Jane Gaines I 181
10
Costume and Narrative: How Dress Tells the Woman's Story
Jane Gaines
In The Costume Designer, a 1949 educational short on Hollywood studio production, Edith Head, in voiceover, describes the thought processes of the designer as film clips illustrating her method of costuming character appear on the screen. Head describes the character as a twenty-year-old woman from a wealthy family. The young woman is dressing for an evening social engagement with an old friend, but she is unhappy because she has just quarreled with the man she loves. "Should the costumer put her in something white, gay, and frothy, or in a more sophisticated and clinging jersey?" asks Head. "But wait a minute! She cries just before she leaves for her date. That's a big emotional moment and deserves all the emphasis it can get," the designer explains. As the costume of the actress on the screen dissolves from a slinky, sequined evening dress to white flowing chiffon, Head describes this softer effect as her solution to the narrative problem.
Edith Head's decision is based on the motion picture costumer's code ^which weighs design against dramatic content. Of all of the designers from the classical period, Head most fully articulated this costume-narrative tension and she continually returned to this theme in numerous interviews and magazine articles as well as in her autobiographies.1 As Maureen Turim emphasizes elsewhere in this collection, Head was often credited v^thjproducing "storytelling wardrobes." Her examples, then, demonstrate more clearly for us what Hollywood creative personnel meant by letting the costume "tell the woman's story," a phrase which appears frequently in screen designers' descriptions of their craft. A closer examination of the work of key designers, however, shows that costume was severely restricted in what it was allowed to "tell," with some exceptions which I will discuss. This restriction has to do with the primary function
of costuming in classical realist cinema where every element in the mise-en-scene—from painted backdrop to prop to lighting cue—serves the higher purpose of the narrative.
In this essay, then, I will consider the constraint on screen design in terms of the antithetical relation between costume and narrative. Clothes, as lower elements in a hierarchy of screen discourses, primarily work to reinforce narrative ideas. On occasion, an accessory planted as a prop may'come to narrative fruition. A telltale glove or shoe (or Head's own example, the scarf with which the heroine is strangled), may start out as part of an ensemble and later serve to advance the narrative.2 But primarily costumes are fitted to characters as a second skin, woricin^njhis^pacjty for the cause of narrative by relayjngjnformationto the viewer about a "pers^.'^Thus, I wilfalso be concerned here with the nineteenth-century ncSoiTof "personality" as it applies to the clothing of the self.
Since the following is the briefest introduction to an area that has thus far been literally untouched by film scholars, I am limiting my examples as much as possible to black and white contemporary dress drama. In the first two parts, I sketch out some of the history of motion picture costume design, taking account of the transition from silent to sound technology. A third part describes the classical star-designer relationship. The special problems posed by color processes and period dress, and the exceptions represented by such genres as the western and the musical are beyond the scope of this basic inquiry.3 However, in the fourth and last section I will make some remarks about melodramatic form as well as the woman's picture as genre since historically it is the woman's story that is told in dress. Although all characters, regardless of gender, are conceived as "costumed" in motion pictures, a woman's dress and demeanor, much more than a man's, indexes psychology; if costume represents interiority, it is she who is turned inside out on screen.
Costume in Silent Film
From the mid-teens, I find such a close link made between dressing the part and playing the part that it is almost as though "acting" for the female silent film player consisted of nothing more than stepping into and out of different costumes. This conception may have to do with the way motion pictures were costumed before the advent of the studio designers. Origi-, nally, the aspiring actress was expected to purchase her own contemporary screen wardrobe which had to include at least one evening dress and an outfit for every possible daytime occasion.4 "Dressing for the Movies," one of the earliest articles on the subject, appearing in Photoplay in 1915, describes the job of the actress as entailing designing as well as sewing the costumes required by the part.5 Interpreting the role for the silent
180
182 I Costume and Narrative
Jane Gaines I 183
actress also meant translating character into costume by means of the "dress plot," a kind of scenario indicating all of her wardrobe changes. In later years, under the studio system, the designer rather than the actress translated the script into a costume plot; after designers were phased out, this "plot" was treated as the kind of continuity record which Elizabeth Nielsen describes in her history of the costumers union included this collection. Before Before Erte, Chanel, and Schiaparelli were hired to bring French couture to the screen, actresses were literally their own designers, although as they became stars they no longer did their own sewing.6 Instead, the stars' original designs were rendered by a personal tailor or modiste.7 Signjfteantjiere is the evolution of the distinction between street wear and motion pjcJun^ostunte,ji^h^ paralleled the development of stardom. The issue of ordinary clothing versus costume was first expressed in the predicament of the struggling picture actresses who, to make ends meet, had to appear in different roles wearing the same gowns, or, had to make their screen clothes double as everyday dress.8 Behind this doubling-up function is the assumption that costume I for contemporary motion picture drama was not different from ordinary I dress, and for a brief period in the early teens, these practices were I defended on the basis of verisimilitude. Very quickly, however, the development of a costume aesthetic specifically fitted to the screen began to ' pose a challenge to the realist position which favored everyday wear over ' special costuming.9 Eventually, discourse about costume would become i locked into this contradiction which came to a head in the mixed message I of costumers' advice to female viewers circulated by publicity departments: Do copy what the stars wear on the screen, but don't wear screen styles because they are too exaggerated.
