Completion norms for 744 French linguistic contexts of differing formats

In 1980, Kutas and Hillyard (1980) reported that the presentation of an unexpected, or incongruous, word within a sentence context is associated with a negative component, peaking 400 ms after word onset: the N400 component. Numerous studies have been aimed at determining the functional significance of this N400 component of the Event-Related brain Potentials (ERPs) and at testing models derived from psycholinguistic research. Results of these experiments have added important information in the area of word recognition, repetition and context effects, and language impairment research. Almost all of these experiments have been conducted in English and have used standardized materials, thus allowing comparison between results obtained in different laboratories. The aim of the study reported here is to provide such materials in French. A total of 744 sentences were constructed and presented to 52 native French speakers. The last word of each sentence was omitted and subjects were asked to complete the sentence context with the first word that came to mind that was syntactically and semantically appropriate. The Cloze probability (Taylor, 1953), that is the probability of a word being given to complete a sentence context, was computed for each sentence. Results are presented as a function of the number of words per sentence. Three set of sentences were created depending upon the terminal word's Cloze probability: low (n = 576), high (n = 58) and very high (proverbs, n = 110). The use of these sentences in experiments on the electrophysiology of language processing and as a database in psychololinguistics and cognitive neuropsychology is discussed.