Bickerton (1981: 58) claimed that, according to his language bioprogram hypothesis, in creole languages the zero-marked verb refers to a present situation (the a-examples) with stative verbs and has a past perfective function with dynamic verbs (examples b) (which corresponds to APiCS value 1, below). The following Belizean Creole examples illustrate this:
This chapter examines how verbs referring to permanent states like ‘can’, ‘know’, ‘love’, or ‘want’ are marked for present reference and how dynamic verbs are marked for (past) perfective aspect (or past reference, if the language has no dedicated perfective aspect marker). If the language has no dedicated perfective aspect marker, a past tense or a perfect tense marker has been retained.
For this feature, it is not important
whether the overt tense and aspect markers are obligatory or optional,
whether only some verbs referring to permanent states are zero-marked for present reference and others are modified by an overt marker (as in the Gulf of Guinea creoles, where some stative verbs are zero-marked and others marked by ka for present reference), and
whether the covert and overt markers retained for this feature have other functions besides referring to present tense in the case of stative verbs and to (past) perfective aspect with dynamic verbs (as with the zero-marked verb in Haitian Creole, which may also refer to habitual situations).
We distinguish the following four values:
|Stative verbs with present reference and dynamic verbs with past perfective reference are both unmarked||38|
|Stative verbs with present reference and dynamic verbs with past perfective reference are marked with the same overt marker||3|
|Stative verbs with present reference and dynamic verbs with past perfective reference are marked differently||26|
|The language has no or only one TAM marker||6|
Value 1 (stative verbs with present reference and dynamic verbs with past perfective reference are both unmarked) is the most widespread value among the APiCS languages and, as mentioned above, corresponds to one of the features of Bickerton’s (1981) language bioprogram hypothesis.
The following Seychelles Creole and Ternate Chabacano examples illustrate value 3 (stative verbs with present reference and dynamic verbs with past perfective reference are marked differently). This value is the normal case in European languages.
Those languages in which the overt marking of the perfective aspect is not obligatory are also subsumed under this value. An example is Gurindji Kriol:
There is a clear areal distribution of the primary contrast between value 1 (unmarked stative verbs have present reference, unmarked dynamic verbs have past perfective reference) and value 3 (dynamic and stative verbs are not marked the same way for the relevant functions): value 1 predominates on both sides of the Atlantic (including the Americas and continental Africa), whereas value 3 occurs mainly in the other areas, from the Indian Ocean to the Pacific – with some exceptions for both values.
Distribution according to the lexifier shows that value 1 occurs in languages with all lexical bases (18 English-based languages, 9 Portuguese-based languages, 6 French-based languages, 3 Dutch-based languages, 4 African or Arabic-based languages). Value 3 occurs somewhat more commonly in the Romance-based languages (11 out of 30) than in the Germanic-based languages (7 out of 30).
From a semantic or functional point of view, value 1 and value 2 are the same: present states and past perfective dynamic situations are marked the same way, covertly with value 1 and overtly with value 2.
Covert marking of both functions is widespread among non-Bantu West African languages (e.g. Yoruba); therefore it is probable that the Atlantic creole languages having value 1 were influenced by their West African substrate languages.
Overt marking of both functions is well represented among Bantu languages: In many of these languages, present states cannot be marked with a present tense marker; they must occur in the Perfect (which is aspectually perfective) or a similar tense. This means that in these languages, states can only be referred to inchoatively: The Perfect tense indicates that the beginning of the state has been achieved. Note that this feature is not restricted to the Bantu-based contact varieties Lingala and Mixed Ma’a/Mbugu, but that it also occurs in non-contact Bantu varieties like Kimbundu, spoken in Angola.
Overtly marked functions are easier to establish than covertly marked functions; therefore it is possible to infer from the languages with value 2 that the languages exhibiting value 1 – or at least some of them – have the same pattern, in the sense that present states can only be referred to inchoatively. This means thus that the zero-marked stative verbs have the same function as the zero-marked dynamic verbs, namely perfective aspect.
The data presented here furthermore show that Bickerton’s feature (which corresonds to APiCS value 1) is the most widespread value among the APiCS languages (about 52%), but that it does not apply to all creole languages: in the APiCS language sample, this feature is absent from 17 creole languages.