The predicates and operators which provide the subdivision of the present section have it in common that they take a bare proposition as their argument. None of them permits marking or changing the clause type of their complement. Such a subordinate proposition may take the form of a complement clause of a non-speech-act verb and consequently a substantive clause unmarked for clause type.

Argument of phase predicate

A phase predicate is a predicate – normally a verb – that takes a bare proposition as an argument and designates a phase of the situation represented by the proposition. There are three principal phase predicates: ‘start, continue, end’. The concept may be expanded to include predicates like ‘happen’, ‘repeat / happen again’ etc. In some languages, these are one-place predicates, as in and .

.k-uchúun-ulinmeet-ik
YucIPFV=SBJ.3start-INCMPL[ SBJ.1.SGmake-INCMPL ]
I start doing it
.k-uts'o'k-olinmeet-ik
YucIPFV=SBJ.3end-INCMPL[ SBJ.1.SGmake-INCMPL ]
I finish doing it

In Yucatec Maya, basic phase predicates are intransitive impersonal verbs. Only if they are causativized may they be constructed personally, as in the English translations, which feature two-place predicates.

The paradigmatic relationship between the construction of and the construction of its English translation may be described by a transformation that has been called ‘subject raising’: What in the Yucatec construction is the subject of the embedded clause, is the subject of the main clause in the English construction.

Argument of modal predicate

A modal predicate is a predicate that takes a bare proposition as an argument and codes the modality of this proposition. There are three principal modal predicates: ‘must, can, want’. In some languages, these are one-place predicates, as in .

.k'abéetinmeet-ik
Yucnecessary[ SBJ.1.SGmake-INCMPL ]
I have to do it
.k-ubeey-talinmeet-ik
YucIPFV=SBJ.3be.possible-INCMPL[ SBJ.1.SGmake-INCMPL ]
I can do it
.taakinmeet-ik
Yucanxious[ SBJ.1.SGmake-INCMPL ]
I want to do it

In Yucatec Maya, basic modal predicates are impersonal, whereas their English translations are all personal and, thus, bivalent. The paradigmatic relationship between the two constructions is as for the phase predicates.

With this difference in construction, both languages obey the same implicational hierarchy:

Implicational hierarchy of personal and impersonal construction of modal verbs
necessity – possibility – volition

The principle of the hierarchy is:

In other words:

See Lehmann et al. 2004 for details.

Propositional argument of a predicate of emotion or volition

Predicates of emotion and volition that take propositional arguments include

The argument of a reactive emotion is factive; the argument of a projective emotion is imagined. is an example of the latter.

.Jill wanted [ Jack to get her an apple ].

Clauses that depend on a predicate of projective emotion typically exhibit an alternation of two syntactic structures:

In the latter case, a construction involving a main predicate of projective emotion may become identical with a construction involving the modal predicate of volition (s. above).

Propositional argument of a causative predicate

An causative verb takes at least two arguments, an animate being as actor, more precisely as agent, and a proposition, more precisely a second-order entity, as undergoer. A causative verb means that the agent brings the situation of the subordinate proposition about. There are two principal ways of doing this, by a speech act or by applying force. A proposition depending on an directive speech act verb is an indirect jussive clause and is treated in the next subsection. Here the verbs of non-linguistic induction remain to be considered. This essentially comes down to causative constructions, as in .

.Linda had Irvin scrub the floor.

Relevant subdivisions here include the following:

  1. degree of control of the causer: coercion () vs. permission (),
  2. degree of direct () vs. distant () involvement of the causer,
  3. degree of control of the causee,
  4. degree of affectedness of the causee (higher in , lower in ).

These parameters are not mutually independent. In particular, the less control the causer exerts (#1), the more control has the causee (#3). And the more directly the causer is involved, the more directly can he affect the causee (#4). This is illustrated by the following example series from Japanese.

.Taroo-gaHanako-okuruma-karaori-saseru
JapTaro-NOMHanako-ACCcar-ABLget.out-CAUS
Taro has Hanako get out of the car.
.Taroo-gaHanako-nikuruma-karaori-saseru
JapTaro-NOMHanako-DAT/LOKcar-ABLget.out-CAUS
Taro lets Hanako get out of the car.
.Taroo-gaHanako-okuruma-karaorosu
JapTaro-NOMHanako-ACCcar-ABLthrow.out
Taro throws Hanako out of the car.

The case variation on the causee NP expresses more vs. less control of the causee, whereas the opposition between lexical vs. regular-derivational causative ( vs. ) expresses degree of involvement of the causer (cf. introductory text on causation).

The most important consequence of parameter #3 is the following implicational hierarchy:

Base predicate hierarchy for causativization
adjectiveinactive intransitive verbactive intransitive verbtransitive verbmultivalent verb

The principle of the hierarchy is the following: If a language has specific causative constructions derived from bases of a category at a given position of this hierarchy, then it has causative constructions on bases of categories to the left of it.

It follows from this hierarchy that causative verbs on inactive bases are relatively widespread, whereas causatives on agentive and multivalent bases tend to be a specialty of languages with a fully productive, possibly purely syntactic, causativization strategy. For derived causative verbs, Yucatec Maya, e.g., is at position 2 of the hierarchy, while Turkish is at the right pole.