mar, the fundamental work in language upon which both dialectic and
rhetoric built. Training in memoria rerum was associated with the tasks of
rhetoric, specifically invention and delivery.42
The fourteenth-century English Dominican Thomas Waleys, a careful
scholar of decidedly non-florid style (Beryl Smalley remarks that, unlike
Robert Holcot, Waleys ‘‘seldom made a mistake’’ in the attribution of a
source), 43 advises against rote memorizing of one’s own composition in his
De modo componendi sermones (c. 1342), for:
114
The Book of Memory
Words fall out of a facile memory, and from such a trivial action, the memory of
what one is saying is thrown in confusion, because the words (rather than
concepts) do not present themselves. Often, from forgetting a single syllable,
one botches the whole thing. Thus, the preacher can be confounded because he
has bound himself to words rather than to their substance.44
Waleys also disapproves of relying on an overly polished style, composing
the whole sermon in rhythm, or too many divisions of the text, for these
devices also engender mistakes of recollection and the consequent confu-
sion of the preacher. So too does citing too many authorities. Forgetfulness
and its attendant embarrassment is the fault of the preacher who strives to
excel in mere ingenuity (‘‘qui in curiositate conatur excellere’’).
To ward off such preacher’s perils (‘‘periculum praedicatori’’), one
should memorize not word for word, but according to the sententia of
one’s authorities (‘‘sententiam auctoritatum plene et distincte retineat’’).
By this advice, Thomas Waleys means that one should remember the most
important words, but not worry about the lesser ones:
And if there are words in these authorities [i.e. the texts] which are singularly
weighty in merit, one should especially strive in memory to retain and speak about
them, taking fewer pains with the others. And this for sure, that there are many
authoritative texts of the saints which, because of their length and obscurity, it is
better and more useful to elucidate according to their sense alone than to recite
word for word. And given that they can be recited word for word [by the preacher],
where the authorities are obscure at all, their meaning can be set forth in other plain
words, for, when they are not understood by the listeners they lack all profit.45
Waleys makes several interesting distinctions in this section, beginning
with a basic one between sententialiter, ‘‘by the sense-units,’’ and verbaliter,
‘‘word for word’’ – a variation on the ancient distinction between memory
for matter and memory for words. Synonyms for these two abverbs are
summatim and verbatim. Waleys’s advice is directed towards the delivery of
an already composed sermon; the preacher should retain his sentence or
subject matters, and deliver his sermon with the aid of his memoria
(‘‘memoriter retinere et dicere’’). This procedure is preferred to exact
recitation for reasons of security as well as elegance.
Sententialiter may also have a technical meaning in the context of
mnemonics, one that links the mnemonic value of the colon divisions
marked off in a written text with the advice to remember ‘‘by the senten-
tiae.’’ A sententia was not merely an impressionistic division, but, according
to a well-known definition by Isidore of Seville, coincides with a colon; it is
a coherent though not complete semantic unit, and a number of such cola /
sententiae make up the completed, distinct thought that is a periodus.46 So
Elementary memory design
115
remembering material sententialiter would mean to remember it in chunks
the equivalent of colon-divisions, by its constituent ideas or sententiae,
rather than by word-for-word reiteration. Two verbs, retinere and dicere,
are used by Waleys with the adverb sententialiter; by contrast, the verb
recitare is used twice with verbaliter. Recitare is the verb also used for the
elementary school-room exercises in which children train their memories
first by rote, word-for-word.
Essentially what Thomas Waleys says here is also Cicero’s advice in De
oratore and elsewhere, that one prepare a speech for delivery by remember-
ing it according to its topics and major parts, rather than word for word.
For Thomas Waleys, the fault of an overly ingenious preaching style is that
one may literally lose one’s way in it, for one must learn an ornate
composition word for word, and that method risks losing everything in
the memory loss of a syllable.
But, Thomas Waleys is quite clear, one does not choose absolutely
between summatim retinere and verbatim recitare. Waleys says that the
preacher knows his memorized texts verbatim, and then adapts them to
the occasion and to the circumstances of his auditors – ‘‘dato quod verbal-
iter recitentur,’’ ‘‘given that [the texts] can be recited word for word [by the
preacher].’’ Whether or not we believe that all preachers conformed to this
behavior, it is significant that Waleys expects it as a norm. Fortunatianus
gave similar advice: we should learn texts verbatim whenever we can, but,
so long as we are careful to convey the substance of our original, we may
accommodate it in our own words to the occasion. Those accustomed to
medieval texts are familiar with the results.
