Technopoetry

Review of: Review of Werner Rammert (ed.), Soziologie und Kuenstliche Intelligenz, 396 p, Frankfurt/New York: Campus, 1995; Bernward Joerges, Technik, Körper der Gesellschaft, 304p, Frankfurt: Suhrkamp, 1996; Barbara Czarniawska and Guje Sevon (eds), Translating Organizational Change, 284p, Berlin: De Gruyter, 1996; Bernward Joerges (ed.): Körper-Technik. Aufsätze zur Organtransplantation, 210p, Berlin: Sigma, 1996.

We need the power of modern critical theories of how meanings and bodies get made, not in order to deny the meanings and bodies, but in order to build meanings and bodies that have a chance for life.

-Donna Haraway [1]

What’s the connecting link between bodies, organs, intelligence (both natural and artificial), software, hardware, wetware, organizations, and virtual realities? Apparently we face a translation problem, a situation of Gestalt Switch, where a cultural resource - technology - appears in various masks and constantly changing roles. So it might be appropriate to begin this essay with a collection of articles on exactly this phenomenon - the translation of free floating ideas into solid artefacts.

Change makes things visible, because it makes things move, and “only what moves is visible”, as Barbara Czarniawska and Guje Sevon formulate the approach to their studies on organizational change (p1). The “language for describing social phenomena itself” reflects this situation. Change “invites the questioning and de-construction of the previous social order”. Which seems to be change itself - “change being the periods during which people begin questioning things that were previously taken for granted.” And “luckily for those studying organizations, change abounds” (p2). So we just have to lurk around until something moves, since reality obviously deconstructs itself - “as the social world undergoes constant construction, yet projecting a strong illusion of stability.” It might be an illusion, but the construction itself is not an illusion. The authors in this volume position the deconstruction in the field itself - as students of organization they just observe this unravelling of previously stable objects; it is not them, who deconstruct them. This is something that has always puzzled me (not just since the science wars). What exactly is the role of the student of science and technology? Is he or she the one questioning the solid facts and matters out there, or is ‘the out there’ itself doing this work of actually undermining meanings, bodies, and technologies?

‘Translating organizational change’ clearly stands for the first approach, and the various contributions to this book prove that assumption. Consider, however, that the visible-is-only-what-moves approach also has a blind spot: How do we make things visible, which do not move? I mean the solid things, the silent technologies underneath our social carpet, who work for us. Luckily for those building societies, things usually do not move. They are stable, they keep their properties, and they are very discreet black boxes. Yet they are in action, they are ‘travelling without moving’, to use a title from Jamiroquai. [2] To make their acting visible, it takes us, the students of science and technology, to make them speak, to lend them a voice.

Barbara Czarniawska and Bernward Joerges call this phenomenon the “materialization of ideas”, “turning ideas into objects and again into ideas” (p13). How does it happen that “ideas are turned into things, and then things into ideas again, transferred from their time and place of origin and materialized elsewhere again?” (p18) As a student of science, in his Wissenschaftslehre Max Weber gives a beautiful example of that. Weber draws a bold line from Italy to the Netherlands, tracing back the origins of modern experimental science to the garages of some tinkering Renaissance piano-makers: “It was the great achievement of the Renaissance that it established the experiment as the main principle of scientific research. The pioneers in this field have been the great innovators in the realm of art: Leonardo and the like, and especially the experimenters in 16th century music with their experimental pianos. From here the experiments travelled into science, which was especially the contribution of Leonardo and - in theory - Bacon; and afterwards it was adopted by the scientific disciplines at the universities of the continent, at first in Italy and the Netherlands.” Definitely a nice story, and a beautiful example of the metaphor of a travelling idea. How did the experiment leave the garage of Zarlino, headed north towards the Alps, crossed them, reached the Netherlands, and became part of the syllabus of science students? Per pedes, on the back of mules and in the form of transportable artefacts.

