few words (the incipit) of each psalm, so that as one visualizes the number
one, one simultaneously visualizes ‘‘Beatus vir qui non abiit’’; upon seeing
the number twenty-two one also sees the text ‘‘Dominus regit me’’; and so
forth. In Hugh’s scheme the images are the written words as they actually
appear in a manuscript and the locus is simply a numbered box, but the
incidental difference of this scheme from the architectural one is less
important than its fundamental psychological similarity; they both employ
a system of consciously adopted, rigidly ordered backgrounds as a grid
which is then filled with the images constituted by the text.
Hugh also counsels that this same method of numerical ordering can be
used to learn the text of an individual psalm. Under number twenty-two,
for example, one might visualize a subsidiary set of consecutive numbers.
102
The Book of Memory
To these one attaches the rest of the text, in short pieces (verses) – however
many one needs to complete the task. The crucial task for recollection is the
construction of the orderly grid through which one can bring to mind
specific pieces of text. This enables one ‘‘when asked, without hesitation
[to] answer, either in forward order, or by skipping one or several, or in
reverse order and recited backwards’’ whatever is in the memorized text as a
whole. And it also enables one to construct mentally a concordance of the
text, thus disputing quickly and surely, making citations to authorial texts
by number alone.9
This scheme will work for any book of the Bible (or for any text at all, for
that matter). A long text must always be broken up into short segments,
numbered, then memorized a few pieces at a time. We have some clue as to
just how short ‘‘short’’ was from the length of verses in the medieval format
of the Psalms, and from the number of words enclosed in cola and commata
divisions. Obviously, optimal length varies slightly from one individual
memory to the next, but the medieval texts of the Psalms generally contain
more verse divisions than do modern texts. Psalm 23 (Vulgate 22), for
instance, is six verses long in the printed King James Bible, but has nine
divisions in the Latin Paris text of the thirteenth century. The longest of
these is the first, containing thirteen words; by contrast the King James
contains thirty words in verse four and twenty-two in verse five. The fewer
number of words per division in the medieval format accords with the real
limits of human working memory, that is, how much one can safely take in
during a single conspectus, or mental glance (to use the terminology of the
memorial artes).10
Hugh’s mnemonic advice is addressed to students, but not to children
just learning to read. This is evident when we ask ourselves at what point in
a monastic education such a work would become useful. The linea or
number grid he describes imagines the psalms in their textual order, from 1
onward to 150. The heavy emphasis which is placed upon the incipit or first
phrase of each psalm is also noteworthy. Though, he says, one may make
subsidiary cells in the grid for each verse of a psalm, its initial words receive
Hugh’s most careful attention, as the cues and starting-off points for
recollection of the whole. Yet, as described, this exercise seems at odds
with what we know of the monks’ usual recitation practices, and thus needs
some further exploration.
At St. Victor, as in all monasteries at this time, novices first learned the
Psalms by heart in the liturgical offices, as they had for centuries, chanting
the words to melodies, perhaps accompanied as well with processional
movements and postures. Entire texts became rote-retained through this
Elementary memory design
103
exercise. The liturgical order of psalms is not, however, that of their textual
order. Rather, the liturgical order groups all the Psalms in weekly cycles,
assigning them to particular offices in a variety of ways that is quite
independent of textual order. So if a novice already had all the Psalm
texts by rote from chanting the liturgical offices, what could be the point of
learning them again in a different, textual order by means of the grid device
Hugh describes? Evidently it was not to learn the Psalms by rote – that had
already been achieved. Those pueri whom Hugh addressed already had
their rote memory of these texts.
The goal must be further study, to be able to find a specific text without
the need to repeat the whole liturgical sequence. And that is what Hugh
counsels: ‘‘dispose it in such a manner that when your reason asks for it, you
are easily able to find it.’’ When your reason asks for it refers to the rational,
investigative procedures of recollection. The device Hugh describes is good
not for habitual, rote retention (melody, gesture, and physical location are
superior for that purpose), but for quick recollection and retrieval, for
creative study and for invention in situations such as preaching, teaching,
and debate. Hugh’s advice is for more advanced students, not for begin-
ners. Indeed, the lists of historical markers in the Chronicle which this
advice prefaces are not matters for beginning readers, but for serious
students of the Bible.
