The ontology of technical systems is still in a rather poor condition. A more general subject, the ontology of artifacts, is, however, a new developing discipline. The first book in this subject was published by Dipert 1993; since its main focus lies on works of art, complex technical systems are rather neglected.
Peter Simons and Charles Dement (1996), in contrast, deal with artifacts by analyzing systems such as airplanes. In their paper they present a mereological approach, in which the concept of a component is defined. They show that different kinds of components have to be distinguished, assembly and disassembly components, for instance, or functional, design and maintenance components. These components are obviously of practical relevance, as is shown by the fact that the manufacturing industry itself has developed a device for representing the mereology of complex technical systems, the so called 'bills of material' (boms). Although the mereological approach appears to be very fruitful, it cannot, however, grasp a very important characteristic of complex technical systems, their multidimensionality.
The present paper deals with this phenomenon,
concentrating on complex artifacts such as telemedical systems. In the
first section the need for a multidimensional ontology is pointed out.
The second paragraph presents the concept of a multidimensional ontology
by discussing problems of Aristotelian ontology and improvements suggested
by Franz Brentano. In the third paragraph a multidimensional ontology for
complex technical systems is suggested
1 The problem: a multidimensional ontology
for complex artifacts
The complexity of an ontology for artifacts such as technical systems can best be exposed by looking at the process of their development. Consider, for instance, a group of scientists trying to develop a telemedical system. Such a system is designed to improve the medical service in regions which cannot afford expensive medical systems. The telemedical sytem connects medical doctors in this location with well-equiped specialists who are possibly hundreds of miles away. A central problem of the developmental process will be the integration of the different perspectives of the developers. Thus the medical doctors, who will use the system, will talk about things such as x-ray photographs and other media for diagnosis. The computer scientists who have to develop the computer programs involved in such systems concentrate on the information processing necessary to make the whole system work. They know nothing about the medical apparatus their colleagues are talking about. The same is true of the engineers who have to maintain the system once it is in operation. But these are not the only perspectives which are relevant for the development of such a complex system. In addition to the practitioners who will directly work with the system or construct it, there are aspects such as data protection or ecology which have to be taken into account. And one should not forget that the whole project is also an economic entreprise, which has to be managed and to turn a profit. Thus the problem of integrating a multiplicity of perspectives arises.
On the first glance it might seem as if this were merely a practical matter since the different perspectives are consequences of the fact that the system is located in different pragmatic contexts. But although a pragmatic approach may describe the differences pretty well, it cannot grasp the fact that the developers intend to refer to one and the same thing. But the mere intention is not enough, since the system has to be developed. This is not a trivial problem. Thus it may well be that the developers do not talk about the same system. This is, for instance, the case if the medical users ask for facilities which cannot be realized in the present state of the technical arts or if economical and technical requirements are incompatible. We see that the developers have to solve an ontological problem, the problem of making sure that they are talking about one and the same system and that this system is a possible entity. The task of developing a complex technical system involves the tak of the unification of objects described from the different perspectives of the different participants. The ontological problem of the unity of an object arises.
The situation is, however, even more complicated. Development is itself a process, in which the object changes. In the different stages of this process the system is represented in the form of ideas, prototypes, alpha-and beta-versions and so on. What we probably don't have is a final version. Even the stage in which the developers can expect a premium for their achievement, in the stage where the system actually runs, improvements and adaptations for new requirements have to be made. Hardware is exchanged, and new versions of programs have to be installed, for instance. Thus we cannot talk about the system as an end product, since it is always evolving in time. Even the destruction of the system is part of this process. It results in the disassembly components (cf. Simons and Dement 1996) which are for ecological and other reasons in need of a special treatment. It is very important that during the different stages of development every participant knows at which stage he is. What is necessary is an abstract theoretical framework which takes into account the many facets of complex technical systems. What we need is an ontology of artifacts.
How is such a framework possible? An important hint is given by Aristotle. He deals with a problem which has some similarity with the one just described. As Aristotle points out, we talk about things and persons in very different ways. Of Socrates, for instance, we say that he is a human being, that he is snub-nosed, that he is in the market place, that he is bigger than Critias and so on. In every instance we are talking about Socrates but in ways which seem to be incompatible with one another. Being in the market and being snub-nosed are completely different things. It was Aristotle's idea to make a taxonomy of the different ways we talk about things. This taxonomy shall inform us of how it is possible to determine the different ways in which we can say of a thing that it is something. These are the kategoremata, the categories, the most general genera. These categories are not only fundamental for logic, but are also central for ontology.
