In example (5) the finite verb mae precedes the subject Americans as is usual
in Welsh, where it is normal for the subject to follow the verb. According to
the MOP, which states that the ML provides the word order, in this example
the ML is Welsh. Additionally, the verb mae has typical Welsh finite subject–
verb agreement, which is identified as an outside late system morpheme and
therefore comes from the ML. Welsh exhibits anti-agreement (Borsley et al.,
2007) so that a plural noun phrase ‘agrees’ with a third-person singular verb.
Note that the Welsh anti-agreement system rather than the English agree-
ment system is being used here, as one might expect in a clause with Welsh
ML. Thus, the SMP also identifies Welsh as the ML. Both principles point to
Welsh as supplying grammatical structure to this clause, with English as the
EL. The following example demonstrates the same model applied to the
Spanish-English data:
(6) ya ahorita estamos almost over
already.ADV in_a_moment.ADV.DIM be.V.1P.PRES
‘We’re almost finished right now.’ (Sastre 1)
In the above example (6) the finite verb estamos shows Spanish
outside late system morphemes (i.e. subject–verb agreement). Note that
Spanish as a pro-drop language allows agreement with a null subject
(Zagona, 2002). This points to Spanish as supplying the grammatical struc-
ture, and therefore the SMP identifies Spanish as the ML with English as
the EL. In (7) below we have an example of English as the ML and Spanish
as the EL.
(7) that car had been chocado de a(de)lante completamente
crash.V.PASTPARTof.PR EPforward.A DVwholly.A DV
‘That car had been completely crushed from the front.’ (Sastre 1)
We extracted all simple clauses (both monolingual and bilingual) from
the 6 conversations, amounting to a total of 1960 finite clauses (1515 mono-
lingual, 345 bilingual) for the Welsh-English conversations and 2611 finite
clauses (2460 monolingual, 149 bilingual) for the Spanish-English conversa-
tions. A quantitative analysis was conducted on the 345 bilingual simple
clauses from the Welsh-English transcriptions and the 149 bilingual simple
clauses from the Spanish-English transcriptions. The dependent variable was
the ML of the clause, with a binary choice of either Welsh or English from
the dataset from Wales and Spanish or English from the Miami dataset. The
independent variables were the extra-linguistic variables listed on the previ-
ous page: age of acquisition of Spanish or Welsh, age of acquisition of English,
language proficiency for Spanish or Welsh, language proficiency for English,
language used in primary school, language used in secondary school, national
identity, and languages of social network.
126 Part 3: Bilingual Language Use