On the subject of travelling ideas we might look back to 1956, on the campus of Dartmouth College. A heterogeneous group of scholars met there for a social gathering that turned out to be the founding conference for a rather busy and noisy interdisciplinary crush of ideas, concepts, objects and artefacts. I speak of Artificial Intelligence (AI). The volume Soziologie und K］stliche Intelligenz, edited by Werner Rammert, contains a comprehensive collection of articles on the delicate relationship between these adversaries, suspiciously eyeing each other. Adversaries? Or even quarrelsome twins? Let’s take a step back, for the story is more complicated. First, the book demonstrates how sociologists observe AI. (It contains no contributions from the AI community - shall we call them the Artificial Intelligentsia?) Thus it’s more a reflection of sociologists on their own way of dealing with the challenge of AI. But why a challenge at all? Why does it seem that sociology is so intimidated by the quite cocky visions of AI?

With all due respect, AI certainly never was a very modest discipline. As Werner Rammert remarks in his introduction, this astonishing network of engineers and professional science barkers always has succeeded in turning its obvious failures to come up with anything useful into even greater promises of future benefits and breathtaking breakthroughs. In their studies of the very few existing expert systems in actual use, Nina Degele and Josef Wehner portray the rather poor general performance of this most prominent technological result of AI research. Even devastating reports on the results of decades of research in AI (like the British Lighthill Report) couldn’t really turn off the never ending warm rain of public money spent on this mollycoddle, as Petra Ahrweiler reports.

On the other hand, this seemed to irritate some social scientists with their mixed feelings of envy, curiosity and indignation at the audacity of AI to claim genuine sociological questions to be actually soluble technical challenges. As Rammert points out, if AI were to succeed with its model of rule-based information as the one and only mode of knowledge, then one of the central concepts of sociology indeed would be ruled out, that is, the social constitution and embeddedness of intelligence and knowing. Confronted with those challenges, first philosophy and then the social sciences reacted and mobilized their own troops, ready to reconquer those and other sociological titles. Although sociology got this job done very well (the articles in this volume are testaments), AI still holds a trump card in its hands, with which it might beat its sociological counterpart: the notorious Turing Test.

Since its popularisation in the ’80s, this intriguing test and its promised result - a sort of waterproof demonstration of machine-based intelligence (though just attributed) - have both fascinated and intimidated philosophers and sociologists. The story of the Turing Test is in itself an exceptional example of the narrative power of a travelling parable, an emblematic episode that seems to have just one implicit message, but actually allows to transport quite diverse meanings or Sinngebungen. As Bettina Heintz remarks in her contribution to this volume, the only way to make sense out of this squabbling between AI and sociology is to trace the way back to Turing’s original papers on this mysterious setting of a human and a machine fighting for a prize for the better social actor. If we carefully read his publications from 1936, the Chinese whispering of his test turns out to be a huge misunderstanding. As such it is nowadays one of the most prominent parables of Science and Technology Studies, since it seems to suggest the breakdown of the last and thus fanatically defended distinction between humans and machines, i.e., intelligence.

Bettina Heintz’s contribution, however, has good and bad news. The good news is that thus far no AI technology can do what humans can do, or at least not all of it, or at least not the best of it, or at least not so good, or at least not really, only by attribution through humans, who might do the same - attributing intention - with even more primitive creatures like dogs or IBM notebooks. Attribution of intention is quite a common habit among us, a working everyday fiction, so why the fuss? Does it shatter my human ego if I respect my cat’s personality? Perhaps because AI technologies are designed to be more than just those ad hoc attributions of specifically human competencies to nonhumans.

