Goldin_Meadow_et_al_2008(2)

8/2/2019 Goldin_Meadow_et_al_2008(2)

1/6

The natural order of events: How speakers ofdifferent languages represent events nonverballySusan Goldin-Meadow*, Wing Chee So, Asl Ozyurek, and Carolyn Mylander** Department of Psychology, University of Chica go , 5730 South Woodlawn Avenue , Ch icag o, IL 60637; Department of Psychology, National University ofSingapore, Republ ic o f Singapore 119077; Department of Linguisti cs, Radboud Universi ty Ni jmegen, 6525 HT Nijmegen , The Netherlands; M ax Pla nckInst i tute for Psycholingui sti cs, 6500 AH Nijmegen, The Netherlands; an d Department of Psychology, Ko c Universit y, 34450 Istanbul, TurkeyEdited by Rochel Gelman, Rutgers, Th e State Universi ty of New Jer sey , Pisca t a w a y, NJ, and approved May 8 , 2008 (received for review November 12 , 2007)

To test whether the language we speak influences our behavior

even when we are not speaking, we asked speakers of four

languages differing in their predominant word orders (English,

Turkish, Spanish, and Chinese) to perform two nonverbal tasks: a

communicative task (describing an event by using gesture without

speech) and a noncommunicative task (reconstructing an event

with pictures). We found that the word orders speakers used in

the ir everyday speech did not influence their nonverbal behavior.

Surprisingly, speakers of all four languages used the same order

and on both nonverbal tasks. This order, actorpat ientact, is

analogous to the subjectobjectverb pattern found in many

languages of the world and, importantly, in newly developing

gestural languages. The findings provide evidence for a naturalorder that we impose on events when describing and reconstruct-

ing them nonverbally and exploit when constructing language

anew.

g esture language genesis sign language word order

Consider a woman twisting a k nob. When we watch this event,we see the elements of the event (woman, twisting, knob)simultaneously. But when we talk about the event, the elementsare mentioned one at a time and, in most languages, in aconsistent order. For example, English, Chinese, and Spanishspeakers typically use the order womantwist knob [actor (Ar)act (A)patient (P)] to describe the event; Turkish speakers use

woman knobtwist (ArPA). The way we represent events in ourlanguage might be such a powerful tool that we naturally extendit to other representational formats. We might, for example,impose our languages ordering pattern on an event when calledon to represent the event in a nonverbal format (e.g., gestures orpictures). Alternatively, the way we represent events in ourlanguage may not be easily mapped onto other formats, leavingother orderings free to emerge.

Word order is one of the earliest properties of languagelearned by children (1) and displays systematic variation acrossthe languages of the world (2, 3), including sign languages (4).Moreover, for many languages, word order does not vary freelyand speakers must use marked forms if they want to avoid usingcanonical word order (5). If the ordering rules of language areeasily mapped onto other, nonverbal representations, then the

order in which speakers routinely produce words for particularelements in an event might be expected to influence the orderin which those elements are represented nonverbally. Conse-quently, speakers of different languages would use differentorderings when asked to represent events in a nonverbal format(the ordering rules of their respective languages). If, however,the ordering rules of language are not easily mapped ontononverbal representations of events, speakers of different lan-guages would be free to use orders that differ from the canonicalorders found in their respective languages; in this event, theorderings they use might, or might not, converge on a singleorder. To explore this question, speakers of four languagesdiffering in their predominant word orders were given twononverbal tasks.

Gesture Task (6). Forty adults [10 English speakers, 10 Turkishspeakers, 10 Spanish speakers, and 10 Chinese (Mandarin)speakers] were asked to describe vignettes displayed on acomputer (some depicting interactions between real objects andpeople and others depicting animated toys) by using only theirhands and not their mouths. The vignettes displayed 36 differentmotion events [see supporting information (SI) Table S1],chosen because events of this type are ones that children t alkabout in the early stages of language learning (1) and thus mayhave a special status with respect to early language. In addition,the events are typically described by using different word ordersby speakers of the languages represented in our sample: (i) 20

events typically described by using intransitive sentences, 7 inwhich an entity performs an action in place (girlwaves) and 13in which an entity transports itself across space (duckmovesto

wheelbarrow); and (ii) 16 events typically described by usingtransitive sentences, 8 in which an entity acts on an objectin place(womantwistsknob) and 8 in which an entity transfers anobject across space (girlgivesflowerto man). To determinethe predominant speech orders speakers of the four languagesuse to describe these particular events, participants were alsoasked to describe the events in speech before describing them ingesture.**