"Onyunusual article from the teens, "How I Teach My Gowns to Act," exemplifies again this idea of the costume as actually taking a role or part, but goes on to attempt a definition of a cinematic costume aesthetic. Actress Marguerite Courtot is quoted on the importance of line, contrast, and tone value. The need to "render" color in black and white, she says, points up the advantage of a specifically cinematic dress as opposed to an ordinary dress which might ^ot make a strong enough screen statement.10 In the process of formulating the new rules of screencostume, theseearly articles espouse two antithetical values at once: cosmmejnust register on the screen at the same time it must recede. However,during^hese~years, costume had more opportunity to register than it would in later years. Silent film costuming was marked by a gratuitous flourishing which was "toned down" after sound. In the earlier period, one finds examples of camera concentration on a costume detail; aesthetic coordination of cos-tume^and sete (Natasha Rambova's art nouveauT costume designs for Camille [1921] and Salome [1923]); and camera movement organized
Figure 10.1 The Kiss (1929). Adrian's costumes for Garbo coordinated with the art deco interior. MGM, courtesy of BFI Stills Archive.
around the display of an entire costume. An especially interesting example of all of these possibilities can be found in Jacques Feyder's The Kiss (1929), in which costume is coordinated with the art deco decor (Figure 10.1), but also, as in the opening sequence where Greta Garbo meets her lover in an art museum, camerawork and editing seek out all of the details of her Adrian-designed fox fur-trimmed coat, afternoon dress, and cloche hat. A clue to the attitude that costume can be a dangerous signifier is the containment rfjinbridJedyisuaXdisplay within dream sequences (Forbid-a^en^ruit^\92l and Male and Female, 1919), and fashion sequences (Fig Leaves, 1926) (discussed elsewhere in this collection by Jeanne Allen and Charlotte Herzog). Outside these conventional boundaries, in the dining room or parlor, the elaborate costume, like poor etiquette at a dinner party, could claim undue attention for its wearer. Thus Courtot, in describing a cerise velvet gown featuring black chiffon sleeves with matching fox fur trim and silver lace at the hem. explains that "one must see to it that these daring bits of fur must never over-actj for "any subtle acting of the wearer would be lost along side of the gown screaming its presence."" And yet the opposite of the detail which "overacted," the dress with jio
184 I Costume and Narrative
distinctive styling or detail whatever, could not "act^at^all according to Courtot"who thus describes the problem:
For instance, the other day I was shown a draped chiffon gown of exquisite orchid shades. Not only were the variations of tone insufficient to register any contrast on the screen, but the ineffectual, caught-here-and-there draperies were meaningless. These hanging lengths of lovely color could neither festoon with joy nor droop for dejection . .
What this suggests, however, is that not only did the screen costume need to exhibit a sufficient contrast of light and dark and a strong line in order to make an interesting photographic composition, but that it had to have the kind of style and fabric combination which served the narrative by restating the emotions which the actress conveyed through gesture and movement. Stepping into a costume, was like stepping mto a^ole.,Cos-tumes, furthermore, were expected to express the same feelings ("joy" or "dejection") called for in the part. And here I emphasize' "same" to call attention to a pair of assumptions underlying these early discussions of character and costume. One assumption is that a role can and should only be costumed with rather than against the personality of the^character. (For Courtot, the design qualities which contradict her conception of her role do not signify disguise or discrepancy, a counter-meaning, but "meaning-lessness.")
In other words, there is a notion of personhood in operation here which assumes a continuity between inner and outer rather than two discrepant parts. But the other related assumption (only too familiar to us at the end of the twentieth century), is that dress is a key to the personality of the wearer. In fact, in this discourse, dress becomes somewhat more than a key or an indicator since "personality" and "dress" are so often confused that it would seem that they have become the same thing. We are, after all, talking about two distinct orders of things—human beings and material goods. Why ascribe the same attributes to clothing as one would ascribe to a person?
