Pertsova (2022): A case for a binary feature underlying clusivity #
[Per22] argues from a 270-language pronoun database that inclusive is a kind of first person distinguished from the exclusive by a binary feature, against containment accounts in which a privative [ADDRESSEE] makes the inclusive a superset of the exclusive. Her feature hierarchy (4) places AUTHOR under +PARTICIPANT and ADDRESSEE under AUTHOR, so a language activates PARTICIPANT, then AUTHOR, then ADDRESSEE, and the categories are the referent sets the active feature values pick out (5). The hierarchy predicts which persons can be fully conflated: activating PARTICIPANT, AUTHOR and ADDRESSEE in turn gives two, three and four persons, and inclusive never conflates with second person, nor second with third, nor first with third (p. 402), conflation arising only from inactive features (p. 403). Against [Har16a] she argues that his author bipartition, second with third against first, is doubtful for pronouns (pp. 425–426), and she grounds the priority of AUTHOR over ADDRESSEE in the privileged status of speakers (p. 426).
Main results #
image_graph_active: the active feature sets the hierarchy allows generate exactly four systems, monopartition, participant bipartition, tripartition and quadripartition.graph_active_ne_authorBipartition: none of them is Harbour's author bipartition.no_partial_conflation: inclusive conflates with second person only with the exclusive too, second with third only with the first too, and first with third only with the second too.
Implementation notes #
The features are those of Harbour2016.ThreeFeature on participant sets, ADDRESSEE being
[hearer]. With no feature active the hierarchy gives the monopartition, which the paper does not
discuss; it is included as the empty initial segment. Number is not modeled, so the referent sets
of (5) appear as their participant sets, me+other(s) as {speaker}. Pertsova's argument from
ABA patterns and derived exclusives (§§2.3, 5) needs vocabulary insertion with blocking and is
not formalized.
References #
The feature hierarchy (4): AUTHOR depends on +PARTICIPANT and ADDRESSEE on AUTHOR, so the active features form an initial segment of participant, author, addressee.
Equations
Instances For
Equations
- Pertsova2022.instDecidablePredFinsetThreeFeatureActive x✝ = id inferInstance
The active feature sets generate four systems: no person distinction, participants against others, three persons, and four with clusivity (p. 402; the first is the empty activation).
Harbour's author bipartition, which [±author] alone generates, is not among them.
Conflation only arises from inactive features (pp. 402–403): inclusive conflates with second person only if the exclusive does too, second with third only if first does too, and first with third only if second does too.
Harbour's bivalent geometry (9) admits the author bipartition, which the hierarchy (4) excludes: the two differ only in whether AUTHOR depends on PARTICIPANT.