2()0 lvlaltnlas rlulz

Waas, M. (1996). Language Attrition Downunder: German speakers in Australia. Frankfurt: Peter Lang. Weinreich, U. ( 1953). Languages in Contact. The Hague: Mouton. In search of the lost language The case of adopted in

Valérie A. G. Ventureyra and Christophe Pallier Laboratoire de Sciences Cognitives et Psycholinguistique, EHESS-CNRS, i Unité de Neuroimagerie Cognitive, Inserm U562, SHFJ & IFR49, Orsay

Introduction

Studies oflanguage attrition have often focused on language loss and interfer- ence in a contact situation and hence involve individuals who at some point could have been considered as 'bilingual'. Under such circumstances it is diffi- cult to determine the degree of 'attrition' due to erosion from disuse and that resulting from interference from another language. These confounding factors complicate the understanding of mechanisms leading to language attrition. Many of these studies are also primarily aimed at evaluating morphosyntac- tic and lexical modifications in the attrited language, excluding phonology, an aspect of language known to be acquired within the first year of life (Werker & Tèes 1983), and therefore quite relevant to the study ofLl attrition' The present study addresses these two issues. Firstly, it considers individ- uals who, at a relatively early age, were extracted from their native linguistic and cultural environment only to be immersed in a radically different mi- Iieu, requiring the acquisition of a new language and rapid adaptation, to the detriment of the native language which had become obsolete. These individu- als were then not fluent in both languages, or at least not for very long, thus minimising the interference factor. In this particular case, attrition may result from 'brain plasticity': the brain's capacity to adapt to a changing environment, which is believed to de- cline with age and brain maturation. The decline in the brain's ability to adapt and to learn is described by the 'critical period hypothesis' as it pertains to In search ofthe lostlangv^ge 2o9 zo8 Valérie A. G. Ventureyra and Christophe Pailier

French language acquisition (Lenneberg 1967). According to this author, a'critical pe- completely forgotten their native language. Eight native monolingual mean 23'5) riod', or a time-frame during which the learning of a given cognitive ability is speakers (2 women, 6 men ranging in age from 22 ro 28 years' = optimal, applies to language and extends from age two years to puberty. Nu- and unexposed to Asian languages, composed the control group' "lan- merous studies ranging from brain-damaged / hemispherectomised children These subjects participated in three behavioural tests. The first was a languages (Korean, (Basser 1962) lo the extraordinary cases of 'wild-children' deprived of language guage identification" test: sixty sentences from five different subjects until relatively late in development (Curtiss 1977), and congenitally deaf in- |apanese, Polish, Swedish, Wolof) were presented one at a time and their de- dividuals having been exposed to sign language at different ages (Mayberry *ere required to give a confidence score between I and 7 indicating was a "word & Eichen 1991; Newport & Supalla 1987), support the claim of maturational gree of certainty that the sentence was Korean. The second test one mechanisms involved in L1 acquisition. recognition" test: on each of 24 trials, the subject was required to choose of The third On the other hand, the maturational hypothesis also suggests that exposure two spoken Korean words as the translation of a given French word. followed by to Ll may leave long-lasting traces in the neural circuits involved in language test used a "speech fragment detection": a sentence was presented fragment processing, and thereby prevent total Ll loss. Thus, according to this hypoth- a 400 ms speech segment and the subject had to decide whether this were esis, we should expect to find traces of the at some linguistic appeared in the sentence or not. One-hundred and twenty-eight sentences Polish' level in our population of Korean adoptees, who were exclusively exposed to presented in four different languages: Korean, French, |apanese and MRI Korean during the first 3 to 9 years of life, and subsequently isolated from This task was performed while the subjects were scanned with functional their Ll during 15 to 30 years. Conversely, the absence of remnants of Ll (Magnetic Resonance Imaging). In the Language would indicate an overwhelming plasticity and replacement of Ll (Korean) by The results from the behavioural tests appear on Figure l. sen- L2 (French). Identification task, the Korean adoptees did not recognise the Korean in Secondly, an important part of this study is devoted to phonology, as it tences. They tended to confound the Korean sentences with those Japanese, is the first linguistic level acquired by infants, and perhaps one of the last to by giving these higher confidence ratings than the Polish, Swedish and Wolof this task followed the be lost. Several studies aimed at determining the benefits of exposure to a sentences. The French control subjects' performance on language in childhood, followed by a long period of disuse and then by re- same pattern and did not differ significantly from that of the Korean adoptees. Korean word learning several years later, have shown a pronounced advantage in perception In the Word Recognition task, requiring the identification of the (560lo) and production in the attrited language (Au, Ifuightly, Iun, & Oh 2002; Oh, corresponding to a given French word, the performance of the Koreans the Frag- )un, Ifuightly, & Au 2003; Tees & Werker 1984) relative to unexposed learners. and that of the French subjects (52o/o) did not differ significantly.In per- The study by Au et al., however, did not reveal any advantage of early expo- ment Detection task, both the Korean and French subjects showed better languages' sure versus late exposure in morphosyntax, suggesting that the mechanisms of formance on the French sentences than on those in the other three the two groups. language loss, as for acquisition, may be different for diverse linguistic aspects, Once again no significant differences were observed between sub- and that phonology may be particularly resjstant to attrition. Two main results emerged from the brain imaging data acquired while jects listened to sentences and performed the Fragment Detection task. The first result concerned the processing ofKorean: no activation specific to Korean Part I: The first behavioural tests and the fMRI study was found in the Korean adoptees and thus, there was no detectable difference (see between the activations elicited by the Polish, japanese or Korean sentences French The study by Pallier, Dehaene, Poline, LeBihan, Argenti, Dupoux, & Mehler Figure 2). The second result indicated that the brain areas activated by (2003) involved eight individuals of Korean origin adopted by French fami- were comparable in both the Korean adoptees and the French control groups. (Wernicke's lies between the ages of 3 and 8 (mean age at adoption: 5.5 years; 2 women Activations were mainly found in the left superior temporal sulcus for Ll and 6 men ranging in age from 20 lo 32 years, mean = 26.8). They were re- area) and in the left inferior frontal gyrus (Broca's area), as is expected cruited with the help of an adoption agency in France. These subjects had not processing. been re-exposed to Korean since their arrival in France, and all claimed to have v4rçrrc paulel ^. u. vctrtur€yra ano \_nrlstophe In searcn ol trre rost larlEua8E zrr

