THE DESCRIPTIVE MODE:

FLAUBERT, VERGA, HUYSMANS, D’ANNUNZIO

A DISSERTATION

SUBMITTED TO THE DEPARTMENT OF

FRENCH AND ITALIAN

AND THE COMMITTEE ON GRADUATE STUDIES

OF STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

IN

FRENCH AND ITALIAN

Amy Lilah Elghoroury

December 2010

http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/yy507wq8379

© 2011 by Amy Lilah Elghoroury. All Rights Reserved.

Re-distributed by Stanford University under license with the author.

This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.

ii

http://creativecommons.org/licenses/by-nc/3.0/us/http://creativecommons.org/licenses/by-nc/3.0/us/http://purl.stanford.edu/yy507wq8379

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Johannes Gumbrecht, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Franco Moretti, Co-Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Laura Wittman

Approved for the Stanford University Committee on Graduate Studies.

Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.

iii

iv

Abstract

Looking at novels by Flaubert, Verga, Huysmans, and D’Annunzio, this project

charts the increasing role of subjectivity in novelistic descriptions from mid- to late-

nineteenth century France and Italy. This extended analysis of literary techniques that

express subjectivity reveals that description was increasingly seen as a field of

experimentation in which the parameters of an individual’s access to the observable

world could be adjusted to the concerns of a particular novel.

Earlier approaches to description in the nineteenth century had used the

proliferation of individual details as a way of ensuring fidelity to the observable world.

An unobtrusive describing voice was considered a mark of the objectivity of a

description. Adaptations of this concept to new types of projects – projects in which the

distinctive perspective of figures within the world of the novel are key to the novel’s plot

– produced different results, however, in the second half of the century: the issue became

not only whether or not one could describe sufficiently and in as objective a way as

possible, but how a description could adapt to accommodate observational subjectivity

while maintaining the referential basis of the descriptive mode. In some cases the use of

techniques that imply the presence of a filtering subjectivity leads to what seems like a

greater impersonality, as the observing descriptor is redistributed and reabsorbed into the

reality of the novel’s characters. In contrast, in others the overabundance of subjectivity

techniques and the influence they have on sentence and paragraph structure overwhelms

the composition and the novel moves into relativism and even idiolect. This dissertation

traces a progression from the former tendency to the latter over the course of the second

half of the nineteenth century in France and Italy, thus showing a continuity between

v

‘realist’ tropes of observation integrated in description and the ‘decadent’ style that

appropriated those tropes in the interest of individual expression.

This project stems from the premise that description is not simply a type of text,

easily isolated from other types; rather, passages, sentences, and phrases considered

descriptive are operating in a mode that prioritizes the durative qualities of objects in the

world of the novel. Descriptions require a pause in novelistic time, and detecting

descriptive pause is the first task in analyzing how the descriptive text functions both

internally and in relation to the narrative whole. Using durative verb tenses as markers of

descriptive pause, descriptive passages were isolated and examined for formal features

that were either repeated throughout the novel or received special emphasis. Some

features appear to be necessary to all approaches to description – the creation of lists, for

example, is a practice common to all four novels – while others are particular to

descriptions of specific kinds of objects or novelistic contexts. The advantage of this

approach is that it looks at trends in descriptive practice as groups of stylistic choices that

enjoyed certain degrees of preference over time, sometimes emerging with greater

emphasis or frequency at certain kinds of moments in a novel (moments of observation,

moments of introspection, moments of adjustment in a character’s expectations, etc.) and,

on a larger scale, at certain times and places in nineteenth century literary history.

The configuration of authors and novels presented here is therefore based on

stylistic tendencies that overlapped across historical and national divisions and not on

periodizations that would isolate these works by other criteria. The use of descriptive

technique to infuse a novelistic world with the subjective observational standpoint of an

integrated observer is shown to have increased in frequency and importance from the

vi

1860s to the turn of the century, but this should not imply causal links between the

successive moments in literary history represented by these novels. Each novel adapts

the possibilities of descriptive language to its immediate aesthetic goals. This

dissertation explains descriptive prose from microscopic observations of phrases,

sentences, and paragraphs in an effort to decode, from individual choices, the workings of

descriptions in themselves and only after that as part of larger narrative wholes.

The first chapter looks at descriptions in Flaubert’s Salammbô as examples of

integration of an observing entity with the world the novel describes through perspectivist

techniques. With the imperfect tense as support for vast descriptive panoramas,

Flaubert’s prose disperses the points of view from which descriptions of ancient Carthage

and its environs emanate. The result is an encyclopedic project that maintains a high

level of documentary detail while accounting for the experience of an observer who

shares a perspectival context with the novel’s characters but is simultaneously not

identified with any single figure in the novel.

The second chapter detects the specific traits of Verga’s use of a choral voice in

Mastro-don Gesualdo. By combining the phrase structure of dialect speech with the

vocabulary of literary Italian Verga’s project achieves an unusual synthesis of describing

voice and described world. The brevity, directness, and expressivity of Verga’s

descriptions collapses the pretense of impersonal objectivity; in particular, Verga’s use of

idiomatic expressions and emotive punctuation in descriptions infuses his describing

voice with the shared emotional, social and linguistic contexts of Gesualdo’s Sicilian

village.

vii

The third chapter looks at Huysmans’s À rebours as an example of a manipulation

of descriptive techniques in the service of the perceptions of a single protagonist. In

rejecting the social and turning Zola’s precision toward an individual mind Huysmans

uses description as a support for interpretive essays on topics that align with the tastes of

his protagonist. Huysmans’s descriptive sentences show a tendency toward superfluous

transformations of syntax and excessive elaboration of the qualities of perceived objects

at the expense of phrasal unity and narrative context.

In the fourth chapter I turn to D’Annunzio’s Il fuoco, a novel in which descriptive

prose becomes poetic prose. D’Annunzio’s poet-describer projects his vision of a

symbolically-charged Venice onto descriptions of the city in a way that integrates the

poet’s feeling of power over his own environment with the descriptive form. In its use of

extended and complex metaphors, emphatic repetition, and sonorous word patterns

D’Annunzio’s descriptive style is a distinct movement away from universal observation

of phenomena toward a solipsistic, individualized use of description that ignores the

epistemological questions that form the basis of literary practices focused on observable

reality.

While Flaubert and Verga provide evidence of descriptive technique being used to

afford greater objectivity through omnipresence of the describing mind, Huysmans and

D’Annunzio prove to be less concerned with the relationship of the individual to the

world the novel describes. The result is a refocusing of the interests of descriptive novels

from the totality of the lived experience shared by inhabitants of a given world to the

interests and perspective of a powerful and alienated individual figure.

viii

For my parents

With special thanks to Sepp, Franco, and Laura

And with thanks to all my advisers, official and unofficial, past and present!...

ix

Table of Contents

Abstract: iv

Introduction: 1

Chapter one: Flaubert 11

Chapter two: Verga 58

Chapter three: Huysmans 108

Chapter four: D’Annunzio 153

Bibliography: 207

1

Introduction

Thinking back on the history of the novel in the nineteenth century, we find it

impossible not to think of long and detailed descriptions: those passages in which human

environments are described in intimate detail as if the entire structure of a novelistic

world rested on the placement of a single object in a scene. In certain novels the

descriptive mode is so overpowering that the novel’s plot seems to fade into the

background, reversing the typical roles of narration and description in a foreground and

background configuration. These are the descriptive novels of the nineteenth century,

and they include not only those novels in which description is quantitatively dominant,

but also those in which description has a disproportionate influence on the relationship of

a reader to the world the novel creates. The nineteenth century saw a shift in the use of

descriptive style from a way of affording seemingly unfiltered access to an objectively

observed reality to a form through which a particular point of view could shape a reader’s

vision of what is described. My argument addresses the nature of this shift and the

function of description in the second half of the nineteenth century, answering the

question of how descriptive prose establishes the parameters of individual access to the

described world in novels that prefer description as primary modality of that access. An

analysis of such novels starting from inside descriptions themselves – from the level of

syntax, lexicon, and punctuation outward to descriptive passages – presents description as

not just the background to narrative’s foreground, but as a stylistically-modulated way of

understanding a novelistic world, a distinct mode of prose usage that determines the

amount and kind of subjective filtration that will inform the statements of a describing

voice.

