Review of D. MacKenzie’s Knowing Machines, MIT Press, 1996
Donald MacKenzie has produced under the title ‘Knowing Machines: Essays on Technical Change’ a wide-ranging collection of his (most recent) previously published work. Not only in terms of topics discussed in the collection, is the book diverse, but also in the nature of the approaches that he has decided to adopt. It is clear that MacKenzie finds the various moves in recent sociology of technology towards ever greater symmetries debilitating. This is reflected in his overall approach, where he is at once a relativist ‘questioning the underlying knowledge/assumptions’ of the building of nuclear weapons (chapter ten -‘Tacit Knowledge and Uninvention of Nuclear Weapons’), and at the same, he is a realist ‘making knowledge claims’ about the prevalence of computer related accidents (chapter nine ‘Computer-Related Accidental Death’). This analytical ‘skirting around’ between approaches provides for a fascinating example of the difficulties of describing just what do we mean when we are talking about the ‘sociology of technology’. If this is a recipe for how to do sociology of technology studies, then it is certainly a heterogeneous one. One possible reading (which I think warrants a mention), while implicit in many of the chapters, but more obvious in others, is MacKenzie’s view of a sociology of technology which informs design.
We need little reminder of the influence of his early work, and of the sway of the social shaping framework. Straight away in the first chapter, ‘Marx and the Machine’, he begins to set out his politics. Here he recounts how Marx’s work on technology has most often been interpreted as an argument for technological determinism. Many Labour Process studies having fed of such accounts, deployed endless descriptions of the power of machines to displace and deskill labour. In arguing against this well held belief, MacKenzie reminds us of the sophistication of those pictures of the early disciplining of workers through technology, an account that shows just how the machine was able to make stable and durable the conditions of work. Nor was Marx arguing that technology was determinant at one remove, in that it solely embodied the ‘interests’ of capital. In MacKenzie’s hands, Marx appears to be arguing that while technology was shaped by the needs of capital, it remained essentially neutral:
It took both time and experience before the workers learnt to distinguish between machinery and its employment by capital, and therefore to transfer their attacks from the material instruments of production to the form of society which utilises those instruments (Marx, quoted in MacKenzie, p44).
If Marx confined his studies to how a neutral technology was later ‘shaped’, then we can rightly ask, ‘what relevance does his work have for a sociology of technology which is attempting to inform design?’ MacKenzie is clear that this approach ‘has relevance’ in looking at, what he calls, the ‘contingency of design’, or, in other words, how the design of technology could have be done ‘differently’. He points to examples from Langdon Winner’s account of the ‘politics’ of bridge building, and to David Noble’s description of the managerial assumptions embedded in machine tools, as landmark studies of identifying these types of contingencies. Once such contingency is identified, MacKenzie argues, there then exists a need to explain ‘why’ certain choices were made. And here Marx’s work is relevant: “…because it does suggest where to look for such an explanation…(p45)”. This is a nice chapter, but, in terms of the sociology of technology, I think that there is little new here. The chapter serves its purpose, which was to correct earlier straightforward technological deterministic interpretations of Marx’s, and at the same time, to remind us of the elegance of his writings. Further, although it does sit fairly uncomfortably among the other papers, it might be said to act as some sort of an introduction to MacKenzie’s own brand of social ‘shaping’ studies of technology. The use of ‘shaping’ as opposed to the ‘construction’ of technology delineates a (sometimes not recognised) difference in the history of the sociology of technology. This can also be characterised by a use of the ‘social’ to explain technical change. By pointing this out, I don’t mean to detract from the sophistication of the approach, and MacKenzie himself is not unproblematic in his use of the social: “This ‘social shaping of technology’… should not be thought of simply as unchanging social relationships causing changes to technology, for heterogeneous engineering involves changes to social relations too.” (p14)
In the next chapter, ‘Economic and Sociological Explanations of Technological Change’, MacKenzie attempts to discover some ‘common ground’ between social studies and economic studies of technology. Amongst other things, he picks up on the notion of ‘natural trajectory’ or ‘technical trajectory’ of technology. The notion, often used by evolutionary economists to explain persistent patterns of technological change, MacKenzie problematises, for the way ‘natural’ seems to resonate with the idea of the mechanical metaphor of ‘trajectory’: “The notion of ‘technological trajectory’ can thus very easily be taken to mean that once technological change is initially set on a given path…its development is then determined by technical forces.” (p55) In referencing studies from the social history of technology (shot) and actor network approaches (ant), he asserts that such studies have shown that while technological change does possess a momentum, it is “never momentum of its own (p55)”. While I agree with his criticism of the notion of trajectory, one of the problems of this kind of ‘outsider looking in’ analysis of evolutionary economics, is that the strawman that MacKenzie sets up (the technological trajectory), is one that does not exist unproblematically in evolutionary economics itself (see Van Lente (1993) for a more elaborate criticism). That said, we are shown the real strength of the collection (and of the potential for crossover) in his discussion of ‘self-fulfilling prophecy’. In economics, such a notion comes under the guise of ‘expectations’. For example, MacKenzie quotes from the economist Paul David’s work on the universal adoption of the QWERTY keyboard: “A particular system could triumph over rivals merely because the purchasers of the software (and/or the hardware) expected that it would do so” (David, 1986 cited in MacKenzie, p57). In a similar way, he uses the notion of self-fulfilling prophecy to add an aspect to the notion of technological trajectory. Namely, that: “[p]ersistent patterns of technological change are persistent in part because technologists and others believe they will be persistent” (p56). In drawing on material form one of the other chapters in the collection, ‘Nuclear Weapons Laboratories and the Development of Supercomputing’, he points to how designers of supercomputers, having estimated the possible future increases in speed of their rivals computers, would aim to build their machine to match or better the estimated speed of their rivals. Therefore, a continued increase in the speed of supercomputers, resulted (partly) from a belief among designers that faster speeds were possible: “The prophecy of a specific rate of increase has thus been self-fulfilling. It has clearly served as an incentive to technological ambition; it has also…served to limit such ambition” (p56). His use of self-fulfilling prophecy (SFP) is, I think, the most intriguing aspect of the collection of papers, and is not something that has been made much of in the sociology of technology. He carries this theme into several of the other papers, but makes most use of it in the chapter on ‘Tacit Knowledge and the Uninvention of Nuclear Weapons’.
This chapter hinges on the argument that the development of nuclear weapons requires both explicit and tacit knowledge for successful construction, and he (tentatively) points to a situation where, in the future, the tacit skills needed to build such weapons may disappear. In his conclusion, he attempts to show how future efforts to build these weapons will be so great (without the tacit skills of the earlier generation of scientists) that it won’t be a simple case of re-building from previous designs but a ‘re-invention’ will be required. And the effort to re-invent, he argues, may be more than most countries are willing to sacrifice. This of course raises an interesting question: could the difficulties of re-invention be such that countries are sufficiently deterred from taking this route? MacKenzie, in a way, dismisses his own conclusions when he invokes the following argument “[i]t is hard to imagine belief in the feasibility of atomic or thermonuclear weapons now disappearing, and that fact alone increases the probability of their reinvention” (p249). For MacKenzie, one of the reasons why it has continued is because opponents think its continuation is inevitable and therefore “[t]heir consequent failure to oppose it has been one factor making it possible” (p57). By positing the idea of the “uninvention” of nuclear weapons, through a loss of tacit knowledge is MacKenzie’s way of countering this pessimism. In taking his argument further, that we can see a time when these weapons might not exist, may enable us to look beyond what now seems inevitable.
If I were to make a criticism here, it would be this: MacKenzie makes recourse to the notion of SFP to explain the durability of nuclear weapons, without really elaborating on the term. The recent trend in the sociology of technology of unpacking reifications, warns us against using such blanket terms, however interesting or useful they appear to be. But of course, recall the introduction and the point about MacKenzie’s aversion to symmetries. Yet I think, his unwillingness to expand on the expression misses out on much of the interesting aspects of what is self-fulfilling about technical change. I am tempted to say that I think he intends a performative use of SFP, rather than the more cognitive (but it is difficult to make this separation without tending towards a dualism). Performative in the sense, that he’s interested in how SFP acts to enrol others. For instance, in the chapter ‘From the Luminiferous Ether to the Boeing’, we hear how firms have to decide whether there is going to be a technological revolution in laser gyroscopes. The risk of being left behind (they believe) is too great not to switch technologies. Therefore many invested heavily, helping the laser gyroscope revolution to become a reality. In the example of nuclear weapons, he might have emphasised just how such SFP’s are constructed and maintained. This could be a description of what “interests” or actors had to be brought together under the same roof to effect such a feeling of inevitability. In a way, it is an obvious point, but I think if we’re to really make use of the notion of SFP, there is maybe a danger in pretending that this is a characteristic of all technologies. When in actuality, such a notion is better thought of as an achievement, rather than as a property of technologies themselves, and in turn, (to be symmetrical) such effects need to be thoroughly described. For example, I like the way MacKenzie starts to outline the possibility of the uninvention of nuclear weapons. Importantly, he begins to describe the initial crumbling of just ‘what’ makes the notion of proliferation inevitable: “The military situation has changed, budgetary constraints have tightened, and parts of the nuclear weapons production infrastructure in both the United States and the former Soviet Union are now, either closed or in physically dangerous condition. Add to this a possible ban on testing and it is far from clear that the governments of the major nuclear states will commission new types of nuclear weapons in the foreseeable future” (p257).
