True <i>wh</i> -movement and <i>wh</i> -in-situ in one language: Evidence from Colloquial Singapore English

Abstract wh -fronting questions (as in English) are analyzed as wh -movement while wh -in-situ questions (as in Chinese) are analyzed as LF movement or unselective binding. Optionality between the two types of questions is observed in many languages, however, upon closer inspection, a stream of previous literature argues that only one strategy is truly available in any given language. Cheng (1991) and Faure &amp; Palasis (2021) argue that wh -fronting languages in Indonesian and Colloquial French are not derived by wh -movement, while Chang (2016) argues that wh -in-situ questions in Colloquial Singapore English (CSE) are not derived from unselective binding or LF movement but are declarative syntax questions. Bobaljik &amp; Wurmbrand (2015) explicitly propose that a language can either have the true wh -in-situ or the wh -movement strategy, but not both. This paper uses CSE as a case study and argues that it allows true wh -movement and true wh -in-situ questions. CSE has been argued to only allow wh -movement by some (Chang 2016) and to only allow wh -in-situ by others (Lan 2016). This study experimentally tests the predictions made by these analyses and shows that the patterns are best accounted for if both ‘true’ wh -movement and ‘true’ wh -in-situ questions exist in CSE (see also Sato &amp; Ngui 2017), thus challenging the previous analyses for CSE, and the cross-linguistic generalization in Bobaljik &amp; Wurmbrand 2015.