Neural Network Acceptability Judgments

Experience-driven acceptability judgments

Linguistic acceptability judgments are a critical part of the toolkit for linguistic investigation. An implicit assumption of much work using these judgments is that the data are stable, presumably reflecting the underlying nature of grammatical representation. Here, I demonstrate that for a range of constructions — including so-called island-violating sentences, those with resumptive pronouns,...

SNAP judgments: A small N acceptability paradigm (SNAP) for linguistic acceptability judgments: Online Appendices

The processing of phonological acceptability judgments

Interpreting Neural Network Judgments via Minimal, Stable, and Symbolic Corrections

The paper describes a new algorithm to generate minimal, stable, and symbolic corrections to an input that will cause a neural network with ReLU neurons to change its output. We argue that such a correction is a useful way to provide feedback to a user when the neural network produces an output that is different from a desired output. Our algorithm generates such a correction by solving a serie...

Magnitude Estimation and the Non-Linearity of Acceptability Judgments

The term experimental syntax – the use of psycholinguistic methodologies for the collection of acceptability judgments – can cover any number of designs, tasks, and statistical analyses (Cowart 1997, Schütze 1996). Over the past decade, one task in particular, the magnitude estimation task, has received significant attention for its alleged ability to provide more accurate data, almost to the p...