FFF CONFERENCE CTF07

Alexander Ziem - Default values. Integrating tacit knowledge in a frame semantic approach to meaning construction

In an early and very influential study, one of the pioneers of cognitive science, Marvin Minsky (1975), observed that every cognitively relevant percept appears as a richly detailed entity, even though the input itself does not offer much information on its own. From this fact he concluded that understanding essentially involves cognitive framing. A percept becomes cognitively relevant if it is embedded in a frame, and we can only access the data provided by a frame if its chunks of knowledge constitute a well structured concept in our long-term memory. Minsky suggests that the structure of frames may be well described in terms of (i) slots and instances of slots, namely (ii) fillers and (iii) default values.    From a linguistic point of view, fillers are new text elements (predications) being explicitly introduced by language users, whereas default values are given elements in the sense that a speaker/writer may presuppose them and a hearer/reader, in turn, may infer them (Lönneker 2003). During language comprehension a word thus provides a semantic potential for evoking pieces of knowledge (cf., for instance, Fillmore 1977); it evokes default values. Viewing frames in this light has fundamental consequences for a cognitive semantic approach to word meaning. If language understanding relies upon schema-driven conceptualization processes, then the main empirical work should consist of investigating default values since they substantially motivate the emergence of meaning. Although many studies have adopted Minsky’s model in one way or the other, the role of default values in frame semantics is widely ignored. How can slots of a frame be determined? And how can default values be identified empirically?    In my talk, I will present both theoretical and empirical evidence for the fundamental role of default values in frame-driven conceptualization processes. On a theoretical level, I first argue for an integrative view of frame semantics and some basic tenets of Cognitive Grammar, such as the symbolic thesis, the entrenchment theorem and the distinction of profile and base. On the basis of a case study, I then go on to illustrate that a so called “hyperonym type reduction” (Konerding 1993, 140-217) provides a suitable device to determine a frame’s slots. My particular object of interest is the word locust, a metaphor introduced by a German politician to stigmatize financial investors. The corpus consists of 200 newspaper articles. The analysis focuses on recurring fillers (predications) being attached to one particular slot of the locust-frame, and it is differentiated between two kinds of entrenchment. Type entrenchment concerns those slots which are significantly more often ‘filled’ than others. Token entrenchment, on the other hand, relates to frequently recurring fillers: default values in Minsky’s sense.

 

References

Fillmore, C. J. (1977): Scenes-and-frames semantics. In: Zampolli, Antonio (ed.): Linguistic Structures Processing. Vol. 5. Amsterdam/New York/Oxford: North Holland,  55-81.

 

Konerding, K. (1993): Frames und lexikalisches Bedeutungswissen. Untersuchungen zur linguistischen Grundlegung einer Frametheorie und zu ihrer Anwendung in der Lexikographie. Tübingen: Niemeyer.

 

Lönneker, B. (2003): Konzeptframes und Relationen. Extraktion, Annotation und Analyse französischer Corpora aus dem World Wide Web. Berlin: Aka.

 

Minsky, M. (1975): A framework for representing knowledge. In: Winston, P. (ed.): The psychology of computer vision. New York: McGraw-Hill, 211-277.