ESHET, Porto (April 2006)
An Analysis of the General Theory Proofing Process
Toshiaki Hirai (Sophia Univ.)
INTRODUCTION
The main purpose of this paper is (i) to examine and to show objectively and exactly how Keynes went on revising the proof of the General Theory, and is (ii) to detect the main features of this proofing process. The period covers the summer of 1934 to February 1936.
If I would be asked what the key points are in this paper, my answer is simple:
There exists no paper which has examined the latter half of Keynes’s developmental process from the Treatise to the General Theory in spite of many studies examining the Keynesian Revolution from a point of history of economic thought have been so far made. The reason is that almost all the studies have been concentrated on the period up to 1933-34, maintaining that the Keynesian Revolution or the General Theory has been, in substance, completed in 1933 (albeit this is right), and that little effort has been spent for clarifying the 18 months leading up to the General Theory with which this paper deals.
What concerns us in this paper is what the major changes are in this proofing process? How and why did Keynes make changes, having made a drastic change from the Treatise to the road leading up to the General Theory up until 1933?
In order to answer this question, we need to investigate the following.
1) To endeavor to describe accurately based on the material concerned how Keynes made changes.
2) To consider the implication and significance which these changes brought about to Keynes’s theory.
The first task is descriptive and explanatory, and the second is theoretical in nature. Through the second task, we may shed some light on the nature and characteristics of the General Theory. (It may not be, for example, a complete book in terms of theory.)
I. AN OUTLINE OF THE PROOFING PROCESS
Keynes had constructed major components of the General Theory (the liquidity preference theory, the marginal efficiency of capital, the fundamental psychological law, the multiplier etc.) before “The General Theory”. In terms of style, however, in this manuscript we first find material incorporated into the General Theory: Ch. 12, “The State of Long-Term Expectation (or Confidence)” is the prototype for GT, Ch. 12, “The State of Long-Term Expectation”; Ch. 10, “The Propensity to Spend” is the precursor of GT, Ch. 8, “The Propensity to Consume: I. The Objective Factors”, and Ch. 9, “The Propensity to Consume: II. The Subjective Factors”.
Having completed almost all the text in the table of contents of “The General Theory” by March 1934, Keynes went to the United States in May and came home in June. He remarked in his letter to Kitoh (22 June 1934) that “it will be some months more before I am ready for printing”.
Insight into the subsequent proofing process is offered by eight pieces of evidence.
[1] The first evidence is a typescript, referred to by Moggridge as “The Pre-First Proof Typescript”4, written in the summer of 1934 (concurrently with the Summer Manuscript). This contains the prototypes of GT, Ch.4, “The Choice of Units”, Ch.5, “Expectation as Determining Output and Employment” and Ch.11, “The Marginal Efficiency of Capital”.
Ch. 6, “The Meaning of Definition”, of “The General Theory” also corresponds to GT, Ch.6, “The Definition of Income, Saving and Investment” and its “Appendix on User Cost”.
In that summer Kahn visited and helped Keynes. In early September6 Keynes wrote to his publisher, D. Macmillan: “I am now fairly well on with my forthcoming Treatise on Economic Theory and … it might be ready for publication sometime next year. ... My provisional title is The General Theory of Employment, Interest and Money.”
After “The Pre-First Proof Typescript” came a set of galleys, which we will call Galley 1, 2 and 3 (June 1935).
(i) Three galleys have the same TOC.
(ii) Galley2 is a revised version of Galley1 while Galley3 revises Galley2. In each galley, only part of the material specified in the TOC was written. Moreover, Galley2 rewrites only parts of Galley1, while Galley3 rewrites only parts of Galley2. Consequently, some sections of the text remained at either Galley1 or 2 stage until what we shall call “The Changeover” (September 1935).
(iii) The text of Galley1 was, of course, written first, but it appeared in October and December 1934, and in the January, May and summer of 1935, so parts of Galley1 postdate Galley2.
(iv) Although Galley1 of GT, Ch.17, “The Psychological and Social Incentives to Liquidity”, and Ch.19, “The Essential Properties of Interest and Money” do not survive, we have confirmation of it in Robertson’s letter to Keynes (3 February 1935) etc.
(v) The TOC differs considerably from that of the General Theory, for Keynes considerably revised Galley3 in August 1935.
