OXFORD LIBRARY OF PSYCHOLOGY

Editor-in-Chief PETER E. NATHAN

The Oxford Handbookof Deaf Studies,Language, andEducationVolume 2

Edited by

Marc Marschark

Patricia Elizabeth Spencer

OXFORDUNIVERSITY PRESS

2010

Early Language Acquisition andCHAPTER

19Rachel I. Mayberry

Adult Language Ability: What SignLanguage Reveals About the CriticalPeriod for Language

Abstract

This chapter examines the critical period for language through the prism of deafness.The first topic isthe concept of critical periods, followed by a summary of research investigating age of acquisitioneffects on the outcome of second-language learning (L2).The phenomenon of late first-language (LI)acquisition among deaf children is then described.The focus of this chapter is on a series of studies thatcompare and contrast the long-range outcome of LI and L2 acquisition in relation to age of acquisition.The effects of late LI acquisition are greater than those for L2 learning. The effects include acompromised ability to process and understand all forms of language. Late LI acquisition hasdeleterious effects on the ability to learn other languages and on reading development.The findingscome from experiments in American Sign Language (ASL) and English using a variety of psycholinguisticparadigms across levels of linguistic structure and include narrative comprehension and shadowing,sentence shadowing and memory, grammatical judgment, and reading comprehension. How thesepsycholinguistic phenomena illuminate the critical period for language is then discussed.Keywords: critical period, sign language, ASL, language acquisition, second-language acquisition, reading,development, psycholinguistics, comprehension, phonology, grammar, education, early intervention,WildBoy of Aveyron, Itard,

In 1806, Jean Itard wrote that adolescence was toolate to learn language. He concluded this in a reportto the government describing his 2-year attempt toteach the famous Wild Boy of Aveyron to speakFrench. hard called the boy Victor because the onlysound to which he responded was the French [l(Itard, 1896/1962:29). Eventually Victor couldcommunicate his desires to a limited degree withactions on objects, as in holding his bowl next to theporridge pot when he wanted more, but he neverlearned French despite intense and structured lessons (ltard, 1896/1962). The idea that languageproficiency is linked to the age when the learningbegins is known as the critical period frr languageand, as Itard’s report demonstrates, is an old idea.Because ltard was the physician For the InstitutionNationaje de Sourdes-Muets, the world’s first school

for deafchildren and the birthplace of one of world’slargest sign language families (Zeshan, 2006), it iscurious that he never tried to teach Victor sign language (Lane, 1976). ‘The school was 40 years old bythen, and the historical record shows that FrenchSign Language was in full bloom (Van Cleve &Crouch, 1989). The idea that sign language islanguage was unheard of in the 1 800s, however, andresearch uncovering its linguistic architecture wouldnot begin for another hundred years (Klima & Bellugi,1979; Stokoe, Casterline, & Croneberg, 1965).

Could Victor have learned French Sign Language,even though he failed to learn spoken French? Thequestion is whether there is a critical period for language acquisition that encompasses sign language.Because professionals who work with deaf childrenoften appear to implicitly assume that there is a

281

critical period for spoken but not sign language,the question is central to clinical and educationalpractice. The question is also flindamental to language theory because it requires that we understandthe complex relations among language acquisition,brain maturation, and sensory-motor modality.Germane to our discussion is research that investigates age of acquisition effects on the outcome oflanguage acquisition, both spoken language andsign language. The linguistic details of these studiesare outside the scope of this chapter, but discussingthe research in broad strokes allows us to see thelarger picture of how age of acquisition affectsthe outcome of language acquisition. To foreshadowthe conclusion, Victor would not have acquiredFrench Sign Language even though he developedrudimentary, nonvocal communication using objects.The goal of this chapter is to examine the reasonswhy. We begin by considering what a critical periodis with respect to language.

Critical Period LearningA critical period is a phase during developmentwhen learning is most efficient. Critical period phenomena were first observed and documented foranimal behavior at a number of levels, for examplein chicks’ identification and attachment to motherbirds (Lorenz, 1965) or in cats’ visual perceptualdevelopment (Wiesel, 1982). Although critical periods are sometimes characterized as being limits onlearning, they represent a unique form of learningin which an interaction between biology and environment produces quick and unconscious learning.Some researchers prefer to replace the term criticalwith adjectives such as sensitive or optimal to denotethe facts that (a) learning can occur after a criticalperiod and (b) that multiple subskills can be involvedin a complex skill, each with its own temporal sensitivity in development (Werker & Tees, 2005).

Language acquisition is an example of a complexskill consisting of numerous subskills. Although scientists have observed and documented time-limitedsensitivity over development for many types oflearning, they disagree as to what the underlyingmechanisms might be, both in terms of what causesthe heightened sensitivity to a particular type ofenvironmental input and why the sensitivity ends.One class of explanations proposes that environmental complexity directly affects the creation of neuralnetworks (Greenough & Black, 1992), although thisneural effect has not yet been found for language.Another type of explanation based on connectionistmodeling proposes that early learning itself affects

a8z EARLY LANGUAGE ACQUISITION AND

subsequent learning (Seidenberg & Zevin, 2005)especially with respect to second-language (L2) vis ivis first-language (Li) learning, as explained here.

Infants and young children acquire the languagnin their environment quickly and effortlessly B)contrast, most adults appear to struggle, and learrnew languages only with sustained and conscioweffort. The neurosurgeon Wilder Penfield, who firsmapped the language areas of the brain during surgeries designed to limit recurrent seizures, observecthat the brain has strong biological biases fo:language in early life. He noted the overwhelmin1tendency of the anterior left hemisphere to controlanguage—except in cases of injury to the lefhemisphere, when homologous regions in the righhemisphere then take control of language, but onl3if the injury occurs in early childhood (PenfieldRoberts, 1959). This suggests some level of plasticity for language in brain development that is onlpresent in early life.

