Workshop on Germanic Infinitives

30.9.- 2.10 2004


H.Haider , Salzburg

Optional Clause Union of Infinitival Complements - an OV effect
Optional clause union constructions for otherwise sentential infinitival complements are well-known form German and Dutch. Descriptively fairly well covered, their grammar theoretic causality has not been modelled satisfactorily so far. It is still unclear or controversial as to

  • why grammar theory allows for a perfectly optional choice between clause union and sentential complementation
  • why this constructional freedom is found in OV languages only
  • how the clause union syndrome is adequately modelled

In this contribution, I shall try to demonstrate that clause union infinitivals (and their syntactic properties) are predictable for OV languages, and excluded for VO languages. Clause union infinitivals are a facet of the verb clustering property of OV language which in turn can be derived from an UG-grounded universal property of syntactic structures in combination with the headedness option (head first - head last).


Jack Hoeksema, Groningen
Nonverbal material in the Dutch verb cluster: a word order pattern in

Particles and small phrases may appear in Dutch verb clusters. Over a period of 3 centuries, several trends can be noticed in the use of this word order option: (1) from the early 18th century onward, this option is mainly found with idiomatic combinations; (2) only certain syntactic phrases are allowed (PPs, not DPs); and (3) phonological heaviness plays an ever-increasing role. I will treat these findings as evidence for a syntactic component with violable output filters.


Hans-Martin Gärtner und JoannaBlaszczak, Potsdam, Berlin
Intonational phrasing, discontinuity, and the scope of negation

We discuss several cases of English and German negative quantifiers taking extended scope. We argue that these scope extensions are sensitive to linear and prosodic continuity, a fact which we capture in terms of "Condition on Extended Scope Taking" (CEST). We provide two formalizations of CEST, one couched in minimalist terms and another within the framework of "Combinatory Categorial Grammar" (CCG). We compare and contrast the resulting systems and suggest that although the differences are clearly dicernible it is too early to judge which of the competitors should be preferred.


Sam Featherston and Wolfgang Sternefeld, Tübingen
The relatedness of the third construction and  long passives in German
In this talk we will report on our studies of coherence phenomena in German infinitival complement clauses. This work has two main aims: first, we wished to explore which factors affect the acceptability of the phenomena known as the "third construction" (1) and the "long passive" (2).
(1) In der Werkskantine heißt es über den Manager, dass er den Neffen bestreitet unbegründet zu bevorzugen.

(2) In der Werkskantine heißt es, dass der Neffe unbegründet zu bevorzugen bestritten wird.

We tested such factors as the NP-type status and Case of constituents, the  adjacency of the verbs, and active/passive status of the infinitive. This  provided a firm basis on which to carry out the second stage.
The second aim was to test whether exactly the same set of verbs permit the two constructions.  For these purposes we carried out two studies collecting judgements of carefully controlled materials, containing 16 verbs on a continuum from "good" coherent verbs (versprechen, beabsichtigen ....) to bad coherent verbs (genießen, zugeben ...).

Ulrike Demske, Saarland University
Infinitival Complementation in Old High German

Ever since the influential work by Bech (1955/57), the concept of (in-)coherence has figured prominently in developing appropriate analyses of infinitival constructions in Present-Day German. In particular, there continues to be debate how the (in-)cohe­­rence of infinitival constructions relates to the categorial status of the infinitival complements in question - current views holding either that there is a two-way distinction between coherent constructions with infinitival VP-complements and incoherent constructions with infinitival CP-complements or that both, incoherent and coherent infinitival constructions select CP-complements with the latter undergoing some kind of restructuring process. Recently, this two-way distinction has been called into question by Reis (2001) and Wurmbrand (2001), both suggesting a graded notion of (in-)coherence resulting in at least a four-way distinction of infinitival constructions, feeding the question of an appropriate analysis anew.
In this paper, I will tackle the question whether we find evidence in Old High German to distinguish different verb classes regarding (in-)co­he­rence or even various degrees of cohe­rence. Evidence to be considered includes extraposition, scrambling, pronoun fronting, adjacency of matrix predicate and infinitive as well as verbal case (= status), theta-marking properties of the matrix predicate and co-occurrence facts regarding auxiliaries and verbs selecting infinitival complements. 

Bech, Gunnar. 1955/57. Studien über das deutsche verbum infinitum. Copenhagen: Munksgaard.
Reis, Marga. 2001. Bilden Modalverben im Deutschen eine syntaktische Klasse? In Reimar Müller and Marga Reis (eds), Modalität und Modalverben im Deutschen. Linguistische Berichte [Sonderheft 9]. Hamburg: Buske. 287-318.
Wurmbrand, Susanne. 2001. Infinitives: Restructuring and Clause Structure [Studies in Ge­ne­rative Grammar 55]. Berlin: Mouton de Gruyter.

