Seminar 3 - some further reading:
this essay, below, may be interesting for those of you who have been inspired to find out more about the early history of computing.
"Machinery Has Been Taught Poetry Instead of Arithmetic": Babbage and the Division of Mental and Manual Labour
Ian Miles
- revised April 2000, originally prepared in 1989 -
This essay seeks to bring to wider attention a very interesting but neglected study in the early history of Information Technology. I also want to make some points about the relation between technology, markets, and rhetoric. The essay was originally prepared in response to a paper by Bruce Berman entitled "The Computer Metaphor", published in Science as Culture no 7 (1989); however I never got around to publishing it at that time, because of getting caught up in the move from SPRU to PREST. I have revised it slightly, and knocked out the footnotes (which HTML publishing via Word seems to lose).
I had met Bruce in the mid-1980s, when he was visiting SPRU (where I was then based). Bruce drew my attention to the remarkable fact that Charles Babbage is known both as an early theorist of both the division of labour in manufacturing and as a pioneer in the development of the computer. In Sweden in 1989, where I was kindly accommodated at Linköping University, I was able to satisfy some of the curiosity this had aroused, both by inspecting a Babbage difference engine at the Stockholm Technical Museum, and by reading some texts that cast light on the relation between these two achievements.
Two conclusions emerged from these investigations. First, Bruce Berman may inadvertently give a slightly misleading impression of the relation between Babbage's work on the division of mental and manual labour. Babbage did not develop the idea of programmable computing from his studies of the division of labour in manufacturing, as might be concluded from the quotation Bruce uses. If anything, the causal sequence is the opposite of this, surprising as it might seem (at first sight) to materialists.
Second, the failure of Babbage's enterprise was a complex and probably inevitable affair: It was not merely technical problems that hindered the development of Victorian computing. Rather, like some computers of more recent provenance, it appears that
Babbage's machine was a solution in search of a problem. The Difference Engine was actually built, though this history has been neglected. While technical problems rendered it rather less effective than its designers had hoped, the more serious problems were a lack of markets for automated forms of mental labour (of specific sorts of mental labour, to be precise) at that point in history.
- The Origins of Babbage's Project
Babbage's On the Economy of Machinery and Manufactures (I will cite from the 1835 edition) is a milestone text on the division of labour. In the preface Babbage provides an account of the origins of this book:
"one of the consequences that have resulted from the Calculating Engine, the construction of which I have been so long superintending. Having been induced, during the last ten years, to visit a considerable number of workshops and factories, both in England and on the Continent, for the purposes of endeavouring to make myself acquainted with the various resources of mechanical art, I was insensibly led to apply to them those principles of generalisation to which my other pursuits had naturally given rise.....Several of the principles which I have proposed, appear to me to have been unnoticed before...particularly...the explanation I have given of the division of labour." (no pagination in the Preface.)
Babbage thus himself claims that he was driven to study (and then to enunciate his conclusions about) the division of manual labour by needing to understand manufacturing processes. He claims, furthermore, that this necessity was imposed by his desire to construct a computer.
He calls this computer the Calculating Engine - or Difference Engine, referring to its mode of operation, and to distinguish it from the later invention of the Analytical Engine. The word "computer" was at the time used for the clerical occupation of calculation, and Babbage speculates that human "computers" will be displaced by technology:
"when the completion of a calculating engine shall have produced a substitute for the whole of the third section of computers, the attention of analysts will naturally be directed to simplifying its application, by a discussion of the methods of converting analytical formulae into numbers" (p195).
In this book, explaining the operation of the Calculating Engine, Babbage draws analogies between it and the division of labour in material manufacturing, comparing formulae to
machines(!). But this an expositional device: while it is quite likely that his study of manufacturing division of labour had some influence upon the analysis of the mental division of labour, and the mechanisation of certain parts of intellectual work, the evidence is strong that his account in the preface is correct. In other words, the theorist of the division of manual labour was born out of the would-be divider of mental labour. Babbage already had conceptualised the division of intellectual labour before he went on to study manual labour.
This would seem to reverse a standard materialist precept, in which consciousness is supposed to follow social action, not vice versa. Closer inspection resolves the paradox. Babbage's analysis of the manual division of labour was a theorisation of what was already happening in workshops. But in supporting this materialist approach, we cannot deny that consciousness certainly can inform social action. Indeed, Babbage’s ideas did have an impact upon subsequent manufacturing practice.
Furthermore, his conceptualisation of the division of intellectual labour was also founded on earlier practice. And it too had an impact on subsequent social action, as we shall see – if not the impact he expected in the quotation above.
