An Introduction to Applied Linguistics

Figure 6.1 MonoConc concordance display of KWIC for the target word like

---

100 An Introduction to Applied Linguistics
of these extended collocations and lists frequent three-, four- and five-word lexical 
bundle patterns by register. Without the use of sophisticated computer programs, 
these patterns would remain undetected.
Working with Tagged Texts
In order to carry out more sophisticated types of corpus analyses, it is often 
necessary to have a tagged corpus. As mentioned in the previous section, when a 
corpus is tagged, each word in the corpus is given a grammatical label. The process 
of assigning grammatical labels to words is complex. For example, even a simple 
word such as can falls into two grammatical categories. It can be a modal – ‘I can 
reach the book’. Or, it can be used as a noun – ‘Put the paper in the can’. By writing 
computer programs that include rules and probability information, computers can 
quite accurately identify the grammatical labels for many words. However, there 
are certain features that remain elusive. For these features, programs that work 
interactively with a user can result in accurate identification. These programs are 
similar to spellcheckers and bring problematic or ambiguous words to the screen 
for the user to select the correct classification. Biber, Conrad and Reppen (1998) 
provide a fuller description of tagged texts and interactive tagging.
Once texts have been tagged it is possible to explore a variety of complex 
linguistic issues. Clusters of features can be counted, thus providing a fuller picture 
of the texts in a register. Rather than information from a single linguistic feature, 
the researcher can explore how features work together in texts. For example, 
studies have shown that interactive spoken texts produced under time constraints 
have particular linguistic features associated with them which are different from 
informational texts. Interactive texts typically have more contractions and a 
greater use of first and second pronouns (for example, I, we, you, my), whereas 
informational texts have relatively few of these features.
Overview of Different Types of Corpus Studies
Over the years, corpora have been used to address a number of interesting issues. 
The question of language change is one that intrigues many researchers, teachers 
and language students. The area of historical linguistics has been well established 
in Europe, with numerous scholars carrying out extensive projects to see how 
language has changed over the centuries.
In addition to exploring changes across the centuries, scholars have used 
specialized corpora to gain insights into changes related to language development, 
both in first and second language situations. These types of studies can provide 
valuable insights as to the linguistic developmental changes that take place as 
individuals acquire their first language and also can provide important insights as 
to patterns of developmental changes that apply to different first language groups 
as they acquire a second language.
Corpora have also been used to explore similarities or differences across different 
national or regional varieties of English. Several collections of corpora that represent 
different varieties of English (Australian English, American English, British English, 
Indian English) have yielded interesting information about the systematic linguistic 
differences that occur in these different regional varieties of English.
There have been large-scale studies to explore the differences between spoken 
and written language. In addition to large-scale comparisons, there have been 

---

101
Corpus Linguistics
descriptions of subregisters, such as newspaper language, or even comparisons 
focusing on different sections of newspapers (for example, news reportage, letters 
to the editor, feature articles, etc.). Many of the patterns of language use discovered 
through corpus studies could not have been uncovered through traditional 
techniques. Prior to corpus linguistics it was difficult to note patterns of use, since 
observing and tracking use patterns was a monumental task. In addition, many of 
these findings run counter to our intuitions of how we use language (for example, 
use of progressive aspect in conversation). For instance, a quick look at most ESL/
EFL conversation textbooks will show an emphasis on the use of the progressive 
aspect. Although the progressive is more common in spoken language than in 
written, its use is relatively small when compared with simple aspect (Biber et al., 
1999).
Describing the characteristics of a particular register can often provide valuable 
resources for teachers and students. For example, MICASE, a specialized corpus of 
spoken academic language, may be used to better prepare students to meet the 
demands of spoken language that they will encounter at university. Teachers can 
use this corpus evidence to develop materials for students that more accurately 
reflect the spoken language tasks they will face in a university setting. In the 
final section of this chapter, we focus on the potential pedagogical applications of 
corpora and corpus linguistics.
How can Corpora inform Language Teaching?
The impact of corpus linguistic studies on classroom language teaching practices 
is already taking shape. The availability of corpus findings, along with the 
increased availability of tools for exploring corpora (for example, MonoConc, 
WordSmith Tools, Paul Nation’s vocabulary programs, the Lextutor website*) is a 
considerable benefit to the language classroom. Corpus-based studies of particular 
language features and comprehensive works such as The Longman Grammar of