Hands are distinguishable; much fair copy is in Engels’s hand with Marx’s corrections	The precise division of several chapters	By chapter where editors have assigned hands, with the manuscript state flagged
The Holy Family	Both men, in assigned chapters	Both, published jointly in 1845	Title page and correspondence allocate chapters; Marx wrote the larger share	Little; the allocation is unusually well documented	As a genuinely joint work, chapter by chapter
Tribune journalism and Revolution and Counter-Revolution	Frequently Engels, under Marx’s contract	Marx as correspondent, often forwarding Engels’s drafts	The published correspondence assigns articles between the two	A residue of pieces the letters do not settle	Article by article from the correspondence, never by signature alone
Capital, volume one	Marx	Marx, through editions he revised himself	Authorial editions and revisions are documented in print	Little on attribution; interpretation is another matter	As Marx’s finished text, edition specified
Capital, volume two	Marx, in manuscripts of different dates	Engels, selecting and combining in 1885	Manuscripts survive and can be compared with the edition	The effect of Engels’s choices among variants	As Marx’s text in Engels’s edition, with the manuscript basis noted
Capital, volume three	Marx, chiefly in a manuscript of the mid 1860s	Engels, constructing the book in 1894	The critical edition prints the manuscript beside evidence of reordering, divisions and smoothing	The significance of the interventions, not their existence	As Engels’s edition of Marx’s manuscript, never as unmediated Marx
Anti-Dühring	Engels	Engels, with a contribution from Marx	Engels acknowledged that Marx read the whole and wrote a section on political economy	The exact extent of Marx’s contributed material	As Engels’s work, with Marx’s contribution noted

Two rows of the ledger deserve a word before the argument moves to its centre. The Holy Family, the partnership’s first joint book, published in 1845 against their former mentor Bruno Bauer, is the control case for genuine joint authorship: the chapters were parcelled out in advance, each man wrote his own, and the correspondence leaves the allocation in little doubt. Set beside it, the later habit of treating every double named text as though it had been produced that way collapses. Joint composition existed in this partnership, it looked like The Holy Family, and most of the famous texts were not produced like that at all. The second row to notice is the first volume of Capital, the other control case. Marx published it in 1867, revised it for a second German edition, supervised its French translation in instalments and treated that French version as having a scientific value independent of the original, and prepared further revisions that Engels later incorporated into the third and fourth German editions. Even here, the careful reader notices, the text is not perfectly simple: the standard English edition descends from a German text partly assembled after Marx’s death. But the assembly followed authorial instructions to an extent that keeps the volume in a different evidential class from its successors, a difference the rest of this article exploits.

The ledger’s rows are not all of the same evidential weight, and its honest use requires the differences to be kept in view. At the strongest end stand the rows settled by convergent documents: the Manifesto, where drafts, letters, a deadline and a participant’s testimony all point the same way; the journalism, where the correspondence assigns articles outright; the first volume, where the authorial editions survive in print. In the middle stand the rows settled by the physical evidence of the manuscripts read together with the correspondence: the division of The German Ideology, the construction of the third volume, where the facts of intervention are certain and only their significance is argued. At the weakest end stand the residues the ledger honestly marks as disputed: the exact extent of Marx’s contributed material inside Anti-Dühring, the division of particular chapters in Brussels, the handful of articles the letters do not reach. A citation practice built on the ledger inherits its gradations. Claims resting on the strongest rows can be made flatly. Claims resting on the middle rows should carry their flags. Claims resting on the weakest rows should be advanced as probabilities and defended as such. Much of the quarrelling in the secondary literature dissolves when the parties are made to state which row, and which grade within the row, their claim occupies, and the dissolution is a further argument for keeping the ledger, rather than any verdict, at the centre of the subject’s teaching.

Capital volume two: building a book from a shelf of drafts

Marx died in March 1883 leaving the continuation of Capital in a state that only one living person could navigate. The first volume had been in print since 1867. Behind it lay a shelf of manuscripts for the second and third books, accumulated across two decades, in varying hands of legibility, at varying levels of finish, some of them superseded drafts of the same material, some partial, one of them, the principal draft for the third book, dating from the middle of the 1860s, before the first volume had even appeared. Engels, at sixty two and with his own projects in hand, took the papers into his keeping, had copies made by assistants because Marx’s handwriting defeated most readers, and set about deciding what could be published. The second volume appeared in 1885, two years after the death, a speed that itself testifies to the relative finish of its materials and to the editor’s restraint in handling them.

What Engels found for the second book was not one manuscript but a series of them, written between the early 1860s and the late 1870s, which he catalogued in his preface by number and date. The drafts differed in scope and in the state of their arguments: some chapters existed in several versions, some in only a fragmentary one, and the final part of the book, on the reproduction of the total social capital, existed in a comparatively polished late manuscript. His method, stated in the preface and borne out by later comparison, was selection and combination. He chose the most developed version of each part, joined the chosen versions, corrected obvious slips, and intervened as little as he judged compatible with a readable book. Where Marx had left alternative formulations standing side by side, the editor picked one, and the picking, though invisible in the printed text, is the kind of decision that later critical editions exist to expose. The manuscripts for the second book reached print in the critical edition in the early twenty first century, allowing the same comparison that had transformed study of the third, and the comparison broadly vindicated the editor’s account of his own restraint, while documenting thousands of small decisions that no preface could have itemised.

What did Engels find when he opened the manuscripts for volume two?

A series of drafts from two decades rather than a single book: several versions of some chapters, fragments of others, and one late, comparatively finished manuscript on reproduction. His preface catalogues the drafts by date and explains that he selected and combined the most developed versions.

The relative innocence of the second volume, editorially speaking, is itself an important datum, because it calibrates the scale for judging the third. Where Marx’s drafts were late, coherent and close to his mature intentions, Engels’s edition stays close, and the scholarly quarrel is muted. The interventions grow as the manuscripts grow older, rougher and more remote from the positions Marx had reached by the time of his death. That gradient matters. It suggests an editor whose hand was forced by his materials rather than an editor imposing a system of his own on whatever came to hand: the same man, with the same loyalties and the same lifetime of ventriloquism, produced a light touch where the paper allowed it and a heavy one where the paper demanded it. Critics of the third volume’s construction must reckon with the second volume’s restraint, and defenders of the third must reckon with the fact that the difference between the two editions is explained by the manuscripts, not by any change in the editor.

There remains, even in this gentler case, the matter of the invisible join. A reader of the second volume in any standard edition reads a continuous argument about the circulation of capital, the turnover of its parts and the reproduction of the whole, and the continuity feels authorial. Yet its joints were chosen by another mind: this version of a chapter rather than its predecessor, this paragraph retained from an earlier draft, that transition supplied because the manuscript broke off. None of this approaches construction in the heavy sense defined earlier, but it does mean that the second volume, too, is an edition and not a transcript, and the attribution practice this article recommends applies to it in the milder form set out in the ledger: cite it as Marx’s text in Engels’s edition, note the manuscript basis where an argument turns on a passage’s exact form, and reserve the strongest form of the rule for the volume to which the next section turns.

One more feature of the second volume’s production deserves record, because it bears on the verdicts collected later. Engels published, in his preface, an open announcement that the papers contained a solution to a riddle about the equalisation of profit rates and prices, set as a challenge to other economists, whose attempts he would judge when the third volume appeared. The gesture tells us something about the editor’s self understanding. He saw himself as a steward presenting his friend’s unfinished achievement to the world, confident enough in the material to stage a public test of it, and conscious that his own name would stand in the prefaces, not on the title page, as guarantor. Stewardship of that kind is neither ghost writing nor co authorship in the ordinary sense. It is a third thing, for which the English language has the plain word executor, and the question the third volume raises is whether the executor’s duties, on worse materials, quietly became an author’s.

Later comparison work on the second volume, undertaken once the critical edition had printed the underlying manuscripts, refined the picture in ways worth recording because they show the comparison method correcting itself. Regina Roth’s editorial studies of the second book’s manuscripts documented the chain of decisions in detail: which manuscript Engels took as his base for each part, where he drew a missing transition from an earlier draft, where the printed text’s paragraphing regularised a manuscript whose author had run discussions together. None of the findings overturned the general verdict of restraint, and several confirmed that where Engels departed from his base manuscript he usually followed a later, superseding intention he could document from the papers. But the studies also established that restraint is a statistical property, not a passage level guarantee. A reader citing a particular formulation from the second volume, a definition of turnover time, say, or a clause in the reproduction schemes where departments exchange, cannot know from the edition alone whether the sentence is the author’s final version or the editor’s selected variant. The practical consequence matches the ledger’s citation column: the second volume is cited as Marx’s text in Engels’s edition, with the manuscript basis noted where an argument turns on exact wording. The volume’s scholars have absorbed this so completely that it rarely needs stating, which is precisely the state the third volume’s scholarship has not yet reached, and the contrast between the two volumes’ citation cultures is itself evidence of what a generation of apparatus work can do.

There is a temptation, resisting it is part of this article’s discipline, to treat the second volume’s relative innocence as settling the editor’s character and the third volume’s difficulties as an aberration explained by bad materials. The temptation should be refused in both directions. Character is not the variable the comparison measures; the relation between a particular manuscript’s state and a particular edition’s finish is. Where the draft was finished, the edition is transparent. Where the draft was exploratory, the edition is constructed, and the construction carries the expositor’s habits into the text regardless of the constructor’s intentions, because no one can arrange another man’s fragments for a decade without the arrangement learning his accent. That formulation, neither accusation nor absolution, is the one the evidence supports, and it sets the terms for the harder case that follows.