[2] The second source is Galley1. By 13 September 1934 Keynes had sent off the first three chapters, the titles of which are the same as those of the corresponding chapters of the General Theory (though Ch.3, “The Principle of Effective Demand”, differs in substance). The galley for them came out on 10 October. Keynes delivered his Michaelmas lectures based on it together with the manuscript for Chs. 4-14.
That Keynes felt this manuscript needed revision is clear from two letters to Kahn. On 18 September Keynes wrote:
“I … have found out one or two interesting novelties. … I’ve solved the riddle of how to define Income in some sort of a net sense…. The deduction from the gross sales proceeds of the output of a given equipment necessary to yield income is that part of the quasi-rent … necessary to induce the entrepreneur not to leave his equipment idle. … [T]he appropriate depreciation allowance is the sacrifice involved in using the equipment as compared with postponing its use, as estimated by the entrepreneur himself”.
Nine days later he wrote:
“I am getting towards the end of the re-writing which you led me into and will show you the new way for dealing with net income in detail next term. It is clumsy …but the best I have done yet”.
These progress reports relate mainly to Ch.6, “The Meaning of Income”, which deals with user cost and Ch.7, “The Definitions of Quasi-Rent, Saving and Investment”, of the TOC.
Keynes then sent the publisher first Chs.4-11, and secondly Chs.12-19, which respectively correspond to GT, Chs. 4-9 and Chs. 11-17. The galleys for these probably came out between early December 1934 and mid-January 1935. This marks the virtual birth of the General Theory. Keynes sent off to Robertson10 Chs.1-19, which we will call “Galley1(I)”.
[3] The third source is the revision of Galley1(I), i.e., Galley2. Ch.17, “The Psychological and Social Incentives to Liquidity”, and Ch.19, “The Essential Properties of Interest and Money”, which correspond to GT, Chs. 15 and 17 respectively, both survive. As the second galleys for Chs. 1-6 were ready on 29 January 1935, some of Galley1(I) must have been returned to the publisher by the end of 1934. The remaining part would have been revised by January 1935, for Keynes wrote a letter to Kahn (15 January):
“I have done two more chapters for you …. [T]hey cover the ground of the philosophical chapter, ninety per cent re-written … concerning the fundamental characteristics of interest … considerably remodeled”.
This would relate to Ch.18, “Philosophical Considerations on the Essential Properties of Capital, Interest and Money”, of “The General Theory”, suggesting that Keynes reworked this as Ch.18, “Sundry Observations on the Nature of Capital”, and Ch.19, “The Essential Properties of Interest and Money”. In response, the publisher returned the remainder of Galley2, i.e. Chs.7-19, to Keynes in April 1935.
[4] The fourth source is the manuscript comprising Chs. 20-25 of the TOC, which cover the same material as GT, Chs. 10 and 18-22. Keynes probably sent this off to the publisher after March 1935, as is suggested by two circumstances.
Firstly, he wrote a letter to Kahn (26 March)12: “I have now finished a full-dress critique of the Prof. [Pigou] to go in as an appendix to the chapter on changes in money wages”. This would indicate that he had already finished writing “Appendix on Professor Pigou’s Theory of Unemployment” to Ch.23, “Changes in Money Wages”, which corresponds to GT, Appendix to Ch.19.
Secondly, this manuscript was sent off to several colleagues in June.
From these elements we deduce that Keynes received the galley in May 1935. We will call this Galley1(II).
[5] The fifth source is Galley3 (June), comprising Chs.2-6, which correspond to GT, Chs.2-5 and part of Chapter 6. In June Keynes sent off to Harrod, Hawtrey, Kahn, and Joan Robinson the following:
(i) Ch.1 of Galley2;
(ii) Chs.7-19 of Galley2, which correspond to GT, some of Ch.6, all of Ch.7, Section I-IV of Ch.8, Chs.9-14, Appendix to Ch.14, and Chs.15-17;
(iii) Galley1(II) (Chs.20 to 25).
This means that what they received was Chs.1-25 of the TOC, corresponding to GT, Chs.1-22.
According to Keynes’s letter to Macmillan (31 July):
(i) he sent the final pages of Galley3 to the printer;
(ii) he expected the book to appear in November.
[6] The sixth source concerns Chs.26-28 of the TOC. Although the galley for these is not extant, the chapter titles are known to us; Ch.26, “Notes on Mercantilism and the Usury Laws”, Ch.27, “Notes on the History of the Notion of ‘Effective Demand’ ”, and Ch.28, “Is an Individualist Economy Capable of Providing Full Employment?”.
