45 years of unergativity and unaccusativity

The terms unergative and unaccusative were coined exactly 45 years ago, on October 30th in 1976. The person who suggested them was Geoffrey Pullum, in a personal letter to Paul Postal. We know about this from a “column” (a blogpost-like publication in a journal) that Pullum contributed to NLLT (Pullum 1988: 582).

As most linguists (and all typologists) nowadays know, these odd-sounding terms refer to two types of intransitive verbs:

(i) an unergative verb is an intransitive verb that has an agentive argument and is atelic

(ii) an unaccusative verb is an intransitive verb that has a patientive argument and is telic

Typical unergative verbs are ‘talk’, ‘laugh’, ‘dance’, and ‘work’, and typical unaccusative verbs are ‘die’, ‘fall’, ‘melt’ and ‘break’. Unaccusative verbs could alternatively be called “intransitive change-of-state verbs”, and “patientive” in the above definition should be understood as ‘undergoing a change of state’.

But how do these weird terms come about? Can they be made a bit more transparent by considering their etymology? The English Wikipedia makes an attempt, in the article unaccusative verb:

“an unaccusative verb is an intransitive verb whose grammatical subject is not a semantic agent… Unaccusative verbs thus contrast with unergative verbs, such as ‘run’ or ‘resign’, which describe actions voluntarily initiated by the subject. They are called unaccusative because although the subject has the semantic role of a patient, it is not assigned accusative case.”

(The same explanation is given in the Polish Wikipedia: “Czasowniki inakuzatywne są tak nazywane, ponieważ ich niewolicjonalne głębokie dopełnienia paradoksalnie nie przyjmują formy biernika.” ‘Unaccusative verbs are so called because their nonvolitional deep objects paradoxically do not take the form of the accusative.’)

But is this more than a popular etymology?. The term pair unergative/unaccusative was first used in public in David Perlmutter’s (1978) paper “Impersonal passives and the Unaccusative Hypothesis”. In footnote 4, Perlmutter credits Pullum, but he gives no explanation for the terms. However, Pullum (1988: 582) gives an explanation:

This can be understood only in the context of Relational Grammar and its jargon, where “1” stands for a subject, “2” for a direct object, and a “stratum” is a level such as the underlying level or the surface level.

So apparently, Pullum’s starting idea was that in a transitive structure with a subject and a direct object, the subject is “ergative” and the object is “accusative”. The semantic composition would then be as follows: “un+ergative” = “not ergative because intransitive (but ergative-like, i.e. agentive)”, and “un+accusative” = “not accusative because intransitive (but accusative-like, i.e. patientive)”.

This can be understood only in a context where ergativity in the sense in which we know the term now (e.g. Coon et al. 2017) was not widely known yet. It was only through Comrie’s (1978) and Dixon’s (1979) papers that the study of ergativity came to be better known among linguists who were not studying ergative languages. (Of course, ergativity had been known for a long time, and there was an a first edited volume in 1950 (Bokarev (ed.) 1950), but this sort of syntactic typology hardly had an impact on American linguistics before the late 1970s.)

Instead, the term ergative was sometimes used for what we now call labile verbs in the 1970s, and this was due to the influence of Lyons (1968). This book was the first comprehensive textbook of linguistics after Chomsky’s impact came to be felt, and it was hugely influential in the 1970s and beyond. John Lyons (1932-2020) was not a typologist, but he had broad interests, and so he introduced the ergative construction of Inuktitut (1968: 342). But he also used the term ergative in a rather different context, namely for labile verbs (1968:352):

So “ergative”, which was originally a term for a case flag of transitive subjects (as in Dagestanian languages, cf. Dirr 1928; Seely 1977), was extended here to transitive verbs which also have a use as intransitive verbs. And it was from this sense that the term ergative came to be applied to all unaccusative verbs. Burzio (1981: 38-39), a very influential MIT dissertation, originally applied ergative only to Italian labile verbs, but the term was extended from these verbs to all intransitive verbs which have a patientive argument (rather than an agentive argument). Burzio’s work was widely read, and his confusing terminology was adopted by many generative linguists in the 1980s (there was even a book with the title Ergativity in German, Grewendorf 1989, which dealt with unaccusative verbs).

It was only after Pullum’s (1988) detailed explanation of the history (including some personal reminiscences), and especially after Levin & Rappaport Hovav’s (1995) book Unaccusativity, that the original Pullum/Perlmutter terminology was generally adopted. (Some typologists like Dixon 1987 were visibly annoyed at this terminological confusion, but it must be kept in mind that there was no internet in those times, and many linguists did not have access to all information, even if they had wanted to pay attention to work outside their narrow circles.)