Marcel den Dikken, CUNY Graduate Center
Some issues in the morphosyntax of Germanic to-infinitives
1          The Frisian alternation in (1) (reported in Den Dikken & Hoekstra 1997:1062) is instructive in a number of ways. First, the fact that the placement of the past participle kontrolearre at the left edge of the verbal cluster is entirely oblivious to what is happening to its right suggests that kontrolearre is not itself part of the verbal cluster (as discussed in Den Dikken & Hoekstra's paper). Secondly, the infinitival marker te brings about a special morphological marking (-n) on the infinitive that immediately follows it, even when te is surface-optional (cf. (1b)). This suggests that the Frisian infinitival marker, in contexts in which it appears to be optional on the surface, is systematically present throughout the morphosyntax, imposing its morpho­pho­nological restrictions on its host, but is subject to deletion in the PF-component under certain circum­stances. Thirdly, the infinitival marker and its associate -n apparently 'migrate' through the cluster, strictly as a couple, showing up consistently on the right-hand member, regardless of whether their host is syntac­tically selected by a verb taking a te-infinitival complement (here, hoeve 'need'; wurde 'become' does not select infinitives).
2            'Migrant' to-marking (and its concomitant morphological marking on the to-marked verb) is not con­fined to clusters with 'inverted' order: it occurs in clusters with 'English' orders as well. The facts in (2a,c) (from Cooper 1989) show this, and also illustrate that Swiss German reproduces the special marking seen in Frisian infinitives preceded by the infinitival marker: the la/laa alter­nat­ion is conditioned in the same way as the i/-n alternation in Frisian. Swiss German also presents a z-omission con­struc­tion that it will be helpful to compare with the te-omission facts in Frisian (1), cf. the pair in (3), where (3a) is a standard German example and (3b) is its Swiss counterpart; (3b) has a different verb order and just a single z, in front of V2.
3          One might take the 'migration' of the infinitival marker as an indication of its syntactic autonomy. However, 'migration' is not the prerogative of the infinitival marker in West-Germanic: the present-participial morph­eme (-end, adorned in (4a,c) with an agreement-schwa) can be found attached to the 'wrong' host as well, as in the Dutch example in (4a), where an 'English-order' verbal cluster is used as a prenominal attributive modifier; the attested form in (4a), while not perfect, is better than any of its alternatives in (4b,d). On minimalism's lexicalist approach to inflectional morphology, present-participial -end cannot be taken to be syntactically autonomous. 'Migration', therefore, is not an argument for syntactic autonomy on the part of the infinitival marker; the parallel with (4a) could in fact be pushed in the opposite direction: 'migration' is an indication that the infinitival marker is part and parcel of the infinitival inflectional morphology (Haider).
4          There are other reasons to believe that the West-Germanic infinitival marker (unlike its English and Scandinavian counterparts) is not syntactically autonomous, instead forming an integral part of a complex head. One is that, on the assumption that the infinitival marker is part and parcel of the infinitival inflectional morphology, a principled account can be given for the correlation, in Dutch (cf. Den Dikken 2003), between (i) the (im)possibility for a double-particle or N,V complex to undergo Verb Second as a unit and (ii) the (im)possibility for a double-particle or N,V complex to have the infinitival marker at the left edge of the complex, cf. (5),(6). We can derive the descriptive generalisation in (5) with the aid of the hypothesis (Bech 1955, Haider 1988, 2002) that the infinitival marker te/zu is an inflectional affix, base-generated inside the complex verb (not the lexicalisation of a VP external functional head). The placement of the infinitival marker vis-à-vis non-inflectional material gives the language user an explicit clue with respect to the location of the inflection inside the complex verb. Only if inflection is peripheral can a complex verb move as a unit.
5          If complex verbs can only move as a unit if inflection is peripheral (as evidenced by the placement of the infinitival marker), that suggests that neither of the alternate word orders of (7a) and (7b) results from movement of the (complex) infinitive or past participle. This conclusion, if upheld under scrutiny, has important repercussions for the analysis of the word-order variation in the West-Germanic verbal cluster and beyond.     

a.   kontrolearre   hoeve      *(te) wurden       3B1BteB2+n   (Frisian)
   check-pptc   need-inf   to  become-inf+n

b.   kontrolearre   wurde      (te) hoeven       3B2BteB1+n
   check-pptc   become-inf   to  need-inf+n
  both: 'need to be checked'

a.         er            hät            versproche,      d            Chind            studiere                        z  laa            (Swiss German)
            he            has            promise-pptc   the            kids            study-inf       to let-inf(long)

b.         er            hät            versproche,      d            Chind            la            z            studiere
            he            has            promise-pptc   the            kids            let-inf            to            study-inf
'he promised to let the children study'

c.         ohni            s            Schtüürrad       mit            bedne            Händ            müese  verlaa
            without            the            steering-wheel  with            both            hands            must-inf to  leave-inf(long)

a.         er            hat            versprochen,    den            Hans            zu            erreichen          zu  probieren
            he            has            promise-pptc   the            Hans            to            reach-inf       to  try-inf

b.         er            hät            verschproche,   de            Hans            probiere            z            erreiche
            he            has            promise-pptc   the            Hans            try-inf              to            reach-inf

a. ?een niet kunnen zingende band (Dutch)
             a            not            can-inf                        sing-presptc-agr   band
'a band unable to sing'

b.             ?*een    niet            zingen               kunnende                    band
            a            not            sing-inf            can-presptc-agr            band

c.            *een     niet            kunnende                    zingen               band
            a            not            can-presptc-agr            sing-inf            band

d.            **een   niet            kunnend          zingene                         band
            a            not            can-presptc            sing-inf-agr                       band

(5)  particles and incorporated nouns can be carried along under V2 iff they can felicitously follow the infinitival marker (surfacing between infinitival marker and V-stem)


double-particle verb type


       V2 of complex verb


inf-marker preceding prt


Type I (voor-aan-melden >preregister=)






Type II (her-af-drukken >reprint=)






Type IIIa (voor-ver-kopen >advance-sale=)
Type IIIb (door-ver-kopen >sell on=)
Type IIIc (over-ver-hitten >overheat=)






Type IV (her-ver-delen >redistribute=)





a.         dat            Jan            z'n            dochter             {wil            vooraanmelden/vooraanmelden            wil}
            that            Jan            his            daughter           wants            preregister-inf            preregister-inf  wants

b.         dat            Jan            z'n            dochter             {heeft            vooraangemeld/vooraangemeld            heeft}
            that            Jan            his            daughter           has            preregister-ptcpreregister-ptchas