That the causative psych-verbs in (484) are morphologically complex is, of course, undisputed for the deadjectival verbs in (488). Note that the prefixes ver- and be- can also express causation if the base adjective does not refer to a mental state, as is in ver-edel-ento ennoble and be-vochtig-ento moisten.

The fact that many of the verbs in (484) are prefixed with ver-, be- and ont- can be better understood by taking these affixes to be responsible for the causative meaning aspect in all these cases.

Some verbs can be used as both unaccusative and transitive verbs. A prototypical verb that exhibits this alternation is brekento break, which can be used both as an inchoative, unaccusative verb and as a causative, transitive verb. It has been claimed that the causer is introduced by a zero-morpheme attached to the (simple) unaccusative verb; cf. Section 3.2.3 for further discussion.

Although Section 2.5.1.2, sub III, has shown that there are only a few unaccusative psych-verbs, the same alternation can be found with psych-verbs. The examples of (490) show this for the verbs kalmeren/bedarento calm down in (469a&b).

The unaccusative verb schrikkento be frightened in (469c) does not participate in this alternation, but it is still possible to derive a causative form from it by using the prefix ver-, resulting in the somewhat obsolete causative verb verschrikkento frighten (known mainly in its adjectival participial form verschrikt in de verschrikte vogelsthe frightened birds and as part of the compound noun vogelverschrikkerscarecrow). The somewhat formal example in (491b) is relevant, however, in that the prefix ver- may be an overt counterpart of the postulated phonetically empty causative morpheme that derives the causative forms in the primed examples in (490).

Unaccusative psych-verbs with particles all have causative counterparts. Since the particle is supposed to function as a kind of predicate, the primed examples in (492) can probably be regarded as equivalent to the causative non-psych-construction Jan breekt het glas in stukkenJan breaks the glass to pieces.

Note that it is not the case that all causative psych-verbs have an unaccusative counterpart; the remaining verbs in (484) do not or only with difficulty.

Many causative psych-verbs are Latinate, or at least Romance, forms ending in -eren. Although there are no attested words from which these verbs are derived, it seems plausible that they are derived from nonverbal stems by affixation with the causative morpheme -eren. Table (493) shows that these postulated nonverbal stems can also be used to derive nouns and adjectives; cf. De Haas & Trommelen (1993:348) and Booij (2002:127-8).

The idea that -eren is or can act as a causative morpheme is supported by the fact, illustrated in (494), that it also derives causative object-experiencer verbs from nouns.

The previous subsections have shown that for many causative psych-verbs we may assume that a causative affix is present and that the verb is therefore complex. Subsection III will show that syntactically these psych-verbs resemble periphrastic causative constructions such as bang makento frighten in (495b) with the causative verb makento make; this can be taken as support for the claim that causative psych-verbs are morphologically complex

Note, however, that the presence of a (possibly phonetically empty) causative morpheme is not immediately plausible in all cases. For example, the psych-verbs in (496) are probably denominal, but as far as we know there is no reason to assume that the verbal ending -en is causative in nature. -

Moreover, the psych-verbs in (497a) do not seem to be derived at all, since there does not seem to be a base form that can be regarded as the input of the verb (at least in present-day Dutch). Of course, we can make a similar assumption for these verbs as for Latinate verbs such as irriterento irritate, and claim that they are derived from stems that occur only as bound morphemes. The verbs, adjectives and nouns in (497) can then all be seen as directly derived from this stem.

On this analysis, the two sets of verbs in (496a) and (497a) would form only a single problem for the assumption that all causative psych-verbs are complex.

The first analysis, according to which (504b') is a passive construction derived from an active sentence with a causer subject, implies that the passive door-PP corresponding to the causer is implied; this leads to the false prediction that (506a) should be acceptable. Another prediction is that the participle is verbal, and thus must be able to appear after the finite verb in clause-final position (which is impossible with predicative adjectives); judgments on example (506b) seem to vary from speaker to speaker, but we tend to think that this prediction is indeed correct. If (506b) is indeed grammatical, we end up with an ambiguous result. One way to resolve this is to say that the unacceptability of (506a) is due to a problem with having two door-PPs in a single clause; we leave open whether such a constraint is desirable.

