In quite a few languages, epistemic possibility may be expressed by the same verb as ability. For example, German kann kommen can mean ‘is able to come’ (ability) or ‘may come’ (epistemic possibility).
By ability, we refer broadly to various non-epistemic possibility types comprising mental participant-internal ability (French savoir ‘know, can’: Il sait nager ‘He can swim’), physical participant-internal ability (‘He can lift 100 kilos’), and participant-external possibility (“She can go to town by bus”, i.e. because there is a bus connection).
Epistemic possibility concerns the speaker’s judgement as to the truth value of the sentence, as exemplified by the English adverbs ‘perhaps’ or ‘possibly’.
See also WALS feature 76 (“Overlap between situational and epistemic modal marking”, van der Auwera & Ammann 2005), which is not concerned only with possibility, but also with necessity.
We distinguish the following three values:
|Ability verb also expresses epistemic possibility||32|
|Ability verb cannot express epistemic possibility||41|
|No ability verb||2|
Value 1 (the ability verb can also express epistemic possibility) occurs in 42% of the APiCS languages. It is present in ten Ibero-Romance-based languages, in seven English-based languages, in six French-based languages, in three Dutch-based languages, as well as in Lingala, Mixed Ma’a/Mbugu, Sango, and Sri Lankan Malay. In the examples below, (a) shows an ability verb, and (b) shows its use to express epistemic possibility.
Value 2 (the ability verb cannot express epistemic possibility) occurs in 55% of the APiCS languages, and value 3, there is no ability verb in the language, concerns only Chinuk Wawa and Yimas-Arafundi Pidgin.
From the Atlantic to South Asia, the distribution of value 1 (ability verb can express epistemic modality) and value 2 (ability verb cannot express epistemic modality) is about equal, but value 1 is virtually absent from Southeast Asia and the Pacific, the only exception being Cavite Chabacano.
Interestingly, not all ability verbs in the English- and French-based languages are derived from the lexifiers’ ability verbs (English can, French pouvoir). In English-based creoles and pidgins, we find sa (< Portuguese saver?) in Nengee and Saramaccan, fit (< English be fit) in Cameroon Pidgin English, Ghanaian Pidgin English, Nigerian Pidgin, and Pichi, save (< Portuguese saber ‘know’) in Bislama, ebul and ell (English < be able) in Krio and Norf’k, and inap (English < enough) in Tok Pisin.
French-based creoles often show verbs derving from French être capable de ‘be able to’: kapab/kab in Haitian, kapab in Louisiana Creole, Reunion Creole, and Seychelles Creole, as well as kapav in Mauritian Creole (see also Kriegel et al. 2003). Tayo shows mwaya (French < avoir les moyens de ‘have the means to’). These expressions all denote only ability in the lexifiers, so the extension to epistemic use must have occurred at a later stage.
In some APiCS languages, the complement of the modal verb may take tense and aspect markers when the ‘can’ verb has epistemic meaning, as in example (1b) or in the following example:
In Early Sranan, according to van den Berg & Bruyn (2013), “[t]he epistemic reading of kan seems possible only with progressives, and perhaps statives, including the copula de.” In other languages, e.g. Papiamentu, the present (imperfective) marker ta, the past imperfective marker tabata, and the perfective marker a may modify the complement of the modal verb; only the future marker lo is precluded (see Maurer 1988: 277-290).
This raises the question of the syntactic status of ‘can’ (and other modal verbs): Is it a verb, or should it rather be considered as belonging to the set of tense, aspect, and mood markers? Of course, there can only be language-specific answers to this question. In Papiamentu, for example, there are two arguments that point to the verbal status of por ‘can’: it may stand alone, and it may itself be modified by the future marker lo or the past imperfective marker tabata (see ex. 1b).
The syntactic status of modals and the possibility of modifying the complement of modal verbs with tense and aspect markers merits a thorough cross-linguistic investigation.