In 1978, Jim Cummins, a rising education scholar, pointed out that first language (L1) development helps rather than hinders second language development (L2), because there is a common knowledge base as well as metacognitive skills that underlie both languages, which he called the Common Underlying Proficiency. This was a blow to the idea that students in the U.S. and Canada should speak English only, think in English only, and use English only when at school, and that their home languages were hindrances to English learning. However, he also made a false claim—related to the Common Underlying Proficiency—that many students from low socioeconomic status (SES) backgrounds have underdeveloped language skills in both their L1 and L2 (“semilingualism”), whereas high SES students with educational advantages often have advanced proficiency in both languages. In 2000, Jeff MacSwan wrote a paper explaining why the “semilingualism” construct, by then already 20 years old, was not only incorrect, but also had negative effects if teachers believed it, because it confused language development and literacy development.
MacSwan, J. (2000). The threshold hypothesis, semilingualism, and other contributions to a deficit view of linguistic minorities. Hispanic Journal of Behavioral Sciences, 22(1), 3-45. https://doi.org/10.1177/0739986300221001
MacSwan’s article is too long to summarize in its entirety, so I will focus on a key point in this post: the important distinction between language and literacy. Language, defined as the ability to talk to other people, is something that every child develops naturally and effortlessly. This is because language is our human inheritance. Unless you have a brain injury, or you were locked in a basement for the first decade of life like a severely abused child named Genie, with no linguistic input at all during the years when you were supposed to be developing your L1, you will have a fully developed L1 (or even L1s). As early as page 5 of the article, MacSwan asserts that “the condition denoted by the term [semilingualism, i.e. having no fully developed L1] does not exist.”
So why do people describe others (or even themselves) as being bad at, or lacking language skills, in their first language? The answer is literacy—or more aptly, LITERACIES. A literacy is a use of language that has to be learned and taught, with the help of expert teachers and a community of practice, just like playing the violin, throwing clay pots, reading MRI scans, operating a forklift, fencing, flying a plane… basically anything that not 100% of human beings naturally know how to do. Examples of literacies include: writing a rap, giving an applied linguistics conference presentation, writing a medical prescription, reciting your genealogy, making a corporate pitch, and talking to 15- and 16- year olds in their slang. It is teachers’ job to teach literacies; every teacher of every subject will teach their students various literacies.
The reason teachers think some of their students are “lacking language” (even in L1) is that schools use literacies mastered by the upper and middle classes as evidence and measurement of language development, while ignoring literacies mastered by the lower classes as evidence of the same. When people grow up being told that they are not good at language because they are not good at the specific literacy practices valued at school, they end up thinking they are not good at language, period. For example, schools use written assessments of literacy far more than oral assessments as evidence of language ability (and even the oral assessments, like debates and Socratic dialogues, are culture-specific), but not all the world’s cultures use print: it is not a necessary aspect of language, despite its being highly valued in European and East Asian cultures. In contrast, other cultures (like Pacific Islander and African cultures) have more sophisticated forms of oral literacy than Western and East Asian cultures.
Furthermore, every language is as grammatically complex and as difficult for non-native speakers to learn as any other language, whether or not it has a written form: “In the early 20th century, Boas (1911) and others painstakingly showed that non-Western languages were every bit as linguistically sophisticated and rich as their European counterparts represented in the universities” (MacSwan, 2000, p. 9). Thus, teachers have often confused literacy differences (which differ at the level of cultures, the level of communities, and at the individual level also) with degrees of ability in language.
MacSwan also points out the false assumption “that the particular linguistic register (or way of talking) that academics use has special qualities” (p. 17). This is not to say that this register didn’t develop to suit its communicative purposes… yes, indeed, it is more suitable for those purposes than other registers that serve other purposes. On the other hand, this register is NOT indicative of more clarity, more precision, more comprehensibility, or more insight. That’s in the eye of the beholder—and everyone finds those literacies that they are most familiar with to be the clearest and most precise, comprehensible, and insightful! Mac Swan jokes: “Place a Harvard professor on a farm in Central Mexico, the way a Spanish-speaking child from Oaxaca might be placed in an academic environment in the United States, and you would find the academic at a considerable loss for relevant vocabulary and speech registers” (p. 17). In addition,
We might imagine a farmer or a skilled boatbuilder, neither of whom has ever been to school. Both will know many concepts and words utterly foreign to academics, just as academics will know words foreign to them. That is no surprise; we naturally expect these differences, given the differences in experiences. … Individual and group differences in language reflect different experiences but not different levels of language development. (p. 18).