The similarity to the problem outlined
abouve is obvious. In the case of the tele-medical system as in the case
of Socrates we talk about them in different ways, which cannot be reduced
to one another. The system described by the medical doctor is in
an analogous way different to the one described by the computer scientist
or the entrepreneur as the description of Socrates as being in the market
and as being snub-nosed. So what we need is a categorial system for complex
technical systems which enables us to distinguish the different ways we
can talk about them. We can expect from such a categorization a practical
advantage which is similar to that which is gained from the Aristotelian
one. Both categorisations can be used as a heuristic. So if we want to
give a complete description of a thing, according to Aristotle, we have
to make statements according to the different categories: what substance
it is, what quality it has, what activities and passions there are, what
temporal and locational determinations can be made, and so on. The same
useful procedure can be expected of a categorial system for complex artifacts.
2 The concept of a multidimensional ontology
Before we return to technical artifacts we have to consider traditional ontology. From this tradition we expect further hints for our ontological project. Thus we use the philosophical tradition as a heuristic. Moreover, we are interested in relating the ontology of complex artifacts to the great metaphysical tradition. As philosophers we are glad to support the man of action with theoretical tools, but we are not really happy until we get new insights, until we understand, for instance, why we need a multidimensional ontology.
The traditional ontology from Plato and Aristotle is monodimensional. According to Aristotle, who dominated this subject until our times, a thing is a unit of material and formal parts. Whereas the formal parts determine what a thing is, the material part individuates the thing. The formal parts of a thing, which are called its 'substantial form', themselves constitute a hierarchy. Thus if something, is a dog this substantial form can be specified by determining what kind of dog it is - for example, a dachshound. On the other hand, there are genera of which the dog is a species, namely animal and living thing. So. if one wants to determine what kind of object one is examining, one has to determine the species specialissima, that is the most specific substantial form. In order to define this, one has to refer to its next genera and the specific difference according to the traditional theory of definition which says: definitio fit per genus proximum et differentiam specificam. So if your determination gives the species specialissima, you implicitly refer to the complete hierarchy up to the most general ontological category which is the category of substance (ousia). Such an ontology I call monodimensional, since there is exactly one line or dimension beginning with most general and terminating in the most specific form which determines the object. The species is not only partially but completely contained in its different genera.(1)
There is, however, one great problem with this ontology: it makes sense only if one accepts Aristotle's doctrine of act and potency. Without this doctrine the notion of substantial form is untenable, since there will be no sharp distinction of substance from accidents. According to Aristotle, the substantial form of a human being is the nous, the intellective soul, which supplies the human being with the logos (zoon logos echein). One cannot, however, find this logos in a new-born child, at least not a complete realization of it. Nevertheless the child is a human being. For Aristotle this is not really a problem. According to his doctrine the child already has thelogos. It is, however, not yet actual but only potentially there. This is proved, according to Aristotle, by the fact that in optimal circumstances the child will gain the logos in its full form. This is not the result of mere education, since the dog which always accomponies the child during his exercises will not be educated. This must have a reason. And this reason is, according to Aristotle, that the potential substantial form becomes actual in the process of development.
The doctrine of act and potency is also necessary for the Aristotelian doctrine of individuation. As already mentioned, according to Aristotle a thing is composed out of form and matter. And it is matter which individuates a thing. This can be made clear by considering the most special species. You can specify a thing in as detailed a manner as you like, but it will always remain possible that there are different tokens of it. Thus the individuation must be made by something other than form, and this job is done, according to Aristotle, by matter. But if a thing consists of form and matter, why is it just one thing and not a collective of two things, namely form and matter? Aristotle's answer is again contained in his doctrine of act and potency, which allows the concept of a perfect fusion of form and matter. Both are unseparable parts which stand in an ontological relation: the matter functions as potency - Aristotle even used the Greek term hyle as a synonym for dynamis, the Greek term for potency. The substantial form, on the other hand, functions as its act (or entelecheia). It makes the matter the thing it actually is. Thus form and matter are one thing, just as a possible chair and the actual chair are not a collective of two things but one thing, although the actual thing has to be possible, in order to be actual.
Unfortunately, Aristotle's doctrine of act and potency is not tenable. It presupposes, for instance, that a pure form functions as a cause. This is, however, not in accordance with physics. This deficiency was recognized also by Franz Brentano, who nevertheless tried to reestablish a metaphysics in Aristotelian style. He tried to establish an Aristotelian ontology without the doctrine of act and potency. Obviously, the abandonment of this doctrine makes a new concept of substance necessary. Such a concept is suggested by Brentano in his later writings which have been (not very happily) characterized as 'reistic'. In these writings Brentano points out that Aristotle was wrong to advance his monodimensional ontology described above. The specific differences are not, as Aristotle taught, subordinated under one category or, as Brentano calls it, they are not homostoichetisch (or monostoichetisch) (from Greek stoichos, line, order). Brentano gives the following example. Consider a red point at location L and another point with the same quality at location O. The two points are not identical, and this makes evident, according to Brentano, that the location is a substantial difference, since there is no other respect in which these two points differ. On the other hand, the point would be a different one if it had another color (quality). Thus the location cannot be the only specific difference of the point. The color of the point or, more general, its quality belongs to its specific difference, too. But quality and location are different ways to classify a thing, or, in the language of Aristotle, quality and location are different categories which are conceived by him as the most general genera. According to Brentano, there are different kinds of substantial differences which cannot overlap. Brentano calls his doctrine that there are different lines of substantial differences subordinated under different general genera heterostoichetisch (or pleiostoichetisch). This is an important deviation from Aristotle. It makes the doctrine of act and potency spurious. For things, which are subject to outer perception, the distinction between subject and accident is given up. Quality, which is traditionally regarded as the prototypical accident, is 'sucked up' by the substance: it becomes a substantial difference. Thus there is no need to distinguish between substantial and accidental forms any longer, so that the doctrine of act and potency loses one of its main jobs.
