As a starting-point, we can explore the origins and uses of the word collocation in English lexis through its treatment in dictionaries. The entries for the word in the dictionaries produced during the time that the word has been current in English and in the Oxford English Dictionary (OED) provide some evidence of the use of the word and the meanings attributed to it at each stage. This is supplemented by an examination of the use of the word in other texts over the period since its introduction into English.

While the dictionary entries and other text sources provide evidence of the existence and use of the word in English vocabulary over the centuries, the practical significance of collocation as a linguistic concept can be more usefully assessed through the use made of the concept of collocation within dictionaries. This establishes the extent to which lexicographers working in different periods have recognised the phenomenon of collocation and the importance (if any) that they have attached to it as a source of information relevant to the words they have documented.

We can get a similar practical demonstration of the use of collocation by considering the development of published concordances – lists of words found in texts showing the environment in which they were used. These were first produced as a form of index to the Bible, and later extended to other texts which were seen as being sufficiently important to justify the production of aids to their analysis. Concordances show a form of practical recognition of the significance of collocation as a tool in the disambiguation of meaning and in the close interpretation of texts through the context of significant words.

Dictionaries and concordances are not the only forms of linguistic guidance that make use of collocation. From the nineteenth century to the present day, style guides have proliferated, providing linguistic advice for those lacking confidence in their use of English, often as a component of guidance on more general matters of etiquette. During the twentieth century large numbers of dictionaries intended for non-native speakers of English were produced. The contents of these dictionaries reveal a shift in the attitude of the lexicographers to collocation, from a reliance on the evidence that it can provide of the meaning of words towards a recognition that it is in itself an important element of the language knowledge that learners need to acquire.

This shift of attitude in turn informed the development of linguistic theory during the twentieth century and led to the eventual identification of collocation as an underlying principle of language production and interpretation.

All of these strands will be examined in the following sections to give as complete a picture as possible of the development of collocation as a recognised phenomenon of word behaviour, a tool for language analysis and an element of linguistic theory.

The Lexicons of Early Modern English (LEME) database is a collection of dictionaries and similar texts produced from 1450 to 1702. It is possible to search for occurrences of specific words within LEME, and a search was carried out using the string ‘collocat*’ so as to find any occurrences of collocate, collocates, collocating, collocation etc. The ‘*’ is a special character which is interpreted by the search mechanism as any characters following the search string. The earlier dictionaries containing this string, published between 1538 and 1587, were bilingual dictionaries of English and Latin, and contained a variant of the Latin form collocatus. The string was also found – in the form collocation – in four monolingual English dictionaries. These were Bullokar’s An English Expositor, published in 1616, Cockeram’s English Dictionarie, published in 1623, Phillips’ The New World of English Words published in 1658 and Coles’ An English Dictionary published in 1676.

Collocation. A placing together.

According to the OED, the word collocation first appears in English in various forms around the beginning of the sixteenth century. It provides a first quotation from 1513 of the verb form, collocate, and gives as its first two senses:

For the primary sense of ordering physical items or facts, ‘placing together with or side by side’ in the OED’s words, the word arrange would seem to be a suitable candidate. Adopted originally from French in the fourteenth century, an earlier period of borrowing frenzy, arrange is originally used in its more specific (and more strictly etymological) sense of ‘draw up in ranks or in line of battle’.

In fact, according to the OED this is a rare word until the nineteenth century. Sense 2a is closest to the meanings already identified for collocate:

In the English definitions for collocation in dictionaries other than the OED quoted above, the word placing commonly forms part of the phrase. It may be that collocation was borrowed because placing had too general a meaning and was being too widely used, while the meaning of the word arrange was too specific until significantly later. The appearance of the word collocation in so many of the early dictionaries suggests that its selection to fill this need was successful from the start.

In the definition of sense 1b the OED recognises the use of the word ‘quasi-concretely’, or almost as a noun in its own right, rather than an action, and includes as an illustrative quotation the passage from Southey’s The Doctor, published in 1836. A fuller quotation from Southey’s text gives the point of his comment more clearly. Speaking of his character Daniel’s high opinion of the seventeenth-century poet Joshua Sylvester, he claims that:

This use by Southey suggests a notion of collocation which emphasises unusual juxtapositions, carefully chosen as part of a literary technique, and it is useful to explore the extent to which this concept appears in other literary texts.

If a search is conducted for all possible variants of the words collocate and collocation in all the texts currently making up the Literature Online resource, 79 instances are found in all, and of these 13 refer to non-linguistic senses of the words, three are instances of Latin texts and a further three are duplicate entries. The remaining 60 occurrences refer to a roughly equal mix of what may be called ‘unusual’ and ‘habitual’ associations between linguistic units. The earliest text in which the linguistic sense is found is The Art of Rhetorick Concisely and Compleatly Handled, by John Barton, from 1634. In this text an example is given to illustrate the notion of composition: