Syntactic studies make use of the minimally pairwise sentences as an argumentation tool, because the pairs allow us to pay attention to the constraints of interest. Likewise, it is helpful to use a set of minimal pairs in deep learning-based experiments for assessing the syntactic ability of neural language models. In this context, this study verifies whether the deep learning Korean model has the ability to properly distinguish the well-formed expressions and the corresponding ill-formed expressions. In the meanwhile, this study serves to examine the feasibility of the language resource constructed by the Korean government for deep learning architecture. The research is three-fold. First, we conducted an acceptability judgment testing to verify whether and how the language resource used in this study is indeed trustworthy. The results indicate that the judgments provided in the language resource converge with the judgments of our own experiment well enough. Second, we employed four Korean models such as mBERT, KoBERT, KR-BERT, KorBERT in order to evaluate how the language resource has a potentiality to predict the well-formedness of Korean expressions. The different models yield different results, the reason of which is fully discussed. Third, we made use of an independent test-set for evaluating the deep learning systems. It turns out that the results are still challenging, which implies that the current Korean models may have room for improvement to understand the syntactic phenomena.

Stripping, also called bare argument ellipsis, is taken to be an elliptical construction that unexpresses everything from a clause except one constituent (e.g., Susan works at night, and Bill too.). It can also accompany the negator not, which we call Not-Stripping (e.g., Alan likes to play volleyball, but not Sandy.). This paper reviews key grammatical properties of Not-Stripping in English and investigates its authentic uses with the online corpus COCA (Corpus of Contemporary American English). The paper then performs some quantitative and qualitative research of the extracted data and discusses theoretical implications. The result from corpus investigation shows that Not-Stripping preferably used as PP remnant. Its register distribution also shows that Not-Stripping is most frequently used in fiction register followed by spoken, and magazine registers which may indicate that Not-Stripping is mainly used in colloquial context. In terms of category match/mismatch types, the construction is overtly used as mismatching relation between the remnant and the correlate. In addition, in case of merger/sprouting types, Not-Stripping seems to favor merger type, but sprouting type can be regarded as frequently used type. The grammatical function of both remnant and correlate are preferably used as modifier. The data for Not-Stripping are dealt with move-and-delete operations to postulate putative clausal source. Observing such attested data challenging the derivation of negative stripping from syntactic movement operations, the paper challenges a direct licensing construction-based analysis of the construction.