The second volume’s three parts differ among themselves in finish in ways that reward a closer look, because they let the reader watch the gradient operate inside a single book. The opening part, on the metamorphoses of capital and their circuit, descends from manuscripts of several dates, and its printed smoothness is the product of Engels’s selection among successive attempts at the same exposition, attempts in which Marx can be seen refining the notation and the sequence. The second part, on turnover, rests on material that includes a manuscript largely devoted to assembling and criticising the literature, from the physiocrats through Smith and Ricardo, and the editor’s task there approached excerpting a treatise from a research dossier. The third part, on the reproduction and circulation of the total social capital, stands at the far end of the gradient: a late manuscript, written with a care that suggests its author knew it for the advance it was, containing the celebrated schemas of simple and expanded reproduction in a state close to publishable. Critics of the edition’s handling have, accordingly, concentrated on the first two parts and found correspondingly less to dispute in the third, while historians of economics have quarried the reproduction schemas for a century with an innocence about their provenance that the manuscripts largely excuse. The pattern confirms the article’s operating principle from within a single volume. Editorial consequence tracks manuscript state. Where the state is good, the editor vanishes from the evidence; where it is rough, he stands in every paragraph, whether he wished to or not.

A related point concerns what Engels did not publish from the second volume’s materials. The manuscripts contain extended engagements with predecessors, repetitions in which Marx rehearses a distinction three ways before choosing, and calculations abandoned mid page. The edition trimmed these to produce a book a reader could finish, and the trimming, uncontroversial as craft, removed from common view the process by which the concepts had been won. Later readers, meeting the concepts in their finished form, have sometimes charged them with being scholastic, not seeing the workshop floor from which the finish was swept. The critical edition restored the floor, shavings and all, and the restoration changed the second volume’s reputation among specialists from a worthy sequel into a visible record of a mind at work. Attribution and appreciation turn out, once again, to be the same discipline practised at different moments: know what kind of object you are holding, and both its authorship and its quality come into focus together.

Capital volume three: the manuscript and the edition

The third volume, published in 1894, eleven years after Marx’s death, is where the authorship question stops being a refinement and becomes a condition of honest argument. The materials were the worst Engels faced. The principal manuscript was a single sustained draft from 1864 and 1865, written before the first volume had been published, exploratory in character, full of calculations, repetitions, false starts and passages where Marx was plainly thinking on paper rather than addressing a reader. Around this core lay later notebooks and supplementary drafts, some from the 1870s, bearing on credit and finance, which the author had never integrated. The manuscript had no authorial chapter divisions for much of its length, its treatments of related themes were scattered, and its most consequential stretch, the analysis of how the rate of profit behaves, existed as a developing investigation rather than a settled exposition. From this, Engels undertook to build the concluding volume of the most ambitious work of social theory of the century, in full knowledge that the world’s judgment of his friend would rest substantially on what he built.

What he built, the printed third volume, has the architecture every reader of Marxism knows: seven parts, proceeding from surplus value and its transformation into profit, through prices of production and the equalisation of profit rates, to the tendency of the rate of profit to fall with its counteracting influences, then to merchant’s capital, to interest bearing capital and credit in a long fifth part, to ground rent, and finally to the revenues and their sources. It reads, for long stretches, like a finished treatise. The comparison that the critical edition made possible asks how much of that finish is Marx’s. The historical critical edition printed Marx’s original manuscript in the early 1990s, roughly a century after Engels’s edition, and scholars including Carl Erich Vollgraf, Jürgen Jungnickel, Regina Roth and Michael Heinrich went through the two texts passage by passage. Their findings, on the factual level, command broad assent even among scholars who disagree about everything else. Engels had reordered material within and between sections. He had supplied the chapter divisions and most of the chapter headings through which readers navigate the book. He had cut, transposed and occasionally rewritten transitional matter. In the fifth part, on credit, where the manuscript was at its most fragmentary, he had in effect composed the arrangement wholesale, selecting among disconnected notes, discarding a substantial quantity of material, and writing the connective tissue that turns a dossier into a part of a book. The printed volume’s appearance of system, its very readability as a treatise, is substantially his achievement.

When could scholars first compare the third volume with its manuscript?

Only after the critical edition published Marx’s original manuscript in the early 1990s, nearly a century after Engels’s edition appeared. Before that, readers had the edition and Engels’s prefaces, but no public access to the draft underneath, so the scale of his construction could not be measured.

The findings need to be held at their exact strength, neither inflated nor deflated, because both errors circulate. Inflation turns construction into invention: it speaks as though the arguments of the third volume were Engels’s ideas foisted on Marx’s name. The manuscript refutes that. The categories, the analyses, the worked examples, the great theoretical moves, the transformation of surplus value into profit, the equalisation of rates, the analysis of rent, are Marx’s, present in the draft, often in formulations the edition reproduces nearly word for word across long passages. Deflation turns construction into secretarial work: it speaks as though Engels merely tidied a book that was already there. The manuscript refutes that too. A draft without its eventual architecture is not the book that influenced a century of argument, and where the architecture does argumentative work, the architect has entered the authorship. Arrangement, as the opening section insisted, is an argument about what depends on what. To gather the scattered remarks on the profit rate into a sequence that presents a law, then its counteracting influences, then the unfolding of its inner contradictions, is to decide how the doctrine will be understood, taught and attacked. That decision was Engels’s, taken conscientiously, announced in general terms in his preface, and invisible in its particulars until the manuscript could be read beside the edition.

Two subsidiary findings sharpen the picture. The first concerns titles and divisions within the profit rate material itself. The chapter headings under which the doctrine is universally cited, the law as such, counteracting influences, the development of the law’s internal contradictions, are editorial formulations, imposed on a continuous draft in which Marx moved between the tendency and its counterweights without the printed book’s clean separation. Scholars who quote those headings as Marx’s own table of contents are quoting the editor’s scaffolding, a small error that stands for the larger one. The second finding concerns the fifth part. Engels himself signalled, in his preface, that this material had cost him the most labour, and the manuscript comparison shows why: the drafts on credit and interest included overlapping treatments written years apart, long excerpts, and plans never executed. His solution was a drastic selection and reordering that produced a coherent part at the price of suppressing the manuscript’s hesitations and duplications. Modern editors, faced with the same dossier, have questioned particular choices, the placement of the material on fictitious capital among them, while conceding that some construction was unavoidable if the part was to exist as a part at all.

It is worth pausing on the human situation, because the neutrality this subject requires is easier to keep with the situation in view. Engels was in his seventies, his eyesight and health declining, working with assistants who transcribed the crabbed manuscripts for him. He believed the third volume completed the theoretical edifice on which his friend’s reputation and, as he saw it, the intellectual honour of the movement depended. He also knew that Marx, had he lived to finish the book, would have rewritten the 1860s draft in the light of twenty further years of study, years in which the author had filled notebooks on Russia, on agriculture, on mathematics, on the very credit system the fifth part treats from older materials. The edition therefore freezes a moving research programme at the moment its author happened to stop, and dresses the frozen draft in the architecture of a conclusion. That is nobody’s fraud. It is the structural predicament of posthumous publication, and it is why the practical rule drawn from this case concerns the reader’s citations rather than the editor’s character.

The reception of the third volume adds a chapter to the case that the manuscript scholars cite with justified emphasis. When the volume appeared in 1894, it was received as the completion of the system, and the most consequential early attack on Marx’s economics, Eugen von Böhm Bawerk’s Karl Marx and the Close of His System of 1896, took the printed third volume as its target and treated its reconciliation of values and prices as the author’s final word, to be judged as the keystone of a closed doctrine. Whatever the merits of that attack, its premise was bibliographical: that the book under review was the book Marx wrote. The premise was universal at the time and unavoidable, since no other book existed to review. But the episode shows how much downstream argument inherits an unexamined attribution. A controversy that structured a century of value theory was conducted, on both sides, against a text whose architecture neither side could inspect. Defenders who answered Böhm Bawerk by appealing to the coherence of the third volume’s design were, in part, praising the editor’s joinery; critics who found the design strained were, in part, detecting the seams of its construction. The manuscript comparison does not award the old controversy to either side. It reveals that the controversy’s shared object was partly an artefact, which is a more interesting result than a verdict, and a standing warning to any field that builds on posthumous texts.

The parts of the volume beyond the profit rate material repay the same attention in miniature, because the editorial gradient within the volume is uneven and instructive. The treatment of ground rent rests on drafts that were comparatively developed, and the comparison shows correspondingly lighter intervention: order adjusted, expression polished, the analysis recognisably presented rather than constructed. The opening part, on the transformation of surplus value into profit and of the rate of surplus value into the rate of profit, stands between the extremes, its manuscript more finished than the credit dossier, less finished than the rent material, its edition bearing the marks of arrangement without wholesale composition. The final part, on the revenues and their sources, breaks off in the manuscript in the middle of the discussion of classes, at the famous point where the text stops just as the definition of class is to be given, and the edition prints the break honestly as a break. That broken ending is worth a seminar’s attention. An editor intent on manufacturing a system had every temptation to round the volume off, and precedents in the editing of other philosophers offered models for a supplied conclusion. Engels printed the fragment as a fragment. The restraint at the volume’s end, like the restraint across the second volume, belongs in the same ledger as the construction in its middle, and any verdict that cannot hold both entries at once is a verdict about a simpler man than the documents contain.

Engels’s decade of construction had a workshop dimension that the prefaces only hint at, and recovering it from the editorial reports helps explain both the achievement and its costs. The manuscripts first had to be read. Marx’s hand defeated most eyes, and Engels, whose own sight was failing, worked through transcriptions made by assistants, among them younger socialists of the London exile circle whose later fame as theorists and party leaders began, in part, at the executor’s desk, copying. Decisions were taken against a mountain of paper that included, beside the principal drafts, excerpt notebooks running to thousands of pages, in which Marx had quarried the literature of political economy, money and banking, agriculture and technology across forty years. Whether a given manuscript passage represented a position or a way station could sometimes be judged only against those notebooks and the letters, and Engels judged with the authority of the one living witness to the author’s intellectual biography. No later editor can replicate that authority, and no later reader should forget that it was personal before it was scholarly: the executor knew what his friend had been reading, arguing and doubting in the years the drafts were written, because he had received the letters in which the reading and the doubting were reported. The construction of the third volume is thus also an act of memory, and its errors, where the comparison finds them, are often the errors of a witness certain he knew what the author meant. Certainty of that kind is the most human of the forces that shaped the edition, and the most resistant to apparatus.