Anyway, the bad news is that a sort of reverse Turing Test might yield rather disconcerting results. As we all know, the Turing Test was designed to give a machine the chance to pass as a human. The price of the game is the label ‘human’. The test focuses on the machine player and the human interrogator who has to be betrayed and fooled by the machine. The part of the second human player is rather strange, as for him or her there is nothing to win. The actual competition happens between the human interrogator and the machine. Put a computer in a room, a human in another room, and connect their terminals with that of the human interrogator. This person may ask the two players various questions like ‘Do you like apple pie?’ or ‘What’s the square root of 4356?’ The interrogator has to find out which answers are of human origin and which are the product of a computer programme. If the interrogator misinterprets the answers, that is, if he or she imagines the computer to be human or the human player to be the machine, then the machine wins.

Oddly enough Turing doesn’t specify the motives of that third player. On whose side is he or she playing? What’s actually the role of this player? Overacting in a very human way, thus supporting the interrogator in fighting the presumptuous machine?

Usually we focus on the position of the human interrogator. This seems to be us, our situation at the end of the millennium, facing the challenges of machines invading our very human realm. A position of the ’30s, however, perhaps intellectually still fascinating, but actually not quite up to date. The astonishing popularity of Turing’s test will disappear as the question behind it (what is the final frontier, the genuine essence of humanity?) forfeits its cultural appeal and relevance. Perhaps we should switch our attention and intellectual curiosity now more to the third player, as this character has nothing to win or lose in this game. She might be an intermediator between the human and the non-human, part of the setting (which is a model of our modern situation), but not of the competition (which never really existed anyway), watching the exchange of properties between the human and the artefact. The two other players are equally hypothetical, artefacts of a philosophical question (or shall I say, a fear?), but no real actors. In my mind, the third player is a student of science and technology.

While Turing’s test might be a fascinating intellectual enigma, its sketched constellation is by no means a model for the basic socio-technical imbroglio of our ‘techno-society’. As Bettina Heintz writes, the raving society of modern technology plays a rather reverse game, though it has the same players. Let’s call this game the Taylor Test. The human player wins this game (and is awarded the Frederick W. Taylor Prize of technical machinations) if he or she manages to act just like a machine, which means: rule-based, repetitive, predictable, definite, trivial, etc. Just perform like a machine, and the interrogator doesn’t know what you really are, human or machine or whatever. To make it a little bit harder for our human player, the machine player probably will overact in an extremely machine-like manner, because again it’s the contrast between challenger and expert that gives the game its spice. And who of us wouldn’t pass such a test? Children and lunatics, maybe. It’s simply much harder for a machine to pass as a human than it’s for a human to pass as a machine. Bettina Heintz suggests to turn the Urszene of the Turing Test upside down:

What humans do when they follow an algorithm can also be done by a machine. That’s the so-called statement of Alan Turing which claimed the theoretical foundations of AI twenty years before its official date of birth. Turing’s formula explains to us why a machine can simulate a chess player or a psychotherapist (ELIZA), but not a simple chat in everyday life. (…) Turing has taught us that the important issue is not the performance of the simulating machine, but the behaviour of the simulated humans. The simulation of human action presupposes rule-based action. Whenever we behave like machines, machines can replace us - that’s the central point of Turing’s statement, formulated from a non-mathematical point of view (p40).

In the beginning, however, there was a productive misunderstanding. Turing never actually claimed that every human competence might be done by a machine. No, just those actions that follow an algorithm, actions that are already machine-like. Which makes it sound like a truism, basically. Though a very successful truism - see Frederick Taylor and the industrialization of human labor. However, we shouldn’t spend too much time on the philosophical debates around AI, as Heintz writes (“We don’t do research on philosophers of technology, we do research on real techniques, we take the empirical problem and run,” says Joerges (p270), quoting Karin Knorr-Cetina), but look at the familiar areas of the social construction of technologies, “regardless of whether we discuss Turing’s ideas in the context of the sociology of industrial relations, the sociology of technology or within the debate between AI and sociology. Instead of focusing on the machines, a sociology of the computer has to interrogate the mechanical parts of human action and their social conditions, beyond all metaphors.” (p53)

How did sociology pick up this idea of the AI community, that is, to create machine-based intelligence? It was not just a counterattack or a retaliation against an unfriendly take-over by a company who was perceived as a poacher in one’s own hunting ground. The initiated redemption of ‘intelligence’ as a genuine social phenomenon was also a chance to rediscover and update one’s own concepts of the debated issues in order to prove that sociology is still the first stop for questions about the social attributes of human life. Presently AI serves as an ‘evocative object’ for sociology; there is not much dialogue between AI and sociology, a situation that might change now with the recent cooperation between sociologist and some representatives of so-called Distributed Artificial Intelligence.

