Book Review |
Man as Machine
This is in many ways an imposing—not to say, daunting—book. The numbers are themselves impressive. The two volumes run to almost 1700 pages. The bibliography alone is 133 pages long. Many an academic author will wonder, somewhat ruefully, just how Oxford University Press was persuaded to publish something so gigantic.This perplexity is reinforced by the fact that it is by no means clear for whom this vast work is intended. It is decidedly not a popular exposition of the development of modern cognitive science aimed at a general readership. The book is too long, too technical and far too expensive for all but the few who have access to a university library. It is, however, no student textbook. Some historians of science and technology may feel obliged to try to read it—although few will have the stamina to stay the course. One is left to conclude that the community, or more accurately communities, of cognitive scientists whose disciplines are under discussion form the projected audience for the work. How many working scientists will find the time to peruse such a work is open to question. Many may, however, find themselves scanning the name index with interest in an idle moment.
Along with the question of who is to read it comes that of how the book is to be read. A work on this scale would normally be classed as an encyclopaedia or textbook that readers would sample or draw upon according to their interests. The author, however, explicitly precludes such an approach to the text. This is, she declares: ‘a historical essay, not an encyclopaedia’ (p. xxxiii). (Most historians, it might be noted in passing, impose a salutary word limit on their students’ essays.) We are also enjoined by way of preface:
My text ... holds together much as a woven fabric does. It's best read entire, as an integrated whole—not dipped into, as though it were a work of reference. Indeed I can't resist quoting the King of Hearts’ advice to the White Rabbit: ‘Begin at the beginning, and go on till you come to the end: then stop’. (p. xxxiv)Only in Wonderland, one suspects, are there many White Rabbits who are likely to heed this somewhat fanciful advice.
Margaret Boden is a distinguished cognitive scientist whose own career spans many of the decades with which the book is chiefly concerned. Mind as Machine therefore belongs to a venerable tradition of histories of scientific disciplines written by practitioners in the field concerned. The history of this genre is itself a subject of some interest. Scientists have been moved to write the history of their subject area for a variety of reasons. Sometimes they feel that it may help to clarify issues of contemporary importance by placing them in an historical context. Or there may be a sense that the contribution of certain individuals or schools needs to be reassessed and assigned its just value. But often practitioner history is overtly polemical in character: the past is invoked to further some cause or argument in the present.
Boden's gives no extended account of her goals in writing this book. She does declare her conviction that: ‘cognitive science today—and, for that matter, tomorrow—can't be properly understood without a historical perspective’ (p. xxxiii). Quite how this historical perspective will serve to illuminate the present state of the field is not, however, made clear. A determination to set the record straight is evidently at work in places. A good deal of space is, for instance, devoted to demonstrating that the work of Charles Babbage should be considered as: ‘superficially close, yet fundamentally irrelevant—to cognitive science’. (p. 131)
The most overtly polemical sections are those that deal with the theories of Noam Chomsky. Boden—seemingly inspired by the historiographic style of W.C. Sellar and R.J. Yeatman—concedes that: ‘for the field as a whole, [Chomsky] was a Good Thing’ (p. 591). But she refuses to accept what is described as ‘the tenfold Chomsky myth’ promulgated by those ‘who uncritically take Chomsky as their scientific guru’ (p. 592). This is a case where Boden's competence as an historian allegedly qualifies her to speak with some authority: ‘it's precisely because I’d done the intellectual history carefully that I’d realized just how questionable, and in certain respects just how damaging, some of Chomsky's claims are’ (p. 594). There is a suggestion that Chomsky has allowed his scientific views to be unduly influenced by his political commitments. Moreover, his ‘personality which thrives on embattlement’ (p. 650) is also to blame for some of the damage he has inflicted. Chomsky, unsurprisingly, has reservations about his: ‘assigned role as the demon who almost destroyed the field’, and is quite up for some polemic of his own (Chomsky, 2007
).
Professional historians of science tend to deprecate practitioner history for its supposed naivety and lack of sophistication. Boden has, however, made an effort to take account of some of the recent trends in the history, philosophy and sociology of science. Many scientists have at least a passing acquaintance with Thomas Kuhn's The Structure of Scientific Revolutions. But Boden also refers to the work of sociologists and historians such as David Bloor, Bruno Latour, Steven Shapin and Simon Schaffer. She somewhat indiscriminately lumps these authors together as belonging to a ‘social constructivist’ school of thought that rejects: ‘realist accounts of science, and objectivist accounts of truth’ (p. 21). Boden makes her own position on these issues quite clear, declaring that: ‘I regard the constructivists’ position as fundamentally irrational, even though they did have some important insights—which need to be remembered if one is to understand the history of this, or any other scientific field’ (pp. 21–22).
She ‘opts’ to reject such ‘constructivist’ approaches in the history of science, in part, because she sees them as incompatible with the notion of scientific progress. Boden maintains that in the history of cognitive science that she recounts ‘scientific advance’ is everywhere evident. In her view, while much controversy remains in the field of cognitive science, much ‘has been found out’. She therefore makes no apology for: ‘the strong whiff of Whiggism that attends many parts of my text’ (p. 20). The reader has, at least, been warned. A more reflexive and critical consideration of such terms as ‘progress’ and ‘scientific advance’ would, however, have enhanced Boden's narrative.
At the same time, Boden sets out to give a more nuanced account of how such advances have been made than is found in the ‘Legend’, or creation myth, that scientists often spin to explain the development of their field. In particular, she insists that: ‘the disinterested pursuit of truth is rarer than many choose to believe, and the conceptually innocent ditto [sic] is impossible’ (p. 26).
