重探修辭記憶術(shù)：記憶的物質(zhì)性

蘇文伶

（臺(tái)灣輔仁大學(xué)）

重探修辭記憶術(shù)：記憶的物質(zhì)性

蘇文伶

（臺(tái)灣輔仁大學(xué)）

［中文摘要］：

一般認(rèn)為西方的修辭記憶術(shù)為口語文化產(chǎn)物，早在印刷術(shù)盛行后過時(shí)而自然淘汰。此外，西方形上學(xué)受柏拉圖影響一向?qū)⒂洃浺暈榉瓷韮?nèi)省，修辭記憶術(shù)屬外部助憶工具，囿于其物質(zhì)性，往往窒礙心靈、造成遺忘。德希達(dá)與保羅德曼認(rèn)為思想須透過符號(hào)中介，因此記憶并無所謂內(nèi)外。修辭記憶術(shù)因此并非次等擬仿物，而是與書寫文字一樣的符號(hào)系統(tǒng)；意義為其結(jié)構(gòu)產(chǎn)物，無關(guān)外緣指涉，且因具延異特性而充滿不確定性。本文采德希達(dá)與保羅德曼的后結(jié)構(gòu)主義觀點(diǎn)重新檢視修辭記憶術(shù)，探討其中主要的三個(gè)組成要素（記憶圖像、場(chǎng)所記憶、記憶順序）在歷史上的流變，尤其聚焦其穩(wěn)定的假面下隱約浮動(dòng)的詭異暗影。換言之，即翻轉(zhuǎn)傳統(tǒng)內(nèi)外偏見，而將人工記憶視為自然記憶原型。

修辭記憶術(shù)；記憶圖像；場(chǎng)所記憶；記憶順序；后結(jié)構(gòu)主義

Notes on Author：Wen-ling Su，female，currently a Ph.D.student in the Graduate Institute of Cross-Cultural Studies at Fu-Jen University，Taiwan，and a lecturer in the English Department at Fu-Jen University，Taiwan.Areas of interest include semiotics，rhetoric，and memory discourse.

Mourning for Paul de Man in Memories，Jacques Derrida paradoxically speaks of an“impossible mourning”that he characterizes as the“essence of memory”.Mourning is impossible because it presupposes an identity between the proper name and the being，a unity across a chasm of unbridgeable differences.The representation concealing the aporia nevertheless still constitutes some kind of a bridge，commonly known as the memory trace. While Plato prioritizes memory as“presence of the absent”and Aristotle maintains that“memory is of the past”（De memoria at reminiscentia，449b 15），Derrida conceives of memory as a futureoriented speech act that promises to remember.Memory is a promise，to the deceased and the bygone，albeit a promise whose fulfillment is uncertain.Much like an allegory，moreover，the promised memory，“always”implies“an unreachable anteriority”that is steeped in its own alterity.The past is indeed a foreign countrywhere surprises abound，such as the posthumous disclosure of pro-Nazi papers that de Man wrote during the Holocaust era andkept quietabout afterwards，an alien Other to the people who used to know him，a concrete presence as indigestible as a fishbone stuck in the throat.Despite the potential of pain resulting from such disruptions and divergences，Derrida insists that we harbor an openness to the specter of the Other in memory“for the right to error，even to an aberration”.

To keep a promise is to keep one's word，like being bound by a contract or by law.Memory is thus shot through and through with textuality，complete with its possibilities and limitations.Much of Derrida's discussion in this regard is indebted to de Man's article“Sign and Symbol in Hegel's‘Aesthetics’”，where de Man differentiates between Erinnerung and Ged?chtnis，a distinction originally introduced by Hegel.While Erinnerung is“recollection as the inner gathering and preserving of experience”，Ged?chtnis，retaining the residual meaning of“thought”from the word root denken，refers to“the capacity of remembering by memorization”.De Man suggests that the opposition between Erinnerung and Ged?chtnis，i.e.，between recollection and memorization，resembles the famous dichotomy Proust establishes between mémoire involontaire and mémoire volontaire.The analogy is illuminating because it reveals an implicit hierarchy inboth pairs of binary opposites.Just as Proust privileges mémoire involontaire，Hegel，“the theoretician of internalization”，favors Erinnerung.De Man，however，is shrewd to point out that“the progression from perception to thought depends crucially on the mental faculty of memorization”.The“machinelike exteriority”of memory，in other words，is constitutive of thought，rather than contingent upon it.This“outward”turn has great ramifications for the post-structuralist view of memory，as it is endorsed by Derrida in Memories and later in Archive Fever，among others.Most significantly，de Man argues that，if thought is never devoid of the exterior auxiliaries of memorization，which Derrida would later call a“prosthesis of the inside”or“hypomnēma”，the word“art”should be restored to its original sense of technè.The study of memory would then become the study of the art of memory，especially that of the artificial mediating mechanism conceived as inscription on the wax block of the mind，with writing serving as aide-mémoire and a sign system functioning as an automatic thoughtgenerating mechanism，viz.as an archive that both produces and conserves.Nevertheless，while Derrida maintains that“There is no archive without a place of consignation，without a technique of repetition，and without a certain exteriority.No archive without outside”，he also argues for an archival violence analogous to the death drive that would usher in a revolutionary beginning amidst the ruins of its own making.