Transparency Task (7). Another 40 adults (10 speakers of each ofthe same four languages) were asked to reconstruct the same

events by using sets of transparent pictures. A black line drawingof the entities in each event (e.g., woman, knob) and a blackcartoon drawing of the action in the event (e.g., an arrowed linerepresenting the twisting motion) were placed on separatetransparencies. Participants were asked to reconstruct the eventby stacking the transparencies one by one onto a peg to form asingle representation (see Fig. 3). Participants were given noindication that the order in which they stacked the transparencies

was the focus of the study; in fact, the background of eachtransparency was clear so that the final product looked the sameindependent of the order in which the transparencies werestacked. The task was designed to test whether speakers wouldextend the ordering patterns of their languages not only to thepictorial modality (where, unlike gestures, all of the elements ofan event are presented simultaneously in the final product), butalso to a noncommunicative situation: the experimenter made it

Author contribution s: S.G.-M . a n d A .O . de signed resea rch; W .C.S. and C.M . performed

re sea rch; S.G.-M . and W .C.S. analyzed data; an d S.G.-M . and A .O . wrote the paper.

The authors de clare no conflict of interest .

Th is article is a PNA S Direct Submissio n.

To whom correspondence should be addressed . E-mail: sgm@uchi cag o .ed u .

**By having participants produce spe ech de scriptions fir st , we recognize that we may be

bi asing them to use their spe ech orders in their gesture descriptions. Th e striking finding

is that they did not .

Th is article contains supporting information onl ine at w w w .pn as.org/cgi/content/full/

0710060105/DCSupplemental.

2008 by The National A cademy of Scie nces of the USA

w w w .pn as.or g cgi do i 10.1073 pn as.0710060105 PN AS July 8, 2008 vo l. 105 no . 27 91639168
http://www.pnas.org/cgi/data/0710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/content/full/0710060105/DCSupplementalhttp://www.pnas.org/cgi/data/0710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/content/full/0710060105/DCSupplemental


2/6

clear that she was busy with another task and not payingattention when the participants stacked the transparencies.

Results

To ensure that speakers of the four languages did indeed usedifferent word orders in speech, we first examined the speechthey produced to describe the vignettes. We focused on theposition of semantic elements traditionally used to characterize

word order in the worlds languages (9, 10): actors (Ar) (whichare typically subjects, S), patients (P) (typically objects, O), acts(A) (typically verbs, V). Speakers of all four languages consis-tently used ArA order when describing intransitive actions bothin place and crossing space (Table 1). However, speakers useddifferent orders to describe transitive events. Following thepatterns of their respective languages, English and Spanishspeakers used ArAP and Turkish speakers used ArPA to de-scribe all transitive actions, both in place and crossing space.

Chinese speakers used ArAP for in-place but ArPA for crossing-space transitive actions.**Participantsused thesame order,ArA, theorder found in each

of their spoken languages, in the gesture strings they producedto describe intransitive actions in place and crossing space (Table1). But participants also used a single gesture order for transi-tive actions, even though their spoken languages used differ-ent orders to convey actions of this type. The predominantgesture order was ArPA, which was identical to the predominantspeech order for in-place and crossing-space actions in Turkishand for crossing-space actions in Chinese, but different from the

predominant speech order for both types of actions in Englishand Spanish and for in-place actions in Chinese. We analyzed

in-place and crossing-space actions separately by using ANOVAswith one within-subjects factor (order) and one between-subjectsfactor (language group). We found significant effects for orderbut not group in each analysis: gesture strings were significantlymore likely to display the ArPA order than the ArAP orderfound in spoken Chinese, English, and Spanish for in-placeactions [F(1,26) 63.18, P 0.00001] and in spoken English andSpanish for crossing-space actions [F(1,16) 49.42, P 0.00001]. Fig. 1 presents examples of ArPA gesture strings.

Fig. 2 presents the proportion of all gesture and speech stringsdescribing transitive actions that were c onsistent with the ArPAorder. We analyzed the data in Fig. 2 by using an ANOVA withone within-subjects factor (modality: gesture vs. speech) and onebetween-subjects factor (language group). We found an effect ofmodality. Proportions were significantly different for gesture vs.

speech [F(1,35) 235.65, P 0.00001] and, as expected, theeffect interacted with language group [ F(3,35) 32.00, P 0.00001]. Gesture was significantly different from speech forEnglish, Spanish, and Chinese (P 0.00004, Scheffe), but not forTurkish. Importantly, there were no significant differences be-tween any pairings of the four language groups for gesture (P0.74). Thus, participants did not display the order found in theirspoken language in their gestures. Instead, the gestures allfollowed the same ArPA order.

A priori we might have guessed that gesturers would begin astring by producing a gesture for the action, as the action framesthe event and establishes the roles that other elements canassume. Indeed, the glass and box gestures in Fig. 1 are verysimilar in form; the fact that these two gestures represent

** According to ref . 31 , Chinese wa s originally an SOV language and becam e SV O; it is

currently in the process of moving back to SOV and thus di splays both orders.