a. Korean sentences identification

7 O E Korean subjects 6 g French subjects O 5 o

6 3 èo I 2 Figure 2. Brain activation patterns ofKorean adoptees while listening to Korean (left), fapanese (middle) and Polish (right). All languages activated the superior temporal gyri Korean Japanese Polish Swedish Wolof and there is no statistical difference according to the language of stimulation. Stimuli

Part II: The newbehavioural tests

b. Korean word recognition c. Speech segment detection (performed during fMRI) Five experiments seeking to reveal any possible traces of Korean were designed r00 presented here. These behavioural tests concern phonology and lexical 100 a and are knowledge: they include Phonetic Perception, Phonetic Retraining, Sensitivity 80 5so .t to Korean Phonotactics, Word and Number Recognition, and Implicit Memory bo These experiments are tested and results are rePorted o60-^ Eoo for Series. Presentlybeing O U here only for the Number Series Recognition Tâsk. All experiments took place I T 840 Y, 40 on a computer which displayed the task instructions, played the stimuli and recorded the responses. 20 -" s20 The participants in these studies comprised eighteen Korean adoptees, in- cluding some who had participated in the fMRI experiment. The new subjects 0 0 were either recruited for the first study via the adoption agency or via a cultural KF French Korean Japanese polish Group adoptees, Racines Coréennes, assembling adoptees from Stimuli association for Korean France, Switzerland and Belgium. Some of the recruits from this latter source Figure l. Performance of the Korean adoptees and native French subjects on three have been re-exposed to Korean, either by formal instruction or by a trip or a behavioural tests. stay in Korea, or both. There were two control groups. The first group consisted of twelve na- In general, the results of this first study seem to imply that French has over- tive Korean speakers who have been residing in France for several months or ridden Korean in the brains ofthe Korean adoptees. However, fMRI technology years and still regularly speak Korean. They were matched for age and edu- has its limits and may not be able to reveal subtle traces of Korean left in the mi- cational level to our group of adoptees. The second control group consisted crocircuitry of the language processing areas, and the behavioural techniques of twelve native French speakers matched for age and educational level to the used in this study may not have been capable of tapping into subthreshold, Korean adoptees. implicit knowledge that requires other merhodologies. The second part of this study is aimed at addressing these issues and further assessing the degree of Ll attrition in our population of Korean adoptees. In searcn or Llls rusL rarr6*-ô'