2

The novels I have chosen to group together here share little in the way of plot, and

I do not mean to imply that their authors shared some motivation for the tendency toward

individual perspectivism in their descriptive styles. Nonetheless the notion of description

as having a certain relationship to our reading of the rest of the novel, of description as an

informational screen that filters our access to the world of that novel is central to the

function of descriptive prose in each.

The descriptive passages I chose for analysis here were selected either because

they contain formal features of description that were noticeably repeated throughout a

novel, or because they receive some kind of special emphasis in the text as a whole (I

have noted these cases). Using durative verb tenses as indicators of descriptive pause, I

isolated he entirety of descriptive material from each novel. I annotated the entirety of

descriptive passages in the novel as a group and marked the elements of the descriptions

that were repeated or received said special emphasis in the text. I was then able to use

frequency of verb tenses, similarities in sentence structure, repeated punctuation marks or

words, and other criteria to create groups of descriptive sentences with shared

characteristics. Those that appeared most frequently or with the most emphasis became

the material of my analysis. The advantage of this approach is that it looks at trends in

descriptive practice as groups of stylistic choices that enjoyed certain degrees of

preference over time, sometimes emerging with greater emphasis or frequency at certain

kinds of moments in a novel (moments of observation, moments of introspection,

moments of adjustment in a character’s expectations, etc.) and, on a more expansive

scale, at certain times and places in nineteenth century literary history.

3

Reading for descriptions

Narration and description, foreground and background. Description is to a

novel’s plot as is the landscape of a painting to the figures whose represented actions are

pushed to the front of the perspectival system; this is Weinrich’s formulation, and it

accounts for most uses of the descriptive in the history of the novel. The spatial

metaphor rests on an assumed prioritization of narrative elements over the descriptive

that is quite natural: narrative, with its focus on the sequential presentation of events

whose organization structures the order of our reading, usually functions to motivate

interest in continued reading. We absorb one narrated event and anticipate the next. In

this system description, an inherently decelerating textual mode, interrupts the flow of

successive points of interest in a plot-driven timeline. It is a structure that decelerates the

time of the novel to a near standstill in the interest of explanation and contextualization.

This is the basic function of description in an otherwise narrative whole. Until recently it

was taken for granted that this system of levels translated into a hierarchy of modes that

implied that what is narrated should, as foreground material, also be prioritized in

reading. I would like to suggest that in certain contexts this is inadequate; one can, and

actually needs to, sometimes, read for descriptions. A case in point is the period to be

examined here, a time during which we find novels packed, if not overwhelmed with

descriptions. While these novels do have plots, it would be off the mark to assume that

descriptions always function in the service of a plot in the light of the simple observation

that plot is not the dominant element in certain nineteenth century novels, the descriptive

novels. Indeed some novels seem to work in precisely the opposite way; a minimal plot

4

is present as a framing structure that accommodates long and detailed descriptions that

function autonomously and sometimes have little connection with narrated events.

This is certainly true of the first novel I have chosen to investigate here;

Flaubert’s Salammbô begins with an entire chapter devoted to long and immensely

detailed description of the appearances and behaviors of the diverse group of mercenary

soldiers feasting outside Carthage without a clear indication of the significance of this

information to the plot of the novel. And as it turns out, this information really is

unnecessary to the progression of historical events that the novel relates; the multiplicity

of types in the mercenary army does not, in this novel, have any causal ties to the

outcome of any of the military events that determine the plot, of which most readers

already know the outcome. This first chapter of Salammbô, however, is by far the most

memorable section of the book. It was in researching the archeological details that

inform the descriptions that Flaubert spent the most time in preparation for writing the

novel. In the light of these observations and others I have concluded that, clearly,

descriptions were not, in the context of late-nineteenth century expectations, mere

supplements to dominant narrative elements. The fact that novels like Salammbô exist

suggests that reading for descriptions was in fact a historical reality, and given the

complexity, number, and length of descriptive passages as they appear in texts from this

period, an aesthetic necessity in the period between 1862 and 1900 and probably before

that.

Although Salammbô is one example of a novel in which description is

quantitatively dominant, this project does not look only to novels in which descriptions

take up more textual space than narrative elements. Rather it is that kind of novel in

5

which description is the focus of compositional energy, in which the work of configuring

and actualizing a novelistic universe is done in the manipulation of descriptive form. In

such novels description is the dominant element in a complex system of textual

modalities that influence a reader’s access to a fictional world and determine the

parameters of that access.

Descriptive perspectivism as an adaptive move

Descriptions in novels took seriously the subjectivity of the describing observer in

the nineteenth century, more so around and after the century’s halfway point. Attention

to description was already present, and heavily so, in Scott’s novels and in Balzac of

course; for them, descriptive detail was a way of foregrounding the material of the

description over the subjective filter of the voice of the author. Detailed descriptions in

novels written in an attitude of realism in the early nineteenth century position groups of

objects and facts in descriptive compositions as if their organization were both necessary

and unchangeable. Descriptions after 1848 use the same basic structure and placement

of descriptive passages, but without the certainty of alignment with a single, extradiagetic

and mostly invisible describing figure. This development aligns with the emergence of

the free indirect style as a major feature in some novels, and with similar effects; what is

achieved, or at least attempted, is some kind of accommodation of individual

observational experience of in the stylistic features of the descriptions themselves.

Whose experience? Usually the answer to this is not any single character or identifiable

observer, but a mixed voice that shares something with the characters but has access to

6

more information than any single character possibly could. The urge to describe in detail

with an appearance of objectivity as it manifests in novels of the early century might have

been a compensatory action against an underlying anxiety regarding the compatibility of

systematic modes of recording knowledge with the everyday reality of individual

perspectivism, an attempt to use precision as a defense against epistemological

uncertainty. The presence of this urge to describe in detail in the early century provided a

context in which mutations of style to reflect individual subjectivity could be

experimented with later on. The issue became not only whether or not one could describe

sufficiently and in as objective a way as possible, but how a description could adapt to

accommodate observational subjectivity while maintaining the referential basis of the

descriptive mode. In some cases the use of techniques that imply the presence of a

filtering subjectivity leads to what seems like a greater impersonality – the observing

descriptor is redistributed and reabsorbed into the reality of the novel’s characters, while

in others the overabundance of subjectivity techniques and the influence they have on

sentence and paragraph structure overwhelms the composition and the novel moves into

relativism and even idiolect. I would suggest that Salammbô and Mastro-don Gesualdo

tend more toward the former, while À rebours and Il fuoco tend more toward the latter.

The use of description in novelistic practice: four case studies

Descriptive prose has been used to present non-narrative elements of literature for

centuries; what was attractive about the sequence of novels I selected here is the intensity

of focus that seems to have been put on descriptive style during this period in these two

7

national literatures. Purposeful manipulation of observational structures seems to have

begun with Flaubert; one thinks immediately of the free indirect style. This, and other

perspectival techniques made Flaubert’s work the logical starting place. Neither Madame

Bovary nor L’éducation sentimentale used description as vigorously as Salammbô, and

Salammbô features the added complication of historical distance from the described

world; Flaubert’s game of omnipresence in the world of his characters took extra

imagination in a representation of a foreign, antique, and relatively unknown subject.

And the techniques are all there – the free indirect style in passages that connect

intimately with the mind of the Carthaginian princess, perspectival shifts akin to

movement-based descriptions of Madame Arnoux in L’éducation sentimentale, great and

serious encyclopedic descriptive treatises on arcane subjects written with the same rigor

as those of Bouvard et Pécuchet. My first chapter looks at Salammbô as a hybrid

historical-realist novel and examines the way in which Flaubert adapted the detail-

oriented style of description used by a previous generation of writers to a more

perspectivist approach to the novel. For example, Flaubert’s relative overuse of the

imperfect in descriptive and passages that combine description and narration neutralizes

narrative time and enables the creation of vast descriptive panoramas so detailed that they

preclude the possibility of creating an effet de réel. Further, Flaubert uses tropes of

observation to introduce descriptions in a way that makes the presence of an observing

subjectivity impossible to extract from the described material. The result is a novel so

committed to representing the total, lived reality of its subject that its plot is

deemphasized while the figure of observer is dispersed over varying levels of integration

with the world of ancient Carthage.

