The purpose of this paper is to (1) introduce ESL instructors to two valuable corpus resources: The Michigan Corpus of Upper-Level Student Papers (MICUSP) and StringNet, and (2) share some corpus activities using these two resources. The activities were developed in an upper-level reading and composition class in a theme-based intensive language center for teaching ESL at a U.S. university. The majority of the students in this class intended to pursue graduate studies. The main writing assignments included writing an argumentative essay on a controversial topic. Students were asked to develop their arguments using logical evidence through synthesizing information from credible sources. They also needed to analyze and refute relevant counterarguments. Another writing assignment was summarizing and critiquing peer-reviewed articles. Students were to evaluate elements such as the author’s support and logic, credentials, consulted sources, bias, and the appropriateness of the vocabulary.

Because such writing assignments demand knowledge of specific language patterns, students were coached to use MICUSP and StringNet directly and independently to (1) explore the language of argumentative essays and critiques, (2) enrich their lexico-grammatical competence, (3) notice language patterns and draw conclusions about their usages, and (4) learn how to use such resources to edit their writing for grammar and syntax. In the following subsections, an overview of MICUSP and StringNet is provided followed by examples of corpus tasks utilizing these two resources.

MICUSP

MICUSP contains a collection of A-graded academic papers written by senior undergraduate and graduate students from various disciplines. It has a user-friendly interface, which makes it easy to navigate. It can be accessed directly without signing up for an account here.

A student can search for recurrent usages of a word or phrase by simply typing it in the search box. Its frequency of occurrence is provided along with statistical visualizations of its distribution across disciplines and paper types (Figure 1).

Figure 1. An example of a search word display.

Possibly has occurred 180 times in 127 papers. The pie chart shows that possibly is most frequent in Reports (45%) followed by argumentative essay (24%). Content can be restricted to student level, nativeness (native- vs nonnative-English speakers), textual features, paper type, and discipline.

Excerpts with the search word highlighted in yellow appear as a list (Figure 2). Clicking the Paper ID provides access to the complete text, which can be saved as a PDF file (Figure 3).

Figure 2. An example of an excerpt with the search word.

Figure 3. An example of the complete text and related information.

StringNet

StringNet is a searchable database of 2.2 billion lexico-grammatical patterns extracted from the British National Corpus (BNC). Students can simply enter a word(s) in the search box. They may click on “Search options” to specify the chunk length and frequency. For example, after entering a word likecontrary (Figure 4), click “Find patterns.” A page with a list of patterns in which the query word is commonly used is displayed (Figure 5). Two clickable part-of-speech options of contrary appear (Figures 5 and 6).

Figure 4. Main page of StringNet.

Figure 5. Patterns containing contrary as an adjective.

Figure 6. Patterns containing contrary as a noun.

By clicking on any word in the chunk, a small box pops up with suggested words that can replace the examined word (Figure 7). Additionally, clicking the magnifier icon under “Examples” leads to a page with a list of sentences from BNC containing the pattern (Figure 8).

Figure 7. Suggested words to replace a word in a chunk.

Figure 8. Examples of sentences from BNC with pattern in blue.

Another interesting option is “Find similar words” (Figure 9). It provides synonyms of the search word (Figure 9). Clicking on the frequency number in the “shared patterns” column displays examples of the synonym in the slot that it occupies in the chunk (e.g., reason in Figure 10).

Figure 9. Results for argument in the “Find similar words” option.

Figure 10. Examples of reason and argument displayed in chunks.

Tasks Based on MICUSP and StringNet

The following tasks were carried out online using Facebook. A closed Facebook group for the class was created. During the first week of class, the teacher introduced MICUSP and StringNet, demonstrated how they can be used, and explained the benefits of using them. A weekly corpus task was posted throughout the session in which students were asked to explore a certain construction and then respond to related questions intended to provoke critical thinking and inductive reasoning. Students had to comment on at least one of their colleagues’ findings. The tasks were graded on completeness and meeting the deadline. Examples of the tasks are shared below.