The idea of dress as the key to the self is not confined to discussions of acting and costuming, but also informs the practice of silent film screenwriting. Frederick Palmer, in the chapter on "Visualization" in his 1922 manual, tells aspiring screenwriters to study character by observing gloves, shoes, and jewelry, because "So much of character is told inone's manner of wearing clothes."13 Clothes and mannerisms in these early manuals are not vehicles for conveying the sense of a "real" person nor are they elements utilized in the craft of character construction, they are "truths" told about persons. Character writing here depends on an idea that real selves (rather than types) can be studied by reading appearance
Jane Gaines I 185
signs which are communicated in public. Another manual contains this advice:
Not one of us but is shouting his or her personal characteristics every minute of every day by the clothes we wear and the way we wear them,and the creative writer can be on the lookout for these tiny indications.14
For the observer as well as the aspiring actress, choosing clothes is conflated into the manner of wearing them. They are simply a matter of the expression of the self. This kind of observation was then a much simpler matter before Erving Goffman called our attention to the difference between the expressions which a person "gives" as opposed to those he "gives off," as well as to the problem of the impression which is received by others.'5
This notion of clothes as self-expression is a product of a combination of the late nineteenth-century sense of public self-in-urban-space developing in Western industrial culture and a Platonic understanding of the matter/spirit dichotomy. Carlyle expresses this in his admission that, although clothes are "despicable" they are unquestionably significant because they are vehicles for the soul. Clothes should thus be seen as the matter necessary to "body forth an idea."16 The problem of the public self which made immediate impressions and the true self which was within, was dealt with in the possibility of personality management and improvement which historians of modern society find emerging at the turn of the 'i century. Not surprisingly, the evolution of the contemporary sense of "how j to" make a personal impression on others coincides with the appearance of! the vanguard institutions of the consumer culture—the department store and the nickelodeon.17
In the century before, as Richard Sennett tells us, the social vogue had been marked by a sense of the discontinuity between the person and the dress, or the belief that the true self was quite safe from public eyes which might see connections between one's body and one's soul.18 At the root of this earlier version of selfhood was the Enlightenment conviction that natural character, which connected all humans, was fixed at birth.1 The exterior, divorced from the interior, could be used ironically in any imaginable way—hence the eighteenth-century moment in costume history in which the body was most literally nothing more than a mannequin on which to hang clothes.20 As Sennett describes the historical abandonment of the permanently fixed natural character and the adoption of the modern personality, the new notion no longer depended upon an idea of separate parts, but on the inextricable connection between the inner and the outer. While it would seem that this understanding of the personality as a con-
186 I Costume and Narrative
Jane Gaines I 187
structed or cultivated self would give individuals a sense of social confidence in facing a world of strangers, the Victorians, as Sennett describes them, assumed a sophisticated anonymity in their dress and demeanor out of concern that their exterior appearance (and details of sexuality and class in particular) would betray them.21
One has to ask if this idea of the involuntary expression is not fundamentally at odds with a new conviction that personality could be acquired and perfected. (But this, as we know, is still an enigma to us in the later part of the twentieth century!) In "Expression of the Emotions," a series of articles by Eugene Brewster appearing in the early years of Photoplay, the author explains silent film acting in terms of the exteriorized social personality which he understands as a combination of contrivance and involuntary expression. He bases a case for the ease with which which audiences followed silent acting techniques on this theory of the personality:
We learn to drive a sharp bargain, to put our best foot forward, to conceal our weaknesses, to exaggerate our good points, and in various ways to create a good impression, all of which necessitate acquiring control of our emotions and of our modes of expressing them. And in so doing we learn to express false emotions.22
We want to ask how the motion picture theatergoer could be assured that he or she could correctly read silent film gesture, posture, or dress by pointing out the incidence of dissemblance in human self-presentation. Although the same notion of personality as a socially decipherable self whose inner and outer layers were in tune informed both social costuming and theatrical stage dress, a dual set of rules was involved. Where the former was thought to involve some subterfuge, the latter was expected to reveal a character (confirming how people of different gender, age, nationality, and social class were thought to be in "real life"), fqr^the convenient comprehension^of an audience.23
In some ways, silent film costuming developed along with the same psychological realism seen in the evolution of theatrical as well as motion picture acting styles in the teens.24 A new_code of naturalism stressed individuated characters over types and emphasizedTfie~fäčeTeyes, and lips as conveyors of "interiority." 5 It is in the discussion of acting (which costuming discourse has historically echoed) that one finds the articulation of the idea that the interior can be registered visibly on the theory that there is a direct line of communication between interior and exterior. Although this early literature does not suggest that an actor or actress drew on previous experience in the portrayal of character emotion (as it later would), acting did involve drawing from an existing reserve within; cos-
\ tl
turning evidenced this evocation at the same time it facilitated it. Clara Kimball Young, in an early Photoplay article describes the way the color in the costume she wore brought out "latent elements" in herself:
To some women red is like a fire, bringing out the glow. To others red is deadening, as if it killed their paler fires. Unless a woman feels that red is essentially related to something within herself, she should avoid wearing it as she would the plague. On the other hand, if she can wear red, she should wear it at such times as she desires to express these qualities that its warmth and richness bring out by its contact with herself.2"
The notion that costuming "brought forth" traits would continue into the sound era and appear repeatedly^as^one of the justifications for the lavishnessof star costuming in the studio era. In the production of black and white films', the special care taken with the use of color was generally attributed to the boon it afforded the actors who could not be expected to project realistic emotions if they were costumed only in shades of grey. A legendary mystique even seems to have arisen around the invisible detail which secretly helped certain actresses to perform, a variation on the actor's practice of putting on a costume as a means of learning about a character. The painstaking work of the costumer—embroidering the star's initial on the inside of a glove, lining a jacket with silk crepe, or adding extra rows of lace to the bottom of a petticoat—was elevated because it aided an actress in the realization of the character.27
Thus in the discourse on costume, dress, like an expressionjjfjmiotion, seemed to grow out of the mysteries bfthe body. This close association witfTthebody helped to construct costume as_behaviori an indicator which in popular, usage could subsume the social, moral, and psychological:
The dowd dresses dowdily, the woman of spirit and originality dresses that way, the business woman dresses in simple tailor-made things, and the adventuress dresses to lure.28
But note here that dress is charged with expressing only one trait or<, reinforcing one quality. While the body was used in acting to express emotional'complexities and to enunciate subtle gradations of feeling, costume was expected to simplify. Costume detail was "fixed" in the Panofskian sense in that it stood, again and again, for the same thing, and could be counted on to provide the most basic information about a character for the spectator, that is, it typified.29 And because of this typifying function, J would contend, film theory has historically lumped costume along with furniture under the category of iconography.
