Grammar of Dynamic Knowledge for Collaborative Knowledge Engineering and Representation

Grammar of Dynamic Knowledge for Collaborative Knowledge Engineering and Representation

Cyril Pshenichny (Geognosis Project, ITMO University, Russia) and Dmitry Mouromtsev (ITMO University, Russia)
Copyright: © 2015 |Pages: 28
DOI: 10.4018/978-1-4666-6567-5.ch016


Constructive discussion must lead to a shared understanding. This understanding is commonly expressed as text; however, for the purposes of collaborative research, the tools of knowledge engineering/knowledge representation look more appropriate. The problem with them is that, to the present day, they are developed largely for the tasks that imply fixed relations between things and their properties, termed here as static. However, collaborative research often deals with fields of knowledge that represent changing environments where these relations cannot be considered fixed, and the tools to capture scenarios of evolution (i.e. the dynamic tools of knowledge engineering) are far from that evolved as static ones, mainly due to the lack of strict logical or mathematical foundation for representation of dynamic knowledge. This chapter presents an attempt to formulate a unified grammar to encode the knowledge of changing environments in any field of science.
Chapter Preview


Constructive discussion must lead to a shared understanding – at least to understanding of points of disagreement. In modern literature, these “milestones of understanding” are often termed as boundary objects defined as “artifacts that aim to bridge concurrent cognitive models through abstraction from all the domains of the partners” (Diviacco, 2012, p. 17). These artifacts range from purely formal (e.g., file structure, blank layers of a geographic information system, keys of a database), which add little to understanding, to some shared meanings. So far, common practice of human discussion has been mainly to come up with a text, be it a manifesto or research paper, which aims to fix and present these meaningful boundary objects (usually in the form of generic terms).

Nevertheless, text has a few properties that undermine its ability to serve this purpose – i.e., unconstrained grammar of a natural language always leaves a room for uncertainty. Verbal narration is unlikely to present all possible consequences from a considered point (that is, to make sure that understanding is really achieved); the way the consequences are made is often unclear and non-verifiable; moreover, to be readable, text should preserve intonations, style and emphases, which may stress some points and obscure others. Finally, text has a beginning and end and thus channels the thought in one given direction, while the vision it strives to communicate may – and commonly does – require not a single-path and one-way, but a network-like, or polytree, pattern of presentation, with an opportunity to travel both ways on each path. Even the hypertext markup language is not free from this disadvantage because it rarely allows to follow several links between the documents simultaneously.

An alternative approach is offered by knowledge engineering/knowledge representation. That is, to bind verbal boundary objects not in a text but in a graphic conceptualization of the considered field of knowledge (e.g., Martin, 2002; Gavrilova et al., 2013).

However, the tools suggested so far (ontologies, conceptual graphs and others) cover only a part of the tasks arising at conceptualization of fields of knowledge – namely, the representation of static knowledge. Static is the knowledge in which the relations between objects are considered fixed and not changing (Pshenichny and Mouromtsev, 2013). This has proven to be efficient for a wide range of tasks like building data models and search algorithms or description of those domains in which things, their properties and relations are considered stable – e.g., production units of a factory or biological species.

The other part of existing tasks, which relates to the knowledge on changing events and implies consideration of processes and scenarios, is regarded as dynamic knowledge (Pshenichny and Mouromtsev, 2013). It remains largely uncovered by existing tools of knowledge engineering and representation. Though, there is quite a research done in the field of dynamic ontologies and ontology evolution (Stojanovic, 2004; Zablith, 2008; Murdock et al., 2010, and others) including that in collaborative environments (Noy et al., 2006), but the approach in this field is quite different and focuses on how to adapt ontology to changing phenomena, while the issue is not how to capture the changes in the phenomena themselves. Speaking about “evolution”, “processes”, “scenarios” and related issues, the researchers in this field mean the change of ontology proper, not of the world it describes, and look for solutions to keep ontology self-consistent despite the changing world rather than to model the changes of the latter.

Meanwhile, the tasks of representation and engineering of dynamic knowledge are especially crucial for collaborative studies where the issue of mutual understanding acquires a key importance. Ignoring the dynamic properties of many environments (e.g., a man is sleeping – a man is vigil and not sleeping), one would not reach understanding but imitate it – obscuring, but not resolving, the contradictions. Therefore, a formal account of dynamic nature of many environments addressed by various disciplines, from geosciences to psychology, is of vital importance for efficient collaborative studies and research networking.

The objective of this chapter is to draft a general approach to collective reasoning and representation of knowledge of changing environments. To do this, we will

Complete Chapter List

Search this Book: