Records, frames and attribute spaces

Type Theory with Records (Cooper & Ginzburg 2015), multidimensional attribute spaces (Gust & Umbach 2015) and frame‐based representations (Naumann, Petersen & Gamerschlag 2017) appear, at first sight, similar in their capacity to encode information and license entailments. However, closer examination reveals relevant differences concerning, e.g., composition, the notion of reference and the type/token distinction.

The idea of this workshop is to compare the three representation formalisms with respect to their capacity to improve processing performance, or achieve non‐trivial analysis of particular linguistic phenomena.

There will be short introductions to each of the three formats, followed by talks addressing particular problems / phenomena.