Task 1: Author

In this class, you will be using the word “author(s)” OR the last name of the author(s) “e.g., Jones (2010)” regularly when you write summaries, critiques, and argumentative essays. Your task is to investigate the frequent chunks that contain the word “author” and the words that occur with it.

1. Go to StringNet
2. Type: author
3. Hit “Find patterns”
4. Answer the following:

1. How frequently does the word “author” appear in this corpus of academic writing?
2. Find three chunks that contain “author + verb” and discuss the following:

1. In which part of the essay would you likely use each chunk (e.g. conclusion, support, examples, argument…)? Why?
2. What word(s) often follow the verb in your chosen chunks?
3. What have you noticed about the grammar of these chunks?

Task 2: Effect vs. Affect

Based on the writings that we have done so far, I noticed that some students confuse the use of “effect” and “affect.” Your task is to investigate how these two words are used in academic writing using MICUSP.

1. Go to MICUSP
2. First type “effect” and hit SEARCH
3. What part of speech often precedes and follows “effect”? Give examples from the corpus.
4. Then type “affect” and hit SEARCH
5. What part of speech often precedes and follows “affect”? Give examples from the corpus.
6. What is the difference between “effect” and “affect”? Use each one in a sentence.
   
   

Task 3: Exploring Argumentative Essays

Using MICUSP, check “Argumentative Essay” only and select an argumentative essay from any discipline. Explore the essay in terms of the following:

1. Thesis statement
1. What is the topic?
2. Is it a pro or con argument?
3. How is the thesis statement structured?
4. Is there a modal verb used in the thesis?
2. Argument
1. How does the author support the argument?
2. Can you identify any counterarguments and refutations? How were you able to notice them?

Task 4: Argumentative Essay Language

Study the underlined constructions from StringNet. Write where you might use these constructions in your argumentative essay. What do you notice about the structure of the sentence?

1. While it is true that village women in India have little power or status in society as a whole, Janet misunderstands what lies behind the caring and “fussing.”
2. Some opponents said later that since none of the houses managed a two-thirds majority, prospects for success in 1991 were thin.
3. It could be argued that the surgeon could face both criminal and civil liability.
   
   

Task 5: Similar Pairs

Using MICUSP, explore each pair in terms of usages, location in the sentence, punctuation, and the context around it. Share your findings and examples from the corpus.

• although and despite
• on the one hand and on the other hand
• such as and for example
• because and because of
  
  

Task 6: Exploring Critiques

Go to MICUSP, check [✔] “Critique/Evaluation” only and select a paper from any discipline.Identify some aspects that the writer has critiqued. This task will be discussed further in class. Please bring your chosen paper and your notes to class.

Task 7: Recurrent Errors

This assignment is based on frequent errors in your papers. Each student is assigned a particular word/phrase highlighted in the sentence to examine using MICUSP and then answer a related question. Write what you notice in a comment and provide examples from the corpus. An example:

Type “whether”: Do all sentences that have “whether” always include “or not”? Explain.

Other Tasks: Self-Editing

Some tasks were tailored according to the individual needs of each student. When giving feedback on the first draft, a teacher can highlight recurrent errors, jot down some questions that scaffold the student to notice and think about his or her errors, and direct them to StringNet to explore correct usages.

Ultimately, such resources are valuable for supporting ESL learners in noticing and discovering language use and promoting learner autonomy.

References

O'Donnell, Matthew B. & Römer, Ute. 2012. From student hard drive to web corpus (part 2): The annotation and online distribution of the Michigan Corpus of Upper-level Student Papers (MICUSP). Corpora 7(1): 1-18. pdf version

Römer, Ute & Matthew B. O'Donnell. 2011. From student hard drive to web corpus (part 1): The design, compilation and genre classification of the Michigan Corpus of Upper-level Student Papers (MICUSP). Corpora. 6(2): 159-177. pdf version

Eman Elturki has recently earned a PhD in language, literacy and technology from Washington State University. She teaches ESL at the Intensive American Language Center of Washington State University.