From ‘Mechanisation of Thought Processes’ to ‘Autonomics’

НазваниеFrom ‘Mechanisation of Thought Processes’ to ‘Autonomics’
Размер1.18 Mb.
  1   2   3   4   5   6   7   8   9   ...   29
Enclosing the Field

from ‘Mechanisation of Thought Processes’ to ‘Autonomics’

David John Clark

A thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science

University of Warwick

Department of Computer Science

September 2002

reprinted with corrections

February 2003

for A. L. R.

Un libro scritto...


Acknowledgements iv

Abstract v

Abbreviations vi

‘MacKay’s Photo’ Cambridge 1952 vii

0. Enclosure 8

1. Prolepsis or, What is a Programme? 11

2. Bildung 34

3. Ratio 37

4. ‘Autonomics’ 37

5. ‘Interlingua’ 37

6. logon techne or, What is a Program? 37

6.9.0 coda 37


A1. Ross Ashby’s Topics 37

A2. Meetings of the Ratio Club 37

A3. The Symposium 37

A4. MoTP58, list of attendees 37

A5. rhetorike 37

Bibliography 37


Firstly, to all those forgotten influences, for all those neglected sources: thank you. But definitely not to be overlooked:

at Warwick: Steve Russ, Martin Campbell-Kelly, David Miller, Greg Hunt, Ian Morley, Annette Rubery;

formerly at NPL: David Yates, the late Donald Davies, Alex Andrew, Mike Woodger, Sylvia Chantler;

elsewhere: Margaret Boden, Jack Copeland, Chris Darwin, Michael Elstob, Donald Gilles, Richard Gregory, Rhodri Hayward, Christopher Longuet-Higgins, John McCarthy;

also archivists at: Cambridge University Library, Public Record Office, Wellcome Library, Bodleian Library, National Archive for the History of Computing, Science Museum.

Postscript: February 2003. This reprint incorporates minor corrections suggested by the examiners: Jon Agar and Aaron Sloman.


It is always difficult to perceive something new except upon the model of what is already known. The result is a partial and incomplete understanding that can only be improved by experience. Our present perception of computing remains strongly influenced by its origin as a mathematicians’ machine. A notion of mathematics as the arbiter of what constitutes a ‘science’ has long been ascendant. As a result, even when viewed from a more empirical standpoint, computing has been influenced by the aspirations of engineering disciplines to acquire the status accorded to a science. Yet there always has been an interest in computing machinery from a different perspective: that of biology—autonomic mechanisms, understood through observation rather than the proving of theorems. But this runs counter to a prevailing trend in twentieth century biology toward a discrete and computational molecular model.

The story of the Autonomics Division of the National Physical Laboratory between 1957 and 1966 provides a particular instance of these tensions. It coincides with a critical point in the history of computing: the time when a distinct discipline of ‘computing science’ took form. The field thus enclosed represented the dominant mathematical and engineering interests, the biological view surviving only in cognitive science.

But these models have not proved adequate for the study of what is distinct and unique to this new science: software—an amorphous concept, at once both intangible and yet, it seems, an engineered product. To resolve this paradox it is suggested that a model of software as a form of literature needs to be developed.


ACE Automatic Computing Engine. (at NPL)

Add8353 Cambridge University Library, Manuscripts Collection: Letters and Papers of Sir Gordon Sutherland

AVIA PRO: Records of the Ministry of Aviation etc.

BT PRO: Records of the Board of Trade etc.

CME NPL: Control Mechanisms and Electronics Division (renamed in 1960 the Autonomics Division)

DSIR PRO: Records created or inherited by the Department of Scientific and Industrial Research, and of related bodies

GC/179 The Wellcome Library for the History and Understanding of Medicine, Archive: Bates, John AV, and the Ratio Club

MIT Massachusetts Institute of Technology

MoTP58 Symposium on the Mechanisation of Thought Processes, NPL Teddington. November 1958

MT Machine Translation

NAHC National Archive for the History of Computing (Manchester)

NPL National Physical Laboratory

NRDC National Research and Development Corporation

PRO Public Record Office

RAE Royal Aircraft Establishment (at Farnborough)

RRE Radar Research Establishment (at Malvern, formerly TRE)

TRE Telecommunications Research Establishment (at Malvern, later RRE)

[Photograph, landscape 25x17cm]

‘MacKay’s Photo’ [Cambridge 2/3 May 1952 ?] From left to right; standing: Giles Brindley, Harold Shipton, Tom McClardy, John Bates, Ross Ashby, Edmund Hick, Thomas Gold, John Pringle, Donald Sholl, Albert Uttley, John Westcott, Donald MacKay; sitting: Alan Turing, Gurney Sutton, William Rushton, George Dawson, Horace Barlow. (Wellcome Library, London)

