More Adventures with Dialects: Convergence and Ambiguity Resolution by Partners in Conversation

More Adventures with Dialects: Convergence and Ambiguity Resolution

by Partners in Conversation

Brennan, Huffman, Hannigan, et al.

Assumptions about dialects

Dialects encode geographic, demographic, educational socioeconomic info

Dialects converge during interaction

Questions about dialects What IS a dialect? Do dialects really converge in conversation? How do listeners cope w/ variability in the

speech signal, including that from dialects? Are all kinds of variability handled by the same

processing mechanisms? Are representations of the sounds of a

language stored as abstract prototypes or as different instances/variants?

Is there parity between comprehension and production when it comes to dialectal variation?

Dialects aren’t all-or-nothing!

Adapting to a partner’s dialect

Do speakers converge with a partner’s dialect? (Brennan & Huffman, in prep.)

Long Island (LI) vs. General American (GA) dialect Design: Session 1 with LI addressee

Session 2 with GA addressee

Sample dialect words to elicit: Consonants

r-drop toaster, saucer, paperclip

Vowels - Monophthongs “bad” plaid, cab, labcoat, bathtub “bat” hat, bat, cat, cap “ale” scale, tail, mailbox “ah” dot, pot, knot, cot

Vowels - Diphthongs aw chocolate, faucet, saw

Task: “Go fish” card game

Subject: “I need a toaster.” Confed: “Here you go, a toaster.”

or “Sorry, no toaster, go fish.” “OK, I need a lab coat.”

<etc.>

Dialect comparisons

LI speech example:

More examples of LI subjects in the context of confederates’ speech

“cab”: LI subj., then GA confed. “saucer”: LI subj., then GA confed. “saucer”: LI conf., then LI subject

(play) (repeat)

Results so far:

Speakers do significantly less r-dropping to GA addressees than to LI addressees

Very small change in monophthongs: F1 goes up slightly, F2 goes down slightly We need to look at differences for individual words

(e.g., PLAID, CAB, LABCOAT, BATHTUB)

Dipthongs: Stay tuned!

Implications

Dialects are clusters of tendencies. Articulatory adaptations associated with dialect

don’t occur as a unit. Some features of a dialect are more flexible

than others. Adaptation may depend on awareness of a

feature!

Adapting to a speaker’s dialect

Do speakers converge with their partners’ dialects? (Brennan & Huffman, in prep.)

Do listeners perceive homophones differently, depending on the dialect of the speaker? And does the listener’s own dialect matter?

(with Stephanie Hannigan & Marie Huffman)

The parity issue, vis-à-vis dialects

The comprehension (input) and production (output) issues are different.

Dialectal variation is particularly hard to deal with on the input side.

When a dialect is unfamiliar, the source of the variation is ambiguous

Tourist: Excuse me, is it pronounced ‘Hawaii’ or

‘Havaii’?

Benny Hill: Havaii

Tourist: Thank you!

Benny Hill: You’re velcome!

(Kraljic, Brennan, & Samuel, under review)

Welcome to Long Island!

(In a deli, ordering a sliced turkey sandwich:)

(example courtesy of Meghan Sumner)

Welcome to Long Island!

(In a deli, ordering a sliced turkey sandwich:)

Would you like white meat or dog meat?

(example courtesy of Meghan Sumner)

Adapting to a speaker’s dialectType A Type B Type C

cod cawed cord

pod pawed poured

tot taught torte

stock stalk stork

sod sawed soared

tock talk torque

cock caulk cork

cot caught court

Adapting to a speaker’s dialect

GA homophones

Type A Type B Type C

cod cawed cord

pod pawed poured

tot taught torte

stock stalk stork

sod sawed soared

tock talk torque

cock caulk cork

cot caught court

Adapting to a speaker’s dialect

LI homophones

Type A Type B Type C

cod cawed cord

pod pawed poured

tot taught torte

stock stalk stork

sod sawed soared

tock talk torque

cock caulk cork

cot caught court

Adapting to a speaker’s dialect

Subjects’ own dialects are testedQuestionnaireRead a story aloud

Exposed to either LI or GA dialect (story) Perform a word selection task Critical trials had either an LI or a GA

homophone competitor

Does perceived ambiguity depend on who the speaker is?

Type A Type B Type C

cod cawed cord

pod pawed poured

tot taught torte

stock stalk stork

sod sawed soared

tock talk torque

cock caulk cork

cot caught court

Does perceived ambiguity depend on who the speaker is? “Click cawed”

stair

cod + cawed

ice

Does perceived ambiguity depend on who the speaker is? “Click cawed”

stair

cod + cawed

ice

(GA dialect)

Does perceived ambiguity depend on who the speaker is? “Click cawed”

stair

cod + cawed

ice

(LI dialect)

Does perceived ambiguity depend on who the speaker is? “Click caught”

told

court + caught

ski

Does perceived ambiguity depend on who the speaker is? “Click caught”

told

court + caught

ski

(LI dialect)

Does perceived ambiguity depend on who the speaker is? “Click caught”

told

court + caught

ski

(GA dialect)

An utterance’s ambiguity depends on the perceived dialect…

If it’s all about priming, then who the speaker is shouldn’t matter!

Type A Type B Type C

cod cawed cord

pod pawed poured

tot taught torte

stock stalk stork

sod sawed soared

tock talk torque

cock caulk cork

cot caught court