The numerical grid provides random, multiple access points to material
already in memory, enabling someone to work with and make something
new from their rote memory store. The mnemonic cue, which is the
incipit of the verse, crucially provides the starting point for recollection of
each rote-retained textual chunk. And if a text has been retained with care
and frequently practiced (conditions provided for by the divine office),
each cue will reliably call forth the whole. 11 The essential task is to impress
rote memories as a complete experience, made as particular as possible. To
that end:
we [should] also pay attention carefully to those circumstances of our matters
which can occur accidentally and externally, so that, for example, together with the
appearance and quality or location of the places in which we heard one thing or the
other, we recall the face and habits of the people from whom we learned this and
that, and, if there are any, the things that accompany the performance of a certain
activity. All these things indeed are rudimentary in nature, but of a sort beneficial
for youths. 12
Advice to pay particular attention to manuscript layout, where and how a
text is presented on a page, is commonplace, and I will discuss other
examples of it shortly. Hugh counsels as well to set our memories within
104
The Book of Memory
the seasons and times, and the performative expessions and gestures that
occured when one first learned the verses. Such experiential cues are
perhaps best explained with reference to liturgical performances, for
which time and season are crucially important. Thus Hugh addresses
situations of memorizing both from written books and also from physical
performance. The two modes reinforce one ano
ther, both providing addi-
tional experiential cues to the rote-retained segments, which help to recover
them richly within one’s investigative grid.
Hugh’s advice to divide the text duplicates that of Quintilian, as it also
articulates the experience of Shereshevsky, who had difficulty remembering
long passages of connected words, but none at all in retaining an apparently
limitless number of short segments. Any long text can be treated as though
it were composed of a number of short ones:
For the memory always rejoices [says Hugh] in both brevity of length and fewness in
number, and therefore it is necessary, when the sequence of your reading tends
toward length, that it first be divided into a few units, so that what the memory
could not comprehend as a single expanse it can comprehend at least in a number. 13
This is advice which Hugh repeated in Didascalicon. In this treatise
Hugh extolls even more the dependency of all wisdom (sapientia) and the
liberal arts upon memory, training which is now sadly decayed because
students do not learn proper habits:
We read that men studied these seven [arts] with such zeal that they had them
completely in memory, so that whatever writings they subsequently took in hand
or whatever questions they posed for solution or proof, they did not thumb the
pages of books to hunt for rules and reasons which the liberal arts might afford for
the resolution of a difficult matter, but at once had the particulars ready by heart.
Hence, it is a fact that in that time there were so many learned men that they alone
wrote more than we are able to read.14
Hugh’s Preface shows this same contempt for the cumbersome, inefficient,
and happenstance method of turning the pages of a book to look for a text
that one needs. Do you think, he asks the students, that people wanting to
cite a particular Psalm turn over the pages of a manuscript hunting for it?
‘‘The labor in such a task would be too great.’’15 It is also striking that Hugh
makes an exact correlation here between the amount stored in one’s
memory and the amount of written composition one produces.
But Hugh’s most practical advice in Didascalicon, a work which teaches
the arts of reading as its title indicates, is to gather (colligere) while reading,
‘‘reducing to a brief and compendious outline things which have been
written or discussed at some length.’’16 Again the principle of dividing a
long text (prolixius) is to be observed, because, Hugh says, the memory is
Elementary memory design
105
lazy and rejoices in brevity.17 Therefore, we ought to gather something brief
and secure from everything we learn, which we can store away in the little
chest of our memory. 18
One should not assume that Hugh meant that one should retain only a
compact summary of what one has read; what he means is that one should
break prolixity, a long text, into a number of short, securely retained segments
which can be gathered in the memory. This method of study certainly leads to
florilegia, as it leads also to Bartlett’s Familiar Quotations and The Oxford
Dictionary of Quotations. The phrase ‘ brevem . . . et compendiosam sum-
mam,’’ ‘ a brief and compendious summary,’’ might seem self-contradictory,
except that Hugh is clearly giving the same advice he spells out more fully in
his Chronicle Preface, memorizing a compendious summation of brief seg-
ments of the text one is trying to master – the scholar’s method of note-taking,
in other words, except written in the memory instead of on note-cards. It is
worth recalling by those who might dismiss such advice as mere florilegiality,
born of distaste for the comprehensive knowledge of a text, that note-taking
and serious scholarship are not exclusive activities. How compendious the
summation of a text might be would depend on the industry and talent of each
individual reader, and the importance to him of a particular text.