Another prediction that would follow from the first analysis is that passivization of a causative psych-verb is only possible if the verb can take a causer subject. Since the verbs in (484b) cannot easily take a causer subject, these verbs can be used to test this prediction. And indeed, it seems that at least some of these verbs categorically resist passivization; the unacceptability of the examples in (507) therefore supports the proposal that (504b') is derived from an active construction with a causer subject. Note that we have placed the participle after the finite verb in the primed examples in (507) in order to exclude the adjectival passive reading.

The fact that the first analysis is (partly) supported by the facts in (506) and (507) does not preclude the analysis of (504b') as an adjectival passive, i.e. a construction in which the past/passive participle is used as a predicative adjective. Such an analysis is certainly viable, since the verb worden is used both as a passive auxiliary and as a copular verb. That this may be the correct analysis in many cases is also supported by the fact that many participles of causative psych-verbs can enter copular constructions with the verb rakento get, which is never used as a passive auxiliary.

That we are not dealing with passive constructions in (508) but with adjectives is also supported by the fact that the participles can easily be coordinated with true adjectives; cf. Jan raakt/wordt [gedeprimeerd en angstig] door dit donkere weerJan is getting depressed and anxious by this dark weather. Note that (508c&d) are causative psych-verbs that (preferably) take a cause subject, which also rules out an analysis of these examples as passive constructions; cf. the discussion of (507).

The discussion in the previous subsections has shown that passivization of causative psych-verbs is only possible when the subject is a causer, not when it is a cause. Note that the crucial fact here is not the animacy of the subject. In (509a), the subject is animate, but what is really being expressed is that it is the whining that irritates the speaker, which shows that the subject is a cause. And (509b) shows that, as long as we stick to this interpretation, the passive construction in (509b) is excluded (but still marginally acceptable if the context makes it clear that the cause of the irritation is something other than the whining). Example (509c) shows that the adjectival construction is perfectly acceptable.

In general, present participles of verbs can be used attributively to modify nouns corresponding to the subject of the verb. The examples in (510) show that the verb can be intransitive, (di)transitive, or monadic/dyadic unaccusative.

Causative psych-verbs simply follow this pattern: The examples in (511) show that the modified noun can correspond to a causer or a cause subject. Since the experiencer object is realized in the primed examples of (511), we can safely assume that the present participles are verbal in nature. This conclusion is also consistent with the fact that these examples are interpreted as referring to an ongoing event; cf. the English renderings of these examples.

Present participles of causative psych-verbs have the special property that they can also be used purely adjectivally, i.e. as property-denoting elements (without any aspectual meaning). In such cases, however, the modified noun cannot correspond to a causer; if the modified noun corresponds to the cause, on the other hand, the result is perfectly acceptable. This is illustrated in (512).

Note that we have added the modifier ergvery to emphasize the purely adjectival nature of the present participle kwetsend. Without this modifier, example (512a) may be marginally acceptable to some speakers with a verbal reading; the marginality is then due to the omission of the experiencer argument. In this context, it may be interesting that we have found on the internet a small number of occurrences of kwetsende ouder(s)/vader/moederhurting parent(s)/father/mother with this agentive reading, in which the omitted experiencer was clearly the implied internal argument of the relational noun ouders, i.e. their child(ren); cf. Section N15.2.2.

That (512a) is unacceptable due to the purely adjectival nature of the present can be supported by the examples in (513), which show that the present participles of a well-defined set of psych-verbs cannot receive a purely adjectival reading. An example is the present participle irriterend, which is interpreted with a verbal reading regardless of whether the arguments are expressed or not; the purely adjectival reading seems to be blocked by the fact that Dutch already has an adjective that expresses this meaning, viz. irritantirritating. The number sign in (513a'') is used to indicate that the example is only unacceptable if the modified noun is interpreted as a causer: it is perfectly acceptable in the cause reading of jongen, which is not intended here. This supports the claim that (512a) is unacceptable due to the purely adjectival nature of the present participle.

