Most languages have valency changing rules. In English and many other languages, we have passive constructions, which change transitive verbs into intransitive ones: "The man ate the hot dog," becomes "The hot dog was eaten," with the agent specified by the adjunct "by the man."

Are there any languages without rules that change the valency of a verb?

  • Passivisation is change of voice, not valency. And 'eat' is a bad example as it's not necessarily transitive. – Gaston Ümlaut Jul 2 '12 at 9:02
  • 5
    @GastonÜmlaut: No, passivisation is a change of valency, because it reduces the verb from two core arguments to one (typically). Voice is a concept which is well-developed in talking about IE languages, but not consistently used for languages in general. Whorf's "nine voices" for Hopi seems to use the word in a different way. – Colin Fine Jul 2 '12 at 10:33
  • 1
    "Valency" is a reasonable metaphor, whatever its terminological history; certainly the analogy with electrons bound to a nucleus works for arguments bound to a predicate, provided one doesn't start positing mesons right and left. – jlawler Jul 2 '12 at 19:40
  • I should add, by the way, that I know of no such languages. Certainly Austronesian languages are striking counterexamples, though the electrochemical metaphor gets a little stretched when it encounters antiergatives and similar delights. – jlawler Jul 2 '12 at 19:43
  • @Gaston: Maybe a better example would be "The machine scanned the man," if we disregard any possible mediopassive reading of "The machine scanned." – James Grossmann Jul 3 '12 at 0:53

If you extend the notion of valency a little, from only verbs to nouns and adjectives (or whatever classes you find in other languages), I guess there's no language without valency-changing rules: Derivation (i.e., change of word class, often signalled in morphology) will probably always involve a change of valency.

  • I don't know why one would extend the definition of valency in that manner. Nouns and adjectives don't take arguments: they take adjuncts, and theoretically limitless numbers of them. Hence "the stone," "the big stone," "the big grey stone," and so on. Hence "happy," "strangely happy," "strangely yet convincingly happy," and so on. What is more, derivation doesn't always change word-class. For example, "happy" and "unhappy" are both adjectives, "sell" and "resell" are both verbs, etc. So I'm afraid I must disagree. – James Grossmann Jan 17 '14 at 0:44
  • One should extend the definition of valency in this way (and even more, if you ask me), because adjunct is a syntactic term, talking about optional appearance, while valency is a semantic term, talking about argument structure and meaning. Every noun derived from a verb takes arguments (argument as in semantics). – virtualnobi Jan 17 '14 at 7:16
  • I'm not sure that the term "adjunct" is exclusively syntactic. This article (en.wikipedia.org/wiki/Adjunct_%28grammar%29) states that “The argument-adjunct distinction is central in most theories of syntax and semantics.” I would also like you to clarify what it means for a noun to have an argument. Would, for example, an obligatory determiner qualify as an argument for a noun? – James Grossmann Jan 18 '14 at 0:25
  • Well, semantics is not my favorite, but are there any obligatory arguments? Using passive, you can get rid of the most important semantic argument... As for arguments on nouns, how would you label "the eating of vegetable"? To me, semantic concepts may be actions, and there are numerous syntactic possibilities (usually verbs, but sometimes nouns and others) to express their meaning. You may certainly posit different concepts for synonyms (because they seldom really have equal meaning), but I cannot convince myself I need different concepts to cover derivation, for example. – virtualnobi Jan 20 '14 at 11:24
  • Syntactically, arguments are necessary by definition: the term applies to constituents that can't be omitted from their matrix constituents. BTW, the importance of the argument omitted in passivization is relative to communicative context and has little to do with either the syntax or semantic structure of the sentence itself. "the eating of the vegetable" is an interesting phrase; is the vegetable an argument of "eating"? I don't know the answer, so you might want to ask this as a question. – James Grossmann Jan 21 '14 at 2:30

Ergative languages like Georgian or Basque have no passive voice, just no voice as a grammatical category. Debates are going on if Sumerian, also an ergative language, had the passive voice, but many agree it had none.

  • 1
    But many ergative languages have antipassive, which is a valency-changing operation. (And of course, there's no reason why an ergative language shouldn't have other types of such operations, e.g. causative or applicative.) – TKR Jan 15 '14 at 22:17
  • In my experience investigating Georgian it's not really an ergative language at all. It has a much more complex mapping between cases and theta roles (if I'm using the terminology correctly). I seem to recall that it has two ways to form passives, with one or both being part of the verb tempate. But while it has several voices I cannot recall whether either or both of its passives are classed as voices. It wouldn't surprise me if such things vary between analyses. – hippietrail Jan 16 '14 at 17:35
  • @ TKR & hippietrail: Thanks for the additional info. – James Grossmann Jan 17 '14 at 0:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.