Conflicting Interests and Language Change GURT 2014 Christopher - - PowerPoint PPT Presentation

Negation Analyses Signaling Cycles Conclusion

Conflicting Interests and Language Change

GURT 2014 Christopher Ahern & Robin Clark

University of Pennsylvania

May 20, 2014

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 1 / 38

Negation Analyses Signaling Cycles Conclusion

Questions

What are the causes of language change? What role might “conflicts” of interest play?

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 2 / 38

Negation Analyses Signaling Cycles Conclusion

Outline

1

Negation

2

Analyses

3

Signaling

4

Cycles

5

Conclusion

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 3 / 38

Negation Analyses Signaling Cycles Conclusion Jespersen

[Jespersen(1917)]

The Negative Cycle

1

N V

2

N V (N)

3

N V N

4

(N) V N

5

V N

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 4 / 38

Negation Analyses Signaling Cycles Conclusion Jespersen

[Jespersen(1917)]

The original negative adverb is first weakened, then found insufficient and therefore strengthened, generally through some additional word, and this in turn may be felt as the negative proper and may then in the course of time be subject to the same development as the original word. Sometimes it seems as if the essential thing were only to increase the phonetic bulk of the adverb...

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 5 / 38

Negation Analyses Signaling Cycles Conclusion Jespersen

[Jespersen(1917)]

Cause and/or effect

Phonetic Syntactic Semantico-Pragmatic

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 6 / 38

Negation Analyses Signaling Cycles Conclusion Emphasis

[Kiparsky and Condoravdi(2006)]

Emphatic negation tends to increase in frequency due to pragmatically motivated

• veruse which is characteristic of inherently

bounded evaluative scales...an obligatory element cannot be emphatic, for to emphasize everything is to emphasize nothing.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 7 / 38

Negation Analyses Signaling Cycles Conclusion Emphasis

[Kiparsky and Condoravdi(2006)]

PLAIN EMPHATIC SOURCE

ου...τι ου-δε...εν Ancient Greek (ου)δεν...τι δεν...τιποτε Early Medieval Greek δεν...τιποτε δεν... πραμα Greek Dialects δεν...πραμα δεν...απαντοξη Modern Cretan

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 8 / 38

Negation Analyses Signaling Cycles Conclusion Emphasis

[Israel(1998)]

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 9 / 38

Negation Analyses Signaling Cycles Conclusion Emphasis

[Davis et al.(2007)Davis, Potts, and Speas]

Quality Threshold [Potts(2007)]

An utterance U by speaker S in context C satisfies quality iff its quality rating, µC(U), is above the quality threshold Cτ for C.

Emphasis

An emphatic utterance U by speaker S in context C conventionally implicates commitment to some higher quality threshold C′

τ > Cτ,

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 10 / 38

Negation Analyses Signaling Cycles Conclusion Emphasis

[Dahl(2001)]

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 11 / 38

Negation Analyses Signaling Cycles Conclusion Information Structure

[Schwenter(2006)]

(The plain-emphatic distinction) is problematic from the present-day perspective of other Romance languages...the post-verbal negative element is heavily regulated by information-structural factors, and specifically by the discourse-old status of the denied proposition.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 12 / 38

Negation Analyses Signaling Cycles Conclusion Information Structure

[Schwenter(2006)]

DISCOURSE NEW

INFERABLE

DIRECTLY ACTIVATED

N V 1 1 1 N V N 1 1 V N 1

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 13 / 38

Negation Analyses Signaling Cycles Conclusion Information Structure

Extend [Davis et al.(2007)Davis, Potts, and Speas]

Inferability Threshold

An utterance U by speaker S in context C satisfies inferability iff its inferability, µC(U), rating is above the inferability threshold Cι for C.

Inferability

Discourse New: µC(U) = 0 Inferable: µC(U) ∈ [0,1] Directly Activated: µC(U) = 1

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 14 / 38

Negation Analyses Signaling Cycles Conclusion Information Structure

The proof is trivial!

