This book critically reviews grammatical research into logical form over the past 20 years and reconsiders some of its major themes in the light of recent theoretical innovations. In the late 1970s generative grammarians proposed the existence of an abstract syntactic level of grammatical representation derived from surface structure which was phonetically invisible. This level, dubbed logical form, has been thought of as the information that the grammar contributes to semantic interpretation. The first part of the book reviews the standard arguments for the existence of LF and its format. Norbert Hornstein focuses especially on Quantifier Raising and a host of conditions that have been proposed to constrain valid LF phrase markers. The second section considers what properties a "minimalist" LF should have. This material is by its nature more speculative. Among the topics broached are anticedent contained deletion constructions, weak crossover configurations and multiple interrogatives.