word omnia. It got entangled in noise and I recorded omnium . . . and the
more people talk, the harder it gets, until I reach a point where I can’t make
anything out’’ (39).
S. had trained his memory ‘‘for words’’ rather than ‘‘for things,’’ in the
ancient sense. His grasp of entire passages of a text, his ability to form
coherent images associated with the meaning of a long passage, was far
from good. If a story were read to him, especially at a rapid pace
(S. required a pause of 3–4 seconds between items in order to form and
place his images), he would complain that, since each word called up an
image for him, they began to collide with one another in a hopeless muddle
(65). He was best at recalling items in a series, or brief connected passages
which he could remember as though they were parts of a scene. His
capacity for recalling thousands of such series, however, was virtually
limitless in space and, over time; indeed, S.’s chief problem, especially
after he became a professional performer, was how to forget, a problem
which presented itself to him as how to erase images he had formed. He
96
The Book of Memory
tried mentally burning an image sequence he no longer needed, and he
tried writing it down (unwittingly putting to the test Thammuz’s fears
about the danger of substituting writing for memory). Thammuz needn’t
have worried. Neither technique had the least effect on S.’s memory.
Finally S. realized that he could simply will a specific series of images not
to appear, and they would not come up (66–71). This clear indication of his
conscious control was a great relief to him.
One test Luria set for S. was to remember the first four lines of the
Inferno. S. knew no Italian, and so made images of Dante’s words from
Russian homophones. This was the method he described using to memo-
rize the first line, ‘‘Nel mezzo del cammin di nostra vita’’:
Nel – I was paying my membership dues when there, in the corridor, I caught sight
of the ballerina, Nel’skaya.
mezzo – I myself am a violinist; what I do is to set up an image of a man who is
playing the violin [Russ.: vmeste], together with Nel’skaya.
del – There’s a pack of Deli Cigarettes near them.
cammin – I set up an image of a fireplace [Russ.: kamin] near them.
di – Then I see a hand pointing to a door [Russ.: dver].
nostra – I see a nose [Russ.: nos]; a man has tripped and, in falling gotten his nose
pinched in the doorway [tra].
vita – He lifts his leg over the threshold, for a child is lying there, that is, a sign of
life – vitalism. (45–46)
His system for the other three lines was similar, though he devised a
different scene of related images for each line. Sixteen years later, without
warning, Luria asked him to repeat these lines for him and he did so,
flawlessly and effortlessly.
Shereshevski’s technique for recalling after so long an interval is also
discussed by Luria. It made no difference to S. whether a series were
presented orally or in writing. He had practiced his imagining technique
to the point where he could find images quickly even for words he had
never heard before; for common words, he tended to use the same images
over again. Once remembered, he had no difficulty in recalling, whether
the interval had been a day, a month, or many years. But he needed first to
place himself mentally back in the situation in which he had first commit-
ted the series to memory. As Luria describes it:
S. would sit with his eyes closed, pause, then comment: ‘‘Yes, yes . . . This was a
series you gave me once when we were in your apartment . . . You were sitting at
the table and I in the rocking chair . . . You were wearing a gray suit and you
looked at me like this . . . Now, then, I can see you saying . . .’ And with that he
would reel off the series precisely as I had given it to him at the earlier session. If
Descriptions of the neuropsychology of memory
97
one takes into account that S. had by then become a well-known mnemonist, who
had to remember hundreds and thousands of series, the feat seems even more
remarkable. (12)
One final characteristic of S.’s method is worth remarking. Luria stresses
that the highly intricate technique he had developed was at its basis entirely
natural to him, essentially a matter of spontaneous recollection by his
memory aided by consciously applied method. The ‘‘techniques he used
were merely superimposed on an existing structure and did not ‘simulate’ it
with devices other than those which were natural to it’’ (63). The author of
Ad Herennium, introducing his strikingly similar mnemonic, comments
that ‘‘the natural memory, if a person is endowed with an exceptional one,
is often like this artificial memory [similis sit huic artificiosae], and this
artificial memory [haec artificiosa], in its turn, retains and develops the
natural advantages by a method of discipline’’ (III. 16.29). The use of the
demonstrative pronoun in the Latin text seems to me significant. This
method, the author is saying, builds upon the memory’s natural disposi-
tions. So too S.’s memory system – to say nothing of that of the Japanese
mnemonist who used a similar one – was natural, developed by practice but
entirely self-generated and self-taught.
