Published by EH.NET (September 2004)

Jon Agar, The Government Machine: A Revolutionary History of the Computer. Cambridge, MA: MIT Press, 2003. viii + 554. $50 (cloth), ISBN: 0-262-01202-2.

Reviewed for EH.NET by Roger Middleton, Department of Historical Studies, University of Bristol.

Jon Agar, sometime Director of the National Archive for the History of Computing, University of Manchester, and now a Research Fellow at the Department of Science and Technology Studies, University College London, here contributes a further volume to MIT’s prestigious series on the history of computing. The government in question is the United Kingdom; the topic the mechanization of its activities; and the time span from c.1798, when a standardized mass questionnaire was first used by agents of the British state, through to the beginning of the twenty first century. So far so good, though the “revolutionary history” of the computer here offered will, I suspect, surprise many in its emphasis as much upon the discursive as the material qualities of the general-purpose machine which comprises the centerpiece of what is now routinely referred to as the second informational revolution, that under which we are currently living. Accordingly, this is most decidedly not the sort of “anorak” account of heroic inventions that we have come to expect of the older style of writings on the history of technology, but instead draws upon insights from business and economic history, the sociology of professionalization and the sorts of approaches to the history of science and technology which will be familiar to readers of journals such Technology and Culture and Enterprise and Society.

The Government Machine is a story about humans and machines, but particularly about humans who promoted machines as members of what he calls expert movements (or what historians of ideas now term, epistemic communities), in this case relatively small communities of academics, civil servants, engineers and others claiming professional expertise who per force had to persuade outsiders of the legitimacy of their claims and did so by championing the introduction of new techniques. We have thus a contribution to on-going debates about the role of experts in British government since the nineteenth century, and there is much here that is pertinent to the debate prompted by MacDonagh (1958) and continued by MacLeod (1988), upon which Agar makes some important revisionist points. But his ambitions are very much greater. The central thesis comes early and typically in provocative mode, when he proclaims “that the apotheosis of the civil servant can be found in a place unfamiliar to most historians and in a form that will surprise them: the computer.” He is explicit that this is “intended to discomfort many readers” (p. 3) and that to convince them of his case requires that two hitherto largely separate disciplines be bridged.

On the one hand, Agar aims to convince political historians and historians of public administration that they have much to learn from the history of technology and the material culture of bureaucracy, with this study highlighting:

* that the changing capacities of government depended on the implementation of new technologies;

* that the civil service contained technocratic movement of experts, the full implications of which can only be appreciated when stereotyped notions of the dead hand of the Treasury are abandoned; and

* that during the twentieth century a very British style solution to the state’s needs for registers of its citizens evolved and is still evolving.

On the other hand, he wants to provoke historians of science and technology to re-examine their understanding of the role of government; in particular:

* that the uptake of office technology depended on a vision of government, even if it was typically commercial business that formed the main market for such technology;

* that the state provided a model of organization so fundamental that considerations of order, framework, structure and machine are inextricably linked with understandings of state or government — in effect, that to study the history of technology is to study the state, and vice versa; and

* that the civil service, cast as a general-purpose universal machine, framed the language of what a computer was and what it could do.

Before assessing the extent to which these ambitions are realized a note is appropriate here about the organization of and sources and inspirations for this study. First, we have a broadly chronological account, though certain topics (Babbage’s analytical engine) and characters (Turing) make more frequent appearances. Second, this is an archive-based study, and particularly rich in the use of official papers deposited in what now calls itself The National Archive (previously, the Public Record Office). Thirdly, and connectedly, while the study does come up to the new millennium, the post-1970 chapters are inevitably affected by the Thirty Year Rule for official papers. Fourthly, Agar’s work borrows and extends into the latter period Mayr’s (1986) thesis about the prevalence of particular metaphors and their profound technological consequences, whilst also building on path-breaking work on office mechanization and data processing, both public and private sector, by Campbell-Kelly. Finally, the text is liberally interspersed with images taken from the National Archive for the History of Computing, appropriately located in Manchester for it was there that the Small-Scale Experimental Machine (known as the Baby, the first machine comprising all of the components now classically regarded as characteristic of a computer) was designed, built and successfully operated on 21 June 1948. As the saying goes, what Manchester does today …