Chs.26 and 27 together would correspond to GT, Ch.23, “Notes on Mercantilism, the Usury Laws, Stamped Money and Theories of Under-Consumption” while Ch.28 to Ch.24, “Concluding Notes on the Social Philosophy towards Which the General Theory Might Lead”.
According to Keynes’s letter to J. Robinson (3 September) he sent the last three chapters off to her, and was still working along the TOC:
I now have … the last three chapters …. The last two chapters are completely unrevised. Roy [Harrod] strongly objects to chapter 26 as a tendentious attempt to glorify imbeciles. ... I have been occupied for several weeks in somewhat re-writing Book I and completely re-writing Book II. In the case of Book II practically not a word of the version you have read has been left standing. … I have somewhat modified my definition of user cost (JMK.13, pp. 650-651).
On the evidence of this letter together with the galleys he sent to some fellow-economists in June 1935 and his letter to Daniel Macmillan (11 September), Keynes’s work on these chapters and their appearance in galley form would have taken place between June and July. We will call this set of proofs Galley1(III).
[7] The seventh source is the revision after Galley1(III), the evidence for which comes from Keynes’s letter to Robinson (3 September) and to Hawtrey (4 September).
The first reveals that he had been engaged for several weeks in rewriting Book I (Chs.1-3) and Book II (Chs.4-9).
The second indicates that Keynes revised Book III (Chs.10-11).
From Hawtrey’s detailed comments, and Keynes’s response (1 October) we know that Keynes had also embarked on rewriting Book IV. We will call it “The Great Revision”.
[8] The eighth source is the revision of Galley1(III). Keynes carried this out between September and October, for we have his letter to Robinson (3 September) and Robertson (10 October), in which he says: “I am now practically finished, and am sending my galleys to the printers to be paged”.
We may also presume that the change-over from the TOC to the GT’s TOC was made in this period. We will call it “The Changeover”.
In his letter to Macmillan (9 January 1936) Keynes writes: “I now have the exact size of the book for the leaflet, namely, xii plus 403”.
After 19 January or thereabouts, the General Theory was out of his hands. It was published on 4 February.
Now we need to keep the following in mind.
First, Galley1 is the most telling of the galleys:
(i) the galley is written along the TOC and laid the foundations for succeeding galleys;
(ii) through an examination of the appendix to JMK.14 (pp. 351-512; hereafter “The Variorum”), we see that in Galleys2 and 3 Keynes made only formal revisions except for a few points.
Second, we need to compare these sets of galleys not only with the preceding developments but also with the General Theory.
For the first comparison we focus on the new ideas in each set of galleys. For the second, together with the new ideas in each set of galleys we will examine formal aspects such as the degree of completion.
II. The Pre-First Proof Typescript
１． The employment function
２． Definitions of some fundamental concepts
III. Galley1(I)
1. The state of Development before Galley1(I)
A. Consumption
(a) Objective factors
However, Ch.10 of Galley1(I) differs considerably from GT, Ch.8, Section II. In “The General Theory” Keynes identified the objective factors as:
(i) the quantity of employment as determining the aggregate current rate of real income;
(ii) the rate of interest;
(iii) the state of long-term expectation.
In Galley1(I) he deleted (i), keeping (ii) and (iii) only, and proceeded to discuss (i) separately, using income rather than the quantity of employment. He then went on to add an argument justifying the substitution in the General Theory.
The analysis in (ii), which is the same as in “The General Theory”, is included as GT’s fourth objective factor, “changes in the rate of time-discounting” while the analysis in (iii), which is the same as that in “The General Theory”, is incorporated to some extent in GT’s third objective factor, “windfall changes in capital-values not allowed for in calculating net income”.
(b) Subjective factors
The argument concerning the influence of distribution on the propensity to consume restates the argument of “The General Theory”, namely that the employment function becomes a straight line with a slope of 45°. In GT Keynes drops it and takes income distribution as given. Concerning “negative saving”, furthermore, in Galley1(I) Keynes adds negative saving by firms and “unemployment relief financed by borrowing” together with the purchase of annuities and death duties in “The General Theory”. In GT, on the other hand, he refers only to “unemployment relief financed by borrowing”.