Marx’s own statements about the unfinished work deserve their place beside the editor’s, because they fix what the editor inherited: not a book awaiting assembly, but a project its author knew to be behind his own development. In letters of the later 1860s Marx described the plan of the whole work and his progress through it with a working writer’s precision, and in the years after the first volume he repeatedly postponed the continuation while he restudied whole fields, agronomy and chemistry among them, telling correspondents that the second book would have to wait for material he was still absorbing. By the late 1870s his health had broken the rhythm entirely. The principal manuscript of the third book therefore predates, in its substance, the author’s final positions on several of the questions it treats, a fact Engels knew and the edition’s readers mostly did not. Nothing in that observation indicts the editor, who published what existed and said that the author had intended reworking. But it deepens the attribution rule’s rationale. A citation of the third volume is a citation of an 1860s draft, arranged in the 1890s, by an author whose thought had continued to move until 1883. Three dates stand inside every such citation, and the rule asks only that scholarship stop pretending there is one.

Within the third volume, one more comparison deserves notice because it shows the editor solving a problem of a different kind: not disorder, but silence. The manuscript’s treatment of the transformation of surplus value into profit begins from the author’s recognition that the categories of the first volume, value and surplus value, appear on the surface of the economy in the transmuted forms of cost price and profit, and the draft explores the transmutation with its characteristic mixture of logical development and historical aside. Engels’s edition preserves the exploration’s substance while regularising its path, and the comparison here acquits him of the heavier charges: the part’s difficulties in the literature descend from the doctrine’s genuine hardness, not from editorial distortion, and scholars who have collated the texts treat this stretch as evidence that the editor’s hand, where the draft was conceptually settled, confined itself to the surface. The finding matters for proportion. A prosecution that exhibits only the fifth part and the profit rate chapters builds a true case about those places and a false impression about the whole. The volume is not uniformly constructed any more than the second volume is uniformly innocent, and the attribution rule’s per passage discipline, tiresome as it can seem beside the pleasures of a sweeping verdict, is precisely the habit that proportion requires.

The rent part supplies the counterweight from the other direction, a stretch where the manuscript’s relative finish let the author’s own organisational intelligence show through the edition undisturbed. Marx’s analysis of differential and absolute ground rent was among the most worked over materials in the whole Nachlass, reprised across the manuscripts of the 1860s, and the printed part’s architecture, different forms of rent distinguished, the historical illustrations ranged, the polemic against Rodbertus woven through, descends from arrangements the author himself had settled in his draft. Readers who want to hear the mature Marx constructing an argument at full stretch, with an editor merely holding the ladder, are directed by the comparison literature to this part, and its presence in the volume does the same work for the defence that the credit part does for the critics. A single volume containing both is the final refutation of every one word verdict, and the reason the article’s ledger gives the third volume not a verdict but an instruction: cite it as an edition, passage by passage, and never as an unmediated whole.

How an edition is read against a manuscript

Because the findings this article relies on are products of a method, the method deserves a section of its own; readers who understand how a comparison is performed are in a position to judge how much weight its results can bear. The critical edition prints, for each work, the text in its documentary state together with an apparatus recording variants, and, for the volumes Engels edited, separate editorial reports in which the modern editors reconstruct, from the surviving papers, what the nineteenth century editor had before him and what he did with it. The comparison scholar’s procedure is then laborious and unglamorous. A passage of the printed edition is located in the manuscript. Its wording is collated. Its neighbours are checked, because a passage’s meaning in an edition often depends on what the editor placed before and after it. The manuscript’s own order, recoverable from its pagination, its catchwords and its internal references, is reconstructed, and the edition’s order is laid over it. Headings are inventoried and classified: authorial, editorial, or uncertain. Cuts are measured by the simple expedient of counting what the manuscript contains that the edition omits, and the omissions are characterised, whether they remove repetition, discard superseded calculations, or set aside material whose absence changes the balance of an argument. Vollgraf and Jungnickel’s studies of the third volume proceeded in exactly this way, section by section, and Heinrich’s work on the edition’s history supplied the interpretive frame within which the measurements acquired significance.

Three properties of the method explain why its factual results command the assent they do. The first is checkability: because the edition prints the manuscript, any claim about an intervention can be verified by any reader willing to do the collation, and several celebrated claims have been revised by exactly that process. The second is granularity: the method works passage by passage, so its results resist summary into slogans, and scholars who agree on a hundred particular interventions may still disagree, honourably, about the hundred and first. The third is asymmetry of surprise. The comparison was available to be run by scholars of every allegiance, including those devoted to Engels’s memory, and its central findings, the reordering, the supplied divisions, the fifth part’s construction, were confirmed across those allegiances rather than imposed by one school. Where the method reaches its limit, honest practitioners say so. It can establish what the editor did. It cannot establish what the author would have done with the same papers given twenty more years, and it cannot weigh, by itself, the philosophical question of how much arrangement a text can absorb before its authorship is shared. Those limits are not defects. They are the boundary between documentary scholarship and judgment, and the verdicts section that follows operates, self consciously, on the judgment side of the line.

The comparison enterprise has also, in its maturity, turned its instruments on the modern editors themselves, and the reflexivity is healthy. The critical edition’s own choices, which manuscript readings to privilege, how to present Marx’s deletions and insertions, how to date undated sheets, are documented in its apparatus and are themselves debatable, and scholars do debate them. Roth’s studies of the third volume include examination of how Engels himself worked from a specially prepared transcript, correcting and annotating the copy rather than the fragile original at several stages, a detail that adds one more physical layer between the printed book and the author’s hand and explains a class of small divergences that pure collation of print against manuscript had left mysterious. The lesson generalises pleasantly. Every edition in the chain, authorial, executor’s, twentieth century, critical, is an argument made by identifiable people under identifiable constraints, and the mature field treats its own instruments with the same suspicion it applies to Engels. Suspicion of that even handed kind is not corrosive. It is what distinguishes a documentary discipline from a cult of either the author or the editor, and it is the spirit in which the verdicts should be read when the article turns to them.

What Engels disclosed in his prefaces

Any fair hearing of the authorship question must begin the defence where the defence begins: with the editor’s own words about his work, which are more candid than the scandal version of the story allows. Engels wrote substantial prefaces to both posthumous volumes, and a reader who studies them will find the method described in general terms with a plainness unusual among literary executors. The preface to the second volume catalogues the manuscripts, dates them, explains the principle of selection among them, and states that the editor had confined himself to choosing the most finished versions and correcting what were plainly slips. The preface to the third volume goes further in the direction of confession. It describes the state of the principal manuscript, acknowledges that whole parts existed only as drafts and notes, distinguishes the parts that cost little labour from the part, the fifth, that cost the most, and warns readers that Marx had intended a thorough reworking of certain sections which death prevented. Engels wrote that his aim had been to produce a coherent and connected book from the materials, and that where he had altered or supplied matter he had kept such intervention within what he judged the author’s sense to require. Nothing in these prefaces is a lie. The question the manuscript comparison raised is not whether Engels deceived his readers about there being an edition. It is whether his general disclosure conveyed the scale of the particulars.

It did not, and the gap between the disclosure and the particulars deserves precise description rather than moral theatre. A modern critical edition accompanies an edited text with an apparatus: a record, passage by passage, of what the editor found, chose, moved and supplied, so that any reader can reconstruct the manuscript state at any point. Engels provided nothing of the kind, and did not claim to. His prefaces operate at the level of method, not of instances. They tell the reader that selection occurred without marking the selections, that arrangement occurred without mapping the rearrangements, that one part was largely constructed without printing the dossier from which it was constructed. Within the conventions of nineteenth century executorship, this was unremarkable; editors of Goethe and of Hegel had done as much and less. Measured against the needs of later scholarship, which would build doctrines on single chapters and quote the editor’s chapter titles as the author’s doctrine, it was a consequential silence. Both measurements are true, and the article’s verdict section will return to their tension. What can be said without controversy is that Engels disclosed the existence and the general character of his interventions while leaving their extent to be discovered by a later century’s editorial labour.

What did Engels say in his prefaces about his own editorial work?

He described the manuscripts and their dates, explained that he had selected among drafts and corrected slips, admitted that the fifth part of the third volume required extensive construction from fragments, and stated his aim of a coherent book. He gave no passage by passage record of his changes.

The disclosure question has a second layer, involving what Engels said about attribution more generally, and here his record is striking for its generosity in the other direction. The 1883 statement assigning the Manifesto’s fundamental proposition solely to Marx has already been noticed. It was not an isolated act. In correspondence and in print, Engels repeatedly subordinated his own theoretical standing to his partner’s, describing himself, in a phrase that has stuck, as playing second fiddle, and insisting that Marx was the genius of the partnership while he was at best a talent. Scholars of a suspicious temper have read this self effacement as a strategy, the construction of a single authoritative doctrine under a single name, with Engels’s own systematising quietly absorbed into it. A plainer reading is available, and the documentary record of forty years of friendship supports it: the subordination was sincere, it was also convenient for a movement that needed a founder, and its long run effect was to hide from view precisely the editorial authorship that the manuscripts later revealed. The man who insisted he was only the second fiddle had, by that insistence, made it harder for posterity to hear his own bow in the posthumous volumes. To study Engels the thinker in his own right, rather than as a hand inside other texts, the companion profile of friedrich engels intellectual profile is the place to go; the concern here is narrower, the reliability of the name Marx as it appears on texts the second fiddle prepared.