In a further chapter entitled, ‘The Charismatic Engineer’, MacKenzie returns the discussion to a subject more familiar to the sociology of technology. That is, in describing the work of Seymour Cray, the engineer whose name has become associated with the invention of the supercomputer, he seeks to delineate some of the heterogeneous work of one of the worlds most distinguished technologists. From Pasteur’s microbes, Bakelite’s plastic to Edison’s electric light, through to talk of charisma, brilliance, leadership, and of great inventors, others have taught us that such attributes are ‘effects’. Invention is the bringing together of many resources and building of heterogeneous networks. Brilliance, invention or great leadership is in placing oneself at the front of these networks. MacKenzie brings another angle to this in his discussion of Cray’s work. Cray, it seems, is a charismatic. He has something special. MacKenzie teaches us of his obsession with control, a fixation that won’t even allow him to transfer work to others, for fear that they won’t be able to deliver on their promises. We hear about his determination to succeed, to make his computers work, every time with more speed. Often his attempts to overcome problems seem so ludicrous, that nobody, not even his colleagues will take him seriously. After hearing one particular suggestion they ‘laughed’ and “rolled in the aisles” (151).
MacKenzie tells us of Cray’s continuous efforts to build the world’s fastest computer, to continually attempt to improve on their speed and, in doing so, he, again and again, places himself at the intersection of what could be described as two contrasting networks. The first is a more stabilised result of previous efforts, where we see the customers of Cray’s machines demanding hardware modifications, software and end-user support, for their newly purchased machines. The second is a more perilous journey with much uncertainty, where to reach his goal of increased speed he continually has to rely on his attempts to enrol the technology and the support of his colleagues. Rather than take the safe route of providing services that his customers need, a route that would provide financial security for the company, he forsakes the security offered here and (again and again) pushes for the second route. In envoking the work of Weber, MacKenzie captures just what is different about Cray:
Just as revelation and the sword were the two extraordinary powers, so were they the two typical innovators. In typical fashion, however, both succumbed to routinization as soon as their work was done…[R]ules in some form always come to govern…The ruler’s disciples, apostles, and followers became priests, feudal vassals and, above all, officials… (Weber cited in MacKenzie, p134).
In the hands of MacKenzie, Cray, at every turn, sees his work, his genius, becoming “routinised” for the needs of capital (to gain a return on investment), for the firm (to build market share), for the customers (to receive increased customer support). In building networks, forming alliances and placing oneself at the front of this network, and when the network begins to stabilise, when there is routinisation, or there is little room for movement, the charismatic engineer can play a “…different role in the new network…” (157). Or, as in the case of Cray, he can “…cut loose from it and begin afresh” (157). This is what is different about Cray. At every turn, he tries to shake of the constraints that might hinder his search for an even faster computer. Cray is not a charismatic, but a single minded engineer, who in searching for that ever faster computer, both built and destroyed an empire.
In concluding, this book has much to offer science and technology studies. MacKenzie shows that there is no such thing as a recipe for doing sociology of technology (something hinted at in other recent collections (Bijker, 1995)). For MacKenzie there are only tools to be used for particular jobs. The critique, often directed at the sociology of technology, that our studies tend towards a political and moral indifference is directly challenged in this collection. Though, where I agree with MacKenzie that the absence of some groups, manual workers or women for example, from the design of technologies is an important, often not discussed issue, I feel a little uneasy about the direction in which he takes some of his analysis. For example, while his argues that a loss of tacit knowledge serves to describe the possibility of a nuclear weapons reduction on two levels, a decline in those who are able physically to make such weapons, and as fuel for his SFP argument, there is also the possibility that his own thesis can be taken up in different ways. In other words, and he notes the point himself, those who wish to continue down the weapons proliferation road may enrol such arguments, as evidence for increased investment in weapons building and testing. The problem of intervening in technology design is an important, and not easily resolved, issue for the sociology of technology. See, for example, the May (1996) edition of the journal ‘Social Studies of Science’ which features a special edition on the ‘politics of SSK’. In the collection, MacKenzie does not deal with the problems that such an intervention might bring, an aspect which is slightly disappointing, given that this theme runs implicitly throughout many of the papers.
Bibliography:
Bijker, W E, (1995), Of Bicycles, Bakelites, and Bulbs: Towards a Theory of Sociotechnical Change, MIT Press.
Van Lente, H, (1993), Promising Technology: the dynamics of expectations in technological developments, WMW-Publikatie 17, Universiteit Twente, Enschede, Netherlands.
author’s address: N.Pollock@Lancaster.ac.uk