Basic Problems of Incremental Monitoring

The problem is that the monitor must be dynamically configured during incremental processing time of single utterances in order to decide

• when the test of ambiguity should take place and
• which partial strings should be revised?

Technically it is possible to check and revise each partial result of the generator. But, without any control, the monitor would try to disambiguate each local ambiguity; it is hard to imagine that the resulting generator would produce anything at all.

Clearly, an utterance can only be said to be (un)ambiguous with respect to a certain context. The assumption is that usually an utterance which is not ambiguous w.r.t. its context will remain unambiguous if it is part of a larger utterance.

It may be possible to restrict the context during the production of a partial utterance to grammatical properties, e.g. to the information associated with the head which selects the phrase dominating this partial utterance. Such an approach can be integrated in head-driven generators of the type described in [Shieber et al.1990].

For example, assume that for each recursive call to the generator the revised monitor is called with an extra argument Head which is to be used as contextual information when the embedded parser is called to test whether the string in question is ambiguous. Thus, suppose we are to generate from the logical form

A head-driven generator first produces the word during as the head. Next an NP with logical form ball' has to be generated. For this logical form the generator chooses the word ball which is however ambiguous. For this partial utterance the monitor is called, using the head information of during. However, being an argument of the head during, only one of the readings of ball is possible. Therefore, the monitor simply `confirms' the choice of the generator. Thus, the assumption here is that this ambiguity will be disambiguated later on by combining this string with its head.

Beside the fact, that this method depends on grammar theories which comes with a notion of head, the main problem of this approach is that

• either each ambiguous partial string has to be revised immediately or
• revision is delayed until the previous recursive call of the generator has been finished.

In the first case the monitor would also revise irrelevant ambiguities. The latter point would mean that revision can only be performed after the whole utterance has been produced. But then the incremental method just simulates the non-incremental method.