For Derrida and de Man，all memory is hence externalized memory by virtue of the technèinvolved.Rather than viewingmemoria，or mnemotechnics，merely as a form of manipulation and abuse，Derrida and de Man suggest treating such artificial memory systems as prototypes of all natural memory，an intriguing perspective that calls for a reevaluation of memoria.Since the publication of Frances A.Yates's pioneering work in the 1960s，considerable scholarship has been devoted to excavating the once prominent tradition，such as the work of Paolo Rossi，Jonathan D.Spence，Mary Carruthers，Lina Bolzoni，Patrick H. Hutton，and Sharon Crowley，to name only a few.With perhaps the exception of The Book of Memory by Carruthers，little effort has been made to integrate the poststructuralist insight shared by Derrida and de Man.Much has been devoted to describing the mnemonic systems，but not the disorienting elements resulting from archival violence.Even Derrida and de Man themselves did not examine memoria in detail either.This paper therefore aims to address this lacuna by a revisit to memoria from antiquity up to the pre-modern period.While delineating the development of memori a in three of its most prominent features：mnemonic images（imagines agentes），memory-places（loci），and the order of recollection，the paper highlights areas where uncertainty is registered in whatappears to be a seamless schema.

The origin of memoria，one of the five major divisions of traditional rhetoric，has come to be well-known.Legend has it that Simonides of Ceos invented the mnemonic technique at a banquet in which he participated.He helped to identify corpses after the roof of the banquet hall suddenly collapsed by remembering where each guest had been seated（Cicero，De oratore，2.86. 352 53；Quintilian，Institutio，11.2. 11 13）.The mnemonist emerges as a Promethean hero in the defiant act of resurrecting the dead and the forgotten；ghosts of death and oblivion nevertheless still lurk in the mnemonic ruins，a loss eternally commemorated as the origin. Ironically，this origin might turn out to be an absent center after all.Quintilian，for instance，suspects the story is mythical，for he finds it odd that Simonides“himself nowhere mentions it，though he would surely not have kept silent on an affair so glorious to himself”（Institutio，11.2.16）.Whether fact or myth，Simonides's story inspired orators in antiquity to follow his footsteps in dividing up speeches into small bits of information that they associated with different images，which were then deposited in specific mental locales to facilitate retrieval.Upon delivering a speech，the orators would take an imaginary stroll in the mental treasure-house，following the same order of deposit，to re-collect the information stored therein.

Often disparaged as mechanical and repetitive，mnemotechnics is nevertheless mechanical in the sense that writing is a writing machine，viz.an autonomous sign-generating system，while its repetition is a symptom of cultural construction and its concomitantrepression，as psychoanalysis proposes.The first Chinese treatise on memoria，Xiguo jifa（Western Mnemonics，〈西國(guó)記法〉），a book by Matteo Ricci，the Jesuit priest that founded the Chinese mission in the sixteenth century，is a case in point.To impress the local literati in Nanchang，China，where Ricci resided from 1595 1598，he demonstrated the ability to recite Chinese classics both forwards and backwards after viewing the text only once.At the request of a local governor，Ricci wrote Xiguo jifa，which details the method of loci that Ricci called“memory palace”.In one of the very few examples showing memoria at work，Ricci creates four mnemonic images—“wu”（military［武］），“yao”（to want［要］），“l(fā)i”（profit［利］），and“ha”（good［好］）—each of which is a Chinese character and is placed in one corner of the room.Though the four words do not compose an idiomatic expression，the selection might not be as random or innocent as it seems，or so Jonathan Spence demonstrates.The most obvious is the last character“hao”（“good”），which，combining the ideograms of a woman and a child，as Spence points out，comes close to the Madonna icon that embodies divine love and salvation，and thus a memory image that Ricci hoped his Chinese readers would commit firmly to memory as he did.The ideological maneuver Ricci might have intended suggests traces of cultural construction，which，as will be shown later，was masked as ethics in medieval times.Spence's rendering of Ricci's memory“work”helps to debunk the myth that memoria，like pronuntiatio or actio（delivery），consists of nothing but trivial and monotonous techniques.In Memory，History，F(xiàn)orgetting，Ricoeur，for instance，likens training in artificial memory to physical manipulation in the behaviorist lab where only animal instinct counts.In contrast，the other parts of rhetoric—inventio（invention），dispositio（arrangement），or elocutio（style，expression）—require more in-depth engagement with ideas.The erroneous impression that mnenotechnics is devoid of intellectual import might very well have led to its relative neglect.

On the other hand，as Ricci's example belongs to memoria verborum（“memory for words”），one of the two kinds of memory included in memoria（the other kind being memoria rerum—“memory for things”），it reveals artificial memory as writing in the most literal sense.In western metaphysics，where memory is synonymous with internalization（i.e.，Erinnerung），writing is said to pose threat to memory.The perspective is most clearly illustrated in Plato's Phaedrus.When presented with the new invention by Theuth，the Egyptian king Thamus comments：

this invention wiII produce forgetfuIness in the minds of those who Iearn to use it，because they wiIInot practise their memory.Their trust in writing，produced by externaI characters which are no part of themseIves，wiII discourage the use of their own memory within them.You have invented an eIixir not of memory，but of reminding；and you offer your pupiIs the appearance of wisdom，not true wisdom....（Phaedrus，275a）

Forgetfulness will result，and by implication，death and annihilationof the soul as well，when inner thought is emptied out and replaced by writing（“external characters which are no part of themselves”）. Unlike the“winged chariot”of natural memory that is capable of soaring and carrying one into transcendence（Phaedrus，246e）—memoria，as an artificial system composed of external cues serving only to remind，is akin to writing，a pharmakon not so much a medicine as a poison to the mind.Implied here，too，is a critique of the sophists，whose memorized speeches made to order，for Plato，boast of false wisdom.