8

My second chapter treats descriptive voicing techniques in Giovanni Verga’s

Mastro-don Gesualdo. Giacomo Devoto and Leo Spitzer suggested that Verga used a

“choral” voice in his prose, a voice that represents the shared mentality of the rural

communities of Southern Italy that are the focus of his novels. In the chapter I try to

detect exactly which formal elements create the sense of a choral voice in Verga’s

descriptions. In Mastro-don Gesualdo as in I Malavoglia Verga attempts to fuse the

linguistic reality of his subject with the indirect non-narrative communications

conventionally rendered in a distanced, literary voice. In character descriptions the

omission of certain information and a distinct informality of address shocks the reader

into intimate closeness with the world of the Sicilian village. A general unconventional

use of idiomatic expressions familiar to oral communication throughout descriptive

passages further deformalizes descriptions and mixes the linguistic tendencies of the

describing voice with those of the novel’s characters. The impression of an involved

observing subjectivity is further created by Verga’s use of unusually expressive

punctuation in descriptions, an importation of graphical markers typically associated with

direct discourse. Verga uses these formal innovations as a way of incorporating the

received ideas of an entire village in the descriptive style. As a move further into

perspectivist description Verga’s achievement can be seen as a transformation of

Flaubert’s intellectual experiment to a national context in which the choice to identify

linguistically with regional dialect speakers was as much a political statement as it was an

aesthetic challenge.

J.-K. Huysmans’s À rebours, of these four the novel probably most readily

associated with excessive descriptive stylization, is the subject of my third chapter. In

9

moving back to France with À rebours I take on the mutation of detailed descriptive style

as it manifests in the thematic realm of decadence. In chapters whose relationship to the

novel’s plot is barely maintained in minimal narrative passages À rebours uses the

temporal scaffolding of the novel form as a support for essayistic descriptions of objects

collected by the novel’s protagonist. Particular rhetorical turns make it possible for the

non-narrative to completely overwhelm the action of the novel: Huysmans’s use of

ornamental elements whose presence is unnecessary to the grammatical functioning of

his sentences, including repetition, manipulation of word order, and an imbalanced use of

verb forms that imply completed action in the past saturate his prose with modifying

language that redirects attention from the described objects to their attributes in the

narrative present.

D’Annunzio, at the end of this sequence, pushes the descriptive outside its

conventional rhetorical boundaries by discarding the notion of an externally positioned

observer and using description as a ground for unhindered individually filtered

projections on the world and poetic expressions quite removed from the possibility of

universal observation. Il fuoco is a novel in which descriptions take on expressive

functions that had hitherto not been activated in descriptive practice, including prominent

use of metaphors and other forms of comparison between the matter of his descriptions

and their idiosyncratic symbolism and systematic repetition of phrase elements with

musical and emphatic effects. D’Annunzio also chooses to employ modal shifts in

descriptive phrases, an innovation that posits an alternate symbolic reality attached

hypothetically to the primary level of observable present of the novel. The result is a

novel in which description begins to look like poetry. The degree to which

10

D’Annunzio’s descriptions express the projections of an individual subjectivity is a far

cry from the distanced and self-effacing style of descriptive prose that had dominated

novelistic practice less than a century before.

The ghost who haunts this configuration is Zola, who did of course sometimes use

a stylized descriptive form and poetic language to magnify the allegorical message of his

images – think of the emblematic description that introduces the death scene of Nana, an

exaggerated historical irony – but doesn’t seem to have actively pursued the perspectival

experimentation I have found in each of the authors’ work I present here.

Whether or not the use of descriptive technique in these novels follows some kind

of progression is unquestionable; certainly, practices changed over time. I do not mean to

suggest, however, that Verghian voicing is a mere transmutation of something that

happened first in France, or that all of this would have been impossible without Flaubert;

there are too many variables, and too many individual innovations, to imagine anything

like it. Each novel adapts the possibilities of descriptive language to its immediate

aesthetic goals. I have attempted to address these in each chapter. Because of the

immense complication of descriptive techniques, some of which overlapped while others

were completely idiosyncratic, I had to approach each novel from the most microscopic

observations, looking outward from individual phrases, sentences, and paragraphs. I

have tried to explain these examples of descriptive prose from inside in an effort to

decode, from individual choices, the workings of descriptions in themselves and only

after that as part of larger narrative wholes.

11

CHAPTER ONE

Descriptive composition and observational positioning: Flaubert’s Salammbô

12

Descriptive composition and observational positioning: Flaubert’s Salammbô

This thesis claims that certain novels written in Europe in the second half of the

nineteenth century used descriptive prose as a field of experimentation in which the

parameters of ‘objective’ representations could be tested and modified. During this

period writers of descriptions in novels had concerns beyond those of detail-oriented

‘realist’ literature of the early century and looked critically at the conditions under which

realistic representations could be made in the context of individual perspectivism. The

goal of this project is to examine actual systems of descriptive practice and to identify the

textual features of perspectivist description as I have briefly introduced it here.

This chapter will focus on a moment in which a shift occurred in novelistic

practice away from descriptions based on the assumption of objective and universally

available observation. Novels written in an attitude of realism toward their subjects in

the first half of the nineteenth century used precision and quantity of details in

descriptions as a way of creating an illusion of objectivity to be carried through the entire

reading experience.1 Such an approach creates descriptions replete with proper nouns

and exacting modifying clauses that overwhelm phrase segments that could be associated

with an observing entity. For the authors commonly associated with the realist paradigm

in France, detailed descriptions added to the sense documentarian objectivity and

authorial detachment. With Flaubert, however, we see the emergence of a new ambition

to not only describe in detail but also to somehow use descriptions to accommodate the

way in which an integrated observer would see the subject from within. The medium of

this expression that accounts for the thoughts of someone within the existential context of

the described world had to be literary style; for as part of the style these expressions 1 See Moretti, The Novel 386-88.

13

could be melded with the structure of the text on the phrase level and would not call

attention to themselves in the composition. This is the source of moral ambiguity in

Madame Bovary, as identified by the prosecutor Pinard: the thoughts of his protagonist

are so well enmeshed with the narrating and describing voice that it was impossible to tell

if that voice condemned Emma for those thoughts. Salammbô was Flaubert’s second

attempt, this time in a context that could leave no question as to the moral opinion of its

author: on 11 February, 1857, Flaubert wrote to Fréderic Baudry: “Quoi écrire qui soit

moins inoffensif que [Bovary]? On s'est révolté d'une peinture impartiale. Que faire?

Biaiser, blaguer? Non! Non! Mille fois non!” (Correspondance 681).2 Flaubert saw the

Carthaginian project, a research-based historical account of a geographically and

historically distanced subject, as a foil to accusations of moral ambivalence following the

publication of Bovary. Archeological detail was to serve as a shield of implied

objectivity that would protect its author’s opinion from being detected to an even greater

extent than had been achieved with descriptions in Bovary, for with Salammbô there

could be no doubt whatsoever that the thoughts and observations of his characters shared

absolutely nothing with the views of its author.

Salammbô is as a result a rigorously descriptive novel, even overwhelmingly so,

and as such an excellent ground for analysis of Flaubert’s descriptive style as a mutation

of objectivity-oriented descriptive practice in nineteenth century novel history.

Descriptions appear with greater frequency, complexity, and relative influence over other

2 See also the letter Madamoiselle Leroyer de Chantepie, Paris 18 March 1857: “Mais je suis bien empêché pour le moment, car je m'occupe, avant de m'en retourner à la campagne, d'un travail archéologique sur une des époques les plus inconnues de l'antiquité, travail qui est la préparation d'un autre. Je vais écrire un roman dont l'action se passera trois siècles avant Jésus-Christ, car j'éprouve le besoin de sortir du monde moderne, où ma plume s'est trop trempée et qui d'ailleurs me fatigue autant à reproduire qu'il me dégoûte à voir.” Flaubert, Gustave, and Jean Bruneau. Correspondance. II, juillet 1851-décembre 1858. Paris: Gallimard, 1980, 691.

