Summary

Our thesis is that all technical artefacts, from stone tools to cars to computers, are "enactive interfaces" that mediate the structural coupling between human beings and the world they live in, and hence bring forth a particular world of lived experience. Thus, the perspective of enaction provides a "middle way" that overcomes the twin pitfalls of a) underestimating the social significance of technology considered as a mere means to a predefined ends ; and b) overestimating the role of technology, by considering that "machines" are animate entities with a will of their own that can dominate mankind.

1. Introduction

The first three texts in this Virtual Workshop have dealt with fundamental theoretical aspects of “enaction”. The aim of the present text is to introduce the question of the significance of the concept of enaction for engineering and engineers.

The basic scheme for considering enaction is the dynamic sensory-motor coupling between an organism and its environment :

Figure 1

The sensory inputs, S, are used to guide the actions A ; the actions A modify the environment and/or the relation of the organism to its environment, and hence modify in return the sensory input. What the world “is” for the organism amounts to neither more nor less than the consequences of its actions for its sensory inputs – what O’Regan and Noe call the “sensori-motor contingencies” ; and this in turn clearly depends on the repertoire of possible actions. Without action, there is no “world” and no perception.

2. Technical Artifacts

The basic scheme of Figure 1 applies already to animals. In the 1920’s the German ethologist von Uexküll characterized “animal worlds” (for example, “the world of the tick”) on the basis of sensori-motor contingencies as they function in ecological context. One of the major characteristics of “human worlds” is that the sensory-motor coupling is mediated by technical artifacts :

Figure 2

For animals, the sensory organs S and the motor organs A are fixed (within any given species), and they are attached to the body. For humans, the mediation of sensory-motor coupling by technical artifacts introduces two radical innovations.

Firstly, the range of possible sensory inputs and the repertoire of possible actions is greatly increased, without any limits other than the invention and fabrication of new artifacts. This is clear for the new possibilities of action which are created by tools, from hammers and screwdrivers to power-tools of many sorts. It is also clear for instruments – microscopes, telescopes, microphones, infra-red detectors, radios and so on resulting in sensory inputs which are strictly impossible without the devices in question. An interesting special case is the sensory substitution devices. More generally, but less obviously, technical artifacts organize sensory experience: think of the world of skier, which is impossible without the artifact. Even when we are not actually skiing, our perception of the mountain is determined by the possibility (i.e. virtual action) of skiing and the correlative sensations. So this first point can be understood more profoundly: in case of contemporary humans, there hardly any “natural” perceptions or relations to the world: our sensory-motor coupling is always fashioned, at least virtually, by technical artifacts.

Secondly, technical artifacts are not irremediably fixed to the body. More precisely, technical artifacts exist in two “modes” : “in hand” and “put down”. When a technical artifact is “in hand”, being used, it becomes a prosthetic extension of the body ; correlatively, the artifact disappears from consciousness, and the attention of the human subject is focussed on the “world” that comes about (think again of the “world of the skier”, for example). Artifacts, like the body, are normally “transparent” to the subject; as Heidegger has pointed out, they are only noticed when they are dysfunctional (a wobbly hammer or a twisted ankle). However, unlike biological organs, technical artifacts can also be “put down” : separated from the body, they can now become objects of attention. In this mode, they can be invented, fabricated, repaired and so on. The back-and-forth movement between these two modes explains the radical innovative potential of technical artifacts. Over several thousand years, and at an ever-increasing rate, technical innovation has radically transformed what the world is for human beings.

3. A categorization of technical artifacts

Before going further, it is useful to give a more complete categorization of technical artifacts, which can be roughly divided into 3 types. The first type, which we have used to introduce the question in the preceding section, can be called “extensions of the body” : tools and sensory instruments. But there is also a second type of artifact, consisting of deliberate modifications of the environment : roads, buildings, fields and so on. It is even more obvious that this second type of artifact also modifies the world that human beings live in. Finally, there is a third sort of artifact that we can call “semiotic artifacts”. Here, the “actions” consist in emitting signals, and the sensory input is specifically geared to the reception of these signals. If the conditions that trigger the emission of a signal and the response of the receiver are appropriate, this leads to a co-ordination of actions, and constitutes the basic form of communication.

All three modes of sensori-motor coupling exist in rudimentary form in animals ; the specifically human innovation is the introduction of technical artifacts to mediate the coupling. Concerning semiotic artifacts, the human inventions are : first of all, language itself; and then a whole series of clearly technical inventions, writing, printing, and in our era computers. Computers are, of course, particularly important for our “Enactive Interfaces” network ; it is important to note that they are not only semiotic artifacts, but also sensori-motor devices. The computer comprises a certain repertoire of real actions (from punching cards in the early computers to mouse movements, joysticks etc.) with, in return, an increasing range of “sensory inputs” (from reading printed output to visual patterns, sounds and so on); regularities are established between action and sensation in this case just as for the first type of artifacts.

This categorization can be useful for analytical purposes; but it is important to note that in practice, technical artifacts do not function in isolation from each other, but form technical systems with a synergy between these three types. For example, roads (type 2) go together with cars and lorries (type 1), their synergy being organized by maps and plans (type 3). A possible use of the term “technology” (techno-logos) is to designate the situation where there is linguistic communication about the design, fabrication and use of technical artifacts.

4. The social dimension

The fact that technical artifacts exist in the mode of being “put down” has an important consequence : the persons who design and make technical artifacts are, generally, not the same as those who use them. Thus, technological development goes together with a division of labour and, correlatively, the development of mechanisms of social synthesis (exchange, market economies) which organize the integration of technical systems as functional wholes.

