Modeling subset-superset relations in shifty variation
Sandhya Sundaresan, Stony Brook University
This mini-course will look at shifty variation in two types of shifty phenomena: (i) perspectival anaphora: i.e. constructions where an anaphor targets the mental or spatial perspective of a salient in- dividual that is not a participant of the utterance-context (Nishigauchi, 2014; Sundaresan, 2018b; Char- navel, 2019), and (ii) indexical shift, delineating cases where the reference of an intensionally embedded indexical pronoun is evaluated relative to the parameters of the intensional predicate rather than wrt. the parameters of the utterance-context (Schlenker, 2003; Deal, 2020; Sundaresan, 2018a).
A striking property of shifty variation is that it is not random but implicationally ordered. For in- stance, the intensional environments that license perspectival anaphora are implicationally restricted in the following sense (an observation going back to Culy, 1994, based on an investigation of perspectival anaphora in 32 languages):
(1) INTENSIONAL HIERARCHY FOR PERSPECTIVAL ANAPHORA:
Speech > Thought > Knowledge > Direct perception
IMPLICATION: if an anaphor is licit in the scope of a certain predicate-class, it is necessarily also licit in the scope of all predicate-classes to its left on the hierarchy.
Analogously, the types of perspective-taking are also implicationally ordered (Sells, 1987):
(2) SOURCE (speaker) > SELF (attitude-holder) > PIVOT (spatio-temporal center):
If a perspectival anaphor in a given language can be bound under a (predicate that provides a) PIVOT, it will also necessarily be licensed under SELF and, in turn, under SOURCE.
The hierarchy in (1) also regulates the availability of indexical shift (Sundaresan, 2012, 2018a; Deal, 2020) crosslinguistically. Variation in which indexicals may shift, both across languages and in a given environment, are also implicationally restricted. As discussed in Deal (2017, 24), there is no language (or individual structure) that shifts ‘you’ to the exclusion of ‘I’ or ‘here’ to the exclusion of ‘you’ (and ‘I’). But the reverse patterns are amply attested.
The shifty hierarchies described so far have all been documented in the literature. We will look at two additional types of implicational dependency which have been significantly less discussed (based on my recent work in Sundaresan, 2021): (a) implicational dependency between perspectival anaphora and indexical shift: I will present evidence showing that the availability of indexical shift in a given environment entails that of perspectival anaphora in that environment, but not vice-versa; (b) subset- superset relation in the internal structures of shifty vs. rigid indexicals crosslinguistically: I will argue (inspired by work on person restrictions in Raynaud, 2020) that shifty indexicals are weak pronouns with a nominal structure that lacks a D layer while rigid indexicals are strong pronouns whose structure subsumes that of shifty indexicals and contains a D layer.
We will look at how these implicational dependencies can be modelled in a selectional, monotonic syntax and explore their consequences for semantics. In so doing, we will also develop a tentative template of attitude shift which can capture these cross-cutting implications, both across the licensing environments and across the shifty elements.