Some years ago Theodore Silverstein recognized that Adelard of Bath
had certainly had direct knowledge of Cicero’s De natura deorum. 47 His
demonstration depended on recognizing that Adelard had reproduced the
res of a passage in Cicero, though he had not quoted it verbaliter. The Latin
of the two texts, Silverstein wrote, is similar ‘‘in the way the ideas . . . march
together step by step, but their language is not quite the same.’’
Nonetheless, Silverstein boldly rejected the notion that an (unknown)
written intermediary had to have intervened between Adelard and Cicero
to account for these differences. He writes, ‘‘such a text . . . has thus far not
appeared in evidence and will, one may guess, be difficult to find.’’ Once
the phenomenon of memoria summatim is recognized by modern scholars,
textual genealogies can allow greater flexibility and perhaps less speculation
than they have in the past..
Even Scripture is altered to accord with the particular occasion a writer is
addressing. An example occurs in a section of Hugh of St. Victor’s De archa
116
The Book of Memory
Noe, a meditation on monastic virtues composed about 1140.48 Hugh’s
theme is that the tree of wisdom is firmly rooted in faith, and he is chastizing
those sinners who flaunt their lack of faith, not fearing the Lord. He quotes
Psalm 72:2–3, but slightly alters the text to make it emphasize his immediate
occasion: ‘ Mei autem pene moti sunt pedes, pene effusi sunt gressus mei;
quia zelaui in peccatoribus pacem peccatorum uidens.’’49 The Vulgate reads
‘‘quia zelaui super iniquos.’ Frequently Hugh will add an inquit or an autem
to a text to improve the transition be
tween his words and those of a
quotation (scholars still do this, of course), or he will change an inflection
to fit the syntax of his own sentence. He also often incorporates fragments of
text into his own words, making them virtually into items of his own
vocabulary, the sense of quotation being almost lost. So Proverbs 8:31,
where Wisdom speaks of herself ‘ ludens in orbe terrarum; et deliciae meae
esse cum filius hominum,’ 50 is quoted as follows: ‘ Hec est sapientia, quae
aedificavit sibi domum, quia ludens in orbem terrarum delicias suas dicit esse
cum filiis hominum.’ Such adaptive freedom is enabled by complete familiar-
ity with the text, the shared memory of it on the part of both audience and
author, and hence a delight both in recognizing the familiar words and in the
skill with which they have been adapted to a new context. It is also evident
that superficial changes in wording like these were not regarded as affecting
the text’s substance, even in the case of Scripture. They fall within medieval
norms of correct quotation and use.
Memorizing sententialiter greatly increases the efficiency of recollection.
George Miller has observed that a hypothetical sentence can be thought of
as 100 letters or 25 words or 3 phrases or 1 proposition. If one considers a
sentence as 1 proposition, one will obviously be able to grasp it more
effectively and securely than if one regards it as 100 letters. It is, says
Miller, ‘‘those larger, subjective units, loosely called ideas, that we must
count to determine the psychological length of any text.’’51 This is what
ancient and medieval writers called memoria rerum.
Hugh of St. Victor, in Didascalicon, describes his own practice as a student,
in terms of his favorite image of the trained memory as an orderly money-bag:
How many times each day would I exact from myself the daily debt of my bits of
sophomoric wisdom [sophismata], 52 which thanks to the principle of shortness
I had symbolized in one or two maxims in abbreviated form, so that indeed I held
in my memory both the payouts [solutiones] and the numerical order of virtually
all the propositions, questions, and objections which I had learned. 53
In this context, solutio refers to the information ‘‘paid out’’ from memory’s
store, as the skillful money-changer pays out coin. This store is in the form
of pieces of wisdom, sophismata or dicta, the bits into which texts are
Elementary memory design
117
divided. And they are marked for recollection (signare) by having been
reduced, according to the rule of brevitas, to one or two sayings ‘‘in
abbreviated form’’ (in pagina) and numbered in order.