The interest in the processes of materialization of meaning is shared by all the contributors to these books as a persisting problem in the social sciences. Change might be a timeless pattern of societies, but globalization and the emergence of large technical systems of global scale are very modern phenomena. Moreover, “the question of how local action emerges and becomes institutionalised on a more global scale” (Czarniawska and Joerges, p13) points to a further specifically modern aspect of travelling ideas. In Weber’s picture the experiment laces up its shoes and crosses the Alps, but at the same time it stays in Italy, and is further developed in the merging local science cultures. Thus, techniques like copying, duplicating, ‘pseudo-imitating’, and especially translating become the focus of the empirical contributions to the volume On Organizational Change. Translation is a complex practise, however. Displacement carries its most distinguished aspects:

Ideas that have been selected and entered the chain of translations acquire almost physical, objective attributes; in other words, they become quasi-objects, and then objects. The simplest way of objectifying ideas is turning them into linguistic artifacts by a repetitive use in an unchanged form, as in the case of labels, metaphors, platitudes. This is an attempt at a reproduction, a mechanical translation, intended to minimize displacement effects (p32). Displacement, disembodying, delocalization, translation, and travelling - I see a common feature among these books as they all deal with attempts to leave behind the constraints of the human condition - that our mind is located physically in a body, that our bodies enact themselves locally, that time tends to freeze our actions in bureaucratic superstructures, and finally that our existence is bound to the fragile balance of this sort of alien nature inside us. Mind, knowledge, and intelligence might be technically disembeddable phenomena, our artificial mind-managers might come up with numerous copies of our cognitive potentials, but the artificial body-recycling-management known as transplantation medicine tries to disembody the body itself - organ-allocation just-in-time is quite a special case of travelling objects.

“This is your captain speaking. We will be arriving at Frankfurt Airport in a few minutes. There are a few other aircraft waiting in front of us, but since we have a kidney from Eurotransplant on-board, they have to wait for us. I hope you have enjoyed your flight.” A typical episode in the European transplantation system. Organs travel hundreds of miles to be delivered to their final body-destination. The book Körper-Technik, edited by Bernward Joerges, contains a number of articles on transplantation medicine. As Joerges mentions in his introduction to the book, transplantation medicine tremendously extended the traditional microsituation of patient, body, and physician. Transplantation medicine is organized as a network of processing and translating disembodied body parts all over Europe in order to re-embody this human material. The articles in this volume present a framework for studying this network by emphazising the function of the technologies that make it work. Although especially the contributions from G］ter Feuerstein and Ingo Braun prefer a very distant style of portraying this system of organ exchange, the collected articles share the idea of multi-authorship in the way they organize their findings. Actors from transplantation medicine and some additional ‘artificial voices’ are incorporated in order to present a polyharmonic dialogue about the delicate subject: “The openness of a dialogue between various voices in our texts should demonstrate (in opposition to the common practise in scientific writing, to close controversies in the field) that our book, in the end, couldn’t resolve this case” (p14). Linda Hogle’s contribution on her fieldwork at a transplantation site in the US demonstrates this openness in a very intriguing way.

