Here’s another fascinating thing I kept meaning to post about: the semantics of control morphology. First, the conventional view, taken from The Languages of Native North America by the great Marianne Mithun.

Salishan Verbal Morphology

The best examples of inflectional and derivational marking of control come from the Salishan languages. In these languages, transitive verbs show obligatory marking of control, as in these examples from Lushootseed:

Many Salishan languages have multiple pairs of verb suffixes primarily distinguished by control, as well as reduplication to make “emphatic non-control”. At the extreme, Thompson (1985) proposes that almost all roots in Thompson River have an inherent control value. But there are interesting hints that control may be less of a pure grammatical voice and more of an aspectual or modal category.

Control as Aspect

For the control as aspect point of view, I’ll summarise the argument from Representing events in Saanich (Northern Straits Salish): the interaction of aspect and valence by Claire Kelly Turner. The crucial point she makes is that, in Saanich, the plain (control) forms imply that an accomplishment culminates, but are compatible with a non-culmination reading:

Culmination as a cancellable implicature only affects durative telic events, not punctual ones (achievements).

The equivalent sentence in English would be nonsensical because verbs like “heal” or “fix” are telic, so the first clause in the unmarked entails that the event culminates, and the only way it could not do so is with some kind of overt imperfective marking.

However, the addition of limited control marking creates transitives whose culmination cannot be cancelled. She quotes Kiyota (2008), who argues that control distinguishes accomplishments (durative, cancellable) from achievements (punctual, non-cancellable) in the transitive system. The following limited control example is ungrammatical:

The semantic range of limited control, including “finally”, “accidentally” etc. can potentially be explained by a difference in aspectual focus from the overall course of an event to its successful conclusion.

Turner says:

“The results of the culmination cancellation test, and the telicity tests (perfect and out of the blue), are fairly clear. Control predicates have culmination implicatures and non-control predicates have culmination entailments, regardless of the volition or agency of their subjects. Thus, this section provides support for the view that “control” is really about situation type (“aspect”), and not agency (Watanabe 2003, Kiyota 2008, Jacobs fc).”

Control As Mood

The previous section mostly focused on the Salish transitive suffixes, but there is a lot of Salishan morphology which has been viewed through the lens of control. One of these is ‘out-of-control’ marking, which is marked by a ka…a circumfix in St’át’imcets. While the transitivity alternation discussed previously is potentially aspectual in nature, this paper by Davis, Matthewson and Rullmann argues that out-of-control marking relates to circumstantial (dynamic) modality.

The authors say:

“The ‘out-of-control’ circumfix ka-…-a in St’át’imcets (Lillooet Salish) expresses an apparently disparate cluster of meanings, including “be able to”, “manage to”, “suddenly”, “accidentally”, and “non-controllable”. In this paper, we present an analysis of the semantics of this morpheme. Our proposal is that ka-…-a encodes circumstantial modality, and that its various meanings all reduce to either an existential (ability) or universal (involuntary action) interpretation.”

The range of meanings covered by ka…a should be vaguely familiar from the previous discussion of transitivity markers:

• Ability
• Manage to
• Accidentally
• Suddenly
• Uncontrolled

They then unify these meanings into two core notions of occurrence without intent (lack of control) and ability. They don’t contend that lack of control is not part of the meaning of ka…a, but the explicitly modal view of its semantics is interesting. Some of their conclusions:

“In doing so, we have also provided support for a striking generalization (reported on in Matthewson et al. 2006, Rullmann et al. 2007) which distinguishes the St’át’imcets modal system from its counterparts in English and other familiar languages. English modals are lexically distinguished by quantificational force (existential versus universal) but are unselective with respect to the modal base. In contrast, St’át’imcets modals show the opposite profile, being unselective with respect to quantificational force but lexically encoding distinctions in the modal base (e.g., epistemic versus deontic).”

“Our conclusions have implications that extend well beyond the grammar of St’át’imcets. To start with, our analysis invites comparison with control phenomena in other Salish languages, which have been regarded as comprising a unified “control system” (see Thompson 1979, 1985). Our work suggests otherwise: it seems unlikely that the modal treatment we have given here for ka-…-a will extend straightforwardly to more typical transitivity-based control alternations, or indeed, to other Salish “out-of-control” phenomena, as exemplified by C2 reduplication (Carlson and Thompson 1982, Kinkade 1982).”

Parallels Elsewhere

Both papers mention an interesting parallel to the semantics of control in Salish. In some Austronesian languages, verbs show a contrast between a ‘neutral’ and ‘ability’ form. Turner gives these examples for Tagalog:

While this alternation doesn’t seem to be as tied to transitivity as some of the Salish morphology, the semantics seem very similar, including a lack of a culmination entailment for the neutral form but not the ability form. What’s interesting here is that the Austronesian terminology is much less focused on control, even though the control implications also appear to be similar.

Conclusions

I’m not sure I have any, except that, like most things in language, the lines are fuzzy and a given form can potentially mark a number of different types of meaning. It seems clear that control in Salish is not just about control, but also interacts with both aspect and wider modal distinctions in a complicated way.

I’d be interested to see the Salish case integrated into a broader, insightful cross-linguistic typology, but I’ve not yet managed to find more than the odd hint of such a thing.