Age ofAcquisition Effects on SpokenLanguage OutcomeScientific investigation of the critical period folanguage requires that we study individuals whosinitial exposure to a language varies temporally ovethe course of human development. One source oinformation comes from children who experienoatypical social isolation during early childhoodincluding feral children such as Victor, or severel’abused children such as the well-known case of Geniwho lived without human contact for the firs13 years of her life (Curtiss, 1977). Genie was ablto learn some spoken English with intense instruction, but she was unable to maintain her languagin adulthood (Garmon, 1994). The multiple cornplications suffered by these rare cases, such as nutritional and emotional deficiencies, means that wmust use caution in interpreting their difficulty witllanguage acquisition because of these confoundinfactors.

The most common means of investigating agconstraints on the outcome of language acquisitiolhas been to measure the language proficiency opeople who learn a second language at varying ageSeveral studies have found a negative correlatioibetween age of L2 acquisition and eventual L2 proficiency and/or significant differences in languagperformance between native and non-native learncr5. These effects have been found across an arraof linguistic structures in morphology, syntax, anphonology, while mostly investigating English as thsecond language, using a number of measuremeor

ADULT LANGUAGE ABILITY

of spoken and written language, including qualityratings by judges, error counts, memory tasks, andgrammaticality judgments (Birdsong & Molis,2001; Flege, Yeni-Komshian, & Liu, 1999; johnson& Newport, 1989).

Although these findings confirm the widespreadfolk belief that learning a second language at ayounger age leads to higher L2 proficiency, comparedto learning it at older ages, researchers disagree as tohow the trend should be interpreted. At issue iswhether the decline in L2 proficiency in relation toage provides evidence for a critical period for language. The data suggest that there is no terminal agewhen a second language can no longer be learned.After the age of 8, the decline in L2 proficiency inrelation to age continues throughout the lifespaninto senescence (Hakuta, Bialystok, & Wiley, 2003).‘Ihis finding is interpreted to mean that some factorother than a critical period causes the negativecorrelation between age of acquisition and L2 proficiency, such as cognitive aging. Moreover, age ofacquisition is not the sole predictor of L2 proficiency. Many L2 learners attain near-native proficiency despite older ages of L2 acquisition. Factorssuch as the amount of education in the secondlanguage, and the linguistic relationship of the LIand L2 also predict L2 outcome (Birdsong & Molis,2001; Flege, Yeni-Komshian, & Liu, 1999).

When weighing the effects of age of acquisitionon L2 proficiency, it is important to bear in mindthat L2 learning entails, by definition, acquisitionof a first language in early life. If there is a criticalperiod for language acquisition, some researchersargue that it should affect acquisition ofthefirstlanguage, rather than the second, based on evidencefrom rare cases of social isolation in early childhood,in which the outcome of Li acquisition is severelylimited (Eubank & Gregg, 1999). However, thesecase studies do not provide straightforward evidencefor a critical period due to the multiple and severedeprivations suffered by these children, as explainedearlier (Mayberry, 1994).

Age of Acquisition Effects on SignLanguage OutcomeVariation in Age ofAcquisitionofSign LanguageA major contribution of sign language researchto co,mtive science is the discovery that linguisticstructure and processing transcend sensory—motormodality (Klima & Bellugi, 1979; Stokoe, Casterline,& Croneberg, 1965). Like spoken languages, signnguages are structured at the sentence (syntax),

word (morphology), subword (phonology), andsemantic (word and sentence meaning) levels. Unlikegesture or pantomime, but like spoken language,sign language comprehension requires the unpacking of meaning via the multilayered and hierarchicalstructure of language. In other words, the meaningof sign language sentences does not come for freeby simply looking at signs as if they were pictures.Rather, sign language comprehension requires knowledge of its linguistic structure. Linguistic structure iswhat young children acquire so readily in early life,leading some researchers to hypothesize that acquisition of linguistic structure is governed by a criticalperiod (Lenneberg, 1967).

That the signer’s mind must use the linguisticstructure of sign language to understand and produce it has been amply demonstrated in psycholinguistic experiments (for a review see Emmorey,2002). Neurolinguistic research has further shownthat the brain treats sign language like spoken language. Consistent with Penfield’s early observationsof how the brain represents spoken language(Penfield & Roberts, 1959), researchers have sincediscovered that anterior regions of the left hemisphere are responsible for sign language processingin deaf and hearing signers, and not regions whichprocess nonlinguistic visual information (for a reviewsee Corina & Knapp, 2006). Damage to these classic language regions cause aphasia in signers comparable to aphasia in speakers (Poizner, Klima, & Bellugi,1987). Thus, language structure and processing areamodal; the mind and brain treat sign language aslanguage because it is language.

Given what is now known about the nature ofsign language, it follows that children’s acquisitionof it from birth is similar to that of spoken languageacquisition with respect to the timing and contentof linguistic milestones (Anderson & Reilly, 2002;Mayberry & Squires, 2006; Reilly, 2006). Althoughsign and spoken language show parallel developmental trajectories when the acquisition begins atbirth, most deaf signers are first exposed to sign language at ages well after birth. The crucial questionis whether variation in age of acquisition affects signlanguage proficiency in adulthood. If so, then thequestion becomes whether such effects are similarto, or different from, those observed for L2 learningof spoken languages described earlier. Before turning to this body of work, it is important to considerwhy age of sign language exposure is heterogeneousamong deaf signers.