B. The Rate of Interest
Ch.17, “The Psychological and Social Incentives to Liquidity”, and Ch.19, “The Essential Properties of Interest and Money”29, of Galley1(I) (corresponding respectively to GT, Ch.15, “The Psychological and Business Incentives to Liquidity”, and Ch.17 (the same title) are not extant. However, we have Keynes’s letter to Kahn (15 January 1935)30, and Robertson’s letter to Keynes (3 February) for them.
The letter to Kahn indicates that Keynes rewrote 90 per cent of these chapters of Galley1(I). All we can say for certain is that Chs.17 and 19 of Galley2, which were completed by 15 January 1935, are identical to GT, Chs.15 and 17, apart from minor changes in wording.
The differences between Galley1(I) and Galley2 concern very trivial points. In the Great Revision, however, both chapters were considerably rewritten before the General Theory. This is confirmed by:
(i) Keynes’s letter to Kahn (27 August), in which he was thinking of completely rewriting the chapters on the theory of interest;
(ii) the fact that of the chapters on the theory of the rate of interest, Keynes rewrote these two chapters only.
Keynes was spurred on to rewrite these two chapters as a result of a discussion with Harrod, which brought his attention to his failure to make his theory of interest immune to misunderstanding.
Although the theme is common to GT, the appendix to Chapter 14 and to Ch.16 of Galley1(I), in formal respects we see two points of difference:
(i) in Galley1(I), Keynes points out that Marshall does not clearly distinguish quasi-rent as a return on assets from interest as a return on money, while in GT he deletes this;
(ii) Keynes’s criticism of Pigou’s theory of interest is more extensively developed in GT, using also Pigou(1927; 1933).
III. Galley1(II)
In Ch.3, “The Principle of Effective Demand”, of Galley1(I), Keynes defined the employment function as D’= F(N) (D’ denotes “the supply price” in the sense of an expectation of sale proceeds), N the volume of employment). However, in Ch.21 of First Galley1(II) he defines it as N = F(D) (where D denotes effective demand).
Then, in the Great Revision, Keynes set about clearing up the confusion by, on the one hand, calling the employment function of Galley1(I) the “aggregate supply function”, and on the other hand, calling that of Galley1(II) the employment function. In the Great Revision, Keynes also used for the first time the concept of the “aggregate demand function”, and re-defined effective demand as exclusive of “user cost”.
From this development we can infer that the employment function, N = F(Dw), of GT is a function of the equilibrium value, Dw (effective demand in terms of wage units). This can be verified by examining the theoretical structure of GT directly.
Ch.23 of Galley1(II) can be traced back to the following:
(i) Keynes launches his attack on Pigou (1933) in September 1933, and enters into fierce debate with Robertson;
(ii) Keynes discusses the book in his first lecture of the 1933 Michaelmas Term; (iii) in a letter to Kahn (26 March 1935), Keynes says that he has finished writing an
appendix for Pigou’s book.
Aside from these three chapters, we find no precursors to Galley1(II).
We see no difference in content between Galley1(II) and the corresponding parts of GT except minor changes in wording. The only exception is Section II of Ch.20, which summarizes the argument in the preceding chapters. It was considerably rewritten, possibly, in the Great Revision.
V. The Great Revision
b. Theoretical and Conceptual Revisions
The most significant differences between the Summer Manuscript and “The General Theory” in the definitions are that in the Summer Manuscript:
(i) windfall profit disappears;
(ii) user cost, Us, is considered to be the difference between effective demand and income (equation (7));
(iii) income equals the sum of consumption and investment (equation (8)).
(ii) Galley1(I) and GT
For a proper understanding of Keynes’s definitions of fundamental concepts we need to attend to how the concept of “user cost” is defined.
We can summarize the relation between the definitions of fundamental concepts adopted in Galley1(I) and in GT as follows:
(i) The difference in the definitions of effective demand, investment and prime cost depends on whether they include user cost (Galley1(I)) or not (GT);
(ii) the definitions of income, profit and saving are the same in the two;
(iii) the equation U2 = U1 - B is vital to the relation between the two sets of definitions.
B. Chapter 3
We say that equations (45) to (47) are the formulation used in GT with the proviso that in GT the definitions of effective demand, investment, income, profit and saving have been changed in such a way that user cost is not included. In GT Keynes explains the reason for this change as follows: “since user cost is obviously dependent both on the degree of integration of industry and on the extent to which entrepreneurs buy from one another, there can be no definition of the aggregate sums paid by purchasers, inclusive of user cost, which is independent of these factors” (GT, p. 24, fn.2).
