...The careful system of discourse has two parts, one of discovery and the other of judgment -- sometimes judgment of the discovery itself, sometimes judgment of the deduction of the discovery, which is the form of an argumentation. The part that teaches about discovery supplies in abundance certain tools for discoveries and is called 'Topics'....The part that has to do with judgment proffers certain rules making determinations and is called 'Analytics'. If it makes observations about the junctures of propositions, it is named 'Prior Analytics'. But if it deals with the discoveries themselves, then the part that discusses the determining of necessary arguments is named 'Posterior Analytics', and the part that discusses false and tricky (that is, sophistical) arguments is named 'Refutations'. The judgment of verismilar argumentations is apparently not dealt with because the nature of judgment concerning the middle is clear and uncomplicated when one is acquainted with the extremes. For if one knows how to judge discerningly what is necessary and is also able to judge false arguments, it is no trouble for him to determine verisimilar arguments, which are in the middle.
[Boethius, Boethius's In Ciceronis Topica, Stump, tr. Cornell University Press (Ithaca, NY: 1988) pp. 27-28.]
This is a very interesting passage. Later medieval scholastics will often take Topics to cover verisimilar (i.e., probable) arguments. Thomas Aquinas agrees with this, but also agrees with Boethius in assigning Topics to the logic of discovery (logica inventiva); then, partly following the Islamic commentators, he also assigns Rhetoric and Poetics to it, taking Topics to be the logica inventiva that concerns belief (which is appropriate to probable argument), Rhetoric to be the logical inventiva that concerns suspicion (as in 'suspecting to be true'), and Poetics to be the logica inventiva that concerns 'estimation according to some representation'.
