I AM A MIDDLE-AGED SPECIMEN of the human race, one who has been raised in the United States with English as my native tongue, and yet, even given the benefit of a university education and years of practical experience using English, I am prone to err in my use of English. Why? But why ask such a question: for is it not human to err? Yes, and if the errors are distributed more or less at random, there would be no need for further inquiry: on the other hand, if a particular error is repeated habitually, then it may be possible to adduce a causal mechanism. This I shall attempt to do for the error of the split infinitive. The split infinitive is a peculiar kind of error, for those who frequently split defend their practice, and in defense point with glee to splits committed by wellknown writers. In many cases the cited author may be dead, and it may be impossible to query the author as to whether he regarded that split as an error or not. In one instance, Sheridan Baker, as recounted in The Complete Stylist and Handbook (New York: Thomas Y. Crowell, 1976), responded to exactly that kind of citation by a student and queried Walter Lippmann as to whether a given split was justified in his mind (p. 213). The response was that the split was a slip which had escaped detection. But is a split really an error? Yes, if we accept the proscriptive rule that one should not split an infinitive. Now, however, I can ask a different question, which, I hope, will get at the heart of the matter. Why has the proscriptive rule, as taught to each student, not induced in the mind of every student a transformation rule such that split infinitives regularly are avoided? That is, one studies grammar in school so as to acquire patterns of speech and habits of thought that will result in the generation of acceptable English. It is my contention that the rule "not to split" is insufficient, by itself, to induce in the mind of every student a generative grammar such that only unsplit infinitives are produced. Some, of course, may learn to avoid split infinitives, but others do not. And for those whose internal generative grammar does not include a mechanism that generates, automatically, unsplit infinitives, the proscription always will require what seems to them an unnecessary after-the-fact patch. The splitters, therefore, defend their behavior because, to them, it seems natural to split. To split or to not split, that is the problem: whether 'tis nobler in the mind to suffer the generation of infinitives outrageously split, or to invent new rules of grammar, and by promulgating end them. But what new rule of grammar could be likely, in union with the existing rules of grammar, to induce in the mind of the