Ars oblivionalis, ars inveniendi:

The Cherub Figure and the Arts of Memory*

MARY J. CARRUTHERS
New York University

Abstract call creative thinking. The need for structured memory stor-
age was understood as a support for making new thought and
To serve their purpose well, many so-called mnemonic im-
composition, not for simply preserving all the past. I will soon
ages in the Middle Ages facilitated meditation and invention by
presenting many rich materials in a highly abbreviated form, return to this point, but I state it clearly now as well: the rote
which could be expanded and recombined for a variety of com- reiteration of a memory store is not what was intended when
positions. To abbreviate fruitfully requires rigorous compres- medieval writers spoke of ars memorativa. In both dialectic
sion and selection, a kind of forgetting that was distinguished and rhetoric, memory craft was practiced as a tool of invention,
both in theory and practice from rote recitation or learning
rationally discovering and selectively recombining things one
by heart. The twelfth- and thirteenth-century diagram called
the Cherub offers an excellent example of how such an image had previously learned for a particular purpose. In that sense,
was used in study and composition. Focusing on six versions recollecting must always also involve some forgetting.
of it, this essay demonstrates that the medieval cherub image During his elegant, eloquent encomium on the extraor-
is not an illustration tied to any particular text but functioned dinary depth and power of his memory, St. Augustine writes:
independently as an analytical tool, an art for inventing argu-
ments, which incorporated the methods of medieval dialectic
and rhetoric. I arrive in the fields and vast mansions of memory, where
are treasured innumerable images brought in there from
objects of every conceivable kind perceived by the
It may seem odd to begin an essay on the arts of memory senses. There too are hidden away the modified images
by talking about forgetting, indeed, about an art for doing so. we produce when by our thinking we magnify or dimin-
Surely, one would suppose, the purpose of an art of memory ish or in any way alter the information our senses have
is exactly to forestall forgetting, to ensure that one had every reported. There too is everything else that has been con-
single experience and bit of knowledge accumulated in a life- signed and stored away and not yet engulfed and buried in
time at one’s immediate call, that one can in fact make of one’s oblivion. . . . The huge repository of the memory, with its
mind a universal encyclopedia of all that has ever been said or secret and unimaginable caverns, welcomes and keeps all
thought, rather as matters put on the Internet today are said to these things, to be recalled and brought out for use when
be incapable of erasure. People sometimes say to me, when needed; and as all of them have their particular ways into
I tell them that I work on arts of memory, “Well, I am more it, so all are put back again in their proper places. . . . This
concerned about forgetting.” They say this as though they I do within myself in the immense court of my memory,
were saying something clever and original. They aren’t: The- for there sky and earth and sea are readily available to me,
mistocles, the great Athenian admiral, supposedly asked to be together with everything I have ever been able to perceive
taught an art of forgetting because his memory was already so in them, apart from what I have forgotten.2
crowded that it needed refreshing. The emphasis in the ancient
accounts is on Themistocles’ remarkably complete and immedi- “Apart from what I have forgotten.”. . . In the cheerful admis-
ate recollection; the admiral’s playful wish to learn forgetting, sion of that phrase lies an essential divide between a modern
in other words, directly attests to the copiousness and secu- and a medieval understanding of the cognitive function of mem-
rity of his mnemonic inventory.1 Forgetting has always been ory. To have forgotten things is seen by us now as a failure of
a necessary part of the craft of remembering. Failing to rec- knowledge, however ordinary a failure it may be, and therefore
ognize this elementary condition indicates a basic misunder- a reason to distrust the power of memory altogether. Yet to have
standing about the purpose and function of mnemonic craft and forgotten some things was understood in Augustine’s culture as
about the nature of the mental techniques involved, including a necessary condition for remembering others.
most famously the making of elaborate mental images placed It is helpful to distinguish two sorts of forgetting, resulting
in multi­chambered constructions. From antiquity, the arts of from different causes. There is the kind that happens because
memory in Europe were conceived of as investigative tools for one failed to record something in the first place, the sort of thing
recollective reconstruction and selection, serving what we now Augustine is talking about here. This should not even be called

GESTA 48/2 © The International Center of Medieval Art 2009 1

forgetting because, as Aristotle makes clear in his treatise But how can this be? Here, I must return to what is meant
On Memory and Recollection, one cannot properly be said to by rote memory. Memorizing by rote or mechanically is a
remember anything until one has a mental image of it impressed human capacity that certainly has its uses, even life-saving
in memory, which one can then later recall.3 The other kind is ones, as when machine operators are trained to act so habitu-
deliberate and selective forgetting, the sort of forgetting that ally that the steps in a safety procedure come to them without
itself results from an activity of memory. One can demonstrate thinking. Those valuable uses, however, are often mistakenly
this from the so-called artes oblivionales found in a few late conflated with accuracy of reproduction, as though rote mem-
humanist treatises on memory art.4 The oblivion discussed has ory and accurate memory were necessarily linked. This is an
to do with how to refresh one’s search networks, not with wor- easy error to make. I have done so several times myself, but I
ries about the accuracy or partialness of one’s memories. As have concluded that it is a seriously misleading error. 10
Lina Bolzoni has commented, “The techniques for forgetting By the phenomenon of rote memory we mean iteration,
handed down by the treatises are testimony to the persistence the ability to repeat things without fail in the same order over
and power of the images,” for they address tasks of sorting out and over. The virtue of rote—in a life-threatening situation, for
and reducing the number of memory cues rather than suppress- example—lies precisely in its mindlessness. Because there is
ing or otherwise editing content one has previously learned.5 only one way to do it, something can be done with great reli-
As the story is told, Themistocles was motivated by this same ability, like a tool that can perform only one task. But this is
need, to simplify his memory store and refresh his recollection because it can be done only in one way. In other words, what is
by reducing its “crowdedness.” In memory technique, crowd- most distinctive about rote is not its accuracy but its iterability.
ing—either by cramming so many memories into stored places We might say it can be accessed only along one line, and, once
that one can no longer “see” them clearly or (even better) by the procedure is started, it will unfold in the same direction in
deliberately covering one set of memories with others so as to the same steps until it stops, over and over. To use an analogy
rearrange their links and pathways—is always presented as the from computer programming, rote memory provides a read-
best way of forgetting something. The underlying assumption only memory, vital, to be sure, for a great many tasks, but nei-
is that human memories, once constructed, cannot ever really ther editable nor inventive. One also requires a random access
be obliterated. They are best forgotten by being repositioned memory; this task is particularly addressed by the structures
and relocated in other networks of associations. 6 and itineraries of ars memorativa.
In the passage above, Augustine is certainly speaking of In his lectures on Aristotle’s discussion of memory and
a consciously trained memory, one whose denizens, like prey recollection, dated 1254–57, Albertus Magnus distinguishes
(for he often speaks of memories as being like animals to be recollective investigation from rote solely in terms of mne-
tracked from their lairs, whose tracks or vestiges are to be fol- monic access.11 Recollection is a rational procedure, involv-
lowed through their familiar pathways in the forest), 7 can be ing associated conceptual sequences, with a variety of starting
rationally sought out via their particular routes when needed points from which one can find (invent) what one is attempting
for use, then returned to their proper places when finished with. to retrieve from memory. There are three basic categories of
But this edifice, this vast treasury, is chosen and constructed. It such starting points: from exact likeness; from opposition; or
is a work of art, using the materials of nature as all arts do, but from some similarity—in Aristotle’s term, “neighborliness”
consciously crafted for some human use and purpose. (súneggus).12 So in any recollection procedure, one can start
In his History, Memory, Forgetting, the late, great French off from at least three different access gates instead of just
philosopher Paul Ricoeur, himself a profound student of Augus- one. “This,” Albertus says, “is how recollecting differs from
tine, complained that arts of memory are “an outrageous denial rote learning [iterato addiscere], since recollection can be set
of forgetfulness and . . . of the weaknesses inherent in both in motion by any” of several means, whether by the logic of
the preservation of [memory] traces and their evocation.”8 In the subject matter or by our own habitual associations. Rote,
a similar vein, Harald Weinrich in Lethe, a book that sweeps though, is not an investigative procedure (and is therefore not
engagingly over the theme of forgetting in canonical Western properly recollection) because it has no such navigational sys-
literature, states that ars memorativa represents itself as “an tems. Someone who knows only by rote, Albertus points out,
art that can serve to overcome forgetting.” He archly observes has no resources if his memory fails. He simply “will not recall
that in its celebrated advice on making multicompartmental or recollect” what he wants.13
structures for a rich trove of remembered matters, “only for- Thus, the distinction between rote recitation and recollec-
getting has no place.”9 Yet, as Augustine makes abundantly tive reconstruction hinges on the issue of mnemonic access.
clear, Weinrich is wrong in saying that. Not only does forget- Discussing rote memories, Hugh of St. Victor says that people
ting have its honored place in an examination of memoria— who have memorized by rote have only one point of access
after all, Augustine devotes a whole section of his discussion to to what they know. Essentially, they are like people trying to
the paradox that he can remember that he has forgotten some- find something in a book without any supporting structures:
thing—for forgetting, of a sort, is essential to constructing an no chapters, no page numbers, no headings, no punctuation,
art of memory in the first place. and certainly no index. One is doomed to start at the beginning

