Discussion

In Levelt's model parsing and generation are performed in isolation using two different grammars (although he only considers generation in full detail). The problem with this view is that generation of un-ambiguous paraphrases can be very inefficient, because the source of the ambiguous utterance is not used to guide the generation process. If, for example an intelligent help-system that supports a user by using an operating system (e.g. Unix, [Wilensky et al. 1984]), receives as input the utterance ``Remove the folder with the system tools'' then the system is not able to perform the corresponding action directly because it is ambiguous. But the system could ask the user ``Do you mean `Remove the folder by means of the system tools' or `Remove the folder that contains the system tools' ''. This situation is summarized in the following figure (LF and LF symbolize two readings of S):

  
Figure 5.3: Relationship between ambiguities and paraphrases.

If parsing and generation are performed in isolation then generation of paraphrases can be very inefficient, because the source of the ambiguous utterance S is not used directly to guide the generation process.