188 I Costume and Narrative
Not only did costume, like decor, provide iconographic cues related to typage and narrative conventions; in the absence of sound it was seen as a substitute for speech. "Dressing for the Movies" goes on to describe how "The elimination of voice from the picture dramas calls the greater attention to movement and raiment."30 Since these elements were conceived as standing in for the missing dialogue, for audiences, attending to costume and gesture was somewhat like listening. Mary Ann Doane describes the displacement of speech in silent cinema, "The absent voice re-emerges in gestures and the contortions of the face—it is spread over the body of the actor."31 If the voice is spread over the body, however, it is spread unevenly, with expressivity concentrated on the upper parts. Yes, vocal enunciation is dispersed over the body, but the uneven distribution of "speech," favoring the fingers, eyes, and lips, suggests the hierarchy of vehicles of expression alluded to earlier. Dress, even in coordination with the mojyement of the torso and arms, is not conceived as having the i eloquence of eyes or lips. If it "speaks," it is treated rather like the flat {character who is never"given more than one line to deliver. In silent Icinema, this line was usually either: "I am the vamp" or "I am the straight
Costume in the Sound Era
The rules of costume and typage—that the dress should place a character quickly and efficiently, identifying her in one symbolic sweep—were held over into the sound era where they coexisted with ideas about a new realism in dress which meant both subtlety and contemporary fashion-consciousness. Edith Head, who began work at Paramount after the silent era, has insisted that costume should carry enough information about characters so that the audience could tell something about them if the sound went off in the theater.33 Since in the sound era narrative came to depend more and more on dialogue to establish identity and reveal interiority,this might have meant that costume could be relieved of its old responsibility. But the discourse on costume still remained centered on characterjcoherence. The use of bold design, enlarged detail, and exagger-aEecTshapes was consistently defended not as aesthetic interest in and of itself but as the need to make a character statement.
The legacy of the silent era was really a tendency toward metaphorical literalization in costume design. Take, for example, costuming for the type of the dangerous woman in both silent and sound periods. In 1914, vampiress Theda Bara, in A Fool There Was, is literally embalmed to the neck in the sack-like full-length coats and dresses she wore in scenes where she attached herself to the throats of her victims. As a man-trap in I'm No Angel (1933), Mae West is costumed by Travis Banton in a black
Jane Gaines I 189
Figure 10.2 /'m No Angel (1934). Travis Banton's rhinestone spiderweb pattern on Mae West. Paramount Pictures, courtesy of the Museum of Modern Art Stills Archives.
chiffon with diaphonous sleeves decorated with rhinestones in the design of a spiderweb (Figure 10.2). In The Little Foxes (1941), Bette Davis, the predatory Regina, wears a huge hat decorated with a literal bird of prey (Figure 10.3). All of Davis's dresses in this period film (designed by Orry-Kelly) feature long tight sleeves, some of which extend beyond the hand like claws.
As Edith Head's commentary on characterization reveals, a designer is
190 I Costume and Narrative
Jane Gaines ! 191
Figure 10.3 The Little Foxes (1941). Orry-Kelly transforms Bette Davis into a bird of prey. Warner Brothers, courtesy of the Museum of Modern Art Stills Archives.
uj^mg^ejaphors in this way every time he or^h^tanslates^^ajcharacter
trait—such, a^jthat, of '^g^^^c^^§-sJ^^^?feSSS£3^9- 9X1 item of apparel. Asked in an interview how she would costume various types, Head does not hesitate, saying that to her, a self-centered woman translated into a "covered-up look," a generous woman into "clothes that look easy and give a little," and a shy or rigid woman into clothing with a collar standing up around her face, cuffs that extend over her hands,
and ruffles.34 What is interesting here is that while this stated theory of costuming was committed to the humanist notion of a unified self, unstated costume practice, when not reined in by producers, could and did suggest characters as complexly divided selves.
The stated theory of character costuming, then, holds that the costumer's creative process works through equivalences and is like other processes ofadgptation in which different artistic systems are consciously compared. Dudley Andrew's work on transposing novels into film is relevant here, particularly as he discusses adaptation as a common aesthetic practice which often goes unnoticed. Essentially, adaptation is nothing more than a kind of borrowing in which one aesthetic domain is raided to enrich another, he says. The search for equivalence, more specifically, has to do with lining up two domains and "matching" the terms in each. Citing the work of art historians E.H. Gombrich and Nelson Goodman, whose analyses of adaptation as systematic exchange predate semiotic theory, Andrew describes adaptation as "searching two systems of communication for elements of equivalent position in the systems capable of eliciting a signified at a given level of pertinence. . ,"35 By pairing signifieds, colors are translated into musical tones, vegetables are understood as animals, and fabrics are transposed into character traits. In the fabric system, for instance, wool tweed corresponds with "serious," black satin with "wicked" or "decadent," and tulle is "lighthearted." Such an equivalence between fabric and mood informs Carole Lombard's warning, "Don't discuss politics intulle!"36
In costume discourse, what is semiotically nothing more than a mismatching of two systems, whether mood, occasion, or personality system, often takes on connotations of moral prohibition and becomes an "inappropriate" choice. "Appropriate" costuming for character is rather like the value of fidelity in literary-film adaptation. It depends on creating the impression of "rightness" by striving for exact connotative equivalence for items drawn from relatively unlike systems. Frequent use of such combinations or coupling of items establishes the rule or code, and eventually this coding becomes "naturalized."
