ABSTRACT
This chapter deals with another harmful consequence of that mistaken view of language that is at least in part due to linguistics, but one so specialized that even many readers who have found the text so far easy going may wish to skip it. It shows that mathematicians are no better than the rest of us at analyzing what lies closest to hand; they have misunderstood the relation between the two disciplines of mathematics and programming, largely because they think that mathematics is a language. The symbolic parts of mathematical texts seem perfectly analogous to natural-language texts, which are likewise sequences of statements constructed from a finite repertoire of symbols, in this case the alphabetic characters. The primitive operations and operands of mathematics are definable only in natural language; composite and higher-level terms are then defined in terms of those primitives.
