This paper examines the goals of prescriptive grammar and the causes and consequences of the rift between prescriptive and theoretical linguistics. It also proposes a principle for guiding prescriptive recommendations in the future as well as a theoretical framework and procedure for predicting the consequences of prescriptive recommendations. The procedure illustrates a hypothetical prescription: the substitution of singular they for prescriptive he. Projected benefits the prescription include neutral connotation, naturalness, simplicity, and lexical availability. Projected costs include covert and overt referential ambiguity; partial ambiguity; conceptual inaccuracy; loss of precision, imageability, impact, and memorability; bizarreness involving certain referents and case forms; distancing and dehumanizing connotations; unavailability of the ‘he or she’ denotation; potentially disruptive and long-lasting side effects on other areas of the language. Procedures are also illustrated for determining the relative frequency of such costs and benefits and for estimating the relative disruptiveness of the costs normal language use. Implications of the data for several issues general interest to linguistics and psychology are explored. (Ambiguity, language change, prescriptive grammar, theoretical linguistics, language planning, pronouns, neologisms.)