The highly ideological process of naturalization, as Stuart Hall describes it, depends on customary use. Or, as he says, the existence of naturalized codes is the result of the "degree of habituation produced when there is fundamental alignment and reciprocity . . . between the encoding and decoding sides of an exchange of meanings."37 The perfect connotative match between systems also makes an easy argument for an existential connection between two terms, and, in the case of costume, this naturalization helps to reassert again that the personality finds its direct manifestation in manner of dress. The naturalization of the costume/character relation is such that the more perfect the fit between the two systems, the less the
192 I Costume and Narrative
Figure 10.4 No More Ladies (1935). Crawford's starched pique collar—Adrian's costume excess unmotivated by character. MGM, courtesy of the Museum of Modern Art Stills Archives.
costume will be seen as costume; at some point, we will see a character as merely wearing clothes.
Film scholars have theorized classical realist cinema as a finely tuned naturalization machine which presents to the viewer a perfectly contained world that is an apparent continuation of the one in which he or she lives. The smooth running of the narrative is produced by erasing any signs that it is different from the actual world and by tightly coordinating its various parts. In this economy of classical cinema, costume signs are rationed for the sake of the unity of the work. As much as is possible, the presence of costume must be justified or motivated by characterization.
To give an example of what can happen when a stylistic flourish is not adequately motivated by character, one could compare two different collars designed by Adrian for Joan Crawford. The starched pique collar seen on the Marcia character in No More Ladies (1935), a farce about infidelity, went so far beyond the function of characterization that fan magazine writers singled it out for humorous commentary, referring to it as a nun's habit turned upside down (Figure 10.4). In contrast, on the stenographer Flaemmchen in Grand Hotel (1932), Adrian's asymmetrical collar is
Jane Gaines I 193
justified, as it helps to create the characterization of a loose-moraled woman whose clothes are always slightly askew (Figures 10.5 & 10.6). Or, designer Walter Plunkett could attach a life-size bird resting on a branch to the bodice of Jennifer Jones's gown in Madame Bovary (1949), since it signifies Emma's misinterpretation of aristocratic tastes (Figure 10.7).
In the service of narrative ideas costume is assigned one main function: characterization. In this capacity, costume also works to blend straggling physiological signifiers so that they contribute to character. Here, Stephen Heath, describing the circularity of the actor, person, and character relation, Jias referred^to costume as well as make-up^s ^sorbing f^cejnd body into charactersignificance^"38 (As I point out in the introduction to this volume, however, the actual process is less benign, since actresses' bodies are often painfully restructured by designers.) Costume assimilates bodily signifiers into character, but body as a whole engulfs the dress. Hence the paradox of costume transparency. Costume historian Anne Hollander explains the invisibility of theatrical dress:
Costumes are so thoroughly identified with bodies that the messages they send are received without acknowledgement, even though an extraordinary emotional power can be generated by the use of very specific, noticeable things—the right use of a black cape, a white scarf, or a pair of bare feet.39
Thus it is that costume is eclipsedj>y both_character and body atjhe expense of developing its qwnaesthetic discourse. Bound to character and body, it isTsocialized, conventionalized, tamed. Like make-up on the face, costume is invisible as it is present,40 and the irony of the concern over costume superfluity is~that fhe real but unforeseen danger is not in too much costume, but in the total absence of it—the body naked.
With the coming of sound, and eventually color, those additional details which seemed to promise the approximation of the real also became a threat to the establishment of the realist aesthetic itself. Here I want to call particular attention to the importance of narrative continuity as it constructs the illusion of the real.,Cosmrne,JikejjO^^ potential to distract the viewer f^n^ thejurrradve, which could result in breaking the illusion and the spell of realism. An early Society of Motion Picture and Television Engineers manual on color technology uses costume as a comparison with color to illustrate the rules of continuity cinema. Color, like costume detail, could become "clutter" on the screen unless restricted and held to its character identification function.41 But costume was to be given little more to do than insure that the character was clearly seen. If costume did not "punctuate" the actor against the backdrop the
194 I Costume and Narrative
Jane Gaines I 195
Figure 10.6 Adrian's asymmetrical collar on Crawford's loose-moraled stenographer— design motivated by character. MGM, courtesy of Academy of Motion Pictures Arts and Sciences Stills Archives.
Figure 10.5 Adrian's asymmetrical collar on Crawford's loose-moraled stenographer—design motivated by character. MGM, courtesy of Academy of Motion Pictures Arts and Sciences Stills Archives.
viewer might not be able to "follow the story."42 Here, costume helps to tell the story merely by insuring that the viewer can distinguish the narrative agent from other elements in the mise-en-scene. In the ecosystem of classical cinema, telling the story requires subordinating an especially evocative aesthetic to narrative designs.