A crude alternative to writing，memoria has been made obsolete by the rise of the printing press since the Renaissance，so much so that it was usually omitted from rhetoric textbooks by the sixteenth century.What with technical advances in exteriorized memory，and what with Plato's disparaging appraisal，the value of memoria as an archetypal semiotic system was not fully recognized until the so-called“l(fā)inguistic turn”in the twentieth century that gave rise to structuralism and post-structuralism.It is from this revisionist perspective that this paper proposes a critical review of memoria's basic tenets.In the three most authoritative classical texts on memoria—Rhetorica ad Herennium by an anonymous author，De oratore by Cicero，and Institutio oratoria by Quintilian—the authors all identified images（imagines agents）and places（loci）as the staple elements of the artificial memory system（Ad Herennium，3.16.29；Cicero，De oratore，2.86.354；Quintilian，Institutio，11.2.22）.Recent scholars highlight a third element，that is，the order of recollection，which would accrue great importance later on in history，especially in the heyday of methodical memory in the eighteenth century.The discussion in the following will therefore be organized around these three key components of mnemotechnics.

Mnemonic Images：Beyond Iconicity

Memoria operates on the premise that memory，as thought of things past，involves images，or，to be precise，lively images（imagines agents）.The classical tradition hence gave rigorous training in rendering words and ideas into such images until one built up a sizable iconological repertory that can be tailored to any occasion，of which the 120 Chinese characters listed at the end of Ricci's Xiguo jifa is a good example.The texts to be committed to memory are then translated into a series of images in a building，like paintings or sculptures in a gallery.The hypothesis that the orator could easily conjure up visual representations in the mind's eye was a commonplace in antiquity and has had wide-ranging influence.As Carruthers observes，the idea that“the human mind requires a sort of image with which to think”in fact characterizes ancient and medieval epistemology.This classical epistemology，which is largely predicated on mimesis，is however contested by poststructuralists like Derrida and de Man.As they adopt an alternative epistemological model that practically amounts to ananti-epistemology，the so-called mnemonic“image”is at best a misnomer for what should be more appropriately called a“name”. De Man，for instance，argues that Ged?chtnis，analogous to memorization，is“entirely devoid of images（bildos）....But it is not devoid of materiality altogether.We can learn by heart only when all meaning is forgotten and words read as if they were a mere list of names”.Identified as empty signifiers as such，the mnemonic“image”is founded on semiotic signification，rather than on mimesis or iconicity.

The poststructualist critique of mnemonic images has been long in the making，but not without precedents.Though often used in combination with one another in actual practice，the various types of mnemonic images are invariably referred to as icons of that which it represents.The anonymous author of Ad Herennium reveals a wider range of possibilities by defining the mnemonic image as a“figure（formae），mark（notae），or portrait（simulacra）”（Ad Herennium，3.16.29），with perhaps only“portrait（simulacra）”fulfilling the description of imagines agents.The iconic or mimetic model is commonly traced back to Aristotle，who，in On Memory and Recollection defines memory as“a condition of mental picture（φαντασμα），related as a likeness to that of which it is an image”（451a 15），that is，an imprint resulting from the“affection”（παθο）of an experience“when time has elapsed”（449b 25）.The wax block metaphor of the mind was a commonplace in antiquity，the mostnotable example of which is a reference to it in Plato's Theaetetus. Plato and Aristotle however attribute different sources to the“affection”on the psyche：while，for Plato，memory images are spiritual imprints—“recollection of the Ideas”or“archetypes of reality”—for Aristotle，the imprint is most likely derived from an embodied experience.Regardless of this fundamental difference，both Plato and Aristotle posit an origin，prior in time and order，of which the mnemonic icon is a derivative copy，hence inferior and secondary.The iconicity of the mnemonic image is challenged from two directions，with one further underscoring the primacy of the origin，while the other undermining it.The former objects to what is akin to a“tyranny of the eye”in the past.In search of authenticity，critics in this camp demand a multi-modal representation of the original experience.The latter，on the other hand，equates mnemotechnics with writing and mnemonic images as arbitrary signs.

Memoria has long been viewed as a primarily pictorial method. For instance，David Bloch asserts，in his latest translation of On Memory and Recollection，that Aristotle's concept of memory involves nothing more than“the actual viewing of images present to the attention of the remembering subject”.Bloch，however，is deliberately arguing against the grain.Even in classical texts on memoria such as De Oratore（2.87.357）and Ad Herennium（3.16.29），the authors would take pains to add hedge words like“a sort of”or“as it were”to the word“picture”when referring to mnemonic images.In fact，in De Oratore，Cicero notes that，though“the keenest of all our senses is the sense of sight，”mnemonic images encompass imprints of other senses.Unlike Bloch，most contemporary Aristotle commentators are apparently aware that the pictorial view of memory is no longer tenable. Richard Sorabji observes that，although many still believe in mental representation of experience，“far fewer are attracted to thinking of this as like a picture”.In addition to visual images，the thought of things past may include many other kinds of sense impressions（aisthêma），such as“simply thinking over，or recounting，or reenacting a childhood scene without imagery，”or“merely finding something familiar when one sees it”.