2
and search through the book linearly and sequentially until one put, and still alive (so long as he was fed), but no one had the
finds what one wants.14 And if, on another occasion, one wanted way to find him again. This conception of forgetting is found
to find it again, one would have to go through the same labori- anciently, notably in Plato, but its even greater significance in
ous process—unless, of course, one had marked it in some way the Middle Ages comes from the fact that it is fundamentally
for quicker retrieval. What the structuring procedures learned biblical, wherein forgetting and remembering both are con-
in ars memorativa support is the mind’s own ability, when aug- ceived as resulting from paying attention, that is, from whether
mented in such a way, to retrieve quickly and securely by look- one is in God’s sight (in conspectu Dei) or not.19
ing for something using a good mental map. Of the Homeric In rhetoric and dialectic alike, ars memorativa was pre-
rhapsodists of the ancient world—those men who were said to sented as an art essential to the making of new composition.
be able to recite all of Homer with great precision from start As invention, it was considered equally a part of dialectic and
to finish—Xenophon commented, invoking a common ancient of rhetoric.20 In his logical work on the topics of argument,
example of this principle, “your rhapsodists are consummate as Aristotle associated it with finding rational arguments, as did
reciters, but they are very silly fellows themselves.” 15 It is not Boethius.21 In ancient rhetoric, it is associated with both inven-
the accuracy of their recitation that is criticized but their general tion and disposition, as a means for an orator to keep his main
idiocy, their foolishness in being unable to put their knowl- subject matters in mind in their best persuasive order while
edge to productive use. They recite only Homer and only by speaking extempore in forum or law court. It came as well to
rote. By contrast, the structures of mnemonic art were intended be a craft most useful for preaching. It might be helpful now to
to provide multiple access routes to the contents of memory, set the arts of memory within a historical frame and to review
supporting the mind’s unaided computational abilities with a how practical memoria adapted to changed circumstances dur-
random-access scheme. ing the thousand-plus years it was considered essential for any
Recollective accuracy is a separate issue. Neither the educated and devout person to acquire, at least in some degree.
method of rote memorization nor the method of making men- The surviving pre-Ciceronian accounts of mnemonic prac-
tal retrieval structures will in itself much affect the accuracy of tices tell some significant stories about their pedagogical func-
one’s memory. A person can inaccurately memorize items in a tion, but none is connected specifically to rhetoric. There are
series by rote as readily as she can accurately recall them within accounts from antiquity of children playing memory games,
a retrieval scheme structured by markers and cues. Faithful and many more of prodigious memory feats involving reciting
accuracy depends more on a well-developed talent for attentive the texts of the poets (the Aeneid backwards, for example). But
observation, continuing practice, craft mastery, and, especially, none of these involve the memoria of rhetoric. The essential
concentration and repetition during the initial stages of memory exercise in ancient and medieval education was to memorize
storing (neuroscience recognizes the value of repetition, nor- quantities of the Poets (or of Scripture, or of both), the founda-
matively three times at ten-minute intervals; “sleeping on it” tional literary texts upon which all further education depended.
measurably aids the secure storage of memories).16 Begun as a child learned to read, exercise in such verbatim
The ancient art of memory was conceived to serve recol- memorization was an aspect of the pedagogy of grammar. It
lection, not retention per se. This is evident from where it is is called recitatio in antiquity and indeed still is (“recitation”),
discussed both in manuals of rhetoric and in the scholastic com- the word-for-word memorization and recollection of text.22
mentaries on Aristotle’s brief treatise On Memory and Reminis- Such elementary memory work, however, should not simply
cence, appended to his work On the Soul. Commenting on this be conflated with the image making counseled in connection
work, both Albertus Magnus and Thomas Aquinas discuss ars with rhetorical memoria verborum such as that described in the
memorativa as a tool of recollection, for memory access and Rhetorica ad Herennium. That is presented specifically as an
retrieval of matters previously stored. In the words of Albertus, exercise for orators to strengthen their memories, and it is in
“recollection is nothing other than the investigation by memory that context that we must seek to understand it.
of what has been forgotten.”17 Recollection is a rational activ- The earliest account of memory techniques we now pos-
ity, it is investigation, and it seeks to retrieve “what has been sess is found in a fragmentary Stoic text called Dissoi logoi
forgotten.” An art of recollection is not an art to overcome for- (literally “Double Arguments,” a Greek version of the curricu-
getting or one that denies forgetfulness, as one might deny the lum exercise of arguing positive and negative sides of a ques-
existence of or obliterate an enemy. Recollection begins with tion, or argumentum in utramque partem), in which a student
what is forgotten and seeks to reconstitute the ways to recover is advised to make rebuslike associations to remember names
it. This description assumes that a memory, once laid down, is and unfamiliar or difficult words.23 Thus, to remember the name
always in the brain, and so can be uncovered by reconstruct- Chrysippus, one may link the syllables to an image—perhaps a
ing its tracks (vestigia). To make a coarse analogy, a forgotten golden horse (chrysos and hippos). The exercise embodied by
memory is like the man in the oubliette, that mythical narrow dissoi logoi was practiced and taught as a device of composi-
chamber to which a prisoner might be consigned and left “for- tion. It is significant, I think, that the other Greek text in which
gotten” because no one would or could investigate his where- an art of memory is mentioned is Aristotle’s Topikōn, where the
abouts.18 He was not annihilated, he was still where he had been method of analysis and composition based upon topoi of argu-

3
ment is likened specifically to the topoi of an art of memory, in privileged medieval position as the engine of invention. Monas-
that arguments are situated in orderly memory places (topoi), ticism developed the craft of meditation upon written, mem-
whence they can be called forth by recollection as needed. To orized texts (sacra pagina) as a self-conscious practical art
investigate exactly what Aristotle meant by such “topics,” and of invention for individual prayer but also, importantly, for
how what he may have meant differed from what Cicero meant homilies, colloquies, meditations, and prayers that were made
when he in turn wrote on the “topics,” and then Boethius in his publicly.31 Within this general craft, a specific art concerned
turn, is not my task here. But it is clear in both these Greek texts with penitential meditation developed, using the topics of sin
that the work of mnemonic techne was associated even this and virtue to arrange one’s thoughts and dispose one’s com-
early with tasks of invention, “finding” hard words or rational position. This meditation involved visualizing scenes, events,
arguments.24 It is also clearly associated with school subjects, and the buildings described in the Bible mentally. One was
especially (perhaps) the arts of the trivium, but by no means admonished to “paint” in one’s mind the pictures which the
confined to these, for memoria was regarded as a discipline texts raise up: enargeia (“bringing-before-the-eyes” or “vivid-
essential to them all.25 ness”) was an especially valued characteristic of style. I have
The continuing influence of the version of memoria arti- written extensively on this “monastic rhetoric” elsewhere, and
ficialis uniquely described in the Rhetorica ad Herennium has it is not my task to replicate my arguments and evidence here.32
been overstated by modern historians.26 It has even been plau- One conclusion seems to me inescapable: in the monks’ medi-
sibly doubted that Cicero ever actually practiced it. Quintilian tational craft—which they speak of as memoria or memoria
describes a simplified, practical mnemonic use of images, men- spiritualis (“pious memory” of divine texts)—the arts of mem-
tally placing a vivid image at strategic locations in a room to ory described by the auctor ad Herennium and by Quintilian
recall the matters of one’s speech (memoria rerum), for exam- played no discernible direct role at all.
ple, an anchor if one’s topics include ships, and so on. He also The late classical source that is truer to the spirit of
suggests annotating specific hard words and important names Ciceronian memoria than the late rhetoricians themselves is
with an image that would help one to remember them (the mat- Augustine, not the work on rhetoric falsely attributed to him,
ter of the memory advice given in Dissoi logoi).27 As Ruth but his work on teaching converts, De catechizandis rudibus
Taylor-Briggs has demonstrated in a set of important articles, (Instructing Beginners in the Faith). In this work, Augustine
immediately after it was composed the influence of the Rhe- uses two technical terms for recollection, ad uerbum and sum-
torica ad Herennium appears to have been slight in classical matim, but distinguishes them clearly from one other in respect
antiquity and through the earliest Middle Ages, until the ninth to their tasks, associating memorandi summatim (which is the
century; it acquired a full commentary only in the eleventh rhetoricians’ memoria rerum) specifically with the task of
century.28 Yet there is abundant evidence that trained memoria composition:
continued to be thought essential to compositional meditation
throughout this same period. What can be the source of such Even if we have memorized [many books of the Bible]
ongoing prestige? verbatim, [when we teach] we should neither just recite
It is not the rhetoricians of the late classical schools. . . . the entire books, nor, by retelling in our own words
Cicero’s De inventione, his most important rhetorical work for [through paraphrase], explicate every single matter con-
most of the Middle Ages, does not discuss rhetorical memoria. tained in these volumes . . . but having grasped them all
The rhetoricians of the late Roman Empire do sometimes dis- [in summary fashion] by their main topics . . . certain
cuss memoria as important in rhetoric, though what they say things can be joined together [by the speaker] which are
bears little resemblance to the exemplary techniques described more worthy to be examined closely, more pleasantly
in the Ad Herennium. The fourth-century rhetorician Julius listened to and arranged in their constituent parts, . . .
Victor, an important source for Alcuin and the Carolingian- dwelling on each for a time as though to loosen it up and
Ottonian courts, conflates memoria in rhetoric with elementary expand it, offering it for inspection and admiration by the
recitatio (recitation), advising that “memory should be trained minds of the audience.33
in learning by heart and word for word as many writings as
possible, both your own and those of others.”29 Although Julius Subject memory is the essential device of composition, the
Victor defines memoria as “the firm mental grasp of matters preacher choosing to dilate upon certain matters in the text
and words for the purpose of invention,” the benefit of such that he has selected for particular inspection by his audience,
training to an orator is that we “will always have with us some- while having the entire text at his disposal in order, learned both
thing to imitate,” on which to model our style and establish verbatim—which is what Augustine means here by ad uerbum
our compositional abilities. He dismisses “many people [who] edidicimus—and by its subject matters or summatim, in topical
offer precepts on places and images, which seem to me not to fashion.34 Verbatim recitation, then as now, meant memorizing
be effective.”30 and reciting texts by rote (ad verbum ediscere, memoriter red-
It is in early monasticism, not in the formal teaching of dere)—this is clear from what Augustine says in the passage
rhetoric, that the recollecting activity of memory assumed its quoted above.

