# Is There A Limit To Valency/Transitivity?

The max valency I ever read about is trivalency. However, hypothetically, can't valency extend to an arbitrily length? To extrapolate on this point, for trivalency, Wikipedia gives examples of:

• I bet him 5 quid on "The Daily Arabian".
• I bet you 2 dollars it will rain.

You could extend these through meta means, or just by adding details though.

Ex.

• I bet him 5 quid that her bet on "The Daily Arabian" would win.
• I bet you 2 dollars it will rain on her.

So, would these qualify as more than trivalent or not? If not, why not?

The verb bet has a valency of four. The roles are

1. The bet maker
2. The other person
3. The stake of the bet

We call these the arguments of the verb. While most arguments are nouns, they can also be prepositional phrases, verb phrases, or whole embedded sentences. So the situation argument can be a long complex embedded sentence. Here's an example with a doubly embedded situation.

Adam bet Beth \$10 that [Carol would lose the bet she made with David that [Evelyn would win the marathon this year]]

But even though the bet is so complicated, the base valency of bet still remains four.

In some sentences the base valency can be increased or decreased. For example, it's very natural to leave out the stake argument in long sentences about bets. I couldn't say for sure that four is the maximum valency, but I don't remember seeing examples of verbs with more.

• I want to make sure I understand something properly: "He kicked her the ball" is ditransitive, right? Is "He kicked the ball to her" also ditransitive? I'm just wondering if functors, word order, or other sentence structure factors can change the transitivity of a verb. Aug 1 '18 at 19:51
• @abcjme Yes I think both are ditransitive. I'm not sure I'd say that they change the transitivity/valency, but that when you want to change the valency structure there are multiple alternative structures you can use. Aug 1 '18 at 21:36
• For the above example couldn't an additional argument be "for s.o."? John [NOM] carried two tons of steel [ACC] from London [ablative] to Manchester [illative] by truck [instrumental] for his friend Sharon [DAT]. And how about another? John [NOM], with his friend Bud [comitative], carried two tons of steel ... Aug 6 '18 at 22:10
• The benefactive would normally be considered an adjunct. Do you have a reason it shouldn't be analysed as an adjunct? Aug 6 '18 at 23:58
• Well, in your above example, you state that "the situation being bet about" is an argument and not an adjunct. I think both your example and mine express a reason for an action and therefore either both are necessary or both are unnecessary. "X bet Y Z dollars" works fine without mentioning what the bet is about, right? Aug 10 '18 at 16:29

Of course, whether a nominal is a complement or an adjunct of the verb makes a big difference as to whether it should be counted in valency.

The artificial logical languages Loglan and Lojban permit five arguments for their semantically primitive predicates. (Lojban at least permits more arguments in compounds with subscripted case markers, but it's doubtful that has happened in practice.) Yes, it is artificial, and idiosyncratic in what it deems a predicate argument; but the listing does rely on some semantic analysis of valency, at least.

In the Lojban list of predicates, 18 predicates have five places. The arguments of Lojban predicates are things that I think most linguists would regard as adjuncts in natural language, and the list includes what languages would typically realise as nouns; so it's not unproblematic as a comparison. For example:

• boxna "wave": wave, medium, wave-form, wave-length, frequency (not particularly natural, and nominal)
• cilre "learn": learner, propositions (things learned), subject matter, source, method ("V learned W about X from Y through Z": Z would typically be analysed as an adjunct and the distinction between W and X does not have to be modelled as two distinct arguments)
• plipe "jump": agent, destination, origin, height reached, propulsion (not clear that the height and propulsion are essential arguments).

From the list, the strongest candidates for 5-valent predicates that I see are:

• benji "transfer": agent, object, receiver, origin, means/medium
• bevri "carry": agent, object, receiver, origin, path (via)
• klama "go": agent, destination, origin, route, means/vehicle
• fanva "translate": agent, source text, target language, source language, translation text
• karbi "compare": observer, comparatum, comparandum, property of comparison, result of comparison (though natural language would not express it this way)

karbi is pedantic, and benji/bevri/klama all involve a route or vehicle (or both), which is arguably an adjunct. The strongest of them all, even if it is a verb of literate culture, is fanva "translate": "V translates W from X into Y as Z".

"To carry":

[John] carried [two tons of steel] [from London] [to Manchester] [by truck].

That's a five-valency verb (and so is, at least, "to bring").

• I think it would be highly conventional to analyse those prepositional phrases as adjuncts. What is the rationale for calling them arguments? Aug 5 '18 at 14:11
• In case grammar, anything can be an argument. Aug 5 '18 at 19:06
• I am not sure that anything can be an argument (for instance, I wouldn't say that "today" is an argument in "John carried the steel today"), but "truck" can certainly be an argument for "carry", as in "the truck carried two tons of steel from London to Manchester". If it is an argument in that example, why wouldn't it be when it is the instrument rather than the agent? Aug 5 '18 at 19:59
• Because two sentences having the same noun phrases don't have to have the same structure. For example, think of passivisation. The valency is reduced and the patient becomes the subject. The actor can be introduced in an adjunct if you want to. Or would you not say that the passive actually reduces valency because the actor can be said in a prepositional phrase, and the passive just changes the order? Aug 6 '18 at 0:51