14

elements in Salammbô than in any other of Flaubert’s novels and most novels of the same

genre written before it. Salammbô also stands at a peculiar point in nineteenth century

literary history after which descriptions would be irrevocably charged with compositional

attention; it seems that, though novels were certainly very descriptive before Salammbô,

never before had descriptions received such dedicated attention as the main focus of

compositional energy in a novel. This novel marks a turning point after which attention

and complexity in descriptions would continue to increase until the turn of the century.

This feature of Salammbô has been the main source of its criticism since the time

of its publication. The novel is too descriptive; in reading time alone we detect some

imbalance between descriptive passages and expectations for narrative movement.3

Critiques of Salammbô have consistently noted the way in which, both quantitatively and

in terms of complexity, the descriptive overwhelms the novel: Sainte-Beuve devoted an

entire section of his already impressive three-article treatment of Salammbô in the

Nouveaux Lundis to descriptions and to style.4 Sainte-Beuve’s generally negative review

of Flaubert’s use of descriptions can be summarized by three main points: 1. The

descriptions are too complete; more is described that could possibly be perceived by an

individual (“Je ne m’accoutumerai jamais à ce procédé pittoresque qui consiste à décrire

à satiété, et avec une saillie partout égale, ce qu’on ne voit pas, ce qu’on ne peut

raisonnablement remarquer,” (88)); 2. Flaubert prefers to describe only the worst

qualities of the fictional world, to exaggeration (89); 3. The style is too closely managed, 3 This aspect of the novel, paired with its exotic subject, has led Flaubert criticism to separate this novel, along with La tentation de Sainte Antoine, from the rest of his work in a biographical binary. Porter, Laurence M. Critical essays on Gustave Flaubert. Boston: G.K. Hall, 1986, 3. In addition, Lukács, on motivation for Salammbô: “It was precisely because of his deep hatred for modern society that he sought, passionately and paradoxically, a world which would in no way resemble it, which would have no connection with it, direct or indirect.” Lukács, György. The historical novel. Lincoln: University of Nebraska Press, 1983, 185. 4 Part three of the third article on Salammbô is titled “Des descriptions et du style,” (88-93).

15

rendering it homogeneous and forced (91). Most interestingly, the critic also protests the

historical and cultural distance between the subject matter and the world of its readership;

as traversal of this distance is further impeded by the superfluity and impossible precision

of details in the descriptions, we can have no communication with the inner lives of the

characters, no “sympathie.” All of these statements add up to Sainte-Beuve’s general

claim that the work is essentially “invraisembable:” a striking assessment insofar as this it

further underlines the disjunction between notions of reality (which, Sainte-Beuve notes,

is in abundance in Salammbô), and the “truth” to which the author would have been

preferred to pursue.

In the same vein, in The Historical Novel Lukács sees Salammbô as a final signal

of the decline of the genre: “historical figures are separated from the real driving forces

of their epoch, and their deeds, thus rendered incomprehensible, acquire a decorative

magnificence by virtue of their very incomprehensibility,” (179). Lukács finds this

“decorative” element to be Salammbô’s main weakness, at least in terms of the historical

trajectory of his critique: “In describing the individual objects of an historical milieu

Flaubert is much more exact and plastic than any other writer before him. But these

objects have nothing to do with the inner life of the characters…the effect of this lack of

connection is to degrade the archaeological exactness of the outer world: it becomes a

world of historically exact costumes and decorations, no more than a pictorial frame

within which a purely modern story is unfolded.”5 Lukács’ criticism does not, like so

many criticisms of the work since, get out of the persistent mimesis question central to

5 For Lukács, a great historical novel in the pre-1848 mode relies on a fusing of individual psychology with a sense of historical particularity. Because Salammbô’s descriptions are “unconnected” with the inner lives of the Carthaginian characters Flaubert’s novel fails the Lukácian test for the truly historically specific. Lukács, György. The Historical Novel. Lincoln: University of Nebraska Press, 1983, 194.

16

many assessments of texts written in a realist paradigm.6 Like Sainte-Beuve before him,

Lukács establishes criteria for a kind of mimetic art – the historical novel in this case –

and then shows that this novel does not meet the requirements. This reading does not

address the non-mimetic operations at play in the Flaubertian descriptive field, nor does it

consider the particular goals of the novelist in taking on a historical subject so incredibly

remote as to be removed from all contemporary cultural reference. Flaubert did not fail

to write a historical novel; he didn’t try to write within the parameters of the historical

such as it had been done up to that point. Unlike most historical novels, Salammbô is not

plot-driven; characterization of historical figures is deprioritized; there is, and it seems

quite purposefully, no possible direct allegorical reading of the novel that would relate it

to contemporary politico-historical events.7 Indeed it is the inverse: Salammbô’s

historical plot is subordinate to the descriptions – not only because descriptions make up

the bulk of the text, but also because the immobility of the historical plot lends itself to

this secondary role – and in turn the motivating force of the novel is located in described

elements (the banquet, the city of Carthage, Salammbô’s costume, the temple at the city

center) instead of in the plot. Description is, in fact, where the action is in Salammbô.

This distinction – that Salammbô was not indeed written to conform to the

generic expectations of the historical novel, but rather uses the narrative structure of its

subject as support for a rich descriptive experience – frees an assessment of this novel

from expectations built on genre parameters. As a descriptive-dominant novel, then,

6 In this I agree with Eugenio Donato. “Flaubert and the Question of History: The Orient,” in The Script of Decadence: essays on the fictions of Flaubert and the poetics of Romanticism. New York: Oxford University Press, 1993, 35-55. 7 Nonetheless this is the thesis of a recent volume: Dürr, Volker. Flaubert's Salammbô: the ancient Orient as a political allegory of nineteenth-century France. Currents in comparative Romance languages and literatures, vol. 107. New York: P. Lang, 2001.

17

Salammbô is a turning point in nineteenth century literary history after which narrative

need not be considered the main focal point of both composition and reading.

The status of descriptions in descriptive novels

Description had increasingly become, around the time of Flaubert’s project, a

focus of compositional and reading attention in novels that approached their subjects with

an affinity for realism and an illusion of objectivity. What is different after the century’s

midpoint is the approach to this illusion descriptive authors took in conceptualizing

descriptive objectivity. Instead of relying only on the effacement of a describing

observer in an overabundance of details in order to make the text appear unfiltered,

Flaubert chose to use details – of a greater quantity, and of a different sort – to open up

the described worlds of his novels to a reader’s perception as if from inside. The idea

that inclusion of details as they would have been seen by an observer within the described

scene would lead one to a greater illusion of objectivity – objectivity through total

description, seeing everything from everywhere – is a mid- to late- century mutation of

the earlier realist impulse to describe. We may see Flaubert therefore as sharing

somewhat in the goals and assumptions of detail-oriented realistic writing of the early

nineteenth century. But the extension of the describing observer into perspectives other

than that of an extra-diegetic, epistemologically differentiated describing entity is a

development that distinguishes Flaubert’s ‘late’ brand of realistic descriptive practice

from that of his predecessors.

18

The descriptive can be seen in this instance as a quite adaptive mode, the form of

which could be adjusted to particular aesthetic goals in particular contexts. The notion of

description itself is in practice relatively flexible in adapting to the conventions and

attitudes of the intellectual and historical milieu with which it communicates at a given

moment in its history, perhaps even more so than narration: while the basic signifiers of

narrative progression are relatively stable (connotation of change through use of

grammatical and syntactical signals of change, i.e. verbs, as a bare minimum),8 the

qualities and signs of the descriptive are relatively changeable. The flexibility of the

descriptive is evinced by the fact that, although the evaluative claims of critical

commentaries have sometimes raised and sometimes lowered the descriptive in relative

merit to narrative prose, criticism rarely sees a difficulty in categorically differentiating

description from narration until at least the nineteenth century.9 Nonetheless, as Phillippe

Hamon notes in a study of the descriptive as a prose genre, individual readers, despite

being able to recognize a description in sharp contrast the story it interrupts, find it

difficult to define the descriptive in terms of its functional characteristics.10 It is clear

that descriptive communications in actual function do not conform to strictly

morphological criteria; if we restricted description to a definition determined only in that

way, we would reduce the descriptive to merely nouns, adjectives, and deictic pronouns

or demonstrative adjectives, as only these grammatical units (when disassociated from

8 According to Claude Bremond’s minimal conditions under which a message can be called narrative: “Que par ce message, un sujet quel-conque (animé ou inanimé, il n’importe) soit placé dans un temps t, puis t + n et qu’il soit dit ce qu’il advient à l’instant t + n des prédicats qui le caractérisaient à l’instant t.” Logique du récit. Paris: Éditions du Seuil, 1973, 100. 9 For a history of critical variations on the notion of the descriptive, see Phillippe Hamon and Patricia Baudoin, “Rhetorical Status of the Descriptive,” Yale French Studies 61, (1981): 1-26. 10 Phillippe Hamon, “Qu’est-ce qu’une description?” Poétique, 12, (1972): 465-85. In this essay Hamon suggests that the minimal characteristics of a communication that can be termed descriptive are: that the textual unit in question be continuous or discontinuous, relatively autonomous in its expansion, more often than not referential, and interchangeable with a noun or deictic pronoun such as this, him, it, etc. (1-3).