There are a number of ways in which Luria’s study of S. can provide useful
insight into the basic neuropsychology of the descriptions we have of the
architectural and other mnemonic schemes. First, Luria was especially
impressed by S.’s synaesthesia, or apparently simultaneous and unsorted
sensory perceptions, especially of ear, eye, and nose (though all five senses
were involved). He ‘ saw’ sounds as visual shapes, and his memory-images
often produced sensations of vibration or smell for him, as well as sight. This
natural tendency was greatly increased by his mnemonics, for everything
coming in to his mind was turned into an image. The essentially visual
nature of these images was psychologically crucial. Sounds were only con-
fusing to S. unless they were translated into particular clear visual images. It
was the images which provided the stability over time to S.’s recollective
power. Medieval memory advice stresses synaesthesia in making a memory-
image. Memory-images must speak, they must not be silent. 131 They sing,
they play music, they lament, they groan in pain. They also often give off
odor, whether sweet or rotten. And they can also have taste or tactile
qualities. I will have more to say about this later, but it is worth commenting
on here, since synaesthesia was considered to be so vital to S.’s success.
In the second place, S.’s process of recollection was a process of percep-
tion; he mentally walked through his memory places and looked at what
was there. This accounts for the essentially perceptual nature of ancient
98
The Book of Memory
advice on the preparation of loci, that they be properly lighted, moderate in
size and extension, different enou
gh from one another in shape that they
can’t be confused, not crowded, that the images in them be of moderate
size and neither too large nor too small, and that they be made clearly
distinct from their background.
Third, S.’s problem with overloading, experienced when words were
read to him too rapidly, or when he tried to grasp a lengthy passage as a
whole, suggests why all the ancient Roman writers, from the Ad Herennium
author through Quintilian, had their doubts about the ordinary utility of
‘‘memory for words’’ when tied to images like these. Fourth, and as a
counterbalance to the previous point, S.’s ability to recall thousands of
short series of words, even though he had trouble grasping long connected
passages, sheds light on the ancient advice to memorize lengthy passages
piecemeal, in short sections which can later be united. As Hugh of
St. Victor writes, the memory rejoices in shortness (‘‘memoria brevitate
gaudet’’); the experience of S. should caution us, however, against assuming
that in saying this Hugh meant that only brief extracts of longer works
could be retained in the memory. Any long work can be considered as a
number of short series joined together – indeed, this is exactly Quintilian’s
advice. And as separate series, a virtually limitless number can be retained
in a good, trained memory.
Fifth, and in some ways the most important for our purposes, S. was
completely uneducated in mnemonics. The system of backgrounds and
images he devised was self-taught, refined by trial and error. He never
seems to have discovered the principle of ‘‘memory for things’’ that every
ancient writer on the subject is at pains to emphasize as the key to the
successful composition and delivery of an oration. This, surely, was because
his prodigious feats of memory were treated as freakish, and he himself as a
vaudeville act, his art perceived to be a mere curiosity without social
usefulness and ethical value. Moreover, he was rewarded only for his
iterative accuracy – creative composition was not his aim. The situation
in the ancient and medieval worlds, needless to say, was far different. By
providing such tantalizing parallels, S. can teach us how the memory
system described in these treatises is true to human psychology, but there
is much about the actual use of memory in education and composition
during the Middle Ages that S.’s experience cannot tell us.
C H A P T E R 3
Elementary memory design
T H E N U M E R I C A L G R I D
The architectural mnemonic was not the only, or even the most popular,
system known in the Middle Ages for training the memory. In her 1936
study of the subject, Helga Hajdu mentions various alternative systems,
including an elaborate digital method of computation and communication
that also served mnemonic purposes, discussed in the treatise De loquela
digitorum ascribed to Bede; various mnemonic verses that serve both
scholars and laity, such as the university students’ mnemonic for remem-
bering the various types of syllogism (‘‘Barbara Celarunt Darii,’’ etc.) and
the execrable rhyming hexameters used by lawyers as a mental index to the
various collections of laws; and various counting devices, like the rosary
and the abacus, which involve manipulating physical objects in a rigid
order as an aide-me´moire in calculation. 1 There are rhyming catalogues of
medieval libraries which were intended to be memorized by the monks, 2
and there are Alexandre de Villedieu’s 4,000 rhymed hexameters setting
forth the rules of Latin grammar, the Doctrinale. One of the earliest and
longest-lived poems in English is the mnemonic for the months of the year:
‘‘Thirty days hath September, April, June and November.’’
