Non-inverted sentences

This section and the sections that follow describe how the English XTAG grammar accounts for properties of the auxiliary system described above. In our grammar, auxiliary trees are added to the main verb tree by adjunction. Figure 20.1 shows the adjunction tree for non-inverted sentences.^20.3

  

Figure: Auxiliary verb tree for non-inverted sentences: $\beta$Vvx

  

Figure 20.2: Auxiliary trees for The music should have been being played .

The restrictions outlined in column 3 of Table 20.1 are implemented through the features <mode>, <perfect>, <progressive> and <passive>. The syntactic lexicon entries for the auxiliaries gives values for these features on the foot node (VP^*) in Figure 20.1. Since the top features of the foot node must eventually unify with the bottom features of the node it adjoins onto for the sentence to be valid, this enforces the restrictions made by the auxiliary node. In addition to these feature values, each auxiliary also gives values to the anchoring node (V$\diamond$), to be passed up the tree to the root VP (VP_r) node; there they will become the new features for the top VP node of the sentential tree. Another auxiliary may now adjoin on top of it, and so forth. These feature values thereby ensure the proper auxiliary sequencing. Figure 20.2 shows the auxiliary trees anchored by the four auxiliary verbs in sentence ((333)). Figure 20.3 shows the final tree created for this sentence.

  

Figure 20.3: The music should have been being played .

The general English restriction that matrix clauses must have tense (or be imperatives) is enforced by requiring the top S-node of a sentence to have <mode>=ind/imp (indicative or imperative). Since only the indicative and imperative sentences have tense, non-tensed clauses are restricted to occurring in embedded environments. Noun-verb contractions are labeled NVC in their part-of-speech field in the morphological database and then undergo special processing to split them apart into the noun and the reduced verb before parsing. The noun then selects its trees in the normal fashion. The contraction, say 'll or 'd, likewise selects the normal auxiliary verb tree, $\beta$Vvx. However, since the contracted form, rather than the verb stem, is given in the morphology, the contracted form must also be listed as a separate syntactic entry. These entries have all the same features of the full form of the auxiliary verbs, with tense constraints coming from the morphological entry (e.g. it's is listed as IT 'S NVC 3SG PRES). The ambiguous contractions 'd (had/would) and `s (has/is) behave like other ambiguous lexical items; there are simply multiple entries for those lexical items in the lexicon, each with different features. In the resulting parse, the contracted form is shown with features appropriate to the full auxiliary it represents.