In chapter 1, Agar considers why machine-like characteristics have been attributed to governments, and it is here that he extends Mayr on the interplay between metaphorical machines and styles of government. The prevalent metaphor gives rise to a mechanical discourse of government which guided and constrained thought and action, though he does not deny that other discourses (government as organic) have featured also. However, what was particularly important in Britain was how the machine metaphor underpinned a triadic division of labor within the civil service (of generalists, the intellectual and upper class; the middle class specialists; and a lower class of mechanicals and clerks) which had particular significance for the technological trajectory of government as influenced by the expert machinery movement he goes on to chart in subsequent chapters. Chapter 2 deepens the historical analysis of the civil service from the late eighteenth century through to Second World War with Turing’s conception of a universal computing machine and Bletchley Park code breaking. An immense repertoire of mechanical language concerning government is unveiled and he shows how this became a resource for the expert movement of mechanizers. One very important conclusion at this stage is that the existing historiography’s portrayal of the generalist’s hostility to specialists, to quantification and to modern mechanization of governmental activities is deeply flawed. Thus, machines must be incorporated into histories of Britain’s civil service, and we need to take seriously the Treasury’s role in the twentieth century as an instrument for labor-saving new office technologies.

Subsequent chapters then develop in broadly chronological order the debates, deliberations and lobbying of this expert movement of mechanicals. Chapter 3 covers the period from approximately the second quarter of the nineteenth through to the early twentieth century, and demonstrates how the movement propagated its message by contrasting the chaotic state of knowledge about contemporary social and economic conditions in relation to the much more rational and ordered state of statistical knowledge assumed in competitor economies. This portrayal of Britain as a laggard then becomes a standard rhetorical technique, while we see — in terms of professionalization — an increasing emphasis on official statisticians claiming privileged expertise in terms of the production and use of statistics, though — as with the economists at this time — they were far from able to exercise monopoly rights. There is much of interest here for economic historians on the lobbying for a Census of Production, successful in 1907; on the campaign to rationalize and centralize statistical production and dissemination through the establishment of a Central Statistical Office, not achieved until 1941; and in contemporary anxieties about the avalanche of numbers and informational overload confronting policy-makers, a concern which added to the rhetorical armory of the expert movement.

In chapter 4 the emphasis shifts from efforts to aggregate for social and economic policy purposes to the demands of the First World War for knowledge, total knowledge for the more ambitious, of the population through the creation of a universal or national register. The informational needs of war quickly raised the issue of whether the number of partial registers in existence might be brought together into one universal register, duly cross-referenced. Had this transpired, the current state of computing technology available (punch cards had been used for the first time in the 1911 census) would have raised interesting questions about how it could have been used, but Agar shows that efforts at this time to bureaucratize the British were a failure because of deep-seated cultural and political antipathies towards the state having such capabilities. Partial registers were thus a hallmark of British government, and have been since as the current Home Secretary is finding in the storm of protests that have erupted against his proposals for a national identity card.

Chapters 5 through 7 are also concerned with the relationship between warfare and different information technologies. In chapter 5, Agar discusses how the upward shift in the mechanization of government activities that had occurred during the First World War was then given a further boost between the wars by the implicit bargain struck between the expert movement and the generalists working to Treasury expenditure control ambitions. Chapter 6 discusses the 1930s as another informational crisis for the British state which was met by the establishment of the Telecommunications Research Establishment (which developed radar) and the much more widely known Bletchley Park. The Second World War also marks the beginnings of an expert movement (a technocratic network) within the Treasury, in the form of Organisation and Methods (O&M) which Agar later argues comprises a non-military counterpart to Edgerton’s (1991) thesis that British governments in the twentieth century pursued a policy of “liberal militarism” which required the creation of technocratic departments of state. The British computer industry of the 1940s and 1950s forms the focus of chapter 7, confirming how the computer shaped and in turn was decisively shaped by the Cold War.