4
 Monasticism was fundamentally text-based. Augustine that the six-winged creatures accompanying the Divine Throne
called memory “a kind of stomach of the mind,”35 and in monas- are seraphs, yet the figure in the medieval manuscripts is regu-
tic writing, chewing and digestion are favored tropes for read- larly labeled a Cherub. For the sake of consistency in this essay,
ing. Yet the conventions of monastic meditation also included I will call the figure a Cherub, even though it has six wings and
vivid imagining, of buildings and other artifacts described in so is, exegetically, a seraph. There is also a Latin text called De
the Bible, for the purpose of composing additional meditation. sex alis [cherubim] with which it seems to be closely associ-
These were sometimes called picturae, as in the elaborately ated, though the nature of that link is more problematic than
imagined verbal pictura (as he calls it) of the tabernacle in once thought, as I will soon discuss.40 The text dates from about
Book 2 of the treatise On the Tripartite Tabernacle by Adam the time of the first extant versions of the drawing, but the two
of Dryburgh, composed sometime about 1180.36 Adam intro- are frequently not found together, even in the earliest manu-
duces his project (which was commissioned by his abbot, John scripts. De sex alis has been ascribed wrongly to Alan of Lille
of Kelso) as a task of painting a picture of the Exodus Tab- (as it is in Migne, PL), but the earliest attributions (1190–1200)
ernacle in order to organize—to dispose—the themes of his are to Clement of Llanthony, a priory near Gloucester, and its
meditation on the body of Christ, the Church, which forms earliest provenance would seem to be English.
the second major division of his composition (i.e., the second The examples of the Cherub drawing still extant are in
“Tabernacle”): manuscripts made after about 1190, though the device itself is
based on earlier practices which we know about only indirectly,
according to your command, holy father, that I should de- through intriguing hints in much earlier texts from the desert
pict the oft-spoken tabernacle in a schematic plan as well, fathers and through the genre of pedagogical allegory popular
so far as I might understand and be able: now I put my in Late Antiquity—for instance, the complex allegorical scene
hand to its visualization, insofar as may be seen in some that opens Martianus Capella’s fifth-century treatise on the ele-
fashion with a bodily sight; both what can be understood ments of an education called The Marriage of Mercury and Phi-
through faith concerning the general Church of the elect, lology.41 There are other, later medieval composition devices
and through meditation be devised about particular ones of this sort, including the various diagrams of the Speculum
of the elect.37 theologie, some of which, like the Cherub, also are commonly
found separately.42
It is impossible to determine from the text alone whether Six different versions of the Cherub image are illustrated
Adam was referring to an actual drawing or to a mental pic- in Figures 1–6. They are quite different from one another,
ture: the words he uses are indeterminate, and if there ever though evidently the same basic design. The texts written out
was a picture accompanying this treatise, it has not survived in on their wings are recognizably the same as well, though often
any manuscript. In meditational writing, the process of imag- differently arranged and articulated, one to another. Their dif-
ining these structures is commonly referred to with the verbs ferences make clear how this type of picture served a mainly
pingo and depingo, as well as fingo and its derivatives. There investigative function rather than a representational or exegeti-
are many examples from the fifth century through the twelfth, cal one. These drawings are highly variable: that is the crux
and even later. Richart de Fournival (act. 1240–60), a canon of for comprehending their genre. Proper investigation proceeds
Amiens, said that all vivid text had peinture (“picture”) as well rationally as one item cues the next in orderly sequence, but
as parole (“speech”), these two being the roads to the portals of those sequences are flexible and plastic and can vary from one
memory.38 The exact source of this recurrent medieval emphasis user to the next as their elements are recombined. That is what
on mental picture-making while reading and meditating is not makes for creative analysis. So also with these particular six
at all clear. Something is surely owed to the cultivation in Late images of the Cherub.
Antique rhetoric of ekphrasis and of stylistic enargeia more Each depicted creature has six wings, two covering its
generally as being cognitively and persuasively valuable, an body, and two interlaced rising above its head (to hide God’s
emphasis that is strong as well in the rhetorical teaching of face), and two outstretched for flying. Most often, naked legs
Aristotle, Quintilian, and Cicero. But the importance of imag- and/or feet of the cherubim are prominently drawn, a detail
ining buildings and their furniture as a meditation device— that does not accord with the biblical text in Isaiah 6:2, which
specifically those various avatars of the Temple that one finds describes the feet of the seraphim as covered by two of their
in the Bible (Noah’s Ark, the Exodus Tabernacle, Solomon’s wings. Hands are also displayed in many but not all of the ver-
Temple, the visionary Temple Mount in Ezekiel, St. John’s sions, a detail taken from the description of the cherubim in
Heavenly City)—is directly owed to traditions of meditation Ezekiel 10:21—but the cherubim of Ezekiel have four wings
and contemplative composition that were adapted into earliest (those guarding the ark are described as having just two). All
Christianity from classical Judaism.39 are pictured with human faces, a detail not given in Isaiah, and
I will discuss just one of these devices in detail, the picture in Ezekiel, the cherubim are described as having four faces,
called the Cherub or the Seraph. The alternatives reflect a nam- those of a man, a bull, an eagle, and a lion. Medieval exege-
ing problem, arising because the text of Isaiah 6:2 stipulates sis consistently makes clear that the seraphs of Isaiah and the

5
cherubs of Ezekiel were distinct (if complementary) orders of a general subject into its constituent topics. In scholastic terms,
angels, so to name a six-winged angel a Cherub is a deliberate the Cherub presents a formal divisio of penitence, each state-
anomaly. ment of the main subject’s topics being arranged in an abbre-
The Cherub figure is not the exegetical illustration of a viated yet clear form, which allows each one to be readily
particular text but follows independent iconographic traditions. expanded by a speaker as suits a particular occasion. The fig-
Indeed, independence from any particular text is a fundamental ure has the basic qualities of any mnemotechnical structure: its
quality of the medieval Cherub, evident well before it came places (topics) are rarus, clarus, and solempnis.45 The Cherub
to be associated with the other Speculum theologie diagrams. demonstrates the logical relations of the subject and its several
The fact that the two earliest versions extant (Figs. 1 and 2) are subtopics (and sub-subtopics) but in a picture rather than solely
linked with two completely different texts (Historia scholastica in words. So, while the figure is, in medieval terms, mnemonic
and De sex alis), though both were drawn in England only a in its purpose, its intention is not simply to recall some specific
couple of decades apart, suggests that Clement of Llanthony’s words but to invent from them, to elaborate and expand the top-
text derives from a picture already in circulation and not the ics in the analytical manner of topical argumentation taught in
other way around. The early occurrence of Clement of Llantho- both dialectic and rhetoric.
ny’s text without the drawing, as a meditation aid introducing The Cherub figure is intended to help its users think and
a Psalter, points to the same conclusion. There are three mani- compose. To use the device, a person would need to internal-
festations of the Cherub in English manuscripts from about ize the picture, remembering the divisions of the subject in
1190–1200, but one has Clement’s text without the drawing an orderly way, as headings of “wings” and “feathers.” With
(Oxford, Bodleian Library, MS Auct. D. 2. 1), one has the draw- the figure in mind (literally), one would readily have the basic
ing but with Peter Comestor’s text (Cambridge, Corpus Christi topics for as many as thirty-six sermon-meditations, nearly
College, MS 29), and one has both the drawing and Clement’s a whole season’s worth, on the general subject of penitence.
text copied together (Cambridge, Corpus Christi College, MS Each preacher would be able to adapt the scheme to the specific
66). What links Peter Comestor’s and Clement’s texts is their occasions of his own speaking, and indeed they did. Bolzoni
purpose: both are important sources for meditation—studious has demonstrated that Bernardino of Siena, the great Italian
reading and composition—meditatio being the initial stage in preacher of the fifteenth century, did just this, mentally adding
planning all sorts of composition, including prayer, colloquy, two more feathers to each of the Cherub’s six wings to make
and homily.43 the total number of topics better suit his Lenten cycle (that is,
From a purely mnemotechnical standpoint, the texts on forty-two rather than thirty-six), and since he began in mid-
the Cherub’s wings could have been written onto any clearly week, on Ash Wednesday, he started with the fourth feather of
organized, rational plan, such as a ladder (one thinks of John the first wing.46
Climacus and Benedict’s Rule), a tree (the lignum vitae or tree Examination of the differences among the versions of
of life device used by Bonaventura and many much earlier the Cherub shown in Figures 1–6 indicates how particularly
writers), or the Tabernacle (used by Adam of Dryburgh and each has been adapted. The earliest surviving version (Fig. 1)
various Victorines). But the winged figure has evident appro- accords fully with the text of De sex alis. It is in an English
priateness as a device for composing meditations on penance manuscript, formerly in the library of the Cistercian abbey of
and for the preparation of sermons and other pastoral materi- Sawley in Yorkshire, though likely made in the scriptorium
als. The medieval Cherub figure conflates commentary on the at Durham about 1190.47 Each wing is clearly labeled on the
four-winged animalia of Ezekiel 1 (called cherubim in Ezekiel coverts with its main topic: confessio, satisfaccio, munditia (or
10:5–22) with the seraphim of Isaiah; with descriptions of the puritas) carnis, munditia (or puritas) mentis, dilectio proximi,
two-winged angels, also called cherubim, that guard the Ark in dilectio Dei (confession, satisfaction, purity of body, purity of
Exodus (Exodus 25) and in the Temple of Solomon (I Kings 6); mind, love of neighbors, love of God). The number of each
and with the animalia that attend the Divine Throne in Revela- wing in the treatise accords with those in this version of the
tion 4, which are adopted from Ezekiel. The cherubim-animalia drawing. The subtopics depend from these wings as feathers
conflation was already made in Ezekiel; Revelation 4 conflates and are so described in Clement’s treatise. In this manuscript,
these further with Isaiah, as the animalia-cherubim sing a dox- the diagram occurs in the middle of a text on confession, “De
ology, as do the seraphim in Isaiah 6:2.44 Jerome glosses sera- confessionem in ultimo,” a letter from Theobald d’Étampes
phim as meaning “ardor, burning” and cherubim as “wisdom.” to Bishop Robert Bloet of Lincoln.48 The letter takes up pages
A seraph touches a burning coal to the prophet’s lips to cleanse 99–101: the Cherub is drawn in the middle of this text, on page
them (Isaiah 6:6) and purify him before he can speak God’s 100. Right after Theobald’s letter (on 102) begins the text of
words to Israel. Thus, to call a six-winged figure a Cherub is no De sex alis, entitled “Incipit descriptio prime ale cherubin.”
mistake but declares plainly the intimate links among peniten- It ends (108), “Explicit descriptio Magistri Clementis prioris
tial purification, ardor, wisdom, and pastoral speaking. canonicorum regularum sancte marie ciuitatis claudiocestrie
The Cherub’s wings lay out the ways of penance by [= Gloucester] de sex alis et xxxta pennis cherubin.” Note that
enabling a particular method of analytical thought that divides the text is identified as descriptio, that is, an ekphrasis, called

6
FIGURE 1. Cambridge, Corpus Christi College, MS 66, p. 100, English, ca. 1190, from Sawley Abbey (Cistercian), but probably made in Durham (photo: re-
produced with kind permission of the Master and Fellows of Corpus Christi College, Cambridge).

