Interdisziplinäres Zentrum für Kognitive Sprachforschung

Links und Funktionen



Abstract Schmid

Conceptual blending, relevance and word-formation

The idea of ‘emergent structure’ being generated in the blended space is one of the corner-stones and major assets of conceptual blending theory. However, it also gives rise to one of its major weaknesses because at least in earlier versions of the theory (and in practical applica-tions) the generation of emergent structure by means of the process of elaboration has ap-peared to be open-ended. This has prompted justified criticism that apparently ‘anything goes’ in blending theory, an objection that was countered by Fauconnier and Turner by the introduc-tion of a set of ‘governing’ or ‘optimality principles’, among them ‘topology’, ‘integration’, ‘unpacking’ and ‘relevance’, which constrain the blending process and the likely amount of emergent structure generated.

In this paper it is argued that the importance of the governing principle of ‘relevance’ has been grossly underestimated in blending theory. The principle plays an essential role not only in constraining the blending process, as Fauconnier and Turner suggest, but more importantly in actually triggering it. Conceptual blending is understood as a manifestation of the general human cognitive propensity for Gestalt formation, whose governing principles (or ‘Gestalt Laws’) such as proximity, adjacency, continuation and closure are also claimed to be based on the irresistible urge to search for relevance.

Novel and established compounds will be investigated as a test case. Data on the interpreta-tion of novel compounds is taken from Ryder (1994) who confronted informants with fabri-cated N+N compounds like bottle-flower or elephant-jar and asked them to comment on their meanings. The answers supplied by the informants reflect their search for ‘optimal relevance’ (in relevance-theoretical terms). It is argued that emergent components of the meanings of compounds are the results of inferred contextual implications answering the question ‘why are these two words presented to me close to each other, adjacent, as one internally continuous and externally bounded unit here and now in this context’.

The paper will end with the claim that word-formation patterns such as the various types of compounds or suffixations, and the ‘meanings’ traditionally attributed to them, are entrenched and fossilised blending routines derived from recurrent implication patterns resulting from searches for optimal relevance. The use of a word-formation item (e.g. John is a weather-complainer) instead of a whole predication (John keeps complaining about the weather) brings up the relevance-based expectation that there is a conventionalised cognitive category of entities denoted by the term.



Ryder, Ellen (1994), Ordered chaos. The interpretation of English noun-noun compounds, Berkeley etc.: University of California Press.