Treasury O&M and its importance for the computerization of postwar government work is then the subject of chapter 8. The installation of punch-card machines, numbering twenty-six installed or planned in 1948, was to expand rapidly thereafter. By 1958, seven departments had installed or were in the process of ordering computers, while by 1965 there were forty-five installed and a program of 250 to 300 more installations planned for the next decade. Treasury O&M staff were crucial in this process whereby the computer was reconceptualized from being an exclusively scientific tool, as it had been say at Bletchley Park, to that of a general-purpose business and office tool. Agar provides in Table 8.1 (pp. 316-17) a breakdown by department and by function of these installation plans which had now extended beyond the traditional areas of government office mechanization (the military, national insurance and other civil registry functions) to encompass most of Whitehall, save one very noticeable exception in the 1950s: the Treasury itself! Treasury officials charged with macroeconomic management did eventually lobby for a Treasury computer for long-term forecasting, but meanwhile for budget forecasting purposes had to rent time on LEO, the world’s first office computer which was owned by J. Lyons & Co, a countrywide chain of tea shops.

In chapter 9 Agar sets out to examine why concerns about privacy, and in particular a perceived threat from government computers, emerged in the late 1960s after the decade had begun so well and was then fortified with the rhetoric of the “white heat of the technological revolution.” In part, of course, the state of technology did now pose more of an objective challenge to established British notions of privacy and data protection, and we are reminded here of the First World War and the antipathy towards universal registers. Agar does not find that explanation wholly satisfactory and argues further that it was at this time that there was a shift in anxieties about threats to the collectivity posed by universal registers to anxieties about threats to individuals which are part-and-parcel of changes in the nature of public trust.

Chapter 9 ends with the Data Protection Act (1984), which was the interim product of these concerns. Chapter 10 extends the story to further data protection legislation which came into force in 2000, but is more concerned with the implications of what political scientists call the hollowed-out state for public sector computer projects where in recent years there have been some spectacular disasters, so much so that in Britain today a successful large-scale IT project is considered something of an oxymoron. Agar’s conception of the hollowed-out state is actually not much informed by the academic literature (Holliday 2000 provides a short survey of the debate) and, very surprisingly, he has nothing to say on the growth of audit and its implications, but the key for Agar is that the hollowed-out British state has lost key capabilities with respect to IT. His expert movement of mechanizers are no longer centered on Whitehall, for large-scale contracting-out has led to the locus of power — based in knowledge, competence and expertise — now residing in private sector specialists like the multinational EDS. Ironically, of course, this trend has increased the threat to privacy as private sector companies are less accountable than government.

Finally, there is a short concluding chapter which inter alia examines in international context whether British governments were backward in office mechanization and asks whether there are links here with the broader decline literature. Rejecting the deadhand of the Treasury orthodoxy, and citing approvingly Campbell-Kelly (1998) and other studies demonstrating that government departments were in the vanguard of productivity enhancing mechanization, he concludes “If there has been a distaste for the machine it has been among writers on government rather than in government itself” (p. 414). This economic historian would have liked rather more evidence on the extent of the efficiency gains attained, rather than those claimed. The Solow paradox is conspicuous for its absence here. Nonetheless, Agar is to be congratulated for his path-breaking study of a fascinating topic which he has opened up in a way which others can now develop.


Campbell-Kelly, M. (1998) “Data Processing and Technological Change: The Post Office Savings Bank, 1861-1930,” Technology and Culture 39 (1): 1-32.

Edgerton, D.E.H. (1991) “Liberal Militarism and the British State,” New Left Review 185 (January/February): 138-69.

Holliday, I. (2000) “Is the British State Hollowing out?” Political Quarterly 71 (2): 167-76.

MacDonagh, O. (1958) “The Nineteenth-century Revolution in Government: A Reappraisal,” Historical Journal 1 (1): 52-67.

MacLeod, R.M. (ed.) (1988) Government and Expertise: Specialists, Administrators and Professionals, 1860-1919. Cambridge: Cambridge University Press.

Mayr, O. (1986) Authority, Liberty and Automatic Machinery in Early Modern Europe. Baltimore: Johns Hopkins University Press.

Roger Middleton’s latest publications include a chapter on government and the economy, 1860-1939 for the recent Cambridge Economic History of Modern Britain (2003), and a co-authored book, Economic Policy under the Conservatives, 1951-64 (2004). He is beginning work on a political economy of Thatcherism and continuing to examine the growth consciousness of the 1960s.