How Bilinguals Move Between Two Grammars

Imagine you are looking at a computer screen that shows two cows on one side and one dog on the other. You hear a sentence, and your job is to pick the one doing the action. The sentence is not quite right, on purpose: “The cows is kicking the dog.”

Who is doing the kicking?

As an English speaker, you would choose the cows, because it's the first noun. A Spanish speaker hearing the sentence in Spanish would choose the dog, because the verb is singular, and in Spanish that points to a single noun.

The grammar does matter to English speakers. Even though they choose the cows, it takes them a little longer. Verbs matter in English. They just matter less than where the noun is.

Evelyn Rodarte is a heritage speaker, someone who speaks one language at home, through her family, and learns another at school, usually the language of the wider society. Evelyn grew up in Los Angeles, the daughter of Mexican immigrants. She joined my laboratory as a PhD student four years ago. Her paper came out on October 14. It used the same task I had first encountered as a graduate student.

The task was developed by Elizabeth Bates and Brian MacWhinney. I first learned about it in 1991. Liz, my doctoral mentor at UCSD, had a testing room at the VA. One day we went to the VA lab, and they showed me the experiments they ran with aphasic patients, people whose language had been impaired by a stroke.

One experiment was a weird setup. Pictures of two animals appeared on the screen as a voice named them out loud: "cow," "dog." Then a sentence would play, such as "The cow is kicking the dog." Some sentences were normal, like that one. Others were strange, like "The cow the dog is kicking." I sat down and tried it myself. Some seemed easy. Others got really hard, and I just had to choose one without knowing why. It was a bit stressful.

After about 10 minutes, I asked, "How on earth do aphasic patients do this?"

The research assistant said, "Oh, they do really well. They actually really like it." Liz and Brian had tested native speakers of many different languages. They found that even aphasic patients still followed the main rules of their language.

When I was 20, I went to Brazil to study abroad for a year and ended up staying two. Coming back to the U.S., I read slowly. Two years later, I was at UCSD trying to do the sentence interpretation task.

It was only later that I understood why it felt so hard. In 1994, Liz, Luis Avila, and I published our version of the experiment with Spanish-English bilinguals, what we would now call heritage speakers. Compared to monolingual Spanish and English speakers, they fell in between. Two of my languages were pulling me toward verbs and one toward order. But we had treated the bilinguals as a single group.

Decades later, when Evelyn came to my laboratory, she was very interested in language proficiency. Most proficiency measures in psycholinguistics look at vocabulary. A vocabulary test only captures the words a person knows. It doesn't capture how they use them. I mentioned the sentence interpretation task. The nice thing about it is that there is no wrong answer. We can measure how Spanish-like or English-like a person is. Evelyn immediately saw that we could use it to test individuals rather than groups. Maybe proficiency in each language would be related to where a person fell.

But it had been 30 years since I first ran it. I contacted Bob Buffington, the programmer who had worked for Liz all those years earlier. He recovered some recordings, but they were not the right ones. We thought about rerecording them but settled on generated voices. Thanks to technology, we could now have the perfect bilingual: the same voice in both languages, sounding like a native speaker of each.

Armed with the new materials and our proficiency measures, Evelyn designed the study and gathered the data. The results were very similar to the ones we had seen in 1994. As a group, bilinguals appeared to blend the strategies from both languages.

But there was more. Evelyn's data showed us individuals. English monolinguals clustered tightly around word order. The bilinguals were spread out, and most of them shifted toward the verb when the sentences were in Spanish.

The clearest predictor of that shift was Spanish vocabulary. Bilinguals who could name more pictures in Spanish relied more on the verb when they heard Spanish. When they heard English, their Spanish vocabulary made no difference. Vocabulary, the measure Evelyn wanted to look beyond, tracked how each person built Spanish sentences. My Nguyen, who had been studying heritage speakers and other bilinguals in a large sample of U.S. adolescents, added a number of analyses that teased out how vocabulary proficiency might link to this grammatical processing. Her work helped bring this point home.

English is the societal language, the language heritage bilinguals are educated in. It is usually the language in which they are more proficient. Spanish depends on how much they are exposed to it. That is why Spanish vocabulary relates to their decisions about grammar.

The attrition literature usually shows that people who become immersed in a second language begin to lose abilities in their first. Many heritage speakers speak Spanish at home. As they are exposed to English at school, their Spanish might appear to be getting worse, and people have called them attriters.

When I came back from Brazil, reading in English was slow and laborious. Eventually it gained speed again. A year into graduate school, I was reading one day and stopped mid-sentence. I had been reading so fast that I had taken halves of words from two different lines and made them into one word. My English had come back, better than it was before.

Evelyn's results point the same way. Bilinguals are managing both strategies.

Technology had allowed us to create the perfect bilingual. But this bilingual was not human. It was native in both languages, and neither one ever leaned on the other. The bilinguals Evelyn tested showed the flexibility that psycholinguist François Grosjean described in 1989, when he wrote so eloquently, "the bilingual is not two monolinguals in one person."

References