Quintilian, Institutio Oratoria, Book 4
Harold Edgeworth Butler, Ed.

[17] For example, this made it necessary for Cicero to digress even in the exordium when he was defending Milo, as is clear from the short speech¹ which he made on that occasion. But the orator who makes some preface to the main question or proposes to follow up his proofs with a passage designed to commend them to the jury, may digress at some length. On the other hand, if he breaks as say in the middle of his speech, he should not be long in returning to the point from which he departed.

IV. After the statement of facts some place the proposition² which they regard as forming a division of a forensic speech. I have already expressed my opinion of this view.³ But it seems to me that the beginning of every proof is a proposition, such as often occurs in the demonstration of the main question and sometimes even in the enunciation of individual arguments, more especially of those which are called ἐπιχειρήματα⁴ But for the moment I shall speak of the first kind. It is not always necessary to employ it.