7
FIGURE 2. Cambridge, Corpus Christi College, MS 29, fol. 8v, English, early 13th cent. (photo: reproduced with kind permission of the Master and Fellows of
Corpus Christi College, Cambridge).

8
in Latin descriptio or pictura, a verbal figure that was regarded The version of the Cherub shown in Figure 2 was made for
as particularly productive of enargeia. It was understood that an English manuscript of Peter Comestor’s Historia scholas-
the details of an ekphrasis could be realized quite differently in tica, dating to the early thirteenth century.51 The drawing is the
each person’s imagination, for its purposes were affective and last in a series of mostly genealogical diagrams that introduce
cognitive, not scientific description. The line between descrip- the Historia: it is on the verso of fol. 8 and the Historia begins
tio (or pictura) as ekphrasis and descriptio (or pictura) as an on fol. 9. The text of De sex alis is not in this manuscript at
actual painting is thin and highly variable in medieval usage. all. Yet the Cherub’s wings are numbered and the texts writ-
In the rubrics in MS CCC 66, the word descriptio refers to both ten as they are described in De sex alis. A user of this book,
the text and the picture.49 then, did not have the explanatio; evidently the Cherub figure
For convenience sake, in the rest of this discussion, I will was familiar enough to be useful without it. Peter Comestor’s
use the numbering of the wings in De sex alis, even though, work was mined for all kinds of compositions; it would seem
as will soon be apparent, the wings in other versions do not that the general virtue of the Cherub for study and recollection
often follow the order. As I argued earlier, all the evidence (of thoughts and matters) was not tied even at this early date
we still have indicates that a mental composition device based to De sex alis.
on a Cherub figure preceded the composition of De sex alis. The pairs of interlaced wings in Figure 2 (one and two and
There are more manuscripts containing only the text of De sex five and six) are not so crowded as in Figure 1, but the writing
alis than ones that have the diagram, too. A copy of the text is more abbreviated. That on wings one and three must be read
made in the early thirteenth century and likely housed in the by moving those wings to the left, and those on wings two and
library of Llanthony itself does not have the drawing, though four by moving them to the right. The upside-down titulus,
it begins “Incipit explanatio Clementis Lantoniensis super alas confessio, on wing one reveals that the scribe had turned the
cherubin & seraphin.” Presumably the scribe thought the pic- page horizontally to write out the texts and neglected to return
ture was familiar enough that it was not necessary to draw it it to the vertical when he copied this titulus; he, of course, was
for its explanatio.50 bound by his vellum in a way that a user’s imagination is not.
In Figure 1, on wing one (confessio) the subtopics are What is significant is that by these motions the scribe assumes
veritas, integritas, firmitas, humilitas, simplicitas (truth, integ- that his readers also will be able to move the image. It has cer-
rity, steadfastness, humility, simplicity). However, as one looks tain specific actions implied within it, which—absent holding
more closely at the device, one is made aware that as a static the physical book at all times—a viewer must perform mentally
object it is rather confusing. To clarify its relationships, the in order to use it well.
wings need to be imagined as moving. For instance, wings one Just how elaborate these implied actions could become is
and two are interlaced in such a way that the words are crowded exemplified by the Cherub shown in Figure 3, which is from an
into one another. (This confusion is compounded because the Italian manuscript of the early fifteenth century and probably
blue ink in which the words on alternate feathers were written Dominican in provenance.52 In this version of the device, the six
has now badly faded.) To read them easily, one needs to imag- wings are attached around the face of the Cherub (red because
ine the wings unfolding so that their feathers are separated and it is inflamed by divine love), pinwheel-like, but with legs and
their topics can be read. But when unfolded, they will obscure feet. Notice the movement implied in this arrangement: to read
wings three and four. And so to read three and four one would the subjects of each wing, the figure must rotate. When this is
need mentally to refold one and two downward and reinterlace done, each wing in turn can be read right-side up. This suggests
them. In the case of wings five and six, at the very top, the top- to me that the users of this device were more concerned about
ics of wing five (dilectio proximi) are written upside down. To the separate topics than about overall coherence, just as one
read them, one needs in imagination to rotate the wing as well would suppose to be the case for someone using the device to
as unfolding it from wing six, and a user of the manuscript page compose many different sermons over a period of time. This
would have to turn the manuscript physically, as the scribe must Cherub stands on a wheel (Ezekiel 10:9), labeled “the works
also have done. Given the mental gymnastics implied in this of mercy” (opera misericordie), on which are written seven
picture, one might conclude that many users, imagining com- additional topics. (The Cherub in Fig. 1 is standing above the
positions from the diagram, would be inclined to think of each golden cloud of the heavens, while that in Fig. 2 stands on
wing as a separate topic of the general subject. In other words, nothing.) The text box on the right explains that the wheel has
the figure invites mental mobility and recombination. It should seven spokes, which are the seven acts of mercy that God will
be noted that the treatise De sex alis does exactly this: it does take into account at the day of judgment; the box on the left
not attempt to make a overarching argument about penance, explains the picture: “this Cherub is depicted in human like-
nor does it contain transitions taking a reader coherently from ness; it has six wings which represent the six ethical actions”
one topic to the next but treats each wing and feather discretely by which a faithful soul may be redeemed.53 In this manuscript,
and rather briefly (with the exception of wing one, confessio). the Cherub is not associated specifically with penance but more
It is a practical manual with some examples, not a developed generally with good conduct and virtuous life, topics suitable
philosophical treatise. for the eclectic audiences of the preaching friars.

9
 Something even more important is shown by the tituli of Cherub found in the Howard Psalter (Fig. 4).56 The wings in
this picture’s feathers. In this version, the wings have no num- this version also have no numbers; each, however, is distinctly
bers. The two wings covering the body are labeled munditia labeled by a text band.57 The texts on the wings and feathers
carnis and munditia mentis, not confessio and satisfactio as in are nearly the same as those of the Cherub in Figure 3. Those
Figures 1 and 2, and in the De sex alis text. The wings of Con- in the text boxes to either side of the angel and on the wheel on
fession and Satisfaction are moved into the flying position, the which the Cherub stands are the same as the ones used in Figure
pair with which the seraph moves. The subtopics of the con- 3, but this Cherub has been imagined very differently from the
fessio wing are “effusion of tears” (lacrymarum effusio), “holy Laurenziana example. The figure is more vertical, the wings
meditation” (sancta premeditatio), “straightforward speech” that cover the body drawn so that one simply overlaps the other,
(simplex locutio), “truthful thought” (verecunda cognitio), and rather than interlacing the feathers (as in Fig. 1)—a design that
“prompt obedience” (obediencie promptitudo). These subtop- makes the words written on them much easier to read. The two
ics are different from those of the confessio in Figures 1 and 2, upper wings also have been simplified in this fashion. But the
culled in part from topics that in those versions appear under texts of the upper right wing are upside down (as in Fig. 3);
Satisfaction and Purity of Mind. Here is another example of a reader would have to manipulate the book or conceptually
this same phenomenon. The legends on the wing called Purity detach this wing from its context in order to read it. The design
of Body (munditia carnis) accord in the earliest versions with of this Cherub, like that in Figure 3, encourages the separate
the divisions recorded in De sex alis and relate to the senses: development of the topics—one does not have to proceed in a
“decorous looking” (visus pudicitia), “chaste hearing” (auditus predetermined order. The Howard Cherub is found with several
castimonia), “modest scent” (olfactus modestia),54 “temperate other of the diagrams in the so-called Speculum theologie, at the
eating” (gustus temperantia), “holy touching” (tactus sancti- beginning of a richly decorated Psalter, made for an aristocratic
monia), but in Figure 3, these categories are placed under the household. The Speculum diagrams occur in varying combina-
Satisfaction wing. The topics of Purity of Body (mundicia car- tions in manuscripts containing study and homiletic materials,
nis) include generous almsgiving, devout prayers, participation often made for preaching and teaching clerics, though notably
in vigils (vigilarum actio), devotional exercises (disciplinarum also for the libraries of lay aristocrats. Such households had
usus), and observing fast days (ieiunium). Two of these tituli many chaplains and other clerics who served the family. One
(prayers and almsgiving) are under Satisfactio in De sex alis, should not assume, I think, that a book like this would be used
but the other three—vigils, spiritual exercises, and fasts—are only by the lay lord who commissioned it, nor that a diagram
not mentioned there at all. Yet these are all topics that might like the Cherub, as I have described it, would be too difficult
be appropriate to the lay and secular audience that friars were for any in such a household to comprehend and use.58
used to addressing, more than they would be to contemplative The diagram for a Cherub in an English manuscript made
monks, who would observe such things as part of their rule. after 1235 for a Dominican friar (Fig. 5) not only does not num-
Major changes like these were made not from ignorance ber the wings but moves each one’s main titulus to an adjacent
or carelessness by a scribe copying some prototype. Rather, circle, evidently to simplify and make less crowded the writing
the divisiones of penitence have been differently analyzed, evi- on each feather.59 This figure is dressed in chain mail visible in
dently to suit the preferences of those using this manuscript the area of its upper legs and appears thus as a militant angel
for the purposes of preaching and counseling. The subtopics of the Apocalypse (Michael?), standing on the seven-headed
appearing under Confession in this manuscript are just as ratio- beast. The conflation of the Cherub device with Apocalypse
nal as the five in the Sawley version (which were veritas, integ- subjects is, as I noted previously, another variation within the
ritas, firmitas, humilitas, simplicitas), but they are not those trope. The matters in this book were all written by three scribes
found in the text of De sex alis. This observation indicates that during the mid-thirteenth century: Michael Evans described
the Cherub figure is no longer thought of (if it ever was) as the it rightly as “an anthology of works that would have served
illustration to a treatise on which it is dependent but instead as a vade mecum in pulpit, confessional and private study”60
as a fully independent investigative and composition device. (though it should be noted that, at 278 by 170 mm, Harley 3244
It also suggests that each wing, or major topic, is conceived of is not pocket-sized). A vade mecum is a familiar study book,
as independent of the others, not as part of a coherently related intended for use by individuals within a “family,” whether a
single analysis of penitence but rather as separate topics or dis- secular household or a religious community.
tinctiones—like a collection of essays on penitence rather than In this manuscript, the Cherub figure is closely associated
a monograph, and thus eminently suitable for development in with two popular works on penance and examination of con-
a series of sermons prepared over a period of time. Conceptu- science, Alan of Lille’s Liber penitentialis61 and the Liber de
ally, one could detach each wing from the figure and examine vitiis (Book of Vices) by a French Dominican, William Peraldus
it as one wished, turning it about, treating its subject matters (Peyraut). The manuscript contains various homiletic materials
at length as the various occasions for sermons might require. 55 besides these, including Honorius’ Elucidarium, an illustrated
This detachable aspect of the wings is even more pro- bestiary, an ars predicandi by Raymond of Thetford, and Gros-
nounced in the fourteenth-century English version of the seteste’s Templum Domini. Evans has described the interesting

