In theoretical linguistics, a converb is a non-finite verb form that serves to express adverbial subordination, i.e. notions like 'when', 'because', 'after', 'while'.

Converbs are differentiated from coverbs, which are verbs in complex predicates in languages that have serial verb construction.

The term converb was coined for Mongolian by Ramstedt (1903) and until recently was mostly used by specialists of Mongolic and Turkic languages to describe non-finite verbs that could be used either for coordination or subordination. Nedjalkov & Nedjalkov (1987) first adopted the term for general typological use, followed by Haspelmath & König (1995).

A converb depends syntactically on another verb form, but is not its argument. It can be an adjunct, i.e. an adverbial, but can neither be the only predicate of a simple sentence, nor clausal argument (i.e. it cannot depend on predicates such as ‘order’ etc.) (Nedjalkov 1995: 97).

Let us examine an example from Khalkh Mongolian:

hün inee-ž ehel-megc zürh anh-and-aa hüčtej cohil-ž aažmaar cohilt-yn hem neg hev-end or-dog baj-na.

human laugh-ž begin-megc heart first-dative-reflexivepossessive strong beat-ž slowly beat-genitive rhythm one form-dative enter-participle_of_habit be-nonpast

“As soon as a human begins to laugh, at first her heart beats strong and slowly the rhythm of the beat assumes one (continuous) form.”

In this sentence, the converb -megc denotes that as soon as the first action has been begun/completed, the second action begins. Thus, the subordinate sentence can be understood as a temporal adverbial. There is no context in which the argument structure of another verb or construction would require -megc to appear and there is no way (possibly except for afterthought) in which a -megc-clause could come sentence-final. Thus, -megc qualifies as a converb in the general linguistic sense. However, from the viewpoint of Mongolian philology (and quite in agreement with Nedjalkov 1995 and Johanson 1995), there is a second converb in this sentence: -ž. At its first occurrence, it is modified by the coverb ehel- ‘to begin’ and this coverb determines that the modified verb has to take this suffix. Yet, this same verbal suffix is used after the verb ‘to beat’ which ends an independent non-finite clause that temporally precedes the following clause, but doesn’t modify it in any way that would be fit for an adverbial. It would even be possible for -ž to mark an adverbial:

Bi … hümüüs-ijn tatgalza-h-yg tevči-ž čada-h-güj gež aj-ž zar-ž ehel-sen.

I people-genitive hesitate-future_participle-accusative bear-ž can-future_participle-negation that fear-ž sell-ž begin-past.

“I started my business, at the very beginning fearing that … I wouldn’t be able to bear the hesitating of the people.”

Such “polyfunctionality” is by no means rare, Japanese and Korean could provide similar examples, and the definition of subordination poses further problems. There are, therefore, linguists who suggest that a reduction of the domain of the term converb to adverbials doesn’t fit language reality (e.g. Slater 2003: 229).

  References

• Haspelmath, Martin & König, Ekkehard (eds.) 1995. Converbs in cross-linguistic perspective. Berlin: Mouton de Gruyter.
• Johanson, Lars (1995): On Turkic Converb Clauses. In: Haspelmath & König 1995: 313-347.
• Nedjalkov, Vladimir P. & Nedjalkov, Igor’ V. (1987): On the typological characteristics of converbs. In: Toomas Help (ed.): Symposium on language universals. Tallinn, 75-79.
• Nedjalkov, Vladimir (1995): Some Typological Parameters of Converbs. In: Haspelmath & König 1995: 97-136.
• Ramstedt, Gustav John (1903): Über die Konjugation des Khalkha-Mongolischen. Helsinki: Suomalais-Ugrilainen Seura.
• Slater, Keith (2003): A Grammar of Mangghuer. London: RoutledgeCurzon.