Table 1. Speech and gesture strings produced by Turkish, English, Spanish, and Chinese speakers categorized according to their fit s

to predominant orders

Types o f act ions d escribed

Sp ee ch strings G esture strings

Predominant

sp ee ch order

Proport ion

co nsistent with

sp e ech orderPredominant

g esture order

Proport ion

co nsistent with

g esture order

Proport ion

con sistent with

sp ee ch order

M e a n SE M e an SE M ea n SE

Ac

to rs,

ac

ts

(intrans

itive)In-place a n d cro ssing-sp ace act ions

Turkish speakers Ar A 1.00 (0.00) A r A 0.85 (0.10)

Chinese speakers Ar A 1.00 (0.00) A r A 0.98 (0.01)

English speakers Ar A 1.00 (0.00) A r A 0.99 (0.01)

Spanish speakers Ar A 0.94 (0.03) A r A 0.97 (0.02)

Acto rs, pat ients, acts (transit ive)

In-place act ions

Turkish speakers Ar P A 0.97 (0.02) A r P A 1.00 (0.00)

Chinese speakers Ar A P 0.88 (0.04) A r P A 0.84 (0.06) 0.30 (0.09)*

English speakers Ar A P 0.98 (0.01) A r P A 0.90 (0.03) 0.20 (0.06)*

Spanish speakers Ar A P 0.92 (0.05) A r P A 0.86 (0.05) 0.34 (0.06)*

Crossing-sp ace act ions

Turkish speakers Ar P A 0.93 (0.04) A r P A 0.69 (0.13)

Chinese speakers Ar P A 0.80 (0.06) A r P A 0.90 (0.08)

English speakers Ar A P 0.

88 (0.

10) A r P A 0.

78 (0.

08) 0.

21 (0.

07)*Spanish speakers Ar A P 0.86 (0.05) A r P A 0.87 (0.04) 0.13 (0.07)*

G estures were produced in place of sp ee ch a n d t h us w e r e n o t accompanied by any sp ee ch at al l. *, P 0.0001, comparing proportion of gesture strings

con sistent with gesture order vs. spe ech order.Proportions were cal culated by taking the number of sp o k e n se n t e nces a part icipant produced that were con sistent with the predominant sp ee ch order and

dividing that number by the t otal number of sp o k e n se n t e nces the part icipant produce d t o d escr ibe the target event .Proportions were cal culated by taking the number of gesture strings a part icipant produced that were co nsistent with the predominant gesture order or the

predominant spe ech order and dividing that number by the total number of gesture strings the part icipant produced to descr ibe the target event. Participants

did not always produce g estures for all three eleme nts when describing transitive act ions (se e Table S1). When ArA strings were produced for a transitive act ion,

w e co u n t e d t h ose strings as con sistent with the predominant order for gesture and spe ech and th us in cluded them in the numerator for both proportions. A blank cell indica t es that the predominant gesture order is identical to the predominant spe ech order for that language group .

9164 w w w .pn as.or g cgi do i 10.1073 pn as.0710060105 Goldin-M ea do w et al .
http://www.pnas.org/cgi/data/710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1


3/6

different objects becomes apparent only after the action ges-tures, tilt-to-mouth and cover, are produced. Despite the factthat an action gesture is often needed to disambiguate an objectgesture, participants in all four language groups produced ges-tures for objects (Ar and P) before producing gestures for

actions. In this regard, it is worth noting that the ArPA orderingpattern we have found is not inevitable in the manual modality.In many conventional sign languages, including American SignLanguage (8), the canonical underlying order is SVO-ArAP inour terms; thus patients do not necessarily appear before acts inall communications in the manual modality.

We turn next to the noncommunicative transparency t askcarried out by a different set of 40 speakers. There was no needto select transparencies in a consistent order because the back-grounds were transparent and the final products looked thesameindependent of the order in which they were stacked. Never-theless, participants followed a consistent order when selectingtransparencies, and the order was the same across all languagegroups (Fig. 3). We conducted separate analyses for intransitive

and transitive events by using ANOVAs with one within-subjectsfactor (order) and one between-subjects factor (languagegroup). We found an effect of order but not group in eachanalysis. Participants stacked transparencies in the ArA ordersignificantly more often than the AAr order for intransitiveactions in place and crossing space [ F(1,36) 203.02, P 0.000001], and in the ArPA order significantly more often thanthe other possible orders for transitive actions in place[F(4,144) 85.01, P 0.000001] and for transitive actionscrossing space [F(5,180) 185.71, P 0.000001].

Here again we found that participants did not display theorders of their spoken languages. Theorder in which participantsselected transparencies was the same across all four languagegroups, and this order was identical to the predominant gestureorder in the gesture task. Across all event types, 81% (SE 6%)of 1,423 transparency trials followed the ArPA pattern, as did90% (SE 3%) of 614 gesture strings.

