Log in / Register
Home arrow Language & Literature arrow Extending the Scope of Construction Grammar

Case study 1: Case frames in Dutch experience predicates

The delicate case-frame network laid out in Table 2 came under pressure when the case system of Middle Dutch broke down in a process that is commonly called “deflection”. When dative- and genitive-marked objects die out, the constructions that deviate from the canonical transitive construction become indistinguishable.

As is well-known, the loss of case in Dutch was compensated for by the increased use of prepositions. It is easy, however, to be misled by this metaphor of “compensation”. In its naive form, it could be taken to suggest that prepositions arose when the case system broke down. This does not conform to the actual facts, however. Prepositions already existed in Middle Dutch, and were used in combination with case. A more accurate way of describing the actual development is that an already existing strategy was exploited by language users. This is a clear case of degeneracy: prepositions were not specifically designed as a trade-off means of expressing case in argument realisation, and the two strategies are not always completely interchangeable, but there is a certain amount of functional overlap so that the language system can sustain loss of one of the strategies. By way of illustration, take example (18), where the impersonal construction is realised by encoding the stimulus as a prepositional object introduced by van which is equivalent to the genitive in other contexts, such as possession, as well.

(18) Hen alien wonderde van dien

they:DAT all:DAT amazed of this:DAT

‘They were all amazed by this.’ (MNW s.v. wonderen)

Still, this strategy was not exploited in full, as the substitution of a dative by a prepositional object introduced by aan or voor does not appear to occur. (If the preposition aan occurs, it is one of the alternatives to the van preposition for encoding the stimulus, see 19 and below).1718

(19) Alle konsten daer men sich aen verwondert

all arts where one himself to amazes

‘All arts by which one is amazed.’ (WNT s.v. verwonderen (I))

There is one construction in present-day Dutch that closely resembles the Middle Dutch impersonal construction, as neither of its arguments is encoded as subject, see (20) and (21). The subject role is taken by a dummy het (‘it’), the stimulus argument is marked by the preposition aan (‘to’) and the experiencer is encoded as an object, with oblique forms in case of pronominal realisation, as in

  • (20) . The construction is odd as it does not occur with those verbs that were most strongly associated with the impersonal construction in Middle Dutch, such as (be/ver)wonderen, but rather with a few verbs that belong to the stimulus- subject category: ontbreken, mangelen, schorten, all of which can be glossed ‘lack’.
  • (20) Het ontbreekt hem aan geld.

it lacks him to money

‘He lacks money.’

(21) Het ontbreekt die mensen aan geld.

it lacks those people to money

‘Those people lack money.’ [1] [2]

To summarise, while degenerate marking by preposition existed and was exploited to partially make up for the gradual loss of case inflection, there was no full hand-off between both strategies: prepositional marking did not reach its full potential. Instead, language users relied on other degenerate strategies.

One of these degenerate strategies was the increased use of the plain transitive construction with experiencer subjects: over the centuries, Dutch experi- encer predicates undergo a drift towards more experiencer-subjects. The drift from dative-experiencers to nominative-experiencers is by no means restricted to Dutch. It is a cross-linguistically rather common tendency and has been analysed from the perspective of grammaticalisation (see Haspelmath 1998: 338-340). From the perspective of Construction Grammar, the process has been ascribed to the extension/schematisation of the transitive construction (Trousdale 2008). For Dutch, this extension of the transitive construction is corroborated by a corpus study on (be/ver)wonderen ‘amaze’, the prime example of an experience verb that tended to occur in the impersonal construction in Middle Dutch (see Figure 6).[3] Using data from two corpora, KLASLIT and LITEROM, the bar chart in Figure 7 shows a diachronic increase in the use of the transitive construction after Middle Dutch.[4] Transitive constructions are those with a subject and a non-prepositional object. Examples are given in (22)-(23). Dutch has not decidedly converged on experiencer-subjects for its transitive construal, but allows stimulus-subjects as well, but Figure 8 shows that within the group of transitive constructions, the experiencer is nevertheless increasingly likely to be encoded as subject - a process which has been observed in English too (see Allen 1995).[5] [6]

(22) Ik bewonder hun schaamteloosheid.