But also the other main function of the doctrine of act and potency, its role in a theory of individuation, becomes useless. Brentano is not in need any longer of matter as a principal of individuation. Since there is more than one line of substantial difference, these lines themselves can individuate a thing. Matter and therefore hylomorphism no longer exist in this new ontology, the problem of unifying form and matter has diasappeared.
This 'new ontology' outlined by Franz Brentano
is in fact already a version of a mulidimensional ontology projected here.
As we have seen, there are, according to Brentano, two lines of classification
(of stoichoi) which are relevant for things given in outer experience,
namely location and quality. One can represent this by building a Cartesian
product, that is, by using the twodimensional system with which we are
familiar from geometry. What results, however, is a qualitative and not
a quantitative space. The possible things can be mapped on a plane in which
each point is determined by its position on a quality axis and on
a location axis. And each thing can be identified by representing its location
on the quality axis and on the location axis. This yields a two-dimensional
ontology. But, depending on the kind of being we are dealing with, higher
dimensional ontologies are necessary. This is the essence of Brentano's
3 A multidimensional ontology for complex technical systems
The questions arises what are the categorial or ontological dimensions, as I call them, for technical systems. A first dimension is the dimension of the different perspectives as described in the first paragraph. In this line we have the telemedical system under the descriptions of the medical doctors, of the computer scientists, the engineers, the entrepreneurs and so on. I call this the cognitive dimension since it is constituted by different cognitive acts in which the system is conceived in different ways.
In each conception the system can be described under different formal concepts, such as consistency, reliability, information, part, and so on. The point is that there is not only one question of consistency, reliability and so on, but as many as there are cognitive perspectives. In the case of the medical doctor reliabiliy refers primarily to the diagnostic procedure, in the case of the computer scientist it refers to a quality of the computer programs, and in the case of the engineer the construction of the hardware is in question. I call this the formal ontological dimension.
I mentioned already that a technical system evolves in time, that there are stages in which the system exists in the form of a mere idea, of protototypes, of alpha- and beta-versions and so on. At each stage you have different cognitive perspectives as well as the fomal ontological dimension. This yields a third dimension, which I dub the progressive dimension.
But this is not yet the whole story of complex artifacts. If you have realized that such a system is a temporal object evolving in time, you will see that there is no straight progression but a quite complicated process, in which there are deadlocks, corrections, new starts with new strategies, recourses to strategies which one had already given up and so on. This dimension, which I call evolutionary dimension, is of greatest practical importance since it mirrors the interactions which have taken place between the participants. The medical doctor, for instance, tells the computer scientist what he wants the system to do. The computer scientist, in turn, translates this into his computer programming language and makes a suggestion. Confronted with this proposal, the medical doctor may realize that he has not been properly understood or that his own ideas have not been clear enough. Thus he rejects the proposal of the computer scientist; he makes a more precise description of the requirements and gives new specifications. This in turn leads to another proposal of the computer scientist and so on. The specifications of the medical doctor are part of the ontology. They form a certain area in the ontological space of the system: they can be related to a viewpoint of the cognitive dimension, to the concepts of the formal ontological dimension and to a certain stage in the progressive dimension. The same is true of all the proposals of the computer scientist. Thus is mapped the cooperation which takes place in the developmental process.
Whether we can manage with these four dimensions is an empirical question. What we are offering is an ontology which calls for an application. One should, however, realize that the four-dimensional ontology here suggested is only a very abstract framework. In this respect it is similar to Aristotle's theory of categorization. For each system one has to determine the specific concepts which are relevant for it. A category such as quality is specified in regard to a sound in different ways than to a visible thing. In the same way the viewpoints relevant for a telemedical system are different to the viewpoints which one has to distinguish in dealing with a transport system.