A final disclosure belongs to the record, the most poignant of them. In his last years Engels was preparing supplementary material of his own for the third volume, essays on the law of value and on the stock exchange, published under his own name in a journal rather than inserted into the book. The restraint is eloquent. Where he wished to add his own arguments to his friend’s volume, he published them separately, signed, after the volume, where no reader could mistake them for the manuscript’s content. An editor bent on smuggling his doctrine into a canonical text had easier instruments to hand, and did not use them. Whatever the verdict on the scale of his construction, the signed supplements show a man who understood the difference between his voice and Marx’s, and who, at the boundary where addition would have become impersonation, stopped.

One more disclosure belongs in this section, because it shows the editor working under a stricter discipline and thriving under it. After Marx’s death, Engels prepared the third and fourth German editions of the first volume, the book Marx had published himself. Here the author’s intentions were documented in unusual detail: Marx had left notes indicating changes he wished to carry over from the French edition, in which he had revised extensively, into the German text. Engels followed the notes, reported in his preface exactly what he had incorporated and where he had stopped, and declined to make further alterations of his own, remarking that he had no right to touch the text beyond the author’s instructions. The result is as close to a controlled experiment as editorial history provides. Given a finished text and documented authorial instructions, the same editor whose third volume construction fills this article behaved with something like modern scholarly restraint, and described his restraint in print. The variable that changed between the first volume’s editions and the third volume’s construction was not the man. It was the state of the paper. That single comparison does more to acquit Engels of the charge of systematic ambition than any character testimony, and it does so while leaving the construction itself, and its consequences for readers, exactly where the manuscripts put them.

Fate, with a sense of symmetry, arranged for Engels to experience editorial intervention from the receiving end before his death. In 1895, months before he died, he wrote an introduction to a new edition of Marx’s Class Struggles in France, reflecting on how the conditions of political struggle had changed since 1848, on the growth of the movement’s electoral strength, and on the diminished prospects of the old insurrectionary tactics in an age of modern armies and rebuilt cities. The party newspaper published extracts in a form that trimmed his tactical argument toward the movement’s preferred militancy, and Engels protested in letters to friends and editors that his meaning had been softened into its opposite, that he had been made to appear a worshipper of legality at any price. The full text reached print only decades later. The episode is the authorship question in miniature and in mirror image: a living author watching an institution edit him toward usability, and objecting, on the record, to precisely the kind of smoothing his own hand had performed on a larger scale and for kindred reasons. Nothing in the episode makes him a hypocrite; his own interventions were declared and his object’s were not. But it demonstrates that he understood, from the inside and in his final year, the principle on which the modern criticism of his editorship rests: that to rearrange and soften a writer’s argument is to change what the writer can be quoted as saying, and that writers, living or dead, have an interest in the difference. The executor of Capital would have recognised the edition attribution rule as a protection he had once wished for himself.

The falling rate of profit as the test case

Every general claim in this article can be tested against one stretch of text, and the discipline of the subject improves when the test is actually performed. The stretch is the third part of the third volume, the treatment of the tendency of the rate of profit to fall, which carries more argumentative weight in later Marxist economics than any other portion of the posthumous volumes. In the printed edition the material forms a stately sequence. A chapter states the law as such: as capitalist production develops, the mass of means of production set in motion by each worker grows, the organic composition of capital rises, and the rate of profit, the ratio of surplus value to total capital advanced, tends to decline even as the mass of profit grows. A chapter then assembles the counteracting influences: increasing intensity of exploitation, depression of wages, cheapening of the elements of constant capital, relative overpopulation, foreign trade, the increase of stock capital. A third chapter unfolds the internal contradictions of the law, its periodic eruptions in crisis. Generations of economists, friendly and hostile, have debated the doctrine in exactly this architecture, attacking or defending the law, weighing the countertendencies against it, and citing chapter and verse from the sequence as from a finished treatise.

The manuscript tells a less architectural story. In the draft of 1864 and 1865, Marx’s discussion of the profit rate moves associatively through the material, raising the tendency, qualifying it, pursuing the cheapening of constant capital back into the argument, returning to the tendency with the qualifications still in hand. The boundary between the law and its counteracting influences, so crisp in print, is porous in the draft, because the author had not yet decided, or had decided against, presenting the countertendencies as a separate systematic block standing outside the law’s own exposition. Engels, building his chapters, separated what the draft interwove. He gathered the counteracting influences into their own chapter and gave the sequence its tripartite form. Heinrich’s studies of the manuscript argue that this was no neutral tidying. Presenting the countertendencies as an external list transforms an open investigation, in which Marx may have been discovering that the tendency and its qualifications are inseparable aspects of one process, into a doctrine with a law on one side and exceptions on the other, a doctrine that invites the arithmetical question, which side wins, that the draft never poses so baldly. Other scholars resist the stronger form of this reading, noting that the draft does at points enumerate the counteracting factors and that Marx himself, in the first volume and in earlier manuscripts, was capable of presenting the tendency in lawlike terms. The dispute is real, and this article does not settle it. What the dispute establishes, beyond reasonable quarrel, is the test case’s lesson: here, at the most consequential point of the posthumous text, the printed architecture and the manuscript architecture differ, and arguments conducted against the printed one are, in part, arguments with the editor’s construction.

Where does the editing of the profit rate material change the argument itself?

At the boundary between the law and its counteracting influences. The draft interweaves tendency and qualification in one investigation, while the edition separates them into successive chapters, giving the doctrine a firmer, more lawlike shape that shapes how critics and defenders frame the debate.

This is the place to be exact about what does not follow. It does not follow that the falling rate of profit is Engels’s invention; the tendency, its mechanism in the rising composition of capital, and the principal counteracting factors are all in the manuscript, in Marx’s hand, argued in Marx’s terms. It does not follow that the doctrine is refuted by its editorial history; a claim’s truth is independent of the tidiness of its first presentation, and the century of debate the doctrine provoked must be settled on economic grounds. The full debate, its mathematical reformulations, its critics from within and outside the tradition, belongs to the companion treatment of the falling rate of profit debate, and the present article deliberately borrows its findings rather than re-arguing them. What follows from the manuscript comparison is a point about attribution, and it is not small. When a critic writes that Marx’s law of the falling rate of profit fails to establish its conclusion, or a defender writes that Marx answered the objection in advance, the subject of those sentences is a composite: Marx’s analysis as arranged by Engels into the form in which the answer or the failure appears. Scholarship that ignores the composite character will sometimes refute the arrangement while believing it has refuted the analysis, and sometimes defend the analysis by pointing to strengths the arrangement supplied. The edition attribution rule exists to prevent exactly those two errors, and this test case is its strongest exhibit.

The test case also illustrates the honest limits of the whole inquiry. Suppose the rule is followed perfectly, and every citation of the profit rate chapters carries its flag: Engels’s edition, manuscript state noted, arrangement distinguished from draft. The underlying question remains whether Marx, completing the book in the 1870s or 1880s, would have kept the tripartite architecture, discarded it, or replaced the whole treatment in the light of his later studies of credit, of joint stock companies and of the mathematics he was teaching himself in his final years. No document answers that question, because the rewriting never happened. Attribution scholarship can separate the hands in the text that exists. It cannot recover the text that does not. Marking that boundary clearly, rather than filling it with confident speculation about what Marx would have written, is part of the discipline this article is recommending, and it applies to defenders of Engels as much as to his critics: the editor cannot be acquitted by appeal to an unwritten final version any more than he can be convicted by one.

A closer look at the counteracting influences shows why scholars treat the separation as argumentative and not merely architectural. In the manuscript’s flow, several of the counterweights are not external accidents befalling a tendency from outside; they are products of the same process that generates the tendency. The cheapening of the elements of constant capital is the clearest instance: the rise in productivity that increases the mass of machinery and materials per worker simultaneously reduces the value of each unit of that machinery and those materials, so the force pressing the profit rate down and the force pressing it up are one development read from two sides. An exposition that gathers the counterweights into a separate chapter, under a heading announcing them as influences that counteract the law, invites the reader to model the outcome as a contest between independent factors, to be settled by asking which list is stronger. An exposition that keeps them inside the investigation invites a different model, in which the tendency exists only in and through its qualifications. Marx’s draft, on the comparison scholars’ reading, leans toward the second model without ever stating the choice. Engels’s chapters institutionalise the first. One need not decide which model is better economics to see that a choice was made, that it was made by the editor, and that generations of textbook presentations, with the law on one page and its exceptions on the next, descend from the editorial choice rather than from any sentence Marx wrote.

The same pattern, at lower intensity, runs through the profit rate manuscript’s treatment of crisis. In the draft, the discussion of how the law’s internal contradictions unfold moves between the tendency, the credit system and periodic breakdown without firm borders, reflecting a research stage at which crisis had not yet been assigned its systematic place. The edition’s third chapter of the sequence gives the unfolding a defined location and a developmental title, which later readers naturally cite as Marx’s theory of crisis in nuce. Crisis theory across the tradition draws on far more than this chapter, and the companion literature treats the whole terrain. But the chapter’s very citability, its existence as a bounded unit with a heading, is an editorial product. Scholars who have compared the texts speak of a difference between a Marx who was assembling materials for a theory of crisis and a volume that appears to deliver one. That formulation, applied with suitable care, may be the fairest single sentence available on the whole third volume question, and it explains the peculiar tone of the best recent literature on the subject: gratitude for the book, unease about its finish, and a determination to cite it, henceforth, as what it demonstrably is.