19

their syntactic value) really achieve “pure” description in the ontological sense. Rather

description should be seen as a mode of communication in the process of which a certain

genre of experience is generated.11 But if the term “descriptive” is not used to designate

a kind of text but rather a mode or style of textual experience, which operations should be

seen as the recurring structural motifs of the descriptive mode? Unlike narration, which

in its inherent diachronicity is congruent to the temporal structure of lived experience,

description requires an extra semanticizing step; or as Gumbrecht puts it: “narration

makes experience accessible in the context of its polythetic constitution, whereas

description makes possible a monothetic comprehension of experience as results of

processes of experiential formation.”12 Description is therefore always already an

interpretive act. This leads naturally to the idea that insofar as reading a description

presupposes this nonconformity to ‘real’ experience through the temporal discontinuity

that the descriptive mode requires, no description is actually mimetic; i.e. there can be no

true realistic description, since description is not a ‘natural’ or ‘real’ mode of relation to

the observable world. The difficulty of description is therefore to somehow generate,

firstly, temporal conditions in which successively evoked objects can be interpreted to

coexist in the same fictional time-space, and secondly, to configure this lapse in the

mimetic process of literary communication in such a way that it has a reflective

relationship to the rest of the narrated world.

11 In this I echo Hans Ulrich Gumbrecht’s postulation: “The complicated circumstances of the varying quality of sense structure resulting from the reception of narrative or descriptive texts can be more concisely grasped when we no longer conceive of narration, description, and argumentation as metahistorical types of discourse but as experiential styles provided by the life world.” Gumbrecht, “The Role of Narration in Narrative Genres.” In Making Sense in Life and Literature. Minneapolis: University of Minnesota Press, 1992, 46. 12 Ibid.

20

It is perhaps this very difficulty – inherent to description and enhanced by the

discreet problems of realistic representation – that attracted writers of the ‘late’ realistic

phase of the European novel to the descriptive genre. In the descriptive, Flaubert and his

generation saw a prime location for testing the parameters of a realist approach to which

they had developed a kind of aesthetically-manifested disdain.13 The inefficacy of an

unexamined and merely thematic approach to writing in a realist mode, paired with

disenchantment with an ethically-charged culture of social responsibility rubbed out by

the fizzling (at least in France) of reformist imperatives of the 1848 revolutions and

subsequent ascendency of the Second Empire prompted a modification both of literary

culture and form. This “aesthetic revolution,” writes Pierre Bourdieu, is inseparable from

the invention of a new lifestyle and social personality – of which Flaubert, artist and

scholar, detached and disdainful of the very bourgeois social conditions from which he

hailed and which made his lifestyle possible – was the ultimate example.14 Typified by

moral impassivity and distance from the dominant culture as well as the prose genres that

culture produced, the “aesthetic aristocratism” of the late nineteenth century (a cultural

strain that persists from late realism to decadence, symbolism, and on to modernism)

could be seen as, in effect, a mutation of the artistic culture and methodology of the

earlier, idealistic social realist period. In this way ‘late’ realism was the enactment of

13 See for example Flaubert’s 1856 letter to Léon Laurent-Pichat in which he responds to his correspondent’s criticism of the “vulgar realism” of Madame Bovary: “Croyez-vous donc que cette ignoble réalité, dont la reproduction vous dégoûte, ne me fasse tout autant qu'à vous sauter le coeur? Si vous me connaissiez davantage, vous sauriez que j'ai la vie ordinaire en exécration. Je m'en suis toujours personnellement écarté autant que j'ai pu. Mais esthétiquement, j'ai voulu, cette fois, et rien que cette fois, la pratiquer à fond. Aussi, ai-je pris la chose d'une manière héroïque, j'entends minutieuse, en acceptant tout, en disant tout, en peignant tout, expression ambitieuse.” October 2, 1856. Flaubert, Gustave, and Jean Bruneau. Correspondance. II, juillet 1851-décembre 1858. Bibliothèque de la Pléiade, 284. Paris: Gallimard, 1980, 635. 14 Bourdieu, Pierre. The Rules of Art: Genesis and Structure of the Literary Field. Trans. Susan Emanuel. Stanford, Stanford University Press, 1996, 107-112.

21

aesthetic experiment – not totally dissimilar to the “experimental” framework conjured

by Zola in the naturalist re-assertion of the legitimacy of the realist project with

naturalism – and description was the prose genre par excellence in which this experiment

could be taken to its limits.15 It is striking that a taste for disinterestedness in art should

be aligned, at least partially, with the themes and techniques of a realism that concerned

itself largely with social reality; although realist thematics would experience a sudden

collapse in the naturalist school with the singlehanded invention of “decadence” (in the

novelistic form at least) by Joris Karl Huysmans in the publication of À rebours, until

1884 the social and largely contemporary world (with the glaring exception of

Salammbô) would be the principle material with which both realist schools would

engage.

In the light of these observations, Flaubert’s 1862 novel Salammbô is a

particularly interesting point of access to the problem of the descriptive in the latter part

of the nineteenth century as it represents the efforts of a novelist to write out, in an

attitude of realism, the material and experiential universe of a people completely separate

from his contemporary situation – the only realist novel up to that point to have been

written about an reality that could not be observed.16 Salammbô, in treating the world

15 At least in the novel. Baudelaire shared in the artistic culture of the late nineteenth century that I am referring to here, but whereas in the novel the extreme formalism that absorbed the energy of the realist impulse was not (yet) completely detached from the moral imperatives of early realist thematicization of the everyday world (or at least didn’t yet find it quite necessary), in poetry the two moves – of condensation of formal precision and of detachment from mundane references – coincided in Baudelaire’s vertical compaction of form and symbol in the concept of “correspondances.” It should also be noted that by 1878 (two years before his death) Flaubert was equally disdainful of the “experimental” process touted by Zola: in a letter to Maupassant, he writes: “Ne me parlez pas du réalisme, du naturalisme ou de l'expérimental! J'en suis gorgé. Quelles vides inepties!” Flaubert, Gustave. Correspondance. V, Janvier 1876-mai 1880. Bibliothèque de la pléiade, v. 539. Paris: Gallimard, 2007, 727. 16 Of course Flaubert’s trips to North Africa and Corsica both in his youth (1840) and during the composition of Salammbô (1858) were, in effect, research opportunities for his work. The observations of contemporary quotidian details of peasant life of the region would be synthesized with archeological and historical details (taken from Polybius’s Histories) from the period of the Punic wars in Flaubert’s writing

22

and worldview of ancient Carthage, attempted to do what had only previously been done

about situations to which contemporary readers could directly relate: to evoke the

thoughts, people the locale, and in doing so, evoke the motivations of a situation and

group of people in their historical, social, and material milieu. To go even further – and

as testament to the suggestion that formal experimentation was at the same time the

impetus and goal of Flaubert’s conception of his “historical realist” novel – we find in