Even Aristotle himself，Sorabji suggests，regards the memory“phantasm”as more than a visual duplicate.Sorabji retains the Greek wordφαντασμα（“phantasm”），for which“image”proves to be a poor translation.John Burchill adopts exactly the same strategy when translating Thomas Aquinas's commentary on De memoria in order to preserve the word's rich polysemy.He claims that phantasms connote the“concrete corporeality of all sense objects”that is involved in the phenomenology of lived experience.A memory phantasm is an imagines agents or lively image indeed. Note that the emphasis lies on a more adequate representation of the original.Burchill also suggests that Aristotle's phantasms contain a certain intentionality that can“move”people so much as to set their recollection into motion.Burchill's point is supported by theanonymous author of Ad Herennium，who gives the same suggestion that the most effective mnemonic images provoke strong emotions（3.22.35）.A novel or striking image would presumably engage all the senses of a person，much akin to synesthesia.Imagine all the bodily reactions accompanying the sense of awe and wonder. Bolzoni hence envisages that memory images confer an aura of a certain“disquieting autonomy”comparable to the“phantoms of Eros”in Plato's Phaedrus.The image as a copy（eik?n）thus proves to be far too flimsy and inadequate a match for what the word“phantasm”entails.In a famous study of the mnemonist S（S.V. Shereshevskii）in the 1920s，for instance，Russian neuroscientist A.R. Luria discovered that S.'s memory phantasms were characterized by a marked degree of synesthesia：“every sound he heard immediately produced an experience of light and color and...a sense of taste and touch as well”.The most effective mnemonic phantasm is a multi modal“image”then，with the so-called“image”serving as a metaphor only.Aristotle's iconic／mimetic model of the psyche is further challenged by a distinction Sorabji makes between phantasia and phantasma：the latter remains“an image like a picture，”whereas the former denoting the faculty producing the latter in memory and the imagination.Phantasia then refers to a cognitive mechanism that encompasses a lot more than what meets the eye.According to the eminent philosopher Martha Craven Nussbaum，phantasia works to grasp experience.In corollary，phantasma should be properly understood as“interpretations of the appearance of things，”viz.as Wiggenstein's“seeing-as，”thereby turning memory into hermeneutics.

Bloch's insistence that Aristotle focuses exclusively on vision in De memoria might very well target a specific contemporary theory：an anti-mimetic theory that conceives of memory as writing.As Sorabji observes，the upshot boils down to the question“whether these postulated brain images are pictorial，or of a more linguistic character”.In The Art of Memory，Yates maintains repeatedly that artificial memory is“inner writing”；its interiority is attributed to its invisibility.In The Book of Memory，Carruther also forcefully argues that memory，as ancient and medieval people conceived of it，is both verbal and visual.Note that Carruthers uses the word“visual”instead of“pictorial，”perhaps in order to steer away from the connotation of likeness or mimesis.On the point of memory being a“visible”writing system，she parts way with Yates.Carruthers contends that memoria is writing per se，implying that no distinction should be made between what is inner and what is outer，a view coming closer to the dictum of Derrida's deconstructionist position—“il n'y a pas de hors-texte”（“there is nothing outside the text”）.As the wax block metaphor indicates，Carruthers argues，“the mind...writes when it stores up its experience in representations”.Mnemotechnics，then，does not so much reproduce things past or perform“rote”memory as function like a writing system，where the so-called mental images are in fact signs serving as cues for things to remember.She observes，

Even the most apparently pictorial of mnemonic systems are based

on principles governing the nature of signs rather than on iterative

copying.Most require that the“picture”relate to the world or

concept it marks for recollection via a pun or homophony.

Memory phantasms therefore“need bear little resemblance to the form in which the information was originally received”.According to Charles Sanders Peirce，the relationship between the sign and the object it refers to in an icon is characterized by resemblance，while in an index by a causal relationship，and in a symbol by conventions.Carruthers hence suggests that，in addition to iconic signs，mental images comprise a larger variety of signs like the index and the symbol.

Carruthers's view is supported by Quintilian in Institutio Oratoria，where，aside from imagines agentes，he also uses signum（translated as“symbol”by Butler）to refer to mental images. Quintilian appears to be a proto-semiotician as he remarks：“thoughts do not call up the same images as material things，and a symbol requires to be specially invented for them”（Institutio，11.2.24）.However，though he endorses memoria rerum（memory for things），Quintilian is critical of memoria verborum（memory for exact words），arguing that it is impossible to create a symbol for all words，such as conjunctions，and that the semiotic system，like the inventory of shorthand writers（notis scribunt），being extraneous to mind，might obstruct，rather than facilitate，thinking（Institutio，11.2.25）.Unlike Quintilian，who is dubious about trained memory，Carruthers regards memory phantasms as heuristic mechanisms.Far from extraneous，these mnemonic symbols are integral to cognitive processes，such as learning，reading andwriting.Books are the most prominent examples of such symbols in the medieval period：“the page as a whole，the complete parchment with its lettering and all its decoration，was considered a cognitively valuable‘picture’”.Another example of medieval mnemonic symbols is“sacred iconography”.In The Craft of Thought，Carruthers points out that“Monastic art is...an art for mneme，‘memory，’rather than one for mimesis，”while memory phantasms range from religious icons to tropes and figures，which in turn constitute the“craft”of monastic reading.Notable among these“memory-resident tools”in the monastic tradition are“elaborate schemes of images of virtues and vices，”which explain why Albertus Mangus and Aquinas considered memoria to be part of Prudence and ethics.

The divide between the mimetic and anti-mimetic theories of mnemonic images widened after the medieval times.The classical model was pushed to the margin with the gradual rise of the scientific method，which leans toward abstraction.In general，memoria bifurcated in two directions during and after the Renaissance，one extending further into the occult，pushing the fantasia composed of images and places to its extreme，while the other underscoring order（logical order，to be exact）in memory，eventually transforming into the scientific method.Yates characterizes the two competitive trends as the“conflict between Brunian and Ramist memory”，namely traditions headed by Giordano Bruno（1548 1600）and Peter Ramus（1515 1572），respectively.In fact，she explicitly identifies them as“two different types of mind”in the Renaissance：the“irrational”occult vs.the“rational”humanist.That being the case，the growing popularity of the latter testifies to the increasing emphasis of memory places over images，and of reason over imagination，that would later lead to the quest for method in the Age of Enlightenment.Despite the charge of irrationalism，the occult turn of memoria moved toward greater autonomy like a well-crafted clockwork.From the systems perspective，I would argue，the aggregate of fantastic images did not become less like a writing machine because of its penchant for mysticism.