Such an ontology is certainly of great philosophical relevance, since it makes clear that we need a new categorial system in dealing with complex technical systems and, more generally, complex artifacts. A philosophy of technology, for instance, will certainly fail if it describes its subject in terms of a thing ontology, or with the help of concepts which are derived from the consideration of tools like hammers. In contrast to the hammer a complex technical system is essentially a multisubjective object, insofar as it is constituted by different cognitive perspectives, which are not necessary in the case of a hammer. Moreover it is essentially a temporal object, in a sense in which an object falling under the traditional category of thing is not.
One may, perhaps, object that we need the concept of a thing for the naive physics project. It is true that the category of a thing is central there. And a thing has always many features or accidents (cf. Herbart) But as philosophers we need not only naive but also more sophisticated ('philosophical') ontologies. The multidimensional ontology can deal with substance/accident. In the case of a hammer, for instance, we can normally neglect the evolutionary dimension. The progressive dimension is radically reduced, only two stages are considered, namely the stage of development and the stage in which there is the working hammer. And of course, only this last stage is regarded as relevant (at least for people who are not tool makers). A similar reduction takes place in the first dimension, in which only one perspective is taken into account, that of the hammering person. If one puts 'Quality' and 'Reliability' on the formal-ontological dimension, then this dimension is needed too. A hammer can then be described as a non-temporal, mono-subjective (in contrast to multi-subjective) thing with many accidents. In other words, the substance/accident ontology can be integrated into the multidimensional ontology. The substance-accident ontology can be conceived as a derived, non-basic ontology.
Also Simons' and Dement's mereological
approach sketched above can be dealt within the multi-dimensional approach.
'Part' is a formal-ontological concept; it is for their approach the only
concept in the formal-ontological dimension. But Simons and Dement take
different pespectives of the cognitive dimension into account. This results
in different 'kinds' of components (parts). The multi-dimensional ontology
can, moreover, be used as a heuristic to improve the mereological approach.
It is quite clear from the structure of the multi-dimensional framework,
that in the case of the telemedical system there must also be components
in the perspective of the medical doctors, and in the perspective of the
computer scientist. In the first perspective there are entities such as
x-ray photographs, in the latter different programs and subroutines. The
multi-dimensional ontology is also of great practical relevance. It can
be used, for instance, as a heuristic in the development of such a system.
It is, actually, inspired by a developmental model suggested by the computer
scientist Bernd Mahr at the Technical University Berlin, which in turn
refers to models like the ISO standard ODP (Open Distributed Processing).
It enables the developers to ask relevant questions, insofar a it helps
to dissect a complex problem into more specific ones. And more importantly,
it improves the cooperation of the participants belonging to a variety
of disciplines and pursuing different interests. Thus, we can put straight
Nietzsche, who tried to philosophize with a hammer, namely by developing
technical systems with Aristotle and Brentano.
Brentano, Franz, Kategorienlehre,
ed. by Alfred Kastil, Leipzig: Meiner 1933; Engl. transl.: The Theory
of Categories, ed. by R. Chisholm and N. Guterman, The Hague: Nijhoff
Chisholm, Roderick, 'Brentano's Theory
of Substance and Accidence', in: Chisholm, R.. Brentano and Meinong
Studies, Amsterdam: Rodopi 1982.
Dipert, Randall R., Artifacts, Art Works,
and Agency, Philadelphia: Temple University Press 1993.
Mahr, Bernd, 'Die Menschwerdung der Maschine',
Kursbuch 128, p.. 15 - 34.
Simons, Peter M. and Charles W. Dement,
'Aspects of the Mereology of Artifacts', in: Roberto Poli and Peter Simons,
ed., Formal Ontology. Dordrecht: Kluwer, 1996, p. 255-276.
Smith, Barry, Austrian Philosophy. The
Legacy of Franz Brentano, Chicago: Open Court 1994.
1. According to Aristotle there are also other categories than the category of substance, as, for instance, the category of quality, quantity, action, passion, relation. They are, however, not of equival value. In contrast to the other categories, substantial forms have a causal power. Therefore science deals only with substances but not with accidents, as Aristotle points out.
2. . Unfortunately, this point has completely escaped the notice of philosophers dealing with the later work of Brentano. The focus is, rather, on Brentano's new interpretation of accidence and substance (e. g. Chisholm, 1978, or Smith, 1994, chap. 3). According to Brentano, examples for accidents can only be found in inner perception. Only in respect to the soul can we speak of accidents, which mean that which intentionally inexists, the intentional content. Brentano's retaining to the ontology of substance and accident in respect to the soul, which he had actually overcome by his multidimensional ontology presented in his later writings, can only be explained by his theological motivations. He wants to retain the soul as a stable entity which can have different 'modifications' without a substantial change. But for this task he needs the old category of a substance. In all areas which are relevant for naive physics, we don't need, however, the concept of substance and accident. Thus one can claim that the importance of Brentano's later ontology consists in the fact that he gets completely rid of the substance - accidence - distinction for ordinary things, like stones, tables and colored points.