A small percentage (less than 10%) of deaf children are born to deaf parents, and this subset of the

MAYBERRY 283

population is likely to be exposed to sign languagefrom birth. For the remaining 90% of deaf children,sign language acquisition begins at various ages. Nosingle factor underlies this variation, even in cases inwhich the deaf child lacks functional spoken language. For example, the child may not have beenenrolled in school until an older age, especially if thefamily lived in a rural area. A school that used signlanguage may not have been in close proximityto the family, and the parents may have been reluctant to send the child away to school or relocate.Alternatively, the family may have insisted that thechild remain in an oral program, despite a notablelack of functional language, in the hope that thechild would eventually develop spoken language withmore time and instruction, fuachers and administrators often share these beliefs. Inaction in exposing adeaf child without functional language to sign language often reflects the assumption that exposingthe child to sign language is “giving up” on spokenlanguage, even though this is a mistaken idea. Thedeaf child’s growing maturity and inability to function at school or home without language is often thecatalyst for the decision to educate the child in signlanguage.

These varying circumstances create heterogeneityin the age of sign language exposure within the deafpopulation. At the same time, these diverse circumstances of sign language acquisition mean that nounderlying pathology covaries and/or worsens withage of sign language acquisition, aside from humandevelopment with incomplete or sparsely developedlanguage. The cognitive consequences of humandevelopment in the face of limited linguistic interaction with the environment have only begun to beinvestigated (Mayberry, 2002; Schick, de Villiers,de Villiers, & Hoffmeister, 2007).

Another factor in the heterogeneity of sign language acquisition among deaf individuals is whetherit is acquired as a first or second language. Somedeaf individuals learn sign language as a secondlanguage after successful acquisition of spokenlanguage, but for other deaf individuals age ofsign language acquisition is more representative ofLiacquisition. This occurs when deaf individuals beginto learn sign language with little or no functionallanguage. Acquiring sign language at older ages withlittle or no functional language is common amongdeaf children, although no currently available datadescribe the frequency of the phenomenon. Someof this LI language delay is educationally induced,for reasons explained earlier. Indeed, only amongdeaf signers do we find individuals who are otherwise

284 EARLy LANGUAGE ACQUISITION AND

intact and were lovingly cared for as children butnonetheless acquired little or no language in earlychildhood.

Experimental Studies ofAge ofAcquisition EffectsSimilar to the L2 research summarized earlier,several studies have found a negative correlationbetween age of acquisition and sign language proficiency. Among college-aged students, accuracy ofnarrative and sentence recall declines as a linearfunction ofAmerican Sign Language (ASL) acquisition between the ages ofbirth and 15 years (Mayberry& Fischer, 1989). The negative correlation betweenage of ASL acquisition and sentence recall accuracypersists in adults who have had 20 to 40 years ofASLexperience (Mayberry & Lichen, 1991), as shown inFigure 19. 1. Similar results have been obtained usinga battery of ASL tasks (Newport, 1990) and a signmonitoring task (Emmorey, Bellugi, Friederici, &Horn, 1995). These results show that age of acquisition affects language outcome independently ofsensory—motor modality, consistent with what wenow understand to be the amodal nature of linguistic structure and processing.

Another question is whether age of acquisitiondifferentially affects the outcome of LI as comparedto L2 proficiency. One study investigated theracy ofASL sentence recall in deaf L2 and Li learnerswho were matched for years of experience and age ofASL acquisition (Mayberry, 1993). The deaf L2learners were born with normal hearing, which theylost in late childhood due to viral infections; theylearned ASL in immersion settings when they becamedeaf By contrast, the deaf Li learners were firstexposed to ASL at the same ages as the L2 learners,but were deaf from birth and had little or no functional spoken language when they began to acquireASL. The L2 learners recalled the ASL sentences withsignificantly greater accuracy than the delayed Lilearners, as Figure 19.2 shows (Mayberry, 1993).

These findings demonstrate that age of acquisition has far greater effects on the outcome of the LIcompared to the L2 proficiency, suggesting that thescope of the critical period for language acquisitionpertains to the first rather than the second languageA corollary implication is that Li acquisition inearly life facilitates later L2 learning. If early acquisition of spoken language facilitates later acquisitionof sign language, the question is whether the reversesituation is true. Does early learning of a sign language facilitate subsequent acquisition of spokenlanguage?

ADULT LANGUAGE ABILITy

Fig. 19.1 Adjusted meanproportion of rccall responses thatwore grammatical as a function of

ag of acquisition and $gti

presentation rate normal andspeeded1. Figure 2 dom \lasberr,R. I.. ELhen. F. B. 1991c Ilielong-lasting advantage of learning

ogn langaage in cnildhoud.\nothct look at the critloal ponodto r langoago a.quis.tion. Joninal of

.ifesns:.zsdLtguage 30.-+5o--l2. ssith permission of the

publishor.

‘The question was investigated by comparing theEnglish syntactic processing of four groups of adultswith contrasting types of Li experience in earlychildhood (Mavberrv & Lock, 2003). One hearinggroup served as the control and consisted of nativeEnglish speakers. The second hearing group consisted of L2 learners of English who were immersedin it in school when they immigrated to Canada inearly childhood; their native languages were Urdu,Spanish, German, and French. ‘The third group washot n deaf and exposed to ASL from birth; they subsoquently learned English as a second language inschool at the same ages as the hearing L2 learners. Itil important to highlight the fact that two different

types of English L2 learners participated in thestudy: one hearing group whose early first languagewas a spoken language other than English, and onedeaf group whose early first language was sign language. The fourth group was also deaf but began tolearn ASL and English in school at the same age asthe other groups. However, their school enrollmentmarked their first experience with fully perceivablelanguage; that is, they entered school with little orno previously acquired functional language.

