Randolph Quirk Fellow Workshop #2: Prof. Martina Wiltschko
When: Tuesday, May 20, 2025, 2:00 PM - 4:00 PM
Where: ArtsTwo 2.17 and online, Mile End Campus
Workshop 2: Does knowledge of language use affect the way we talk to ourselves?
Click here to join online.
What do humans know when they use language?
One of the core goals of the generative enterprise (and other cognitive approaches towards language) is to explore what humans know when they know a language. To achieve this goal, generative linguists typically explore properties of individual languages, the commonalities across languages, as well as the range and limits of observed variation. What has long been ignored in this approach are aspects of language restricted to language in use. Two assumptions that prevail the generative enterprise have conspired to this status quo:
(1) Linguistic competence is the object of investigation and is to be distinguished from linguistic performance.
(2) The sentence is the unit of grammatical analysis.
On this view, units of language (UoLs) that appear at the periphery of sentences and which serve to regulate linguistic interaction (oh, huh, well ...) are not considered to fall within the purview of grammar. However, over the past two decades there is a growing consensus that the distribution of such UoLs can be successfully analyzed on the hypothesis that they occupy the very top of syntactic structures. If so, this has profound implications for the two assumptions above as summarized in (1') and (2').
(1') UoLs that serve to regulate interaction should be considered part of linguistic competence.
(2’) Grammatical analysis should not be restricted to the classic notion of the sentence.
While there are several competing proposals to model the integration of aspects of language in use into our knowledge of language, in this series of workshops I focus on one such proposal: the Interactional Spine Hypothesis (Wiltschko 2021), according to which grammatical knowledge includes not only knowledge of how to construct the propositional aspects of language (p-language) but also its interactional aspects (i-language).
Once we acknowledge that (at least some) conditions on language use are part of our grammatical competence, new research questions emerge as well as the need for new methodologies which go beyond the exploration of sentences in isolation. During the three workshops, I explore some of these questions and methodologies.
Workshop #2 Abstract:
Most frameworks that assume aspects of language in use to be part of syntactic structure minimally include a representation of the interlocutors. The Interactional Spine Hypothesis is no exception. Interactional roles are assumed to be introduced in dedicated grounding phrases, where speaker and addressee are represented as the holders of their individual grounds (a placeholder for their epistemic states). If grammatical knowledge includes representations of the (epistemic states of) speaker and (what the speaker assumes to be the epistemic state of their) addressee, then it is predicted that utterances in interaction are sensitive to the identity of the speaker and addressee, and the relation between them. This in turn requires a methodological shift in that the well-formedness of utterances must be judged relative to a particular context that takes the identity of the interlocutors into consideration. Much current research focusses on the use of honorifics and other addressee-oriented formality strategies to empirically explore the nature of these representations.
In this workshop I explore this question from a novel angle, namely self-talk. That is, if the addressee role is part of the grammatical representation of i-language, then the question arises as to what happens in situations when people talk to themselves, i.e., when there is no addressee. I demonstrate that self-talk provides us with striking evidence for the grammatical representation of the (epistemic state of the) addressee role.