Loading...
Publication Date
2020
Start Date
14-11-2020 1:40 PM
End Date
14-11-2020 1:45 PM
Description
While code-switching (CS) is highly studied, common quantifiers such as 'extensive’, ‘abundant,’ ‘frequent’ describing CS1 remain uninterpretable in the absence of numbers, as no clear unit of measurement for CS has been established. Using the New Mexico Spanish-English Bilingual (NMSEB) corpus2—transcribed in Intonational Units (IUs)—we employ prosodic patterns to detail the frequency of CS3.
We compare the number of CS that occur within versus at the boundary of prosodic sentences (Fig. 1, N=3,650) to a baseline rate of all IUs within or at the prosodic sentence boundary (Fig, 2, N=45,578). This comparison demonstrates that our speakers prefer to switch at, rather than within, the boundary of a prosodic sentence. Further, even within the prosodic sentence, speakers capitalize on the prosodic boundaries of the IU. Comparing switches that occur within IUs themselves to those that occur at the boundary of IUs (Fig. 3, N=3,569), we find that 63% occur at this prosodic boundary.
These parallels between prosodic and CS structure support adopting the IU as a basis for calculating an approximate rate of CS. By adding the number of CS that occur at prosodic boundaries (from Fig. 1) to a weighted approximation of the number that occur within the IU (~1/3 of all boundary switches; 1,304/3650) and dividing it by the total IUs in the corpus (from Fig. 2), we arrive at an estimated rate of ~10%. While not intended as an exact rate, it provides a replicable quantification that allows for comparison of CS across communities.
See full abstract linked below.
Video transcript
Trawick, LaCasse-mACOL2020-abstract.docx (35 kB)
Full abstract
Adding Quantities to Quantifiers: Measuring Frequency of Code Switching
While code-switching (CS) is highly studied, common quantifiers such as 'extensive’, ‘abundant,’ ‘frequent’ describing CS1 remain uninterpretable in the absence of numbers, as no clear unit of measurement for CS has been established. Using the New Mexico Spanish-English Bilingual (NMSEB) corpus2—transcribed in Intonational Units (IUs)—we employ prosodic patterns to detail the frequency of CS3.
We compare the number of CS that occur within versus at the boundary of prosodic sentences (Fig. 1, N=3,650) to a baseline rate of all IUs within or at the prosodic sentence boundary (Fig, 2, N=45,578). This comparison demonstrates that our speakers prefer to switch at, rather than within, the boundary of a prosodic sentence. Further, even within the prosodic sentence, speakers capitalize on the prosodic boundaries of the IU. Comparing switches that occur within IUs themselves to those that occur at the boundary of IUs (Fig. 3, N=3,569), we find that 63% occur at this prosodic boundary.
These parallels between prosodic and CS structure support adopting the IU as a basis for calculating an approximate rate of CS. By adding the number of CS that occur at prosodic boundaries (from Fig. 1) to a weighted approximation of the number that occur within the IU (~1/3 of all boundary switches; 1,304/3650) and dividing it by the total IUs in the corpus (from Fig. 2), we arrive at an estimated rate of ~10%. While not intended as an exact rate, it provides a replicable quantification that allows for comparison of CS across communities.
See full abstract linked below.