But is that possible, can we avoid in our scientific writing to do the same thing with our objects of observation as the transplantation surgeons do with their transplants? We cut our objects out of their life-worlds, and transmogrify these reality bites into short stories, parables, cases, and data, thus translating them into transferable quasi-objects. Linda Hogle writes about the difficulties to explant and move parts of the human body as material objects. They still represent at least something of their former social host - the (in most cases) dead donor, thus perhaps threatening the identity of its recipient who is usually left alone with the task to re-embody the gift. In transplantation medicine “organs are literally changed into transferable objects, and that is only possible, if we manage to separate materiality and culture” (p189ff). Human bodies and their inner parts are always both - “at the same time they are cultural and material objects … That makes them into something in-between, they have an in-between status, neither ‘living’ nor ‘dead’, not totally ‘human’ and not entirely ‘technical’” (p186). Human tissue has that strange quality, which makes the procedures that reveal these hybrid qualities such beautiful opportunities for studies in the anthropology of science. However, it is a quality of all objects, including our own textual representations of them. We might distinguish matter and culture, but we can not separate them. The tricky thing in doing science and technology studies is that we always have to write about objects as if we could separate culture and matter. If only we could listen to matter itself, without writing technopoetry by lending them our voices.

In a collection of articles he published since 1979, Bernward Joerges uses the metaphor ‘body of society’ to formulate the inseparability of culture and matter. It might not have been very clever to suggest the tongue-twister prosopopoietic systems as the new approach to study technologies … However, according to the Encyclopaedia Britannica prosopop琶e is “a rhetorical figure by which an inanimate or abstract thing is represented as a person, or with personal characteristics …” (p265). Sound familiar? Right, there it is again, Turing’s game, but this time society itself is the playground, or what we are still used to calling society - the site where the symmetries and asymmetries of technology and society are enacted on the stage called nature. It might be a difference between the German and the Franco/Anglo-Saxon constructivism, that the former, as Joerges writes (p278), is more interested in detecting (or shall I say ‘critically defending’?) the differences, whereas the latter emphasizes the associations of hybrids, the imbroglios, the melting of the differences in what we might call - following Bruno Latour - techno-social collectives. Joerges raises the question whether we should really leave this problem to the philosophers of technology. If we don’t, how shall we empirically study these constructions of the difference, the Anderssein of nature? The transmogrifying procedure is the prosopopoiesis of technical systems: “Technical action is prosopoietic insofar as it makes something present, which is not present in the social. It makes it present, it incorporates and enrols it in the social, but at the same time in a different way, as something different. Due to the lack of a better shorthand symbol, we might as well call this absent being nature” (p283).3 It’s all about the (political) status of differences, how to respect them without using absolute categories. Constructivists seem to be obsessed with these differences, although it appears to me that constructivism in science and technology studies actually represents a project of flattening them. Following Joerges, the more hybrids of nature, social, technical and artificial are around us, the more it becomes important to follow the practises which seem to blur these differences. How to speak correctly about differences might become (or perhaps already is) the global etiquette of our techno- cultivating modern society. Maybe we simply forget or unlearn certain differences? And it might be the foremost job of constructivism in science studies to support the invention of new vocabularies for the time ahead.

Notes

1. The quotation is from Situated Knowledges: The Science Question in Feminism as a Site of Discourse on the Privilege of Partial Perspective, in: Feminist Studies 14 (3), 1988, 575-99. The title Technopoetry refers to the art exhibition Technopoetry - When Tekkno Turns To Sound Of Poetry, which was held at the Kunst-Werke, Berlin in the Summer of ‘95.

2. Or was it actually from Frank Herberts SF-novel Dune?

3. For those of you who can read and enjoy German science prose: “Technisches Handeln ist insofern prosopopoietisch, als in ihm sozial Abwesendem, man kann das in Ermangelung eines besseren Körzels auch Natur nennen, zur Anwesenheit, zur Mitwirkung verholfen wird, aber immer auch zu einer ganz bestimmten Eigenart, einem Anderssein. Technische Systeme, könnte man mithin sagen, sind prosopopoietische Systeme in dem Ma&slig;, in dem beides gelingt.”

author’s address: gerald@chem.uva.nl