Wednesday, December 12, 2007

Final Paper: The Long-term Effects of Early Language Exposure After Language Attrition Has Occurred

The long-term effects of early language exposure after language attrition has occurred

Zenas Lee
Psych 17N Final Paper
Stanford University

1. Introduction
Babies have the seemingly magical ability to acquire language without effort. The widely accepted critical period hypothesis claims that after a certain age, people lose this ability. This theory has been extended to second-language acquisition. Supporters of this extended theory point to evidence of age as a factor affecting achievement of native-like fluency. Children are able to learn a second language “quickly, automatically, effortlessly, and to a level indistinguishable from that of native speakers,” while for adults, second-language learning is perceived to be “slow, effortful, and often less than perfectly successful” (Snow, 1987). Language acquired at a young age can also be forgotten. There are many people who claim to have been fluent in a certain language when they were young but cannot remember a word now. Such language loss or attrition occurs when exposure to that language is cut off or outweighed by another language, such as in the case of international adoption or immigration to a foreign country. Given these two characteristics of language (the critical period and language attrition), this paper aims to explore the long-term benefits of being exposed to a language in the developmental ages. First, a brief overview of language attrition will be followed by research on the effects of early language exposure. Then, the results of the various studies will be discussed. Finally, a conclusion will be drawn from the discussion and the author’s personal insights.

2. Brief Overview of Language Attrition
“Language attrition” refers to loss or attrition of skill in one’s native language (L1) or a second or foreign language (L2). The phenomenon can be observed in both large groups and individuals, but this paper refers to the latter form of language attrition. Individual language attrition can be subdivided into four categories: L1 loss in an L1 environment (e.g. aphasia), L1 loss in an L2 environment (e.g. immigrants), L2 loss in an L1 environment (e.g. L2 learners), and L2 loss in an L2 environment (e.g. older immigrants who revert to L1). This paper does not consider L1-in-L1 loss, as it has more to do with physical reasons than influence of another language. The degree of attrition in the other three categories depends on various factors, such as age, skill level, amount of other language exposure and use, etc. Although a highly interesting topic in itself, this will be the extent of the introduction to language attrition, as it is sufficient to understand how one could “lose” language, thereby allowing us to consider the effects of early language exposure after attrition has occurred.

