This study investigates how (im)politeness strategies are used in the performance of four different speech acts in modern Korean business correspondence. The data consists of 194 authentic business texts of internal (77 e-mail messages) and external correspondence (117 formal letters) collected from two Korean companies a food company (corporation A) and a pharmaceutical company (corporation B). The corpus of Korean business correspondence provides evidence for the four different types of speech acts: disagreement, giving bad news, request and compliment, Brown and Levinson’s (1987) theory of politeness was revised and applied to examining text data. A three-level text analysis is used as the primary data analysis method in this study: (1) a hand-tagged functions-analysis, (2) a qualitative methods: and (3) a quantitative method. Major findings in my investigation of (im)politeness strategies realized in Korean business correspondence are summarised as follows. 1. The rhetorical style in Korean business correspondence can be disconnected by ‘external modifications’ for politeness concerns. 2. Solidarity enforcement strategies are divided into two types of strategies: ‘show interest’ and ‘use of in-group language’. Conflict avoidance strategies are divided into three groups ‘be conventionally indirect’. ‘mitigating devices’, and ‘defocusing the writer/reader or action from the face-threatening act’. Off record strategies are divided into two types: ‘reasonableness’ and ‘act in question’. 3. The most frequent (im)politeness strategy across corporations and genres is ‘conflict avoidance’. The second most frequent strategy (internal communication of corporation B being the only exception) is ‘solidarity enforcement’. The least frequent strategy is ‘off record’ in the internal communications of both corporation and ‘bald on record’ in the external communications of both corporations. 4. A static or traditional view of culture cannot be always adapted to accommodate the data from Korean business correspondence. It is because each corporation has its own distinctive corporate culture/norm (which might be different from national culture), and each corporate culture/norm affects the choice of (im)politeness strategies (e.g. the heavy use of bald on record strategy by less powerful people in corporation B, the heavy use of conflict avoidance strategy by more powerful people in corporation B, and the heavy use of off record strategy by more powerful people in both corporations). A most important message I convey in this study is that it is dangerous to make any generalizations about the characteristics of languages and cultures using cross-cultural data (e.g. the conventional rhetorical structure in Korean business correspondence; hierarchical Korean society in the realization of (im)politeness strategies). Since results can be substantially different from the expectations one might have from conventional or traditional perspectives of a culture, we need to overcome cultural stereotyping in dealing with politeness phenomena or any other pragmatic phenomena.