Chapter 0. Enclosure

Some of the history of a subject may be revealed in its bibliography. The frequent reference to certain authors or publications, whether by way of relevance or reverence, has something to say about what is significant in that particular field. The bibliography of Artificial Intelligence does not lack references to the published proceedings of the Symposium on Mechanisation of Thought Processes that was held at the National Physical Laboratory in Teddington, SW of London, in the autumn of 1958. The particular papers referenced will vary according to preferred flavour: McCarthy’s ‘Programs with Common Sense’, Selfridge’s ‘Pandemonium’ are but two of the most frequently cited; there were also papers from Marvin Minsky, John Backus and others—some less familiar to the student of computing’s history. Yet, while histories of computing never fail to mention the meeting at Dartmouth in the summer of 1956 at which ‘Artificial Intelligence’ is said to have been founded, nor do they neglect the volume Automata Studies, edited by McCarthy and Shannon and published shortly before, the 1958 Symposium is rarely mentioned. The same contributors appear—and more besides. Such gatherings were as rare in 1958 as in the years before, yet history and the bibliographical record seem to differ. A curiosity—this research began as an attempt to discover why.

But this inquiry led to a broader question: what is computing science? Or rather why, when the evidence of that 1958 symposium attests to a very broad interest in the intellectual challenge of computing, should computing have developed to be for the most part an amalgam of mathematics and engineering? To point to the origin of computing machinery in the mid-twentieth century, as a tool for applied mathematicians and as an engineered artefact, will not suffice. The new machinery, both its potential uses and the philosophical questions raised, attracted wide interest; and not all from an ill-informed public dazzled by ‘giant electronic brains’.

A new phenomenon is perceived, explained, and assimilated by analogy with what is already known. A restricted range of analogues, a poverty of models, may constrain the understanding that is eventually achieved. When, through familiarity, the computer ceases to be ‘a sort of...’ and is just an everyday object, the place of that object in our world may depend on the route by which we have come to understand it.

In the 1940s and 50s there were other models of computing and automated machinery, analogies that took their inspiration not from engineering, physics, or mathematics, but from biology; from empirical observation rather than a priori reasoning; from a study of creatures in the world, doing and being done by, rather than commanded by instructions in code. This approach retains a following, indeed it shows signs of renewed interest, but it is peripheral to what presently constitutes computing science; cognitive science is, perhaps, more a sub-field of psychology than computing. And new fields such as bio-informatics are certainly not in the tradition of biology: rather, they show the adoption (some might say usurpation) of the reductive and discrete traditions of physics by the life sciences.

If we talk of an enclosure of computing science, of its assimilation into a predominant scientism—reductive, mechanical, the purity of its science assayed by the density of its mathematical content—then the traditions of natural history are not the only exclusion. There was, for example, from the beginning, a significant interest in computing applied to natural language. This has had less lasting influence; unlike genetics and molecular biology the enthusiasm for machine translation did not lead to a notable success for a mathematico-logical approach. But what of a counter influence? At the time of the Teddington symposium, there were significant developments in programming languages, Fortran, Cobol, Algol, and Lisp, all represent significant strands in the move toward high-level languages and conceptions of a virtual machine. That is, a machine instantiated by an effort of imagination. Yet though we talk of programming languages, the analogy of software-as-fiction appears to have attracted little attention. By and large, we study the language of programs in a manner which the experience of machine translation might have led us to distrust. Though programs are written, and the text enacted, the analogy of a literature of programs seems to have attracted only passing attention.

In 1958 then, there was an open field: computing as an intellectual challenge knew no disciplinary boundaries. Within ten years, there was a distinct discipline of computing science, yet one that seems to lack a ‘cognitive content’ of its own. A study of computing machinery is engineering, a study of algorithms, mathematics: what is unique to computer science? The answer I suggest is software, but not as the coding of an algorithm, nor the specification of a virtual machine; outside the field enclosed by engineering and mathematics, there is not only a natural history but also a humanist perspective: a study of the virtual worlds conjured up by programming.

In chapter one we will consider the problem of what we see when we encounter a thing for the first time. In chapter two, the origin of the deep cultural bias that favours the abstract, the ‘useless’ and mathematical knowledge. The third chapter looks at the computing and automation in the biological tradition. Chapter four is an account of Autonomics at NPL, a research programme that owes much to that biological influence. Examining the context of one project in that programme, machine translation, the dominance of a mathematico-logical model is evident. Finally, in a contrasting view of the nature of language, chapter six attempts to trace the gradual coming into consciousness of programming as a literary activity.

This work thus falls into three parts. The first two chapters are concerned with theory, a history of ways of seeing. The second and major part, which extends from chapter three into chapter five, is a history of a particular perception of computing and computers. The final part, taking chapter five forward into six, retraces that history of computing to detect the emergence of yet another perception, one whose status—mainstream or meander—is as yet undecided.

