Quintilian Institutio Oratoria 11.2.17-26 English translation by Butler (Loeb, 1290-22).
(17) Ex hoc Simonidis facto notatum videtur iuvari memoriam signatis animo sedibus, idque credet suo quisque experimento. Nam cum in loca aliqua post tempus reversi sumus, non ipsa agnoscimus tantum sed etiam quae in iis fecerimus reminiscimur, personaeque subeunt, nonnumquam tacitae quoque cogitationes in mentem revertuntur. Nata est igitur, ut in plerisque, ars ab experimento.	The achievement of Simonides appears to have given rise to the observation that it is an assistance to the memory if localities are sharply impressed upon the mind, a view the truth of which everyone may realise by practical experiment. For when we return to a place after considerable absence, we not merely recognise the place itself, but remember things that we did there, and recall the persons whom we met and even the unuttered thoughts which passed through our minds when we were there before. Thus, as in most cases, art originates in experiment.
(18) Loca discunt quam maxime spatiosa, multa varietate signata, domum forte magnam et in multos diductam recessus. In ea quidquid notabile est animo diligenter adfigunt, ut sine cunctatione ac mora partis eius omnis cogitatio possit percurrere. Et primus hic labor est, non haerere in occursu: plus enim quam firma debet esse memoria quae aliam memoriam adiuvet.	Some place is chosen of the largest possible extent and characterised by the utmost possible variety, such as a spacious house divided into a number of rooms. Everything of note therein is carefully committed to the memory, in order that the thought may be enabled to run through all the details without let or hindrance. And undoubtedly the first task is to secure that there shall be no delay in finding any single detail, since an idea which is to lead by association to some other idea requires to be fixed in the mind with more than ordinary certitude.
(19) tum quae scripserunt vel cogitatione complectuntur [et] aliquo signo quo moneantur notant, quod esse vel ex re tota potest, ut de navigatione, militia, vel ex verbo aliquo: nam etiam excidentes unius admonitione verbi in memoriam reponuntur. Sit autem signum navigationis ut ancora, militiae ut aliquid ex armis.	The next step is to distinguish something which has been written down or merely thought of by some particular symbol which will serve to jog the memory; this symbol may have reference to the subject as a whole, it may, for example, be drawn from navigation, warfare, etc., or it may, on the other hand, be found in some particular word. (For even in cases of forgetfulness one single word will serve to restore the memory.) However, let us suppose that the symbol is drawn from navigation, as, for instance, an anchor; or from warfare, as, for example, some weapon.
(20) Haec ita digerunt: primum sensum [bello cum] vestibulo quasi adsignant, secundum (puta) atrio, tum inpluvia circumeunt, nec cubiculis modo aut exhedris, sed statuis etiam similibusque per ordinem committunt. Hoc facto, cum est repetenda memoria, incipiunt ab initio loca haec recensere, et quod cuique crediderunt reposcunt, ut eorum imagine admonentur. Ita, quamlibet multa sint quorum meminisse oporteat, fiunt singula conexa quodam choro, nec onerant coniungentes prioribus consequentia solo ediscendi labore.	These symbols are then arranged as follows. The first thought is placed, as it were, in the forecourt; the second, let us say, in the living-room; the remainder are placed in due order all round the impluvium and entrusted not merely to bedrooms and parlours, but even to the care of statues and the like. This done, as soon as the memory of the facts requires to be revived, all these places are visited in turn and the various deposits are demanded from their custodians, as the sight of each recalls the respective details. Consequently, however large the number of these which it is required to remember, all are linked one to the other like dancers hand in hand, and there can be no mistake since they what precedes to what follows, no trouble being required except the preliminary labour of committing the various points to memory.
(21) Quod de domo dixi, et in operibus publicis et in itinere longo et urbium ambitu et picturis fieri [spieri] potest. Etiam fingere sibi has +imagines+ licet. Opus est ergo locis quae vel finguntur vel sumuntur, et imaginibus vel simulacris, quae utique fingenda sunt. Imagines voco quibus ea quae ediscenda sunt notamus, ut, quo modo Cicero dicit, locis pro cera, simulacris pro litteris utamur.	What I have spoken of as being done in a house, can equally well be done in connexion with public buildings, a long journey, the ramparts of a city, or even pictures. Or we may even imagine such places to ourselves. We require, therefore, places, real or imaginary, and images or symbols, which we must, of course, invent for ourselves. By images I mean the words by which we distinguish the things which we have to learn by heart: in fact, as Cicero says, we use "places like wax tablets and symbols in lieu of letters."
