Using a concordancer

Concordancers and corpus linguistics in vocabulary learning and teaching

Using a concordancer

Download AntConc

AntConc is a free corpus analysis toolkit developed by Laurence Anthony. It can be used for concordancing and for analysing collections of spoken or written texts.

Learning to use AntConc

These YouTube tutorials introduce the main features of AntConc:

Manual

See the AntConc manual: Concordance for detailed instructions.

Corpus

To experiment with a concordancer, you need a corpus – a collection of texts that can be searched electronically. You can download the Brown Corpus of standard American English as a text file.

Why use a concordancer?

A concordancer can help you discover the vocabulary and patterns of English that are most useful for a particular purpose. You might want to find the words most frequently used to refer to a subject, discuss a topic or express an idea. Teachers can use the same approach to identify vocabulary and language patterns that are worth including in lesson plans and practice materials.

Rather than relying entirely on intuition, a concordancer allows you to examine how words and expressions are actually used in a large collection of texts.

The best way to understand what a concordancer can do is to try one. Some online corpus tools require simple registration before you can search them.

Examples include:

  1. British National Corpus – simple search
  2. British National Corpus – full search
  3. Corpus of Contemporary American English

You normally begin by entering a word or vocabulary item into a search box, much as you would enter a query into a search engine. The concordancer then returns examples from the corpus showing how the word is used in context. Depending on the tool, you may also be able to search for longer strings of words, examine particular parts of words, or restrict your search to particular types of corpus data.

The size and composition of the corpus matter. If, for example, you restrict your search to a very small collection of texts, a particular search string may produce very few results. A larger and appropriately selected corpus is more likely to provide useful evidence about general patterns of usage.

If learners have access to the Internet, concordancing can also provide useful classwork or homework activities. Tasks can focus on vocabulary, such as words occurring in collocation; grammar, such as syntagmatic relationships; or the interaction of syntax and meaning, as in the difference between “I think” and “I am thinking”.

Concordancers can also search for strings containing more than one word. Depending on the software and the search options available, it is possible to investigate parts of words, morphemes and even particular letter sequences, such as difficult consonant clusters.

Sample tasks to complete with a concordancer

  1. [LEXICAL] Your teacher tells you that the word “sill” usually occurs after “window”. Is this true? Select a corpus large enough to provide several examples containing “sill”. If the concordancer provides an is equal to operator, use it when you want to search for the whole word rather than a string contained within a longer word.

  2. [LEXICAL] Find an English vocabulary practice book dealing with collocations, such as The LTP Dictionary of Selected Collocations, edited by Jimmie Hill and Michael Lewis. Some authors select collocations on the basis of their teaching experience and intuitive knowledge of what learners need to know. Others use corpus evidence to check how frequently particular combinations occur.

    Take some of the collocations in a practice book and test them with a concordancer. Choose an appropriate corpus and examine the individual constituents. Do the words frequently occur together? Are you being taught combinations that are relatively rare? How useful does the practice book appear to be in the light of the corpus evidence?


  3. [GRAMMATICO-SEMANTIC] Your teacher tells you that the verb have usually expresses ownership, possession or embodiment when used lexically in simple tenses, but can express several different meanings – such as experiencing, taking or eating – in progressive or continuous forms.

    Use a concordancer and corpus data to investigate the claim. Compare the results for the search strings “has” and “is having”. Does the evidence support your teacher's statement, contradict it, or suggest that the situation is more complicated?


  4. [PHONOLOGICAL] Your teacher tells you that there are only a few English words ending in the consonant cluster “pths”. Is this true? Select an operator such as contains the string when searching for a sequence of letters occurring within words.

These activities illustrate an important advantage of corpus work: instead of simply accepting a statement about English usage, learners can investigate the evidence for themselves.

Using a search engine as an alternative

It is also possible to use a large search engine such as Google instead of a dedicated concordancer. Familiarity with Google's search operators can make this approach more effective.

One possible language-awareness activity is to search for combinations of consonants or alphabetical sequences that are unlikely to occur in your own language. This can lead learners to investigate foreign-language websites and compare the possible sound and spelling patterns of different languages. Such an activity could form part of a basic language-awareness course or be used to introduce the concept of complementary distribution to undergraduate students of linguistics.

Vocabulary building with other computer utilities

Wildcard searches can also be useful when creating word-building exercises.

Older computer users may remember DOS-based WordPerfect 5.1, which was particularly popular during the late 1980s and early 1990s. It included a useful utility called SPELL.

This was the software I used when researching the lists of Final Consonant Clusters offered in the pronunciation section of this site.

The utility allowed searches of the large number of words contained in WordPerfect 5.1's spell-check files. It was also useful for solving crossword clues when the answer could not otherwise be identified.

For example, entering the search string *pth* would return words such as:

  1. depth
  2. depths
  3. ophthalmitis
  4. upthrow
  5. upthrows
  6. upthrust
  7. upthrusts

The asterisks in *pth* are known as wildcard characters. They represent any number of letters that may occur before or after the specified string, provided that the resulting sequence forms part of a word.

Question marks can be used in a similar way to represent individual unknown letters. For example, the search string ???ses asks for six-letter words ending in ses, with exactly three unknown letters preceding them. The SPELL utility returned a list of 51 such words. Crossword enthusiasts can combine wildcard characters with letters already established from other clues to narrow down the possible answers.

Corpus linguistics

Concordancing is part of the wider field of corpus linguistics, which involves using collections of spoken or written language to investigate how English is actually used.

The British National Corpus

The Collins COBUILD Project – Looking Up: An Account of the COBUILD Project in Lexical Computing, edited by J. M. Sinclair