Section Three

Writing (and Writing Instruction) Benefits from Attention to

Language-Level Features
All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

Language-level features generally receive little attention in current discussions of students’ writerly
development, and they are not listed among threshold concepts. This is not altogether surprising given the
complicated relationship between writing studies and linguistics. Yet it is worth pausing to consider why a
language-level approach to writing is not seen as foundational. After all, as Charles Cooper claimed about
strategies such as sentence-combining and the generative rhetoric of the sentence, “No other single
teaching approach has ever consistently been shown to have a beneficial effect on syntactic maturity and
writing quality” (72). George Hillocks, after an extensive review of research on language-level pedagogies,
affirmed and amplified Cooper’s claim. Study after study showed that students’ writing improved when
they received language-level instruction. During the period between the mid-1960s and the early 1980s,
teaching approaches that drew on linguistics enjoyed wide approval. Articles on language-level approaches
appeared regularly in journals, and several conferences on sentence-combining were held at Miami
University.
A quick look at the history of writing studies provides one explanation of why such successful
instructional practices would have been banished from the field. In the mid-1960s, as the field of writing
studies or composition and rhetoric took shape, the generative rhetoric of the sentence as articulated by
Francis Christensen received a great deal of attention for its capacity to help students create periodic
sentences. This was followed by sentence-combining, which Kellogg Hunt and John Mellon showed to be
highly effective in enabling students to produce more complex syntax. The work of linguist Noam
Chomsky provided a theoretical basis for sentence combining, lending it further stature, and many in the
field received training in or claimed affinity with linguistics. Tagmemics, introduced by Richard Young,
Alton Becker, and Kenneth Pike, added a rhetorical dimension to a language-level approach, but it proved
difficult to teach. As language-level writing instruction was taking shape, the field of writing studies began
moving in another direction. Proponents of a process-based approach to writing, such as Janet Emig,
Donald Murray, and James Moffett, criticized language-level approaches for removing context from
writing. Their desire to push language-level approaches aside was made easier by the turn of linguistics
toward ideal rather than actual users of language. To the extent that linguistics or language study received
attention from scholars in writing studies after the early 1980s, it took the form of a sociolinguistic focus
on issues surrounding African American vernacular as articulated by Geneva Smitherman, or discussions
Copyright 2019. U OF M DIGT CULT BOOKS.