Hollywood directors in the sound era, when asked about costuming, have consistently insisted that the costume's contribution to telling the story called for subservience to it. George Cukor, who can be associated with some of the most interesting designing for black and white sound film, such as Adrian's work on Camille (1936) and The Women (1939), is quoted in an interview as saying that he did not have a favorite costume because the ideal costume was the one that most perfectly "suited the If the costume "knocked your eye out," he goes on, it was not
scene.
good for the scene or for the entire film." The rule is that costumes, should be modulated, much like orchestral underscoring for Hollywood melodrama which was so carefully matched with emotional connotations that it was heard but not noticed. The costumer's formula, as Alice Evans I Field defines it, requires that clothes be "harmonized to the mood, be it
196 I Costume and Narrative
Figure 10.7 Madame Bovary (1949). The awkward and ridiculous bird perched on the bodice of Etnma's ball gown—Walter Plunkett's design for Jennifer Jones. MGM, courtesy of Academy of Motion Pictures Arts and Sciences Stills Archives.
comedy, tragedy or romance; they must add subtly to the grace of the wearer; and they must enhance the rhythmic flow of the story. Never must they call undue attention to themselves, unless for sharp definition of character, and they must have originality of detail within the certain bounds of good taste."44 Narrative realism dictates that costume be curtailed by conventional dress codes; continuity requires that it be monitored for the telltale continuity error; economy requires that it reinforce causality.
The directorial practices informed by this notion of costume as servant of narrative ideas, then, were at odds with any vision of costume as spectacular design that might have no other purpose than to feed a visual appetite. Directors might plan compositions that would lop off exquisitely detailed handiwork around a hem, or block a scene wiirTinďSeřěncé to the back interest focäľpoiňt of a gown, as is the case with Adrian's bias-cut satin gown with cut-out sleeves, scarcely seen on Jean Harlow in China Seas (1935) (Figure 10.8). The practice of cutting films for narrative coherence and visual continuity meant that close-ups which featured design interest around the collar disappeared in the final print and even that many costumes created for the production would be edited out of the film
Figure 10.8 China Seas (1935). Back interest in Adrian's bias-cut dress for Jean Harlow. MGM, courtesy of Academy of Motion Pictures Arts and Sciences Stills Archives.
but would show up in studio publicity stills.44 As a guide to costume in motion pictures in the studio era, old fan magazines, star biographies, and glossy cocktail table books are often misleading since they reconstruct the golden age from production stills which are pre-release records of the publicity image of a new production.In comparison with these production, stills, the costumes as filmed in final scenes are often a disappointment. | The fashion coverage in popular magazines from the period, as well as the nostalgia books read in retrospect, portray the contemporary trend-
198 I Costume and Narrative
Jane Gaines I 199
setting on the screen as unproblematic, as though all Hollywood creative personnel had as much enthusiasm for fashion developments as the women in the audiences. Directors, however, disassociated themselves from this trend-setting, and generally held the belief that motion pictures should not "cater" to fashion.46 In addition, the designers themselves often denied that they were trying to "set styles." Costume extravagance came to be associated with consumerism, which meant that it was a double threat— to narrative coherence as we have seen—and to a notion of art that stands apart from the vicissitudes of commerce.
■ Certainly there is a contradiction here since Hollywood production s have historically been unabashedly commercial. The motion picture industry in this period, however, represented its product as uplifting entertainment
! which stood outside time and was never "dated." The industry avoided tying
in so closely with the woman's fashion trade that it would be required to refer to the seasonal shifts which stimulated the retail clothing business. The 1930s phenomenon described in Charles Eckert's article reprinted here was an exception. In part, the retail tie-in was cooperative advertising, an effort to counteract the economic effects of the Depression. However, it is best to see the collaboration between Hollywood and Seventh Avenue as a publicist's vision, a reality at the distribution, not the production, end. Jtecause of the tensions created within the^fihnjtse^ycompetln^dis^ourses—the rhetoric of costume and contemporary style consciousness asagainst narrative coherence and economy—the interest in style and the work of the studio designers had to be deflected elsewhere. And here the institution of stardom worked to divert the superfluity which could not be contained in the ninety-minute feature.
Star Designing
The institution of stardom directed the superfluity of costume (and the crasser extremes of commerce) away from the film itself and then redirected interest back to it. As I see stardom in relation to costume in the 1920s through the 1950s, the institution establishes a marketable entity which could carry the extra charge of glamour and the overload of emotion. Building on the work of Richard Dyer (who originally theorized the star actor and the sum of his or her star vehicles as a continuous text), I am suggesting that there was an easy carry over of the costume excesses from film to star.47 The "clothes horse" stars, Joan Crawford and Kay Francis, were thus produced as style-setting tangents to their star vehicles. A fan could even follow these stars or the lesser fashion plate figures—Claudette I Colbert, Loretta Young, and Delores Del Rio—without ever seeing them in their motion picture roles, for they were each constructed as much by publicity releases and women's page fashion features as by their
performances. Looking at the phenomenon of high visibility celebrity, semiotic theory sees the constant deferral of meaning in which the multiplicity of determinations keeps the star semiotically active. Stephen Heath's theorization of the "shifting circulation" between narrative agent, character, person, and image, in which no term "settles" or exclusively determines the contents of stardom, explains why commonsense knowledge finds it so difficult to reduce the star to a final key to the meaning of contemporary society.48 But an understanding of stardom as circulating also explains why different interests have been able to exploit the star from so many angles while still maintaining the fiction of a person and the requisite single point of coherence.
