Evolutionary Change in Motion Event Encoding in Indo-European

8/12/2019 Evolutionary Change in Motion Event Encoding in Indo-European

1/41

EVOLUTIONARY CHANGE ININDO-EUROPEANMOTION EVENT ENCODING

Annemarie Verkerk

SCCR 22/25-02-2011


2/41


3/41

TALMYS MOTION TYPOLOGY

Spanish (verb-framed)La botella entr a la cueva flotandothe bottle moved.into the cave floating PATH MANNER

Talmy (1985: 69)

Dutch (satellite-framed)De fles dreef de grot inthe bottle floated the cave into MANNER PATHThe bottle floated into the cave


4/41

The typology of motion expressions revisited1

JO HN BEA V ER S

Department of Linguistics, The University of Texas at Austin

BETH LEV I N

Department of Linguistics, Stanford University

SHI A O W EI THA M

Department of East Asian Languages and Literatures, Wellesley College

(Received 20 March 2008; revised 15 January 2009)

J. Linguistics46 (2010), 331377. f Cambridge University Press 2009doi:10.1017/S0022226709990272 First published online 30 November 2009

The Many Ways to Search for a FrogLinguistic Typology and the Expression of Motion Events

Dan 1 Siobin

1 INTRODUCTIONThe chapters in this volume, along with the extensive list of frog-story studies inAppendix II provide a rich database for the exploration of particular questionsof language use and acquisition. The studies reported in Part I reflect a range oflanguages of different types, making it possible to focus on the role of linguistictypology in narrative construction. l A recurrent concern in those studies is theexpression of motion, which is one of the dominant themes of Frog where areyou? In one way or another, all of the studies confront Talmy's by now familiartypology of verb-framed and satellite-framed languages (Talmy 1985, 1991,2000b). Briefly, the typology is concerned with the means of expression of thepath of movement. In verb-framed languages ( V-languages ) path is expressedby the main verb in a clause (,enter', 'exit', 'ascend', etc.), whereas in satellite-framed languages ( S-languages ) path is expressed by an element associatedwith the verb ('go in/out/up', etc.). This dichotomy has engendered a good dealof esearch and debate in the literature on motion-event descriptions over the pastdecade or SO.2 n this concluding chapter on typological perspectives I suggestthat several different sorts of factors conspire to produce a range of frog-storyvarieties. These varieties result from combined influences of linguistic structure,on-line processing, and cultural practices. Talmy's typology was designed tocharacterize lexicali zation patterns, and it has provided important insights into theoverall set of structures that define individual languages. However, the typologyalone cannot account for discourse structures, because language use is determinedby more than lexicalization patterns. t is striking how much has been learned byapplication of the V-Ianguage/S-Ianguage contrast, and it still plays a part in themix of factors considered here. But a fuller account of narrative organization willrequire attention to a range of morphosyntactic, psycho inguistic, and pragmatic

Kyoko Hirose Ohara, Seiko Fujii and Charles J Fillmore

to solve as you m ight have thought. Presented at the Fifthtructio n Gramm ar, Austin, Texas.. Grammatical constructions and linguistic generaliza-uction. Language 75,1-33.ot as hard a problem to solve as you might have thought.o ognitive grammar: Volume I: Theoretical prerequisites..

nalysis of Pawnee kinship usage. Language 32,158-194.Sweetser (Eds.), Spaces, worlds, and grammar 124-156).ess.rammar, and multilinguality in the Japanese FrameNet.

l Conference on Language Resources and Evaluation.io Ohori, Ryoko Suzuki, Hiroaki Saito Shun Ishizakiroject: An introduction. Proceedings o he 4th Interna-ources and Evaluation.minology in Hebrew: A study in Lexical Semantics. Dis-

Revising Talmy s typolog ical classificationo complex event constructionsWilliam Croft, Johanna Barodal, Willem Hollmann,Violeta Sotirova, and Chiaki TaokaUnivers i ty of New Mexico, USA, University of Bergen, Norway,Univers i ty of Lancaster, UK, University of Nottingham, UK,and Kobe Col lege, Japan

1. Introduction


5/41

Alices Adventures inWonderland (Lewis Carroll)

Through the Looking-Glass andwhat Alice found there

O Alquimista (Paulo Coelho)

308 motion sentences

PARALLEL CORPUS


6/41


7/41

MOTION ENCODING

STRATEGIES Alice laughed so much at this, thatsatellite-framed: she had to run back into the wood

for fear of their hearing her;

verb-framed: she had to enter the wood running / at a run / quickly

path-only: she had to enter the wood

manner-only: she had to run in the wood

deictic: she had to go into the woodcoordination: she had to run and go back into the wood

subordination: she had to run to go back into the wood

manner+path verb: she had to run+enter the wood

other: she was in the wood


8/41

MOTION ENCODING IN IE

o ri gi na ls r us si an p ol is h l it hu an ia n s we di sh g er ma n d ut ch l at vi an e ng li sh i ri sh s er bo cr oa ti an gr ee k i ta li an p or tu gu es e r om an ia n f re nc h h in di p er si an a rm en ia n a lb an ia n n ep al i

proportion

persentence

0.