of writing instruction for students who use English as an additional language by scholars such as Ilona
Leki, Alister Cumming, and Tony Silva.
Robert Connors offers another explanation for the decline of language-level writing instruction. He
argues that critics framed language-level approaches as formalist, behaviorist, and empirical, setting them
in opposition to more holistic, student-centered, and process-focused strategies. Connors goes on to claim
that the reason colleagues in writing studies became so actively antiformalist, antibehaviorist, and
antiempirical was because they found their “departmental home in the same place its primary course
identity—first-year composition—resided[; . . .] the graduate students after 1975 who would make up the
EBSCO Publishing : eBook Open Access (OA) Collection (EBSCOhost) - printed on 4/13/2022 1:29 AM via EBSCO OPEN SEARCH
AN: 2033512 ; Anne R Gere.; Developing Writers in Higher Education : A Longitudinal Study
Account: ns316374.main.eds 94
 core of composition studies were, for better or worse, English graduate students, and they would go on to
become English professors” (121). The current number of writing departments now separate from English
departments raises questions about Connors’ claim, but English departments still house many scholars of
writing studies. Given that current language-level approaches to writing instruction call on computer-aided
approaches, and given English departments’ increased interest in the digital humanities, there may be
reason to think that attitudes and perspectives in today’s English departments have shifted.
In any event, it is time to bring language-level approaches back into the field of writing studies, and the
two chapters in this section demonstrate what our field can learn from doing so. During the time that we in
writing studies have been looking elsewhere, linguistics has developed theories and practices that can
inform our approaches to writing. Among other things, helping students develop an understanding of how
certain features of language can shape the larger effect of a given piece of writing gives them a
metacognitive perspective that may be easily transferred to other rhetorical contexts. Language-level
attention to writing can help address common writing problems such as overgeneralization, use of
ineffective words, and the struggle to assume an authoritative stance. Moreover, language-level approaches
can be effectively combined with the more familiar rhetorical genre studies (Miller; Russell; Devitt).
Linguistics has also developed new methods of analysis using computer technologies that reveal patterns
of language use that are impossible for an individual reader to discern. Like Franco Moretti’s distant
reading, computer-aided analysis or corpus linguistics makes different, large-scale, aspects of writing
visible. Corpus linguistics begins by creating a collection or corpus of texts in a principled way, such as
copies of student essays from the same course or institution, and then, often, comparing this corpus with
another, perhaps student essays written by a different group of students or an established reference corps
such as the Contemporary Corpus of American English or the Michigan Corpus of Upper-Level Student
Papers. In a comparative study, a tool or concordance such as AntConc (Anthony) or WordSmith Tools
(Scott) identifies keywords or words that are distinctively salient in each corpus, and statistical analysis
determines if the differences between the two sets of keywords occurred by chance or were significant in
the statistical sense. In some cases, a corpus is divided into subcorpora to examine differences such as
gender or major in a large body of writing. A concordance can also generate lists of collocations or groups
of words that occur together in a given corpus. Reflecting important patterns and distinctions across texts,
both keywords and collocations can show textual choices that have a considerable effect on entire pieces of
writing.
In recent years, corpus analysis has been used to study first-year student writing (Aull; Aull and
Lancaster) and writing constructs valued in US composition instruction (Dryer). Corpus studies show that
teachers and students often respond to patterns that are tacit and invisible, and they can reveal unconfirmed
intuitions about writing. For example, Zak Lancaster’s recent corpus study tests the intuition-driven
writing templates in the widely used textbook They Say, I Say. He finds that the templates do not reflect
discourse practices of published academic writing; for instance, the templates encourage students to
directly entertain objections, using phrases such as many will probably disagree, while academic articles
and essays instead favor indirect phrases such as it could be argued that (251).
Both of the authors in this section analyzed corpora or collections of texts written or spoken by students
in our study using a concordancer. Each of the authors addressed a different research question, but they
both used similar methods in that they created corpora, used a concordance to identify words or phrases
that were distinctive in two or more corpora, and compared the results with another corpora. Applying this
method of analysis to various collections of student writing, sorted by categories such as gender, minor or
nonminor, major, or student level in college can show how language-level choices contribute to much
larger effects.
Laura Aull’s chapter takes on the problem of overgeneralization, as it frequently appears in the writing
of relatively inexperienced students, signaled by words such as every, always, and people, among others.
These uses are analyzed alongside assertions of certainty. By considering the appearance of such language
in writing sorted by genre, discipline, and student level, she provides insights into the writing development
of seven students in our study based on analysis of the entire collection of their writing across their
undergraduate years. She begins with the Directed Self-Placement (DSP) essays written by each student as

