Published online by Cambridge University Press: 19 February 2009
Sociolinguists disagree on how to characterize diglossia with respect to the structural relatedness of the H(igh) and L(ow) varieties: Ferguson 1959, 1991 holds that H and L should be distinct but related varieties of language, while others maintain that a continuum model is more appropriate. Both discrete models (Gair 1968, 1992) and continuum models (De Silva 1974, 1979) have been proposed for Sinhala, as spoken in Sri Lanka. In this article, I employ a computer-generated multidimensional graph of relations between varieties of Sinhala to show that the distribution of H and L grammatical features in a sample of naturally occurring texts supports the discrete H and L model more than the continuum model. A rigorous characterization of diglossia as a distinct type of language situation is proposed, based on the notion “functional diasystem.” (Diglossia, Sinhala, Sri Lanka, diasystem, hybridization, continuum, South Asia, standardization)