Cao one language influence anorher ene structurally? Or, put differently, can languages borrow from each ether? This issue has been hotly debated, both in historical linguistics and in language contact studies, and na consensus bas been reached. Doe ofthe reasons for this is that there are widely divergent views on what language is really like. At opposite ends we find the ‘system’ view end the ‘bag of tricks’ view. The system view holds that languages, or more specifically grammars, are tightly organized wholes, of which all elements are related by complex syntagmatic and paradigmatic re1ationships. A prominent advocate ofthe system view was Ferdinand de Saussure, the founder of structuralism, who claimed that language was a system ‘ou tout se tient’ (where everything hangs rogerher). The bag of tricks view holds rhat languages are primarilv complex tools for referring to the world and for communication, and that these tools easily adapt to new communicative and referential needs. A prominent proponent of the bag of tricks view was Hugo Schuchardt, the creolist referred ro in chapter I. Schuchardr (1914) went as far as proposing Ianguage chemietry for handling creole data: the ‘formula’ for two creole languages spoken in Surinem was CEP ,D for Sranan end CEP, D for Saramaccan (C = CreoIe, E = English, P = Portuguese, D = Durch), and rhe different numerical subscript for P indicares the different amount of Portuguese present in both creoles.

On purpose we have cited two scholars from the early twentieth century as holding these rather opposite points of view. In modern linguistics the distinction is not as clear,as we will see be1ow. One would tend to associate Chomsky and the generative tradition with the system view, and Hymes and other functionalists with the bag of tricks view. The notion of system itse1f, however, has undergone important changes. The holistic systems of early structuralism have been replaced by modularized systems in modern grammatical theory. These systems contam a number ofindependent components: rhe lexicon, the phonological component, etc. The impheation of this conception of the grammar for borrowing is thar borrowing a word does not imply necessarüy thar the sounds of which the word are composed are bcrrowed in the same way. Of course words are phonologically adapted in the process ofborrowing.