Codeswitching in diglossia

John Paolillo

Sun. 11-1:05 C

Myers-Scotton (1993a,b,c) proposes two theoretical models to account for codeswitching: an account of its structural constraints, the Matrix Language-Frame (MLF) Model, and an account of its social constraints, the Markedness Model. Myers-Scotton intends these models to apply broadly to situations including dialect, register and style switching (1993a:2). Yet most code-switching research, including Myers-Scotton's, addresses only language-language codeswitching.

In this paper I consider register codeswitching in Sinhala diglossia. The data are texts collected from contexts in which register switching occurs, such as formal speaking situations, radio and TV broadcasts, personal letters, and children's literature. These are analyzed according to the MLF and Markedness Models. The Sinhala data support the Markedness Model: register switching occurs where it is predicted to from the Rights and Obligations (RO) sets (the social conditions on the use of a particular code) of the two registers. But the data are problematic for the MLF Model: there are numerous counterexamples to the System Morpheme Constraint, under which system morphemes (i.e. grammatical or closed-class morphemes) of the Embedded Language and Matrix Language are not expected to co-occur. The following example from a Sinhala television documentary narrative illustrates the co-occurrence of unambiguously Colloquial (Col) and Literary (Lit) system morphemes:

Col Lit

awurudu de- daas pan -siiyaka-Ta issara Kuweeni-ya

year two-thousand five-hundred-dat before Kuweni -acc

Lit Lit Lit Lit Lit

nul keT -iim-ee yed -ii siT-i bawa lankaa itihaasa

thread spin-ger-loc employ-ppl be -adj that Sri Lanka history

Col Col

kathaa-wee sa~dahan wena-waa

story -loc indicated be-fin

"It is indicated in the story of Sri Lanka's history that Kuweni was engaged in spinning thread two thousand five hundred years ago."

I propose that the apparent discrepancy between language-language codeswitching and the diglossic register switching in Sinhala can be resolved by recognizing that RO sets in the two situations are different: in the Sinhala case, the RO sets of the two registers index social identities that are largely congruent, whereas in a language-language codeswitching case, e.g. Swahili and English in Tanzania, they index distinct and incompatible social identities. Since the RO sets of language pairs are associated with system morphemes in the lexicon (Myers-Scotton 1993b:238), the effect is the same as the System Morpheme Constraint for the language-language case, while less constrained mixing is allowed in the register-switching case. This minor revision of the Markedness Model thus subsumes the MLF Model, resulting in a simplified account of codeswitching that accounts for both language-language switching and diglossic register switching.

References

Myers-Scotton, Carol. 1993a. Social Motivations for Codeswitching. Oxford: Oxford University Press.

Myers-Scotton, Carol. 1993b. Dueling Languages. Oxford: Oxford University Press.

Myers-Scotton, Carol. 1993c. Common and Uncommon Ground: Social and structural factors in codeswitching. [[florin]] 22.4:475-504.