An exploratory description of classifiers in South African sign language

Abstract

This dissertation explores the phenomenon of classifiers in South African Sign Language (SASL). In sign language research, the term classifier has been used to refer to linguistic expressions that describe the specific appearance, movement or location of objects or people (e.g. CAR TRAVELS-ALONG-A-WINDING-ROAD). Due to their highly complex nature, classifiers pose a challenge for categorising them according to existing lexical categories (such as nouns, verbs, adjectives, and so on).

Building on Schembri’s seminal work on classifiers - polycomponential verbs (2003) - this research attempts to offer insight into the correlation between some of the formational features of classifiers in SASL and the three concept classes identified by Meir (2012) – Entity, Event and Property. They roughly correspond to noun, verb, and adjective, respectively, but the use of concept classes is preferred in order to avoid the transfer of pre-existing grammatical categories identified in spoken languages to sign languages.

I collected an overall of 50 classifiers from five SASL signers who are native or near native in their language competency. These classifiers are divided into three groups – one-handed classifiers (Group A), two-handed-classifiers with one referent (Group B) and two-handed classifiers with two referents (Group C). I used the Berkley Transcription System (BTS) to transcribe the data. The properties I have chosen to look at are: handshape, eyegaze focus, face and mouth actions, and movement. I also looked at three traditional categories of classifiers (handling, entity and SASS). The research was two-fold: to provide detailed description of the formational features of classifiers in SASL, and to examine them in terms of three concept classes. Although this study is small-scaled and does not comply with criteria for inferential purposes, it suggests that it is possible to distinguish the three concept classes of classifiers based on some of their formational features. The features which seem to contribute most are eyegaze focus and face and mouth actions. Some face and mouth actions only occur in one particular concept class – such as the combination of ‘blowing gently’ and ‘squint’ only appear in Entity (for classifiers indicating quantity and extent) and ‘expel air’ can only be seen in Event.

This research also highlights some of the issues and challenges posed by the BTS. BTS follows a semantic approach, making provision for meaning components. However, some of their descriptions of handshape are phonologically based, and this inconsistency impacted on the identification of concept classes in my data. Based on this experience, I suggest the importance of maintaining the integrity of the BTS by using semantically based symbols for all handshapes