“…Nevertheless, the data we find in this way are striking counterexamples to the above generalization and form the basis of alternative theories for licensing such partial NP constituents (De Kuthy and Meurers, 2001;De Kuthy, 2002): To overcome the shortcomings of the crude approximation we used in our search pattern for this example, one has to rely on more richly annotated corpora, such as the treebanks mentioned at the end of the previous section. To search in such treebanks, query languages and tools which can refer to syntactic structure or dominance relations have been developed (cf., e.g., Pito, 1994;Brew, 1999;Rohde, 2001;Mckelvie, 2001;König and Lezius, 2000;Kallmeyer, 2000;Steiner, 2001;Kepser, 2003). 21 For example, to search for example patterns such as the one in this section, Kallmeyer (2000) defines a formal language which can encode the search for ''a prepositional phrase modifying the accusative object and preceding the finite verb (i.e., in the so-called Vorfeld), and an accusative object between finite verb and non-finite forms (i.e., in the so-called Mittelfeld).''…”