Recall that half of the vignettes depicted crossing-spaceactions and thus contained endpoints (e.g., the scaffolding in Fig.3). Here, too, the language that participants spoke did notinevitably determine their orderings on the gesture or transpar-ency tasks. When describing the vignettes in speech, participantsfollowed the patterns of their languages: Turkish speakers

produced words for endpoints before words for actions in 97%of their spoken utterances; Chinese, Spanish, and English speak-ers produced words for endpoints after words for actions in 88%,94%, and 100% of their spoken utterances, respectively. How-ever, participants in all four groups performed similarly on thegesture and transparency tasks. Participants tended to placegestures for endpoints at either the beginning (Turkish 25%,Chinese 40%, Spanish 26%, and English 37%) or end (58%,48%, 56%, and 48%, respectively) of their three-gesture strings

There were no ordering differences in vignettes portraying real vs. toy acto rs. In addition ,

acto rs pre ce d e d acts w h e t h e r t h e acto rs were animate or inanimate , and preceded

patients whether the patients were animate or inanimate , suggest ing that participants

orders were based on the semantic roles, not animacy, of the enti t ies involved.

BOY GLASS TILTS TO MOUTH

CAPTAIN PAIL SWINGS

CAPTAIN PAIL SWINGS

GIRL BOX COVERS

Fig. 1. Examples o f A rPA g esture strings produced by speakers of al l four

languages. (To p) The pictures s h o w a Spanish speaker descr ibing the boy tilts

gl ass vignette. (Middle ) The pictures show an Engl ish speaker and a Turkish

sp e a k e r d escr ibing the ca pta in sw ings pail vignette; n o t e t h a t t h e E n g l ish

speaker (Uppe r Middle ) conveys th e captain by produci n g a g esture for his

cap , the Turkish speaker (Low e r Middle ) by pointing at the still picture of the

captain. (B ottom) T h e p ictures sh o w a C h i nese sp e a k e r d escr ibing the girl

covers bo x vignette.

Speech strings Gesture strings

0.0

0.2

0.4

0.6

0.8

1.0

Turkish speakers

Chinese speakers

English speakers

Spanish speakers

ProportionofstringsconsistentwithArPAorder

Fig. 2. Proportion of spe ech (Lef t) and gesture (Right ) strings produced by

speakers of Turkish, Chinese, English, a nd Spanish t o d escr ibe transitive ac-

t ions that were co nsistent with the ArPA order. Included are bot h in-place and

cro ssing-sp ace transitive act ions.

Goldin-Meadow et al . PN AS July 8, 2008 vo l. 105 no . 27 9165


4/6

containing endpoints and select the endpoint transparencyprimarily at the beginning (78%, 69%, 71%, and 83%, respec-tively) but also at the end (10%, 25%, 20%, and 12%, respec-tively) of the transparency stacks containing endpoints; the fewremaining endpoints were placed in the middle of a gesturestring or transparency stack. The interesting generalization isthat gestures and transparency selections tended to be positionedoutside of the semantic core: only 15% (SE 5%) of gesturesand 8% (SE 2%) of transparencies for endpoints were placedbetween gestures or transparencies for Ar, P, and A, reinforcingthe notion that events are built around these semantic elements,not endpoints. It is intriguing that languages also seem toprivilege Ar, P, and A, typically encoding them as S, O, and Vand relegating endpoints to the linguistic periphery as, forexample, indirect (as opposed to direct) objects or objects of apreposition. This pattern may reflect another cognitive prefer-ence that languages co-opt and build on.

To summarize both the gesture and transparency tasks, wenote four striking findings: (i) Participants adhered to a consis-tent ordering even though consistency was not demanded byeither task. (ii) The ordering was the same across participants,both within and across language groups. ( ii i) The ordering wasthe same across tasks, both communicative (gesture) and non-communicative (transparency). (iv) The ordering was not nec-

essarily the same as the ordering in the participants spokenlanguage.

Discussion

What might account for the particular ordering we observedacross language groups and tasks? On the basis of the gesturetask alone, we might hypothesize that the participants arrive atthe ArPA order because of communicative pressure; the ArPAorder might, for example, be particularly easy for listeners todecode. But data from the transparency task weaken thisargument. Actors and patients also preceded acts in the trans-parency task, even though the goal was not to communicate (andthe medium was not gesture).

We therefore speculate that, rather than being an outgrowthof communicative efficiency or the manual modality, ArPA mayreflect a natural sequencing for representing events. Entities arecognitively more basic and less relational than actions (9), whichmight lead participants to highlight entities involved in an actionbefore focusing on the action itself, thus situating Ar and P

before A. Moreover, there is a particularly close cognitive tiebetween objects and actious (10), which would link P to A,resulting in an ArPA order.

