World Knowledge Representation Research Articles

Cognitive Research Methods in Linguistics: Conceptual-Inferential Analysis

In studying language modern linguistics takes into account the users’ personality and their role as interpreters, arguing that language does not only convey a person’s knowledge of the world in one format or another, but also interprets it in a certain way. The leading role of the anthropocentric factor in linguistic research, as well as focusing on the cognitive and linguistic mechanisms of meaning construction makes the problem be discussed in the study up-to-date and significant one within the framework of the cognitive-discursive paradigm. The authors put forward an idea of conceptualinferential analysis as a new research method which can be effectively applied to language data within the framework of a cognitive study of language. The aim of this article is to substantiate the possibility of using this method in modelling the processes of interpretation and representation of world knowledge in language. The application of the method of conceptual-inferential analysis is illustrated by the example of quantitative interpretation of quality in language as a complex mental processing of quantitative-qualitative relations in the world. Based on the analysis of the linguistic means of the quantity concept representation, the article gives a deeper insight into processes of quantitative interpretation of quality, describes the repertoire of potential inferenced qualitative meanings and demonstrates their dependence on the cognitive context. In addition to the proposed method, the authors also use other research methods and techniques of conceptual analysis and conceptual-definition analysis. The evidence data (12000 examples) is retrieved from the works by modern authors. The integrated application of conceptual-inferential and conceptual-definition analysis methods reveals a high interpretive potential of quantitative meanings in representing quality in language. It also helps identify cognitive mechanisms of the quantitative-qualitative meanings construction, and to detect various types of knowledge available to an individual as an interpreter of quantitative-qualitative relations and their representation in language.

Deaf Adolescents’ Representations of Taxonomic Relations

Because deaf people differs from hearing people in so many ways,such as using a different language and having a relatively isolated way of living,a reasonable hypothesis is that their representation of world knowledge might have some unique characteristics compared to hearing people.In the current study we sought to investigate their representation of taxonomic categories.Our hypothesis was that deaf adolescents had stronger tendency of imagery representation of exemplars and weaker awareness of taxonomic categories than their intelligently-matched hearing counterparts. In experiment 1,15 deaf adolescents and 15 intelligently-matched hearing counterparts(31 months younger) were recruited to complete a semantic categorization task.There were two groups of experimental stimuli.One was high-typicality group composed of 16 taxonomic categories,in which 3 exemplars of higher typicality and 1 of lower typicality were selected as prime and target.The other was low-typicality group again of 16 taxonomic categories,in which 3 exemplars of lower typicality and 1 of higher typicality were selected as prime and target.All the exemplars from 25 taxonomic categories were presented in forms of words and pictures and subjects' reaction time and accuracy were recorded.In experiment 2,16 deaf adolescents and 16 intelligently-matched hearing counterparts(38 months younger) were recruited to complete a false memory task.There were two kinds of experimental stimuli.The first group was composed of exemplars from 13 categories,in which 3 exemplars of high typicality and 1 of low typicality were used as study item and lure.The second group was composed of exemplars from another 13 categories,in which 3 exemplars of low typicality and 1 of high typicality used as study item and lure.All the exemplars were presented in forms of words and pictures and the purpose was to further test our hypothesis. The results showed for experiment 1,both deaf and hearing adolescents responded faster to the stimuli sets in which the primes were exemplars of higher typicality than lower typicality.But deaf adolescents responded slower to picture stimuli than words while hearing adolescents' reaction time remained the same.Besides,deaf subjects made more errors when the primes were exemplars of higher typicality.Similar results were obtained from Experiment 2.Deaf adolescents had lower recognition accuracy when target items were of lower typicality while hearing subjects' recognition accuracy was not influenced by exemplar typicality.For the recognition accuracy of lures,deaf subjects made more error than hearing subjects.Furthermore the hearing subjects' recognition was influenced by exemplars typicality and item presentation mode while the deaf subjects were not. To sum up,the current study demonstrated that deaf and hearing adolescents had similar typicality awareness.But due to their persistent dependence on the use of sign language,deaf adolescents had weaker awareness of a taxonomic category than their hearing counterparts.Instead,they had a stronger tendency to use imagery representation of taxonomic categories than their hearing counterparts.

The emergence of high level structure in classifier systems - a proposal

While Artificial Intelligence approaches to the representation of world knowledge has taken an unashamedly structured approach, Artificial Life workers have shunned such forms (Maes, 1992). Recent problems in the production of symbolic reasoning within classifier systems (Forrest, 1991), the debate between the ‘flat’ and ‘hierarchical’ schools of thought within autonomous agent development (Tyrell, 1992a,b). and developments in coevolving classifier systems (Dorigo & Schnepf, 1992) have fuelled new interest in higher level structures linking multiple classifier systems. It is shown that further work is required in order to understand what form such structures might take. and proposals are presented for the development of these structures

Knowledge representation support

The representation of world knowledge in a machine translation system is discussed, especially in terms of the interlingua text that is adopted in KBMT-89. While knowledge about the world is at the core of a knowledge-based machine translation program, it cannot be extracted from a text without knowledge about how languages encode it syntactically. The paper therefore also presents the grammar formalism with which language-specific syntactic knowledge is presented. Details are presented of our version of unification-based grammar, its basis in linguistic theory and its role in a knowledge-based machine translation system.