Both groups of L2 learners, regardless of whetherthey were hearing or deaf, or whether their first language was spoken or signed, performed at near-nativelevels across the English syntactic structures tested,

Fig. 19.2 The mean proportion of the subjects’total responses that were grammatically acceptableand setnanticallv parallel ro the stimuli for subjects

grouped by age of American Sign Languageacquisition and first versus second language

.Ioquisirion. Figure 4 from Mayberry, R. 1. (1993).i rsc-language acquisition after childhood differs

Itont second-language acquisition: The case of-\Itierican Sign Language. Journal cf_Speech ansi

Ki” Reu,aruh. 16, 1258—12Th, nith permissionof the pttblisher.

0—3 5—8 9—13Native Childhood Late-First

Age of ASL AcquisitionGroup

• Normal Rate—•— Speeded

a

C.)C.)

0aCeC-)a

0)aC)Ca

ato-J0)

1.00

0.90

0.80

0.70

0.60

0.50

0.40

0.30Native 4to6 9to 11

Age of ASL Acquisition

Sentence Level Characteristics of Responses

——--- Grammatical

°- Semantic Relations

1.00

Ca)0.80

0a.Ca)

‘ 0.60

0

00.40

0.20 -I-9—15

Late-Second

‘CIAYBERRY 285

as Figure 19.3 shows (Mayberry & Lock, 2003).These results are consistent with earlier describedstudies investigating the effects of age of L2 acquisition on spoken language, namely that L2 learnerscan sometimes attain near-native levels of L2 proficiencv. ‘The results slims that this is true even whenthe first language is an early-acquired sign language,as was the case in this study. These results also replicate the previous finding that delayed L,I acquisitionimpcdes language proficiency in adulthood: the deafdelayed LI learners entcred school with little functional language. A dearth of language acquisitionin early childhood has deleterious effects on the outcome of all subsequent language learning in laterlife (sign language and written and read language);neither the first nor the second language is acquiredto near-native levels in adulthood.

In another experiment, the English task waschanged from grammatical judgment to sentence-to-picture matching. Despite the added nonverbalcontext of pictures, the delayed Li learners, all ofwhom had normal nonverbal IQ, performed at lowlevels on several English structures including conjoined, passives, and relative clauses. Replicating theresults of previous experiments, the two L2 groups,

one hearing and one deaf, again performed at mnative levels. ihus, three experiments in tv’.o 1guages (ASL and English) show, first, that ag.acquisition effects arc robust and persist into adihood for Li acquisition. Second, the results showLi acquisition in carlx life supports and facilit:subsequent L2 learning independently ofmotor modalirs’ \lavherrv, 200Th.

When Li Exposure is Delayed UntilAdolescence: Effects on Language OutcomExperimental studies of groups of deaf adults sItthat the scope ol’ the critical period fr langupertains to the first language, and that a lack’language acquisition in early life impedes the ahito learn language throughout life. The few avaiL’case studies of deaf individuals’ Li acquisition bein adolescence corroborate the main findings oft1experiments. First, the rate of delayed Li acqution is significantly slower than that of timelyacquisition. Second, adult language proflcicncsignificantly limited in comparison to earlyacquisition or later L2 acquisition.

Two studies longitudinally followed the spolLi acquisition of a deaf adolescent and an at

C.)a)0

C.)C’,‘aC.)Ca)C

0)

0C0t00.0

0

Type of Early Experience

Fig. 19.3 F 0 a’ a earl’. e’periaJ.e ufl later Ianguaec learn n g. a. rkinenr’a Sign I ,inguae ‘\St a’. rIot ‘i a ne a do t’. ‘di’.

F ad e’.peri’.n.ed no language in earl’. hf.,. uhjwts wer., :std using a task r._q’.siring ren,JlI ot .On’.f’i’. , 551 seiit.O’.C. 5. t

performanse of deal adults who had had experien’.ed d’.’I in infanse, and of hearing adult’. isho had n’.pcii’.’nrd p’ •i’’.a. linguae

other titan English in infania. Suhjsns were tested using a task requiring judgments otv.hether ‘.un’.pkx Euglidi ,eiiet,,, g”.en ii

print were grarnmatwa]h orreer; chance performance is 5O’.’. Figure 1 from Ma’. herr’.’, R. I.. I ock, F.. g Kaimi, II OO2

Linguistic ability and earl’.’ languagr exposure. ,‘\ano, ‘sI —‘ 35. with permission of th publisher.

Later Language:American Sign Language

Effects of Early Experience on Later Language Learning

Later Language:English

No Early Early Spoken No Early Early Signed Early SpokenLanguage Language Language Language Language

286 EARLs L5.’sGI’SGE ta QuIsil 1ON \ND ,DULI LA’SGU\nt SEILI iv

who had no exposure to sign language. in one study,the spoken Spanish acquisition of an adolescentboy was observed once he obtained hearing aids(Grimshaw, Adeistein, Bryden, & MacKinnon,1998). After 4 years, he acquired limited vocabulary, could combine single words with gestures, butproduced only a single two-word utterance. Anotherstudy followed the spoken English acquisition of awoman whose hearing loss was not identified untilshe was 31 years old. After 10 years of hearing aiduse and spoken English instruction, she was reportedto have a vocabulary of approximately 2,000 wordsand a grade 2—3 reading level, but her spoken utterances were described as being ungrammatical(Curtiss, 1988). The slow rate of delayed LI acquisition observed in this adolescent and adult contrastssharply with the rate of LI acquisition in young deafand hearing children. Early childhood language acquisition is characterized by voracious vocabulary learning (Anderson & Reilly, 2002; Bates & Goodman,1997). By the age of 6 years, hearing children cancomprehend as many as 14,000 words and most grammatical structures of their language (Clark, 2003).