3. Research on the Effects of Early Language Exposure
To properly discuss the effects of early language exposure, it is necessary to first compare language acquisition between adults and children. One prediction of the critical period hypothesis is that second language acquisition will be relatively fast, successful, and qualitatively similar to first language only if it occurs before the age of puberty, when cerebral lateralization is about complete. Thus, to test the hypothesis, many studies have focused on the age factor.
Studies examining the correlation between age of exposure and language ability have looked at immigrants and their language skills. One particular study compared the perceived foreign accent and grammaticality scores of 240 native Korean speakers who acquired English as a L2 at different ages (Flege et al., 1999). The participants’ age of arrival (AOA) in the United States ranged from 1 to 23 years, and their mean length of residence was 15 years. They were first required to complete a language background questionnaire so that the study could be analyzed using different variables as controls. Subsequently, they took two tests: 1) a test asking them to repeat English sentences that contained a wide variety of English vowels and consonants, and 2) a grammaticality judgment test. Results showed that as the AOA increased, the strength of perceived foreign accent increased and the grammaticality scores decreased, but when variables confounded with AOA were controlled, grammaticality scores turned out to be less relevant to AOA and more dependent on education and use of English/Korean. The findings from this study actually did not support the critical hypothesis theory because if the hypothesis were correct, there should have been a correlation between AOA and L2 performance for individuals who began learning their L2 before the age of puberty (12 years), but not for those who began learning their L2 later in life. However, the AOA–foreign accent correlations and the AOA–morphosyntax correlations were significant for both Koreans with AOAs of 2–12 years and those with AOAs of 13–23 years. These findings did, however, lend credibility to the theory that with increasing age, the ability to acquire or perceive phonemes that are not in our L1 decreases.
Now that we’ve examined the age factor, we can turn to research on the effects of early language exposure. Research on the beneficial and lasting effects, if any, of exposure to language during childhood is too limited to provide any conclusive evidence. However, recent studies have ventured into this area, and one study interestingly examined the effects of overhearing a language during childhood (Au et al., 2002). They tested the idea that having auditory exposure to a language during childhood may help people learn to speak that language with a more native-like accent later on as adults when they start studying it. The study compared adult learners of Spanish who had overheard Spanish regularly during childhood with those who had no regular exposure until around age 14. Both groups were students enrolled in a 2nd-year Spanish class at the University of California, Los Angeles. The overhearers had heard native spoken Spanish for at least several hours each week for at least 3 years between birth and age 6. The exposure decreased significantly after this period, and they had spoken and been spoken to in Spanish minimally (occasional Spanish words or short phrases embedded in English sentences) until they started taking Spanish classes around age 14. The L2 learners had minimal (the occasional word or short phrase in Spanish) or no regular exposure to Spanish until they started taking classes around age 14. Similar to the Flege et al. experiment, participants were asked to read Spanish sentences aloud. These were recorded and later phonetically (voice onset time) and phonologically (phonemic patterns and phonological rule) analyzed to rate the accents of the participants. Additionally, a second test was run with the same group of participants, examining whether morphosyntax could be more readily acquired by overhearers. The participants were asked to verbally complete five simple jigsaw puzzles, which were designed to test number and gender agreement among determiners, adjectives, and nouns in Spanish. Parallel to the Flege et al. study, the results of this study showed that overhearers were able to produce more native-like accents than L2 learners but did not have any measurable advantage in morphosyntax.
The third study that this paper will examine is similar to the previous experiment but differs in that it looks at childhood speakers of languages (Oh et al., 2003). The participants in this study were enrolled in 1st-year college Korean language classes and divided into three groups: 1) childhood speakers who had spoken Korean regularly for a few years during childhood; 2) childhood hearers who had heard Korean regularly during childhood but had spoken Korean minimally, if at all; and 3) novice learners who had no prior Korean experience. The first two groups both had experienced a sharp drop in hearing and speaking Korean around age 5 to 7. A fourth group of native Korean speakers were also tested for comparison. The participants’ first task was to listen to a Korean word and identify the word from several similar sounding choices. The second task was to read a Korean sentence aloud. The first task tested phoneme perception, and the second task tested phoneme production. The results of the study showed that childhood speakers were as good as native speakers at hearing the phonemic contrasts of Korean, outperforming the novice learners. They were also able to produce native-like accents, outperforming both novice learners and childhood hearers. These findings bring to light the benefits of childhood speaking experience and suggest that the benefits of early language experience are long-lasting even with little or no subsequent experience with the language. The nature of early language experience also seemed to be important: childhood hearers outperformed novice learners in the perception but not production of Korean phonemes. This evidence seems to point to early hearing experience helping later perception, and early speaking experience helping later production. However, the study recognizes that this cannot be the simple, general conclusion because the previous study showed that early hearers had a benefit in phonology production in Spanish.
There has also been research presenting evidence contrary to those of the studies above. A study examined the language abilities of monolingual adult speakers of French who had been adopted as monolingual Korean speakers by French families from Korea to France between 3 to 8 years of age and never again exposed to Korean (Pallier et al., 2003). The adoptees took two behavioral tests: 1) listening to Korean, Japanese, Polish, Swedish, and Wolof sentences and assessing whether they were Korean or not; 2) choosing the translational equivalent of a French word after auditory presentation of two Korean words. Also, while they listened to Korean, Japanese, French, and Polish, their brain activities were also analyzed using fMRI. Results of the behavioral tests showed that the adoptees were not able to better distinguish Korean from other languages or identify Korean words in a forced-choice task than native French speakers who had no prior exposure to Korean. Moreover, event-related fMRI activation patterns did not differ while listening to Polish (an unfamiliar language) and Korean, and adoptees' activation patterns for French and Korean did not differ from those of the native French speakers.

4. Discussion of the Research
The findings of the research explored above seem to generally show that at least the phonological aspect of language is retained after exposure to the language in the developmental ages. This seems like a plausible result when we consider the fact that babies are born with the ability to differentiate all phonemes but only acquire the ones of his/her native language and lose the rest. The studies suggest that if a person is sufficiently exposed to a language as a young child, the phonemes of that language are kept even after he/she has experienced language attrition. However, early exposure seems to offer no advantage for the morphosyntactic aspect of language. This suggests that morphosyntax is something you can acquire with effort when studying a language. While this seems logical, real-life experience and personal observation of Koreans studying English lead the author to believe that more effort is required of people without early exposure. One explanation for this is the difference between certain languages. While it is easier for Spanish speakers to learn French because of the similar characteristics in the two languages, it is more difficult for Korean speakers to learn French or Spanish because Korean does not assign gender to its vocabulary other than when identifying a man or woman and has different sentence structure.
The suggestion by the Oh et al. study of a difference in benefits arising from the nature of language experience in childhood seems logical, but as they pointed out themselves, doesn’t exactly agree with the Au et al. study. However, this disparity can be explained by several possible factors affecting the study. First, the Spanish overhearers had had limited but some experience speaking Spanish words, which means they were able to produce the phonemes they had acquired through auditory means. Second, the length of language education was longer for the Spanish overhearers than for the Korean hearers. The Spanish overhearers had studied Spanish since the age of 14, while the Koreans were taking first-year classes in college. If the Korean hearers had studied Korean for a longer time, they might exhibit better phoneme production.
The last study also contradicts the suggestion of an existence of beneficial effects from early language exposure. This can be plausibly explained. The study assumed that because they had known Korean as children, the adoptees would be able to recognize Korean. However, after experiencing language attrition (L1-in-L2), the adoptees could not be reasonably expected to identify what Korean sounded like without knowing what it was. They also would not know Korean words they had not used in over 20 years or ever (they may have never learned the word as a child). The adoptees could still potentially be better at acquiring Korean than native French speakers with no Korean exposure. A one-time experiment may not be sufficient enough to test these potential abilities. In the other studies, the participants were students who were taking Spanish or Korean classes, which could have triggered their abilities and lead to better performance on the tests. Therefore, a new study comparing the perception and production abilities of the adoptees and native French speakers after they have both taken Korean classes should be run.