Textbook transmission deserves its own paragraph in the test case, because the edition’s architecture reached most readers at third hand, through summaries that preserved the shape while discarding everything else. The standard textbook presentation of Marxian crisis economics, across much of the twentieth century, opens with the law of the falling rate of profit, proceeds to its counteracting tendencies as a checklist, and closes with the system’s inner contradictions erupting in crisis. That is, to the paragraph, the sequence of Engels’s three chapters. A student could absorb the entire structure, reproduce it in examinations and build arguments on it, without once encountering a sentence of the manuscript, in which the tendency surfaces inside a thicket of qualifications and the checklist does not exist as a list. The point is not that the textbooks were wrong to summarise; summary must choose a shape, and the edition’s shape was the only one on offer. The point is that a civilisation’s worth of teaching can inherit an editor’s arrangement as completely as it inherits an author’s argument, with no marker in the transmission to distinguish the two bequests. Markers are what citation practice supplies. A textbook that named its source texts honestly would cost its author a footnote, and would save its readers the recurring shock of discovering, in the critical literature, that the ground plan of a doctrine was drawn by the doctrine’s first editor.

Three verdicts on the editor

Scholarly judgment on Engels’s editorial work has sorted itself into three broad verdicts, and the reader is best served by meeting each at its strongest rather than by a composite that flatters the article’s own conclusion. The verdicts agree on the documentary facts established since the manuscript of the third volume became available. They disagree on what those facts amount to, which is another way of saying that the authorship question has matured from a dispute about evidence into a dispute about the meaning of authorship itself.

The first verdict is the faithful executor. On this account, Engels performed the office of a literary executor to the highest standard his century knew. He took unpublishable paper and gave the world the books, promptly for the second volume, at the cost of his final decade for the third. He disclosed his method in his prefaces, signed his own additions separately when he made them, and subordinated his name throughout to his friend’s. The interventions the manuscripts reveal are, on this reading, the necessary minimum of construction any editor would have supplied, and their scale looks large only when counted in an apparatus rather than weighed against the hundreds of pages reproduced with fidelity. The strongest evidence for this verdict is comparative and biographical: the treatment other nineteenth century executors gave their charges, the completeness of Engels’s preservation of the manuscripts themselves, without which no comparison would be possible, and the second volume’s demonstrated restraint where the materials were good. A destroyer of evidence does not spend his last strength preserving the drafts by which his own work will one day be audited.

The second verdict is the constructive editor. On this account, associated with the scholars who performed the manuscript comparisons, Engels did substantially more than the executor’s minimum, and the more that he did has direction. His constructions consistently strengthen the systematic, finished and lawlike appearance of Marx’s economics: divisions sharpen, hesitations are smoothed, an exploratory draft emerges as a doctrine with parts and chapters. The strongest evidence is the pattern itself, visible wherever the comparison is run, in the profit rate chapters, in the fifth part’s architecture, in the very decision to issue the 1860s draft as the third volume of a finished work rather than as the draft it was. The constructive editor verdict does not allege bad faith. It alleges temperament, the expositor’s temperament described at the outset, applied for a decade to another man’s unfinished system, and it invites readers to discount the finish accordingly wherever an argument leans on it.

The third verdict, the least common and the most unsettling, is co author in effect. On this account, at least for the third volume as the world has received it, the distinction between editing and authorship collapses under the weight of the construction, and the honest attribution of the received text is a dual one, however the title page reads. Its strongest evidence is the counterfactual test: remove Engels’s arrangement, his chapters, his selection and his connective work, and the influential object, the book that movements studied and economists attacked, does not exist; what exists is a manuscript that specialists can read. Influence ran through the constructed text, so construction shares the authorship of the influence. The verdict’s weakness is equally clear. It stretches the word author past the point where it distinguishes the man who conceived and drafted the analyses from the man who built them into a book, and it would, applied consistently, make co authors of the editors of every posthumous classic. Most scholars who find the manuscript evidence damning to the simple picture stop at the second verdict for this reason.

Each verdict also implies a different instruction to the reader, which is worth making explicit because the verdicts are often debated as though they were only judgments of a dead man’s desert. If the faithful executor verdict is right in its strong form, the reader’s duty is discharged by gratitude and a general awareness that editions mediate; the text can be argued from as it stands. If the constructive editor verdict is right, the duty sharpens: arguments that lean on the system’s finish, on the separation of law and counterweight, on the fifth part’s coherence, carry a surcharge of caution, and the surcharge is payable in the citation. If the co author verdict is right, the duty changes in kind: the third volume should be catalogued, taught and cited as a work of two hands, and the long scholarly habit of writing Marx’s third volume without qualification should end. The present article’s rule was designed to be acceptable to holders of the first two verdicts and to move practice most of the way toward the third verdict’s demand without requiring agreement with its premise. That design is deliberate. Attribution practice should not have to wait for metaphysical agreement about what authorship is; it should encode, in the footnote, exactly what the documents show, and let the verdicts fight over the encoding’s meaning.

The verdicts also distribute differently across the corpus, and forcing one verdict to cover every text is a reliable sign of a weak treatment. The executor verdict fits the second volume so well that even the comparison’s authors qualify their findings there with respect. The constructive verdict fits the third volume’s middle parts and the profit rate sequence. The co author intuition stirs only at the extreme, in the fifth part, where the editor’s selection among fragments determined not merely the order of an argument but which materials would count as the argument at all. A map with three regions is less satisfying than a verdict with one word, and the literature’s polemics reflect the dissatisfaction, since polemics prefer a single Engels, faithful or faithless, to a gradient. The gradient is nevertheless what the manuscripts show, and the ledger’s citation column, which varies its formula row by row, is the article’s attempt to print the gradient where readers can use it: at the point of citation, where a verdict must in practice be chosen every time a text is named.

The strongest objection to the whole line of inquiry must now be stated as its best defenders state it, because the article’s rule is worth having only if it survives the objection. The objection runs: Engels worked from unfinished manuscripts with the explicit purpose of producing a usable book, he said openly that he had made editorial decisions, and any editor faced with the same material would have had to impose structure, so describing his work as intervention imports a modern editorial standard into a century that did not hold it. Judged by the standards of his own time, there is nothing to explain and nothing to forgive. The answer begins by granting the objection’s history. Executor conventions of the period were loose, and Engels was, by those conventions, exemplary. It then distinguishes two kinds of structure an editor can impose. Necessary structuring makes a text legible: it orders fragments, resolves obvious disorder, marks the joints the author left implicit. Doctrinal structuring makes a text systematic: it decides that a tendency shall be presented as a law with external counterweights, that scattered notes shall constitute a part with a thesis, that a research programme shall wear the dress of a completed theory. The first kind any editor must supply and no rule of attribution begrudges. The second kind changes what later arguments take the author to have claimed, and it is there, and only there, that the attribution rule bites. The objection is also answered, finally, by the purpose of the rule itself. The rule is not a retroactive judgment on Engels, whom it leaves, on the evidence, an honest and laborious executor. It is a present tense instruction to readers, who are free to use modern standards for their own citations even while judging the dead by the standards of their day. Conceding that the line between the two kinds of structuring is a matter of judgment, as this article does, is not a weakness in the rule. It is the rule’s operating condition, since judgment, exercised case by case over documents, is what attribution scholarship is.

The three verdicts have a history of their own, and tracing it guards against mistaking the present distribution of opinion for a permanent settlement. While the manuscripts remained inaccessible, the faithful executor verdict held the field by default, not because it had been demonstrated but because the only evidence was Engels’s prefaces, which no one had reason to doubt and no means to test. The presumption was reasonable and it lasted a century. Its collapse, when the third volume’s manuscript appeared in the early 1990s, was therefore not the exposure of a century’s gullibility; it was the predictable consequence of evidence arriving in a field that had organised its beliefs around the evidence it had. The constructive editor verdict is the child of that arrival, and it commands the documentary centre of the field at present. The co author verdict remains a minority position, held most often by scholars working on the theory rather than the documents, who feel the counterfactual argument’s force. A reader orienting in the literature should also distinguish the authorship findings from a neighbouring thesis with which they are often confused: the claim, associated with Terrell Carver and others, that Engels’s own independent writings systematised and in places altered the doctrine that goes under Marx’s name. That thesis concerns works Engels signed himself, and its merits belong to the study of his independent contribution. The manuscript findings are narrower and stronger: they concern what happened inside books signed by Marx. Keeping the two claims separate, as the careful literature does, prevents the documentary findings from being either inflated into a general theory of the partnership or dismissed as an episode of it.

Comparison with the century’s other great posthumous constructions calibrates the verdicts better than any abstract standard, because it shows where the spectrum of editorial conduct actually runs. At one pole stands the case of Nietzsche’s Nachlass, where the author’s sister and her collaborators assembled a book, The Will to Power, from notebook fragments, presented it as the philosopher’s crowning work, and managed the archive in ways that later scholarship, armed with the manuscripts, judged to be falsification in the plain sense: arrangements designed to manufacture a doctrine and an image. At the spectrum’s middle stands the editing of Max Weber’s writings by Marianne Weber and others, where arrangement decisions shaped a generation’s reading of an unfinished masterwork and where scholars, again armed with a critical edition, have spent decades disentangling construction from text without impugning the editors’ honour. Engels belongs on the Weber side of the spectrum, and the distance between his construction and the Nietzsche case measures the distance between intervention and falsification that popular accounts keep collapsing. His manuscripts were preserved entire, his method was declared in general terms, his own additions were signed, and his arrangement answered to a genuine, if contestable, reading of the author’s design. The scholars who press the constructive editor verdict are not moving him toward the falsification pole; they are mapping, with new instruments, a position in the spectrum’s honest half that his first century of readers could not see. Popular writing that borrows their findings to move him across the spectrum misreads them, and the misreading is one this article exists to correct.

