The University Record, January 11, 1999

Project probes the mysteries of academic language

By Rebecca A. Doyle

Tape recorders have been softly clicking on all over campus for the past year, gathering up words to be studied by the researchers involved with the Michigan Corpus of Academic Spoken English (MICASE) project.

Project members have collected and transcribed samples of academic spoken English through recordings of as few as 20 minutes from a small biology class activity to a two-hour philosophy seminar that netted more than 20,000 words.

Their goal is to collect, stratify and analyze speech in an academic setting in order to determine whether current grammar and vocabulary texts truly reflect speech patterns in such a setting, or are taken from written text. No such database exists presently, a situation that Sarah Briggs, co-director of the project and social science research associate at the English Language Institute (ELI), says is most interesting. Briggs hopes that data collected and analyzed by the project will help improve the reliability and validity of comprehension testing at ELI.

ELI director John M. Swales, professor of linguistics; Rita C. Simpson, ELI research fellow and project director for MICASE; and research assistant Janine Ovens explain the project and its importance.

MICASE wants to track changes in speech patterns as faculty, staff and students gain experience in a university culture, which may be evidenced by the differences between speech patterns found in a freshman seminar and those found in a dissertation defense, or between junior faculty and senior faculty members.

Recordings to date cover a wide range of topics and include a lecture by the provost, academic advising sessions, a music dissertation defense, an immunology lab meeting, hydraulics problem-solving lab section and astronomy peer tutoring session. More than 355,000 words have been recorded and transcribed, about 20 percent of the total goal, Simpson says. ELI has committed resources to the project, as has the University Library. A search engine to be constructed by Chris Powell of the Humanities Text Initiatives section of the Library will be used to find occurrences of specific words.

For instance, notes Simpson, in speech, Americans use contractions often. “When are non-contracted forms used?” she asks. “Do they give the word more emphasis?”

If that is the case, some non-native speakers of English may use the full form because it is easier to pronounce and give it a meaning they don’t intend,” Briggs says.

Swales notes that one of the current applications is an analysis of the word “just” in spoken language. “Its softening use is not well understood,” Swales says. Another part of the project focuses on understanding the meanings of similar words in an academic setting, and the adjectives that consistently are used with them. For instance, “problem,” “issue,” “concern” and “question” all are used to mean the same thing, but have subtly different meanings and occasionally are replaced in an academic setting by a temporarily popular substitute, such as “challenge.”

Simpson says that certain adjectives are paired with nouns to give subtly different meanings to words that are listed as synonyms. “They may be synonyms, but they are not interchangeable,” she notes. As examples, she listed “vital,” “important,” “crucial,” “major” and “significant.”

Applications for findings of the MICASE project may include the following, but would not be in practice for some time, the researchers note:

• A training vehicle for linguistics and other students in transcription and the analysis of spoken discourse.

• A source for ESL listening tests.

• A source for new pedagogical grammars of spoken academic English and other kinds of textbooks and study-aids.

• A database that non-native speaker academecians and graduate students could use to check such things as phraseological patterns.

• A resource for those involved in translating academic speech from other languages into English, or vice-versa.

• A resource providing flagged examples of what appear to be successful examples of spoken communicative events (such as a “good” anecdote, or a “good” defense of one’s position) that can then be used for modeling.

For more information about the project, its researchers and the database, visit the MICASE Web page at www.lsa.umich.edu/eli/micase/micase.htm.


You can always drop us a line: urecord@umich.edu.