Part I

Guildenstern: A man breaking his journey between one place and another at a third place of no name, character, population or significance, sees a unicorn cross his path and disappear. That in itself is startling, but there are precedents for mystical encounters of various kinds, or to be less extreme, a choice of persuasions to put it down to fancy; until—“My God,” says a second man, “I must be dreaming, I thought I saw a unicorn”. At which point, a dimension is added that makes the experience as alarming as it will ever be. A third witness, you understand, adds no further dimension but only spreads it thinner, and a fourth thinner still, and the more witnesses there are the thinner it gets and the more reasonable it becomes until it is as thin as reality, the name we give to common experience. … “Look, look!” recites the crowd, “A horse with an arrow in its forehead! It must have been mistaken for a deer.”

Tom Stoppard Rosencrantz and Guildenstern are Dead

Chapter 1. Prolepsis or, What is a Programme?

Commencing a lecture on Faraday’s Lines of Force in 1873, James Clerk Maxwell invited his audience to look at a phenomenon at once both familiar and yet mysterious in the singular, and at that time still unorthodox, mode devised by Faraday. He forewarned his listeners:

In order to do so I must have the indispensable help of your own best powers of thought. If I had merely to describe to you some new discovery in science, I should be able to avail myself of your previous knowledge as a foundation, and to erect thereon a representation of the new fact which you were to place beside those old ones which you knew before. The greater your previous knowledge, the easier would be my task. But what I have to do is something quite different. I have to shew you the facts with which you are already acquainted in a light of a different character from that which the most illustrious philosophers have shed upon them—a light which the wisest among them would probably have avoided as deceptive and misleading had it been presented to him in his own time, because the slow yet steady progress of science has only in more recent times prepared us for its reception.1

Maxwell reminds his audience of the counter-intuitive notions of Newtonian theory: objects do not move only if pushed, our motion is not always accompanied by the sensation of motion. “It is impossible to overestimate the influence which the experience of smooth sailing has had on the minds of men in enabling them to get rid of these habits of thought.”2 In this way understanding depends on general concepts and experience that may not be universal. Faraday, self taught, had a limited knowledge of mathematics and thus insulated from the mathematical treatises of Poisson and Ampère “was obliged to explain the phenomena to himself by a symbolism which he could understand, instead of adopting what had hitherto been the tongue of the learned.”3 Maxwell’s influence and contribution to the mathematicising and quantifying tradition of physical science—that worldview in which a computing machine is a desideratum—will be considered in a later chapter, here my concern is the problem of the understanding or even merely the perception, of phenomena encountered for the first time.

Another instance is presented by Umberto Eco:

Often, when faced with an unknown phenomenon, we react by approximation: we seek that scrap of content, already present in our encyclopaedia, which for better or worse seems to account for the new fact. A classic example of this process is to be found in Marco Polo, who saw what we now realise were rhinoceroses on Java. Although he had never seen such animals before, by analogy with other known animals he was able to distinguish the body, the four feet, and the horn. Since his culture provided him with the notion of a unicorn—a quadruped with a horn on its forehead, to be precise—he designated those animals as unicorns. Then, as he was an honest and meticulous chronicler, he hastened to tell us that these unicorns were rather strange—not very good examples of the species, we might say—given that they were not white and slender but had “the hair of the buffalo” and feet “like the feet of an elephant”. 4

Rather than add to a catalogue of the vivid, Marco Polo modified the bestiary that he and his audience already held in common: he “corrected the contemporary description of unicorns, so that if they existed, they would be as he saw them and not as the legend described them.”5 Eco develops his theme by considering the duck-billed platypus; a ‘water mole’ to the first Australian settlers it took over eighty years to create a secure place for it in a taxonomy that had never envisioned possibilities other than mammal, bird or reptile. Such technical disputations contribute to the shaping of a discipline. The platypus “had set itself athwart the path of taxonomy to prove its fallaciousness.”6

Now, is this just a problem of composing a description or illustration in terms that will make sense to someone who was not there, or is it the perception of the witness that is constrained by experience? Did they actually
  1   2   3   4   5   6   7   8   9   ...   29


From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconFrom ‘Mechanisation of Thought Processes’ to ‘Autonomics’

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconEnclosing the Field from ‘Mechanisation of Thought Processes’ to ‘Autonomics’

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconEnclosing the Field from ‘Mechanisation of Thought Processes’ to ‘Autonomics’

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconEnclosing the Field from ‘Mechanisation of Thought Processes’ to ‘Autonomics’

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconA thought would be necessary that would no longer be constructed as a relation linking the thinker to the thought, or in this thought a relation without

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconHistory of geographical thought and research methodology unit-i thought

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ icon2. 12. Discussion: Innovation processes and creative problem solving processes as a part of the service design activities 39

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconOnno Boxma, Department of Mathematics and Computer Science Eindhoven University of Technology, Two-dimensional workload processes and two-dimensional insurance processes II

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconM odern Political Thought

From ‘Mechanisation of Thought Processes’ to ‘Autonomics’ iconThe History of Management Thought

Разместите кнопку на своём сайте:

База данных защищена авторским правом © 2014
обратиться к администрации
Главная страница