The edition attribution rule in practice

The edition attribution rule can now be stated in its working form, with the cases behind it. Any claim about what Marx argued in the second or third volumes of Capital should specify whether it rests on Engels’s edition or on the manuscript, because for several important arguments the two differ in structure and emphasis. The rule has three operations, and each is cheap to perform. Cite the edition: a reference to the third volume should name the text actually consulted, Engels’s edition in its standard translation, or the manuscript in the critical edition, and not float above both as though they were interchangeable. Name the editor’s hand where the edition shows it: where a passage’s position, heading or formulation is Engels’s construction, the prose that uses the passage should say so once, plainly, rather than letting the quotation marks imply unmediated authorship. Flag the constructed passages in quotation: never quote an Engels constructed passage, a chapter heading, a transition, a passage from the fifth part’s arranged dossier, as Marx’s unmediated words without the flag. Performed together, the three operations take a sentence or two per citation, and they convert an invisible vulnerability in an argument into visible scholarly hygiene.

Working habits make the rule easier to keep than it sounds. A researcher assembling notes on the posthumous volumes does well to keep the edition and the manuscript state side by side in the notes themselves, recording for each extract whether the passage stands in the draft, and in what form, which is exactly the kind of parallel note taking that the VaultBook study notes tool is built to hold, two texts and their divergences kept in a single private file rather than trusted to memory. Where the work involves handling the documents themselves, comparing edition against manuscript across long stretches, a document tool for working with source texts and editions earns its place in the workflow: the ReportMedic document explorer serves here as the article’s recommended instrument for opening, searching and moving between the files of an edition while the comparison proceeds. Neither tool settles an attribution. Both serve the unglamorous truth of the subject, which is that attribution errors are usually filing errors first: a quotation detached from its edition, a heading mistaken for a text, a note that forgot to record which object it came from.

The rule extends naturally beyond the two volumes that motivated it, and the ledger shows where. For the Manifesto, the rule asks the citer to remember the scaffolding: the programme’s framework descends from Engels’s drafts even where the sentences are Marx’s. For The German Ideology, it asks for chapter level attribution where the editors have separated the hands, and for honesty about the manuscript’s unfinished state where a famous passage is quoted as doctrine. For the journalism, it asks for article level attribution from the correspondence, and treats signature as no evidence at all. For the first volume of Capital, it is satisfied by naming the edition, since the authorial text there is as stable as this subject affords. A full treatment of citation mechanics across the whole corpus, including the collected works and their tangled publishing history, is given in the companion guide to citing marx and engels, which this article’s rule presupposes and extends. The rule’s spirit is uniform even where its strictness varies: the reader of your citation should be able to tell, from the citation itself, how many editorial hands stand between them and the author’s desk.

What the rule refuses is as important as what it requires. It refuses to rank the texts into genuine and spurious, because the edition is not a forgery and the manuscript is not a scripture. It refuses to allocate percentages of authorship, because the contributions differ in kind and the arithmetic would be theatre. It refuses, above all, to let the attribution question be used as a shortcut in arguments where it does no work. A passage’s editorial history neither validates nor refutes the economics in it. The rule disciplines claims about who argued what, and leaves claims about what is true to the arguments themselves, where they belong.

A worked example shows the rule’s cost and yield better than another statement of it. Consider a sentence of the kind that appears in essays and articles every season: Marx argues in the third volume of Capital that the rate of profit tends to fall as the organic composition of capital rises, before counteracting influences, which he lists, come into play. Under the rule, the sentence acquires a clause and loses an implication, emerging as: Marx’s manuscript for the third volume argues that the rate of profit tends to fall as the organic composition of capital rises; the counteracting influences are gathered as a separate chapter in Engels’s edition, and are interwoven with the tendency in the draft. The second version is longer by one clause. It is also a different claim, narrower, defensible under cross examination, and immune to the objection that felled the first, which attributes an architecture to an author who did not build it. Multiply the example across a bibliography and the rule’s real yield appears: a literature in which the words Marx argues are a reliable guide to where the argument stands in the documentary record, and in which the editor’s substantial, honourable and consequential labour is neither hidden nor mistaken for the author’s hand.

One more mediation stands between most readers of this article and the texts it discusses, and completeness requires naming it: translation. The standard English versions of the second and third volumes descend from Engels’s German editions, the third volume’s translation having been made from the text he constructed, so that an English quotation from the profit rate chapters carries a double mediation, the editor’s arrangement and the translator’s choices, beneath its quotation marks. The German reader of the standard edition stands one step closer and no more. Only the critical edition removes the first mediation, and it does so at the price of presenting a draft rather than a book, with an apparatus in German that makes its own demands. The attribution rule, applied consistently, therefore governs translation practice too: a scholar quoting the third volume in English on a point where arrangement matters should check the passage against the manuscript or say that the check was not made. Honesty about an unmade check is not a confession of failure. It is the ordinary currency of a field that has learned, expensively, what unexamined texts can cost.

The papers themselves completed a journey in the twentieth century that rounds off the archive’s story and explains the long delay between the questions and their answers. After Engels’s death the literary remains passed through the keeping of Marx’s daughters and the German party’s archive in Berlin, where Eduard Bernstein and others who had known the executor guarded them as relics and instruments at once. Copies and originals diverged, some material travelling to Moscow in the era of the first collected edition, whose heroic editor, David Riazanov, began publishing the manuscripts in their documentary state before his arrest in 1931 ended the enterprise and nearly ended the archive’s accessibility for a generation. The bulk of the papers found their durable home in Amsterdam, in the International Institute of Social History, whose purchase of the collection in the 1930s is one of the great rescues in scholarly history. There the manuscripts waited out the century’s storms until the second historical critical edition, a German led international project begun in the 1970s, could do what Riazanov had attempted: print everything, in order, with apparatus. The authorship question as this article has treated it is a direct dividend of that rescue and that edition. It could not have been asked in its present form in 1950, and its having become askable is a reminder that scholarly progress is sometimes a matter of paper surviving in the right cellar.

Anthologies and quotation collections are where the rule does its most neglected work. The famous sentences of the corpus circulate in excerpted form, in readers for courses, in dictionaries of quotations, in the epigraphs of books on every subject, and each act of excerpting is an attribution decision made invisibly. An anthology that prints the profit rate chapters’ headings as Marx’s own summary of his doctrine has made a finding about authorship without knowing it. A quotation dictionary that assigns a sentence of the fifth part’s constructed prose to Marx, unqualified, has done the same. Editors of such collections cannot be asked to collate manuscripts, and the rule does not ask it. It asks for the cheap signals: the edition named in the source line, the constructed volumes marked as edited by Engels wherever the excerpts are drawn from them, the chapter headings of the third volume credited to the edition rather than the author. These are the courtesies that scholarly publishing already extends to other posthumous classics, whose readers expect to be told when they are reading a text an executor assembled. Extending the same courtesy across this corpus requires no new scholarship at all, only the decision to treat its texts as documents rather than as scripture, a decision whose lateness, in this one field, is the historical accident this article has been describing.

The stakes of the whole inquiry can be stated without exaggeration by following the constructed text into the hands of its most consequential readers. The generation that built the mass workers’ movements of the early twentieth century learned its Marxism substantially from the posthumous volumes as Engels issued them, and the revolutionary leaderships of the Russian movement, Lenin among them, studied the third volume as the completed keystone of the system, citing its chapters in polemics whose outcome shaped a century. None of this makes their readings worthless; intelligent readers can learn from an edited text, and did. But it means the doctrine that entered history under Marx’s name was, in its economic architecture, partly the doctrine as arranged in a London study in the early 1890s, and that the movements’ confidence in the system’s finished character rested on a finish the editor had supplied. Historians of the movement ignore the attribution findings at the price of misunderstanding their subject’s own sources. The priesthood of a text should always be told when the text’s binding was done by another hand than the author’s, not to diminish the book, but to understand the faith that was placed in it.

Digital circulation has given the attribution problem a new scale and a new carelessness, and a treatment of citation practice that ignored it would be advising a vanished world. The texts of the corpus now travel chiefly as files: scanned editions, transcribed archives, searchable collections in which a passage can be summoned in seconds, stripped of its edition’s title page, its preface and its provenance. The third volume circulates online in a translation made in the first decade of the twentieth century from Engels’s German text, so that the global reader’s Marx arrives through the executor’s construction and a Edwardian translator’s English, with neither mediation marked at the point of use. Search compounds the invisibility, since a search result displays a sentence, not a bibliography. None of this is an argument against the digital corpus, which has democratised access to the manuscripts’ transcriptions as well as to the editions, and which puts the critical edition’s findings within reach of readers the print apparatus never served. It is an argument for carrying the attribution rule into the file itself: edition statements in the headers of online texts, constructed passages flagged in the transcription, the manuscript state linked at the chapter head. The technology that dissolved the title page can restore the information the title page carried, in better places. Until it does, the reader’s own discipline, of the kind this article’s rule codifies, is the only apparatus most quotations of the posthumous volumes will ever receive.

Even the world’s libraries encode the question in their quiet way, and a researcher browsing a catalogue is reading attribution decisions whether they know it or not. The third volume is catalogued under Marx as author, with Engels recorded in the editor’s role, a statement of the orthodox position embedded in the infrastructure of scholarship itself. The Manifesto is catalogued under both names as joint authors, which the documents, as this article has shown, support only with the scaffolding qualification the catalogue record has no room to express. The German Ideology’s records vary by edition, some naming the twentieth century editors whose arrangement the volume embodies, most passing over the arrangement in silence. None of this is a reproach to cataloguers, whose rules require a main entry and reward decisiveness. It is a reminder that attribution is not a scholarly luxury practised in articles alone; it is load bearing in the systems by which texts are found, shelved and cited, and errors or simplifications at that level propagate silently into thousands of footnotes. The critical edition’s volumes, by contrast, catalogue themselves with fastidious precision, distinguishing author, editor, manuscript and construction in their very titles. A reader who has learned to notice such things has acquired, almost by osmosis, the habit the rule prescribes: before asking what a text says, ask what it is, whose hands made it, and which of its features belong to which hand. The habit survives every particular controversy in this article and applies to every posthumous classic on the shelf, which is why an inquiry that began with two names on a pamphlet ends as a general education in reading.

