Different methods of learning

WORD CONCORDANCE IN A TEXT
The use of concordances is essentially a manual task for human analysts, unlike the use of many computer algorithms to automatically extract information about the occurrence and co-occurrence of words in texts. Automatic extraction of wordlists, collocate lists, etc. can lead the analyst to deal only with words abstracted from the texts where they occur, and taken away from the place where meaning is created. Reading concordances means looking at words in their context of occurrence in texts, and allows the analyst to see what the meaning of the word is in the text, and to see how that meaning is created in the particular case.
Furthermore, reading concordances allows the user to examine what occurs in the corpus, to see how meaning is created in texts, how words co-occur and are combined in meaningful patterns, without any fixed preconceptions about what those units are. It can be a method of approaching the corpus in a theory-neutral way, and is what Tognini-Bonelli (2000) calls corpus-driven linguistics.
However, interesting results do not spring out as soon as the corpus is loaded into the software. To generate a concordance, the user must select what to search for, and this means approaching the corpus with some pre-conceptions about what words (or other features) will be interesting to look at. One way of avoiding this bias is to make use of a function which some programs have to provide a complete concordance of a text or a corpus. This can be useful for a text, and was the traditional way of making concordances for the study of literary or religious works before the era of the computer. However, a complete concordance of a corpus will usually produce more data than human analysts can cope with. Even major lexicographic projects are likely to be selective with what words to search for and how many examples to look at in a large corpus. There are other functions, described below, such as making wordlists, collocates and keywords, which can be used as starting points which allow the corpus to suggest things to look for and investigate.