Semantic Invariance and Variance in Linguistic Analyses

Abstract

This paper was written for a symposium on invariance (The Invariance Condition in Educational Research: Invariance Between Groups, Instruments, Language and Across Time). The philosophical genre of hermeneutical phenomenology provided a perspective for examination of invariance in scientific research and linguistic analysis that applies the Natural Semantic Metalanguage (NSM) Approach. In both instances, a medium (theory and instruments) is constructed a priori on the assumption it will display invariance when taken out of the laboratory. The real world then inscribes the medium in accordance with qualitative differences (variance) in the phenomenon of interest. In this study, the medium is the Natural Semantic Metalanguage Approach and the phenomenon of interest are three Japanese ʻhearsayʼ markers - rashii, sooda and tte. The NSM Approach uses a set of 64 universal and culture-independent concepts. These are termed ʻsemantic primesʼ because they represent innate meanings that are fundamental to human thought. They are indefinable, their meanings so basic that they cannot be broken down any further.The raw data for this study are the meanings of rashii, sooda and tte as expressed in a corpus of eight novels written in Japanese and with English translations. Using the NSM Approachʼs syntactic rules, a combination of primes was used to define each marker. Reductive paraphrases that are simpler than the original words were identified by a process of semantic reduction. The resulting definitions comprised discrete components that defined the respective markers. This NSM Approach analysis illustrates how explicating the differences between similar terms in one language and across more than one language, needs a common medium with specific attributes. The medium requires that meaning be reduced to a level beyond which further simplification is not possible. This medium also limits the number of semantic primes to 64. It is the invariant nature of the NSM Approach that provides definitions that can accurately and consistently reveal qualitative differences between the terms - linguistic variance.