The antecedents to 20th century PI are not well defined.
The rhetoricians of democratic Athens in 400 B.C., and later the catechumens of the middle Ages, approached an early counterpart to the small steps, sequential instruction, and question-and-answer pattern of 20th century programmed instruction. Their teaching procedures included the determination of subsequent instruction as a result of the user’s answers to the questions. PI, however, is basically a 20th century phenomenon. In its current forms, its history comprises only 30 years.
However, significant contributions were made by Sidney L. Pressed as early as 1915 in his efforts at the Ohio State University to build a simple machine for testing comprehension of material that had been taught. These crude machines presented multiple-choice questions to users while providing immediate knowledge of their results. Only later did Pressed conceive of their usefulness as instructional devices.
It is important to note that there early teaching machines represent what Pressed called adjunct auto-instruction i.e. the use of test questions presented after conventional instruction. (Hartley and Advice) This means that the machines were not integrated into the instructional material; rather, they were added to (or adjunct to) traditional instruction (usually text), much the same as adjunct questions.