The rise of occultism in Renaissance memoria signifies a revolt against Aquinas's Aristotelian leaning，in fact against anything smacking of dark medievalism.The revival of Ciceronian oratory in Italy，especially in Venice，was closely associated with a“mysticmagical artificial memory”.Instead of corporeal similitude，mnemonic images dotting the celestial field were infused with mysterious spiritual energy.The encyclopedic knowledge stored in the memory vault was no longer static but took on a life of their own，as if assuming three-dimensionality，like the illusion created by means of linear perspective in Renaissance paintings，or simply like characters walking on the stage.Among Bruno's most illustrious precursors is Giulio Camillo，whose Memory Theatre，as legend has it，was an actual wooden architectural structure crowded with images that represents the peak achievement of Renaissance memoria.Images in Camillo's theatre were characterized as“talismanic，astralised mythological”.Typical examples ofmnemonic talismans are images of the stars，such as“an image of Venus as the goddess of the planet Venus，or an image of Apollo as the god of the planet Sol”.In the words of a visitor to Camillo's theatre：Camillo

pretends that all things that the human mind can conceive and which we cannot see with the corporeal eye，after being collected together by diligent meditation may be expressed by certain corporeal signs in such a way that the beholder may at once perceive with his eyes everything that is otherwise hidden in the depths of the human mind.And it is because of this corporeal looking that he calls it a theatre.

The claim of all-inclusiveness is a major characteristic of a reductive system.The mnemonic images in Renaissance's artificial memory，as embodied in Camillo's Memory Theatre，then function as semiotic symbols.Following the occult code，the meaning，i.e.，the mystery hidden beneath the sign，can be unveiled and deciphered.Renaissance memoria also turns the mnemonist into a magus who is an expert reader of signs，aside from being a sorcerer，a wise man，and a man of great imagination at the same time.For the Italian humanist，a mnemonist would have been a man capable of miraculous in-sight into the hidden truth，one who is able to“see into the life of things.”The insight，however，is derived not so much from light，as from shadows.There is no transparency of meaning，even for the magus-mnemonist.

這樣的情況比比皆是，集中出現(xiàn)于17世紀(jì)早期《圖像學(xué)》版本的文本中。在對(duì)待前人著述中的“既有”成果方面，里帕作為一名“知識(shí)搬運(yùn)者”，保持了文藝復(fù)興時(shí)期對(duì)古典文化一以貫之的尊崇態(tài)度。

For the occultists，the memory trace consists primarily in shadows.For light does not reveal itself directly，but only throughits silhouette.But what are shadows if not the binary opposites of light？What are shadows if not different shades of light diffused and deferred in space？The occult sense of light is then shot through with Derridean différance.Created independently of Camillo's Memory Theatre，Bruno's artificial memory systems，also indebted to the Hermetic tradition，demonstrate the same emphasis on secret knowledge，which could be uncovered only by the initiated.In Bruno's book De umbris idearum（Shadows），as a Brunian follower put it，“The‘umbra’or image is as a shadow of the light of the divine mind which we seek through its shadows，vestiges，seals”. Bruno's images of the stars might be flimsy like shadows，but they mediate between Ideas and phenomena，as substantial as dream work. Bruno's phantasms function as mediation between the mind and the world：“‘seals’，signs and figures which，together with gestures and ceremonies...seen as the basic elements of a mystic-ritual language which opens the way for divine colloquies”.The mnemonist that arranges and manipulates the signs then is engaged in an activity similar to writing，which is a way to“act on”the world.The world of writing“shadows”the material world and attains virtual reality.

Camillo's and Bruno's Hermetic memory，though popular in Italy and France，was seriously challenged by the more“rational”camp of humanists，such as Erasmus and Ramus，whose memory systems are precursors of modern semiotics.As Yates shrewdly observes，“whilst the controversy is always ostensibly about two opposed arts of memory，it is at bottom a religious controversy”. The occult images，in particular the animation of the images by the magus-mnemonist，come too close to idolatry，which is strictlyprohibited in Christianity.Embedded in the Ramist camp，which later evolved into the scientific method，is a Puritan urge of iconoclasm.By integrating memoria into logic，the Ramist method dispenses with both loci and imagines agentes，as well as any association with the scholastic insistence on corporeal similitude.Though often blamed for bringing about the demise of the classical art of memory，the Ramist method in fact preserves it as a spatial method，as will be demonstrated below.Yates is right，however，in noting that“The dialectical method was emotionally aseptic”.The major difference between memorizing Ovid's lines through logical disposition and doing that with the Ovidian images lies in the absence of affect.Yet is it true that as images are gone，so is imagination？If by“images”Yates refers only to“pictorial images，”which I think she does，then the answer is a resounding NO.Memory phantasms are semiotic symbols，while memoria as externalized memory is a kind of writing.From this perspective，artificial memory might morph into other semiotic systems，with or without the pictorial element involved.

