Using a logic grammar to learn a lexicon

It is suggested that the concept of "logic grammar" as relation between a string and a parse-tree can be extended by admitting the lexicon as part of the relation. This makes it possible to give a simple and elegant formulation of the process of infering a lexicon from example sentences in conjunction with a grammar. Various problems arising from implementation and complexity factors are considered, and examples are shown to support the claim that the method shows potential as a practical tool for automatic lexicon acquisition