Friedhart Klix

The Method of Analogy Recognition for the Determination of Semantic Relations in Long-Term Memory

Publisher Summary This chapter discusses the method of analogy recognition used for the determination of semantic relations in long-term memory (LTM). It is known that information stored in memory do not exist independently but they enter into various different relations—a reflection of really existing connections between objects, events, and scenes. These relations between units in semantic LTM such as the concepts are referred to as semantic relations. Different approaches are used to classify these semantic relations. The existence of two distinguishable classes of semantic relations in LTM can be verified in psychological experiment. The sub-division of semantic relations is differentiated into inter- and intra-concept relations by the method of analogy recognition. An analogy is present when there is at least one identical relation between original and derived domain. The semantic analogy exists between concepts of natural language which can be defined by word marks. The property sets and connections existing between these properties form the basis for semantic analogy formation. The properties belonging to the property set of a concept as being equal are considered.

Semantic Relations: Recognition Effort and Pupillary Reaction

Two kinds of knowledge are differentiated: event-related, stationary knowledge and property-related, procedural knowledge. Experiments are carried out to check hypotheses concerning the recognition effort for both kinds of knowledge. Indices of recognition effort are recognition time and pupillary response. Analysed influencing factors are complexity and emotional evaluation of knowledge.

The Storage of Concepts and their Cognitive Application

Concepts are independently formed classes of objects and phenomena of our environment that are represented as cognitive units in our memory. They are the basis of most different cognitive activities. The present investigations analyze the use made of conceptual knowledge in the conceptual identification of objects, the verification of relations between sub- and superconcepts and the fulfilment of conceptual analogy tasks. It can be shown that the cognitive effort required in coping with conditions is systematically dependent on the specifics of the feature characteristics of concepts.

ON HUMAN PARSING

A model of sentence processing is outlined. It reflects to some extend human strategies in text comprehension. The main differences in comparison with AI approaches are (1) in the immediate access of words (or word groups) to their conceptual representation in memory and (2) in how conceptually organized background knowledge interacts with the text inputs.

On Microanalyses of Cognitive Performances: Correspondence with Psychological and Psychophysiological Parameters

A model of human knowledge representation is outlined. Two kinds of knowledge are distinguished: procedurally derived vs. stationarily stored knowledge. Some implications have been tested in a series of experiments. Predictions as to the necessary amount of information processing steps were investigated by means of the infrared pupillographic method. The data obtained correspond to the model; the comparisons between the data and the model give rise to new hypotheses on recognition processes.