Well before Babbage had begun his studies, people had been wrestling with the problems of simplifying complex computations (required for engineering, navigation, etc.) Babbage describes in Chapter XX one French effort that resulted in the production of 17 large folio volumes of mathematical tables (the Tables of Cladastres). This involved new methods for the production of such tables, involving a detailed division of labour between:
- 5 or 6 eminent mathematicians, whose job was to select among alternative analytical expressions of functions, choosing those most easily adaptable to a high division of labour in their calculation (incidentally, this involved a form of human parallel processing!);
- 7 people of high mathematical knowledge who would convert these formulae into numbers - a big job - and verify the calculations by "certain means" other than checking them all;
- 60 to 80 calculators, who used only simple addition and subtraction to calculate interim and final results.
Thus it appears that the division of both mental and manual labour in practice led Babbage first to theorise the former - and then, as a result of his efforts to mechanise the former, to theorise the latter. No doubt history was vastly complex than this summary account suggests. For example, there were calculating machines before Babbage's Engine, and it requires detailed study to specify what was new in his device, and to identify (as far as possible) where the notions came from. In Stockholm's Technical Museum, the Difference Engine was (in 1989 anyway) situated in the same room as earlier calculating machines and later electronic computers, suggesting a long (if often interrupted) tradition of work. Also in the room is a picture of the design for Babbage's later, unbuilt Analytical Engine - which was a far more radical proposal, much more like the modern programmable computer. This involved separate units carrying out arithmetic, data storage, and data output (printing), and using punch cards for programming. The Analytical Engine’s design is displayed alongside a Jacquard Loom. This makes the genealogy of Babbage’s notions of programs clear. Not only were punch cards used in the loom, but Babbage also labelled of part of the Engine as the "mill".
Babbage drew explicit analogies between the automation of, and division of labour in, "mechanical and mental operations". The division of labour in both cases:
"enables us to purchase and apply to each process precisely that quantity of skill and knowledge which is required for it."(p201)
Interestingly, he goes on to rephrase Smith's well-known comment about the division of labour being limited by the size of the market: "division of labour cannot be successfully applied unless there exists a great demand for its produce; and it requires a large capital to be employed in those arts in which it is used."(p201) This statement throws a great deal of light onto what was to become the fate of the Difference Engine in practice.
Babbage's book is well worth reading, not least because of its futurological elements, and the author's unembarassed statement of his ideology. (It also contains a splenetic early
account of the "British Disease", as he complains about the Presidency of the Royal Society being taken by a member of the aristocracy rather than by a member of the "nobility of science"; and a discussion of energy crises!)
Babbage sees science as key to technical progress, and mathematics as the ultimate science:
"the further we advance from the origin of our knowledge, the larger it becomes, and the greater power it bestows upon its cultivators, to add new fields to its domains. Yet...the whole, already gained, bears a constantly diminishing ratio to that which is contained within the still more rapidly expanding horizon of our knowledge... another and a higher science, itself still more boundless, is also advancing with a giant's stride, and having grasped the mightier masses of the universe, and reduced their wanderings to laws, has given us in its own condensed language, expressions, which are to the past as history, to the future as prophecy. It is the same science which is now preparing its fetters for the minutest atoms which nature has created.... It is the science of...calculation which becomes continually more necessary at each step of our progress, and which must ultimately govern the whole of the applications of science to the arts of life." (pp387-8)4
These laws of mathematics must have been designed, according to Babbage, who sees this as evidence of a creator: in a manner resonant of some recent commentators, for example Rudy Rucker, the Universe is seen as a vast programmed system.
- The Difference Engine in Practice
Before my Swedish trip I had not realised that several working versions of the Difference Engine were actually constructed. Indeed, I had fallen for the line later recycled when the Science Museum displayed their model of the Engine, that Babbage’s device had not previously been constructed till their effort. However, in Sweden, I was lucky enough not only to see one, but also to come upon a rich account of their development. Unfortunately this study is not as well known as it deserves to be.
Michael Lindgren's (1987) book Glory and Failure provides a fascinating account of Babbage's work, of the people that created more-or-less viable products out of it, and of the fate of these products in the Victorian market. The book contains discussions of many
themes, but here I concentrate on points concerning the origins and fate of the Difference Engine. (Page references from here on are to this book.)
Lindgren notes that Babbage was a collector of tables, with a longstanding interest in the topic (he had visited Paris in the 1810s and '20s and seen The Tables de Cadastres there5). He also had a practical concern with generating them - in 1824 he had prepared mortality tables while helping to found a life insurance company. In 1822 Babbage was arguing that the progress of science would be retarded by "the accumulating labour which arises from the arithmetical applications of mathematical formulae"(p44) Whether this was a motivating factor in the development of the Difference Engine, or a marketing rationalisation is not clear.
Babbage was already working on the initial idea in 1821, and by 1822 a rough prototype had been constructed. One question that will strike the modern reader concerns why
the name "Difference Engine" was chosen. "Difference" comes from the calculation procedures followed: the method of differences. But it was unusual at the time to find "engine" applied to machinery, notes Lindgren, and it was new to apply it to calculating machinery. Hoiwever, the use of the term "steam engine" was widespread, and a technological revolution was underway through the use of steam power at the time Babbage was writing. Babbage had envisaged the Difference Engine being steam-powered. The steam power would not only drive the calculating unit, but also a printing unit, achieving mechanised typesetting of the tables, so they could be reproduced reliably. Lindgren comments that this latter feature was in many respects the most radical aspect of the Engine: at that time newspapers were being printed by steam, but not set. However, Lindgren (pp39-41) also plausibly suggests that Babbage was hoping that the term would highlight the technologically revolutionary implications of his idea. Parallels with present-day "hyping" of products are readily apparent.
Babbage had some success in promoting his idea: in 1823 the Astronomical Society (a body with interest in calculation of planetary tables, etc.) awarded him a medal for invention, and the British government donated £1500 towards the costs of building a working model. Another £1500 was stumped up in 1829, and Lindgren estimates the total cost to the government as some £17,000 in the end: a considerable amount of money for the time.
But Babbage failed to deliver the goods. Much of Lindgren's study is concerned with delineating the reasons for this. These reasons include Babbage's perfectionism (the best illustration of which is his almost fanatical self-imposed research project studying the readability of different coloured inks on different coloured papers, with a huge range of unlikely and implausible combinations!), his underestimation of expenses, intervening illnesses, personality clashes with his (also perfectionist and expensive) engineer, repeated modifications of the design, and so on. It was actually in the course of trying to get workshops to meet his exacting standards that Babbage was driven to study the division of manual labour.
By 1833 only one section of the Engine was completed, less than one seventh of the total unit, and of little practical use by itself. The printing unit was never constructed. Babbage
halted work in 1834, turning his attention to the Analytical Engine. And this is where the story usually ends: the failure of the Difference Engine was a technical or personality problem. If it could have been constructed, the implication goes, it would have found wide application. This theme is pursued in a highly interesting and well-informed alternate-history novel, The Difference Engine by William Gibson and Bruce Sterling, pioneers of the "cyberpunk" genre of science fiction. This book explores the implications of an IT revolution happening simultaneously with the industrial revolution. For my review of this – and why its founding assumptions are shown by Lindgren’s work to be wrong, see Miles (1991.)
Lindgren documents that a Swedish teenager actually did construct Difference Engines (including the one that I saw). He shows, too, that they proved practically impossible to commercialise.
Georg Scheutz was a Swedish innovator - a great deal of his work was a matter of technology transfer and diffusion. Much of his writing also finds curiously echoes in the 1990s, although he castigates Sweden for things which are now more commonly described as flaws of the British. Thus in 1834 he noted that Sweden lacks:
"the known talent of bringing...goods to the market, which the foreigner usually possesses to a high degree and which we...almost despise" (p91).
Scheutz described Britain in terms now reserved for Japan: there was :
"in England a special class of industrialists, who made their money from other people's inventions...tirelessly read foreign journals, reports, etc... special agents who studied the patent offices and workshops in foreign countries" (p134).
This made him think twice about patenting in the UK, even though he saw more chances for diffusing new products in the UK than in Sweden. Scheutz shared with Babbage an interest in mathematical tables and calculating machines. An avid reader, he encountered
Babbage's writings on the Difference Engine. Concluding that there were flaws in Babbage's design, he interested his son Edvard in the project - and Edvard eventually produced a hand-driven Difference Engine cheaply, initially as a hobby, at home. Does this remind you of the garage computer industry of California in the 1960s?
The first machine was completed in 1843. It was smaller and lighter than Babbage's machine, and embodied solutions to some printing problems that Babbage had floundered on. A second machine was finished in 1853: this one was built in a modern factory workshop and incorporated some new features: the Scheutz's were hoping to commercialise their achievement. A third Engine was completed in 1859. Lindgren describes a long catalogue of efforts by the Scheutzs, with considerable support from Babbage, to market the Difference Engine. It was patented in the UK, exhibited in London and Paris.