Another study longitudinally observed the Liacquisition ofASL begun at age 13 by two deaf adolescents who had received no special services in theirhome countries prior to immigrating with theirfamilies to North America (Morford, 2003). After31 months of ASL exposure, both adolescents hadreplaced the bulk of their gestures with ASL signs.No vocabulary assessment was made, but comprehension tasks given after 7 years of exposure showedboth adolescents to suffer from severe comprehension deficits in ASL (Morford, 2003). These comprehension deficits caused by delayed Li acquisitioncorroborate the experimental findings summarizedearlier (Mayberry, 1993; Mayberry Lock, & Kazmi,2002).

Deaf individuals who are not exposed to language in childhood are not completely devoid ofcommunication with those in their surroundings.Deaf children who acquire little functional, spokenlanguage in early childhood have been observed togesture for communicative purposes with theirfamilies, a phenomenon called home sign in ASL.Detailed analyses of home sign show that deaf children combine points with iconic gestures in a rule-governed fashion to name things in their environment,make comments, and make their needs known tosome extent (Goldin-Meadow, 2003). This line ofresearch shows that the deaf child creates the gesturesystem and, although they understand it, hearingparents do not appear to use it with the deaf child

instead of speech. Home sign has been documentedto occur cross-culturally among deaf children whohave otherwise not acquired a spoken or sign language (Goldin-Meadow, 2003). Home sign was themeans of communication used by the two adolescent Li learners ofASL studied by Morford (2003).Although home sign displays many language-likefeatures, it does not appear to function as early Liacquisition in the critical period phenomenon. Thedeaf delayed Li learners who participated in theexperimental and case studies described here didnot attain near-native levels of ASL proficiencydespite years of using it, even though they had usedhome sign as young children.

Together the results of these diverse studies indicate the effects of critical period for language to bethreefold: (1) early Li acquisition leads to native-like language proficiency in adulthood; (2) early Llacquisition supports and facilitates subsequent L2learning, often leading to near native-like L2 proficiency in adulthood; and (3) a lack of early Liacquisition impairs the ability to learn languagethroughout life, that is, Li acquisition that beginsuncommonly late in human development leads tolimited language proficiency for any language inadulthood.

The critical period phenomenon described heremay explain the frequent clinical anecdotes fromseveral countries, including the United States,Canada, Norway, Denmark, and Sweden, that youngdeaf children who have acquired sign vocabulary aremore successful with cochlear implants than thosewith no sign vocabulary. Early language acquisitionfacilitates subsequent language acquisition cross-linguistically (Mayberry Lock, & Kazmi, 2002).The first stage of language acquisition is vocabularylearning. Computer modelling shows that theamount of language input and size of early vocabulary affects the organization of semantic categoriesacross the early lexicon (Borovsky & Elman, 2006).Vocabulary acquisition in sign may thus help theyoung child with a cochlear implant identify themeaning of distorted acoustic stimuli and bind it toalready acquired word meaning and semanticcategories. This is an area where research is neededdue to the substantial rise in the number of deafchildren receiving cochlear implants (see Chapters 9and 29, this volume).

Preliminary research suggests that delayed Liacquisition has significant neural consequences aswell. The degree to which the classic language areasof the anterior left hemisphere are activated duringsign language processing are negatively correlated to

MAYBERRYj

287

the age at which the first language was acquired

in childhood (Mayberry, Klein, Witcher, & Chen,

2006).

Sign Language Skill and ReadingDevelopmentGiven that early Li acquisition facilitates subse

quent L2 learning to near-native levels, but late LI

acquisition impedes it, the next question is whether

this critical period phenomenon relates to reading

development in the deaf population. Some, but not

all, theories of reading development posit that the

reader must be able to speak the language repre

sented in the written text in order to comprehend it.

In these models, recognizing word meaning occurs

only after written letters have been mentally trans

formed into the speech sounds they represent. Such

theories predict that readers who are deaf and do

not speak well will have difficulty reading. Note that

this prediction is at odds with the research findings

summarized earlier in this chapter, in which early

and robust language acquisition supports other

kinds of language acquisition independently ofsensory—motor modality. Early spoken language

acquisition supports later sign language acquisition

and vice versa, including written representations of

a)>a)-Ja)

(!30)C‘0(aa)C(aa)

spoken language. This line of research suggests analternative explanation for low, median literacylevels in the deaf population. If Li acquisition inearly life scaffolds subsequent L2 learning, then it

should support L2 reading too, even when the firstlanguage is ASL and the second is English. Many

deaf individuals have wealdy developed languageskills in any language, for all the reasons just described.

The question is how such weakly developed language skill relate to reading achievement.

To answer the question, adult deaf signers wereclassified as having either strong or weak ASL skillsas measured by grammatical judgment and narrative

comprehension tasks (Chamberlain & Mayberry,

2008). The two groups of signers were also givenstandardized reading tests. The results were striking.

There was a bimodal distribution of reading achievement between the two groups with no overlap, asFigure 19.4 shows. Average English reading achievement for the group with strong ASL skill was

between the grade 10 to college level, depending

upon the particular reading test. By contrast, theaverage reading achievement of the group with weakASL skill was between grade 3 and 4 (Chamberlain& Mayberry, 2008). Thus, sign language proficiency

is a strong predictor of reading achievement among

Fig. 19.4 Mean performance of deaf adults grouped by American Sign Language proficiency level, skilled and less skilled, on the

reading comprehension subtest of the Stanford Achievement Test and the Gates-MacCinitie passage and vocabulary comprehension

subtests of the Gates-MatCinitie. Data from Chamberlain, C., & Mavberrs; R. 1. (2008). ASL syntactic and narrative comprehension

in skilled and less skilled adult readers: Bilingual-bimodal evidence for the linguistic basis of reading. Applied Psycholinguistics, 28,

537—5’s9.

