2.4 Summary

The qualitative analysis clearly shows some of the problems of the search with Google. The results retrieved with the advanced search option have to be dealt with most carefully, as there are far too many distorting factors which influence the search. When Google displays 11,600 hits for lest he be, there are four primary reasons why this does not mean that all the results can be used and worked with:

1. There are numerous duplicate documents, where the same construction can be found on different pages.
2. Although some pages may contain the exact words of the search term, these may be used in a different way so that their meaning differs from what the construction under investigation aims to express.
3. Spelling errors, especially in blogs and forums where everyone can post an entry, are another problem: sometimes people forget letters or use abbreviations so that these entries are displayed among the search results although they have nothing to do with the construction which is actually being examined.
4. The user essentially has no way of knowing according to which criteria Google selects its results. Its search-technique is widely non-transparent, which leads to a number of unanswered questions: How does Google know what the most relevant results are and according to which pattern does Google pick these results?

Therefore, although the Web-as-corpus approach does prove a helpful tool in terms of providing general information about certain word combinations, its search-procedure lacks an underlying differentiating system, which makes it practically impossible to make reliable and absolute statements regarding the frequency of single words, phrases and grammatical constructions.

Created with the Personal Edition of HelpNDoc: Produce electronic books easily