The ArPA order found in our participants gesture strings andtransparency selections is analogous to SOV word order inspoken and signed languages. In principle, all six possibleorderings of S, O, and V should be found equally often in thelanguages of the world. However, two orders predominate; theorders are about equally frequent and together account for90% of the worlds languages. SOV is one of those two orders(SVO is the other) (11, 12). In addition, although direction ofchange is difficult to assess over historical time, SOV has beenhypothesized to predominate in the early stages of spoken (13,14) and signed (15) languages. Even more relevant to our study,

If we look only at two-ge sture strings, wh ich , of cou rse, do not allow a middle re spo nse,

we find t hat Turkish , Chinese, an d Spanish participants produced gestures for endpoints

about equ ally often before (51%, 56% , and 32% , re spe ctively) and aft er (49%, 44% , an d

68 % , re spe ctively) gestures for actions. English speakers pl aced endpoints before act ions

in 94% of their relevant t wo-gesture strings. Note that these ge sture patterns do not

conform to the typical pattern in spe ech for any of the group s: English speakers tend to

pla ce endpoints after actions, as do Chinese and Spanish speakers; Turkish speakers tend

to place them before actions.

ArA

AAr .

0

5

10

15

20

IntransitiveActions

MeanNumberofTimesTransparencies

WereSelectedinEachOrder

Predominant

Order: Ar A

ArPA

PArA

ArAP

PAAr

APAr

AArP

0

2

4

6

8

Transitive ActionsIn Place



Predominant Order:

Ar P A

ArPA

PArA

ArAP

PAAr

APAr

AArP

0

2

4

6

8

Turkish speakers

Chinese speakers

English speakers

Spanish speakers

Transitive ActionsCrossing Space



Predominant Order:Ar P A

Transparencies when stackedTransparencies in the arrangement participants saw

Fig. 3. Performance on the transparency t ask. (Uppe r) Di splayed is t he set of transparencie s for the ma n carries chick e n tosca ffoldingvignette (a transitive

act ion cro ssin g spa ce) , with transparencie s as participants s aw them (Lef t) and transparencie s after being st acked (Right ). (Low e r) The graphs di splay the mean

number of t imes speakers of Turkish , Chinese, English, a nd Spanish sel ected transparencie s in each of the possible orders for acto rs and acts in intransitive act ions

(Lef t) and act or s, acts, and patients in transitive actions in place (Center) and cro ssin g sp ace (Right ).



5/6

SOV is the order currently emerging in a language createdspontaneously without any apparent external inf luence. Al-Sayyid Bedouin Sign Language arose within the last 70 years inan isolated community with a high incidence of profoundprelingual deafness. In the space of one generation, the languageassumed grammatical structure, including SOV order (16). Inaddition, deaf children whose profound hearing losses preventthem from acquiring a spoken language and whose hearingparents have not exposed them to a conventional sign language

invent their own gestures to communicate, and those gesturesdisplay a consistent OV order in both American (17) andChinese (18) deaf children [the deaf children typically omitgestures for transitive actors, the S (10)].

If SOV is such a natural order for humans, why then arent allhuman languages SOV? Languages respond to a variety ofpressures, for example, the need to be semantically clear, pro-cessed efficiently, or rhetorically interesting (19). We suggestthat, initially, a developing language co-opts the ArPA orderused in nonverbal representations and uses it as a defaultpattern, thus displaying SOV order, which may have the virtueof semantic clarity. But as a language community grows (20) andits functions become more complex (21), additional pressuresmay exert their influence on language form, in some casespushing the linguistic order away from the semantically clear

ArPA (SOV) order.Our findings bear on the Sapir-Whorf hypothesis (22, 23), the

hypothesis that the language we speak can affect the way weinterpret the world even when we are not speaking. Thishypothesis has been tested with respect to a variety of linguisticconstructions with mixed results (2429), but has not been tested

with respect to word order. Our data suggest that the orderingwe use when representing events in a nonverbal format is nothighly susceptible to languages influence. Rather, there appearsto be a natural order that humans (regardless of the languagethey speak) use when asked to represent events nonverbally.Indeed, the influence may well go in the other direction; theordering seen in our nonverbal tasks may shape language in itsemerging stages.

In sum, we have shown that speakers of languages that differin their predominant word orders do not invoke these orders

when asked to describe or reconstr uct events without speaking.Thus, the ordering found in a speakers habitual talk does notinevitably influence that speakers nonverbal behavior. More-over, the ordering found in nonverbal tasks appears to be morerobust than the ordering found in language; speakers of fourdifferent languages used different orders in their spoken sen-tences, yet all displayed the same order on two different non-

verbal t asks. This order is the one found in the earliest stages ofnewly evolving gestural languages and thus may ref lect a naturaldisposition that humans exploit not only when asked to representevents nonverbally, but also when creating language anew.

Materials and Methods

D a t a w e r e collect e d i n Istanbul for Turkish speakers, in Bei j ing for Chinese

speakers, in Chicago for English speakers, and in Madrid for Spanish speakers.

Participants were drawn from urban universities in each city for both gesture

and transparency t asks. Ea ch participant was initially given co nsent forms a n d

a language interview . Only monol ingual speakers were included in the study.

None of the part icipants w as conversant in a conventional sign language.

