What exactly constitutes academic-level syntax? What corpus-based data show

Ask international students coming into the university writing center what they want to focus on in their tutoring session, and many of them are likely to utter a sentence containing a word feared and loathed by writing tutors everywhere.  That word, of course, is grammar.  The many facets of grammar in ESL student writing are far beyond the scope of this blog.  Prescriptive grammar and minor grammatical errors do not concern me here.  Rather, the question I am concerned with in today’s entry is quite specific, and it’s a doozy:  from a grammatical standpoint, how do we know whether university students whose native language is not English are ready to enter an academic discourse community and participate in the conversation as full communicative partners?

According to Biber, Gray and Poonpon's (2011) article, studies of grammatical complexity in writing have long been used as a measure to assess writing development.  Grammarians and L2 writing scholars traditionally have relied heavily on a measurement instrument called a T-Unit, which measures the amount of subordination within independent clauses.  The argument goes, the more dependent clauses that occur per independent clause, the more complex the sentence, and therefore the more advanced the writer.  However, the researchers, using corpus data, show that the T-Unit may not be an appropriate measure of syntactic complexity in academic writing, since this kind of subordination is actually much more common in conversation than it is in academic writing.  Instead, the researchers argue, scholars and teachers wishing to do this type of assessment would be better served to use a different measure to determine syntactic complexity in academic writing: the proportion of dependent phrases that function as constituents of noun phrases.

(charts taken from: Biber, D., Gray, B., & Poonpon, K. (2011). Should we use characteristics of conversation to measure grammatical complexity in L2 writing development? TESOL Quarterly, 45(1), 5-35.)

The article appears to be well-researched; one of the arguments that makes using corpus data so appealing is that it is based on actual language usage, rather than native speaker perceptions about writing, which are often wrong.  One drawback is that the type of corpus-based research engaged in by Biber, et al. (2011) is quantitative in nature, which makes interpreting the data challenging at times; the sophisticated statistical instruments the authors use (and the charts generated therefrom) are unintelligible unless the reader has some background in interpreting statistics.  Nevertheless, most people understand bar graphs; the tables reproduced below clearly show that conversation and academic discourse tend to employ different types of complex structures.  The authors argue that acknowledging this fundamental difference between speech and academic writing constitutes a “weak” interpretation of their findings; a “strong” interpretation implies an acquisition order for complex grammatical structures in English, from the least complex complement clauses in conversation to the highest levels of phrasal embedding in academic writing.

The pedagogical applications of the current research are less clear than we might wish, since the authors stop short of offering recommendations on how teachers can use this research to benefit their students.  What we, as instructors of college students writing in a second language, can draw from Biber et al.’s (2011) study depends on which view of the findings we take:  if we take only the weak view, then we are merely confirming what linguists, according to Lippi-Green (1997), already know: that speech and writing are fundamentally different linguistically.  On the other hand, if we take the strong view of the current findings, we have a potential new tool in our box as writing instructors, even if it is, at the present time, unsharpened and somewhat unwieldy.

In my opinion, the utility of this tool is to help teachers gauge whether, from a syntactic standpoint, students with English as L2, whether they are in a bridge program, a basic writing class, or even a mainstream composition course, are ready to enter an academic discourse community using written English as the medium of communication.  This is not to say that it is practical, or even desirable, for teachers to employ sophisticated statistical instruments to measure the frequency of this type of construction in their students' writing.  Nor does it imply that the use of advanced grammatical structures in writing should be the only criterion for whether students with English as an L2 are ready to enter an academic discourse community at the undergraduate or graduate level. However, assuming a familiarity with the types of structures discussed here, teachers of L2 writing can at least “eyeball” their students’ compositions in order to roughly gauge whether they contain the same level of phrasal embedding as the compositions of many of their native English-speaking peers.  In other words, the applications of the research as it now stands are limited, but the implications for the future of both L1 and L2 grammar pedagogy as it pertains to college student writing are significant.