EBSCOhost - printed on 4/13/2022 1:29 AM via EBSCO OPEN SEARCH. All use subject to https://www.ebsco.com/terms-of-use 95
 part of their matriculation into the university, and she concludes with papers written in the senior year. As
she looks at each category of student writing, she compares it with a similar corpus. For example, in
analyzing the DSP essays, she compared her sample with a larger corpus of DSP essays collected between
2009 and 2013. In analyzing more advanced student writing she uses the Michigan Corpus of Upper-Level
Student Papers for comparison. For the most advanced student writing the Contemporary Corpus of
American English, which includes published academic writing, served as the reference corpus. While not
exactly parallel to the student writing under consideration, these three corpora provide a means of
comparing students’ use of language that marks generality and certainty across three levels of writing.
In addition to words that mark generality, Aull’s study focuses on qualified generalizations such as
almost all, virtually every, and some people, along with lists of hedges, words such as perhaps,
approximately, and plausible that qualify claims, and boosters, words such as conclusively, extremely, and
doubtless that amplify claims to show how students’ use of such language shifts across their undergraduate
writing careers. Not surprisingly, generality markers along with boosters appeared most frequently in the
writing of first-year students, and advanced student writing showed more hedges and qualified
generalizations. These variations suggest that the epistemic stance created by the languages of generality
and certainty serves as an indication of broad patterns of writerly development.
To show further nuances in the ways students use markers of generality and certainty and to deepen her
analysis, Aull considers subcorpora of particular genres; discipline-specific texts; three divisional groups
of humanities, social sciences, and natural sciences; and student and professional writers. This analysis, in
addition to that focused on the full corpus, makes visible the differences between the writing students do
when they arrive at the university and when they leave. Aull’s analysis implies, among other things, that
genre and discipline can have a shaping effect on language-level choices made by student writers, and
these choices, in turn, contribute to rhetorical constructs such as audience and purpose.
Zak Lancaster also employed automated text analysis, but instead of focusing on student writing he
examined what students say in interviews about the metalinguistic constructs of style and voice.
Specifically, he analyzed responses to questions raised in 131 interviews about students’ views of “good”
writing as well as their writerly identities and goals. Using the concordancing software AntConc, he
identified 454 instances of students using style and voice in responding to interview questions. After
differentiating what he calls the individualist (expression of the writer’s unique inner self) and social
(culturally embedded language performance, sometimes called stance) views of voice, he reviews
scholarship on the various ways these constructs have been described, asserting that most scholars who
take one view or the other would probably agree that “voice is best understood in dialogic terms, through
specific discoursal interactions” (p. 167). Lancaster does not review all the scholarship on style, but he
notes a similar division between viewing style as prose unique to an individual author and style as more
socially constructed. Most importantly, he observes that writers draw on linguistic resources to construct
community-valued voices (p. 167), thereby making clear the need to learn more about what resources
student writers draw on and how they deploy them.
Combining two groups—minors and nonminors—Lancaster reports that style appears more commonly
than voice in the responses of the entire group, and he suggests that its frequent appearance may result
from the multiple meanings students attach to it. These meanings include style as individual language use,
as register, as genre, and as usage conventions. Significantly, minors and nonminors apply specific
meanings at different frequencies, with nonminors increasingly referring to individual style, perhaps using
it as a substitute for voice, and minors describing style as register. Lancaster observes that the minors’
description of style as register is “more congruent with current theoretical conceptualizations” (p. 171).
This observation suggests that the curriculum experienced by minors may have enhanced their ability to
think about style in more complex terms, especially since minors referred to voice more than twice as
much as nonminors during their entry interviews. Without the minor curriculum, this group might have
continued to focus on voice in relatively simplistic terms.
With regard to voice Lancaster found that individualist terms tended to surface in discussions of specific
forms such as assignments, while students used social terms in more general discussions of writing goals.
Both minors and nonminors described voice less frequently in individualist terms as they moved toward