This principle of grouping or ‘‘gathering’ respects the limits of working
memory. It is called ‘ chunking’ in neuropsychology now.19 While the
storage capacity of memory is virtually limitless, the amount of information
that can be focused upon and comprehended at one time is definitely
limited, to a number of units somewhere between five and nine; some
psychologists express it as a law of ‘ Seven plus-or-minus two.’ 20 So one of
the fundamental principles for increasing mnemonic (recollective) efficiency
is to organize single bits of information into informationally richer units by a
process of substitution that compresses large amounts of material into single
markers. In this way, while one is still limited by one’s capacity to focus on
no more than six or seven units at a time, each unit can be made much richer.
As the psychologist George Miller has written (without being aware that he
was echoing one of Hugh of St. Victor’s favorite images), if my purse holds
only six coins I can carry six pennies or six dimes; similarly, it is as easy to
memorize a list containing a lot of information coded into ‘‘rich’’ units as it is
to memorize one containing ‘ poor’ units, for the limiting factor is the
number, not the nature of each item. Miller describes grouping in this way:
The material is first organized into parts which, once they cohere, can be replaced
by other symbols – abbreviations, initial letters, schematic images, names, or what
have you – and eventually the whole scope of the argument is translated into a few
symbols which can be grasped all at one time.21
106
The Book of Memory
This is exactly what Hugh counsels doing when he substitutes number-
coordinates for the verses of a Psalm – the active, working memory first
focuses on the number, and then that numerical address leads to the text
placed within it, itself composed of a few words at a time, grouped into
phrases. In recollection, one first focuses on the informationally richest
sign, say ‘‘Psalm 23.’’ That stands in for a set of six sub-units or verses (in the
King James Version), and one might focus on one number amongst those,
perhaps the number two. That sign in turn both stands for and cues the
words ‘‘He maketh me to lie down in green pastures; he leadeth me beside
the still waters,’’ themselves grouped into five syntactic sub-units or
phrases. A fairly common method of citing texts in commentaries and
other learned material, especially Psalm texts (the most widely known by
heart), was by the initial letters of each word, so that, for example, ‘‘Beatus
uir qui non abiit’’ (Ps. 1:1) could be written ‘‘b. u. q. n. a.’’ This is another
substitution process of the sort Miller adduces, a means of making psycho-
logically richer units. It saved work for the scribe, but it also served as a
memory note of effective brevity, for readers who were expected to know
the words referred to such signs.
Because of the substitution process that creates rich units, one can skip
material, rearrange it, collate it, or whatever, simply by manipulating a
few digits mentally – recalling the second verse of ev
ery Psalm, perhaps, or
reciting a couple of Psalms by alternating verses from one with the other,
maybe one in ascending order and the other in the reverse. Any number of
impressive parlor-tricks (ancient and medieval pedagogy would have
called them exercises) can be played, for one is actually just counting a
few digits at a time. One can also use the ability conferred by this process
for serious ends, such as marshaling texts on a particular topic, as Hugh
suggests in his Preface, by proferentes numerum, ‘‘fetching forth . . . its
number,’’22 that is, memorizing only the numerical coordinates to a text
under the topical key-word. These coordinates then trigger recollection of
each separate text. It would indeed be possible to conduct a substantive
discussion of doctrine by numerical citations alone among scholars who
knew their Bible in this fashion.
Division and composition
There is a trope in late medieval university prose called a divisio or a
discretio: we might call it a summary outline or even a set of study notes.
Such a ‘‘division’’ comprises the main topics of a particular subject under
study, and one finds them in the margins of academic manuscripts such as
Elementary memory design
107
the Decretals and the Sentences, works designed for commentary and
disputation. A fine example is in the fifteenth-century Bolognese manu-
script of the Decretals (a canon law text) reproduced in figure 4. This
divisio concerns papal powers, and the subdivisions of the topic spew forth
from the beak of a memorable imperial eagle, each written on a wavy line
emanating like speech from the eagle’s mouth. The whole image provides a
unique, succinct, and effective cue for these subject matters, indicating how
they are related in content and what their underlying theme is, for these
powers all relate to the crucial fourteenth-century tension between the
papacy and the lay monarchs who both sponsored it and fought each other
through its offices and influence.
Two late Roman grammarians, Consultus Fortunatianus and Julius
Victor, define the process and goal of division clearly, and since each was
quite influential throughout the history of medieval pedagogy, it is worth
The Book of Memory Page 20