Participants i n b o t h t asks were sh o w n 3 6 v i g n et t es o n a computer . A

q u a r t e r o f t h e v i g n e t t es depict e d i n t e ract ions b e t w e e n r e a l p e o p l e a n d

objects; the remaining vignettes were animations involving toys representing

objects a n d p e o p l e. Table S1 lists the events depicted in the vignettes, t he

types of objects pl aying the acto r, patient, or endpoint roles in each vignette ,

and whe ther or not the vignette depicted real objects and people or toys.

Gesture Task. Protocol. The participant and experimenter were po sitioned in a

natural conversational grouping with a laptop computer between them . Th e

entire sessi o n w as videotaped, and the camera was po si t ioned so that i t had

a g o o d v i e w o f b o t h t h e p a r t ici p a n t a n d t h e computer screen. Participants

were told that they would se e a series o f short v ideotaped vignettes a nd , after

e ach, were to tel l the experimenter what happene d in the vignette. For the

part icipants to get a better at look at the entitie s pictured in the vignettes, still

pi ctures of the ini t ial scene of the event including al l of the e nti t ies involved

in the event (act or s, patients, endpoints) were provided for ea ch vignette;

participants occasi o n a ll y r e f e rr e d t o t h e p ictures w h e n d escr ibing the vi-

g n e t t es. Two practi ce vignettes were run before the participant began the se t

of 36 vignettes. A f t e r d escr ibing al l of the vignettes in sp ee ch, participants

were told that they would se e t h e same vignettes again but , th is t ime, they

were to tel l the experimenter what happened u sing only their hands and not

spe ech. Th e stil l pictures were sh o w n d u r i n g t h e g esture alone description s

and part icipants occasional ly pointed at the pict u r e t o r e f e r t o o n e o f t h e

elements (see Fig . 2).

Codingandanalysis. Sp ee ch w as transcr ibed and coded by native speakers of

e ach language . Ge sture was de scribed in terms of hand shape , palm orienta-

t ion, motion, pl acement (e.g., neutral sp ace a t che st level, on the body, near

an object) , articulator (e.g., r ight hand , left hand, head), a nd size of motion

(e .g., 2 in ch es, 25 i nche s, 5 in ch es). G estures were divided into strings by

using motoric criteria; str ing breaks were co d e d w h e n p a r ticipants relaxed

their hands b e t w e e n g estures (30). Ge stures were cla ssified as either pointing

or ico nic ge stures and given a meaning gloss. Pointing gestures were rarely

use d b u t w h e n t h e y w e r e, they ei ther indica t e d a n e n t i t y o n t h e computer

screen or the still pictures (e .g., point at the man on screen, gl ossed as ma n) o r

an enti ty in the room that wa s simi lar to or had an interpretable relat ion to

one in the vignett e (e .g., point at self whe n referring to the girl in the vignet te ,

gl osse d as girl; point behind self when referr ing to the chick e n w h o stood

behind the captain in the vignette , gl osse d as chicke n). Icon ic g estures werepantomimes used to represent ei ther the action in the vignett e (e.g ., t w o fists

m o v e d a w a y f r o m t h e che st in a l ine , gl ossed as push in the man push es

garbage can to motorcycle man vignet te) or an entity (e.g., t w o fists held at

che st level rotated as although revving a motorcycle , gl ossed as motorcycle

ma n in this same vignette). Other types o f g estures (e .g., beats, n od s) were

transcr ibed but did not represent vignette act ions or enti t ies a n d t h us were

not included in the analyses. Reliability for gesture coding ranged between

87% and 91% agreement between coders depending on the category.

We analyzed only those strings that descr ibed the target event , i.e ., strings

co n t a i n i n g a w o r d o r g est u r e f o r t h e act ion displayed in the vignette . In

addit ion to containing a word or ge st u r e f o r t h e act ion, t he str ing had to

in clude at least one other word or gesture representing an e nti ty playing an

actor or patient role in the act ion. For example, a cap ge sture referring to the

captain (actor) in the ca pta insw ings pailvignette produced in se q u e nce with

a sw ing ge sture (act ion) would be included in the analyses; however , either

g esture produce d o n i ts o w n w o u l d n o t.

Moreover,

if a string contained ag esture for the captain (actor) and one for t he pai l (patient), but no gesture

for the swinging mot ion (act) , t he string would not be included in the analyses

simply beca use, without the sw ingg esture, w e could not be certain that the

part icipant meant to be referr ing to the captain and t he pai l in their roles as

actor and patient, re sp ectively. When describing events in spe ech, participants

typical ly mentioned al l of the relevant elements in the event: two elements

(act or , act) in intransitive act ions and three (act or , act , patient) in transitive

act ions. When describing events in gesture alone, participants often omitted

elements; they produce d 5 0 1 g esture strings containing two relevant ele-

ments and 113 containing three , all of which were included in the analyses. W e

cla ssi fi e d strings t h a t d i d n o t h a v e g estures f o r t h e f u l l co m p l e m e n t o f

elements according to the elements that were prese n t i n g esture. Th us, f or

g esture strings conveying transitive act ions, w e cla ssified three types of strings

as co nsistent with an ArPA pattern (ArA , PA , A rPA) and seven as in co nsistent

( A A r, AP, A rA P, A Ar P, APA r, PAr A , PAAr). In addit ion, w e conduct e d separate

analyses o f g esture strings containing endpoints. Participants produced 97

strings containing gestures for the endpoint and action and 121 containing

g estures for the endpoint, act ion, and one or two other element s (actor and/or

patient).

