It is not immediately obvious how or why a concordance program should be used in language teaching. At its heart, a concordancer is a simple program which allows searching of large text files for a word or phrase. So why use a glorified search program in the language classroom? The answers to this question given here are all variations on a theme, namely that it is important for students to work with instances of actual language usage. This sounds all well and good in the abstract, but the problem of working with actual language is that any lexis or construction that is of interest will occur only rarely and be obscured by masses of other language data. How do we create a focus on some particular input? This is where the concordancer shows its mettle. A concordancer is an excellent tool that makes "working with the language" possible; it allows the user to work with actual data and at the same time obtain a "concentrate" of the items under investigation. Furthermore, a crucial difference between a concordancer and a general search program lies in the possibilities, discussed below, for sorting the results of the search in order to more easily observe patterns in the data.

Let us start off with a well-accepted idea: context in all its forms is a good thing. Narrowly defined, context is equivalent to co-text, i.e., accompanying words, and an even narrower notion of co-text is collocations, i.e., recurrent word associations such as a good thing. Knowledge of a wide range of collocations is now seen as an important part of competence in a language, and it is actually quite difficult to get access to collocations without using a text analysis program such as a concordancer. Wider notions of context such as the type (genre) of the text, or the purpose of the text, or the setting for a text will not be discussed here.

Everyone is in favour of context; whether it is the use of context to "guess" unknown words or to help determine the appropriate circumstances of use for a word, phrase or construction. One important advantage of using a concordancer is that context is always present. When a word (or phrase) is searched for, the results always include the context of that word. Thus if we search for a word such as since in newspaper texts, we get results in the KWIC format shown below. To save space only a small selection of instances are shown and, here, the number of characters surrounding the search term has been greatly reduced to allow the displays to fit in the two-column newsletter format.

While the search results in this format reveal something about the nature of since, notably the co-occurrence of dates, it is worthwhile looking for patterns (or having students look for patterns) of language use, i.e., phrases or constructions based on since. A simple but telling reorganisation of the concordance lines involves re-sorting the lines in alphabetical order of the word following since. This operation, known as a right sort, leads to the output shown below.

Using this data students will see not only the use of since with dates, but also the more common use with a descriptive nominal or with a verbal phrase referring to an event: since his student days, since the Comedy Store opened. It may also be useful to draw students’ attention to the use of since in which the anchoring event is less clearly defined and may simply be implied. This use is often associated with the present perfect and is represented in this sample by the phrase a party chief who has since been redeployed. (This use can be found by using a left sort and looking for instances of since preceded by has/have.)

This focus on the association of since with temporal reference points then leads naturally to an examination of other uses of since including the phrase ever since and the logical connective since. In the case of the latter, attention can be drawn to the comma preceding since in this use.

In presenting these examples, I have not discussed how these activities are to be executed, but I assume that however the data is presented to the students, their attention will be focussed through appropriate questioning on particular phrases in the concordance results and that as a result of following a sequence of questions, students will come to understand for themselves several uses of since. In providing the nineteen examples shown above, I did not delete any examples, but culling of examples will sometimes be necessary to tailor the data to the students’ needs. (And here I am simply assuming that the corpus itself is suitable in its form and content for the students.)

As a final exercise on this topic, we can easily produce a reconstruction exercise based on for and since in the following manner. Search first for since and stop the search after getting twenty or so examples, then perform a search append for for and halt the search after about fifty examples. The search append command adds the results of the second search to the existing concordance lines. Remove unwanted items such as instances of for that do not indicate a duration of time. Next, perform a right sort in order to intermix the instances of for and since. Then choose "Conceal hits" from the Display menu of MonoConc and print the results---shown below---or save them to a file. In this format the students have only the context to guide them and from this they must reconstruct the missing word and so determine whether for or since (or both) can fit in the gap.

The example above indicates an inductive–or what Tim Johns calls a data-driven–approach to language learning. Students are presented with data, i.e., concordance output, and are guided by questions that direct their attention to certain aspects of that data in order to help them understand the usage of the term since and at the same time increase their "feel" for the common patterns of the language, i.e., the way that things are commonly expressed in the language.

In the above example, all uses of since were presented at the same time. Alternatively, the teacher could produce a series of worksheets in which the different uses of since are separated and are presented to the student sequentially. In this mode, the teacher would first select the temporal uses of since, and then move on to other uses.

While I have a preference for an inductive language learning methodology, data from concordance output can also be used in a deductive teaching style. The teacher can present a pattern or rule and the student can check the data for instances, counter-examples etc. This approach does not allow the students to construct their own rule based on their experience of the data, and so is not likely to be as beneficial. However, one deductive approach that might be interesting is to present a grammar rule taken perhaps from the student’s coursebook, and have the student work with a few examples to assess the goodness of fit between the rule and the examples. This can be followed by a brief discussion of those examples which do not fit the rule, if the examples are enlightening.

It is impossible to talk about language and culture in a paragraph or two, but I will recklessly forge ahead and present some data that may be useful in stimulating ideas concerning corpus-based explorations of cultural themes. The simple notion introduced here is that one way to get a handle on some cultural associations is by searching for colour terms and observing what collocates commonly occur with these terms that might furnish clues to cultural connotations.

First, let us look at the instances of red in the Arts section of the Independent newspaper. Some results are shown below. While these results are not culture-free (note the standard "traditional" red telephone kiosk), many of the examples refer straightforwardly to the colour red

Turning to the business section of the newspaper and searching for red, we get a much more culturally interesting set of uses: red ink, red rag, and red tape. I chose to give examples from the Arts section and Business section to illustrate the variation that can be found with different text types, but the main point I want to make is that a search for colour terms in a variety of texts can be used to focus attention on cultural aspects of language.

This article is too limited to give more than a flavour of teaching with a concordancer like MonoConc. (I have not even mentioned collocate frequency or word frequency calculations.) For a more thorough account of the use of concordancing in the classroom, the best source is Concordances in the Classroom by Chris Tribble and Glyn Jones, which has been revised and re-published by Athelstan.