Traditionally, social and political science (with the exception of Marx) has not paid much attention to technology, which is usually considered as a “black box”, as intrinsically neutral means to pre-defined ends. The approach outlined here leads to a new perspective in which technology occupies a central position. The work of engineers has immense social significance because, in fine, the choices of technological devices and systems fashion the human condition itself, by constructing the world that human beings live in, and particularly by manufacturing interfaces that change the means of action, and influence sensations. Thus, in our society, any really serious political debate necessarily involves debate on technological choices.

5. Polemical consequences

The preceding considerations may not seem particularly controversial, but they have some controversial consequences. The term “Interface” appears in the title of our Network of Excellence, and is thus clearly of central importance. However, the term itself is the vehicle of an ambiguity that requires clarification : “interface” between what and what?

As we understand it, the term “interface” is properly used as the interface between an organism (human or otherwise) and its environment. Thus, the basic “interfaces” are the biological sensory and motor organs (section 1) ; for humans, technical artifacts are extensions to these basic interfaces, but they remain interfaces (section 2). New technical devices constitute new “worlds” : think for example of the “world of the car-driver”, or the “world of the skier”, or the “world of the violinist”. But note this : we do not talk about the “interface” between the man and the ski (or car or violin) ; the ski is the interface between the man and the snowy mountain, or better still between the skier and the “ski-ing world” that is brought forth.

Does this change in the case of computers? Our point of view is that computers are basically technical devices, and should be treated in the same way as other technical devices. Certainly, they are devices of a special sort, and the “worlds” that are brought forth when a human being uses them are a special sort of “world” ; but the interaction that occurs (that is mediated by the machine) is between the human being and this “world”; it is not an interaction between the human being and the machine. Thus, there is something deeply wrong in the very phrase “Human-Computer Interface”. Of course, “HCI” has become a hackneyed term, but this engrained (mis)-use does not make it correct. The basic problem lies in the implication that human beings and computers are entities of the same sort, so that they could “interact” on a basis of equality. This would only be correct if one whole-heartedly embraces the “computational theory of mind” according to which humans function like computers ; but as we understand it, an important common point among members of our network is that we reject this classical paradigm in cognitive science. If we are wrong about this, we certainly need some more fundamental discussion about the epistemology of cognitive science.

To introduce this discussion, a few words on the question of “symbols” may be in order. Conventional “wisdom” (as exemplified in our own collective text presenting the “Enactive” network) considers that there is “complementarity” between enaction and the symbolic domain. This may be diplomatically prudent (in an area where the primacy of computers and formal symbols is taken for granted), but fundamentally it seems to us to be a mistake. What are “symbols” symbols of ? And more fundamentally, what are “symbols” ?

We take the position that the grounding of all cognition lies in the dynamics of sensory-motor loops. As we have briefly indicated in section 3, a first stage in (biological) evolution consists of enlarging the range of “actions” to include the emission of “signals” ; and enlarging the domain of sensory inputs to include appropriate reactions to the reception of such signals. A paradigm case is “swarm intelligence” in colonial insects such as ants, termites and bees. There are beautifully worked-out examples which show that “signals” of this sort (pheromones in this case) lead to very effective co-ordination of actions (and the building of nests, trails, etc). We can define this as basic “animal” communication. There is then a big question in cognitive science as to the nature of (human) language ; this is not the place to go into this question (although maybe we will need to at some point), suffice it to say that language is a meta-communication involving intentionality and hermeneutics (the whole question of interpretations). Writing, and mathematics, come after that. The take-home point is that formal symbols (invented by Hilbert at the beginning of the 20th century!) are the end-result of a long evolutionary and historical process. The big mistake of the computational paradigm in cognitive science is to take formal symbols as theoretical “primitives” ; by so doing, this paradigm obfuscates the grounding of symbols in sensory-motor loops. In other words, the “enactive” approach is not “complementary” to the conventional “symbolic” approach ; rather, it leads to a radical reversal in perspective.

The take-home conclusion of this is that we should (in our view) eschew talk about “thinking machines” or “communicating machines” : no machines, not even computers, really think or communicate. They mediate interactions between humans and their world, or mediate communication between humans and humans.

Finally, an interesting question that arises is the status of “virtual reality”. In this case, it does seem as though the computer is playing the role of “the world”, by providing the sensory consequences of actions on the part of the human being. But even here, note that the experience of a human being immersed in a “virtual reality” is not that of interacting with a computer ; the human interacts with the entities that populate the “world” that has been brought about. We only become conscious of the computer (the interface) when a malfunction triggers the switch to the “put-down” mode ; in normal functioning (the “in-hand” mode) the computer-interface disappears from consciousness.

This remark is in no way meant to decry the interest of “virtual realities” ; on the contrary, such experiments are deeply revealing. What they show is that in order to create a “virtual reality”, it is neither necessary nor sufficient to compute (in all its gory detail) the total physical reality – an impossible task anyway, as shown by flight simulators that have to fall back on analog models ; what is required is neither more nor less than to provide the appropriate sensory returns to human actions. This helps, greatly, to bring home the point that what human beings experience in “natural” situations is not “the world in itself”, but the sensory-motor contingencies of their embodied situation.

To sum up : the title “Enactive Interfaces” is dangerously misleading. It is not the interface that is (or is not) “enactive” : it is the human subject, using an (appropriate, well-designed) interface who enacts a world. Thus, although an interface in itself never enacts anything, the role of interfaces is absolutely crucial : interfaces can permit (or not) humans to enact the world, and the world we live in depends on their design.