Loci：Toward a Systems Perspective

Images have always been subordinated to locality in memoria. Simonides's invention is，after all，a re-construction of the collapsed banquet hall，i.e.literally a“place”，and thus it is predominantly a method of loci.Just as Ricci securely harbors the Madonna and Child in the northwestern niche of the room，so Marcel deposits，in the memory vault of the town Combray，the enticing colors andfragrance of the hawthorn blossom that bring him to his first love in Proust's A la recherche du temps perdu.In the most authoritative texts on memoria in antiquity，localities are invariably advised to be set up first，i.e.，before mnemonic images or symbols are selected. Why emphasize the priority of loci over imagines agentes？If，according to Carruthers，mnemonic images are semiotic symbols acting as retrieval cues，then what is the function of the loci in which these images are embedded？Are there any difference between loci in natural memory and artificial memory？In Remembering，which includes a pioneer study on place memory，phenomenologist Edward S.Casey comments that“memory of place，of having been in a place，is one of the most conspicuously neglected areas of philosophical or psychological inquiry into remembering”.He proposes that，while images align memory with imagination，place memory underscores the situatedness（being-inthe-world）of experience.As if echoing Cicero's view in De Oratore that“a material object without a locality is inconceivable”（2.87. 358），Casey observes that“as embodied existence opens onto place，indeed takes place in place and nowhere else，so our memory of what we experience in place is likewise place-specific：it is bound to place as to its own basis”.Much as Casey sheds light on how place is an integral part of Dasein，his view does not account for the primacy given to locality in artificial memory systems.As the discussion on mnemonic images reveals，maybe we have to think outside the box of mimetic representation.

Practical concerns over realism predominate as regards loci in classical memory treatises.The focus invariously falls on thetechnical aspects，such as the layout of the architectural design.In De Oratore，Cicero only mentions that“one must employ a large number of localities which must be clear and defined and at moderate intervals apart”（2.87.358）.Confusion results from blurry and indistinct boundaries；vagueness and obscurity should by all means be avoided.To make an image stand out from its background，a certain drama will be ideal.Differential relationships are hence accentuated.The mnemonist S.studied by Luria intuitively stresses the same principles governing the physical features of his mindscape—“clarity，contrast，the ability to isolate a figure from its background，the degree of lighting available，etc.”.The emphasis on clarity in perception indicates that the spatial method was initially heavily reliant on vision and iconicity.

As memoria gradually broke away from realistic considerations in history，locality took on increasing structural significance.For one，a memory locale can be revisited.It has to allow for reiteration so that recollection could take place.For another，it has to accommodate infinite variations with limited resources.After all，as Yates notes in The Art of Memory，most people only have a limited repertoire of places they are familiar with，so much so that the same set of loci have to be recycled for remembering different material.Rossi proposes that the difference between memoryplaces and images is“the same as the difference between the fixed and the non-fixed”.The locality is like a wax writing tablet whereas the images are like the letters written on it（Cicero，De oratore，2.87.354；Ad Herennium 3.17.30；Quintilian，Institutio，11.2.21）；“the images，like letters，are ef faced when we make no use of them，but the backgrounds，like wax tablets，should abide”（Ad Herennium，3.18.31）.Insofar as the same place could serve as the background for many different sets of images，memory-places help to establish the sense of grounding or stability，that is，the immutability of mutability.In terms of semiology，memory loci could then be likened to the axis of syntagm，the memory trace upon which mnemonic images functioning as signs could be freely associated and combined.In terms of cognitive poetics，on the other hand，the relationship between memory phantasms and places pertains to the figure-ground construction in Gestalt psychology，with the figure having more salient formal features than its background，such as characters versus their settings in drama. Peter Stock well，nevertheless notes that sometimes“the setting can thematically become the figure，emerging out of the background to assume a figure status in the text，”such as Egdon Heath in Thomas Hardy's The Return of the Native.Reuven Tsur also observes that in music，painting，and literature very often the figure-ground distinction either oscillates or does not exist.M.C.Escher's paintings are a case in point.In comparison，memoria，as traditional rhetoric prescribes it，abides by very mechanical figure ground relationship，which might prove to be a major limitation that has，for one，resulted from the heavy reliance on vision，and，for another，hampered the development to its full potential.Nevertheless，parallel to the linear perspective，which had dominated western art since the Renaissance until the rise of Cubism in the twentieth century，the stable and solid foundation of loci in memoria has repeatedly been challenged throughout history，especially in the form of the uncanny，or das unheimliche.

Whereas Luria attributes the rare instances of S.'s omission in recall to“defects”of perception，traditional rhetoric evidently made more positive use of shadows and concealment in memory places.One such instance is an inventional device called the topics（topoi in Greek，topica in Latin，meaning“places”）or the commonplaces（locis communis in Latin）.In Ancient Rhetorics for Modern Students，Crowley observes that the“topoi used by persons with a trained memory must be mnemonic loci”.After all，ideas are to be found where they are input and deposited.Memoria and inventio thus fold into each other.Just as Mnemosyne is the mother of the Muses，so memoria is the fountainhead of inventio；conversely，the output of inventio feeds back to memoria，forming a loop.The convention of topoi，however，reveals the obverse side of memory loci.According to Crowley

Ancient rhetorician often described the places as though they were hidden away.Quintilian，for example，defined the topics as“the secret places where arguments reside，and from which they must be drawn from”（Vx 20）.Just as hunters and fishermen need to know where to look for specific kinds of prey，rhetoricians need t be skilled at tracking down suitable proofs.

