Data from a wide variety of sources are cited in support of the cloze technique as a global measure of language proficiency. Cloze tests as measures of readability, reading comprehension, quality of translations, pragmatic expectancies due to textual illustrations, attitudes, and ethno‐linguistic background are discussed. In second language applications of the cloze procedure the effects of variations in difficulty, scoring methods, and the language used are considered. Analyses of response types and frequencies are examined. Correlations with external criteria and other theoretical considerations suggest that short term memory constraints are invoked by the cloze procedure. It is argued that such constraints are limitations oncompetencerather than mereperformance.It is claimed that cloze tests tap this underlying competenc