Exploring the grammar of perception
A case study using data from Russian
In this paper, I pursue the distributional hypothesis that the meaning of a word is derived from the linguistic contexts in which it occurs and apply it to verbs of perception. Differently from NLP implementations of the distributional hypothesis, I explicitly limit the range of variables to the grammatical domain and chart the way in which verbs of Vision, Hearing and Touch are used, morphologically and syntactically, in a representative sample of corpus data. Some aspects of experience are so central and pervasive that reference to them has grammaticalized (Divjak 2010; see also Janda & Lyashevskaya 2011; Newman 2008). The aim is, firstly, to determine to which extent a verb’s grammatical context alone allows us to classify utterances according to the perception type, and, secondly, to chart the similarities and differences in the verbs’ preference for morphological markers and syntactic constructions. If contexts are highly specialized, language structure, as it is witnessed in use, could assist sensory impaired speakers in building up viable representations of concepts, even if sensory experience is lacking. If, in addition, similarities between certain sensory perception verbs are high, sensory impaired speakers could use these similarities to perform analogical mapping across senses and ground concepts relating to the impaired sense in a cognate sensory experience. The findings are relevant for concept acquisition and representation in general and for concept acquisition and representation in sensory impaired populations, such as the blind, in particular.