Preposition drop and ditransitives in British English: a corpus and survey approach

The present thesis maps geographical variation in the relative frequency of
variants of two syntactic variables: ditransitives [DATALT] and preposition
dropping [P/D-Drop]. Several data sources are used: large scale geolocated
Twitter corpora, combined with traditional corpora.
These data are analysed in the context of models of diffusion, and patterns
of historical migration. Comparing the present-day distribution to historical
linguistic data, migration data and social history, allows for the estimation
of the likely time periods of variant innovation and diffusion. Similarly, un-
derstanding the present-day geographical variation of linguistic features also
provides a window into historical migration patterns.
Using mapped data, it is possible to find sharp boundaries, where the rela-
tive rates of variants change abruptly over the space of a few miles, indicating
multiple transition zones. These boundaries are shown to correspond to other
linguistic features. Most importantly, the isoglosses of various manifestations
of definite article reduction [DAR], found in the Survey of English Dialects, are
shown to correspond tightly to northern variants of [P/D-Drop], suggesting a
relationship that has hitherto not been made. The data additionally provide
quantitative evidence to support the diagnostics used in contemporary syntac-
tic studies of the variables in question, and is able to locate, the likely place
where a particular grammar should exist.
The linguistic boundaries also provide numerous opportunities for further
investigation, using additional methodologies. In the present work, a further
study is carried out in the North West of England, using a syntactic judgement
survey, delivered through sixth form colleges in the region. The survey data
are used to triangulate the Twitter data and to test for the existence of the
grammar predicted from the mapped usage data.