Valérie A. G. Ventureyra and Christophe Pallier

7 trFr I Kor Experiment 1: Series of numbers in dffirent languages by rehearsal' More- Number series are learned early in life and are memorised language' Thus' numbers over, number words are quite frequent in everyday 6 maybeexpectedtob"mo,"resistanttoattritionthanothertypesofwords'The followingexperimentspecificallytestsfortherecognitionofKoreannumber series among series from several languages' 5

Stimuli:Twelveseriesofnumbers(oneineachoftendifferentlanguagesand systems) of the numbers I two in Korean, which has two different counting 4 tol0spokenbynativespeakerswereeitherrecordedordownloadedfromthe mainly for counting ob- Internet. Korean has one counting system proper used (Sino-Korean) jects. The other system used in Kàrean is derived from chinese ) and telephone numbers' and is used for years, months, days, minutes' money Asbothofthesesystemsareessentialtodailylife'wecansafelyassumethat languages used were chosen on the young children acquire both. The other 2 (most series were non- furi, of th.i, unfamiliarity for average French speakers excluded as some of our Indo-European). Some Asian languages had to be (fapanese' Vietnamese' Man- !bD Korean udàpte.s had been in contact with these >ÉE oÉ'î' E4 o .= ë Hungarian' ! ten series of numbers were in: Greek' Turkish' où€ loO darin). The other d!& 'Ë inr I 9.d F and Cantonese' ll- a1 Finnish, Hebrew Farsi, Quechua, Tagalog' Thai e9 ÊL =

Srno-Kor lagalos I har I irrkrsh [u]. The second syllable was constant: 'ma'. (Examples of pseudowords: pima, 7 @ o o kama, suma). 6 oo o @ 5 o @ o responses o o o Procedure: The stimuli were presented through headphones and 4 æ oo o o oo session with twelve pairs of pseu- @ made on keys on the keyboard. A training J @ o o oo dowords of French occurred first. The first member of the pair was always 2 o oo G o o o a speaker and the second by a female speaker. Then, the o pronounced by male I o @ 138 pairs of Korean pseudowords were presented. As in the training session, the male speaker appeared first followed by the female speaker. The second stimu- o 7 lus of the pair appeared 500 ms after the first. The following pair was presented o o 6 ô o o o o 1000 ms after the subject's response. Subjects were required to indicate whether X o o 5 o @ the two words of the pair were the same or different by pressing a given button @ 4 o o on the keyboard as quickly and accurately as possible. There were ûve condi- o 3 o tions: Same, and four types of Difierent, according to the consonant or vowel o o 2 o o æ @ contrast in question. They were distributed as follows: (1) 48 pairs of Same @ @ @ o o I Canfonccc Korenn Farsi Finnish (P); (Z) 48 pairs of Different Vowel (DV); (3) 18 pairs of Different Consonant 7 oo lenis-fortis contrast (DCl); (4) 12 pairs of Different Consonant lenis-aspirated 6 o o @ o (DC2) and (5) 12 pairs of Different Consonant fortis-aspirated (DC3). 5 æ o o o o o o o @ o 4 o o o o @ o o o o o o 3 o o o caæ oo E:-:::â r o o ._"1 2 o o o o o

I o oo oo o oo Fr Kor Fr Kor Fr Kor Fr Kor

Figure 4. Individual confidence ratings for the recognition of number series (Experiment 5). o g+U^ Experiment 2: Phonetic discrimination The phonetic/phonemic categories of Ll are acquired in the first years of life. ô] It is also known that adults have difûculties perceiving some non-native con- trasts. Korean voiceless stop consonants use contrasts which are particularly difÊcult for monolingual speakers of French. In this experiment, we test the adoptees' ability to perceive such contrasts. F.DCI K.DCI F.DC2 K.DC2 F.DC3 K.DC3 F.DV K.DV F.P K.P

Figure 5. Percentage correct responses in the Phonetic Discrimination task (Experi- stimuli: Twenty-four cvcv pseudowords of Korean were recorded by each of ment 1) in tvvo control groups (native Koreans (K), native French (F) speakers) and two native speakers of the Seoul dialect of Korean (one female and one male). the different types of contrasts. (P = Same, DV = different vowel, DCI = lenis-fortis, The first syllable always started with a consonant characteristic of Korean: lenis, DC2 = lenis-aspirated, DC3 = fortis-aspirated) fortis or aspirated [k], [p], or a lenis or fortis [s], and was followed by [a], [i] or v. vLulurL/r4 4rru vurrsfuPus I drnsl lll scalell ul Llrç ruùr r4rr5u4Ë,c zt/

