Linguists call the total word stock of a language its lexicon. The linguist would define a dictionary as an objective record of the lexicon of a language.But the lexicon of English is open-ended. It is not even theoretically possible to record it all as a closed system;so the task of the dictionary maker is to determine, a selection of items from the lexicon that form a reasonable group for a purpose. Such a group is called a corpus. (This problem faces all dictionary makers.) There is a notion in America that a so-called unabridged dictionary records the unabridged lexicon, that is, all the words of the language. This is nonsense. Philip Gove, editor of the largest American dictionary now in print, wrote in his preface that it could have been many times larger, but it would have been impossible to bind.His predecessor, John P. Bethel, once told me that it would be possible to fill a dictionary the size of the, largest "unabridged"with names of compounds of carbon alone. How then does the lexicographer select his words? In the past there was only one way that offered any objectivity-making a file of citations from reading. All respectable dictionaries have done and still do this, though no one will ever again compile such a file as the great Oxford English Dictionary,which has extensive lists of citations for each word to establish its history.