Present participles of most verb types cannot be used in predicative position, as shown in (514) for the same set of present participles used attributively in (510).

The present participles of causative psych-verbs, on the other hand, do allow predicative use of the present participle. Given our conclusion from the previous subsection that present participle of causative psych-verbs can be purely adjectival, this is not remarkable as this simply predicts that present participles like kwetsendhurting can be used in the same way as an adjective like irritant. In (515), the noun phrase that the adjective is predicated of is necessarily interpreted as a cause, just as in (512) and the doubly-primed examples in (513).

Nevertheless, it should be noted that there are a small number of cases on the internet that seem to contradict our judgments about the (a)-examples; we will consider these to be metaphorical (totum pro parte) expressions for the moment (and perhaps there are also some occasional linguistic errors). To give an example based on two fragments from the internet (March 17, 2024): the subject in [Hij] is onnodig kwetsend tegen de islamHe is needlessly offensive to Islam may stand as a shorthand for the more extensive subject in ... wat [hij] zoal over moslims zegt is kwetsend... what he says about Muslims is hurtful/offensive.

Example (516) shows that the past/passive participles of causative psych-verbs can be used attributively to modify a noun corresponding to the experiencer object in the corresponding verbal construction. Both the causer and the cause subject of the verb can optionally be expressed as the complement of a door-PP.

Since attributively used past/passive participles are normally used to modify a noun corresponding to the theme argument of the verb, the question arises as to whether objects in causative psych-verb constructions should be characterized as experiencers or whether it would be more appropriate to simply characterize them as themes. This question seems to become more urgent when we take into account that past/passive participles of nom-dat verbs modify the theme-subject, not the experiencer object of the verb; cf. examples (107) and (108) in Section 2.1.3, sub D.

However, this question is perhaps premature: objects of causative psych-verbs may in fact be neither experiencers nor themes. This claim is related to the proposal, discussed in detail in Section 2.5.1.3, sub V, that causative psych-verbs have an underlying structure similar to that of periphrastic causative constructions such as in (517a). If it is true that causative psych-verbs are always morphologically complex, it seems plausible that the object is not an internal argument of the verbalizing suffix -eer, but an inherited external argument of the nonverbal stem irrit-; this gives rise to the underlying structure in (517b). This structure is very similar to that of the periphrastic causative construction in (517a), in which the object is also an external argument of the adjective kwaadangry, and not an internal argument of the causative verb maken. Since this decomposition analysis of the causative psych-verbs deprives the question as to whether we are dealing with a theme or an experiencer of any theoretical or descriptive significance, we will not pursue this issue any further and will simply continue to use the descriptive label “experiencer” for the object of these causative psych-verbs.

For completeness’ sake, example (518) shows that the complex phrase boos gemaaktmade angry can also be used attributively and behaves in this respect like geïrriteerd in (516b), which is also a complex phrase according to the proposal under discussion.

We will not discuss here the various technical ways of explaining that the stem irrit- and the suffix -eert in (516b) appear as a single verb form; Subsection III will return to a possible syntactic proposal involving incorporation, i.e. head movement of the stem into the verbal affix. Other proposals are based on the hypothesis that vocabulary items are inserted post-syntactically on the basis of more abstract information provided by the syntax; cf. Noyer (1999).

Past/passive participles can be used predicatively, provided they are truly adjectival; cf. Section A31.3. The examples in (519) show that virtually all past/passive participles of causative psych-verbs have this option; since rakento get is not a passive auxiliary but a copular, the participles in these examples must be truly adjectival. The examples in (519b&c) further show that these adjectival participles differ from causative psych-verbs in that they can take a PP expressing the subject matter of emotion; cf. the discussion of the examples in (485).