Memory loci widened in scope over time in history.Manifested in the metaphors used to describe them is a growing capacity to contain plenitude，so much so that they border on the space of the imaginary，a semiotic space through and through.It was Quintilian who began to seriously conceive of memory as a massive treasurehouse，that is，in the direction of copia，“an abundant and ready supply of language”.As an example，he names Metrodorus as a contemporary prodigy who boasted of having constructed，in his memory，“three hundred and sixty different localities in the twelve signs of the Zodiac”（Quintilian，Institutio，11.2.22）.Until the mnemonic scheme ultimately sprawled into an all-encompassing，capacious archive，the locus of memory must first and foremost be able to accommodate variety and allow for infinite extension.The question is how it is accomplished.Since the mnemoscape is essentially open-ended，the key to its design might lie in the contiguous connections between its constituents.A master mnemonist might not command a large memory，but he or she is perceptive into what Roman Jakobson calls the combinational relations in the signifying chain，and is able to be creative with them.In Institutio oratoria，Quintilian notes that each place（such as a room），real or imaginary，serves as a module that can becombined with others into a larger unit（such as a palace，a city，a journey，or a picture），divided into submodules（such as the four corners of a room），like a genus being subdivided into species.

Foreign influences emerging during the Middle Ages added complexity to the intricate design of memory sites.Carruthers and Ziolkowski observe that，while the metaphor of memory as a treasure-house remained，along with its variants，such as“waxtablet，belly，bookchest，chest，heart，or mirror，”

the domestic and familiar spaces of a Roman house，the type of architecture most commended for memory work by ancient writers，were replaced by divine structures derived from descriptions in the Bible，such as the Ark，the Tabernacle，the Temple，the Heavenly City，the map of the world，the cosmos itself.

Memoria，like the other parts of rhetoric，no longer contributed exclusively to persuasion，such as forensic oratory，but much more so to personal piety and homiletics.As if based on the same belief as Augustine，who sought to find God in his memory（Augustine，Confessions，11. 24 27），these medieval locales were rife with religious symbolism and“decidedly...more fantastical”as an abode harboring the divine.For example，in a treatise on the meditational exercise memoria spiritalis（“remembering God”），where，by way of illustration，Hugh of St.Victor teaches how to mentally construct Noah's Ark，there is an unorthodox passage devoted to compartments for amphibious animals not mentioned in“Genesis”（42 43）.Besides，there is a profusion of allegorical and moral meanings to the architectural structure beyond its literaldescription.The most typical are a series of ladders symbolizing spiritual progress.Hierarchical structures also characterize the thenpopular devotional practice of remembering Paradise and Hell that would later found artistic expression in Dante's Divine Comedy.

As classical rhetoric was revived in the Renaissance and imagines received major influence from Neoplatonism，memory loci continued to proliferate on a cosmic scale.In the same way that the Gothic cathedral，as Yates proposes，might have served as a memory locus for Aquinas to memorize his own Summa Theologiae，Renaissance memory-places were all-encompassing architectural structures that housed encyclopedic knowledge，except with increasingly more humanist and occult influences.Camillo's Memory Theatre，for example，was meant to house all archetypal Forms，as well as“the whole realm of nature and of man”. Amongst many other things，it is in human memory that divinity is manifested.In addition to expanding memory loci into structures of celestial proportions，Renaissance also introduced motion into memory sites that had previously always been static.Bruno，for instance，borrowed from Ramon Lull（c. 1232 1316）and devised revolving wheels to serve as loci on which to deploy striking animated figures.In the classical memoria，the mnemonist followed a pre-determined schematic arrangement，but in the Renaissance occult memory，the order is dictated by chance，by magic，and by mysterious forces of the unknown.The dynamics introduced to the loci，via such ingenious devices as Lull's revolving wheels，brings to memoria a sense of indeterminacy，which suggests the presence of the Other，thereby adding to its mystery and to thehermeneutic interest.While in classical artificial memory，the orator masterminded the entire recall process；in occult memory，the magus-mnemonist's path of discovery almost always has to be revealed.For a non-believer，the name of the game is variously called“uncertainty，”“suspicious，”or“chance，”a theme that will later be picked up by post-structuralists.

Order：“Like Dancers Hand in Hand”

The order of recollection has always been an integral part of memoria，though it was subsumed under the memory place as a path within it.In De Oratore，Cicero points out that one of the key functions of loci is that they“preserve the order of the facts”（2.87）.The anonymous author of Ad Herennium also asserts that it is“obligatory to have［the］backgrounds in a series”（3.17.30）. With a fixed order that provides contextual cues to facilitate retrieval，we could be easily reminded of the images stored in the backgrounds，whether going forward or backward or starting somewhere in the middle.An order is a tightly knit bundle of associations，a ritual，and a habit.In the Middle Ages，the scholastics revised the classical place rules by underscoring the order of recollection；nevertheless，according to Yates，this is due to an uncharacteristically offhand mistake of Aquinas when reading Ad Herennium.He took the word“solitude”for“solicitude，”which，connoting“cleaving with af fection，”contributes to the devotional use of mnemonics.The（mis）reading strays from the classical rules regarding memory loci；as a result，“The emphasis of the Thomistrules is on order，and this order is really the order of the argument”.What was a pre-inventio internal process turns into outward arrangement，viz.disputio，which is characterized by total visibility and accessibility.The order that was an insubstantial habit hence solidifies into law.The search of the natural law during the Age of Reason hence turns God into the greatest mnemonist that never gets lost in the labyrinth of his mindscape.An order，as a result，represents a memory trace marked by full presence and iterability.Yet this understanding of mnemonic order runs counter to that of psychoanalysis，which has amply demonstrated that the memory trace is often a treacherous path，the repeating of which is often displaced and disguised，as well as beyond conscious control. The order of recollection is a game of subterfuge.The insight derived from psychoanalysis into the mask of the Name of the Father，however，has remained virtually unexplored in the scholarship on memoria.