0

0.

2

0.

4

0.

6

0.

8

1.

0


9/41

MOTION ENCODING IN IE

o ri gi na ls r us si an p ol is h l it hu an ia n s we di sh g er ma n d ut ch l at vi an e ng li sh i ri sh s er bo cr oa ti an gr ee k i ta li an p or tu gu es e r om an ia n f re nc h h in di p er si an a rm en ia n a lb an ia n n ep al i

proportion

persentence

0.

0

0.

2

0.

4

0.

6

0.

8

1.

0

satellite-framed

verb-framed

path-only


10/41

PHYLOGENETICCOMPARATIVE METHODS


11/41

PHYLOGENETIC COMPARATIVE METHODS


12/41

PRINCIPAL COMPONENT

ANALYSIS

65% of the variance is

explained by the firstprincipal component

Take the score of each

language on the firstprincipal component

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3

-0.

20

-0.

15

-0.

10

-0.

05

0.

00

0.

05

0.

10

First principal component

S

econdprincipalcomponen

t

Russian

Polish

Lithuanian

Dutch

German

English

Irish

Greek

French

Portuguese

Armenian

Hindi

Albanian

Persian

Nepali

Romanian

Swedish

Serbo-Croatian

Italian

Latvian


13/41

TESTING HISTORICAL SIGNAL

Phylogenetic trees:

- from Dunn et al. (2011)

- build on lexical data (Swadesh lists)

- estimated using a Bayesian MarkovChain Monte Carlo approach

Tocharian BTocharian A

Albanian G

Greek ModAncientGreek

Armenian Mod

FrenchProvencalWalloonLadin

Italian

Catalan

SpanishPortuguese ST

Sardinian CRumanian List

Latin

Old NorseDanish

Swedish List

Riksmal

FaroeseIcelandic ST

German STDutch List

FrisianFlemish

Afrikaans

Pennsylvania DutchLuxembourgishEnglish ST

Old English

Gothic

Old Church Slavonic

SlovenianSerbocroatian

MacedonianBulgarian

Lusatian ULusatian LSlovak

CzechUkrainianByelorussian

RussianPolish

LatvianLithuanian ST

NepaliBiB

Kashmiri

Kurdish

TadzikPersian List

Baluch

Scots GaelicIrish B

Welsh N

Breton STCornish

Hittite

Indo-European


14/41

DATA + TREES

Albanian 0.21

Persian 0.12

Armenian 0.13

Modern Greek 0.06

Nepali 0.23

Hindi 0.06

Serbo-Croatian 0.04

Russian -0.25

Polish -0.17

Latvian -0.11 Lithuanian -0.16

English -0.05

Swedish -0.20

Dutch -0.17

German -0.15

Portuguese 0.12 French 0.15

Italian 0.10

Romanian 0.12

Irish -0.10

1

0.43

0.37

0.75

0.44

11

1

1

10.7

0.4

1

1 1

0.4

1

0.50.6


15/41

HISTORICAL SIGNAL


16/41

the likelihood of real trees is significantly different fromlikelihood of trees with zero lambda (p < 0.01)

(Pagel 1999)


Albanian

Persian

Armenian

Modern Greek

Nepali

Hindi

Serbo-Croatian

Russian

Polish

Latvian

Lithuanian

English

Swedish

Dutch

German

Portuguese French

Italian

Romanian

Irish

Albanian

Persian

Armenian

Modern Greek

Nepali

Hindi

Serbo-Croatian

Russian

Polish

Latvian

Lithuanian

English

Swedish

Dutch

German

Portuguese French

Italian

Romanian

Irish


17/41


Blomberg et al. (2003)