Resuhs of the two control groups: As shown in Figure 5, both the P and DV the l5 training sessions. The first test involves new real words of Korean be- categories were easi-ly identified by both native Korean speakers (n = 12) and ginning with the different t1ryes of [p] and is composed of two blocks- One native French speakers (n = 12). AII three DC contrasts were well discriminated block is composed of words spoken by one of the speakers heard in the train- by the native Korean subjects and poorly discriminated by the native French ing sessions and the other is composed of words pronounced by a new speaker. speakers with no overlap between groups, thus showing that the DC contrasts The second test is the Phonetic Discrimination described above and the third are indeed difÊcult for native French speakers with no previous exposure to involves words and pseudowords of lenisifortis/aspirated [t]. Korean. The DCI contrast (lenis-fortis) is apparently the most difficult for the French subjects. Experiment 4: Sensitivity to phonotactics Consonant clusters (e.g.Pr, sl.. .) such as those existing in English or French Experiment 3: Phonetic retraining do not exist in Korean or in |apanese. It has been demonstrated that fapanese Assuming that the results of the previous experiment show that the adoptees speakers perceive an illusory vowel within foreign consonant clusters (Dupoux, are not better than native French speakers at perceiving the Korean contrasts, it Kakehi, Hirose, Pallier, & Mehler 1999). This phenomenon is known as vowel may still be possible that the adoptees will learn such contrasts faster than the epenthesis.It is due to the fact that the human speech decoding system uses the native French speakers with an intensive phonetic training program. phonotactic constraints imposed by Ll to decode utterances. The properties of Korean phonologylead us to expect that native speakers of Korean will perceive Stimuli: Four minimal triplets of Korean words and four minimal triplets of an illusory vowel ([ui]) within consonant clusters of French words. If such is Korean pseudowords, all commencing with either a lenis, fortis or aspirated the case, it would be interesting to see whether the adoptees perceive French [p], were chosen as the stimuli for this experiment. Each of the 8 words/pseu- consonant clusters like native Koreans or like native French speakers. dowords was recorded by 6 native speakers of Korean (3 female and 3 male). There were 4 tokens of each stimulus per speaker. Thus, a total of 96 different Stimuli: Four pseudowords of Korean were recorded by two native Korean stimuli were recorded by each speaker (8 words x 3 types x 4 tokens). The vowel speakers (one female and one male). Two pseudowords involving a change in following the [p] was variable. vowel(ibakilibuki) areused in the control condition and two pseudowords pre- senting a consonant cluster and the epenthetic vowel (astalasuta) are used in Procedure: There are l5 training sessions in total. Each session lasts approxi- the experimental condition. Each word was modified giving rise to six tokens mately 30 minutes. The sessions are spread over 3 weeks, totalling 5 sessions per varying in pitch and speed. week. During the course of each session subjects are presented with 6 blocks of 64 stimuli each. Each block trains the subject to identifr one of two types of [p], Procedure: Each condition is composed of a training session for familiarisa- presented at the beginning of the block. Hence, only one consonant contrast tion with each stimulus, presented one at a time, and its corresponding button (lenis-fortis, lenis-aspirated or aspirated-fortis) is presented in a given block in on the keyboard. Following the training session, series of four consecutive stim- order to facilitate the learning of these different types of [p]. There are thus 3 uli are presented aurally. There are twenty-four such series in each condition. different blocks per session, each repeated once: a total of six blocks per ses- The subject's task is the identification of the sequence heard during each trial sion. The first six sessions are all single speaker sessions. The following sessions (ex. asta, eswta, asurta, asra). Feedback is given after each trial in order to become progressively more difÊcult by a gradual mixing of stimuli by different encourage learning. speakers: two speakers for Session 7 and all six speakers by Session 15. In each block and session, the subject's task is the identification of