Role 6 of quotations: to provide the raw material for a dictionary
Johnson's remarks quoted in the preceding section (on how 'quotations...will often exhibit...diversities of signification') point to a crucial function of quotations in a dictionary, one that now seems much more important, in a modern lexicographical enterprise, than the others reviewed above. This is that quotations provide the empirical evidence on which the dictionary is constructed: it is the use of words in context that determines what a word means. A dictionary constructed in some a priori fashion, from a list of words rather than a list of quotations, is conceptually mistaken (words only 'mean' in context) and moreover unlikely to be able to reveal and identify the full range of meanings that words have in practice. As we have seen (role 5 of quotations above), this can be simply demonstrated by comparing Johnson's entries with those of his predecessors. 

Johnson identifies this function almost incidentally, as a way of defending his inclusion of large numbers of quotations, perhaps because it is not obviously congruent with what he seems to have regarded as their most important role, that is to provide pleasure and instruction. No doubt Johnson's study of words in their contexts also explains the aptness of his definitions, which has often been remarked on, and their consequent attractiveness to the OED lexicographers who adopted so many of them.

For Murray's use of Johnson, see Silva 2005; and for a discussion of some of the differences and similarities between Johnson and OED1 in terms of their attitude towards usage, see Mugglestone 2004.

Providing the raw material for the dictionary is the most important role of the quotations so far as OED is concerned. The development in lexicographical methodology between Johnson and the OED in this respect, so that the primary (announced) role of quotations in OED is to supply evidence of usage, instead of, as in Johnson, aesthetic and/or moral example, is representative of the changes in intellectual culture between the eighteenth and nineteenth centuries (see OED and nineteenth-century scientific methodology).

Note: the term 'raw material' is used several times by the OED lexicographers to describe their collection of quotations, for example Herbert Coleridge: 'the raw material of the Dictionary, the words and authorities, are being brought together by a number of independent collectors' (go here for the context of this remark), and C. T. Onions: '[the quotations are] the raw material of most of the articles in the Dictionary' (Onions 1928a: 15)