EBSCOhost - printed on 4/13/2022 1:29 AM via EBSCO OPEN SEARCH. All use subject to https://www.ebsco.com/terms-of-use 96
 graduation and increasingly framed it in social terms. Still, though, minors referred frequently to voice in
their entrance and exit interviews, and it was usually represented in individualist terms, while their
references to style usually emphasized social terms. From a developmental perspective, one of the most
interesting findings with regard to students’ use of both voice and style is the extent to which students
move between individualist and social meanings for both, sometimes in the same sentence. This fluidity
suggests that even graduating seniors are still developing concepts of writing and of themselves as writers.
Lancaster’s analysis of two individual selections of writing offers an intriguing discussion of the
resources and strategies these students call on to create a writerly self-image. Through careful reading and
identification of both linguistic and rhetorical features, Lancaster shows how these student writers establish
roles for themselves and their readers at the same time that they enact some of the meanings of style and
voice that they articulated and/or contradicted in their entrance and exit interviews. Through this analysis,
Lancaster further demonstrates the enormous complexity that surrounds students’ negotiations with taking
a stance in writing.
Together these two chapters demonstrate how much we can learn about writerly development by
looking through a language-level lens at both the writing students produced and what they say about
writing and themselves as writers. By looking at the large patterns made visible with automated text
analysis, we can begin to understand how developing writers actually progress from one type of writing to
another because language-level changes contribute so much to the overall effect of a piece of writing.
Features such as boosters and hedges may seem relatively incidental, but as Aull shows, they help shape
the stance of the writer regarding the extent to which the writing does or does not conform to the
discourses of the academy and/or confer authority on the writer. If a goal of college writing is to guide
student writers to take up academic discourses, the tools of corpus linguistics can indicate the
developmental levels achieved by a group of students, particularly with regard to a threshold concept such
as writing is a social and rhetorical activity. Students’ use of hedges and boosters can, among other things,
provide an indication of students’ rhetorical sensibilities as well as the extent to which they seek to engage
the reader in dialogic terms.
Similarly, linguistic analysis of the language students use to talk about their own writing, especially if
done in concert with analysis of samples of their writing, can provide insights into their understandings of
writing. As Lancaster showed, the terms students use and the meanings they apply often vary from those
used by professionals in writing studies, but nonetheless they offer a window into the thinking that
contributes to writing choices. Of course, this does not mean that there is always symmetry between what
students say about writing and what they actually write, as Lancaster’s analysis of Joe shows. Even though
Joe talks about voice in individualistic terms, he uses patterns of language and rhetorical moves that invite
interaction with the reader.
Contradictions such as these point to another aspect of language-level analysis. Even though large-scale
patterns and smaller-scale examination of linguistic features can provide some insights into students’
writerly development, it is never steady or uniform development. As the variations between minors and
non-minors, different genders, and lower and upper division students, to say nothing of the variations
within groups, show language-level development is irregular and dynamic; we cannot point to stages or
levels of linguistic development in student writers. But we can point to the value of giving students
opportunities to develop metalinguistic awareness. Such awareness can lead to productive discussions
about writing that capitalize on students’ uneven and shifting writerly development.

Works Cited

Adler-Kassner, Linda, and Elizabeth Wardle, editors. Naming What We Know: Threshold Concepts of Writing Studies. Utah
State UP, 2015.
Anson, Chris, and Jessie L. Moore. Critical Transitions: Writing and the Question of Transfer. CSU Open Press, 2017.
Available at wac.colostate.edu/books/ansonmoore/

EBSCOhost - printed on 4/13/2022 1:29 AM via EBSCO OPEN SEARCH. All use subject to https://www.ebsco.com/terms-of-use 97
 Anthony, Laurence. AntConc (version 3.4.3) (computer software). Waseda University, 2014.
Aull, Laura L. First Year University Writing: A Corpus-Based Study with Implications for Pedagogy. Springer, 2015.
Aull, Laura L., and Zak Lancaster. “Linguistic Markers of Stance in Early and Advanced Academic Writing: A Corpus-Based
Comparison.” Written Communication, vol. 31, no. 2, 2014, pp. 151–83.
Chomsky, Noam. Aspects of the Theory of Syntax. MIT Press, 1969.
Christensen, Francis. “A Generative Rhetoric of the Sentence.” College Composition and Communication, vol. 14, no. 3. 1963,
pp. 155–61.
Connors Robert J. “The Erasure of the Sentence.” College Composition and Communication, vol. 52, no. 1, 2000, pp. 96–128.
Cooper, Charles R. “Research Roundup: Oral and Written Composition.” English Journal, vol. 64, 1975, pp. 72–74.
Devitt, Amy. Writing Genres. SIU Press, 2004.
Dryer, Dylan B. “Scaling Writing Ability: A Corpus-Driven Inquiry.” Written Communication, vol. 30, no. 1, 2013, pp. 3–35.
Emig, Janet. The Composing Process of Twelfth Graders. National Council of Teachers of English, 1971.
Hillocks, George, Jr. Research on Written Composition: New Directions for Teaching. National Council of Teachers of English,
1986.
Hunt, Kellogg. “Teaching Syntactic Maturity.” Applications of Linguistics, edited by G. E. Perren and J. L. M. Trimm.
Cambridge University Press, 1971.
Leki, Ilona, Alister Cumming, and Tony Silva. A Synthesis of Research on Second Language Learning in English. Taylor and
Francis, 2008.
Mellon, John. Sentence-Combining: A Method of Enhancing the Development of Syntactic Fluency in English Composition.
National Council of Teachers of English, 1969.
Miller Carolyn. “Genre as Social Action.” Quarterly Journal of Speech, vol. 70, no. 2, 1984, pp. 151–67.
Moffett, James. Teaching the Universe of Discourse. Houghton Mifflin, 1968.
Moretti, Franco. Distant Reading. Verso, 2013.
Murray, Donald. “Teaching Writing as Process Not Product.” The Leaflet, Fall 1972; rpt. in The Essential Don Murray: Lessons
from America’s Greatest Writing Teacher, edited by Thomas Newkirk and Lisa C Miller. Boynton/Cook, 2009.
Russell, David. “Activity Theory and Its Implications for Writing Instruction.” Reconceiving Writing, Rethinking Writing
Instruction, edited by Joseph Petraglia. Taylor and Francis, 1995, pp. 51–78.
Scott, M. Wordsmith Tools (version 6.0) (computer software). Oxford University Press, 2014.
Smitherman, Geneva. Talking and Testifying: The Language of Black America. Wayne State UP, 1986.
Young, Richard, Alton L. Becker, and Kenneth L. Pike. Rhetoric: Discovery and Change. Harcourt, 1970.

