DEFINING THE TARGET LANGUAGE IN LANGUAGE GENESIS 65

 lexemes in MC in order to support the superstrate hypothesis. In order
 to communicate within this heterogeneous language community,
 immigrants had to find words which other members of the community
 would understand. French was the only language to which the entire
 community had some varying degree of limited exposure. Logically,
 French words were more likely to be understood in encounters between
 people who did not share a common ancestral language. In other words,
 the majority of the words used within this community were likely to
 have been French regardless of the linguistic aims of the community.
 "It is not the proportion of MC words which are of French origin which
 is significant so much as the ways in which those words differ from
 French in pronunciation, form and function" (1994:69).
 Baker claims that phonological, morphological and sociohistorical
evidence tends to support his conception of creole genesis (i.e. the
'creativist' model). Evidence of substrate influence noted is mainly the
phonemic inventory, which appears to represent a selection of common
sounds to the majority of languages represented in Mauritius in the
18th century. This strategy would have minimized the number of novel
phonemes to be acquired by the linguistically diverse population. But
there is no evidence to suggest that non-Francophone immigrants
relexified their original languages with French words. Baker notes that
while the MC article system has French etyma, the system as a whole
is clearly not derivable from any French variety, nor does it closely
resemble any language introduced by the CFP. In other words, the article
system was clearly created in Mauritius and supports the creativist
model. A subset of MC's TMA system can be attributed to East African
Bantu languages, partially supporting the substratist position. But the
TMA system allows the possibility of combining two or more preverbal
markers in a single predicate, which provides some support for the
universalist position. The MC verbal system, which emerged during
the first half of the 19th century, has two forms, neither of which
encodes any temporal or aspectual information. When the verb governs
an overt element, the short form is used. The long form is found in all
other contexts (e.g. mo sante 'I sing sing/sang/(am/was) singing' vs.
mo sant en sante 'I sing a song'). Texts indicate that this was clearly an
innovation by the CFP.
 After reviewing a range of features from MC, Baker concludes that
the creativist model was far more consistent than any of the other
three positions. The three features attributed to Bantu and Malagasy
origin were all valuable vocabulary building devices. That is to say,
non-French strategies were employed precisely where they expanded