The Ramist method brought to the fore the issue of order in mnemotechnics.The rise of Ramism during the Renaissance represents a critical moment in the vicissitudes between rhetoric and logic.To be specific，Ramus further removes memoria from rhetoric and integrates it into logic，namely，from doxa（“belief”or“opinion”）to logos（“true knowledge”）.He proposes to restrict rhetoric to elocutio and pronuntiatio，while allotting memoria，inventio and dispositio to dialectic.This is indeed a revolutionary move as Aristotle specifically refers to rhetoric as the counterpart to dialectic in On Rhetoric（1.1354a）：dialectic appeals only to syllogisms，whereas rhetoric uses ethos and pathos，in addition tologus.Rossi observes that the three parts of dialectic（memoria，inventio and disposition）are often treated interchangeably in Ramist texts，as they all have to do with“placing or arranging the‘things found’（res inventas）in a rational order”.Advocates of the scientific method such as Bacon，Descartes and Leibniz，therefore，“saw memory as one of the primary divisions of the new logic”and regarded their respective“method”as a“classification of reality，”a sort of“universal syntax”.

Charges of logocentrism and na?ve optimism notwithstanding，Ramism and the subsequent Enlightenment discourse contain seeds of a new structural concept：intertexutality.For the Ramists，artificial memory contains both encyclopedic knowledge and the ordering of that knowledge，both the sciences and a science of the sciences，or an“art of arts”.As a model of the psyche，memoria represents an active mind—or“an able gymnast of invention，”in the words of Bolzoni—that is composed of a“matrix of a reminiscing cogitation，”constantly engaged in“shuffling and collating‘things’stored in a random access memory scheme，or set of schemes—a memory architecture and a library built up during one's lifetime with the express intention that it be used inventively”.The concept of memoria as an automated meaningmaking machine that Carruthers presents here comes close to a semiotic system.With schematic structures，visual apparatuses（such as“wheels，diagrams，and tables”）that came into widespread use following the rise of the printing press，are no sprawling labyrinths like traditional memory loci.Instead，they function like maps thatnavigate us to“find our bearings in the vast seas of logic and encylopedism”and“provide access to the rhetoric machines”of textual archives.The new art of memory，Bolzoni argues，is akin to games in that“can function only in a closed world made up of clearcut rules known to all and easily remembered”.These structures mediating between the old texts having been read and the new texts to be created，i.e.，between reading and writing，anticipate the virtual space of intertextuality that would be posited by structuralists and poststructuralists in the twentieth century.

Other scholars，however，are much more dubious about the escalating attention given to the order of recollection.Studying the development of rhetoric during the seventeenth and eighteenth centuries，Crowley argues that an introspective theory of the mind is implied in memoria influenced by the Ramist or scientific method，or what she calls，the“methodical memory”.While，in the seventeenth century，method was taken to be a universal faculty，like reason，in the eighteenth century，“method became a theory of composing，”with the author engaged，not in exploring ideas like Montaigne in his essays，but in reflecting the goal-oriented mental processes leading to the inductive or deductive conclusion.Most rhetorical theorists at this time“took for granted the mind's ability to record and faithfully repeat a process of investigation”.A well-organized argument is taken to represent a clear mind，whereas a poor organization indicates a confused thinker. For pedagogical purposes，effective structural features were later generalized into strict，formulaic rules prescribed for all in the Current-Traditional Rhetoric that dominated from the late nineteenth to the earlytwentieth century.The major problem with the methodical memory，however，is that“No account can be taken of the fact that language sometimes fails—or refuses—to represent an author's intention with accuracy or clarity”.The focus on formal features believed to represent orderly thinking，in other words，conceals an inadequate theory of mind.

Though memoria might seem to have been superseded by the printing press ages ago，arguably it survives in other forms in modern technology，such as e-mail，as Derrida suggests in Archive Fever.While constrained by the technéinvolved，memory also thrives in the virtual space it opens up，each a memory palace haunted by its shadows，and a game field that allows for infinite play.In“Semiology and Rhetoric，”citing W.B.Yeats's“Among School Children，”Paul de Man famously asks，“How can we know the dancer from the dance？”Indeed，insofar as we cannot separate the dance from the dancer，style is constitutive of thought.The ambiguity of the rhetorical question—both a question（“How can we know...？）and a declarative statement（“We cannot know...”）—is integral，not extraneous，to meaning.From this perspective，memoria could serve as the prototype of the human psyche.All natural memory is artificial memory in that memory as Ged?chtnis is always mediated，be it visual，spatial，sequential，or other semiotic representations.

Memoria Revisited：The MateriaIity of Memory

Wen-ling Su
（Fu-Jen Catholic University，Taiwan）

Memoria，the rhetorical art of memory，is commonly believed to have been superseded by the printing press.Its demise can also be attributed to Western metaphysics，which，under the influence of Plato，identifies memory as introspection.Mnemotechnics，as externalized memory，is disparaged for being bogged down by its materiality.Jacques Derrida and Paul de Man however make no distinction between interiorized and externalized memory，for both kinds of memory are subject to the mediation of semiotic systems.In fact，they suggest the possibility to view artificial memory as a prototype of the human psyche.For them，mnemotechnics is a semiotic system like writing，where meaning results from differential relations and is thus arbitrary and forever deferred.It is from this poststructuralist perspective that this paper presents a critical review of memoria.While tracing the development of memoria's three major components—mnemonic images，memory loci，and the order of recollection—this paper highlights areas where uncertainty is registered in what appears to be a seamless schema.

memoria，mnemonic images，method of loci，order of recollection，post-structuralism

蘇文伶，女，臺(tái)灣輔仁大學(xué)跨文化研究所博士生，輔仁大學(xué)英文系講師，主要研究領(lǐng)域?yàn)榉?hào)學(xué)、修辭學(xué)與記憶論述。電子郵箱：wling1＠m(xù)s23.hinet.net。