The positive corollary of this revision of the Legend is that an adequate account of any scientific discipline needs to be sensitive to the interests of those who cultivate it. It also implies that any history predicated on the assumption that science proceeds by means of the discovery of facts, unmediated by any kind of conceptual framework or set of presumptions, is hopelessly naïve. An adequate account of the history of a scientific discipline must, therefore, pay close attention both to its social and cultural context and to broader contemporary intellectual movements and tendencies.
Boden, however, declares at the outset that she does not ‘explore socio-political influences at any length’ (p. xxxiv). She is somewhat more interested in how cultural trends—or ‘Zeitgeist’—might have impacted on her subject. Thus, she maintains that: ‘In the second half of the twentieth century ... counter-cultural ideas and values had a significant effect on the history of cognitive science’ (p. 36; see also pp. 1042–1043). But for the most part, this is a resolutely internalist, largely narrative, history.
In a chapter on the ‘origins’ of the idea of mind as machine, Boden seeks the roots of cognitive science in antiquity and follows the story into the 17th century. Predictably, she pays special attention to the work of René Descartes—whom she tentatively identifies as ‘the ancestor nerd’ (p. 68). Such flights of fancy occasionally enliven the text. Boden also discusses the Kantian and neo-Kantian ‘alternative’ to the Cartesian worldview, seeing a recurrent dialectic in the history of cognitive science between these two competing understandings of the mind. The 18th and 19th centuries are also scanned for anticipatory signs of an emergent cognitive science—often with digressions on topics that may seem of only tangential relevance.
But all this is by way of an extended preamble to what is Boden's main theme: the emergence in the later 20th century of the notion that the ‘mind’ could be understood as a machine. The idea of ‘Man the Machine’ was not new. In the middle decades of the 20th century, however, the view that mental processes could be understood in purely mechanical terms became for the first time widely disseminated. The sources of this conceptual leap lay in technological innovation, and above all in the invention of the first digital computers. Every era had its favoured metaphors for the understanding of the workings of mind and brain, which were often derived from prominent technologies of the day. In the 19th century, for example, the operations of the brain were sometimes compared with a telegraph system. By the early 20th century, the telephone exchange provided an alternative figurative resource (Otis, 2001). After 1945 these earlier models for cognition were replaced by the dominant metaphor of the mind as a computing and/or cybernetic mechanism. Boden insists that: ‘even merely thinking "metaphorically" about the mind in terms of computers helped psychologists to think clearly’ (p. 308). In due course, however, such ‘mere’ metaphorical thinking led to serious attempts to model mental processes in computational terms.
Although it was possible to attempt this without reference to any physical substrate, full implementation of the mind as machine concept required more; it called for embodiment. From the mid-1950s efforts were made to map such concepts onto the nervous system—in short, to find the ‘hardware’ that underlay these mental computations. Boden is unabashed in claiming that ‘computational neuroscience’ existed before the name (p. 1113). These efforts gained considerable impetus by the revival of connectionism in the 1980s. Connectionist ideas of how information could be stored, organized and used to guide behaviour were more easily applied to neuronal nets than the symbolic processing approach of what Boden calls Good Old Fashioned Artificial Intelligence (GOFAI).
Boden's claim that the machine metaphor has become the prevailing resource for understanding mental processes is open to criticism. One might, in particular, question just how central the computational metaphor has been to the neurosciences. Moreover, even her view of the scope and concerns of cognitive science itself betrays a perspectival bias. Boden does acknowledge the challenges and criticisms that the computational understanding of mind has aroused and still provokes. While some have welcomed the notion of the mind as a computer as—perhaps counter intuitively—liberating, other psychologists found the computing metaphor a stifling influence that ignored questions of meaning, intentionality and responsibility. None the less it is beyond question that in the period following World War II the notion of mind as a machine has proved a powerful resource and instigator of numerous research programmes. Boden devotes more than 1000 pages to detailing these developments.
The reader is often at risk of being overwhelmed at the weight of detail provided. Moreover, while we are told much of the how of modern cognitive science, the why is more elusive. Boden does offer some hints as to why the Artificial Intelligence approach to understanding mental processes arose when it did and flourished for as long as it has. Although she eschews the detailed consideration of ‘socio-political influences’, she is not entirely oblivious to the context in which her main actors operated.
She is, for instance, sensitive to the part that the demands of the military have played in breaking down the distinction between man and machine. To quote one of Boden's section headings, Artificial Intelligence was ‘nurtured in war’ (p. 825). Moreover, the fortunes of cognitive science as an academic discipline have, especially in the United States, been directly correlated with the amount of money that the field could attract from military budgets. Military planners have a vested interest in the emulation and enhancement of natural sensory-motor systems through the creation of evermore sophisticated cyborgs. One might add that commerce and industry have, since the middle of the 20th century, also come to demand mechanisms that can either replace human agents or with which humans can interface in order to boost performance in a wide range of tasks. For all its length, however, the book lacks any sustained development of such suggestive themes.
I began this review by wondering for whom Mind as Machine was written. Different readers will no doubt assess its value and derive benefits in divergent ways. For the historian of science, the book is perhaps best viewed as a primary source, providing the particular perspective of someone who was herself involved in the development of the field. Some of the most interesting passages—such as Boden's account of her first encounters with the concept of Artificial Intelligence in a Cambridge orchard—are also the most personal. Historians with an interest in the sciences of mind will also find the massive bibliography that graces this book a valuable resource.
Wellcome Trust Centre for the History of Medicine
University College London
United Kingdom
References
Chomsky N. Symposium on Margaret Boden, Mind as Machine. Artif Intell (2007) 171:1094–103.[CrossRef]
Otis L. Networking: communicating with bodies and machines in the nineteenth century. (2001) Ann Arbor: University of Michigan Press.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||