the type Experiment 5: Word recognition of [p] heard, by pressing the corresponding key on the keyboard. Feedback is The previous study (Pallier et al. 2003) implemented a word recognition test given after each trial in order to enhance the learning process. in which the participants had to choose the translation of a given French word Three tests of generalisation to other words using the Korean [p]'s as well among pairs of spoken Korean words. The participants Performed at chance as to other consonants sharing the same characteristics are administered after level, suggesting a radical loss of lexical knowledge. To test this hypothesis zt8 Valérie A. G. Ventureyra and Christophe pallier In search ofthe lost language 2tg further, we impremented a new test with more trials and words taken from Conclusion a corpus of Korean children's productions (see Appendix). In this new task the subject was with one Presented Korean word per trial and had to choose the The study of Korean adoptees having been completely cut-off correct translation among two French words. from their native language and immersed in an L2 environment relatively early in life presents an stimuli: sixty common interesting case to the field of language attrition. A possible mechanism for lan- Korean words, thought to be known by children (52 guage attrition due to erosion of which were raken from a corpus of an unused language is brain plasticity, and is compileà by choi a copnit< 1995) and recorded by a native supported bythe experiments described in the first part of this paper. However, speaker ofKorean, were presented aurally. one-hundred and twenty common French fMRI technology has its shortcomings and thus, the behavioural tests presented words, sixty of which were the translations of the Korean words and in the second part of this article, and presently being tested, are aimed at a more sixty of which were matched for written lexical fre- detailed assessment of quency (New, Pallier, Ferrand, & Ll attrition in our population of Korean adoptees. so Matos 200r) were presented visually on the computer screen. far, the data obtained on the recognition of the number series suggest that the adoptees have a somewhat more precise notion of the sound pattern of Korean Procedure: In than the native French, but no explicit access to knowledge of Korean lexical a given trial, subjects first heard a Korean word twice and then items. were presented with two French words simultaneously on the screen. The task was a forced choice recognition whereby the subjects indicated which of the two French words was the translation of the Korean word. References Experiment 6: Implicit memory for series of days of the week, months and numbers Au, T. K., Knightly, L. M., Iun, S-A., & Oh, I. S. (2002). "Overhearing a language during childhood'l P sy ch olo gical S cien ce, 1 i (3), 23 S-243. Experiment l and 5 used recognition tests which assess explicit memory for Basser, L. s. (1962). "Hemiplegia of early onset and the faculty of speech with special Korean words. we tried to devise a test that,would tap into implicit memory reference to the effects of hemispherectomy". Brain, 75,427460. for series learned in early childhood: choi' s., & Gopnik, A. (1995). "Early acquisition of verbs Korean: days of the week, months and numbers. in A cross-linguistic The test is based on the "Mere sttdy". Iournal of Child Language, 22,497-529. Exposure Effect" which reveals a preference for (1977). york, familiar items vs. unfamiliar items (Kunst_Wilson Curtiss, S. Genie: A psycholinguistic stuily of a modern-day'wild chitd'. New Ny: & ZajonclgS0). Academic Press. Dupoux, E., Kakehi, K., Hirose, Y., Pallier, C., & Mehlea J. Q999). "Epenthetic vowels in stimuli: Recordings of days of the week in Korean were made. Each day was fapanese: A perceptual illusion'i Journal of Experimental psychology: Human perception isolated and later combined with two other and Performance, 5, 1 568-1 57 8. days. series of three days were cre_ ated' Half of the series were regular (Monday, Kunst-Wilson, W. R., & Zajonc R. B. ( 1980). 