ASL Proficiency and English Reading Achievement

Skilled Less Skilled

ASL Proficiency Groups

288 EARLY LANGUAGE ACQUISITION AND ADULT LANGUAGE ABILITY

deaf signers, just as language proficiency has beenfound to be a strong predictor of reading achievement in the hearing population (Dickensen, McCabe,Anastasopoulos, Peisner-Feinberg, & Poe, 2003).Note that the average reading level of the signerswith weak ASL skill was identical to the medianreading achievement reported for the deaf school-aged population in the United States, namely betweengrades 3 and 4 (Traxler, 2000). This provides preliminary evidence that delayed LI acquisition maybe a significant factor in the attenuated readingachievement in the deaf population (Chamberlain& Mayberry, 2008).

Implications of the CriticalPeriod for LanguageThis review of research investigating age of acquisition effects on Li and L2 outcomes aliows us toreconsider the question of whether Victor couldhave become proficient in French Sign Language inthe early 1800s had the idea to use it occurred tohard. Victor was believed to be about 12 years oldwhen Itard began to teach him French (Frith, 2003).Itard described the boy as having no gestures and nolanguage, although eventually, after being socialized, he spontaneously used objects to communicate some of his needs (Itard, 1896/1962). Toddlersuse objects to communicate with their caretakersand then begin to use their hands and arms to pointat objects and people before they speak their firstwords (Bates, 1979). This means that Victor achieveda prelinguistic level of communication. The researchsummarized here completes the picture by suggesting that Victor would not have become proficient inFrench Sign Language because he had grown intoadolescence without having acquired any language.Scholars have suggested that Victor may have beenautistic. (Lane, 1976). Socially isolated children,including feral children, often develop autistic-liketendencies, but young autistic children can and doacquire sign and spoken language (Frhh, 2003).

If the critical period phenomenon for languagedescribed here is robust and a key factor in adultsign language proficiency, why are incomplete language acquisition and low sign language proficiencyamong deaf signers not more often identified andurgently treated by clinical and educational professionals? One reason is a lack of educational or clinical models of normal versus delayed sign languageacquisition, such as are in place for spoken languageacquisition. including assessment tools and specialists. Educators and clinicians need to be trained toassess language development in sign.

A clinical anecdote illustrates the serious natureof the problem. An adolescent boy who had beeneducated in total communication in a public schoolfor deaf students was brought to a university clinicfor an educational assessment. His ability to readEnglish was limited, as was his ability to comprehend simple sentence structure in ASL or signedEnglish. When asked about his sign language skills,his teachers uniformly replied that they could notunderstand him and attributed this to his quicksigning rate. His parents, who did not sign, wereaware of his limited reading ability but assumedhim to be a proficient signer, again due to his fastsigning rate. In short, neither his teachers nor hisfamily were aware of his aphasic-like symptoms insign language. Magnetic resonance imaging (MRI)scans revealed numerous and widespread brain lesionsdue to birth complications. The lesions explainedthe aphasic-like symptoms in his signing, but theywere invisible to professionals unprepared to ascertain the difference between normal and aphasicdevelopment in sign language.

In conclusion, research investigating whether acritical period exists for language acquisition hasused several kinds of variation in the temporal onsetof language acquisition over human development:spoken language, sign language, case studies ofchildhood social/linguistic isolation, and two typesof acquisition, Li and L2. L2 research with sign andspoken language has found a negative correlationbetween age of acquisition and L2 attainment.Nonetheless, near-native skills are often achieved byolder L2 learners depending upon amount of education undertaken in the second language and thelinguistic similarities between the first and secondlanguages. However, unlike the findings for L2acquisition, delayed Li acquisition impedes theultimate proficiency attained in any language,signed or spoken.

AcknowledgmentsThe research reported here was supported by grantsfrom the National Institutes of Health, the NaturalSciences & Engineering Council of Canada(#171239), and the Social Science and HumanitiesResearch Council of Canada (#410-2004-1775).The author thanks Deaf communities in Chicago,Detroit, Quebec, Ontario, Alberta, and Nova Scotiafor their invaluable assistance with this research;Drucilla Ronchen, Patricia Viens, and PamelaWitcher for their ASL insights and testing help; andAmy Lieberman and Carl Vonderau for helpfulcomments on earlier versions of this chapter.

MAYBERRY 289

References5.nderson. I).. & I/sills, J. (2002;. Ihc MacArthur

(.onini un leans c I )cs elopnsssc t 1 is en 51st 5: )\srmatis e Liar,,

[or American ‘sign Language. J,ns ‘s, or J)e,zf Srssslws cind

Do’ art/en. —, S i—1(t6.

Bares, E 1 ‘)9 . 1/se em, Gin, ef ;,i/so (s”siiio,s aad ,os,s

isiunistissrs sr i’ssm Ness Son A .s’(cmic Press

Bases, I ., & (soodtn sn J. (2 1 9’l,. 4 )n the inscpatahilmiy at

grammar and tic’ I-vs _omi I id net front acqinsition aphasia

and real-tins’ processing. Lasscsssp.e ,osl ( gsiiGts P,snc, ss,

12. 5)G .S3.

