This study is based on a novel model of analysis proposed in (Author 2018) that combines results from experimental research in theory of mind (ToM) (Goldman 2006; Apperly 2010; Wilkinson & Ball 2012) with the notion of intersubjectivity in usage-based linguistics (i.a. Verhagen 2005; Nuyts 2012; Traugott 2012). The present approach to intersubjectivtiy is based on a mismatch between interaction as mere ‘co-action’ vs. interaction as spontaneously communicated awareness of an(other) mind(s). We provide two case studies centred on the first language acquisition of the aspectual/evidential marker 过 guo and the sentence-final particle 吧 ba in Mandarin. A combination of multiple correspondence analysis and mixed effects logistic regression of spontaneous use of the two markers indicates that, beyond expressions of joint attention, children’s ToM ability progressively underpins ‘ad-hoc’ generalised instantiations of extended intersubjectivity. Extended intersubjectivity underpins the socio-cognitive skill to overtly problematise what a general persona would act, feel, know, or potentially think in a specific context (Author 2018). This usage-based model further supports the evolutionary hypothesis of a shift from triadic to collective intentionality (cf. Tomasello 2019: 7).