Occasionally participants produce d strings in spe ech or gesture describing

a n act ion that did not match t h e i n t e n d e d act ion displayed in the vignette

(e .g., the garbage can moved to the motorcycle man rather than the man

p ushed the garbage can to the motorcycle man). In t h ose cases, we recla s-

si fi e d t h e str ing to fit the part icipants interpretat ion and analyzed i t along

with strings conveying the act ion i t described (in this case, as an intransitive

act ion rather than a transitive action). Degrees of freedom vary for stat istical

analyses o f some of the events b eca use participants who produced the recla s-

si fi e d strings did not then have a dat a point for the originally inten ded act ion.

If r ecla ssi fi e d strings are omitted from the analyses, the resul ts are unch a n g e d.

Proportions were su b m i tt e d t o a n a rcsine transformation before stat isti cal

analysis.

Goldin-Meadow et al . PN AS July 8, 2008 vo l. 105 no . 27 9167
http://www.pnas.org/cgi/data/0710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0710060105/DCSupplemental/Supplemental_PDF#nameddest=ST1


6/6

Transparency Task:Protocol andCoding. Participants were told that they would

b e w a tching a series of video cli ps and that , after each vignette, they were to

re con str uct the event by puttin g a set of transparencie s on a peg one at a time .

They were told that i f they needed to see the vignette again, they could click

on the repeat button on the computer . After two practi ce trials (the experi-

menter did not model the task for the part icipant), the experimenter played

the first v ignette on the computer and then placed the transparencie s for that

vignette on the table in a tr iangular configuration , beginning at the part ici-

pants right and ending at his or her left. For example, the four transparencie s

pr ese n t e d t o t h e p a r t ici p a n t f o r t h e ma n carries chick e n t o sca ffolding

vignette were placed on the table in the fol lowing order (Fig . 3): 1) arrow

denoting the trajectory of the moving a ct ion, 2) chicke n, 3) scaffolding, an d

4) man. Transparencie s were laid down in the same order for each participant ,

and orders were randomized over the 36 vignettes.

Participants were discouraged from talking during the study. To encourage

t h e m t o t r e a t t h e t a sk as no ncommunicative , the experimenter occupied

h er se l f w i t h a n o t h e r t ask a n d d i d n o t p a y a t t e n t io n as participants pi ck ed

up the transparencie s. When part icipants fi ni sh ed st acking the transparencie s

o n t h e p e g , t h e y a l e r te d t h e e x p e r im e n t e r w h o t h e n collect e d t h e m a n d

started the next v ignette on the computer .

Transcribers w at ch e d t h e v i de o t a p es o f e ach sessi o n a n d r eco r d e d t h e

order in which the part icipant placed the transparencie s on the peg for each

vignette.

ACKNOWLEDGMENTS. W e t h a n k S. Beilock , L. Gleitman, P. Hagoort , P. Inde-

frey,

E.

N e w p o r t,

H.

N usbaum,

S.

Levine,

and T.

Regier for comments and A.

S.

Acu na , T. Asan o, O . E. Demir, T. Go ksu n, Z. Joh nson , W . X ue , and E. Yalabik forhelp with data collect ion and coding. Th is re sea rch w as supported by NationalInst i tute on Deafness a n d O t h e r C o m m u nication Disorders Grant R01 DC00491(t o S.G.-M .).

1. Brown R (1973) A First Lan g u ag e(Harvard Univ Pre ss, Cambridge , M A) .

2. Greenberg JH (1963) Some universal s of grammar with particular reference t o t h e

order of meaningful elements. Universals of Lan g u ag e, ed Greenberg JH (M IT Pre ss,

Cambridge , M A) , pp 73113 .

3. Ha spelmath M , Dryer MS, Gil D , Comrie B (2005) T h e W o r l d A t l a s of Lan g u ag e

Str uctures (Oxford Univ Pre ss, Oxford).

4. Perniss P, Pfa u R, Steinbach M (2007) Can t you see the d i fference? Sou rces of variation

in sign language str ucture . Visible Variation:Cross-LinguisticStu d i esonSig nLan g u ag e

Str ucture, ed s Perniss P, Pfa u R, Steinbach M ( Mouton , Berlin), pp 134.