I admire their shamelessness

‘I admire their shamelessness.’ (21st century, LITEROM)

(23) Neen, maar het verwondert me niet.

no but it astonishes me not

‘No, but it does not surprise me.’ (21st century, LITEROM)

The diachrony of the transitive construction with (be/ver)wonderen in Modern Dutch

Figure 7: The diachrony of the transitive construction with (be/ver)wonderen in Modern Dutch

Experiencer-subject vs. stimulus-subject in (be/ver)wonderen

Figure 8: Experiencer-subject vs. stimulus-subject in (be/ver)wonderen

A closer look at the diachrony of the argument realisation of (be/ver)wonderen shows, however, that the transitivity drift is not the only thing that is going on. Dutch turns out to deploy other degenerate ways to maintain the meaning differences primarily carried by case frames in Middle Dutch. Rather than yielding completely to an undifferentiated transitive construction, Dutch exploited existing voice distinctions to re-establish the middle ground that was formerly covered by (double-)oblique case frames. More specifically, there are two constructions that take over this function, namely the reflexive, exemplified in (24), and the static passive, exemplified in (25). Again, we are dealing with degeneracy here, rather than pure renewal, as the reflexive and the (static) passive already existed in Middle Dutch, and just extended their use (frequency and context) in Dutch.

(24) Ik verwonder mij daarover.

I amaze myself there_about

‘That amazes me.’ (21st century, LITEROM) [7]

If we plot the distribution of all these constructions through time (see Figure 9), they appear to become less popular, but our view is clouded by the fact that morphology plays an important role here.[8] The apparent decrease in the use of the reflexive and static passive is due to the rise in the transitive construction. The latter construction, however, cannot be used without altering the morphology of the verb wonderen. In order to use it transitively, it has to be preceded by a prefix ver- or be-. The prefix be- has an applicative function, turning a predicate from intransitive to transitive (e.g. be-zingen ‘sing about’). In this sense, it has, to a certain extent, the same value as the nominative-accusative case frame in Middle Dutch. As becomes clear in Figure 10, the use of this applicative prefix rises precipitously over the centuries - indeed, plain wonderen is not attested anymore in present-day Dutch.[9] The rise of the applicative shows that language users increasingly encode the process of amazement as transitive. Still, in those cases where they do not use the be- prefix, there does seem to be a rise in the use of the other voices (reflexive + static passive voice combined), at the expense of the plain active-transitive, which is also an option here, as illustrated in (23). This is shown in Figure 11, which gives the diachronic distribution of the constructions ignoring the observations with a be- prefix.[10] This shows that although language users increasingly rely on the transitive construction, and on the experiencer-subject version of it, they also increasingly use voice- based constructions when they want to deviate from the plain transitive construction.

Voice-based constructions with (be/ver)wonderen in Modern Dutch

Figure 9: Voice-based constructions with (be/ver)wonderen in Modern Dutch

Prefixes in Modern Dutch

Figure 10: Prefixes in Modern Dutch

Voice-based constructions with (ver)wonderen in Modern Dutch

Figure 11: Voice-based constructions with (ver)wonderen in Modern Dutch

Another degenerate strategy compensating for the loss of the fine-grained case-based argument realisation for experience processes in Middle Dutch seems to be an increase in the lexical expressions. Next to (be/ver)wonderen, present- day Dutch has verbs like verbazen ‘amaze’, versteld zijn ‘be amazed’, verrassen ‘surprise’, which did not occur yet in Middle Dutch. This lexical proliferation can be used to express different meanings. It remains to be seen whether this tendency holds up under a more systematic study, but a similar lexical increase has been observed for Middle Dutch vergeten ‘forget’ (see Van de Velde 2004: 70).