10
FIGURE 3. Florence, Biblioteca Medicea Laurenziana, MS Plut. 30. 24, fol. 3v, Italian, ca. 1410, probably Dominican (photo: reproduced with permission of
the Ministero per i Bene e le Attività Culturali).

11
FIGURE 4. London, British Library, MS Arundel 83-I, The Howard Psalter, fol. 5v, East Anglian, ca. 1310–20, made for the family of John Fitton(?) (photo:
© The British Library Board).

12
structural features of the manuscript. The Liber penitentialis the (early?) fifteenth century, a half leaf was sewn in between
is written in fols. 19–26v, and, with the Elucidarium, takes up fols. 7 and 8, onto which are copied abbreviated materials on
the first two quires as the book is presently put together (it is confession and on the seven deadly sins and their derivatives.
not now in its thirteenth-century order, as the original quire The scribe of the Cherub is not the same one as that for the other
numbers, still mostly visible, indicate that, before the early diagrams; its exact date is disputed,65 but there is no dispute
fifteenth century, these two quires were moved from the end that it was a different scribe who made it—likely “Hermanus
of the book to its beginning). On fol. 27 is a full-page picture custos,” the abbot identified below the angel figure. The text of
of a Dominican kneeling before Christ seated on a throne, and De sex alis begins on this page and continues on the verso side,
carrying a banderole, which states, “Son, entering the service in a highly abbreviated and variant version (compared with that,
of God, prepare against temptation.”62 Fols. 27v–28 contain for example, in MS CCC 66).
the picture of the Knight against the Vices, which occupies In this book, the Cherub addresses the compositional
the full opening: it is titled “the life of man upon the earth is a needs of contemplative monks, not preaching friars (Kempen
struggle,”63 a theme that accords well with the Peraldus text but was the home of Thomas à Kempis). But it is not odd to find the
also looks to the subject of penance and temptation emphasized diagram in a Cistercian book—Sawley Abbey, which housed
in the Liber penitentialis and the picture of the kneeling friar. the Cherub shown in Figure 1, was a Cistercian foundation as
The paintings are on a separate bifolium, which originally was well. As presented on the page, the Kempen version is more
at the beginning of the book, with the painting of the kneel- complex and more inclusive of chains of other texts and sub-
ing friar its frontispiece. The text of the Liber de vitiis begins jects than some of the other versions, and it includes psalm
on fol. 29; opposite it, on fol. 28v, immediately following the verses especially associated with contemplative life. Herman
Knight against the Vices opening, is painted the Cherub (Fig. the abbot is painted under the angel’s feet—the book he holds
5). Liber de vitiis ends in column b of fol. 33v; immediately has written on it the text of Psalm 16 (17):8, one especially
following it in column b is the text of De sex alis, beginning related to contemplative life and to the Cherub.66 Notice how
“De prima ala que dicitur Confessio & eiusdem ia penna. que the angel appears to spring up out of his head, as indeed was
dicitur Veritas.” This text is almost complete, following the the intended case. Yet the most startling feature of this draw-
numbered textual order through wing six, feather four, and ends ing is how the feathers of wing two have been rotated outward
on fol. 35v with some penwork. Thus the whole sequence in from behind wing one, so that they can be easily perceived and
the present book, from fol. 19 through fol. 35v including the read in their proper order, depicting graphically one important
pictures in fols. 27 through 28v, makes a thematic unit. Most mental task the picture required. Whenever my students look
important for any preacher, it provides rich materials for study at this drawing for the first time, at least one will judge that the
and composition on many pastoral subjects. Whoever moved painter was simply incompetent to render this wing normally—
the gatherings containing the Liber penitentialis to the front whatever a seraph’s wing normally looks like.67 Competence
of the book understood and emphasized the thematic progres- in drawing is not the issue here, however, for Hermanus has
sion, which the first maker had clearly intended, and adapted it deliberately exploded this wing outward and backward at the
further to his own uses. wrist to make it clear and easy to read and thus to recollect its
The great variation among the drawings found in these topics or loci in the investigative, rational manner described
examples indicates how widely the figure was adapted for prac- by Albertus zMagnus. It is a picture that was used in difficult
tical use. It was treated, not like a picture illustrating or com- mental work, as were the other diagrams in this booklet, some
menting on a particular text passage, but like a tool. Scholars of which have additional texts and citations written onto them
found the picture useful without a written explanation of it and by later users.
also determined that the verbal ekphrasis was useful without In all these versions of the Cherub figure, the matters of
its being materially realized; both situations are common in the penitence are recorded in summary form, as is appropriate to
manuscripts. A final example will reinforce this point. Figure an instrument for invention, following the advice of Augustine
6 displays yet another conformation of the seraph. The manu- to focus on abbreviated highlights rather than trying to speak
script in which this drawing is found was in the library of the about everything equally. The tituli are given summatim, a
Cistercian monastery at Kempen in the Rhineland.64 It is only a set of shorthand notes for later development in a full-fledged
booklet, of eight bifolia, evidently left unbound for quite some composition. Scholars are inclined to think a device like this
time, for its outer pages are badly rubbed and faded. It consists was used to teach: well, yes, but only as it was useful to the
of the diagrams of the Speculum theologie, but there is a curios- chaplain/confessor/preacher in inventing his own composi-
ity about it. The first seven folios contain the diagrams (not all, tion—these are not aphorisms dumbed down to satisfy minds
to be sure, for the selection is particular to each manuscript, as too novice to learn anything better. In other words, the pithy
Lucy Sandler has documented), one to a page, all written in the subtopic written on each feather is not “content” as modern
same hand of the fourteenth century. The Cherub was drawn audiences understand it, but a clue that leads to much greater
later by someone else and painted on the recto side of the last and fuller content stored in other places in the preacher’s mind
leaf (8), which had previously been blank. At some point in (and even the layperson’s mind). Abbreviation is necessary to

13
 FIGURE 5. London, British Library, MS Harley 3244, fol. 28v, made in England after 1235, for a Dominican friar (photo: © The British Library Board).

14
FIGURE 6. New Haven, Beinecke Library, MS 416, fol. 8, Rhineland, Cistercian monastery of Kempen, ca. 1425 (photo: Yale University, Beinecke Library).

