"The term ‘formulaic sequence’ was coined to refer to a wide range of subtypes of multi-word strings that “[are] or appear … to be, prefabricated: that is, stored and retrieved whole from memory at the time of use, rather than being subject to generation or analysis by the language grammar” (Wray, 2002b, p. 9). Formulaic sequences have long been recognized as a feature of language disorders. Early clinicians remarked on their resilience in aphasia, where they often remain when other linguistic capabilities have been lost. However, they occur in a number of other conditions too. Only by putting all the different manifestations together, and locating them within the nature of normal communication practice, can we see the whole picture. It is for theoretical models to explain how and why certain word strings should have the status of ‘formulaic’ when others do not, and why different subtypes are found in different language disorders. In clinical linguistics, as in second language acquisition research, identifying formulaic sequences is assisted by the contrast between normal (or native-like) language and abnormal (or non-native-like) language."