In a number of creole languages, temporal clauses can be expressed by constructions in which the verb is copied or reduplicated, as in (1) and (2).
Verb-doubling constructions occur widely in Atlantic creoles and pidgins to express focus (see Chapter 105), but they are rare in the function of marking temporal clauses. In our sample, they are found only in seven languages (plus an obsolete variety of an eighth language, see below):
Verb doubling is possible in temporal clauses | 7 | |
Verb doubling is not possible in temporal clauses | 67 | |
Representation: | 74 |
Verb doubling in temporal clauses is found in five Atlantic languages, as well as in Sango and Sri Lankan Malay (see (2) above). In addition to Haitian Creole, we find it in Guadeloupean and Martinican Creole:
Verb doubling is found in Berbice Dutch:
In Nicaraguan Creole, it is not found in the modern language studied by Bartens (2013a), but it was attested at an earlier stage:
Finally, we find it described for a variety of Cape Verdean Creole:
According to Lefebvre (2011: 21-22), such constructions are also found in Saramaccan and Papiamentu, but our contributors on these languages did not confirm this.
The verb doubling constructions illustrated by (1)-(6) are fairly heterogeneous. In particular, the Sri Lankan Malay pattern in (2) is divergent, as it shows adjacent doubling, i.e. a pattern that is best described as reduplication. (Note that in Chapter 105 on verb doubling and focus, we also find reduplication as a special rare type.)
The other five constructions share the order “verb1 – subject – verb2”, but the simplest pattern is found only in the French-based creoles of Haiti, Guadeloupe and Martinique. In the Nicaraguan example (5), we see tense marking in the first instance of the verb, which is impossible in Haitian (Lefebvre & Ritter 1993: 68).
In Berbice Dutch, the first instance of the verb seems to be a kind of nominal form, as it is preceded by the definite article di and followed by the relative marker wat, so (4) is literally ‘the turning that you turned’, which looks like a kind of cognate nominalized object. However, the relative clause marker is not obligatory (Kouwenberg 2013a):
In Haitian Creole, too, the first instance of the verb can take a definite article (postposed la/a) (cf. 8a), so Lefebvre (1998: 369) regards it as a kind of nominalized form. However, the situation in Haitian is complex, and the definite article may also follow the second instance of the verb (cf. 8b).
A nominalization interpretation is readily available also for the Cape Verdean Creole example in (6), which can be translated literally as ‘in the eating that they ate’. However, it must be kept in mind that this nominal occurs only in this construction (otherwise ‘eating’ is kumida).
It should be noted that at least in Haitian Creole, such verb-doubling constructions can also have a causal sense (‘because Jean arrived’) and a factive sense (‘the fact that Jean arrived’) (see Lefebvre 1998: §12.6).
According to Lefebvre (1998: §12.6), very similar patterns are found in Gbe languages, as shown by (9) from Fongbe (a Kwa language spoken in Benin) (see also Lefebvre & Brousseau 2002: ch. 16):
We know of no comparative research on this kind of doubling construction in West African languages, and it remains to be seen whether the substrate explanation extends to the other Atlantic creoles. For Sri Lankan Malay, Nordhoff (2009: §7.5.4) notes that this pattern has been attributed to Sinhala influence, so here we have a non-lexifier source as well.