In sum, deflection in Dutch was not as detrimental to the case-based argument realisation of Dutch experience predicates as one might think. Thanks to existing degenerate strategies, the loss of the case system could be sustained without losing the capability of expressing semantic nuances in experience processes. Several alternative strategies were exploited, such as (i) prepositions, (ii) generalisation of the transitive construction, (iii) voice-based distinctions, (iv) applicative prefixes, and (v) lexical differentiation - all of which serve functions outside the argument realisation of experience processes. The actual facts are thus vastly more complex than a simple “transition from cases to prepositions” or a simple “extension of the transitive construction”. The multifaceted nature of the diachrony of argument realisation in experience predicates shows how

Dutch, as a complex adaptive system, benefits from degeneracy both in terms of robustness and evolvability: language users can do away with something as central to grammar as case without running into problems.

  • [1] The analysis of (19) is complicated because the complement of the preposition is daer,constituting what is called in Dutch grammar a (separable) pronominal adverb.
  • [2] Other prepositions occur as well: af, in, om, in, over (see WNT s.v. verwonderen I).
  • [3] The prefixes be- and ver- have an impact on the meaning of the verb, of course. Bewonderen,with the applicative be- prefix involves higher agentivity on the part of the experiencer. Thequestion is whether to include bewonderen in the corpus study. Its applicative nature implies astrong tendency to occur in the transitive construction, and it could be argued that it wouldbe better to leave the instances of bewonderen out of the corpus study, as it risks to overestimatethe transitivity drift. Still, there are good reasons to retain bewonderen in the counts. First, itsemantically overlaps with (ver)wonderen: on the one hand Middle Dutch (ver)wonderen couldcarry the meaning of present-day Dutch bewonderen (‘admire’), as in the following example:Nochtan so verwondert die natuer die suverlijcheit (‘Still, nature admires pureness’) (MNW s.v.suverlijcheit). On the other hand, Early Modern Dutch bewonderen does not consistently mean‘admire’, but can be used in the sense of ‘amaze’ as well (see WNT s.v. bewonderen). Moreover,it does not consistently occur in the plain transitive experiencer-subject construction, butoccurs in the inverted source-subject construction (Het bewonderde mij ‘it APPL=amazed me’(WNT s.v. bewonderen)) and in the reflexive as well, as in: Zoo men (...) binnens-Lands nogsommige Kaakenbeenderen van Walvisschen, ‘t zy in Klai, Veen of Zand, verkalkt of versteendmogte ontdekken, behoeft men zig . . . deswegens niet te bewonderen (‘If one would discoversome calcified or petrified whale cheek bones inland, be it in clay, peat or sand, one shouldnot be amazed by it’) (WNT s.v. bewonderen). In short, there is no hard and fast distinction,neither in meaning nor in argument realisation, between bewonderen on the one hand and(ver)wonderen on the other hand. This is the reason why we treat bewonderen and verwonderentogether in the corpus study. I will return to the role of the prefixes below.
  • [4] The correlation is statistically significant: Kendall’s Tau-b 0.28, ASE 0.07. Chi-Square: p < 0.0001.
  • [5] The correlation is statistically significant: Kendall’s Tau-b 0.56, ASE 0.04. Chi-Square:
  • [6] p < 0.0001.
  • [7] ... dat je over alles even verwonderd was. that you about everything equally amazed were ‘... that you were equally amazed by everything.’ (21st century, LITEROM)
  • [8] Making the “construction” variable binary by amalgamating the reflexive and the static passive in one category, as opposed to the active transitive, yields the following results: KendallTau-b: 0.56, ASE 0.10. Chi-Square: p < 0.0001.
  • [9] The reason why Figure 10 plots more observations is that the non-finite contexts (and imperatives) could be included as well, which were obviously ignored in the figures on argumentrealisation above. The figures have been analysed with a Cochran-Mantel-Haenszel statistic,which tests a matrix with an ordinal variable in the columns and a nominal variable in therows for the alternative hypothesis that the row mean scores differ. In the case at hand, it testswhether the use for either of the three forms (0-wonderen, ver-wonderen, be-wonderen) shiftsthrough time. The test value corresponds to a p-value < 0.0001.
  • [10] The results show that the correlation observed in Figure 9 disappears: Kendall’s Tau-b 0.06(ASE 0.10). Chi-Square p = 0.12. (But note that absence of evidence of an association is not thesame as evidence of absence of an association.)
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science