Acceptability Judgments in ChatGPT (Study 2)

In this study, I am looking at structure dependence, which is an important concept for generative syntax. Based on a standard poverty of stimulus argument, structure dependence is taken to be part of UG (or derived from third factor considerations).

Acceptability Judgments in ChatGPT

On a dreary Saturday afternoon, I decided to ask ChatGPT a bunch of standard acceptability judgments. My prompt was this:

For each sentence, tell me if it is acceptable in spoken English or not. Do not give any explanations.

After that, I just typed in the sentences one after the other. And it spit out either 'Acceptable' or 'Not Acceptable.'