In §16.2 we noted that it is common for a single modal form to be used for several different types of modality; but there are also many languages where this does not occur. Even in English, not all modals allow both epistemic and deontic uses. Might is used almost exclusively for epistemic possibility, at least in main clauses.8 Can is used almost exclusively for root modalities, although the negated forms cannot and can’t do allow epistemic uses. What these examples show is that it is possible, even in English, for both strength and type of modality to be lexically specified.
Matthewson (2010) shows that in St’át’imcets (Lillooet Salish), clitic modality markers are lexically specified for the type of modality, with strength of modality determined by context; see examples in (25). In this regard, St’át’imcets is the mirror image of English.
(25) a. wá7=k’a s-t’al l=ti=tsítcw-s=a s=Philomena
be=epis stat-stop in=det=house-3sg.poss=exis nom=Philomena
‘Philomena must/might be in her house.’ [only epistemic]
b. lán=lhkacw=ka áts’x-en ti=kwtámts-sw=a
already=2sg.subj=deon see-dir det=husband-2sg.poss=exis
‘You must/can/may see your husband now.’ [only deontic]
The St’át’imcets data might be analyzed roughly along the lines suggested in (26): the modal markers =k’a and =ka are both defined in terms of a quantifier which is underspecified for strength, but they lexically specify different types (or flavors) of modality:
(26) a. Epistemic = k’a:
‘Philomena must/might be in her house.’ (25a)
[all/some w: (w is consistent with the available evidence) and (the normal course of events is followed as closely as possible) AT_HOME(p) in w
b. Deontic = ka:
‘You must/can/may see your husband now.’ (25b)
[all/some w: (the relevant circumstances of the actual world are also true in w) and (the requirements of the relevant authority are satisfied as completely as possible in w)] hearer sees husband in w
This contrast between St’át’imcets and English provides additional support for the conclusion that either strength or type of modality, or both, may be lexically specified. It is possible for both patterns to be found within a single language. The Malay modal mesti ‘must’ has both epistemic and deontic uses, like its English equivalent. The Malay modal mungkin ‘probably, possibly’ has only epistemic uses, but the strength of commitment is context-dependent, much like the clitic modality markers in St’át’imcets.
Van der Auwera & Ammann 2013 report on a study of modal marking in 207 languages, focusing on the question of whether a single modal form can be used to express both epistemic and deontic modality. They report that this is possible in just under half (102) of the languages in their sample: in 105 of the languages, all of the modal markers are lexically specified as either epistemic or deontic/ root, with no ambiguity possible. Only 36 of the languages in the sample are like English and French, with markers of both possibility (may) and necessity (must) which are ambiguous between epistemic and deontic readings. In the remaining 66 languages there is a modal marker for one degree of strength, either possibility ‘may’ or necessity ‘must’, which is ambiguous between epistemic and deontic readings; but not for the other degree of strength.
The 36 languages which have ambiguous markers for both possibility and necessity are mostly spoken in Europe, and most of them express modality using auxiliary verbs; but neither of these tendencies is absolute. West Greenlandic (Eskimo) is a non-European member of this group which expresses modality with verbal suffixes. The suffix -ssa ‘must’ has a deontic/root necessity reading in (27a) and an epistemic necessity reading in (27b). The suffix -sinnaa ‘can’ has a root possibility reading in (28a) and an epistemic possibility reading in (28b).
(27) West Greenlandic9
a. Inna-jaa-ssa-atit.
go.to.bed-early-nec-ind.2sg
‘You must go to bed early.’ [deontic]
b. Københavni-mii-ssa-aq.
Copenhagen-be.in-nec-ind.3sg
‘She must be in Copenhagen.’ [epistemic]
(28) Timmi-sinnaa-vuq.
fly-can-ind.3sg
‘It can fly.’ [root]
(29) Nuum-mut aalla-reer-sinnaa-galuar-poq …
Nuuk-allative eave-already-can-however-3sg.ind
‘He may well have left for Nuuk already, but…’ [epistemic]
Most of the research on modality to this point has focused on languages of the European type. There is no obvious reason why modal markers in other types of language should not also be analyzed as quantifiers over possible worlds, since (as we have seen) lexical entries for modal markers can specify strength, type of modality, or both. However, this is a hypothesis which should probably be held lightly, pending more detailed investigation of the less-studied languages.
8 In indirect speech-type complements, might can function as the past tense form of may, e.g. Mary said that I might visit her. In such contexts the deontic reading is possible. (See Chapter 20 for a discussion of the “sequence of tenses” in indirect speech complements.)
9 Examples from Fortescue (1984: 292–294, p.c.); cited in van der Auwera & Ammann 2013