Salammbô the configuration of a narrated and described universe that is not only

completely dissimilar to the familiar milieu of the bourgeois social drama but that also

represents an experience that no longer existed and may not even be accessible to

contemporary language codes, social imagination, or cultural reference. Finally, to push

the experiment to the limits, Flaubert denies us access even to the consciousnesses of the

novel’s protagonists (Salammbô is famously “empty” as a fictional persona17) and opts

instead to evoke the Carthaginian world through the perspective of an observing

subjectivity that is an outsider to Carthage (the observer position of a mercenary soldier,

and that unfixed). These choices, along with the particular stylistic manipulations of

point-of-view and descriptive organization presented here, make Salammbô the ultimate

experiment descriptive style and structure, and not just the fanciful and exotic foray into

process. Flaubert noted the difficulty in one of many letters to Ernest Feydeau on the subject of the seeming impossibility of this combinatory synthesis of observation and “archeological” detail: “C'est une œuvre hérissée de difficultés. Donner aux gens un langage dans lequel ils n'ont pas pensé! On ne sait rien de Carthage. (mes conjectures sont je crois sensées, et j’en suis même sûr d' après deux ou trois choses que j'ai vues). N'importe, il faudra que ça réponde à une certaine idée vague que l’on s'en fait. Il faut que je trouve le milieu entre la boursouflure et le réel.” Flaubert, Gustave, and Jean Bruneau. 1980. Correspondance. II, juillet 1851-décembre 1858. Paris: Gallimard, 1980, 279. 17 “For there exists no “contact” in Salammbô,” writes Victor Brombert. “…the enigmatic nature of the characters seems only to underscore our basic indifference to the ruthless struggle between Carthage and its Mercenaries, as well as our almost total lack of knowledge of the society Flaubert set out to resuscitate.” Brombert’s judgment echoes decades of disappointment on behalf of Salammbô’s critics claiming that a certain kind of communication between a fictional interiority and the reader is summarily lacking in Salammbô. Brombert, Victor H. The novels of Flaubert; a study of themes and techniques. Princeton, N.J.: Princeton University Press, 1996, 93.

23

Flaubert’s own orientalist fascinations that literary history and criticism has been apt to

label it.

Tropes of observation

As evidence of the degree of description’s dominance over this novel, I begin

with the very first chapter, the majority of which is devoted in its entirety to a complex

description of the feast of the mercenary soldiers outside Carthage. In it we find the

language, customs, cuisine, attire, and comportment of this ethnically diverse group

explicated in microscopic detail. Flaubert mixes the detachment implied by the scholarly

precision of his description with a situational framework that integrates his details with

the perceptual ordering of a roving observer. The diegetic status of this observer is

mixed; its perspective is not completely integrated with the world of ancient Carthage,

but certain features (directionality of observation, limitations to knowledge, implied

cultural biases) place local limitations on the content of descriptions that would not be

shared by an external, omniscient describing voice in the conventional descriptive style.

This curious mix first appears in the novel’s initial sentence:

C'était à Mégara, faubourg de Carthage, dans les jardins d'Hamilcar.18

The sentence begins with a demonstrative adjective, ce, a feature that appears frequently

in Flaubert’s more succinct (and usually totalizing) phrases. Appearing as it does at the

18 This and all subsequent citations of Salammbô refer to the following edition: Flaubert, Gustave. Salammbô. Préface de Henri Thomas; introduction et notes de Pierre Moreau. Paris: Gallimard, 1974. This citation, p. 43.

24

very beginning of the novel, can only refer to “the feast” itself, indicated in the chapter

heading just above.

LE FESTIN

C'était à Mégara, faubourg de Carthage, dans les jardins d'Hamilcar.

The sentence shows the dual personality of the Flaubertian observer; on the one hand, as

a textual deixis ce precludes any rhetorical link with the described world, but on the

other, the succinctness of phrasing here implies previous familiarity with the subject.19

By combining a distancing technique with information presented in a more familiar mode

this sentence initiates an unusual relationship between describer and described that will

oscillate between registers (of integration, of implied objectivity) in various descriptive

contexts throughout the novel. The demonstrative adjective reminds us that description

inherently requires an extra semanticizing level that removes us from the observed world,

engaging a hermeneutics of observation before the descriptive act has even been initiated.

To continue:

Les soldats qu'il avait commandés en Sicile se donnaient un grand festin pour

célébrer le jour anniversaire de la bataille d'Éryx, et comme le maître était absent

et qu'ils se trouvaient nombreux, ils mangeaient et ils buvaient en pleine liberté.

19 Leo Spitzer explains this interpretation of the demonstrative adjective in “Racine’s Classical Piano.” In the case of the introductory sentence of Salammbô the demonstrative prompts a distancing and subsequent traversal of levels, from diegetic to extradiegetic. In Essays on Seventeenth-Century French Literature. Translated and edited by David Bellos. Cambridge: Cambridge University Press, 1983, 1-113.

25

Les capitaines, portant des cothurnes de bronze, s'étaient placés dans le chemin du

milieu, sous un voile de pourpre à franges d'or, qui s'étendait depuis le mur des

écuries jusqu'à la première terrasse du palais; le commun des soldats était répandu

sous les arbres, où l'on distinguait quantité de bâtiments à toit plat, pressoirs,

celliers, magasins, boulangeries et arsenaux, avec une cour pour les éléphants, des

fosses pour les bêtes féroces, une prison pour les esclaves. (43)

The description uses some of the major descriptive ordering techniques that are typical

for Flaubert, in this novel and others. The first phrase uses the plus-que-parfait to

establish a causal background to the presentation of the feast: “Les soldats qu'il avait

commandés en Sicile se donnaient un grand festin pour célébrer le jour anniversaire de la

bataille d'Éryx, et comme le maître était absent et qu'ils se trouvaient nombreux, ils

mangeaient et ils buvaient en pleine liberté.” After that, the sentence beginning with

“Les capitaines, portant des cothurnes de bronze…” uses phrases introduced by verbs in

the imperfect as a brace for modifying clauses that accumulate details specific to each

group (les capitaines: voile de pourpre à franges d’or, etc.) and aligns each with certain

features of the camp according to an inscribed hierarchy in which captains inhabit a space

beyond the stables and common soldiers are mixed with animals and slaves. The

sentence structure evoking a topographical arrangement suggests omniscience and

intellectual detachment; the fact of the division between officers and soldiers is presented

as if the information were necessary for a total understanding of the scene from a certain

distance. On the other hand, two features of this brief introduction to the mercenary army

suggest some peripheral involvement on behalf of the describing entity: first, the verb

26

s’étendre, implying the view of the tents stretches out from some visual standpoint within

the scene, and second, the observer trope “où l’on distinguait” in the fourth phrase, a way

of positing a neutral observer within the described scene.

This kind of observer trope appears frequently in Salammbô and deserves special

attention.20 In lieu of the passive “il y avait,” a phrase that would merely render present a

described object without acknowledging the implied presence of an internal observer,

here “l’on distinguait” posits an observing subjectivity, although impersonal, that is

present in the descriptive past of the feast and looks on from a diegetically-inscribed

epistemological position.21 In fact Flaubert almost never initiates a description without

attributing an act of observation to the observing on, a pronoun signaling an entity that

simultaneously inhabits the fictional sphere without acting on it – a non-figure who

nonetheless has the perceptive powers of subjectivity. This observing on deserves some

special attention here: Flaubert uses it in pairings with verbs that indicate perception. In

the first chapter alone, these observer tropes appear frequently and use a variety of modes

of perception to introduce described material: “On voyait entre les arbres courir les

esclaves” (43), “On entendait, à côté du lourd patois dorien, retentir les syllabes celtiques

bruissantes comme des chars de batailles” (45), “…et l’on voyait au milieu du jardin,

comme sur un champ de bataille quand on brûle les morts, de grand feux clairs ou

rôtissaient des bœufs” (45), “On entendait à la fois le claquement des mâchoires, le bruit

20 I have observed the same tendency toward overt use of such tropes of observation in other works by Flaubert as well. For instance, “on voyait” or “l’on voyait” both appear more frequently in Madame Bovary (seventeen occurrences) than in either Salammbô (nine), L'éducation sentimentale (thirteen), or Bouvard et Pécuchet (nine), though Éducation is the longest. Other modes of sensory input appear in similar phrases with similar frequency, with the exception of Bovary, which seems to privilege hearing: “on entendait” appears in descriptions twenty-two times in Bovary, nineteen times in Salammbô, fifteen times in L'éducation sentimentale, and only six times in Bouvard et Pécuchet. There is a general decrease in use of “il y avait” in descriptions over the course of Flaubert’s career. However, the phrase appears with about the same relative frequency in the first Éducation as in the 1869 version. 21 Flaubert prefers this to conventional presentifying tropes such as “il y avait,” and its variants.