5. Steele S (1978) Word order variatio n: A typological study. Universals o f H u m a n

Lan g u ag e, ed s Greenberg JH, Ferguson , CA , Moravesi k E A (Stanford Univ Pre ss,

Stanford, CA ), pp 585 623 .

6. Goldin-Meadow S, McNeill D, Singleton J (1996) Si lence is l iberating : Removing the

handcuf fs on grammatical expression in the manual modal i ty. Psych Rev 103:34 55 .

7. Ge rshkoff-St o w e L, Goldin-Meadow S (2002) Is there a natural order for expressin g

semantic relation s? Cognit Psycho l 45:375412 .

8. Liddell S (1980) American Sign LanguageSynt ax(Mouton , The Hague, The Netherlands).

9. Gentner D, Boroditsky L (2001) Individuation , relativi ty and ear ly word learning .

Lan g u ag e Acqu isi t i o n an d Co n cep tu al Devel o p m en t, ed s Bowerman M , Levinson SC

(Cambridge Univ Pre ss, New York), pp 215256 .

10. Goldin-Meadow S (2003) Resi l ience of Lan g u ag e(Psychology Pre ss, New York).

11. Baker MC (2001) Th e Ato m s of Lan g u ag e(Ba sic Books, New York).

12. Dryer M ( 2005) Order of subject , object a n d v e r b . Th e Wo rl d Atl as of Lan g u ag e

Str uctures, ed s Ha spelmath M , Dryer MS, Gil D , Comrie B (Oxford Univ Pre ss, Oxford),

pp 330 333 .

13. Newmeyer FJ (2000) On the recon str uction of proto-world word order . The Evolu-

t i o n ary Em erg en ce of Lan g u ag e, ed s Knight C, Studdert-Kennedy M , Hurford JR

(Cambridge Univ Pre ss, New York), pp 372388 .

14. Givon T (1979) On Un d er stan d i n g Gram m ar(A cademic, New York).

15. Fishe r S (1975) In fl u e nces on word order change in American Sig n Language. Wo rd

Ord er an d Wo rd Ord er Ch an g e,

ed Li CN (University of Texas,

Au stin),

pp 125.

16. Sandler W , Meir I, Padden C, Aronof f M (2005) The emerge nce of grammar: Systematic

str ucture in a new language . Proc Natl Acad Sci USA 102 :26612665.

17. Goldin-Meadow S, Feldman H (1977) The development of language-like communica-

tion without a language model . Scie nce 197:401403 .

18. Goldin-Meadow S, Mylander C (1998) Spontaneous sig n syste ms cr e a t e d b y d e a f

children in two cultures. Natu re391:279 281 .

19. Slobin DI (1977) Language change in childhood and history. Lan g u ag eLearn i n g an d

Th o u g h t, e d M acnamara J (A cademic, New York), pp 185214 .

20. Senghas A , Kita S, Ozyurek A (2004) Children creating core properties of language:

Evidence from an emerging sign language in Nicaragua. Scie nce 305:1779 1782.

21. Sa n k o f f G, Laberge S (1973) On the acqu isi t ion of native speakers by a language .

Kivung6:32 47.

22. Sapir E (1921) Lan g u ag e(Harcourt , Bra ce, & World , New York).

23. Whorf BL (1956) Lan g u ag e, Th o u g h t, an d Reality: Sel ected Wri ti n g s of BenjaminLee

Wh o rf , ed Carroll JB ( MIT Pre ss, Cambridge , M A) .

24. Gumperz JJ, Levinson , SC, ed s (1996) Reth i n ki n g Linguistic Relativity(Cambridge Univ

Pre ss, New York).

25. Pederson E, et al . (1998) Semantic typology and spatial con ceptual izatio. Lan g u ag e

74:557-589.

26. Gentner D , Goldin-Meadow, S, ed s (2003) Lan g u ag e i n Mi n d : Ad van ces i n th eStudy of

Lan g u ag e an d Th o u g h t(M IT Pre ss, Cambridge , M A) .

27. Lucy J (1992) Gram m ati cal Categ o ri es an d Co g n i tio n (Cambridge Univ Pre ss, Ne w

York).

28. Papafragou A , Ma ssey C , Gleitman L (2002) Shake , rattle, n roll: The representation of

motion in language and cognition . Cognition 84:189 219 .

29. Gilbert A , Regier T, Ka y P, Ivry R (2006) Whorf hypot hesis is supported in th e right visua l

field but not the left . Proc Natl Acad Sci USA 103 :489 494 .

30 . Goldin-Meadow S, Mylander C ( 1984) Gestural communica t i o n i n d e a f children .

Mo n o g rSoc Res Child Dev 49:1151 .

31 Li CN , Thompson SA (1981) Man d ari n Ch i n ese: A Funct i o n al Referen ce Gram m ar

(University of California Pre ss,

Berkeley).


Goldin_Meadow_et_al_2008(2)

Documents

Transcript of Goldin_Meadow_et_al_2008(2)