15
any expansion; it is necessary to contract before one can dilate erful finding tool. It is precisely because one does not have to
creatively, as Augustine makes clear, following ancient advice start at the beginning and go through it in one way, over and
that became standard in the Middle Ages. The fundamental over again, that it proved so useful and so popular. Redactions
pulse beat from abbreviation to amplification, repeated through of this figure can be found in preaching and devotional materi-
endless variations, drives virtually all medieval compositions. als well into the seventeenth century.
And to abbreviate one must condense and leave out—one must Through these examples, I hope I have been able to dem-
forget many things in order to recollect more, distill more onstrate that the craft of memory, like its successor investi-
fruits of study more nourishingly, more originally. Forgetting gative art, systematic logic, is not fundamentally an overly
is necessary to the rhythm of remembering. complicated procedure for preparing to pass examinations or
Albertus Magnus described recollection as a rational for memorizing random facts or for reciting hundreds of verses
search, following the steps of an orderly series to matters that by heart. It is in fact ill-suited to rote repetition, as people who
are forgotten (in the sense of that prisoner in the oubliette). have written about it have been pointing out for more than two
The wings and feathers of the Cherub map out precisely such thousand years. Perhaps it is time we paid attention to what they
a search. But this is not the only, or even the most power- said. For them, it was a craft for the creation of new knowledge,
ful, cognitive virtue of the figure. It presents multiple points thinking new thoughts, and for investigating difficult subjects in
of access—thirty-six at least, grouped for easy recollection in the forums of debate and commentary, preaching, counseling,
six groups of six topics. Any one of these can be found and and contemplative prayer.
accessed independently and immediately. The device is a pow-

NOTES
* I have been greatly helped by the careful editorial comments of Clark 4. For example, Lambert Schenkel, Gazophylacium artis memoriae (1595;
Maines and Anne D. Hedeman during the preparation of this essay, which Strasbourg, 1610), counsels imagining that a great wind has blown
would be much poorer without their advice. It was first presented at the through the chambers of your memory places and carried all their images
University of Illinois Urbana-Champaign Center for Medieval Studies in away, or that a servant has swept all the rooms entirely clean. Similar ad-
September 2005. I am deeply grateful to my colleagues at Illinois for ar- vice occurs in a late-sixteenth-century memory treatise by the Dominican
ranging this meeting, a singular honor, and for their generous hospitality friar Cosmo Rosselli, Thesaurus artificiosae memoriae. These and other
during my two visits to the university as a Mellon Distinguished Scholar sixteenth-century examples are discussed by L. Bolzoni, The Gallery of
in 2005, in particular to Anne D. Hedeman, Stephen Jaeger, Martin Ca- Memory (Toronto, 2001), 139–45. It should be noted in all this advice
margo, Danuta Shanzer, Karen Fresco, and Herbert Kellman. I also want that, although the particular images are cleared away, the basic structure
to thank my excellent colleague, Lucy Freeman Sandler, for teaching me of the memory places remains secure and intact. Erasure of images from
so much over the years about the Cherub. the memory places is assumed in the ancient model of the memory places
1. The story is recounted by Cicero, De oratore 2.299–300, and cf. 2.351. as being like wax tablets and the images like the stylus-incised letters
Cf. also Quintilian, Institutio oratoria 11.2.50, who praises Themistocles’ erased from the tablet when they are no longer needed; see Rhetorica ad
unheard-of feat in acquiring fluent Persian within a year. Herennium, 3.18.31.
5. Bolzoni, The Gallery of Memory, 144.
2. Augustine, Confessiones 10.8.12–14: “Transibo ergo et istam naturae
meae, gradibus ascendens ad eum, qui fecit me, et uenio in campos et 6. The way in which the monumental map of Late Antique Antioch was
lata praetoria memoriae, ubi sunt thesauri innumerabilium imaginum de “relocated” by the Christians during the struggle between the bones of
cuiuscemodi rebus sensis inuectarum. Ibi reconditum est, quidquid etiam blessed Babylas and the spring of Apollo in Daphne is a good case in
cogitamus, uel augendo uel minuendo uel utcumque uariando ea quae point; see M. Carruthers, The Craft of Thought: Rhetoric, Meditation, and
sensus attigerit, et si quid aliud commendatum et repositum est, quod the Making of Images, 400–1200 (Cambridge, 1998), chap. 1, esp. 24–59.
nondum absorbuit et sepeliuit obliuio. . . . Haec omnia recipit recolenda, The seminal study by M. Halbwachs of “communal memories” based on
cum opus est, et retractanda grandis memoriae recessus et nescio qui se- the holy places in Jerusalem is also relevant: La topographie légendaire
creti atque ineffabiles sinus eius: quae omnia suis quaeque foribus intrant des Evangiles en Terre Sainte (Paris, 1941; rpt. Paris, 1971).
ad eam et reponuntur in ea. . . . Intus haec ago, in aula ingenti memoriae 7. On the trope of recollection as hunting for prey, see M. Carruthers, The
meae. Ibi enim mihi caelum et terra et mare praesto sunt cum omnibus, Book of Memory: A Study of Memory in Medieval Culture, 2nd ed. (Cam-
quae in eis sentire potui, praeter illa, quae oblitus sum.” Translated by bridge, 2008), 78, 323–24. All further references to The Book of Memory
M. Boulding (Hyde Park, NY, 1997), from the Latin text of L. Verhejien, are to the second edition.
CCSL 27, revised in 1990 (Turnhout, 1996), which is quoted above. 8. P. Ricoeur, History, Memory, Forgetting (Chicago, 2004), 67.
3. Aristotle, De memoria et reminiscentia 1.450a–b, discussing why people 9. H. Weinrich, Lethe: The Art and Critique of Forgetting (Ithaca, NY,
vary in their abilities to remember; see also his comments at 453a. I have 2004), 10.
used the Loeb Classical Library edition, translated by W. S. Hett from the 10. I have argued the case against such a conflation in Carruthers, The Book
1898 text of W. Biehl (Cambridge, 1936). of Memory, xii–xiv, 100–106.

16
11. Albertus Magnus, Commentary on Aristotle’s De memoria et reminis- story of Cornelius the pious centurion, whom an angel directs to St. Peter
centia; trans. J. Ziolkowski from Albertus Magnus, Opera omnia, ed. A. because his good deeds are seen and remembered by God (Acts 10:4; also
Borgnet (Paris, 1890), 9:97–118, in The Medieval Craft of Memory: An Acts 10:31; Douay-Reims translation): “orationes tuae et elemosynae
Anthology of Texts and Pictures, ed. M. Carruthers and J. Ziolkowski tuae ascenderunt in memoriam in conspectu Dei” (Thy prayers and thy
(Philadelphia, 2002), 118–52. alms are ascended for a memorial in the sight of God). A variation of the
12. Aristotle, De memoria et reminiscentia 451b. 20; see also R. Sorabji, trope involves God turning his face toward or away from something, to
Aristotle on Memory (Providence, RI, 1972), 42–46. remember or forget; cf. Psalm 33 (34):16–17, Psalm 9B (10):11, Psalm
108 (109):14–15.
13. Albertus Magnus, Commentary on Aristotle’s De memoria et reminis-
centia, tractatus 2, capitulum 3 (ed. Borgnet, 112); trans. Ziolkowski, 20. I discussed this curricular situation in M. Carruthers, “Rhetorical memo-
143: “Et ista est differentia in qua reminisci differt ab eo quod est iterato ria in Commentary and Practice,” in The Rhetoric of Cicero in its Medi-
addiscere, cum reminiscentia possit moveri quodam praedictorum mo- eval and Renaissance Commentary, ed. V. Cox and J. O. Ward (Leiden,
dorum in id principium quod est ante quaesitum jam in memoria, sive 2006), 205–33.
ex parte rei, sive ex parte consuetudinis. Iterato autem addiscens talibus 21. See the preface by Eleonore Stump to her edition of Boethius’s De topicis
non movetur. Cum vero non investigat et movetur per aliquod principium, differentiis (Ithaca, NY, 1978), 15–17, 165, referencing in particular what
tunc non recordabitur vel reminiscetur.” Boethius says at the start of Book 1. Aristotle’s advice is in Topikōn 8.14
14. Hugh of St. Victor, De tribus maximis circumstantiis historiae (with (163b), a text taught commonly in the medieval Organon just after the
tables and a diagram); trans. M. Carruthers in Carruthers and Ziolkowski, Sophistical Arguments (Sophistici Elenchi); see Stump, Dialectic and
The Medieval Craft of Memory, 32–40 (reprinted as Appendix A in Car- Its Place in the Development of Medieval Logic (Ithaca, NY, 1989), esp.
ruthers, The Book of Memory, 339–44), from the edition of W. M. Green, 11–30. Aristotle also describes the recombinative virtues of a topical
Speculum, 18 (1943), 484–93. memory scheme in De memoria et reminiscentia 2.452a; it should be
15. Xenophon, Memorabilia 4.2.10; trans. E. C. Marchant, Loeb Classical noted that both these discussions relate to discovering materials (argu-
Library (London, 1923). They are called akribountas, “precise in speak- ments, examples) for compositions.
ing.” The comment is made by the young man Euthydemos, in response 22. These various exercises are described in Carruthers, The Book of Mem-
to a query by Socrates, who asks if he wishes to become a rhapsodist ory, 20–23, and in eadem, “Rhetorical memoria.” Aspects of childhood
since he possesses a complete copy of Homer. This is one of several memory training in Hellenistic antiquity are described by R. Cribiore,
specialized professions that Socrates suggests to him as a goal for his Gymnastics of the Mind (Princeton, 2001), 164–67; and see also J. P.
education; no, he says. Socrates offers him instead an education that Small, Wax Tablets of the Mind: Cognitive Studies of Memory and Lit-
makes good governors and judges as well as a persuasive speakers. The eracy in Classical Antiquity (London, 1997), esp. 126–31.
distinction being made is between skill and wisdom (or, as we would now 23. Traditionally thought to have been composed in the fourth century bce,
say, between a technical and a liberal education). On this ancient debate the dating of the Dissoi logoi is now unsettled and may be considerably
in the fourth century bce, see D. S. Hutchinson, “Doctrine of the Mean later than previously thought; see M. Burnyeat, “Dissoi Logoi,” in The
and the Debate Concerning Skills in Fourth-Century Medicine, Rhetoric Routledge Encyclopedia of Philosophy (London, 1998), 3:106–7. The
and Ethics,” Apeiron, 21 (1988), 17–52. See also Plato, Gorgias 463B; treatise has been translated by R. K. Sprague in Mind, 77 (1968), 155–67.
and Aristotle, Metaphysics 1.1, on the distinction between “a knack” The exercise of argumentum in utramque partem continued through the
(empeiron) and “an art” (techne), or experimentum and ars in the Latin Middle Ages; see M. C. Woods, “The Teaching of Poetic Composition
translation of Aristotle’s Metaphysics by James of Venice (the earliest in the Later Middle Ages,” in A Short History of Writing Instruction, ed.
medieval translation, ca. mid-twelfth century). J. J. Murphy, 2nd ed. (Mahwah, NJ, 2001), 123–43.
16. See Y. Dudai, Memory from A to Z (Oxford, 2002), s.v. “Acquisition,
24. Aristotle, Topikōn 8.14 (as in n. 21 above). See Stump, Dialectic and Its
Consolidation”; and see also the observation of Quintilian, Inst. orat.
Place in the Development of Medieval Logic.
11.2.43.
25. Discussed in the general introduction to Carruthers and Ziolkowski, The
17. Albertus Magnus, Commentarium in De memoria et reminiscentia, trac-
Medieval Craft of Memory, esp. 9–17.
tatus 2, capitulum 1 (ed. Borgnet, 107); trans. Ziolkowski, 136: “remi-
niscentia nihil aliud est nisi investigatio obliti per memoriam.” See also 26. See Carruthers, “Rhetorical memoria.” The memory craft developed in
Thomas Aquinas, Sentencia libri . . . de memoria et reminiscentia, lectio monastic meditation is, for the Middle Ages, of far greater importance,
6, in Opera omnia 45.2 of the Leonine edition of Aquinas (Rome, 1985). as I demonstrated in The Craft of Thought.
Aquinas’ commentary has been newly translated from this edition by K. 27. Quintilian, Institutio oratoria 11.2.18–31; a new translation in five vol-
White and E. M. Macierowski, Commentaries on Aristotle’s “On Sense umes (Cambridge, MA, 2001) by D. A. Russell is available in the Loeb
and What Is Sensed” and “On Memory and Recollection” (Washington, Classical Library, based on the edition of M. Winterbottom, 2 vols. (Ox-
DC, 2005). A translation of the Marietti edition of this commentary (ed. ford, 1970).
R. M. Spiazzi, Turin, 1973) is in Carruthers and Ziolkowski, The Medieval
28. The conclusions of these essays are brought together in R. Taylor-Briggs,
Craft of Memory, 153–88 (the differences between the editions are slight).
“Reading between the Lines: The Textual History and Manuscript Trans-
18. The oubliette as an instrument of torture belongs to the fevered “Gothick” mission of Cicero’s Rhetorical Works,” in Cox and Ward, The Rhetoric
imagination of the late eighteenth and nineteenth centuries in Europe. The of Cicero, 77–108. The text seems to have been edited in north Africa
earliest attestation of the word’s use in English is 1777, in an account of in the fourth century and thence brought perhaps to Milan and the circle
a visit to sites in France and Spain. Unsurprisingly, the French and other around Ambrose, though its circulation was restricted. Medieval manu-
tourists found oubliettes in England at about the same time, and Walter scripts of the Rhetorica ad Herennium are known from the ninth century
Scott’s Ivanhoe (1819) irrevocably brought them into popular conscious- (representing at least two different stemmae), and glosses are a feature
ness. The narrow-mouthed, subterranean or semisubterranean structures of many of these, some of which may derive from Late Antiquity; a
to which the term is applied in tourist brochures were mostly cellars for complete commentary certainly existed by the later eleventh century, at-
cool storage, though a few actually seem to have been used as prisons, tributed to a “magister menegaldus,” possibly Manegold of Lautenbach:
as is evidenced by their grafitti. see Taylor-Briggs; and also J. O. Ward, “The Medieval and Early Renais-
19. In conspectu Dei/Domini is a common biblical phrase indicating that sance Study of Cicero’s De inventione and the Rhetorica ad Herennium:
God witnesses and so remembers. An example of the trope occurs in the Commentaries and Contexts,” in Cox and Ward, esp. 25–29.