The companion institution and ideology of the star designer—the artist genius who invented the actress much as the Svengali-likedirector—was instrumental; he helped to formulate a point of coherence and to channel the costume excess into the star as a marketable entity. I say "he" because the Svengali model developed in the classic cases (Travis Banton-Marlene Dietrich, Orry-Kelly-Bette Davis, Gilbert Adrian-Joan Crawford and Greta Garbo) remained the dominant one. Edith Head, in contrast, is not seen as having created Dorothy Lamour, Barbara Stanwyck, or Elizabeth Taylor in the same way, and although Mae West would design her own image, it still was executed in costume terms by Travis Banton and Edith Head. _^
What I would argue is that the star designer had a particular role in the maintenance of two mutually supportive monopolies—the studio's exclusive control of the star property and the star's exclusive right to the construction which he or she claimed as a "self." While the fact that the studio exercised complete control over the action during this period has been well established, the idea of the star's control of the self as a "personal monopoly" comes from Barry King's recent work on the star actor. As King analyzes the economics of the star system, the "personal monopoly" is a response to an oversupply of actors and significant here is the development of an inextricable link between the person and the screen image. This unity, King says, depends on the creation of an off-screen life or public "conduct" predicated on the existential connection between person and image.49 King replaces Richard Dyer's concept of "star image" in his theory with the notion of persona (the person and image united), which has the advantage of positing this synthetic amalgam as yet another character/^
The persona is the marketable character which takes material form in~l the star designer's costuming; the work of the designer was especially important here, for he created the overriding persona through the medium i of the star's personal wardrobe, one which was a continuation of the \ costumes the designer made for the star's characters. The earliest view ofj costuming, the silent film dictum that motion picture costumes for "true-
200 / Costume and Narrative
Jane Gaines I 201
to-life" characters should be no different from street clothes was then reversed. The off-screen wardrobe now had to be similar to the on-screen costume in its exaggerated qualities and had to carry over a definitive style, testifying that the woman who wore the off- as well as the on-screen clothes by the star designer was, in fact, the same individual. In the service of the star's personal monopoly, personality (as expressed through dress) takes on a new importance as the guarantee of the real person in the role— the insurance that the off-screen and on-screen characters are the same.
In the era of the studio designer, costuming no longer "brings out" traits in the actress of its own accord (as Clara Kimball Young described the actress-costume relation in silent film). Star designing effects thesynthesis between character and actress, intermingling "traits in such a way, that TfieTwo become indistinguishable. By way of example, let me quote a representative description of the star designing process as explained to the fan magazine reader:
First: Take stock, not of your body, but of your mind! That is what Adrian does to each star with whom he works. Even before attempting sketches he talks with her, gets her slant on life. Not, if you please, the life she totes out for the benefit of her public, but her private one which is the key to her personality. He tries to feel what is really going on inside that beautiful head of hers. Then, if he designs her clothes for private life, they are the expression of that, and that alone. For picture purposes, of course, Adrian must also determine how the character she is to portray thinks, and then blend the two—for that is how she will appear on thescreen.5'
To illustrate this, Adrian explains his conception of the character in Queen Christina (1933) as a "blend" of Garbo's "originality" and the real queen's "cleverness."52
But why the insistence that the "truth" of the self is a product of the mind rather than the body? For one thing, this insistence affirms the need for the designer, the expert who will interpret the person and whose costume diagnosis will "bring out" hidden traits. This emphasis on personality over body also stresses uniqueness and individuation which serves to divert our attention from the way the star always functions as a type, one of the most important assertions of Richard Dyer's work on star acting.53 Furthermore, to add to Dyer, the notion of a real personality as opposed to a body type masks the way in which star image/persona is often a kind of caricature or emblem derived from the actress's underlying skeletal structure. The characterization of Garbo as having the "quality of old world repose," for instance, is an articulation of the famous slouch,
and Norma Shearer, who was physically short and compact, became known for her "tailored personality," which combined her fashion diagnosis based on body type with an idea of a real self.54 The irony of the use of costuming to demark personalities is that at the same time the artist designer individuated actresses, he created for an ideal impossible body. Adrian's sketch of a gown for Norma Shearer in Escape (1940) is the long-limbed, slim-hipped, broad-shouldered shape which fashion myth holds he derived from Joan Crawford's body (Figure 10.9). Actually, Adrian sketched this same 1940s silhouette for every model for nearly thirty years.
