|The types of scalar inference can be divided into these two: scalar entailment and scalar implicature. According to Gazdar (1979), Levinson (1983, 2000) and Horn (1985, 1989, 2004), assuming a scale < e1, e2, e3 … en >, where e1 scalar-entails e2, e2 also scalar-entails e3, etc, but not vice versa. On the other hand, uttering a sentence including (en) scalar-implicates the negation of a sentence including (en-1). These scalar inferences can be the evidence verifying that ordinary language users observe the Maxim of Quantity proposed by Grice (1975), and these pragmatic principles seem to always enable the ordinary language users to foresee regular conclusions, in which any scalar utterances occur. In our ordinary language use, however, sometimes there might be some exceptional cases where the hearer cannot properly interpret the speaker`s exact intention, if these scalar inferential principles are to be mechanically applied. The reason is due to the point that various non-linguistic factors such as language users` intuition and their socio-cultural environments can also be involved in the process of the interpreting the scalar utterances, besides the linguistic principles proposed by Horn (1989, 2004) and Levinson (2000). These non-linguistic factors are very significant in that they might influence the ultimate meaning intended by the speaker; in this paper, I discuss what they are, and account for how the scalar inference connected to them can be treated in the discourse.