Byron Ahn (Princeton University) presents:
Syntactic Influences on Accent Placement
In this talk, I will present data on the distribution of certain pitch accents in English. We will look at pitch accents that associate with discourse-neutral phrasal stress (a.k.a. ‘New’) and those that associate with con- trastive focus stress (a.k.a. ‘Focus’). We will conclude that placement of New pitch accents is constrained by syntax, and that Focus pitch accents and abstract Focus marking may diverge in what they mark, in a highly constrained way, again mediated by syntax.,
A longstanding generalization is that New pitch accents seem to fall on the final member of particular prosodic domains in English (e.g., Chomsky and Halle 1968’s Nuclear Stress Rule):
(1) a. Hazel glued Kenneth to the chair.
b. We thanked Sarah.
c. Every morning, Bob turns on the radio.
There are apparent exceptions to such a generalization:
(2) a. Hazel glued Kenneth to herself.
(Reflexive Pronoun – not pitch accented)
b. Sarah thanked us.
(Non-Reflexive Pronoun – not pitch accented)
c. Every morning, Bob turns the radio on.
(Verb Particle – not pitch accented)
Moreover, there are exceptions to these exceptions:
(3) a. Hazel glued Kenneth to himself.
(Reflexive Pronoun – pitch accented)
b. Sarah thanked Bob or us.
(Non-Reflexive Pronoun – pitch accented)
c. Every morning, Bob turns it on.
(Verb Particle – pitch accented)
We will look closely at the case of the reflexives, and draw conclusions on the syntactic influences on pitch accent placement. This will lead us to a model in which syntactic structure (and not linear order) plays a critical role in the distribution of New pitch accents. This will lead us to (independently corroborated) predictions about the (b-c) examples above.
Another longstanding generalization is that Contrastive Focus pitch accents fall on elements that serve to answer a (tacit) Question Under Discussion (in the sense of Roberts 1996).
(4) a. Hazel glued KENNETH to himself.
(QUD: Who did Hazel glue to themselves?)
b. We THANKED Sarah.
(QUD: What did we do for Sarah?)
c. Every morning, BOB turns on the radio.
(QUD: Who turns on the radio every morning?)
This generalization, known as Question Answer Congruence, is remarkably robust. For this reason, the follow- ing is very surprising:
(5) Hazel glued Kenneth to HERSELF.
a. Apparent QUD1: Who did Hazel glue Kenneth to?
b. Apparent QUD2: Who glued Kenneth to Hazel?
(6) We thanked OURSELVES.
a. Apparent QUD1: Who did we thank?
b. Apparent QUD2: Who thanked us?
The latter apparent QUD is unpredicted on the basis of standard assumptions about the syntactic and semantic structures posited for sentences like (5) and (6). I argue that this is an issue of syntax (Contrastive Focus is marked in the syntax; e.g., Selkirk 2007), and that what is really focused in the contexts of (5b) and (6b) is a silent semantic operator. A theory of QAC predicts such ‘mismatches’ to occur (if there are silent operators with semantic content), and this data informs what kinds of analyses are possible in ‘mismatch’ conditions.
Thursday, April 12, 2018, 4:15pm - 6pm Graduate Center, 365 Fifth Ave, room 6417
All are welcome!
Refreshments to follow in room 7400
Dinner to follow refreshments. Please RSVP to firstname.lastname@example.org