The best-known differential-coding scale is the well-known “extended animacy hierarchy” in (1)

(1) 1st/2nd > 3rd > human > animate > inanimate

This scale can be taken as a basis for two well-known generalizations about semantic-role coding, about differential (or split) object marking (DOM, e.g. Bossong 1991; Aissen 2003) and split (or differential) ergative marking (e.g. Dixon 1994). Informally and in a simplified way, they can be stated as in (2) and (3).

(2) Accusative marking is more likely on nominals on the left of the scale.

(3) Ergative marking is more likely on nominals on the right of the scale.

For this reason, we can call such scales differential-coding scales. They do not predict anything by themselves, but they can be crucial ingredients of generalizations about asymmetric differential coding such as DOM and differential ergative marking.

In the same spirit, Becker & Malchukov (2022: 41) advance the “actionality hierarchy” in (4), a ranking of the four Vendlerian actionality classes.

(4) achievements > accomplishments > activities > states

Like the extended animacy scale, this scale can be taken as a basis for two generalizations about aspect coding, formulated here in (5) and (6) (these formulations seem to be implicit in their paper):

(5) Perfective marking is more likely with verbs of the left-hand actionality classes.

(6) Imperfective marking is more likely with verbs of the right-hand actionality classes.

The paper does not provide a lot of evidence from languages with “differential aspect coding”, but an example comes from Ewenki (a Tungusic language), where the perfective meaning is unmarked with telic verbs such as ‘come’ (example (7), Becker & Malchukov 2022: 47). By contrast, an unmarked verbal form does not have a perfective (and past) interpretation with stative verbs.

On the other hand, to get an imperfective interpretation, quite a few languages use progressive or habitual markers with accomplishment verbs (e.g. English I am writing a letter), while no such marking is needed (or even possible) with stative verbs such as I know the answer (cf. *I am knowing the answer). In their paper, Becker & Malchukov present an image (Figure 8) showing the scale with three cut-off points, implying its generality.

But assuming that the coding tendencies are indeed general across the world’s languages (at least as tendencies, always showing the same direction of asymmetric coding when the coding is not symmetric), what explains them?

Becker & Malchukov make reference to “local markedness”, but they also say that the asymmetries have to do with frequency of use (“local markedness can be observed in the relative frequencies of such value combinations, which is why we also expect marked combinations to be less frequent in language use than unmarked combinations”), and in §4, they go on to test the idea that greater likelihood of coding is due to the low frequency of a grammatical meaning given a lexical class.

That perfective meaning is much more frequent with achievements and accomplishments, while imperfective meaning is much more frequent with activities and states is an important (though unsurprising) result of their corpus study of Russian, Czech, Hungarian and German, summarized in their Figure 9.

Given the efficiency theory of asymmetric coding (Haspelmath 2021a), the tendency for asymmetric coding of perfective and imperfective aspect in accordance with the actionality scale is precisely what we expect. Similarly, the tendencies for asymmetric differential coding of patients and agents (DOM and differential subject marking, as mentioned above) have been successfully explained by asymmetries of usage frequency (Haspelmath 2021b).

In the first part of their paper, Becker & Malchukov mention several additional “interaction hierarchies” (or asymmetric-coding scales, as I would call them), such as the “person hierarchy for imperatives”, as shown (in a simplified way) in (8).

(8) 2SG > 2PL > 1PL.INCL > 3SG/PL

Again, we observe differential coding of imperative governed by “interaction” with person: Second person imperatives are very often zero-coded, while third-person imperative forms (e.g. let her sing!) tend to require longer marking.

And Malchukov (2015) presented evidence for the scale in (9) governing the use of ergative and accusative patterns.

(9) resultative > perfect > perfective > imperfective

It seems to me that the relevance of all these scales for differential-coding tendencies is probably explained by frequency asymmetries, just as in the case of differential object marking and (apparently) differential aspect marking. I do not see any need to refer to “local markedness”, or to “naturalness”, or to “infelicity” (cf. Malchukov 2011) (or to “prototypes”, as in Croft 2003). It is true that we sometimes have a clear understanding why certain associations between grammatical meanings and lexical meanings are used frequently and others are used rarely, so terms like “natural correlations” or “felicitous combinations” may seem intuitive. But ultimately, it seems that what is going on is merely frequency of use which leads to predictability and thus efficient asymmetric coding.

Doing the relevant corpus counts is tedious, and associated with some difficulties (e.g. how exactly does one determine perfective readings in German?) – but I found the results of the present paper very important, because they seem to point in the right direction: Asymmetric differential coding is ultimately explained by efficiency of use, and for this, corpus frequencies provide a crucial source of evidence.

The paper appeared in a special issue on coexpression diagrams  (or semantic maps), so Becker & Malchukov spend some time discussing the similarities between differential-coding scales and “semantic maps”, but I do not see many. It is true that the adjacent meanings on the differential-coding scales often share properties, but this need not be the case, as can be seen in the “person hierarchy for imperatives” as in (8). I would also say that the crucial property of coexpression diagrams is that they diagram coexpression patterns, not that adjacent meanings share features. So I do not really see differential-coding scales (“interaction hierarchies”) as intermediate between implicational scales and semantic maps (and I think that Croft’s 2003 decision to conflate them to a significant extent was not the right way to go).

References

Aissen, Judith. 2003. Differential object marking: Iconicity vs. economy. Natural Language & Linguistic Theory 21(3). 435–483.

Becker, Laura & Malchukov, Andrej. 2022. Semantic maps and typological hierarchies: Evidence for the Actionality Hierarchy. Zeitschrift für Sprachwissenschaft 41(1). 31–66. (doi:10.1515/zfs-2021-2044)

Bossong, Georg. 1991. Differential object marking in Romance and beyond. In Kibbee, Douglas & Wanner, Dieter (eds.), New analyses in Romance linguistics, 143–170. Amsterdam: Benjamins. (https://www.rose.uzh.ch/dam/jcr:ffffffff-c23e-37d9-0000-00006e1a9200/Bossong_80.pdf)

Croft, William. 2003. Typology and universals. 2nd edition. Cambridge: Cambridge University Press.

Dixon, R.M.W. 1994. Ergativity. Cambridge: Cambridge University Press.

Gorelova, Lilia M. 1979. Kategorija vida v evenkijskom jazyke [The category of aspect in the Even language]. Novosibirsk: Nauka.

Haspelmath, Martin. 2021a. Explaining grammatical coding asymmetries: Form-frequency correspondences and predictability. Journal of Linguistics 57(3). 605–633. (doi:10.1017/S0022226720000535)

Haspelmath, Martin. 2021b. Role-reference associations and the explanation of argument coding splits. Linguistics 59(1). 123–174. (doi:10.1515/ling-2020-0252)

Malchukov, Andrej. 2011. Interaction of verbal categories: Resolution of infelicitous grammeme combinations. Linguistics 49(1). 229–282. (doi:10.1515/ling.2011.006)

Malchukov, Andrej L. 2015. Towards a typology of split ergativity: A TAM-hierarchy for alignment splits. In Bornkessel-Schlesewsky, Ina & Malchukov, Andrej L. & Richards, Marc (eds.), Scales and hierarchies: A cross-disciplinary perspective, 275–296. Berlin: Mouton.

---

---