A Faroese recipe in a cookbook explains how to catch a puffin before you roast it. Like a cook, a linguist studying language change must first gather together the basic ingredients. In the case of the linguist, the facts must be collected and pieced together before they can be interpreted. How is this done?

There are basically two ways of collecting evidence, which we may call the ‘armchair method’ and the ‘tape-recorder method’ respectively. In the first, a linguist studies the written documents of bygone ages, sitting in a library or at a computer, and in the second he or she slings a tape recorder over one shoulder and studies change as it happens. Both methods are important, and complement one another. The armchair method enables a large number of changes to be followed in outline over a long period, whereas the tape-recorder method allows a relatively small amount of change to be studied in great detail.

The armchair method is the older, and the basic techniques were laid down in the nineteenth century – as is shown by the quotation above from Dickens' novel Dombey and Son which was published in 1847–8. Let us therefore deal with it first.