VI. The Changeover
As we explained in Section 1, “The Changeover” designates the revision work of Galley3 and the change from TOC to the table of contents of GT, which took place between September and October 1935. Evidently, up to September Keynes continued to structure his work in line with TOC. In his letter to Harrod (17 August 1935) he writes: “Here are the last two chapters of my book. ... But chapter 26 is too long, .... In chapter 27 the emphasis hasn’t worked out” (JMK. 13, p. 542). Further, in his letter to Hawtrey (4 September 1935), he reports: “I have ... completed my re-writing of the first three books, namely chapters 1-11” (JMK. 13, p. 576).
Even as late as September, Keynes continued to rewrite the chapters treating the rate of interest. In his letter to Harrod (10 September 1935), we read: “I shall be here from September 22 … up to the end of the month. ... by then I shall have finished re-writing the chapters dealing with rate of interest” (JMK.13, p. 559). This bears fruit in GT, Ch.14, “The Classical Theory of the Rate of Interest”.
However we know neither when Keynes compiled the table of contents of GT, nor how he rewrote the galley after September. August was the period of the Great Revision. At that stage Keynes may well have been simultaneously engaged in rewriting the galley making a marked departure from TOC, although we lack the evidence. Apart from this, it was probably in September and October that he rearranged the whole galley in line with the table of contents of GT. Hawtrey, in his letter to Keynes (19 December 1935), made comments on the proofs based on the table of contents of GT, for we find such a passage as “Own―rates of interest. A renewed study of chapter 17 leads me ...” (JMK.13, p. 625).
It should be noted that Chs.6-10 (in TOC) undergo substantial further change to reach the form of GT, and that there the arrangement again changes as a result of Ch.22 (of TOC) being shifted to become Ch.10 of GT.
VII. CONCLUSION
The following points emerge from our analysis:
(1) Galley1(I) represents the most considerable piece of revision work carried out on the topics covered in TOC. Galley2 and Galley3 represent stylistic revisions of Galley1(I).
(2) Galley1(I) is composed of Chapters 1-19. Of these, Chs. 4, 5, 12 and 13 had been fashioned into completed texts of the corresponding parts of GT before Galley1(I). The other chapters are completed both in contents and at the stylistic level, as far as TOC is concerned. However, the theory of consumption, the theory of liquidity preference, and the definitions of various fundamental concepts had been completed in substance before Galley1(I).
(3) The largest change in the proofing process from Galley1(I) to Galley3 takes place in the Great Revision. The definitions of some fundamental concepts (such as effective demand, investment, and the prime cost) change as a result of both the change in the definition of “user cost” and the change in its treatment.
“The Monetary Theory of Production”: Before April 1932. JMK.13, pp.381-396.
“Historical Retrospect” (1932): JMK.13, pp.406-408.
“The Parameters of a Monetary Economy” (1932):JMK.13, pp.397-405.
“The First Manuscript” (1933): JMK.29, pp.62-66.
“The Second Manuscript” (1933): JMK.29, pp.63, 66-73, 87-92, 95-102.
“The Third Manuscript” (1933): JMK.29, pp.76-101, and JMK.13, pp.421-422.
“The First Undated Manuscript”: the end of 1932 - the beginning 1933. JMK.29, pp.102-111.
“The Second Undated Manuscript”: the end of 1932 - the beginning 1933. JMK.29, pp.111-120.
“The General Theory”: Spring in 1934. JMK.13, pp.423-456.
The Summer Manuscript (1934): JMK.13, pp.471-484.
The Pre-First Proof Typescript: Summer in 1934. See JMK.14, p.351.
TOC: the table of contents from Galley1 to Galley3.
Galley1(I), Galley2, Galley1(II), Galley3 and Galley1(III): See LKE and SKE, Table 13-1.
The Great Revision: August 1935. See LKE and SKE, 14, 1(A).
The Changeover: September and October 1935.
LKE: Hirai T. 2003, Looking at Keynes’s Economics from Multiple Points of View, University of Tokyo Press (in Japanese). LKE, x, y:LKE, Ch.x, Sec. y. (The short and updated version is to be published from Routledge.)
SKE: Hirai T., December 1997-March 1999, “A Study of Keynes’s Economics” (I)-(IV), Sophia Economic Review, 43(1[67-136], 2[13-121]), 44 (1[35-127], 2[29-96]). SKE, x, y: SKE, Ch.x, Sec. y.
