Dimensions of Iconicity

Abstract

This thesis investigates how cross-linguistic phoneme distributions of 56 fundamental oppositional concepts can reveal semantic relationships by looking into the linguistic forms of 75 genetically and areally distributed languages. Based on proposals of semantic primes (Goddard 2002), reduced Swadesh lists (Holman et al. 2008), presumed ultraconservative words (Pagel et.al. 2013), attested basic antonyms (Paradis, Willners & Jones 2009) and sense perception words, a number of semantic oppositional pairs were selected. Five different types of sound groupings were used dividing phonemes according to; the frequency of vowels' second formant and consonants' energy accumulation (Frequency), sonority (Sonority), a combination of the aformentioned two (Combination), general phonetic traits, e.g. voicing (General), and lastly incorporating all traits of the four presented groupings (All). These were analyzed by means of cluster analyses creating biplots, illustrating the phonological relatedness between the investigated concepts. Also, the phoneme distributions' over- and underrepresentation from the average was calculated defining which sounds represented and were lacking for each concept. Significant semantic groupings and relations based solely on phonological contrasts were found for most investigated concepts, including the semantic domains; Small, Intense Vision-Touch, Large, Organic, Horizontal-Vertical Distance, Deictic, Containment, Gender, Parent and Diurnal, and the sole concept OLD. The most notable relations found were; MOTHER/I vs. FATHER, a three-way deictic distinction between I, indicatory deictic concepts and THERE, and a dimensional tripartite oppositional relationship between Small and (possibly with Intense Vision-Touch), Large-Organic and Horizontal-Vertical Distance. Embodiment, benefits of oppositional thinking and evidence for more general concepts to precede complex concepts were proposed as explanations for the results.