Now rhetoric as we have discussed it in relation to the lovers consists of truth plus its artful presentation, and for this reason it becomes necessary to say something more about the natural order of dialectic and rhetoric. In any general characterization rhetoric will include dialectic,[9] but for the study of method it is necessary to separate the two. Dialectic is a method of investigation whose object is the establishment of truth about doubtful propositions. Aristotle in the _Topics_ gives a concise statement of its nature. “A dialectical problem is a subject of inquiry that contributes either to choice or avoidance, or to truth and knowledge, and that either by itself, or as a help to the solution of some other such problem. It must, moreover, be something on which either people hold no opinion either way, or the masses hold a contrary opinion to the philosophers, or the philosophers to the masses, or each of them among themselves.”[10] Plato is not perfectly clear about the distinction between positive and dialectical terms. In one passage[11] he contrasts the “positive” terms “iron” and “silver” with the “dialectical” terms “justice” and “goodness”; yet in other passages his “dialectical” terms seem to include categorizations of the external world. Thus Socrates indicates that distinguishing the horse from the ass is a dialectical operation;[12] and he tells us later that a good dialectician is able to divide things by classes “where the natural joints are” and will avoid breaking any part “after the manner of a bad carver.”[13] Such, perhaps, is Aristotle’s dialectic which contributes to truth and knowledge. But there is a branch of dialectic which contributes to “choice or avoidance,” and it is with this that rhetoric is regularly found joined. Generally speaking, this is a rhetoric involving questions of policy, and the dialectic which precedes it will determine not the application of positive terms but that of terms which are subject to the contingency of evaluation. Here dialectical inquiry will concern itself not with what is “iron” but with what is “good.” It seeks to establish what belongs in the category of the “just” rather than what belongs in the genus _Canis_. As a general rule, simple object words such as “iron” and “house” have no connotations of policy, although it is frequently possible to give them these through speech situations in which there is added to their referential function a kind of impulse. We should have to interpret in this way “Fire!” or “Gold!” because these terms acquire something through intonation and relationship which places them in the class of evaluative expressions.