Hugh uses the word pagina in a specialized manner here that is recog-
nized only in dictionaries which cover late Roman and medieval Latin: Du
Cange and the Thesaurus linguae latinae. Pagina, like (s)cedula (discussed in
the next part of this chapter), refers both to a type of physical surface and,
by extension, to the sort of thing written on such a surface. Latin pagina is
derived from the verb pango, ‘‘to fasten,’’ and refers to the rectangular
boards covered with wax that, fastened together, made up a set of tablets.54
The constraints of the surface required writing in columns. Obviously,
pagina refers to such physically shaped surfaces, as when Cicero, for
example, writes of ‘‘filling a page’’ (complere pagina).55
But what one wrote on a pagina was what we call notes, that is,
ephemeral memoranda, abstracts, computations, the rough drafts of com-
positions that were to be fully composed later, either in oral delivery or
written onto papyrus or parchment in fair hand. The meaning of pagina
was extended in later Latin to include this other meaning – a memoran-
dum, something that is in a form abbreviated for further study and
application.56 Du Cange cites a letter of Ambrose, in which he says to his
correspondent that he is sending him a little present, this short letter,
because his friend wanted him to paginare, ‘‘examine in brief and informal
form,’’ some things concerning the interpretation of the Old Testament. 57
Similarly, Paulinus of Nola wrote that the laws of God were written on
stone in pagina, meaning they were set down breviter, without all the
commentary. 58 In medieval Latin, one finds the word paginator, the person
who ornamented the pages with pictures and drawings. This task is
associated with the work of the rubricator or miniator (the word from
which Modern English ‘‘miniature’’ derives), the person who worked with
red pigments to mark the text with short-form summaries, digests, and
pointings of its contents, in pagina. It is worth noting that rubricare and its
vernacular derivative rubric had dual meanings also. 59
After one has divided a text, clearly marked its constituent pieces, and
placed the segments onto one’s numerical grid, the next principle, accord-
ing to Hugh’s Preface, is to memorize each bit in such a way that one sees it
clearly and distinctively. One should always read a text from the same
codex, so that the features of the page on which the particular segment of
text appears become part of one’s mnemonic apparatus. ‘‘Indeed I consider
nothing is so useful for stimulating the memory as this,’’ Hugh writes.60
The layout of the many folio leaves of names, places, and dates which
118
The Book of Memory
comprise his Chronicle was evidently designed by Hugh himself. Each page
is in four columns, with headings in red introducing groups of ten items or
less (figure 5).
Formatting the page of memory
The lay-out Hugh adopted for his Chronicle was clearly influenced by that
of the Canon Tables of Eusebius, added to virtually all early medieval
Bibles.61 Compiled in Greek around 331, they were later translated into
Latin. The elements of their design, which show parallel passages in the
four Gospels, are very old and scarcely vary over their history (a ninth-
century example is reproduced in figure 6). Passages are cited by number
(including numbers over 200 – instead of chapters and verses, Eusebius
divided the Gospels into sections) without any written text. 62 There are ten
tables in all, the first showing concordant passages in all four Gospels; the
next three show concordances among three of the four; the next five those
among any two; and the last one lists passages unique to each of the four.
The tables are laid out in columns (one meaning of in pagina), the numbers
listed one after another vertically, and architectural columns are drawn to
separate the four main vertical spaces on the page, together with arches and
other architectural elements representing a classical fac¸ade.
It has been suggested that, in this context, an arcade motif may derive
from the ancient mnemonic advice to use buildings – including interco-
lumnia, the spaces between columns – as backgrounds for things to be
remembered. 63 Certainly intercolumnia is one of the most enduring types
of memory locus. Within each rectangular space made by the columns in
the Eusebi
an Tables, the name of the gospel is written at the top, and then
the chapter numbers of the synoptic passages are recorded. Horizontal
lines, sometimes colored, are drawn between every four numbers (in the
Greek text) or five (in the Latin); the effect is to divide the page into a series
of small rectangular bins, none holding more than five items. Such a layout
is clearly designed for mnemonic ease; as Hugh of St. Victor tells his
students about his own similar design: ‘‘Now indeed you have enough to
do to imprint in this fashion in your memory the matters which are written
out below, according to the method and diagram for learning by heart
demonstrated to you earlier.’’64
The presentation of the manuscript page in medieval Bibles provides
one of the most interesting contexts for understanding Hugh’s injunctions
to his young pupils concerning the need to divide material for memorizing,
the mnemonic value of textual lay-out, and the requirement always to use
Elementary memory design
119
5. Paris, Bibliothèque nationale de France MS. lat. 15009, fo. 3v (French, twelfth century).
Hugh of St. Victor’s column format for his Chronicle. The headings and section markers
are in red and blue ink, alternating. From the library of St. Victor, this is likely to be
either the archetypal manuscript of this text, or very close to it.
120
The Book of Memory
6. New York, Pierpont Morgan Library MS. M. 860, fo. 8v (France, between 857
and 862). A page of Canon Tables from a Gospels made in the monastery of
St. Martin in Tours.
Elementary memory design
121
the same codex. Typical of so-called ‘‘Paris format’’ is an English Bible of
the late thirteenth century, now in the Huntington Library (EL 9 H4).
Each book is divided into numbered chapters and subdivisions called cola
and commata, using a scheme devised in Paris at the end of the twelfth
century by the English scholar, Stephen Langton, later cardinal and arch-
bishop of Canterbury.65 Each chapter begins with a colored initial, alter-
nately red and blue. The text of each chapter is presented in prose form,
unbroken except for the punctuation of the cola and commata. Each
The Book of Memory Page 22