17
29. “Exercenda est memoria ediscendis ad verbum quam plurimis et tuis 38. This now-famous passage in the preface to Richard de Fournival’s Bes-
scriptis et al.ienis”; Julius Victor, “On Memory” (Ars rhetorica, cap. 23; tiaire d’amour, has been often discussed; see esp. V. A. Kolve, Chaucer
ed. C. Halm, Rhetores Latini Minores [Leipzig, 1863], 440); trans. J. Zi- and the Imagery of Narrative (Stanford, 1984); S. Huot, From Song to
olkowski, in Carruthers and Ziolkowski, The Medieval Craft of Memory, Book (Ithaca, NY, 1987); and E. Sears, “Sensory Perception and Its Meta-
298. phors in the Time of Richard of Fournival,” in Medicine and the Five
Senses, ed. W. F. Bynum and R. Porter (Cambridge, 1993), 17–39.
30. Julius Victor, capitulum 23 (ed. Halm, 440); trans. Ziolkowski, 297–98:
“Memoria est firma animi rerum ac verborum ad inventionem perceptio. 39. Pictura can refer equally to both ekphrasis and to paintings; so can de-
. . . Ad hanc obtinendam tradunt plerique locorum et simulacrorum quas- scriptio. The many plans and sketches in Richard of St. Victor’s literal
dam observationes, quae mihi non videntur habere effectum. . . . Ita enim exegesis of the Temple compound in Ezekiel are fully copied in all the
confirmabimus memoriam et adsuescimus optimis, semperque habebimus manuscripts. This work contains several schematics of the sort Adam
intra nos quod imitemur.” seems to have in mind. Hugh of St. Victor’s “picture” of the ark, a work
31. J. Leclercq, The Love of Learning and the Desire for God (New York, found in many manuscripts, never contains drawings, nor does Adam of
1961). Others have discussed these practices since, but Leclercq’s study Dryburgh’s tabernacle. Both Hugh’s and Adam’s works are presented as
remains the best place to begin. See also, on Hugh of St. Victor, I. Illich, ordering schemes for ethical and theological meditation. See Carruthers,
In the Vineyard of the Text (Chicago, 1993) and, particularly, B. Stock, The Craft of Thought, esp. 241–54. A different interpretation of Hugh’s ark,
Augustine the Reader (Cambridge, MA, 1996). as instructions for making a fully drawn and colored chart from which Hugh
taught, has been put forward by C. Rudolph, First, I Find the Center Point:
32. Discussed at length in Carruthers, The Craft of Thought, chaps. 2–4. Reading the Text of Hugh of Saint Victor’s “The mystic ark,” Transactions
33. Augustine, Instructing Beginners in Faith (De catechizandis rudibus) of the American Philosophical Society, 94 (Philadelphia, 2004). W. Cahn
3.5; ed. J. B. Bauer, CCSL, 46 (Turnhout, 1969); trans. R. Canning (Hyde has studied examples of exegetical as well as meditational picturae; see
Park, NY, 2006): “non tamen propterea debemus totum pentateuchum, “Architecture and Exegesis: Richard of St. Victor’s Ezekiel Commentary
totos que iudicum et regnorum et esdrae libros, totum que euangelium et and Its Illustrations,” AB, 76 (1994), 26–49; and “The Allegorical Meno-
actus apostolorum, uel, si ad uerbum edidicimus, memoriter reddere, uel rah,” in Tributes in Honor of James M. Marrow, ed. J. F. Hamburger and
nostris uerbis omnia quae his uoluminibus continentur narrando euolu- A. S. Korteweg (Turnhout, 2006), 117–26. The latter is a meditational dia-
ere et explicare; . . . sed cuncta summatim generatim que complecti, gram that focuses on the menorah candlestick, one of the temple furnishings
ita ut eligantur quaedam mirabiliora, quae suauius audiuntur atque in that commonly served as an organizing device for meditation and study, as
ipsis articulis constituta sunt, . . . aliquantum immorando quasi resoluere in Bede’s De templo Salomonis liber (Migne, PL 91).
atque expandere, et inspicienda atque miranda offerre animis auditorum.” 40. A translation by B. Balint of both the diagram and the treatise of which
My changes are indicated in brackets. See also Carruthers, The Craft of it forms part is in Carruthers and Ziolkowski, The Medieval Craft of
Thought, 62–66. Topical invention, the focus of many studies of oral ver- Memory; 83–102; see also the accompanying preface and bibliography.
nacular composition, is a different phenomenon from this learned medi- As edited in Migne, PL 210, 269–80, the treatise begins with a long
tational tradition, at least in the earlier part of the Middle Ages, though, section copied from Hugh of St. Victor’s De archa Noe (formerly De
since it also involves controlled recollection, it is not wholly unrelated; arca Noe moralia) and then turns to demonstrating the use of the Cherub
see D. F. Kelly, “Topical Invention in Medieval French literature” in Me- diagram as a device for meditation and composition. This version is rarely
dieval Eloquence, ed. J. J. Murphy (Berkeley, 1978), 231–51. found in the manuscripts, however, which usually copy only the expositio
34. As this passage clarifies, the Ciceronians’ memoria verborum is not the or descriptio (both words are regularly used in the title) of the Cherub’s
same as the verbatim rote memory that Augustine mentions here: in rheto- wings and feathers.
ric, memoria verborum is applied to remembering a few difficult words, 41. The wheel commonly shown below the Cherub’s feet is the chariot wheel
names, and phrases by associating their syllables with punning images. described in Ezekiel 10, which moves with the cherubim. Meditation on
One should not suppose that Augustine had learned the Bible by heart the seraphs and cherubs was a feature of desert monasticism: a Syriac ver-
(verbatim) by associating every word in it with a punning image. It should sion is extant of a meditation on the angels’ wings ascribed to Evagrius,
be noted, as the quotation in n. 29 above shows, that Julius Victor uses but its subject is not penitence. There is evidence of Carolingian medita-
the same phrase, “ediscendis ad verbum,” for learning by heart, that is, tion using such a figure in Hrabanus Maurus’ In laude crucis, a devotional
memorizing texts by rote. Learning by heart, though, did not require mak- poem that remained popular for a long time.
ing images for each syllable of text memorized, only for the particular
few an individual might have trouble with. It was, however, a lurking 42. An important discussion of these and related diagrams in medieval Italian
confusion in discussions of memoria verborum in later rhetorics as it is devotional works is L. Bolzoni, The Web of Images (Aldershot, 2004),
for modern histories. 41–81. Bolzoni has identified a clear path of transmission from the de-
vices common in twelfth-century monastic meditations to thirteenth-cen-
35. Augustine, Confessiones 10.14.21: “quasi uenter est animi.” tury and later vernacular devotional works mainly by friars, including
36. See Carruthers, The Craft of Thought, 246–54, for a full discussion of the Jacopone da Todi and Simone da Cascina (with links to the circle of
nature of the pictura referenced in Adam of Dryburgh’s text. De tripartito Catherine of Siena). On the Tower of Wisdom itself, see L. F. Sandler, The
taburnaculi is in Migne, PL 198, 609–796; all translations from it are Psalter of Robert de Lille (London, 1983), and her separate demonstra-
mine. tion of the turris sapientia diagram in Carruthers and Ziolkowski, The
Medieval Craft of Memory, 215–25. A number of other diagrams used
37. Adam of Dryburgh, De tripartito tabernaculi 2.1.77 (Migne, PL 198,
for meditation are translated and demonstrated in The Medieval Craft
683C): “[J]uxta vestram, pater sancte, jussionem, ut saepe dictum tab-
of Memory. See also M. Evans, “An Illustrated Fragment of Peraldus’s
ernaculum in plano quoque, quantum sciero et potero, depingam; jam
Summa of Vice: Harleian MS 3244,” JWCI, 45 (1982), 14–68.
manum appono, quatenus per corporalem etiam aliquatenus cerni possit;
quod et de communi electorum Ecclesia intelligi debet per fidem, et in 43. Carruthers, The Craft of Thought, 7–30, 198–209; and eadem, The Book
singulis electis construi per meditationem.” Corporalis simply means of Memory, 202–17. The Cherub diagram was discussed with several
sense-based and could refer either to a drawing or to a mental image cre- others in the seminal article by F. Saxl, “A Spiritual Encyclopedia of the
ated through verbal ekphrasis, even some combination of the two, like Late Middle Ages,” JWCI, 5 (1942), 82–139. Saxl regarded them as peda-
a simple drawing meant to be enlarged on in meditation. But no extant gogical simplifications of complex theology made for novice students, an
manuscripts contain such a drawing. assumption with which few would now agree.

