Join us this Friday, September 23, from 3:40-5:40PM in Hamilton 717 for our first speaker of the year! Chris Barker, professor of linguistics at NYU, will be speaking on the pros and cons of viewing actions as semantic objects, as informed by the computational, logical and linguistic literatures.
On ontologically parsimonious approaches (Portner, Schwager), actions
are characterized indirectly, via propositions. Then an imperative
such as “Eat an apple!” places an obligation on the addressee to bring
it about that we inhabit a world in which an apple has been eaten.
On certain computational and logical approaches (Pratt, Harel,
Segerberg; Wikipedia topic Dynamic_logic_(modal_logic)), actions are
relations over worlds: eating an apple in world w changes it into
world w’. Then “Eat an apple!” is an instruction to change the world
in an apple-eating way. In this style of dynamic semantics, actions
update worlds (rather than updating discourse states, as in more
familiar dynamic semantics such as DPL). In the linguistics
literature, as far as I know, only Lascarides and Asher propose an
analysis of imperatives involving actions.
In this talk, I will consider some of the pros and cons of recognizing
actions as a legitimate kind of semantic object. On the pro side,
besides providing a simple and appealing picture of the meaning of
imperatives and related expression types, actions also provide a
natural account of the interaction of imperatives and deontic modality
with disjunction. This problem is known in the philosophical
literature as Ross’s paradox, and in the linguistics literature as the
problem of free-choice permission: from “John may eat an apple”, it is
not valid to infer “John may eat an apple or a pear”, despite the fact
that eating an apple entails eating an apple or a pear. On the action
view, since the composite action of eating an apple or eating a pear
is a larger relation than the action of eating an apple, no inference
The cons include some difficulty seeing how to gracefully embed a
logic of action within a more general grammatical framework. At stake
is the proper conception of the basic logical operation of
disjunction, as well as the degree to which the various uses of
disjunction in natural language can be unified.