5. Conclusion
The questions initially posed by this paper were: Does language exposure in the developmental ages help in acquiring that language later on in life? If so, which aspects of the language are retained? After exploration of related research, the author is convinced that early exposure to language leaves an imprint of phonemes that one can retain later on in life even after language attrition has occurred. This seems to correspond with the author’s personal experience of speaking both English and Korean more predominantly at different times in life as well as the author’s observation of friends who had varying degrees of early exposure to English. The author had exposure to both Korean and English at a young age. However, because of the occurrence of language attrition, the author had transition periods when she had to take ESL classes in kindergarten in America after living in Korea from age 3 to 5 and the ESL-counterpart of Korean in sixth grade when she moved back to Korea. She is now perceived by both native English speakers and native Korean speakers as a native of both languages. However, the author recognizes that her case is specific to her and cannot be generalized without further studies. The author would like to see future studies better examining the effects of early language exposure, so that this paper’s questions can be answered more conclusively and also further questions, such as how many different languages can an infant be exposed to and reap the benefits of such exposure from without it negatively affecting his/her linguistic abilities.

Au, T. K., Knightly, L. M., Jun, S. & Oh, J. S. Overhearing a Language During Childhood. Psychological Science, Vol. 13, No.3. pp. 238-243. (May 2002)
Flege, J. E., Yeni-Komshian, G. H. & Liu, S. Age constraints on second-language acquisition. Journal of Memory and Language, 41, pp. 78–104 (1999).
Oh, J. S., Jun, S., Knightly, L. M., & Au, T. K. Holding on to Childhood Language Memory. Cognition, Vol. 86, No. 3, pp B53-B64. (Jan. 2003)
Pallier, C., Dehaene, S., Poline, J.-B., LeBihan, D., Argenti, A.-M., Dupoux, E. & Mehler, J. Brain Imaging of Language Plasticity in Adopted Adults: Can a Second Language Replace the First? Cerebral Cortex. pp. 155–161, 1047–3211 (Feb. 2003)
Snow, C. E. & Hoefnagel-Höhle, M. The Critical Period for Language Acquisition: Evidence from Second Language Learning. Child Development, Vol. 49, No. 4. pp. 1114-1128 (Dec. 1978)

Tuesday, December 4, 2007

#17. Dreaming in Another Language

A 67-year-old man talks in Spanish in his sleep. This doesn't seem weird at all. But take into account that he has forgotten almost all his Spanish, which he learned as a child until the age of 10, when he moved to the Netherlands and started learning Dutch and other languages. He says that Spanish is the only language he was ever fluent in but now all that is left of it is only in his dreams. (link to article below)
In Korea, dreaming in English is people's "dream." They think that if they dream in English, it means that they have achieved a significant level in English. This doesn't seem totally nonsensical, so I wanted to find out how much correlation, if any, there is between dreams and language ability.
Steve Kaufmann, author of The Linguist, founder of, and speaker of nine languages, offers his opinion on the topic: "In my experience it does not mean much. It sometimes happens to me, I think it reflects the fact that we are really committed to the language, listening to it, and wanting to speak it. In our dreams we are not inhibited so we do well." In an online forum (, people seem to share the opinion that dreaming in a foreign language doesn't mean you have the language down solid. Citing their experiences, they seem to agree with Kaufmann that people dream in the foreign language when they're studying it intensely.
Other than other personal accounts, I couldn't find any research that pointed to a correlation between dreaming in a foreign language and ability in that language. Apparently, we're "smarter" when we're asleep, so we can speak the foreign language better in our dreams. However, some people have experiences of dreaming in a language they never learned, which they perfectly understand while they're asleep. A lot of sites that I came across as I researched this topic were about interpreting dreams, premonitions (dreams predicting the future), and sleep disorders like somniloquy (sleep talking). So it seems that a lot of the dream world is unknown, because we do things in our sleep that seem unexplainable like carrying out a conversation with someone.
As for me, I don't think I'm very conscious about which language I dream in, but I do find it odd when my English-speaking friends and my Korean-speaking friends come out together in my dreams and seem to communicate perfectly with each other. I think I've had dreams where I understood French perfectly, but the memory is vague, so I'm not really sure. My roommate tells me I sometimes say random words in Korean while I sleep, but as she doesn't know Korean, I'm not sure if it's just gibberish or actual Korean. Although I didn't find any substantial evidence, the majority opinion seems to be that one's actual ability in a foreign language and dreaming in that language have little correlation. Just because you dreamt in German, that doesn't mean you've achieved fluency or even proficiency in it.