It is a little more difficult to show that past/passive participles can be truly adjectival with verbs such as zijnto be, which can be used both as a copular verb and as a passive auxiliary; cf. the discussion of worden in Subsection D above. Recall from Subsection A that the verb interesserento interest is special in that it allows a voor-PP expressing a target of emotion; we show this again in (520a). Now consider the construction with zijn in (520b), in which it is also possible to use the preposition in to introduce a target of emotion (we will return to the reason for the marked status of the voor-PP in the next subsection). The fact that in is the only option in the adjectival passive construction in (520c) supports the conclusion that the past/passive participle in (520b) can be truly adjectival.

Further evidence that the choice of PP signals whether we are dealing with a verbal or an adjectival past/passive participle can be found in (521). If we are dealing with an in-PP, the participle clearly shows adjectival behavior: it can be modified by the adverbial modifiers heelvery and zeervery, as in (521a), and allows a comparative/superlative form, as in (521b); it can be prefixed with the negative affix on-, as in (521c); finally, the PP-complement in zijn verhaal can be placed between the participle and the finite verb in clause-final position, as in (521d), which is never possible when the participle is verbal. All examples in (521) become unacceptable when the preposition in is replaced by voor.

The question mark in (521c) indicates that some speakers consider ongeïnteresseerd in marked compared to niet geïnteresseerd in: however, since the first form is common on the internet, the choice between these forms may depend on individual preference.

The previous two subsections have shown that past/passive participles of causative psych-verbs can have either a verbal or an adjectival reading. However, there is a strong tendency to construe the participle as nonverbal (which explains the marked status of example (520b) with the preposition voor). To show this, we will discuss the results of two tests, discussed in detail in Section A31.3.1, sub I, to distinguish between the two readings.

The first test involves temporal modification. The main difference between verbal and adjectival past/passive participles is that the former denote perfective events, whereas the latter denote a property of a noun phrase. This is reflected in the fact that the two types of participle co-occur with different kinds of temporal adverbial phrases; verbal participles can be combined with adverbial phrases such as gisterenyesterday, which refer to a certain time interval during which the event was completed, whereas adjectival participles instead combine with adverbial phrases such as al jarenfor years, which refer to a larger continuous span of time during which the property denoted by the participle holds. If we now consider the data in (522), the degraded status of the examples with gisteren shows that attributively used participles of causative psych-verbs are clearly construed as adjectival.

The verbal reading of the attributively used participles may arise when they are accompanied by an agentive or causative door-PP, but even then the examples in (523) seem marked.

However, not all past/passive participles of psych-verbs allow modification by al jaren. Exceptions are the verbs rakento affect and treffento move and verrassento surprise, which may be related to the fact that these psych-verbs are semelfactive, i.e. denote events that happen instantaneously and have no extension in time: since properties usually last for a longer period of time, the formation of an adjectival participle based on these verbs arguably results in a semantically incoherent meaning. The number sign in (524a) indicates that gisteren is possible if the verb is interpreted as “hit (by e.g. a bullet)”; an interpretation of the verb as a causative psych-verb seems impossible.

The second test concerns the position of the participle in clause-final position. Examples with the verb zijnto be are expected to be ambiguous between a passive reading and a reading in which the participle is predicative; cf. Subsection D. In the passive reading we are dealing with a verbal participle, and we therefore predict that it can be placed after the finite verb in clause-final position. However, the examples in (525) show that this prediction is incorrect: the participle must precede the finite verb, from which we can conclude that it is adjectival.

The unacceptable examples in (525) improve somewhat when a passive door-PP is added, as in the primeless examples in (526), but even then the result is often dubious. The primed examples show that the addition of a causative door-PP cannot be used to evoke the verbal reading of the participle. It seems that acceptability judgments vary from case to case and from speaker to speaker.

The passive reading is only perfectly acceptable if the passive auxiliary worden is used. Note that the passive door-PP is then optional. The primed examples show that the use of a causative door-PP often has a degrading effect on the passive construction. This seems to support our earlier conclusion from Subsection D that causative psych-verbs with a cause subject cannot be passivized; the cases in the primed examples that are acceptable are adjectival passives.