Agents in Science, Technology and Theory

Agents have been used with convincing results in science, technology and in theory. It is useful to distinguish between the deployment of agents in these three fields in order to avoid misunderstandings and to reduce unreasonable expectations.

1. Agents in the Empirical Sciences

Empirical sciences investigate real systems and develop models for them. This means that empirical sciences find their objects of cognition and have to take into account the conditions that they encounter in the real world. System theory has shown that numerous real systems despite superficial differences have the same deep structure. For all models which resemble each other in this way uniform basic structures of models can be designed. Such models, which can be applied to a clearly-defined class of real systems, will be referred to as reference models.

Once reference models are developed on the basis of agents, the internal structure of the agents must conform to the objects in reality which are to be depicted by means of agents. The qualities and modes of behaviour with which an agent is endowed are determined by the qualities and the behaviour of the real objects which are modelled by means of agents.

This is also true of real systems in which human beings are involved and for which agents are deployed as models for real human beings. The agents acting as representatives for real human beings in the model must be as similar as possible to real human beings. They should possess similar qualities and display similar types of behaviour.

It will become clear that human behaviour does not consist solely of cognitive components but is to a significant extent emotionally conditioned. It therefore follows that agents whose function is to simulate human behaviour must natural also possess emotional capacities.

We can see that in order reliably and conscientiously to model human beings and human behaviour we cannot avoid posing the question of the nature of human beings.

2. Agents in Engineering Sciences and in Technology

Engineering sciences and technology have the task of designing and constructing new systems, systems which have not previously existed. In contrast to the empirical sciences, engineering sciences and technology are not primarily concerned with accuracy of depiction and structural resemblance to an already existing system. Here the emphasis is on function and performance. When developing and implementing new technological systems the engineer has to meet the following requirements: to produce a model which is clear, simple, functional and reasonably cheap. Apart from these requirements, no limits are imposed on his inventiveness and his creativity.

The engineer confronted with such tasks also has standard solutions at his disposal: they know for example how a suspension bridge or a tunnel is to be built. These standard solutions will be referred to as architecture patterns.

It has proved useful to have recourse to architecture patterns based on agents for certain tasks. One thinks in this context of the design of robots or of complex software systems. When agents are deployed in this context, their internal structure and internal architecture conform only to the requirements of engineering science. It is homo faber who operates in this field.

This means that the question of whether robots or software agents should be provided with emotions is solely a question of whether such agents function better and can be more cheaply produced than a design with a differently structured architecture pattern.

3. Theory and Virtual Worlds

The human mind is free and cannot be constrained by the instructions of nature for which a model is to be produced or by the requirements of a technical device which is to be constructed. It is capable of thinking up new realities and of trying out new possibilities. This does not mean that the virtual world created in this way must immediately be regarded as a model for actual real systems or as a blueprint for technical systems. Human beings develop systems with arbitrary qualities in a playful, intention-free manner and curiously study the modes of behaviour these new systems display. We are dealing here with homo ludens.

Nobody would dream of criticising such systems because they did not replicate reality or were useless and without practical application. If a system that originated in the virtual world contains agents, these agents can be however their designer wishes them to be. They can contain emotions or no emotions, as he wishes. The entertainment and science fiction branches make liberal use of this possibility. Criteria for the meaningfulness of such activity can be originality, creativity, imaginativeness and entertainment value.

4. Synergies

So far we have stressed the fact that the deployment of agents is meaningful in the empirical sciences, in technology and in the creation of virtual worlds but that in all three cases different types of demands are made on the agents. In this context it seems useful to sum up the essential points:

This could create the impression that the three areas of deployment are completely watertight and have nothing to do with each other. This is not the case. There are interactions and reciprocal influences. For example empirical sciences frequently find inspiration as blueprints for reference models in constructs of theory which at first seemed to have no application. Engineering sciences try to adopt construction schemata from models of nature developed by empirical sciences and then try to use them as the basis for architecture patterns. Socionics and bionics are examples of this. Theory in turns takes up conceptions from the empirical sciences or from technology and develops them further.

It is therefore advisable, whatever field one is working in, to risk taking a look over the garden fence in order to find inspiration and enrichment of one’s ideas.

Fields of application for agents