ABSTRACT

Pseudocoordinations are a classical topic in the grammar of Norwegian, Swedish and Danish. These constructions occur with a small number of first verbs, and look like VP coordinations, cf. (1)-(2). However, they do not have the syntactic properties that are typical of coordinations. For example, the two verbs cannot behave as a unit in c-structure, cf. (3)-(4).

Pseudocoordinations raise problems for grammatical theory. In the eighties and nineties, it was established that pseudocoordinations are not coordinations. Anward 1988 and Wiklund 1996 suggested that pseudocoordinations are subordinations, where the first verb governs a complement headed by the second verb. (This means that the morphological form of the second verb is due to "feature copying".) Wiklund 1996 also proposed that this complement has a PRO subject. Bodomo 1997, on the other hand, suggested that pseudocoordinations are complex predicates with the properties of serial verb constructions.

These analyses assume that pseudocoordination is a unitary phenomenon. I will show that pseudocoordinations are not grammatically unitary. The complex predicate analysis is correct for pseudocoordinations with _ta_ 'take', as in (2), but other pseudocoordinations are biclausal. A small group is raising constructions (as shown by the fact that the first verb can take an expletive subject selected by the second verb, as in (5)). However, the main group (for example (1)) is control constructions. An important problem for their analysis (which has never been mentioned in the literature) is the fact that they are grammatically different from regular control constructions. For example, control pseudocoordinations allow the presentational focus construction (cf 6), while regular control constructions don't, (cf. 7). Regular control constructions allow their complement to topicalize (cf. 8), while control pseudocoordinations don't (cf. 9). Regular control constructions allow passivization, with the verbal complement or an expletive as a passive subject (cf. 10-11). Control pseudocoordinations don't allow the corresponding passives (cf. 12-13). Control pseudocoordinations only allow a rather peculiar passive like (14), in which the superordinate verb shares an expletive subject with a subordinate verb.

These differences between control pseudocoordinations and regular control constructions might seem to make a control analysis of pseudocoordinations impossible. However, they find a natural explanation within LFG's theory of control. LFG distinguishes between functional and anaphoric control. A functionally controlled verbal complement is a VCOMP (also called XCOMP). On the other hand, an anaphorically controlled verbal complement can be a COMP or an OBJ, depending upon its external syntactic properties. The realization of verbal complements vary, both with different languages and different governing predicates.

In Norwegian, verbal complements usually have the syntactic properties of OBJs. The verbal complements of regular control verbs are OBJs with anaphorically controlled SUBJs (as in Icelandic, cf. Andrews 1982). The verbal complements of control pseudocoordination verbs are different, however. I will propose that they are VCOMPs with functionally controlled SUBJs. This accounts for the differences mentioned between control pseudocoordinations and regular control constructions:

The presentational focus construction gives the verb an expletive subject and an OBJ. Regular control verbs already have an OBJ, and do not allow this construction. (Cf. 7.) Control pseudocoordination verbs allow it; the OBJ then controls the verbal complement, as predicted by the classical Lexical Rule of Functional Control in Bresnan 1982. (Cf. 6.)

Topicalization is generally possible with all constituents, except COMP and VCOMP ("Higgins' generalization"). The verbal complements of regular control verbs can topicalize, since they are OBJs (cf. 8), while the verbal complements of control pseudocoordination verbs cannot, since they are VCOMPs (cf. 9).

Passivization is generally not possible with obligatory subject control verbs, since passivization would leave the verbal complement without a realized controller ("Visser's generalization"). This is the reason control pseudocoordination verbs do not allow regular passivization (cf. 12-13). Regular control verbs passivize freely, since their verbal complements are OBJs with anaphorically controlled subjects (cf. 10-11.)

The peculiar passive that is allowed by control pseudocoordination verbs (cf. 14) also has an explanation. A verb like _sitte_ 'sit' selects an expletive subject in the passive. If the subordinate verb also selects an expletive subject, the two verbs can share one, since functional (as opposed to anaphoric) control can share expletives.

We see, then, that the properties of pseudocoordinations follow from the assumption that their complements are functionally controlled, differing from other control constructions in Norwegian. Pseudocoordinations thus give new evidence for LFG's classical distinction between functional and anaphoric control.

DATA (Translations are literal, and do not render the aspectual contributions of the first verbs of the pseudocoordinations.)

1 Han sitter / ligger / staar og leser avisen
he sits / lies / stands and reads the-paper

2 Han tok og ringte til henne
he took and phoned to her

3 Der sitter han og synger
there sits he and sings

4 *Der sitter og synger han
there sits and sings he

5 Det drev og kom ned noe hele tiden (authentic example from the www)
there kept-on and came down something all the time
(i.e. Something kept coming down all the time)

6 Det sitter en mann og synger her
there sits a man and sings here

7 *Det forsoeker en mann aa synge her
there tries a man to sing here

8 Aa synge har han forsoekt
to sing has he tried

9 *Og sunget har han sittet
and sung has he sat

10 Aa synge ble forsoekt
to sing was tried

11 Det ble forsoekt aa synge
there was tried to sing

12 *Og sunget ble sittet
and sung was sat

13 *Det ble sittet og sunget [*with non-passive sunget]
there was sat and sung

14 Det ble sittet og sunget [OK with passive sunget]
there was sat and sung

References:

Andrews, Avery 1982 The representation of case in modern Icelandic. In Joan Bresnan (ed.) The Mental Representation of Grammatical Relations. Cambridge, MA: The MIT Press.
Anward, Jan 1988 Verb-verb agreement in Swedish. McGill Working Papers in Linguistics. Special Issue on Comparative Germanic Syntax, 123-55.
Bodomo, Adams B. 1997 Paths and pathfinders: Exploring the syntax and semantics of complex verbal predicates in Dagaare and other languages. Dissertation. Trondheim, Norway: The Norwegian University of Science and Technology.
Bresnan, Joan 1982 Control and complementation. In Joan Bresnan (ed.) The Mental Representation of Grammatical Relations. Cambridge, MA: MIT Press.
Bresnan, Joan 2000. Lexical-Functional Syntax. Blackwell Publishers.
Wiklund, Anna-Lena 1996 Pseudocoordination is subordination. Working Papers in Scandinavian Syntax 56, 29-54.
Wiklund, Anna-Lena 2001 Dressing up for vocabulary insertion: The parasitic supine. Natural Language and Linguistic Theory 19, 199-228.

Back to Abstracts index