27

des paroles, les chansons, des coupes…” (47); in descriptions of Giscon: “À travers les

déchirures de sa tunique on apercevait ses épaules rayées par de longues balafres” (49),

“On n’apercevait que sa barbe blanche, les rayonnements de sa coiffure…” (51); in a

description of Narr’Havas, looking at Salammbô: “…l’étoffe, bâillant sur ses épaules,

enveloppait d’ombre son visage, et l’on n’apercevait que les flammes de ses deux yeux

fixées” (59), in a punctual moment, Mâtho, running: “on le vit courir entre les proues des

galères, puis réapparaitre le long des trois escaliers…” (61), “On entendait dans le bois de

Tanit le tambourin des courtisanes sacrées…” (63), “On apercevait dans les greniers

ouverts des sacs de froment répandus…” (66). Another phase Flaubert employs often in

similar contexts, “on aurait dit que…” gestures to the interpretation of the subjective

observer: “On aurait dit quelque rosse idole ébauchée dans un bloc de pierre…” (85),

“Au delà on aurait dit un nuage où étincelaient des étoiles; des figures apparaissaient dans

les profondeurs de ses plis…” (83), “Sous les évolutions rapides, des portions de terrain

encore dans l'ombre semblaient se déplacer d'un seul morceau; ailleurs, on aurait dit des

torrents qui s'entre- croisaient, et, entre eux, des masses épineuses restaient immobiles.

Mâtho distinguait les capitaines, les soldats, les hérauts…” (note here a shift from “on” to

Mâtho, 50), “On aurait dit que les murs chargés de monde s'écroulaient sous les

hurlements d'épouvante et de volupté mystique…” (113). In all of these instances

Flaubert’s descriptor uses specific lexical choices in the interests of establishing, however

temporarily, the stylistic flexibility that would enable a configuration of experiential

perspectives to be simultaneously engaged in functionally similar but structurally

differentiated descriptive acts in a single piece of prose. The frame of observation being

thereby installed, the description can now proceed to arrange the objects of observation in

28

their detail and heterogeneity of form into the larger syntactic components of the total

descriptive passage.

Word choice in phrases indicating observation is a second aspect of Salammbô’s

descriptions that opens up the descriptive frame to a mixed observer position. In

particular, Salammbô features descriptions in which the directionality of the observation

is inscribed in the selection of verbs connected with described objects. In these phrases

sensory information is a key indicator of the experiential coordinates of the describing

entity. A paragraph-length description that initiates the novel’s third chapter, also titled

“Salammbô,” features information associated with a variety of sensory stimuli in a

composition that relates impressions made by the described objects on an unmentioned

observer. This long description features a number of elements that function to integrate

describer with described world, and is exemplary of many shorter descriptive passages

dispersed throughout the novel.22

La lune se levait au ras des flots, et, sur la ville encore couverte de ténèbres, des

points lumineux, des blancheurs brillaient: le timon d’un char dans une cour,

quelque haillon de toile suspendu, l’angle d’un mur, un collier d’or à la poitrine

d’un dieu. Les boules de verre sur les toits des temples rayonnaient, çà et là,

comme de gros diamants. Mais de vagues ruines, des tas de terre noire, des

jardins faisaient des masses plus sombres dans l’obscurité, et, au bas de Malqua,

des filets de pécheurs s’étendaient d’une maison à l’autre, comme de gigantesques

22 I would like to note that Salammbô is rather imbalanced with regard to the distribution of its descriptions; the majority of the novel’s long, complex descriptive passages appear in the first half of the novel. This imbalance is represented here by the large number of passages taken from Salammbô’s first five chapters. While descriptive passages appear at intervals throughout, the most exemplary and compositionally dense passages appear at the beginning of the novel.

29

chauves-souris déployant leurs ailes. On n’entendait plus le grincement des roues

hydrauliques qui apportaient l’eau au dernier étage des palais; et au milieu des

terrasses les chameaux reposaient tranquillement, couchées sur le ventre, à la

manière des autruches. Les portiers dormaient dans les rues contre le seuil des

maisons; l’ombre des colosses s’allongeait sur les places désertes; au loin

quelquefois la fumée d’un sacrifice brulant encore s’échappait par les tuiles de

bronze, et la brise lourde apportait avec des parfums d’aromates les senteurs de la

marine et l’exhalaison des murailles chauffées par le soleil. Autour de Carthage

les ondes immobiles resplendissaient, car la lune étalait sa lueur tout à la fois sur

le golfe environné de montagnes et sur le lac de Tunis, où des phénicoptères

parmi les bancs de sable formaient de longues lignes roses, tandis qu’au-delà,

sous les catacombes, la grande lagune salée miroitait comme un morceau

d’argent. La voûte du ciel bleu s’enfonçait à l’horizon, d’un côté dans le

poudroiement des plaines, de l’autre dans les brumes de la mer, et sur le sommet

de l’Acropole les cyprès pyramidaux bordant le temple d’Eschmoûn se

balançaient, et faisaient un murmure, comme des flots réguliers qui battaient

lentement le long du môle, au bas des remparts. (98)

In this description the verbs associated with aspects of the city do not describe how these

aspects are, but rather how they would seem relative to a situated and perceptually active

individual: “brillaient” (les points lumineux, des blancheurs), “rayonnait” (les boules de

verre), and “faisait des masses plus sombres” (de vagues ruines), in the first sentence, and

“s’allongeait” (l’ombre des colosses) in the second; “resplendissaient,” (les ondes

30

immobiles), “formaient de longues lignes roses,” (les bancs de sable), and “miroitait,” (la

grande lagune salée) in the third all emphasize the impression of light made by the

objects in the description.23 Using these verbs activates the visual elements of the city as

they would have been seen from outside the city walls (i.e. by a visitor or mercenary

soldier).24 The collection of lexicographical choices in this representation of Carthage

maintains the exoticism and unfamiliarity of the city – its light is mysteriously animated

as if by an unknown inner source, much like the temple at its center – while indicating in

their impressionistic quality the subjectivity of the describer. This is further emphasized

by the amalgamation of sense impressions presented in the description – sight, smell,

(“des parfums…”) and sound (“les cyprès…faisait un murmure”). Certain rhetorical

choices suggest the description is based on subjective reactions to these sensory

phenomena: the use of “çà et là,” for instance, in the first sentence, or the adjective

“vagues” paired with “ruines” in the second, “quelquefois,” indicating repetition and

duration, but imprecise; everywhere, Flaubert’s use of indefinite articles with described

objects adds a note of inexactness to an otherwise detail-filled description. Finally, the

phrase “on n’entendait plus le grincement des roues hydrauliques” fits the ‘observing on’

23 Light and light-effects are important aspects of descriptions in Salammbô, especially in descriptions of either the city, the sacred veil, or the woman. Salammbô seldom ever is in a scene (we never encounter a pithy “C’était Salammbô,” fully revealed) but rather appears, and the light that issues out from around her is often represented using similar techniques as found in this description of Carthage: see, for example, the intermittent description of the princess as Mâtho encounters her in her chamber with the veil, which uses a swaying lamp to successively reveal and ensconce in shadow the sleeping Salammbô, whose luminous body seems to intermix materially with the atmosphere that surrounds her on pages 148-150. 24 The association between the describing subjectivity viewing Carthage in this way and the figure of the mercenary soldier is even clearer in a description of the city written from the point of view of Mâtho: “En face de lui, dans les oliviers, les palmiers, les myrtes et les platanes, s’étalaient deux larges étangs qui rejoignaient un autre lac dont on n’apercevait pas les contours. Derrière une montagne surgissaient d’autres montagnes et au milieu du lac immense, se dressait une île toute noire et de forme pyramidale. Sur la gauche à l’extrémité du golfe, des tas de sable semblaient de grandes vagues blondes arrêtées, tandis que la mer, plate comme un dallage de lapis-lazuli, montait insensiblement jusqu’au bord du ciel,” (169-170); again here the choice of verbs and rotates the directionality of the description toward the observing outsider.

