Every speech however consists at once of that which is expressed and that which expresses, that is to say of matter and words. Skill in speaking is perfected by nature, art and practice, to which some add a fourth department, namely imitation, which I however prefer to include under art.

There are also three aims which the orator must always have in view; he must instruct, move and charm his hearers. This is a clearer division than that made by those who divide the task of oratory into that which relates to things and that which concerns the emotions, since both of these will not always be present in the subjects which we shall have to treat. For some themes are far from calling for any appeal to the emotions, which, although room cannot always be found for them, produce a most powerful effect wherever they do succeed in forcing their way.

The best authorities hold that there are some things in oratory which require proof and others which do not, a view with which I agree. Some on the other hand, as for instance Celsus, think that the orator will not speak on any subject unless there is some question involved in it; but the majority of writers on rhetoric are against him, as is also the threefold division of oratory, unless indeed to praise what is allowed to be honourable and to denounce what is admittedly disgraceful are no part of an orator's duty.

It is, however, universally agreed that all questions must be concerned either with something that is written or something that is not. Those concerned with what is written are questions of law, those which concern what is not written are questions of fact. Hermagoras calls the latter rational questions, the former legal questions, for so we may translate λογικόν and νομικόν.

Those who hold that every question concerns either things or words, mean much the same. It is also agreed that questions are either definite or indefinite. Indefinite questions are those which may be maintained or impugned without reference to persons, time or place and the like. The Greeks call them theses, Cicero Top. xxi. 79. propositions, others general questions relating to civil life, others again questions suited for philosophical discussion, while Athenaeus calls them parts of a cause.

Cicero Top. 81; Part. Or. xviii. 62. distinguishes two kinds, the one concerned with knowledge, the other with action. Thus Is the world governed by providence? is a question of knowledge, while Should we enter politics? is a question of action. The first involves three questions, whether a thing is, what it is, and of what nature: for all these things may be unknown: the second involves two, how to obtain power and how to use it.

Definite questions involve facts, persons, time and the like. The Greeks call them hypotheses, while we call them causes. In these the whole question turns on persons and facts.

An indefinite question is always the more comprehensive, since it is from the indefinite question that the definite is derived. I will illustrate what I mean by an example. The question Should a man marry? is indefinite; the question Should Cato marry? is definite, and consequently may be regarded as a subject for a deliberative theme. But even those which have no connexion with particular persons are generally given a specific reference. For instance the question Ought we to take a share in the government of our country? is abstract, whereas Ought we to take part in the government of our country under the sway of a tyrant? has a specific reference.

But in this latter case we may say that a person is tacitly implied. For the mention of a tyrant doubles the question, and there is an implicit admission of time and quality; but all the same you would scarcely be justified in calling it a cause or definite question. Those questions which I have styled indefinite are also called general: if this is correct, we shall have to call definite questions special questions. But in every special question the general question is implicit, since the genus is logically prior to the species.

And perhaps even in actual causes wherever the notion of quality comes into question, there is a certain intrusion of the abstract. Milo killed Clodius: he was justified in killing one who lay in wait for him. Does not this raise the general question as to whether we have the right to kill a man who lies in wait for us? What again of conjectures? May not they be of a general character, as for instance, What was the motive for the crime? hatred? covetousness? or Are we justified in believing confessions made under torture? or Which should carry greater weight, evidence or argument? As for definitions, everything that they contain is undoubtedly of a general nature.

There are some who hold that even those questions which have reference to persons and particular cases may at times be called theses, provided only they are put slightly differently: for instance, if Orestes be accused, we shall have a cause: whereas if it is put as question, namely Was Orestes rightly acquitted? it will be a thesis. To the same class as this last belongs the question Was Cato right in transferring Marcia to Hortensius? These persons distinguish a thesis from a cause as follows: a thesis is theoretical in character, while a cause has relation to actual facts, since in the former case we argue merely with a view to abstract truth, while in the latter we have to deal with some particular act.

Some, however, think that general questions are useless to an orator, since no profit is to be derived from proving that we ought to marry or to take part in politics, if we are prevented from so doing by age or ill health. But not all general questions are liable to this kind of objection. For instance questions such as Is virtue an end in itself? or Is the world governed by providence? cannot be countered in this way.

Further in questions which have reference to a particular person, although it is not sufficient merely to handle the general question, we cannot arrive at any conclusion on the special point until we have first discussed the general question. For how is Cato to deliberate whether he personally is to marry, unless the general question whether marriage is desirable is first settled? And how is he to deliberate whether he should marry Marcia, unless it is proved that it is the duty of Cato to marry?

There are, however, certain books attributed to Hermagoras which support this erroneous opinion, though whether the attribution is spurious or whether they were written by another Hermagoras is an open question. For they cannot possibly be by the famous Hermagoras, who wrote so much that was admirable on the art of rhetoric, since, as is clear from the first book of the Rhetorica of Cicero, de Inv. i. 6. he divided the material of rhetoric into theses and causes. Cicero objects to this division, contends that theses have nothing to do with an orator, and refers all this class of questions to the philosophers.

But Cicero has relieved me of any feeling of shame that I might have in controverting his opinion, since he has not only expressed his disapproval of his Rhetorica, but in the Orator, Orator xiv. 45. the de Oratore and the Topica de Or. iii. 30; Top. 21. instructs us to abstract such discussions from particular persons and occasions, because we can speak more fully on general than on special themes, and because what is proved of the whole must also be proved of the part.

In all general questions, however, the essential basis is the same as in a cause or definite question. It is further pointed out that there are some questions which concern things in themselves, while others have a particular reference; an example of the former will be the question Should a man marry? of the latter Should an old man marry? ; or again the question whether a man is brave will illustrate the first, while the question whether he is braver than another will exemplify the second.