Since 1778, scholars have disagreed about the relationship between Shakespeare’s Macbeth and Middleton’s Witch. Allegedly empirical studies of the authorship of passages of Macbeth (3.5 and 4.1) have reached opposed conclusions. Although modern databases and more sophisticated attribution methods have succeeded in reaching consensus about whole plays and even plays divided between two authors, especially plays dating from the seventeenth century, they have been less successfulwith smaller blocks of text. Taylor’s essay presents the first survey of all lexical data (words, word strings of various lengths, collocations) in control passages in King Lear, Pericles, and A Mad World, My Masters, checked against databases of the entire Middleton and Shakespeare canons and then against databases of all early modern drama. Establishing the reliability of such comprehensive tests in identifying short passages of known authorship, the essay tests the suspect passages in Macbeth 3.5 and 4.1, using the same procedures and databases, concluding that there is overwhelming evidence against Shakespeare’s authorship and for Middleton as the adapter. Taylor relates this empirical conclusion to larger issues of style and epistemology in the Shakespeare and Middleton canons, concluding that Macbeth is the tragedy of a man who believed what he was told.