'Affective discrimination of stimuli that cannot Tiresday, wednesday) and the other half be recognized". Science, 207, 557 were irregular (Monday, Thursday, Sunday). -555. Lenneberg, E.H. (1967). Biological Foundations of Langnge. york, The same procedure was used for months New Ny: fohn Wiley. or irr. year and numbers in ( Korean. Mayberry, R., & Eichen, E. B. 1991). "The longJasting advantage of learning sign language in childhood: Another look at the critical period for language acquisition". Journal of Memory and Language, 30,486-512. Procedure: In each experiment, (days, New, B., Pallier, (2001). months and numbers each gave rise to C., Ferrand, L., & Matos, R. "Une base de données lexicales du one experiment), a regular français contemporain sur internet: Lexique". L'Année psychologique, 10j,447462. series was matched with an irregular series. on u.gi"_." trial, the subject hears Newport, E. L., & supalla, T. (1987). A critical Period Effect in the Acquisition of a primary a pair of regular-irregular series and indicates which one he/she Language. Unpublished ms. prefers by pressing on u b-.rtto.r on the keyboard. (The regu- Oh, J., Iun, S-A., Knightly, L., & Au, T. K. (2003). "Holding on to childhood language lar series appear first on halfofthe trials and last on the other half.) There are memory'l Cognition, 86, 53--64. 14 trials for days, 24trials for months and 14 trials for numbers. 22t In search of the lost language Pallier zzo Valérie A. G. Ventureyra and Christophe (Park) (shoes) parc 27 chaussures D., Argenti, A-M., Dupoux, E., & Mehler' I. frrit (fruit) Pallier, C., Dehaene, S., Poline, J-B., LeBihan, clef (key) second language 28 (wall) ofiu"guage plasticity in adopted adults: Can a mrr (2003). "Brain i-ugi"f 29 nez (nose) f55-161' bol (bowl) replace the frrsT?." Cerebral Cortex' 13' horloge (clock) or recovery of the 30 (sleeve) e' (iSs+)' "Perceptual flexibility: Maintenance matche Tees, R. C., & Werker, l. j8, 31 balle (ball) sounds". canadian lournal of Psychology, puzzles (Puzzles) ability to discriminate spàch bonbons (candY) ""n_n*ir. 32 c'est fatigant (it is tiring) (it is noisY) 579-590. the t) c'est bruYant (it (1983)"'Developmental changes across childhood in c'est noir isblack) Werket I.K., & Tees, R'C' c'est beau (it is beautiful) 34 ie suis seul (I an alone) perceptionofnon-nativerp.".hrontdrl'canadianlournalofPsychology'37'278J86' c'estbeaucouP (it is alot) 35 premi" r fit is the frst) bon (it is not good) .'"rt l. 36 c'est n'est Pas encore (again) c'est (it is not alot) 37 Peu tu peux! (You can!) regarde ici | (Iook here!) 38 coiffe-toi (comb Your hair!) Appendix (get dressed!) 39 habille-toi (cahn down!) (close the door!) calme-toi! 40 ferme la Portel you!) on va te punirl (wc will punish ListlFrenchtranslationsofthewordsthatwereaurallypresentedinKorean 4l rangel (Put Your things awaY!) (Exp' 4) prendsJe! (takeit!) in the Word Recognition Task pousse! (Pashl) inside) 42 i" u.*.rr,r., (I wûflt to conrc 43 jejoue I PlaY) the List2FrenchwordsmatchedinwrittenfrequenrytothewordsinListl u.* l. Ptu- (I want Prizc) j'ai peur 1l am scarcd) l. 44 je (I tookit!) (I do not want to!) I'ai Prisl List 2 45 l. n" u"* Putt List 1 je souris (I smile) ie fraPPe (I ftit) 46 je le sens (I it) fauteuil (couch) do notknow) fecl I ctrapeau (hat) 47 l..,..uis Pas I (trY!) matin (morning) jettel (throw!) 2 Iumière (light) 48 "ttutst (quiet!) (day) silence! jov 1on rst (give!) 3 eau (water) 49 (I am the frst!) jeu (game) je suis le Premier! (eye j'ai mall (I amhurting!) 4 oe\l ) 50 je comPte (I count) cheverx (hair) jebois (I drink) 5 tr^iî (truin) 5l (pull!) rideat (curtain) 1ilgt poisson (fsh) 52 lève-toil (get uP!) 6 (ant) tiensl. (hold!) (medication) forrmi viensl (come!) 7 médicament 53 (swallow!) doigt (rtnger) (sit down!) avalet 8 lait (milk) 54 assieds-toi! (some) (socks) quelques :uns (choPsticks) chaussettes c'estgrand (itisbig) 9 baguettes 55 (less) (grasshoPPer moins (cookie) sauterelle ) jamais (never) 10 biscuit 56 (and so) nouveau (it is new) et alors (it is good) c'est toujours (alwaYs) ll c'est bon 57 comPris (I understood) c'est vrai (it is true) j'ai (l am cold) f'ai t2 c'est petit (it is small) 58 froid (it is rluick) (it is c'est raPide (itis sPicY) c'est gelé frozen) j'ai chaud (I amhot) 13 c'est piquant 59 (itkfull) c'est triste (it is sad) (itisred) c'estPlein t4 c'est lourd (it is heavY) 60 c'estrouge c'estplat (itisflaù c'est pareil (it is the same) 15 (wash hands!) (oPen the door! lave tes mains Your 16 ouvre la portet ) emporte-lel (take it!) t7 réveille-toil (wake-uP! ) tourne! (turn around!) 18 descendl (come down!) laisseiel (leave it!) 19 va-t'en! (go awaY!) devine! (guessl) 20 sorsl (go out!) je suis fier (I am Proud) 21 j'aisotf (I am thirstY) je suis tombé (I/ell) 22 j'ai faim (I amhungrY) famille (familY) 23 école (school) algeîT (moneY) 24 Iivre (book) sac (bag) 25 orellle (ear) balr:' (bath) 26 couteat (knife)