Uncertainty terms (e.g., some, possible, good, etc.) are words that do not have a fixed referent and hence are relatively ambiguous. A model is proposed that specifies how, from the hearer's perspective, recognition of facework as a potential motive for the use of an uncertainty term results in a calibration of the intended meaning of that term. Four experiments are reported that examine the impact of face threat, and the variables that affect it (e.g., power), on the manner in which a variety of uncertainty terms (probability terms, quantifiers, frequency terms, etc.) are interpreted. Overall, the results demonstrate that increased face threat in a situation will result in a more negative interpretation of an utterance containing an uncertainty term. That the interpretation of so many different types of uncertainty terms is affected in the same way suggests the operation of a fundamental principle of language use, one with important implications for the communication of risk, subjective experience, and so on.