The editors edited in their turn

The chain of mediation did not end with Engels, and its continuation holds a final irony that the article would be incomplete without. Engels appointed as his own literary executors men of the next socialist generation, and the edition of the Marx and Engels correspondence that August Bebel and Eduard Bernstein published in 1913, the very edition whose revelations reassigned the Tribune journalism, was itself an edited object in the strong sense. The executors cut passages they judged imprudent: sharp judgments of contemporaries, personal matter, reflections that sat awkwardly with the movement’s mature pieties. The omissions were discovered as fuller texts became available through the Moscow edition and the later collected works, and the restored passages are now part of the standard record. Nobody familiar with the constraints of 1913, with a mass party’s inheritance to protect and living persons to consider, finds the cuts inexplicable. But the episode completes the article’s demonstration. Editing toward usability is not a vice of one executor or one movement; it is the gravitational pull on everyone who has ever held famous papers, and the only known counterweight is the survival of the originals together with a scholarly culture willing to print them whole. Engels preserved Marx’s originals, and thereby armed his own critics. His executors, in cutting his letters, repeated his century’s habits rather than his best example. The difference between the two conducts, preservation with intervention against intervention with suppression, is the difference on which the whole ethics of this subject turns, and it leaves Engels, on the final accounting, on the right side of the line that matters most.

The correspondence episode also refines the popular notion of a documentary golden age in which the truth simply waited in the archive. The archive’s contents reached the public through editions, and the editions had editors, and the editors had situations. The 1913 correspondence was a monument and a filter at once; the Soviet era editions added their own filters, selecting and framing under pressures that need no elaboration; the Western collected works, for all their fifty volume completeness, arranged the two men’s writings in an interleaved chronology that quietly teaches the brand’s lesson, one river, two names, indivisible. Even the critical edition, the hero of this article, is an edition, with principles that future scholarship will annotate as it annotated Engels’s. The mature conclusion is not despair about mediation, which would be a counsel of ignorance. It is the recognition that mediation is the permanent condition of textual knowledge, that it can be made visible, dated and checkable, and that the authorship question, pursued in that spirit, is simply the name this corpus gives to the universal discipline of knowing what one is reading.

What remains unknown

An honest article on attribution should end its investigation by inventorying its own ignorance, because the map of the unknown is one of the findings. Three territories remain dark, and they differ in kind. The first is accidental loss: not everything the partnership wrote survived, letters are missing from runs that are otherwise complete, some manuscripts Marx is known to have drafted have not been found, and every attribution resting on the surviving archive carries the silent caveat that the archive is a residue. The second territory is the same room problem. The Brussels years, when the two men worked in daily contact, generated few letters between them, and the division of labour in works of that period, The German Ideology above all, rests on hands and inference where later periods rest on correspondence. The third territory is in principle unknowable: the unwritten revisions. Marx’s late notebooks, on mathematics, on geology, on the Russian commune and its possible paths, show a mind still in motion on questions the posthumous volumes treat from manuscripts of the 1860s. What the finished Capital would have been, had its author lived to rebuild it from those notebooks, is not a question documents can answer, and the discipline of the subject consists partly in declining to answer it while continuing to describe, precisely, the books that exist.

The critical edition, meanwhile, continues, and its remaining volumes will extend the comparison method to materials this article has touched only in passing. The correct posture toward its findings is therefore per volume and per passage rather than global: established where the collation has been done, provisional where it has not, and always open to the revision that a new collation may bring. That posture, unspectacular as it sounds, is the mature form of the authorship question. The question began as a scandal in search of a villain and has ended as a bibliography in search of accuracy. The transformation is itself the strongest evidence that the subject was never really about two men’s reputations. It is about whether readers of powerful books will take the trouble to know what kind of object they are holding.

What future work can still settle should be distinguished from what it cannot, in closing the inventory. The collation of the second volume’s manuscripts against the edition is younger than the third volume’s, and particular passages there may yet change their attribution status as the editorial reports are absorbed and contested. The great survey manuscripts of the early 1860s, from which Kautsky built his rearranged fourth volume, are now available in their own order, and the systematic comparison between Marx’s design and Kautsky’s remains a live exercise with direct lessons for how arrangement arguments should be framed. The journalism, too, retains a residue of unattributed and misattributed pieces, in the provincial press especially, where the correspondence thins and stylometric argument must carry weight it carries badly. Each of these is a soluble problem, soluble by the ordinary means of the discipline, patience and paper. They stand in instructive contrast to the insoluble residue, the counterfactual Capital that died with its author in 1883. A field is healthy when it can tell its soluble problems from its insoluble ones and spends its passion on the first list. By that measure the authorship question, once a byword for speculative quarrelling, has become one of the healthiest corners of Marx scholarship, and the least dispensable, since every other corner builds on its texts.

Two incentives keep the popular version of this subject cruder than the scholarship, and naming them is a service to readers who will meet the crude version first. The devotional incentive, strongest inside traditions that treat the texts as foundations, prefers a single seamless author, because a doctrine’s authority is easier to invoke when its source speaks with one voice and no editorial static. The scandal incentive, strongest in journalism and polemic, prefers a hidden hand and a betrayed trust, because exposure is a better story than apparatus. Between the two incentives the documentary record, with its gradient of interventions and its honest executor, is squeezed from both sides, and each new scholarly finding is promptly conscripted: the faithful quote the restraint, the scandalised quote the construction, and the collation that established both sits unread. The corrective is not a midpoint between devotion and scandal, which would merely be a third distortion. It is the habit this article has tried to practise throughout: state the manuscript facts entire, distribute the verdicts at their strengths, and let the practical rule, which depends on none of the polemics, do the quiet work of improving the next generation of citations.

The findings of the whole inquiry can be gathered, finally, into the short form in which they deserve to travel. The Manifesto is Marx’s text built on Engels’s drafts, and its doctrine was jointly owned thereafter. The German Ideology is a genuinely collaborative manuscript whose hands can be separated sheet by sheet and whose ideas cannot always be separated with them. The journalism is a firm’s output, attributable article by article from the correspondence and never by signature. The first volume of Capital is the author’s own finished work. The second is his text in a restrained edition. The third is his analysis in a constructed one, the construction documented, consequential at identifiable points, and neither a falsification nor a secretarial tidying. Engels disclosed his method in general and withheld its particulars, by the standards of his century honourably, by the standards of citation now unavoidably. The remedy costs any writer one clause per citation: name the edition, name the editor’s hand where the edition shows it, and never present a constructed passage as unmediated. Scholarship that pays that small price can go on arguing about everything else in these books with a cleaner conscience and better aim, which is all the edition attribution rule has ever asked.

A last word belongs to proportion, since this article has spent many pages on distinctions that a hostile summary could mistake for antiquarian fuss. The distinctions matter in proportion to the uses made of the texts, and the uses made of these texts, in scholarship, in teaching and in the self understanding of political movements, are among the largest any books of their century sustain. Where a text is load bearing for an argument about how economies work or how history moves, the question of whose hand shaped its canonical form is not a footnote to the argument. It is part of the argument’s foundations, and foundations are the one place where fuss is a virtue. Readers who finish this article and thereafter write Engels’s edition, once, where they would have written Marx, unqualified, have not been recruited to a faction. They have been handed a tool, and the tool will keep working long after the verdicts in the case have been rearranged by the next generation of collation.

For students and exam candidates

Your practical exposure to this question is narrow and manageable: essays and examination answers go wrong on it in predictable ways, and the fixes are short enough to memorise. Never write that Engels wrote the second and third volumes of Capital; he edited and constructed them from Marx’s manuscripts, and the distinction is exactly what a good answer states. Never attribute the chapter headings of the third volume to Marx; the divisions are the editor’s. If your argument quotes the falling rate of profit material, one clause noting that the chapters rest on Engels’s arrangement of an 1860s draft will mark your script as better informed than most of the cohort. In citations, give the volume and the edition you used, and where a question invites you to assess the partnership, reach for the documented cases rather than generalities: the Manifesto’s drafts, the Tribune ghost writing, the prefaces’ disclosures. Examiners reward the student who knows that attribution in this corpus is a finding, not a formality, and they penalise the confident sentence that treats two names as one author. Keep the summary straight and the rest of the topic becomes much easier to handle under time pressure.

For teachers and seminar leaders

This topic teaches research method better than it teaches doctrine, and a seminar built around the documents will outperform a lecture built around the verdicts. Set the two texts of the profit rate material side by side, the manuscript passage and the edition’s chapter, and let students find the seams themselves; the exercise makes the abstract claim that arrangement is argument into something a class can point at. A second productive exercise is the byline problem: give students an unattributed Tribune article and the correspondence extracts, and ask them to argue the attribution and state their confidence, which teaches graded evidence better than any methodological sermon. Handle the scandal framing firmly when it appears, because students arrive with it from popular sources. The record shows an honest executor, a difficult inheritance and a genuine scholarly problem about significance, and modelling that three part judgment, facts agreed, meaning contested, editor judged by his century while citations are judged by ours, is the transferable skill the topic offers. Assessment should credit students who flag editions in their citations and should not demand a verdict on Engels where the evidence supports a range.

