Knowbotic Interface Project
Transforming reality into a symbolic representation: Some examples
This
is a working paper which will probably be rewritten several times in the next months
AUTHOR: Dr. Gerd Döben-Henisch
COLLABORATORS: Leo Pos und Thore Swindall
FIRST DATE: December 27, 1995
DATE of LAST CHANGE: March-25, 1996
A short look into the literature -and into the history- of software engineering shows, that a uniform attitude can not be found regarding this task.
D.TEICHROEW and E.A.HERSHEY used for the display of a conceptual model a language called Problem Statement Language [PSL] (D.TEICHROEW/ E.A.HERSHEY III 1977:397) and they presupposed, that the problem space is a system with the general structure of having objects with properties, and the properties can have values. The objects are furthermore interrelated by relationships (D.TEICHROEW/ E.A.HERSHEY III 1977:398).
Somehow analogous is the argumentation of D.T.ROSS/ K.E.SCHOMAN Jr. (D.T.ROSS/ K.E.SCHOMAN Jr. 1977). For them the problem space consists of things that perform activities that interact with other things, each such interaction constituting a happening. They state explicitly that it is a generic universe of things and happenings which one has to presuppose and they give the list "objects - operations, data - activities, nouns - verbs, informastion - processing, substances - events, passive - active" (D.T.ROSS/ K.E.SCHOMAN Jr. 1977:370).
H.ÖSTERLE, who is focussing his analysis on problem spaces which are whole companies (H.ÖSTERLE 1980), speaks of informations understood as knowledge, processors understood as machines and humans, and methods, which will be applied from the processors to handle the informations (H.ÖSTERLE 1980:150). The transformation of such a system into a useful and reliable piece of software has to be done by humans which have only limited perceptual and cognitive capabilities to analyze such complex systems. In his analysis of the problem space he looks for 'components' (what is a component?), analyzes the hierarchy of relationsships between the components, defines the interfaces, defines a precedence-graph and a correlated controll-tree, which is extended with a marked network to handle parallel processes (H.ÖSTERLE 1980:153ff).
For DILLON and TAN consists the problem space of a portion of the real world, which has to be analyzed (DILLON/TAN 1993:24). For the conceptual model a set of structures has to be chosen which will be determined by real-world factors like entities with static and dynamic properties, relationships, transformations, constraints, imprecision and uncertainty, states, and time (cf. DILLON/TAN 1993:25f). For the construction of the conceptual model no assumptions have to be made about the software structure (DILLON/TAN 1993:24).
RAASCH assumes that the problem space is characterized by systems in environments. Events in the environment can trigger activities of the system which can lead to certain responses. Events can be context-free or context-sensitiv (J.RAASCH 1993:59-61). Analogous to H.ÖSTERLE sees J.RAASCH the process of transformation of real systems into some formal piece of software as a process of abstraction: a human person selects important features of the real system and constructs an idealized mental model (J.RAASCH 1993:65). This model has to be articulated with some symbolic means in a way that another person can interpret this symbolic representation with respect to the modeled reality (J.RAASCH 1993:66f). The model should be organized in a way that it allows the fragmentation of reality to an extend that a human person can 'grasp' the content. To get a complete model one needs a process with many stages (J.RAASCH 1993:70).
A new accent has been introduced into the discussion by MARTIN and ODELL (MARTIN/ ODELL 1995). The problem space is not any longer some portion of the 'real world', but the mind of the people whose content are ideas and concepts (MARTIN/ ODELL 1995:4f,11). In accordance with psychology they assume that it is a basic ability of humans to perceive regularities, to form concepts, to label concepts with symbols, and to use symbols in communication (cf. MARTIN/ ODELL 1995:12f). A concept is an idea or a notion that we apply to things/ objects in our awareness. Concepts shape our perception of reality (cf. MARTIN/ ODELL 1995:13f). We can also possess concepts of things that do not have sensory qualities (cf. MARTIN/ ODELL 1995:14). Although we can connect concepts with symbols, there is no intrinsic need for names to use concepts (cf. MARTIN/ ODELL 1995:16f). In our awareness are all objects instances of concepts; no object can exist without at least one concept that applies to them (cf. MARTIN/ ODELL 1995:26). In our mind we can select a domain as an area of interest (sphere/ perspective/ context). Because such a domain is connected with a certain set of concepts we can construct a domain specification by organizing the related concepts (cf. MARTIN/ ODELL 1995:20f). This can result in a conceptual model. The 'real world' is a certain subset of our awareness, which is dealing with the 'real world' mediated by concepts (cf. MARTIN/ ODELL 1995:337f).
In these examples one idea is -more or less explicitly- common to all of them: it is a human person which has to transform certain aspects of 'reality' into a 'symbolic representation'. ÖSTERLE, RAASCH, and MARTIN/ ODELL state explicitly that this process of transformation is bound to the perceptual and cognitive capabilities of human persons. MARTIN/ ODELL say, that it are the concepts of our mind which shape our perception of reality. Thus 'reality' can not be separated from the way our awareness is using concepts. Reality is given to us only within the conceptual framework of our mind. Thus the analysis of some portion of reality within our mind is not only guided by the correlated concepts, but it also produces again only concepts, or -as RAASCH states it- an idealized mental model. A Communication about such mental models is possible because we can introduce symbolic representations of such models and because different people can share the same concepts.
This analysis leads immediately to the point of view which is adopted in KIP II.
The KIP II point of view
In KIP II we assume, that a complete picture of 'reality' can not be achieved by looking to reality only 'from the outside' of an observer. A 'most complete view' is only availabe from a 'first-person view' as NAGEL calls it (cf. NAGEL 1986) or from a phenomenological point of view as it is stated in the philosophical introduction to KIP II theories. 'Empirical Reality' is in this view a proper subset of the set of all knowable phenomena D-PHEN, also known as the 'consciousness' or the 'mind'.
For the task of the transformation of 'reality' -or some other relevant parts of the consciousness- into a conceptual model this does mean that within KIP II certain subsets of the consciousness have to be analyzed and then to be mapped into a symbolic representation. This kind of mapping is a special case of the more general task of the selfdescription of the consciousness. The general reflections of this topic can be found in the paper Phänomenologische Struktur des Bewußtseins.
At the moment it is still an open question which in fact are the structures of our consciousness, and, depending on this, the question, which are the appropriate symbolic means to represent these only vaguely known structures. Therefor is the decision at the moment a bit arbitrary, what kind of symbolic means one will chose.
Comments are welcomed to kip-ml@inm.de
Daimlerstrasse 32, 60314 Frankfurt am Main, Deutschland. Tel +49- (0)69-941963-0,
Tel-Gerd: +49(0)69-941963-10