EBSCOhost - printed on 4/13/2022 1:29 AM via EBSCO OPEN SEARCH. All use subject to https://www.ebsco.com/terms-of-use 98
 Chapter Five

Generality and Certainty in Undergraduate Writing Over Time

A Corpus Study of Epistemic Stance across Levels, Disciplines, and Genres

Laura L. Aull

Consider the following sentences written by incoming college students:

Since the beginning of time, people have feared new technologies.

Our culture today stresses perfection and nobody wants to be less than perfect.
Pressure always brings out the worst in people.
Every mature, established person has control over what they say and do.

Sentences such as these provoke the common lament that students overgeneralize, making claims they
cannot support, or worse, intimating a presumptuous, homogeneous view of the world. The sentences
apply observations and predictions across time, contexts, and people— suggesting not only that pressure
brings out the worst in people, for example, but that it always does.
Of particular concern are students transitioning into higher education. Instructional materials from
across the United States warn new college writers against generalizations as a fallacy. In just two
examples, Bowdoin College and the University of North Carolina (UNC) cite generalizations as among the
most common characteristics of ineffective introductions: Bowdoin warns students against “absurdly
general phrases” such as “humans have always . . .”; UNC encourages students to avoid openings
addressing “human beings” writ large.1 These recommendations indicate that such generality is viewed as
imprecise, unnecessary, or not credible to instructors. Instead, as noted in the Writing Program
Administrators Outcomes Statement, students must learn to craft credible, “appropriately qualified”
claims.2 Due to the importance of college writing in educational access and attainment, these expectations
are crucial for traditional and nontraditional students alike (Berlin; Penrose). They especially matter in a
time of divisive and uncivil media discourse in light of the ethical implications of generalizing, since
responding effectively to diverse perspectives requires global citizens who can distinguish when views
apply across people and contexts and when they do not.
Curiously, generalizations are addressed in instructional materials far more often than in research. The
research we do have suggests that many students either perceive they should generalize or struggle to
avoid generalizing relative to more experienced writers. For instance, Ellen Barton’s discourse analysis of
professional versus undergraduate essays shows that the greatest distinction between the two is
generalizations. All of the student essays in the study contain generalizations, not only in introductions but
also, in “striking contrast” to the professional essays, in idea development within body paragraphs (763).
Corpus analysis of writing by students transitioning into university likewise shows that incoming

EBSCOhost - printed on 4/13/2022 1:29 AM via EBSCO OPEN SEARCH. All use subject to https://www.ebsco.com/terms-of-use 99