The Evolution of Complexity - Abstracts. The Evolution of Complexity - Abstracts.
Science has evolved through a sequence of paradigm changes which affected not only the method but also the choice of questions studied Ku:Stru. The original `natural philosophy', which attempted to capture the workings of the Universe by a handful of rational principles, turned into a gigantic jigsaw puzzle, as the original principles got refined into complex and detailed models stated in a technical, specialized language (on a low representation level). The focus of science has thus drifted away from the simple and broad terms close to daily-life language (high-level terms), by which essential questions and answers of general interest are most naturally expressed.
The sense of clarity and power of reason which science originally brought to the humankind has diminished. At the same time, the ascent of science has weakened the trust in other sources of high-level information. A peculiarly nearsighted culture resulted, with the low representation level where technological means are found in its eye's focus, and the high representation level where meaningful ends are to be sought in its eye's blind spot.
Polyscopic modeling seeks to lay a foundation for a balancing cultural impulse (the word `modeling' is used in a broad sense roughly equivalent to `information representation'). It encompasses an attitude} towards information, a set of principles and a collection of techniques.
In contempt of a common attitude by which any accurate piece of information is at best useful and at worst irrelevant, polyscopic modeling views information as a subtle cultural pollutant destroying interest and meaning, unless the information is carefully structured and cautiously used. Well-structured information provides a perspective by which only that what is broadest and most essential meets the eye. Multiple simple and concise models (called views) are preferred to a single large model. Essential high-level views are emphasized, nonessential and detailed views are subdued and made available on demand.
Simplicity, however, should not result by adherence to some chosen metaphysics or tradition. Harmonizing with the view that is common in modern science (see, for example, Auto), polyscopic modeling views information as a collection of models rather than as a unique or exact rendition of reality. By disassociating modeling from metaphysics, in particular from the mechanistic metaphysics that was often implicit in classical science, an essential obstacle to scientific high-level modeling is removed. Diverse related models are needed to represent multiple levels of detail, viewpoints, paradigms, traditions and styles.
Information is viewed as having two inseparable parts, a static part which can be stored on some physical medium, and a dynamic part which is unavoidably embodied by living humans. The primary purpose of static information is seen in supporting a natural development of dynamic information. Static information has a meaning only when the corresponding dynamic information is present.
Freedom from the constraints of tradition or metaphysics reintroduces the problem of determining the modeling paradigm. The techniques of polyscopic modeling are defined via formal polyscopic modeling languages and studied theoretically. The role of a modeling language/theory pair in polyscopic modeling is analogous to the role that mathematics has with respect to physics.
At present, the basic concepts and techniques of polyscopic modeling are defined by the ab polyscopic modeling language. The language includes operations for joining the information of multiple models into a single model, for computing the information which is common, for computing high-level views and others. The language also includes relations or properties of models such as coherence, compatibility and inclusion.
Modeling languages can also be used directly as formal languages for defining models, similarly as computer programming languages are used for defining algorithms. Information technology and the available techniques for information processing (computer programming) can then be applied in information representation.
Patterns or generic models are a central polyscopic modeling tool. They vaguely resemble scientific laws in that they express common aspects of seemingly disparate phenomena or models. They thus serve as natural building blocks for high-level modeling. By exhibiting similarity of what is unfamiliar with that what is familiar, patterns facilitate understanding. Furthermore, patterns are used for high-level terms (whose meanings are often faded through centuries of use) based upon their relationships with other concepts and the roles they have with respect to the whole. Intuitive techniques such as ideograms and metaphors are used for describing patterns along with formal modeling languages.
The paper introduces and motivates polyscopic modeling and surveys its principles and techniques. The ab language and the corresponding theory are illustrated by an example. Theoretical and applied polyscopic modeling research in progress are outlined. The format of the paper represents an attempt at polyscopic text formatting, where text is structured as a collection of interdependent modules.