Interesting as these particular items are, however, they are too limited in
purpose or too bound to a particular situation, such as the monastery or the
law court, to be of much interest in considering how memories were
trained by a wide variety of people to make a securely searchable inventory
of a variety of texts and other material (what Albertus Magnus lumps
together as negotia, a word usually understood as ‘‘business,’’ but in the
context of memory-training it translates better as ‘‘matters,’’ ‘‘content,’’ or
even ‘‘stuff’’). Other systems of memory training clearly surpassed that
described in the Ad Herennium, in both their longevity and their broad
applicability to all spheres of learning. And, unlike the Herennian mne-
monic, they were widely disseminated in the Middle Ages (not just the
99
100
The Book of Memory
humanist circles in papal Avignon and in Tuscany), and find their way into
practical writings on the subject of memory well before the fourteenth
century. 3 In this chapter, I will be concerned with two simple schemes used
in memory training. They employ the most common of locational heu-
ristics, the order of numbers, and that of the alphabet. In order to
demonstrate their ubiquity, I must appeal to the reader’s patient willing-
ness to pick the way through the eclectic evidence that I have gathered up in
this chapter.
Learning by heart: Hugh of St. Victor on memorizing the Psalms
How to use a number scheme is described by Hugh of St. Victor in a text
virtually overlooked by modern scholars, the Preface to his Chronicle, also
called ‘‘De tribus maximis circumstantiis gestorum,’’ translated in
Appendix A. Hugh’s description is one of the fullest, and clearest, of any
mnemonic system. The method utilizes psychological principles similar to
the method using images in background places, but in this case the system
of loci is numerical, and the images are short pieces of text written into the
numbered backgrounds, as though within the cells of a grid. The images of
written text are impressed also as they appear in the particular codex from
which they were first memorized, including their location on the page
(recto, verso, top, middle, bottom), the shapes and colors of the letters
themselves, and the appearance of each page including marginalia and
illuminations, to make a clear visual experience. Finally, Hugh advises that
the physical conditions under which one had memorized the original
material should also become part of one’s total recollection of it – in
modern psychological terms, to make of it an ‘‘episodic’’ not just a ‘‘seman-
tic’’ memory.
This Preface was composed about 1130, and is addressed to novices
(pueri) studying Scripture in the cathedral school of St. Victor. It precedes
a Chronicle of Biblical history, set out as columns – seventy folio pages in
the fullest versions – of names, dates, and places, which the students were to
memorize as an elementary part of their education in sacra pagina, or the
study of the Bible. The Preface was unpublished until 1943, when it was
edited by William M. Green (there appear to have been no previous
printings at all).4 The manuscripts suggest that the preface, while having
some local success, never
achieved the wide dissemination of Didascalicon,
the work which earned Hugh a reputation for being an important propo-
nent of ars memorativa in the later Middle Ages and Renaissance.5 There
are several manuscripts of the Preface alone; Green comments that, while
Elementary memory design
101
portions of the Chronicle tables were ‘‘often omitted with much resultant
confusion and variety in the manuscripts, the text of the prologue . . . is
usually well preserved.’’6 There are 34 surviving manuscripts of the Preface
(compared to 125 of Didascalicon), nearly all written in the twelfth and
thirteenth centuries, though there are 2 from the fourteenth, and 1 from the
fifteenth century. Of those whose provenance is known, most are claus-
tral.7 The largest number of manuscripts is French, as we would expect,
including several from St. Victor. So, the evidence suggests that this
particular treatise was not regarded as major or original enough to deserve
wide dissemination, despite its author’s eminence, and that it sank into
oblivion by the early fourteenth century, because it had been superseded by
or incorporated into other pedagogical tools. Its very ordinariness, how-
ever, makes it important to my study.
Hugh’s method displays the principles basic to classical mnemonics, as
we find them described by Aristotle, Tullius, Cicero, and so many others.
One must have a scheme of places in a rigid, easily retained order, with a
definite beginning. Into this order one places the components of what one
wishes to memorize and recall. As a moneychanger (nummularius)
separates and classifies his coins by type in his money-bag (sacculum,
marsupium), so the content of wisdom’s store-house (thesaurus, archa),
which is the memory, must be classified according to a definite scheme.
Without retention in the memory, says Hugh, there is no learning, no
wisdom. ‘‘Indeed, the whole usefulness of education consists only in the
memory of it.’’8
The example Hugh gives is how to memorize the Psalms. There are 150
in all, and to learn them one constructs a series of mental compartments,
numbered consecutively from 1 to 150 – in other words, a rigid system of
cells that has a definite starting-point. To each number is attached the first
The Book of Memory Page 19
