The last example shows that comparative adjectival phrases cannot distribute quite as freely as comparative nominals. The analysis of elliptical comparative adjectives follows closely to that of comparative nominals. We build them up by first adjoining the comparative element to the A node, which then signals to the AP node, via the compar feature, that it may allow a than-clause to adjoin. The relevant trees are given in Figure 23.8. CARBa is anchored by more, less and as, and axPnx is anchored by both than and as.
 |
 |
|
| (a) CARBa tree | (b) axPnx tree |
The advantages of this analysis are many. We capture the distribution exhibited in the examples given in ((483)) - ((488)). With CARBa, comparative elements may modify adjectives wherever they occur. However, than clauses for adjectives have a more restricted distribution which coincides nicely with the distribution of AP's in the XTAG grammar. Thus, by making them adjoin to AP rather than A, ill-formed sentences like ((488)) are not allowed.
 |
There are two further advantages to this analysis. One is that CARBa interacts with nxPnx to produce sequences like more exquisite horse than Black Beauty, a result alluded to at the end of Section 23.3.1. We achieve this by ensuring that the comparativeness of an adjective is controlled by a comparative adverb which adjoins to it. A sample derivation is given in Figure 23.9. The second advantage is that we get sentences such as ((489)) for free.
Since better comes from the lexicon as compar+ and this value is passed up to the AP node, axPnx can adjoin as desired, giving us the derivation given in Figure 23.10.
 |
Notice that the root AP node of Figure 23.10 is compar-, so we are basically saying that strings such as better than Bill are not ``comparative.'' This accords with our use of the compar feature--a positive value for compar signals that the clause beneath it is to be compared against something else. In the case of better than Bill, the comparison has been fulfilled, so we do not want it to signal for further comparisons. A nice result which follows is that axPnx cannot adjoin more than once to any given AP spine, and we have no need for the NA constraint on the tree's root node. Also, this treatment of the comparativeness of various strings proves important in getting the coordination of comparative constructions to work properly. A note needs to be made about the analysis regarding the interaction of the equivalence comparative construction as ... as and the inequivalence comparative construction more/less ... than. In the grammar, more, less, and as all anchor CARBa, and both than and as anchor axPnx. Without further modifications, this of course will give us sentences such as the following:
Such cases are blocked with the feature equiv: more, less, fewer and than are equiv- while as (in both adverbial and prepositional uses) is equiv+. The prepositional trees then require that their P node and the node to which they are adjoining match for equiv. An interesting phenomena in which comparisons seem to be paired with an inappropriate as/than-clause is exhibited in ((492)) and ((493)).
Though prescriptive grammars disfavor these sentences, these are perfectly acceptable. We can capture the fact that the as/than-clause shares the equiv value with the latter of the comparison phrases by passing the equiv value for the second element to the root of the coordination tree.