Conclusion

The computer analysis of corpora provides us with profiles of the occurrence and co-occurrence of textual features and these serve as a norm of what is customary against which any particular instance of usage can be compared. 

Text analysis can only tell us about texts, language that people produce in the process of communication. It cannot tell us about the process itself, about how people negotiate a relationship between text and context in order to bring about a degree of discourse convergence appropriate to their purpose.

These textual facts cannot account for all the other factors we have considered in this book that come into play in making meaning, and it would be a mistake to claim that  they can. 

What they can do is to alert us to possible intentions and interpretations which we might otherwise not be aware of, and so provide a basis, and a stimulus, for further empirical enquiry into the pragmatics of discourse and the nature of human communication.