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 15 / 38

Negation Analyses Signaling Cycles Conclusion Causes

[Grice(1975)]

I am, however, enough of a rationalist to want to find a basis that underlies these facts, undeniable though they may be; I would like to be able to think of the standard type of conversational practice not merely as something that all or most do in fact follow but as something that it is reasonable for us to follow, that we should not abandon.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 16 / 38

Negation Analyses Signaling Cycles Conclusion

Crucial points for Games

1

Signaling Games

2

Nash Equilibria

3

Evolutionarily Stable Strategies

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 17 / 38

Negation Analyses Signaling Cycles Conclusion Signaling Games

[Lewis(1969)]

One if by land, two if by sea.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 18 / 38

Negation Analyses Signaling Cycles Conclusion Signaling Games

δ S R 1,1 aland 0,0 asea mone R 1,1 aland 0,0 asea mtwo tland S R 0,0 aland 1,1 asea mone R 0,0 aland 1,1 asea mtwo tsea

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 19 / 38

Negation Analyses Signaling Cycles Conclusion Signaling Games

Sender observes some state of the world, t ∈ T, given probability distribution over states, δ. Sender chooses message, m ∈ M, based on strategy s ∈ [T → M]. Receiver interprets message with action, a ∈ A, based on strategy r ∈ [M → A]. US and UR are the utility functions that define preferences over T ×A. Expected utility of sender and receiver: EUS(s,r) = ∑

t

δ(t)·US(t,r(s(t))) EUR(s,r) = ∑

t

δ(t)·Ur(t,r(s(t))) (1)

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 20 / 38

Negation Analyses Signaling Cycles Conclusion Nash Equilibria

[Nash(1951)]

A strategy profile s∗,r∗ is a Nash equilibrium if and

• nly if:

∀s ∈ S, such that s = s∗, EUS(s∗,r∗) ≥ EUS(s,r∗) ∀r ∈ R, such that r = r∗, EUR(s∗,r∗) ≥ EUS(s∗,r) “Something that it is reasonable for us to follow.”

• Grice

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 21 / 38

Negation Analyses Signaling Cycles Conclusion Evolutionarily Stable Strategies

[Maynard Smith and Price(1973)]

An Evolutionarily Stable Strategy is a strategy that, if all the members of a population adopt it, then no mutant strategy could invade the population under the influence of natural selection

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 22 / 38

Negation Analyses Signaling Cycles Conclusion Evolutionarily Stable Strategies

[Selten(1980)]

In the class of asymmetric games, a strategy is evolutionarily stable if and only if it is a Strict Nash Equilibrium. A strategy profile s∗,r∗ is a Strict Nash equilibrium if and only if: ∀s ∈ S, such that s = s∗, EUS(s∗,r∗) > EUS(s,r∗) ∀r ∈ R, such that r = r∗, EUR(s∗,r∗) > EUS(s∗,r) “Something that it is reasonable for us to follow, that we should not abandon.” -Grice

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 23 / 38

Negation Analyses Signaling Cycles Conclusion

Crucial points for Cycles

1

Preferences

2

Simulations

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 24 / 38

Negation Analyses Signaling Cycles Conclusion Preferences

Emphasis and Information Structure

Interpretation of game

T : Quality of utterance, Inferability of utterance M : Different forms of negation A : Quality of utterance, inferability of utterance

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 25 / 38

Negation Analyses Signaling Cycles Conclusion Preferences

Emphasis and Information Structure

Interpretation of "conflict"

Tendency to exaggerate quality threshold diminishes information conveyed Tendency to assume inferability threshold is met diminishes information conveyed

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 26 / 38

Negation Analyses Signaling Cycles Conclusion Preferences

[Crawford and Sobel(1982)]

Divergence

US(t,a) = −(a−t −(1−t)b])2 UR(t,a) = −(a−t)2 (2)