Albanian

Persian

Armenian

Modern Greek

Nepali

Hindi

Serbo-Croatian

Russian

Polish

Latvian

Lithuanian

English

Swedish

Dutch

German

Portuguese French

Italian

Romanian

Irish

Albanian

Dutch

Persian

Serbo-Croatian

Modern Greek

Hindi

French

Nepali Polish

Lithuanian

Romanian

English

Armenian

Swedish

Irish

German

Russian

Portuguese

Latvian

Italian

the kappa score provided by this analysis shows that historicalsignal is present


18/41

Albanian 0.21

Persian 0.12

Armenian 0.13

Modern Greek 0.06

Nepali 0.23

Hindi 0.06

Serbo-Croatian 0.04

Russian -0.25

Polish -0.17

Latvian -0.11 Lithuanian -0.16

English -0.05

Swedish -0.20

Dutch -0.17

German -0.15

Portuguese 0.12 French 0.15

Italian 0.10

Romanian 0.12

Irish -0.10

1

0.43

0.37

0.75

0.44

11

1

1

10.7

0.4

1

1 1

0.4

1

0.50.6

TESTING FOR HISTORICAL SIGNAL


19/41

go back in time

ANCESTRAL STATE ESTIMATION


20/41


PIE?Albanian

Persian

Armenian

Modern Greek

Nepali

Hindi

Serbo-Croatian

Russian

Polish

Latvian

Lithuanian

English Swedish

Dutch

German

Portuguese

French

Italian

Romanian

Irish


21/41


PIE = satellite framed?(Talmy 2007, Acedo Matelln

and Mateu 2008)


22/41


Albanian

Persian

Armenian

Modern Greek

Nepali

Hindi

Serbo-Croatian

Russian

Polish

Latvian

Lithuanian

English Swedish

Dutch

German

Portuguese

French

Italian

Romanian

Irish

PIE?

??

?

?

??

??

?

??

?

?

?

?

Maximum Likelihoodtransition rate:

-0.50.5

Russiant ua n a n

Armenianr s

Armenianr s

Russiant u an an


23/41

ArmenianGreek

RussianAlbanian

Armenian

Greek

Albanian

ArmenianGreek

RussianLithuanian

EnglishDutchGerman

PortugueseFrenchIrish

Albanian

ArmenianGreek

RussianLithuanian

English

DutchGerman

PortugueseFrench

Irish

AlbanianArmenian

Greek

RussianLithuanian

English

utcGerman

Portuguese

French

Irish

Albanian

Armenian

Greek

RussianLithuanian

EnglishDutcherman

PortugueseFrench

Irish

Albanian

ArmenianGreek

ArmenianGreek

Russian

Albanian

ArmenianGreek

RussianLithuanian

EnglishDutchGerman

PortugueseFrench

Irish

Albanian

ArmenianGreek

RussianLithua

EnglishDutchGerman

PortugueFrenchIrish

Albanian

ArmenianGreek

RussianLithuanian

EnglishDutcherman

PortugueseFrench

Irish

AlbaArmenian

Gree

RussianLithuania

English

utcGerman

Portugu

French

Irish

or ugueseFrench Dutch

ermanDutcherman or ugues

French

n. 346n. 345 n. 347 n. 348

n. 619 n. 621n. 620 n. 622

n. 887n. 886 n. 889n. 888

?

?

?

?????

?-0.50.5


24/41

ANCESTRAL STATE

ESTIMATION

-0.25 0.23

Russian

Swedish

Polish

Dutch

Lithuanian

German

Latvian

Irish

English

Serbo-Croatian

Hindi

ModernGreek

Italian

Persian

Romanian

Portuguese

Armenian

French

Albanian

Nepali


25/41

ANCESTRAL STATE

ESTIMATION

-0.25 0.23

Russian

Swedish

Polish

Dutch

Lithuanian

German

Latvian

Irish

English

S

erbo-Croatian

Hindi

ModernGreek

Italian

Persian

Romanian

Portuguese

Armenian

French

Albanian

Nepali

Root estimate PIE:

between -0.02 and 0.09


26/41

INCORPORATING INFORMATIONFROM ANCIENT LANGUAGES

PIE?Albanian

Persian

Armenian

Modern Greek

Nepali Hindi

Serbo-Croatian Russian

Polish

Latvian

Lithuanian

English Swedish

Dutch

German

Portuguese French

Italian

Romanian

Latin

Irish

Sanskrit


27/41

THE EVOLUTION OF THE

PIE PREVERB SYSTEM

PIE Latin Romance Prep.+Noun ... VerbPreverb ... Verb... Preverb Verb

... prefix-Verb ... VerbWatkins (1964), Vincent (1999), Iacobini & Masini (2006)

! !


