My forthcoming book, Lexicon and Grammar: the English Syntacticon, focuses on the need for a formal, highly constrained and empirically revealing lexical theory. It defends a syntactic approach in terms of c-selection (subcategorization) for capturing patterns projected from the lexicon, arguing that thematic grids or lexical conceptual structures are neither empirically revealing nor truly formal. For example, the book argues against widely and uncritically accepted semantic selection analyses of propositional complements, discourse anaphora, "unprojected arguments" and "light verb" theta-grids.
The lexicon in this extended theory of c-selection consists of two quite different components, a "Syntacticon" of closed class items and a Dictionary of more contentful open class lexical items. Each item in the Syntacticon is a unique combination of values from a small set of cognitive features F used in syntactic derivations, while Dictionary entries have in addition purely semantic features f. The Syntacticon is regulated by a newly developed theory of multi-level lexical insertion: an item's feature composition determines at which level it satisfies c-selection: at the outset of computation on a domain ("deep structure"), during computation of the domain ("in syntax"), or after Spell Out of the domain ("in PF").
The book also argues against any autonomous morphological component. Rather, the only statements of morphology are those of phonological effects conditioned by syntactic factors. E.g: "Bound forms in the English Syntacticon have no inherent stress." Within syntax proper, bound morphology is simply a special case of compounding a Syntacticon item (one lacking semantic features f). The theory of multi-level insertion applied to bound morphology provides novel explanations for previously unrelated complex syntactic patterns in e.g., Germanic and Romance gerunds, passives and perfects.
Late insertion of Syntacticon entries also induces "flat structures" for certain syntactic constructions. The books shows that the similar, recalcitrant properties of causative, restructuring, light and linking verbs all result from these flat structures. These explanations crucially exploit a newly developed notion of "empty underlying head."
In general, the book demonstrates that familiar classes of English and Romance complements instantiate all and only the predicted c-selection frames. It thus identifies the limits which Universal Grammar imposes on the construct "possible lexical entry."
Dictionary Syntacticon a. Items with cognitive features F used in syntax: yes yes b. Cognitive features F canonically realized on UG-defined hosts: yes yes c. Insertion possible at the beginning of a syntactic derivation: yes yes d. Grammatical categories in the inventory: N, V, A, P All e. Items with purely semantic features f (not used in syntax): YES NEVER f. Open classes; coining for adult speakers: YES NEVER g. Bound forms have inherent stress and head true compounds: YES NEVER h. Interface with non-linguistic memory and culture: YES NO i. Full suppletion inside paradigms: NEVER YES j. Certain phonetically zero morphemes: NEVER YES k. Items must conform to core vocabulary phonology: NO YES l. Items with alternatively realized features: NEVER YES m. Insertion possible during syntax and at PF: NEVER YES