Linguist challenges language learning assessment

In his research within the field of applied linguistics, ASU professor Jeff MacSwan recently examined a language assessment which identified some children as “non-nons” who, according to the test, know neither English nor Spanish.

“The test result is really unexpected, because linguists see language acquisition as just a natural part of being human. We acquire languages because we are biologically endowed with a language faculty,” MacSwan said. “Everybody learns a language effortlessly and without instruction, so why would we find kids who don’t learn a language at all? The consequences are huge for these kids. They very often end up in special education programs because schools don’t know what to do with them.”

Understanding this dilemma is intimately related to policy issues and educational issues facing English language learning.

“Whoever developed the tests to reach this determination has some concept of what it means,” said MacSwan, who is one of the nation’s leading scholars in applied linguistics. “We need to ask: Is it a good concept? Is it a defensible theory of linguistic knowledge or is it not?”

To challenge this popular assessment, MacSwan gave children a wordless picture book and asked them to tell the story. He wrote down their responses word for word and used a linguistic coding system to analyze the syntactic error rate and morphological error rate to find evidence of language deficiency. Instead, 98 percent of the children scored in the proficient range with the natural language sample compared to the commercial language tests that labeled 90 percent of them “non-nons.”

“The tests were dramatically out of synch with the natural language abilities of these kids,” MacSwan said. “The language sampling approach is informed by sound linguistic theory and the commercial tests were designed by people who have a very crude understanding of language proficiency.”

On the surface, MacSwan’s abstract linguistics approach to the study of bilingualism seems unrelated to his work in ESL policies and classroom practice. But his inquiry into the nature of bilingualism also informs perspective and understanding of language as he probes questions in applied areas related to policy, curriculum, teaching and learning.

“For me, they’re tightly connected because when we discuss bilingualism and language minority education, we bring along a set of assumptions about the nature of language and the nature of bilingualism," he said. "Those assumptions have very dramatic consequences on the way we think about why some students do better than others in English language learning programs.”

MacSwan has approached linguistics research through a disciplinary lens then applied it to support English language learners in schools. At ASU, his scientific work has focused on the basic linguistic structure of code-switching, in which speakers interchange words from dual languages.

“I probe the question of grammatical structure or the system of grammatical rules,” he said. “It’s been very strongly established that code-switching is governed by a system of grammatical rules just like monolingual language. The question for researchers becomes, what is the system of rules?”

“When people say a sentence is wrong, they are thinking of prescriptive grammar. We’re talking about an empirical discipline where we have evidence that a sentence is not consistent with an underlying rule system that is mentally represented in our subconscious minds,” MacSwan said. “It starts getting very, very complicated because it matters which language pairs you have and it matters what particular grammatical context is used.”

MacSwan is the program director for ASU's interdisciplinary Applied Linguistics PhD program and a member of the graduate faculties of Applied Linguistics, Language and Literacy, English, Speech and Hearing Science, and Educational Leadership and Policy Studies. He received his PhD from UCLA in 1997 and in 2003 was selected as a National Academy of Education/Spencer Postdoctoral Fellow. He also was a visiting scholar in the Linguistics Department at the Massachusetts Institute of Technology and a guest scientist at the Center for the Study of Multilingualism at the University of Hamburg in Germany.

His 2000 article on the underlying mental architecture of bilingualism and code-switching is anthologized in a prestigious international book series published by Routledge, titled "Bilingualism: Critical Concepts in Linguistics,” and has been translated and reprinted in international journals.

MacSwan’s interest in code-switching was piqued while teaching English as a second language to bilingual students at a high school in California. The experience ultimately became the subject of his dissertation.

“I saw my kids code-switching like this all the time. I had a background in linguistics, so I found it very intriguing," MacSwan said. "My focus was on immigrant kids and English language learning, and it seemed like a nice marriage to use linguistics to study bilingualism."

He began using the linguistic framework of Noam Chomsky’s Minimalist Program, which theorizes the rules that govern the structure of language and has radically changed thinking about linguistics since the 1990s. From this research, MacSwan developed a model of intra-sentential code-switching which explored consequences of Chomsky’s minimalist approach.

“One important aspect of classical minimalism is lexicalism, the idea that language differences are encoded in words rather than the syntax,” MacSwan said. “The old idea was that there was a parameter, or switch, that tells the learner what the syntactic rule is. In code-switching, that kind of a theory gets a little difficult to unravel because if you have a mixed sentence, how do you know which syntactic system is contributing the rule for basic word order?

“With lexicalism, you no longer have the idea that parameters are set in syntax as it was developed to look at monolingual language,” MacSwan continued. “There is some property of a word that has the consequence of one word order or another. For code-switching researchers, that was extremely valuable because now we can pin down what properties a word should have. Now we have a tight relationship between the language and the specific linguistic contributions it’s making to an expression.”


Written by Verina Palmer Martin