At present, we only deal with the adjectival metalinguistic comparatives as in ((447)). The analysis given here for these can be easily extended to prepositional phrases and nominal comparatives of the metalinguistic sort, but, as with coordination in XTAG, verb phrases will prove more difficult. Adjectival comparatives appear to distribute with simple adjectives, as in the following examples:
This patterning indicates that we can give these comparatives a tree that adjoins quite freely onto adjectives, as in Figure 23.1. This tree is anchored by more/less - than. To avoid grammatically incorrect comparisons such as more brighter than dark, the feature compar is used to block this tree from adjoining onto morphologically comparative adjectives. The foot node is compar-, while brighter and its comparative siblings are compar+23.1. We also wish to block strings like more brightest than dark, which is accomplished with the feature super, indicating superlatives. This feature is negative at the foot node so that ARBaPa cannot adjoin to superlatives like nicest, which are specified as super+ from the morphology. Furthermore, the root node is super+ so that ARBaPa cannot adjoin onto itself and produce monstrosities such as ((455)):
Thus, the use of the super feature is less to indicate superlativeness specifically, but rather to indicate that the subtree below a super+ node contains a full-fleshed comparison. In the case of lexical superlatives, the comparison is against everything, implicitly. A benefit of the multiple-anchor approach here is that we will never allow sentences such as ((456)), which would be permissible if we split the comparative component and the than component of metalinguistic comparatives into two separate trees.
We also see another variety of adjectival comparatives of the form more/less than X, which indicates some property which is more or less extreme than the property X. In a sentence such as ((457)), some property is being said to hold of Francis such that it is of a kind with stupid and that it exceeds stupid on some scale (intelligence, for example). Quirk et al. also note that these constructions remark on the inadequacy of the lexical item. Thus, in ((456)), it could be that stupid is a starting point from which the speaker makes an approximation for some property which the speaker feels is beyond the range of the English lexicon, but which expresses the supreme lack of intellect of the individual it is predicated of.
Taking our inspiration from ARBaPa, we can handle these comparatives, which have the same distribution but contain an empty adjective, by using the tree shown in Figure 23.2.
This sort of metalinguistic comparative also occurs with the verb phrase, prepositional phrase, and noun varieties.
Presumably, the analysis for these would parallel that for adjectives, though it has not yet been implemented.