... argument.1
The alternative is to give verb clusters lexical status as well. We discuss this option below. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... items.2
In southern dialects of Dutch, the possibility of NP or PP arguments violating cross-serial word order is clearly not restricted to idiomatic expressions only [9]. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... verb-cluster.3
Most analyses outside the categorial tradition assume a right-branching structure for the verb cluster. Evers [7], [2], and [5] argue for right brancing structures, but some of their arguments are questioned in [12], who present a left-branching analysis instead. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...coord).4
Example (28b) is taken from [15], who presents it as an argument against [2], who consider a structurally similar example ungrammatical. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... argument5
Note that a `side-effect' of this assumption is that it eliminates a spurious ambiguity that might otherwise arise if we derive a phrase consisting of a transitive verb governed by an inverted modal or auxiliary:

\begin{tabular}[t]{ll} i. & {\sc \begin{tabular}[t]{rccc} \ldots & {\em een bo... ...-vc]}}$} \\ \cline{3-4} && \multicolumn{2}{c}{vp} \end{tabular}} \end{tabular}

The right-branching derivation in (ii) is ruled out if willen requires a verbal complex as argument. 

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Noord G.J.M. van
1998-09-29