MacSwan goes on to say: “This middle-class advantage relates not to some presumed superior quality of the oral language of middle-class children, but to the special alignment of their particular home experiences and speech registers with those encountered at school” (p. 18). The linguistic anthropologist who demonstrated this most famously was Shirley Brice Heath (1983) in a book called Ways with words: Language, life, and work in communities and classrooms, by comparing the home and school practices of children from White middle-class, White working-class, and Black working-class families. Thus, MacSwan argues that the only valid evidence for semilingualism would be if a person did not acquire the literacies of the communities he/she/they grew up in. So far, we have never found such a case.
He concludes that “confounding literacy and language in this way forces a number of disastrous conclusions” (p. 23). School literacies and reading and writing are not a further development of talking in L1 or L2, any more than playing the violin or fencing is a higher level of talking. Scholars whom MacSwan cites show that “in contrast to the way in which children learn to read and write, a native language is acquired effortlessly and without instruction by all normal children. [The scholar who most famously demonstrated this was the linguist Jean Berko Gleason, with her adorable Wug Test.] Success in reading and writing, however, is dependent upon direct or tacit instruction, practice, and effort” (p. 25).
“So what about standardized test results?” people might argue. “Don’t they show differential levels of L1 development?” MacSwan responds that students’ L1 abilities “can be no more meaningfully informed by the extent to which they can read, write, or do well on tests of synonyms [i.e., the common tasks on standardized tests of language achievement in L1] than they can be informed by the extent to which children know Morse Code, can recite the words of ancestors, or are good at writing rap songs” (p. 27).
In other words, while people’s performance on each literacy varies, “all normal children effortlessly acquire the language of their own speech community” (p. 31). Yet it is middle- and upper-class people who speak the dominant societal language in certain ways who define what literacies count as evidence of general language proficiency. This puts students from other backgrounds at a decided educational disadvantage, as they have to work even harder outside school to acquire literacies they would otherwise learn subconsciously and relatively effortlessly if they were born into a different family. In fact, middle- and upper-class people who are bi/multilingual often have a higher chance of mastering academic literacies in L2 English than low SES L1 English speakers, because of privileges and resources that allow the former to harness the common underlying proficiency in learning L2 literacies. Furthermore, many middle-class academic practices (e.g., monolingual textual literacies and test-taking skills) are similar across cultures.
If these systems of disempowerment are the real sources of educational inequality, not “semilingualism,” then teachers believing in semilingualism can only disadvantage educationally marginalized students further, as these children and youth may internalize teachers’ and society’s low expectations for them. Getting rid of such ideas will not fix these students’ troubles with poverty, but it may at least rid teachers of wrong assumptions about the children they teach: “It must be shown that the language and cultural practices of reputed semilinguals is deficient, not different, and this has not been done” (Mac Swan, 2020, p. 33).
Cummins has since distanced himself from the term “semilingualism,” as many other people have criticized it besides MacSwan. However, MacSwan concludes with a caveat:
Although I have levied some very strong criticisms against the work of honest and extremely well-intentioned colleagues who are genuinely concerned with the education of language minority children, I hope that these criticisms will be understood as being made in the interest of scholarly and moral progress aimed at improving the lives of children in school. Jim Cummins, in particular, is a giant among us in his tireless defence of language minority children and is responsible for much progress in creating policies and practices that have benefited language minority children in the United States in the last quarter of the 20th century. (p. 37)
Boas, F. (1911). Handbook of American Indian languages. Smithsonian Institution, Bureau of American Ethnology. Washington, DC: Government Printing Office.
Cummins, J. (1976). The influence of bilingualism on cognitive growth: A synthesis of research findings and explanatory hypotheses. Working Papers on Bilingualism, 9, 1-43.
Heath, S. B. (1983). Ways with words: Language, life and work in communities and classrooms. Cambridge University Press.