Birdsomsg. [).. & Malls. 51. 2001 ((is the c Dense [or matura

in nal constrai :sis in ecs nd I nwc’ ig Sc W hiatt. J’m s.a/ of

,‘sIsrnos c’. La.gmsscge, ii. 235 2 s’i

Borovske A.. & Llnian. 1. 2006). 1_ sngslage inpin and sensantic

s.ategorics’ 8 rdation hetsseen cognition and earle ssord

learning. Jcurn.il ef(issld 1,? is no or .16, V) 9)i.

t_bansberlain. (1. & MaCwits. 0. 1. (loBS,. ASL ‘entactic

and narratis corn prehension in skil led and less skilled adult

readers: Bilingual_bimodal evilence tot the linguistic basis of

reading. Applss’d ISysiiolsssgnssils’. 28. NA.’ 840.

Clark. K. 8. (2003i. Firt Iasigisigs .icgsisIsrtscii. Ness York:

Cambridge Univccsits Irc’s.

Corina, D., & knapp, H. (2000). Sign language processing and

the mirror neuron system. ( gee,. i2, 829 539.

Curtiss, S. (1 97). Genie: A psyehsslznguistic study a_f a mode rn_day

“wild child. “Ness’ York’. .‘kcademnsc l’s’is.

Curtiss, S. 11988). Abnormal language acquisition and the mod—

ularit> of language. In F. J. Nesvnteyer (I do Linguistic cheesy:

E.srensicssis and smplicatwns (Vol. 2, pp. 96 1161. Ness’ York:

Cambridge University Press.

Dickensen, D. K., McCabe, A., Anastasopoulos, L.. Peisner

Frinbcrg. E. S., & Put, I). M. (2003. Bringing it all together.

The multiple origins, skills, and ens’ironmncntal supports of

eat lv literacy. l.ests ning Dt55tbil1ti Reiesoch and Ptsict,ce, 16.

186-202.

Emmorev, K. (2002). Language cognitiosz csndtJ,e Mahwah.

NJ: Lasvrenee Erlhaum Associates.

Erninotev K.. Bellugi. C., Fniederici. A. 0.. & Horn, P (1995). Effecrs

of age of acquisition on grammatical sensitis By: ISv idence from

on-line and off’line tasks. lpphsdPsyc/olsngiustscs, 16 I—23.

I ubank, L., & Gregg. K. K. (1990) Critical periods and (secondl

language acquisition: Divide er inspera. In D. Birdsong (Ed.),

Second language acsjuisinon sisid the critical period hypothesis

(pp. 68 99). Mahw,th. NJ: Lawrence P rlhaum 5sswsc iares

Flege, J. F., Yrni-Kom’loan, C. H., & Liu, S. (1999). Age eon

straints on second_language acquisition Journal cs_f dlesnors’

audi angucige. 41. 8 IOn.

Frith, U. (2003). A’tttsrn 1 sjslsiinsssg the cnigsna, 2nd edition.

Oxford: Blackwell.

(,a[tnomi, I . (‘sVrtter( (I 9’Ot(. Ihe si_tACt of the svild child. In

F. Garmon (l’roducer;, Sot is. 1 ‘if.: 58 CR11 Boston.

Goldin-Sleadoss. S. (21)0)). Ifs’s, thesis a/Lsngtsiçs. Ness York:

I’s’schcslogr I’scss.

(.reenough. 55.) F. b Black. J 1 . (1 ‘50 . lmssletctissn of brain

structutc f’;, spc 6 r, e sulist, stes 55,1 am’ isis c des_h spnss’nt.

In \‘ 1/ tAsc,mna & p .o. AL. so . Pds... Ii, ‘is ‘5,55, ‘scsI

5’ .5. 555515 cs, s is 2. s ‘U’i Ian;5 Isle Al:

I ass’eise Filbisit 5

(.,rsnssh.sss. Os \i Sdelstc(ss S.. Bssd’ i 55. P. iS. \l,sckinsiosi

1 . 109’s . lb it angu sc 55.’ p.s ‘‘its’ r in Ds.ili_ scenes:

Es ide ice for a i_sit sc si period tar cOst 1 sngisage des clop

tsscnt. /? sssss size1 I sngssags. 6/, .‘3

I )alut,s. K.. Bialvstssk. F.. & 55 ils.s., 1. (21)03’.) titie.tl ci idence

A test of the critical period 1sy pcsmhescs [or second languag

acquisition .l’rsi’sslogis ii sc sesscr. in. 3 I —.38.

hard, J. SI. (5. (189o’1962s. The s lid ho3 oJ le’evsesn ‘Len_fan

‘list a/s’. ‘Rqmosrs e! ,ncssss’t, tAo I etsssstg’ ‘55’ Th11 e’:,o,,

C. Hunsphres &M. Huniphrs.p.lrans. . Ness York: iSppleton

P enturv Crafts

lohnson, . & Ness port, E. I . 198). (. I tic ii period etlects is

weond 5,’nguage frill 1n0’:.dse .otluence at mattcrat2sss.

iate sir the acquisitioss of i-.slish as a second lan5uage

(og.s:ti r fc’,,i’s,L;;, ,.‘. (() (s’5

klima. E. S.. & Bellugs. 4. (0’) lie .,ess’ of Lt,,s,,,

(.,soils ridge: Har; .srd Lii a cr/st (‘tess

Lose. H. I oThs. f/se ,s lid °c. s I e, son. (..ambridge: Harvs.r

Ln is ersirs Press.

lessnehc’rg. III. l)6’, BsoI’sgscriIrs’sisi..rioss. uI lasgsssssc’c. Ac’

York: joli n T ilrv A. Sons. Ins..

I orenz K. i’ 96’s. Es ol:srtoss ‘ni sits,,!, fi_,sts’ss; c/ hJ,ss zo

(_lucago: Utaversiry set 5. hieaga ltress.