28/41


29/41

THE EVOLUTION OF THE

PIE PREVERB SYSTEM

PIE/Sanskrit later Sanskrit modern lang.Prep.+Noun ... Verb lost

Preverb ... Verb Noun+Post. ... Verb... Preverb Verb

... prefix-Verb ... VerbWatkins (1964), Bloch (1965)

! !


30/41


31/41

INCORPORATING INFORMATIONFROM ANCIENT LANGUAGES

Albanian

Persian

Armenian

Modern Greek

Nepali Hindi

Serbo-Croatian Russian

Polish

Latvian

Lithuanian

English Swedish

Dutch

German

Portuguese French

Italian

Romanian

Latin

Irish

Sanskrit

PIE?

??

?

?

?

?

?


32/41


33/41

CORRELATED EVOLUTION language encoding manner verb class1. Russian -0.23 332. Swedish -0.18 233. Polish -0.15 264. Lithuanian -0.14 265.

Dutch

-0.14

216. German -0.12 30

7. Latvian -0.09 268. Irish -0.08 149. English -0.03 3010. Greek 0.09 1611. Hindi 0.08 1612. Italian 0.13 1613. Persian 0.14 1514. Portuguese 0.15 2015. Armenian 0.15 1516. French 0.18 1317. Albanian 0.24 11


34/41

CORRELATED EVOLUTION

English

Dutch

RussianGerman

Lithuanian

Portuguese

IrishArmenianGreek

Albanian

0.2 0.3 0.4 0.5 0.6 0.7 0.8

5

10

15

20

25

30

35

scale of encoding patterns

s

izeofmannerverbclass

Polish Latvian

Swedish

FrenchHindi Persian

Italian

GreekRussian

t ua n a n

AlbanianArmenian

r sArmenian

Greek

r s

GreekRussian

t u an an


35/41

ArmenianGreek Albanian

Greek

Albanian

ArmenianGreek

RussianLithuanian

EnglishDutchGerman

PortugueseFrenchIrish

Albanian

ArmenianGreek

RussianLithuanian

English

DutchGerman

PortugueseFrench

Irish

AlbanianArmenian

Greek

RussianLithuanian

English

utcGerman

PortugueseFrench

Irish

Albanian

Armenian

Greek

RussianLithuanian

EnglishDutcherman

PortugueseFrench

Irish

AlbanianGreek

ArmenianGreek

Albanian

ArmenianGreek

RussianLithuanian

EnglishDutchGerman

PortugueseFrench

Irish

Albanian

ArmenianGreek

RussianLithua

EnglishDutchGerman

PortugueFrenchIrish

Albanian

ArmenianGreek

RussianLithuanian

EnglishDutcherman

PortugueseFrench

Irish

AlbaArmenian

Gree

RussianLithuania

English

utcGerman

PortuguFrench

Irish

or ugueseFrench Dutch

ermanDutcherman or ugues

French

n. 346n. 345 n. 347 n. 348

n. 619 n. 621n. 620 n. 622

n. 887n. 886 n. 889n. 888

Phylogenetic Generalized Least Squares


36/41

CORRELATED EVOLUTIONp = 0.05

probability of slope

frequency

0.0 0.2 0.4 0.6 0.8 1.0

0

50

150

250


37/41

CONCLUSION

An approach to motion events that takes into accountpatterns of usage gives us a more fine-grained and productive

perspective

Patterns of motion encoding diversity are not random buthistorically patterned, and comparative analysis needs to take

this into account

In order to take into account this history we need ways tocombine traditional historical linguistic methods withphylogenetic comparative methods


38/41

Thank you!


39/41

GEOGRAPHICAL DISTANCE

Mantel test (Spearman correlation):

Mantel coefficient 0.095

Two-tailed p-value: 0.369


40/41

GEOGRAPHICAL DISTANCE

1.5 2.0 2.5 3.0 3.5 4.0 4.5

0.

0

0.

1

0.

2

0.

3

0.

4

0.

5

log.lang.dists

data.

dists


41/41

CORRELATED EVOLUTION

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 2.73756 0.31165 8.7841 1.041e-05 ***log(encoding) -1.01505 0.36695 -2.76620.02189 *

---

Signif. codes: 0 "***#0.001 "**#0.01 "*#0.05 ".#0.1 "#1

Residual standard error: 0.01612 on 9 degrees of freedom

Multiple R-squared: 0.4595,Adjusted R-squared: 0.3995

F-statistic: 7.652 on 2 and 9 DF, p-value: 0.01144

Evolutionary Change in Motion Event Encoding in Indo-European

Documents

Transcript of Evolutionary Change in Motion Event Encoding in Indo-European