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 27 / 38

Negation Analyses Signaling Cycles Conclusion Preferences

Game structure

Set of states T = {t0,...,tn}, ti < ti+1 States equiprobable, δ(t1) = ... = δ(tn) = 1

n

Set of messages M = {m1,...,mk} Set of actions A = {a0,a1

0,...an}, ai < ai+1 i

< ai+1 Divergence b ∈ [0,1]

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 28 / 38

Negation Analyses Signaling Cycles Conclusion Simulations

Emphasis (n = 2)

−1.0 −0.5 0.0 0.5 1.0 25 50 75 100

Time Difference

Message m0 m1 m2

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 29 / 38

Negation Analyses Signaling Cycles Conclusion Simulations

Information Structure (n = 3)

tH−tM tM−tL −1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0 25 50 75 100

Time Difference

Message m0 m1 m2 m3

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 30 / 38

Negation Analyses Signaling Cycles Conclusion Questions

Questions

What are the causes of language change?

Phonetic Syntactic Semantico-Pragmatic

What role might “conflicts” of interest play?

Pragmatic pressures decrease information conveyed Lack of information allows for change

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 31 / 38

Negation Analyses Signaling Cycles Conclusion Future Directions

Future Directions

Theoretical

Discrete versus continuous type space Different game dynamics

Empirical

Distribution of types from corpora Comparison of historical changes Cycles versus stability

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 32 / 38

Negation Analyses Signaling Cycles Conclusion Future Directions

Thanks!

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 33 / 38

Negation Analyses Signaling Cycles Conclusion Future Directions

Bibliography I

Vincent P Crawford and Joel Sobel. Strategic information transmission. Econometrica, pages 1431–1451, 1982. "Osten Dahl. Inflationary effects in language and elsewhere, pages 471–480. John Benjamins Publishing Company, Philadelphia, 2001. Christopher Davis, Christopher Potts, and Margaret Speas. The pragmatic values of evidential sentences. In Proceedings of SALT, volume 17, pages 71–88, 2007.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 34 / 38

Negation Analyses Signaling Cycles Conclusion Future Directions

Bibliography II

H.P. Grice. Logic and conversation. In Studies in the Way of Words, pages 22–40. Harvard University Press, 1975. Michael Israel. The rhetoric of grammar: Scalar reasoning and polarity sensitivity. PhD thesis, University of California, San Diego, 1998. Otto Jespersen. Negation in English and other Languages. Host, 1917.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 35 / 38

Negation Analyses Signaling Cycles Conclusion Future Directions

Bibliography III

Paul Kiparsky and Cleo Condoravdi. Tracking Jespersen’s Cycle. In Mark Janse, editor, Proceedings of the 2nd International Conference

• f Modern Greek Dialects and Linguistic Theory. University of Patras,

2006. David Lewis. Convention. Harvard University Press, Cambridge, 1969. John Maynard Smith and George Price. The Logic of Animal Conflict. Nature, 246:15, 1973.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 36 / 38

Negation Analyses Signaling Cycles Conclusion Future Directions

Bibliography IV

John F Nash. Non-cooperative games. The Annals of Mathematics, 54(2):286–295, 1951. Christopher Potts. Conversational implicatures via general pragmatic pressures. In New Frontiers in Artificial Intelligence, pages 205–218. Springer, 2007. Scott A Schwenter. Fine-tuning Jespersen’s cycle. In Betty T. Birner and Gregory Ward, editors, Drawing the boundaries of meaning: Neo-Gricean studies in pragmatics and semantics in honour of Laurence R. Horn, pages 327–344. John Benjamins, 2006.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 37 / 38

Negation Analyses Signaling Cycles Conclusion Future Directions

Bibliography V

Reinhard Selten. A note on evolutionarily stable strategies in asymmetric animal conflicts. Journal of Theoretical Biology, 84(1):93–101, 1980.

Ahern & Clark (UPENN) Conflicting Interests and Language Change May 20, 2014 38 / 38