A discussion of text translatability in simultaneous interpreting (SI) looks at semantic redundancy, the repetition of semantic components essential to creating an utterance, and offers some classroom techniques for teaching interpreting skills. It is proposed that the translatability of a text in SI should be studied in terms of the experiential and cognitive models used in the making of its sentences, i.e., whether they are the result of one-domain mapping (iconic or metonymic) or two-domain mapping. A text can be handled optimally in SI if the predominant types of mapping in the making of its utterances are one-domain iconic mapping; one-domain metonymic mapping, with the same part; whole relationship in the two languages; and two-domain metaphorical mapping with some restrictions. Most SI training is oriented toward increasing knowledge of the language used in some major fields, coping with grammatical difficulties, and developing memory and attention. However, the semantics of the utterance should also be addressed by helping students cope with language-specific cognitive mapping, particularly in a workshop format. Class exercises using: (1) exemplary analysis with instruction for individual work; and (2) report-with-discussion are useful for this type of training. (MSE)