What was the market for a Difference Engine? Georg Scheutz himself suggested in 1833 that:
"even if the larger and more complete engine should later on remain the only example of its kind, it would suffice for the needs of the whole world". (p272)
In other words, all the needed tables could be produced by one machine and sold in bulk. This sounds very reminiscent of suggestions made in the early days of electronic computers that probably one or two would suffice for all the computing requirements of a country; and these requirements were naturally more numerous and varied than those envisaged for the difference engine. But later, he was won to the idea of a small and portable machine, and in 1857 wrote to a mailing list of 400 arguing that the Engine would be:
"a necessity wherever there is an observatory, a university, or a press dealing with mathematical works... it will not take so many years for this to be the case" (p209).
Lindgren's analysis suggests that the earlier judgement was the more accurate - and that, since the needed tables had already largely been calculated to the number of digits which were the limit of the machine, the outlook was even less promising. (Still, there was the matter - which Babbage considered rather important - of avoiding the introduction of new errors in the reprinting of tables.)
The second Engine was eventually sold in 1857 for much less than hoped, to an observatory in Albany, USA. It was never used beyond its first trials, for a combination of reasons. These seemingly include Scheutz's refusal to release details of the operation to them, in order to guard his intellectual property. The result was that tables were produced, but never published.
The manufacture of the third engine was, after long efforts, supported by Britain's
General Register Office: the agreement was made in 1857 and the Engine completed in 1859. Life Tables were published from this Engine's work. But its UK manufacturer made financial losses on the project. Worse, the Engine misbehaved after delivery, and required costly maintenance. Its use was eventually abandoned.
Lindgren shows that the conventional account is misleading. Certainly Babbage failed to produce his Difference Engine, but this was not because it was technically too difficult. (Even the Analytical Engine might have been completed.) Indeed, it was possible for a gifted teenager to successfully produce this early, hand-driven computer. The final product was cumbersome and rather unreliable - but so are first versions of most innovations. More importantly, there was not the market for computers which might have provided the economies of scale in manufacture, and focused more engineering attention on the problems of the device, so as to overcome the problems that users did confront.
The forms of mental labour which the Difference Engine challenged were specialised ones, with limited users. It might be further argued that the development of firms, corporations, and government agencies was not yet of a form which dictated the application of more general-purpose computer power. Beniger’s The Control Revolution makes good reading on this issue.
Many themes could be extracted from the early history of the Difference Engine. It could be read, for instance, in terms of an innovation research framework: supply-side enthusiasm taking precedence over proper analysis of user requirements, for
instance. It could be read in terms of the evolving nature of demand for arithmetical calculations and mental labour of related types, and related to the changing nature of the so-called "information economy". The story of Babbage himself is often told in terms of a classical scientific hero-figure (or in more demonological terms in a left-wing version, perhaps), but it too can be reworked in terms of a treatment of science as culture.
Despite considerable change in the social role of science and technology, Babbage's tireless promotion technical solutions (the Difference Engine) to economic and organisational problems, and his willingness to make claims for his invention that far
exceeded its actual achievements, bears a dramatic similarity to certain more recent technological developments - for instance nuclear power and weapons (consider Walter Marshall and Edward Teller). Even the way in which he deploys arguments about the
fundamental role of mathematics, to bolster his invention, is echoed in, for instance, the rhetoric about unlocking the power of the basic building blocks of matter. The symbolic and practical dimensions of science and technology appear to be inextricably intertwined, at least when it comes to forging support for new products and programmes.
Charles Babbage, 1835 edition, On the Economy of Machinery and Manufactures (4th ed) London, Charles Knight; 1986 reprint by Augustus M Kelley, Fairfield NJ
James Beniger, 1986, The Control Revolution Cambridge, Mass, Harvard University Press
Bruce Berman, 1989, "The Computer Metaphor: Bureaucratizing the Mind" Science as Culture no 7 pp 7-42
Michael Lindgren, 1987, Glory and Failure: the difference engines of Johan Muller, Charles Babbage and Georg and Edvard Scheutz Linkoping Studies in Arts and Sciences,9 Stockholm Papers in History and Philosophy of Technology, 2017 ISBN91 7870 146 5 ISSN0282 9800
Ian Miles, 1991 "Past Tense, Futures Imperfect?"(review of William Gibson and Bruce Sterling, The Difference Engine) Futures vol 23 no 5 pp552-555
