Quintilian, Institutio Oratoria, Book 10
Harold Edgeworth Butler, Ed.

6. Having dealt with writing, the next point which claims our attention is premeditation, which itself derives force from the practice of writing and forms an intermediate stage between the labours of the pen and the more precarious fortunes of improvisation; indeed I am not sure that it is not more frequently of use than either. For there are places and occasions where writing is impossible, while both are available in abundance for premeditation. For [p. 129] but a few hours' thought will suffice to cover all the points even of cases of importance; if we wake at night, the very darkness will assist us, while even in the midst of legal proceedings our mind will find some vacant space for meditation, and will refuse to remain inactive. [2] Again, this practice will not merely secure the proper arrangement of our matter without any recourse to writing, which in itself is no small achievement, but will also set the words which we are going to use in their proper order, and bring the general texture of our speech to such a stage of completion that nothing further is required beyond the finishing touches. And as a rule the memory is more retentive of thoughts when the attention has not been relaxed by the fancied security which results from committing them to writing.

But the concentration which this requires cannot be attained in a moment or even quickly. [3] For, in the first place, we must write much before we can form that ideal of style which must always be present to our minds even when engaged in premeditation. Secondly, we must gradually acquire the habit of thought: to begin with, we shall content ourselves with covering but a few details, which our minds are capable of reproducing with accuracy; then by advances so gradual that our labour is not sensibly increased we must develop our powers and confirm them by frequent practice, a task in which the most important part is played by the memory. [4] For this reason I must postpone some of my remarks to the portion of this work reserved for the treatment of that topic.¹ At length, however, our powers will have developed so far that the man who is not hampered by lack of natural ability will by dint of [p. 131] persistent study be enabled, when it comes to speaking, to rely no less on what he has thought out than what he has written out and learnt by heart. At any rate, Cicero records that Metrodorus of Scepsis,² Empylus of Rhodes,³ and our own Hortensius⁴ were able to reproduce what they had thought out word for word when it came to actual pleading.

[5] If, however, some brilliant improvisation should occur to us while speaking, we must not cling superstitiously to our premeditated scheme. For premeditation is not so accurate as to leave no room for happy inspiration: even when writing we often insert thoughts which occur to us on the spur of the moment. Consequently this form of preparation must be conceived on such lines that we shall find no difficulty either in departing from it or returning to it at will. [6] For, although it is essential to bring with us into court a supply of eloquence which has been prepared in advance in the study and on which we can confidently rely, there is no greater folly than the rejection of the gifts of the moment. Therefore our premeditation should be such that fortune may never be able to fool us, but may, on the contrary, be able to assist us. This end will be obtained by developing the power of memory so that our conceptions may flow from us without fear of disaster, and that we may be enabled to look ahead without anxious backward glances or the feeling that we are absolutely dependent on what we can call to mind. Otherwise I prefer the rashness of improvisation to the coherence given by premeditation. [7] For such backward glances place us at a disadvantage, because our search for our premeditated ideas makes us miss others, and we draw [p. 133] our matter from our memory rather than from the subject on which we are speaking. And even if we are to rely on our memory and our subject alike, there are more things that may be discovered than ever yet have been.