.‘tla, herr’,, K. I. 1 ‘(9) . Fiist lancuage acquisitissn after chile

hood dOfets from ‘ecomid language acq.sisiti.sn: Ilse ..as,.

Antericams Sign Language. J’cr,.t,’ of Spes.c/s ,m,.’t’ fL,,

Reseush, 36. 1258 1

‘dat beers: K. I. I I 99.N. The importance of ehildhooel to lan

guage acquisition: Insights from American Sign I.anguage. I

j. C. Goodman A. El. 0. Nusbaum (Edw s, The dei c/op o,ent

speech pesceptiost: Th tranrsttos from speech sanods to ieee

(pp. 5-90s. Cambridge: MIT Press.

Macberry, K. 1. (2(102>. L.ogninse development of deaf childret

The interface of language and pes’eption in neurtspsveho

og). In S. J. Segalou its A. I. Rapin lFds.), Handbook of nit

rssps,ychaecsgy, 2nd edition. ((ui. a, pp. 1—10). Amsierdan

Elses’ier.

Maybcrrv, K. I. (200”). When tinting is every thing: Age ofhrs

language ac4uisition effects on second-language learnims

Applied Psycholinguistics, 28. 5 V’—549.

Mavberry, K. I., & Lichen, F. B. (1991). The long-lasting aduam

rage of learning sign latiguagc in childhood: Another look

rite cs iticai period [or language e quisitis.sn. Jan, ,issi ssf’.ifesssss

and Language, 30, 486 512.

Ma’. berrs, R. I.. & Fischer, S. D. (1989L I csoking through phs

nological shape to lexical meaning: The bottleneck of font

arise sign language processing. c’tletiioim asmd I ogs000si. 1

7n0 754,

,Vlayherre, K. 1., Klein, 0., \X’irchei. P. & ( hi_ti. j. K. (2904

Netssal ststesns ussderlvzsig ca It lassgisage e.spertence. Paper pm

sensed as rite Societs for Neuroseiemiee, Atlanta. GA.

Mavherrc, K. 1., & Lock, I . (200 51. Age cnnsmraitmts 5sf lit

versus secesisd language acqttisitiomm: Fs idence foe littguism

plastieit\ and epigenesis. Bsatn asi,l Ictssgtsrtge. 8°. 369. 38.

Ylayberry, 1/. 1.. Lock, I., & Kasmi, H. 20U2s. I.itigeusnic abili

amid ea eli langciage expsssum e. . si —. 38.

Mayherrs: R. I., A. Squires. B. ()000). ‘sign language aeqossimio

in K. Brass n (Edt, JOists lo,eedss sfi.tsss.issss.e iisi,gnsstic’. A

edition. (Sol. 11, pp. 291 l’t6). Oxford: Elsesiet.

S lot f,srcl. I. 2 j5)) (s taim nsatie.s I des tis spnis.i. m i is .ie(olcSec

f rst lango.sgs. less sets I ii)l . ‘1 , 1 , liSt I

\.,p,u1i L55’s’ss .‘vlstitr,r’oI_s.’nst..,s.ssjss’.’:si_,,,,

Itarsi ng C o_,ss,” i S s s . ‘ ., ( 1 28.

f’cntsslci. \5,,,,

I.’ •‘s$’t . s,,,,,

l’rsisecrssn: Pt ii celois sit sci.i s’ l’s c ss.

I’ s)etsc. Ii., Kli,n.., I S e\ lsllssi. I . : or” . 0;

). / , . I stssbrsctg. . ‘ii 5: “,il I Press

290 1 (1513 LA/c 5.505 5sQI.t’ssttt)\ .5”D AtOLL) t.(’,t,t...(c,F BiLit’

Reilly, J. (2006). How face comes to serve grammar: The developmenr la nonmanual morphology in American SignI anguage. In B. Schick, M. Marschark. & P. Spencer (Eds.),Advances in iign language development of deaf children(pp. 261—290). New York: Oxford University Press.

Schb.k. B.. de Villiers. P. de Villiers. J. & Hoffmeistcr, R.(200”). Language and theory of mind:A studvofdeafchildrcn, Child Development, 8. 3-6—396.

Seidenberg, M. S.. & Zeviti, J. D. (2005,. Connectiottist modelsin dexelopmental cognitist neuroseience: Insights aboutcritical periods. In 1. \lunakata & SI. H. Johnson Eds.l,iticnOO,( and performance XXI: Proce,es of change in

brain and cognicac detelopmenc. Oxford Oxford UniversityPress.

stokor. SC C.. Casterline. D. C.. & Croneberg, C. C. (1965).A dictiona p ofAmet icon Sign language on linguistic primiples.Washington DC: Gallaudct University Press.

Traxler, C. B. (2000). ‘The Stanford Achievement Test, 9thEdition: National nornsing and performance standards fordeaf and hard..of.hearing students. Journal of Deaf 5rudzesand DeafEducation, 5, 337—348.

Van Cleve, J. 55, & Crouch. B. A. (1989). A place of their own:creating the deaf community in .4merica. Washington DC:Gallauder Unisersirv Pless.

Werker, J., & Tees R. C. (200). Speech perception as awindow fur uttderstandiitg plasticity and commitment inlanguage systems of the brain. Dci elopmesttal Pcchobiology.46.233—251.

Wiesel. ‘L N. I 1982. Postnatal development of the visualcortex and the influence of the environment. Nature, 299,583—.S91.

Zeshan, L. (2006). Sign languages of the world. In R. E. Asher(Ed.), Encyclopedia of language and linguistics, 2nd edition.(pp. 335—365). Oxford: Elsesier.

MAYBERRY 291