31

model mentioned above. This description, with its mix of precise detail and certain word

choices indicating subjectivity is an apt example of the mixed status of Flaubert’s

descriptor in Salammbô.

Movement-descriptions

The changeable status of Flaubert’s observing agent is also evident in

Salammbô’s movement-based descriptions, descriptions in which the amount and kind of

information available to the descriptor changes in tandem with a changing temporal-

spatial situation – in most cases, as the descriptor moves toward or around the described

object. The aim of these descriptions is to mimic in prose the observational process of a

situated, but unidentified (or at least not aligned with a single postulated observer),

observing agent.

Though the main text I will be dealing with here is Salammbô, a brief look at

another description in Flaubert’s work will support my claim that this was in fact a

primary concern for Flaubert throughout his career. Even one of the most famously

synthetic descriptions of Flaubert’s oeuvre, the initial observational encounter of Frédéric

Moreau and Madame Arnoux in L’éducation sentimentale of 1869, demonstrates a

discursive complexity that the notion of a Flaubert responding only to a crisis of

representation does not adequately explain:

Ce fut comme une apparition: elle était assise, au milieu du banc,

toute seule; ou du moins il ne distingua personne, dans l'éblouissement

32

que lui envoyèrent ses yeux. En même temps qu'il passait, elle leva la

tête; il fléchit involontairement les épaules; et, quand il se fut mis plus

loin, du même côté, il la regarda. Elle avait un large chapeau de paille,

avec des rubans roses qui palpitaient au vent, derrière elle. Ses bandeaux

noirs, contournant la pointe de ses grands sourcils, descendaient très bas et

semblaient presser amoureusement l'ovale de sa figure. Sa robe de

mousseline claire, tachetée de petits pois, se répandait à plis nombreux.

Elle était en train de broder quelque chose; et son nez droit, son menton,

toute sa personne se découpait sur le fond de l'air bleu.

Comme elle gardait la même attitude, il fit plusieurs tours de droite

et de gauche pour dissimuler sa manœuvre; puis il se planta tout près de

son ombrelle, posée contre le banc, et il affectait d'observer une chaloupe

sur la rivière. Jamais il n'avait vu cette splendeur de sa peau brune, la

séduction de sa taille, ni cette finesse des doigts que la lumière traversait.

Il considérait son panier à ouvrage avec ébahissement, comme une chose

extraordinaire. Quels étaient son nom, sa demeure, sa vie, son passé? Il

souhaitait connaître les meubles de sa chambre, toutes les robes qu'elle

avait portées, les gens qu'elle fréquentait; et le désir de la possession

physique même disparaissait sous une envie plus profonde, dans une

curiosité douloureuse qui n'avait pas de limites. (22-3)

Right away we see notice the abundance of sensory information here, and as has

been noted, Flaubert emphasizes the visual impression made on the observer – here,

33

Frederic – with a heavily visual lexicon (“apparition,” “distingua,” “regarda,” even “se

découpait,”).25 But the descriptive passage does not limit itself to the visual. A spatial

element is also present; Frédéric sees Madame Arnoux in a visual context made possible

by a literary geometry that comes out in the diachronic presentation of a continuum of

observational presence through time; the effect is distinctly cinematic. But the workings

of space-time, while rarely as complex as in this example, have always caused difficulties

for the descriptive author; Flaubert’s modulation of observational temporo-spatiality is

hardly unprecedented in the history of the novel. A description may have the marks of a

dominant universal variable (like space or time), but no description is only spatial or only

visual – this is simply the rule of observation in a relativistic universe. The observational

foci of the description, therefore, are not merely “moments,” or “positions,” (temporal

and spatial metaphors, respectively), but effects of the inscribed relationships between

these variables and an observing and/or describing subjectivity, a descriptor. The

perspective we get of this scene is the product of a cooperation of discursive elements

(syntax, ordering), and interpretive acts (observer, descriptor, and finally reader) and can

only be called experiential. In Flaubert’s descriptions, the discursive elements have been

enhanced to the point that they take on some of the hermeneutic responsibilities of the

implied subjectivities that should be involved.

25 As notes Sara Danius in a recent study of the “visual” in nineteenth century realism: “The woman on the bench is presented as though in a full-length portrait. Devouring her being from head to toe, Frédéric’s gaze moves from her straw hat down to the lining of her dress. The description of her external appearance is exclusively visual and meticulously detailed…Flaubert’ aestheticized representation of the unknown woman quickly turns into an independent entity. It is an island floating around in the narrative that surrounds it, self-contained and self-sufficient. What unfolds before the reader is not so much a description of a woman as rather an image – an autonomous image that has been carefully inserted into the diegesis.” In The prose of the world: Flaubert and the art of making things visible. Uppsala: Uppsala Universitet, 2006, 26.

34

The degree to which certain stylistic and lexical choices emphatically undermine

the fiction of objective, unifocal description is increased by an inscribed flexibility of

describing subjectivity. The description is told from the “point of view” of an impossible

observer who is both factually omniscient (descriptions have more information than any

single character could) and stylistically subjective (they are written with the tonal

inflection of an invested subjectivity). The passage begins with the announcement: “Ce

fut comme une apparition.” If this passage had been about the observed objects

(Madame Arnoux and her many qualities), it would read “elle fut comme une apparition,”

but the demonstrative adjective “ce” refers to something else: the paragraphs to follow,

the descriptive experience itself.26 The person who says “ce,” therefore, cannot be inside

the diegesis, but rather shares some experience both with the descriptor and the reader.

However, the clause “ou du moins il ne distingua personne, dans l'éblouissement que lui

envoyèrent ses yeux,” temporarily fixes the descriptive-experiential subjectivity in the

person of Frédéric, whose imagination is poetically obfuscated by the perceived splendor

of this appearance. Again, with “Elle était en train de broder quelque chose,” the

“quelque chose” implies an observer who either can’t see or doesn’t care what Madame

Arnoux is embroidering. This example demonstrates that Flaubert’s descriptions can be

read without having to settle on an observing subject position. Indeed much effort is 26 Indeed the entire passage persistently employs the demonstrative in such a way as to remind the reader of the very situatedness of the current observation: “Jamais il n'avait vu cette splendeur de sa peau brune, la séduction de sa taille, ni cette finesse des doigts que la lumière traversait. Il considérait son panier à ouvrage avec ébahissement, comme une chose extraordinaire.” One will recall Flaubert’s use of the demonstrative adjective to elevate, and to thereby lend a level of reflexivity that often implies irony, in Madame Bovary: “C'était un de ces sentiments purs qui n'embarrassent pas l'exercice de la vie, que l'on cultive parce qu'ils sont rares, et dont la perte affligerait plus que la possession n'est réjouissante,” (Madame Bovary 123). Genette notes a similar usage of the demonstrative in his study of Stendhal in Figures II: “Le cas du démonstratif (“Cette malheureuse”) dont Stendhal fait un usage très marqué, est un peu plus subtil, car s’il s’agit essentiellement (abstraction faite de la valeur stylistique d’emphase, peut-être italianisante) d’un renvoi anaphorique du récit à lui-même (la malheureuse dont il a déjà été question), ce renvoi passe nécessairement par l’instance de discours et donc par le relais du narrateur, et par conséquent du lecteur, qui s’en trouve imperceptiblement pris à témoin” (190).

35

made on the part of the descriptor to destabilize any features of an observation that would

match, conclusively, the entirety of a description with a single observing subjectivity.27

The complexity of Flaubert’s perspectivism is not just sensorial; in its stylistic

manifestation it becomes attitudinal, it assigns values, it is experiential. It strives not

only to relate a set of objects to a set of attributes, but also to model an organic process of

observation.

Though the metaphorical notion of perspective in literature originated in the

visual arts, in the literary field perspective has become one of the most elaborated

concepts of literary technique available to the critic of nineteenth century literature.

Terminology varies, but almost all versions of the idea of perspective stem from a

metaphorical usage of physiologically determined modalities (with the exception of

Todorovian “aspect”): “point of view,” “voice,” “vision,” “focalization.” Of central

importance to these definitions is the concept of a position (as if on a coordinate plane) –

aligned with variables of space, of time, of knowledge, experience, interest, of some

combination of all of these -- from which a narration (or description) can be made.28

27 Timothy Unwin offers a concise description of th