For researchers, historians, and journalists

The research frontier on this question is documentary, and it remains open in specified places. The critical edition’s publication of the manuscripts behind the second volume extended the comparison method to a text long assumed to be editorially innocent, and the detailed mapping of Engels’s choices among the second volume’s variants is still being absorbed into the secondary literature. For the third volume, the passage level findings are established in their factual content, while their significance for particular doctrines, crisis theory above all, is contested and likely to remain so; contributions that separate those two levels before arguing will age better than contributions that blur them. Historians should note the value of the case as a general instrument: the partnership is an unusually well documented instance of posthumous construction, with the editor’s prefaces, the author’s drafts and the correspondence all surviving, and it can anchor comparative work on executors from Hegel’s editors onward. Journalists face the sharpest practical version of the rule. A claim attributed to Marx in reporting should be traceable to a volume and an edition; where the claim comes from the third volume, the honest sentence is that Marx argued it in a draft his executor arranged, and where space forbids the qualification, the claim should be quoted from the first volume or dropped. Attribution in a headline is a promise. This corpus punishes promises made carelessly, and rewards, with unusual consistency, the writer who checked which text the famous sentence actually stands in.

Frequently Asked Questions

Q: Who actually wrote the Communist Manifesto?

Marx wrote the published text, working in Brussels through December 1847 and January 1848 after the Communist League commissioned a programme. Engels supplied the scaffolding: two earlier drafts in catechism form, the second being the Principles of Communism, which fixed much of the agenda and some of the material the pamphlet then recast in continuous prose. In the preface to the 1883 German edition Engels stated that the work’s fundamental proposition belonged solely to Marx, a judgment the surviving drafts support. The accurate description is therefore Marx’s text on Engels’s scaffolding, not a pamphlet composed jointly sentence by sentence.

Q: How much of Capital did Engels write?

None of its arguments and a great deal of its final form, with the proportions differing sharply by volume. The first volume is Marx’s own published and revised text. For the second volume, Engels selected and combined Marx’s drafts with a comparatively light hand. For the third, he constructed the book from a rough manuscript of the mid 1860s and later fragments, supplying the chapter divisions and headings, reordering material, and composing the arrangement of the part on credit. His signed supplementary essays were published separately. So the sentences and analyses are Marx’s, while the architecture of the posthumous volumes, above all the third, is substantially the editor’s work.

Q: Did Engels change Marx’s manuscripts?

Yes, substantially in the later volumes, as comparison with the manuscripts published in critical editions from the 1990s demonstrates. The documented changes include reordering material within and between sections, adding chapter divisions and headings absent from the drafts, cutting and transposing passages, smoothing exploratory writing into systematic exposition, and, in the fifth part of the third volume, selecting and arranging a dossier of fragments into a coherent part. The changes are editorial rather than authorial in intent: the categories and analyses remain Marx’s. But where the arrangement itself carries argumentative weight, as in the profit rate chapters, the changes reach into what readers take the doctrine to be.

Q: Which works are genuinely joint?

Genuinely joint composition, planned in advance and written in shares, is rarer in this partnership than the shared title pages suggest. The clearest case is The Holy Family of 1845, whose chapters were parcelled out between the two men, with Marx writing the larger share and the correspondence documenting the allocation. The German Ideology was jointly produced in the looser sense of two authors drafting, copying and correcting at the same table, with a contribution from Moses Hess, but it was never finished or published by them. Most other double named texts divide asymmetrically on inspection: the Manifesto is Marx’s text on Engels’s drafts, the posthumous volumes are Marx’s manuscripts in Engels’s editions, and the journalism is article by article attribution from the letters.

Q: Can you tell their handwriting apart in the manuscripts?

Yes, and the difference is one of the partnership’s scholarly fortunes. Marx’s hand is compressed, heavily corrected and notoriously hard to read; Engels wrote a clear, regular hand and habitually served as copyist, which is why much of the fair copy of The German Ideology stands in his writing with Marx’s additions in the margins. Editors of the critical edition use the hands, ink and paper to assign drafting sheet by sheet. The caution is interpretive rather than technical: a sheet in Engels’s hand may record joint discussion or dictation, so handwriting establishes who held the pen, and the correspondence and internal evidence must help establish whose formulation it was.

Q: Did Engels edit the third volume heavily?

By the standard of the second volume, very heavily, and the manuscript comparison leaves little room for a gentler description. The principal manuscript was a single exploratory draft from 1864 and 1865 without settled divisions, accompanied by later unintegrated notebooks. Engels imposed the seven part architecture, supplied most chapter headings, transposed and cut material, and in the fifth part, on interest bearing capital and credit, built a coherent sequence out of fragments written years apart, discarding a substantial quantity of the dossier. He acknowledged in his preface that this part had cost him the greatest labour. What the comparison adds to his acknowledgment is the map of the particulars, which his preface did not provide.

Q: Does the editing change any Marxist argument?

It changes the presentation of several and the effective content of at least one famous case. In the treatment of the falling rate of profit, the edition separates the law and its counteracting influences into successive chapters, where the manuscript interweaves them in a single investigation. That separation gives the doctrine a firmer, more lawlike shape, and much subsequent debate has been conducted against the firmer shape, by defenders and critics alike. Elsewhere the effects are smaller: emphasis, order, the suppression of hesitations. Nothing in the comparison shows an argument of Marx’s replaced by a different argument of Engels’s. The change is in how doctrines present themselves, which is why the remedy is a citation practice rather than a dismissal of the text.

Q: Was Engels a faithful editor?

The honest verdict is a divided one, and the division is instructive. By the standards of nineteenth century executorship he was exemplary: he preserved the manuscripts entire, disclosed his method in general terms, signed his own additions separately, and gave his final decade to the task. By the standard of a modern critical edition he falls short on one point, the absence of any passage level record of his interventions, which left readers unable to distinguish the author’s text from the editor’s construction for a century. Scholars who call him faithful usually mean the first standard; scholars who qualify the praise usually mean the second. The documentary facts beneath both judgments are agreed, and the dispute concerns which standard a dead editor should be held to.

Q: Did Marx write articles that Engels signed?

The documented traffic runs overwhelmingly the other way. Engels wrote a large share of the journalism that appeared under Marx’s name as the New York Daily Tribune’s correspondent, including the military articles and the series later known as Revolution and Counter-Revolution in Germany, whose authorship the published correspondence settled in 1913. Marx had the contract and the editorial relationships; Engels had the speed, the salary that freed his evenings and the military expertise. Isolated pieces moved in the reverse direction during their earlier newspaper ventures, and each man drafted material the other revised, but no comparable body of Engels signed journalism has been shown to be Marx’s composition. Signature, in this corpus, is evidence of almost nothing.

Q: Does it matter who wrote what?

It matters wherever an argument depends on attribution, which in this corpus is often. If a doctrine’s canonical form was given by an editor, criticism aimed at that form may miss the author’s analysis, and defence of the analysis may silently rely on the editor’s strengthening. It matters for intellectual history, because the two men had distinguishable habits of mind, the explorer and the expositor, and blending them produces a composite thinker who never existed. It matters least where it is most often invoked, as a weapon: a passage is not refuted by showing that Engels arranged it, nor vindicated by showing that Marx drafted it. Attribution settles whose claim a text is. The truth of the claim is a separate trial.

Q: How does the edition attribution rule change a footnote in practice?

It adds two pieces of information and forbids one implication. A citation of the third volume should name the text consulted, Engels’s edition in the translation used, or the manuscript in the critical edition, rather than citing the volume as though only one text existed. Where the passage relied on is one whose placement, heading or arrangement is the editor’s construction, the note or the surrounding sentence should say so once. What the footnote may no longer do is imply, by bare quotation, that a constructed passage, above all from the credit part or the profit rate chapters, reached the reader unmediated from Marx’s desk. In practice the change costs a clause, and examiners, editors and reviewers who know the corpus notice its presence.

Q: What should a researcher check before quoting a passage from the third volume of Capital?

Four checks, in order. First, whether the passage stands in Marx’s manuscript at all, which the critical edition allows the researcher to verify. Second, whether its wording there differs from the edition in ways that affect the use being made of it. Third, whether its position and heading are the author’s or the editor’s arrangement, since a passage quoted as the conclusion of an argument may stand mid investigation in the draft. Fourth, whether the passage comes from the fifth part, where construction was heaviest and the flag is most necessary. A passage that survives the four checks can be quoted with confidence and cited to the edition; one that fails any of them should be quoted with its manuscript state named alongside.

Q: Why do the manuscripts let scholars separate the two hands at all?

Because the partnership wrote on paper, in distinguishable hands, and the paper survived. Marx’s handwriting is idiosyncratic to the point of illegibility for most readers, while Engels’s is regular and clear, and the two appear in different inks, in draft and fair copy, in body text and marginal correction. Survival was not guaranteed: the manuscripts passed through executors, family keeping and party archives, and their preservation, in which Engels himself played the central part, is the precondition of the entire inquiry. Where sheets are lost or a passage survives only in a copyist’s transcript, attribution weakens to inference from style and correspondence, and the careful literature marks those cases as probable rather than proven.

Q: How much rewriting counts as editing rather than co-authorship?

There is no word threshold, because the distinction is functional rather than quantitative. Correction, selection among variants and the ordering of fragments remain editing, however many hours they take. Rewriting crosses toward co authorship at the point where the second hand begins to determine what the text argues: where it decides the architecture by which a doctrine will be understood, resolves the author’s open questions by arrangement, or gives a research programme the finished form in which alone it will influence readers. Measured by that test, Engels remains the editor of the second volume without strain, and stands at the boundary in the third, which is why the literature’s middle verdict, constructive editor, has proved more durable than either acquittal or promotion.

Q: What did Engels disclose about his own editorial interventions?

More than the scandal version of the story allows, and less than a modern reader needs. In the preface to the second volume he catalogued the manuscripts, dated them, and explained that he had selected the most finished versions and corrected slips. In the preface to the third he described the rough state of the principal draft, warned that Marx had intended reworkings that death prevented, and acknowledged that the fifth part had required the most extensive construction. He stated his aim as a coherent and connected book. What he did not provide was any record at the level of passages: which headings were his, which orderings were imposed, which material was set aside. The disclosure was honest at the level of method and silent at the level of instances, and that combination is exactly what the manuscript comparison, a century later, supplied.