(22) llud quoque ad verbum ponere optimum fuerit: "locis est utendum multis, inlustribus, explicatis, modicis intervallis: imaginibus autem agentibus, acribus, insignitis, quae occurrere celeriterque percutere animum possint." Quo magis miror quo modo Metrodorus in XII signis per quae sol meat trecenos et sexagenos invenerit locos. Vanitas nimirum fuit atque iactatio circa memoriam suam potius arte quam natura gloriantis.	It will be best to give his words verbatim: "We must for this purpose employ a number of remarkable places, clearly envisaged and separated by short intervals: the images which we use must be active, sharply-cut and distinctive, such as may occur to the mind and strike it with rapidity." This makes me wonder all the more, how Metrodorus should have found three hundred and sixty different localities in the twelve signs of the Zodiac through which the sun passes. It was doubtless due to the vanity and boastfulness of a man who was inclined to vaunt his memory as being the result of art rather than of natural gifts.
(23) Equidem haec ad quaedam prodesse non negaverim, ut si rerum nomina multa per ordinem audita reddenda sint. Namque in iis quae didicerunt locis ponunt res illas: mensam, ut hoc utar, in vestibulo et pulpitum in atrio et sic cetera, deinde relegentes inveniunt ubi posuerunt.	I am far from denying that those devices may be useful for certain purposes, as, for example, if we have to reproduce a number of names in the order in which we have heard them. For those who use such aids place the things which have to be remembered in localities which they have previously fixed in the memory; they put a table, for instance, in the forecourt, a proof in the hall and so on with the rest, and then, when they retrace their steps, they find the objects where they had placed them.
(24) Et forsitan hoc sunt adiuti qui auctione dimissa quid cuique vendidissent testibus argentariorum tabulis reddiderunt, quod praestitisse Q. Hortensium dicunt. Minus idem proderit in ediscendis quae orationis perpetuae erunt: nam et sensus non eandem imaginem quam res habent, cum alterum fingendum sit; et horum tamen utcumque commonet locus, sicut sermonis alicuius habiti: verborum contextus eadem arte quo modo comprehendetur?	Such a practice may perhaps have been of use to those who, after an auction, have succeeded in stating what object they have sold to each buyer, their statements being checked by the books of the money-takers; a feat which it is alleged was performed by Hortensius. It will, however, be of less service in learning the various parts of a set speech. For thoughts do not call up the same images as material things, and a symbol requires to be specially invented for them, although even here a particular place may serve to remind us, as, for example, of some conversation that may have been held there. But how can such a method grasp a whole series of connected words?
(25) Mitto quod quaedam nullis simulacris significari possunt, ut certe coniunctiones. Habeamus enim sane, ut qui notis scribunt, certas imagines omnium et loca scilicet infinita, per quae verba quot sunt in quinque contra Verrem secundae actionis libris explicentur, [ne] meminerimus etiam omnium quasi depositorum: nonne impediri quoque dicendi cursum necesse est duplici memoriae cura?	I pass by the fact that there are certain things which it is impossible to represent by symbols, as, for example, conjunctions. We may, it is true, like shorthand writers, have definite symbols for everything, and may select an infinite number of places to recall all the words contained in the five books of the second pleading against Verres, and we may even remember them all as if they were deposits placed in safe-keeping. But will not the flow of our speech inevitably be impeded by the double task imposed upon our memory?
(26) Nam quo modo poterunt copulata fluere si propter singula verba ad singulas formas respiciendum erit? Qua re et Charmadas et Scepsius de quo modo dixi Metrodorus, quos Cicero dicit usos hac exercitatione, sibi habeant sua: nos simpliciora tradamus.	For how can our words be expected to flow in connected speech, if we have to look back at separate symbols for each individual word? Therefore the experts mentioned by Cicero as having trained their memory by methods of this kind, namely Charmadas, and Metrodorus of Scepsis, to whom I have just referred, may keep their systems for their own use. My precepts on the subject shall be of a simpler kind.
	[DRB 10.2.2009]