So there is a kind of happy end now, because the definitions that I gave above are now very widely accepted. Originally, of course, the definition of “unaccusative verb” was “a verb with a single argument which is underlyingly a direct object”, and the two classes of verbs were identified in practice in terms of “unaccusative/unergative diagnostics”, such as the occurrence with particular auxiliaries (e.g. sein vs. haben in German; cf. Haider 1985, a paper that Haider was working on while he was my undergraduate syntax teacher at the University of Vienna); or the availability of adjectival past participles, or the notorious ne-cliticization in Italian. However, linguists soon discovered “unaccusative mismatches” (e.g. Levin 1985; Levin & Rappaport 1989; Gerdts 1991), and the definition in terms of an underlying direct object was not always helpful in practice. By contrast, a distinction using a semantic definition (in terms of agent role and aletic actionality, and patient role and telic actionality, as above) turned out to be useful in many languages. So we now have a useful term pair which can be used as comparative concepts across languages, using the same criteria in each language. This is not what Perlmutter, Postal and Pullum had envisaged in the 1970s, but it can be said to be a useful outcome.

The original “Unaccusative Hypothesis” was of course formulated in such a theory-internal form (“Certain intransitive clauses have an underlying object but no underlying subject”, Perlmutter 1978: 160) that it could not be tested objectively, and it is very weak anyway (“certain intransitive clauses”). But given the comparative concepts, we can make some proposals for testable universals, e.g.

If a language allows postverbal position of an unergative verb’s subject, it also allows postverbal position of an unaccusative verb’s subject.

Thus, 45 years of unergativity and unaccusativity is a reason to celebrate regardless of one’s methodological preferences, and it is good reason to congratulate David, Paul, and Geoff on their lasting contribution.

References

Bokarev, Evgenij Alekseevič (ed.). 1950. Èrgativnaja konstrukcija predloženija. Moskva: Izdatel’stvo Inostrannoj Literatury.

Burzio, Luigi. 1981. Intransitive verbs and Italian auxiliaries. Massachusetts Institute of Technology. (PhD thesis.) (http://hdl.handle.net/1721.1/15849)

Comrie, Bernard. 1978. Ergativity. In Lehmann, Winfred P. (ed.), Syntactic typology: Studies in the phenomenology of language, 329–394. Austin: University of Texas Press.

Coon, Jessica & Massam, Diane & Travis, Lisa deMena (eds.). 2017. The Oxford handbook of ergativity (Oxford Handbooks). Oxford, New York: Oxford University Press.

Dirr, Adolf. 1928. Einführung in das Studium der kaukasischen Sprachen. Leipzig.

Dixon, R. M. W. 1987. Studies in ergativity: Introduction. Lingua 71(1). 1–16. (doi:10.1016/0024-3841(87)90065-9)

Dixon, R.M.W. 1979. Ergativity. Language 55. 59–138.

Gerdts, Donna B. 1991. Unaccusative mismatches in Halkomelem Salish. International Journal of American Linguistics 57(2). 230–250.

Grewendorf, Günther. 1989. Ergativity in German (Studies in Generative Grammar 35). Dordrecht: Foris.

Haider, Hubert. 1984. Was zu haben ist und was zu sein hat – Bemerkungen zum Infinitiv. Papiere zur Linguistik 30. 23–36.

Levin, Beth & Rappaport Hovav, Malka. 1995. Unaccusativity: At the syntax-lexical semantics interface. Cambridge, MA: MIT Press.

Levin, Beth & Rappaport, Malka. 1989. An approach to unaccusative mismatches. Proceedings of the 19th Annual Meeting of the Northeastern Linguistics Society, 314–328.

Levin, Lori. 1985. Operations on lexical forms: Unaccusative rules in Germanic languages. Cambridge MA: MIT.

Lyons, John. 1968. Introduction to theoretical linguistics. Cambridge: Cambridge University Press.

Perlmutter, David M. 1978. Impersonal passives and the unaccusative hypothesis. Proceedings of the 4th Annual Meeting of the Berkeley Linguistics Society, 157–189. Berkeley: Berkeley Linguistic Society.

Pullum, Geoffrey K. 1988. Citation etiquette beyond Thunderdome. Natural Language and Linguistic Theory 6(4). 579–588.

Seely, Jonathan. 1977. An ergative historiography. Historiographia Linguistica 4(2). 191–206. (doi:10.1075/hl.4.2.04see)


1 thought on “45 years of unergativity and unaccusativity

  1. Thank you! I could never get my head around the terms “unaccusative” and “unergative”. As you noted, Wikipedia doesn’t help much. Now I have something I can hang a peg on to figure out the meaning.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.