This study investigated the judgments of politeness made by both native speakers of American English and nonnative ESL learners with varied language backgrounds. The judgments were made on eight different request strategies in English, using the method of rank orderings in a contextualized condition. The eight strategies tested were systematically varied across three syntactic/semantic features generally regarded in the theoretical literature (e.g., Lakoff 1972, 1973b. Jespersen 1964, Quirk and Greenbaum 1973) as varying with politeness‐namely, imperative/declarative/ interrogative mood, presence or absence of modals, and tense of modals (“can” vs. “could”). Previous work in this area in ESL, e.g., Rintell(1979), Scarcella (1980). Waiters(1980). while a fruitful beginning, has analyzed spontaneously produced data and has not explored the syntactic/ semantic distinctions systematically. By contrast, this study analyzes theory‐based distinctions via judgments elicited under controlled conditions.Results indicated that the responses of both native speakersand ESL learners fell into a hierarchy similar to that hypothesized. There was also a high and significant correlation between the natives and nonnatives in their politeness judgments, a finding similar to that of Walters (1980). In general, however, the ESL learners tended to perceive more distinctive levels of politeness than the native speakers, reflecting a kind of “over‐sensitivity” to syntactic/semantic form distinctions. Examination ofthe native politeness hierarchy showed each of the three syntacticlsemantic features contributing differentially to politeness; this suggests that in teaching the pragmatics of politeness, different grammatical patterns should not be given equal