18
44. The liturgical significance of the conflation is important too: in Isaiah device because his sermons (he preached as many as five a day) are filled
the seraphim sing the Sanctus, and in Revelation the cherubim sing the with remarks that indicate firm control over the order and placement of
Gloria; patristic exegesis had introduced both as singers on each occa- his main topics; see Carruthers, The Book of Memory, 255–57.
sion. More evidence of the conflation can be found in Carruthers and 56. The Howard Psalter is London, British Library, MS Arundel 83-I, an
Ziolkowski, The Medieval Craft of Memory, 84–86. East Anglian manuscript made about 1310–20, for an aristocratic house-
45. See ibid., 1–23. hold; see the catalogue description of L. F. Sandler, Gothic Manuscripts,
46. In The Web of Images, 117– 35, Bolzoni discusses the use Bernardino 1285–1385, A Survey of Manuscripts Illuminated in the British Isles, 5
made of the Cherub diagram during a set of Lenten sermons he preached (London, 1986), no. 51.
in 1424. The topics are appropriate to penitence, in keeping with the 57. Clark Maines suggested to me that these textual bands are in fact “ab-
liturgical season, but their content is quite different from those in the stracted” from the separation between the shorter and longer feathers
Cherub drawings we still have. of a bird’s wings, an area that was used as a field for the titulus in other
47. Note that the manuscript has been numbered by pages rather than folios. Cherub drawings, as, for example, in Figure 3. I thank him for this excel-
On this manuscript, see the catalogue description of C. M. Kauffmann, lent observation.
Romanesque Manuscripts, 1066–1190, A Survey of Manuscripts Illu- 58. L. F. Sandler discussed the possible uses of the Speculum theologie dia-
minated in the British Isles, 3 (London, 1975), no. 102. On the Durham grams in The Psalter of Robert de Lisle, see esp. 32–34, 82. Though
origin of the manuscript (which initially included what is now Cambridge apparently collected up in the late thirteenth century by John of Metz, a
University Library, MS Ff.1.27), see C. Norton, “History, Wisdom and Franciscan friar working in Paris, and often incorporating favored Fran-
Illumination,” in Symeon of Durham, ed. D. Rollason (Stamford, 1998), ciscan material like the Lignum vitae of St. Bonaventure, complex dia-
61–105; and P. Binski, Becket’s Crown (New Haven, 2004), 54–62. grams, some to be imagined even with moving parts, are rather common
48. The text is in Migne, PL 163, 759–763, there dated ca. 1108. A Norman invention devices; see Carruthers, “Moving Images in the Mind’s Eye”
cleric from Caen, Theobald d’Étampes is the earliest named magister for several examples from the twelfth–thirteenth centuries especially,
(whatever that may mean at this time) in Oxford. Robert Bloet was bishop some drawn, some presented as only verbal ekphrasis.
of Lincoln from 1094. See R. W. Southern, “From Schools to University,” 59. On this interesting manuscript, see the catalogue description of N. Mor-
in The Early Oxford Schools, ed. J. I. Catto, vol. 1, The History of the gan, Early Gothic Manuscripts, 1190–1250, A Survey of Manuscripts
University of Oxford, ed. T. H. Aston (Oxford, 1984), 5–6. Illuminated in the British Isles, 4 (London, 1975), 1: no. 80. As well as
49. On this point, see M. Carruthers, “Moving Images in the Mind’s Eye,” the Cherub, it contains the Knight against the Vices figure, discussed at
in The Mind’s Eye: Art and Theological Argument in the Middle Ages, length in the context of this manuscript by Evans, “An Illustrated Frag-
ed. J. Hamburger and A.‑M. Bouché (Princeton, 2006), 287–305; and ment of Peraldus’s Summa of Vice.” Evans dates the manuscript to about
Carruthers, The Craft of Thought, 116–70. 1255. Peraldus’ text was popular for pastoral care of the laity; it is one
50. Oxford, Bodleian Library, MS Auct. D. 2. 1, dated to about 1200 and source of Chaucer’s “The Parson’s Tale.”
described as “possibly” from the library of Llanthony; see O. Pacht and 60. Evans, “An Illustrated Fragment of Peraldus,” 38. The structure of the
J. J. G. Alexander, Illuminated Manuscripts in the Bodleian Library, present manuscript is described in detail by Evans, 43–45; its original
Oxford, 3 vols. (Oxford, 1973), 3: no. 226. The text occupies fols. 2–6, order is discussed, 38–41. A list of the contents in their present order on
the rest of the book being a Psalter with the (abbreviated) commentary fol. 1, is in an early-fifteenth-century hand.
of Gilbert de la Porrée—material helpful in study and preaching.
61. The text of De sex alis is found often with Alanus’ Liber penitentialis,
51. Cambridge, Corpus Christi College, MS 29. Many thanks to Grover Zinn especially as it came to be attributed in the thirteenth century to Alan
for giving me this reference. of Lille; see especially M-T. d’Alverny, “Alain de Lille: Problèmes
52. Florence, Bibliotheca Medicea Laurenziana, MS plut. 30. 24. In The Web d’attribution,” in Alain de Lille, Gautier de Châtillon, Jakemart Giélée
of Images, Bolzoni discusses the figures in this manuscript at length, the et leur temps, ed. H. Roussel and F. Suard (Lille, 1980), 27–46. It is,
Cherub and also several trees, including Bonaventura’s Lignum vitae, and however, associated in a few manuscripts with a section of Hugh of St.
many of Joachim of Fiore’s apocalyptic images. An apocalyptic theme Victor’s De archa Noe. See Carruthers and Ziolkowski, The Medieval
runs through many of these meditation devices, unsurprisingly, since the Craft of Memory, 83; this is the version published in Migne, PL.
Apocalypse was such a major site in monastic and later devotions for 62. “Fili accedens ad seruitutem dei. preparate ad temptacionem.”
meditational image-making.
63. “militia est uita hominis super terram.” The miniatures in this manuscript
53. On the left, “Cherub iste in humana / depictus effigie. / Sex habet alas /
are available on the British Library’s online catalogue.
que sex actus morum re / presentant. Quibus debet / fidelis anima redi-
miri. / Si ad Deum per incrimenta uir / tutum uoluerit peruenire.” On the 64. All the images in this manuscript are available online through the The
right, “Rota sub pedibus che/rub. Septem habens / radios. Septem opera/ Beinecke Library’s site.
misericordie. designat / Que dominus commina/tur se improperaturum/ 65. L. F. Sandler identified the script as early fifteenth century in The Psalter
in die Iudicii negglige/ntibus remissis.” Except for minor spelling varia- of Robert de Lisle, Appendix 3, no. 15, p. 136. R. Rouse has argued it
tions, these are the same texts as those occupying the same positions in could be a fourteenth-century hand contemporaneous with that of the dia-
the Howard Psalter Cherub (Fig. 4). But the Howard and Laurenziana grams in fols. 1–7, but less formal. For both, see B. Shailor, A Catalogue
Cherubs are not otherwise related. The Latin texts are identical also to the of Medieval and Renaissance Manuscripts in the Beinecke Rare Book and
Cherub figure of the DeLisle Psalter, transcribed and translated by L. F. Manuscript Library, Yale University (Binghamton, NY, 1987), 2:329– 30.
Sandler, The Psalter of Robert de Lisle in the British Library (London, 66. “Sub umbra alarum tuarum protege me,” a text that is often included in the
1999), 82, 106. explicit of De sex alis, though not in the text in this manuscript. Instead,
54. The text makes clear that this virtue refers to moderate and decorous use it has been “displaced” to the titulus of Hermanus’ meditational source,
of perfumes; it is interesting that all of the sensory virtues described here the book he holds in his hands.
have to do with giving an appropriately modest and controlled social 67. I discussed this image at greater length in my essay “Moving Images
impression. in the Mind’s Eye,” in particular stressing how, as the visualization of
55. As is the case with the Lenten sermons of friar Giordano of Pisa; he did a complex concept, it implies movements that must be realized in the
not appear to use the Cherub figure, but he evidently used some such viewer’s imagination.

19