Star designing as it supports star acting, then, creates a personal monopoly as it articulates a quickly comprehensible character, the entity which includes person and image and guarantees that an individual with a distinct personality stands behind this public character. Since star acting as a mode of performance often appears to be self-evident because of its prevalence, let me suggest how it compares with another model of performance, and, finally, how star designing as opposed to other costume traditions and practices serves particular economic interests. Pertinent here is the principle Richard Dyer has called the star-character "fit," a principle which has undergone some refinement in Barry King's analysis of star acting.55 King makes a distinction between the theatrical ideal of impersonation (in which the actor's performance so successfully foregrounds the role that the actor "disappears" into the character) and star acting (which is effectively personification). Schools of acting (and audiences who subscribe to the notion of believability) value impersonation, the performance in which the actor is effaced. These naturalistic schools hold that physiogonomy as well as technique must evaporate on stage. Personification, held in less esteem because it is popularly understood as an actor "playing himself or herself," is based on the coincidence of character and person and thought to require no "acting" in the traditional histrionic sense.55
Although star designing is thoroughly committed to foregrounding the star, the value of impersonation over personificationhas still held. The ideal in this way becomes the transformation of the star that plays on a fascination with masquerade while remaining a transformation that stops short of complete disguise. Edith Head expresses this contradictory ideal in her comment that what the designer strives for is to "shock" the viewer into saying, "I don't believe that's Grace Kelly after all."57 The job of the star designer, then, is to be sure that the actress does not completely disappear into the character, for the eclipse of the star in the film is akin to marketing a consumer good without prominent placement of the manufacturer's trademark. Something of the way the star design functioned historically to secure the monopoly the studio held on the star
product is suggested in Jack Warner's worry that Paul Muni was so convincing in his role in Juarez (1939), that audiences would not recognize the expensive Warner Brothers property.58 Costume as total disguise, as a misleading indicator of identity, as remnant of an older theatrical tradition, of carnival and ritual, is also incompatible with the narrative rule of coherence. As tangent to the film, then, the star carries off the artifice and the extravagance where it can be relished and reclaimed by female fans,
Jane Gaines I 203
costume collectors, and her gay camp following, thus recycling the excess which returns these enthusiasts to the films for years later.59
Costume and Melos
Finally I want to suggest what happens in some genres during particular periods in cinema history when costume has not been reined in,and where the relative sartorial effulgence suggests a larger connotative investment in this layer of mise- en-scene. This is the case with the melodramas produced by the major studios in the 1920s through the 1950s,'and I"would argue that the work of the designers on these films can be read as the most fully de^elqr^ Aetftric of motion p^cjtoreajrtume. The costumes in many of the films which have come to constitute the canon for melodrama studies are the work of the legendary design talent who, with the studio resources behind them, saw their wildest visions and most outrageous whims made into clothes.
To give only a few examples, Now Voyager (1942) and Dark Victory (1939), as well as the majority of the Bette Davis star vehicles at Warner Brothers, were designed by Orry-Kelly. Travis Banton, Paramount's star designer, discussed by Gaylyn Studlar within this collection in relation to his costumes for Marlene Dietrich, also designed Imitation Of Life (1934) and Letter From An Unknown Woman (1948). Howard Greer and Walter Plunkett (best remembered for his work on Gone With The Wind 1939), both designed costumes for Christopher Strong (1934), and Jean Louis (recalled for Rita Hayworth's Gilda dress) was responsible for Lana Turner's wardrobe in Imitation Of Life (1958). At MGM, Gilbert Adrian designed Camille (1936) and Anna Karenina (1935), and later, after he had left to start his own couture business, Warner Brothers hired him to create some of Joan Crawford's costumes for Humoresque (1946) and Possessed (1947).
Although feminist critics have often remarked about the costuming in these women's pictures, the work on costume and melos still lags behind the work on musical scoring in the genre. The comparison with music is an important one, and not only because music has also been seen as a subordinate system in the classical narrative scheme. Like the musical code, the vestural code has a basic typifying function as well as an elaborative function, the later accessing the realm of emotion and compensating for the expressive deficiencies of the dialogue.60 At the basic level, musical motif and costume motif alike borrow popular conceptions to locate types. For instance, the wide apron identifies the mammy, the feather boa the floozie, and the turban with bananas, the Latin American rhumba dancer—visual shorthand which depends, like musical typage,
204 I Costume and Narrative
i on ideological premises lodged in this iconography/1 It is this "type" costuming which best exemplifies the ideological dimension of naturalization as I have referred to it, that is, costuming which tends to disappear b as it confirms commonsense notions. And corollary to this, costuming which exceeds the typifying function can constitute a threat to the narrative—like the virtuoso composition which suddenly calls attention to itself in a scene and is heard as music.
For the most part, in Hollywood films of this period, generic convention keeps costume as spectacle in check by motivating it as show business in the musical, or as visual pun, parody, or humorous incongruity in comedy (as in Bernard Newman's designs for Theodora Goes Wild [1936], Travis Banton's for My Man Godfrey [19361, and Adrian's for Bombshell [1935] and The Women [1939]). But classic Hollywood domestic melodrama offers a somewhat different case in point. On the one hand, plotwise and stylistically, it depends on the extraordinary; on the other, it consistently keeps one fooTiti everyday life (as signified by its verisimilitudinous mise-en-scené)."The'costuming extremes that I am intěrestěďiřThere escape from the^strict realism of contemporary dress. Often they eschew historical
> realism in the representation of period clothes. Sometimes they even exceed the realism of social class, that is, they exhibit a wealth beyond the means of the character. These, then, are the residues that represent
* the "much too much" sensibility which characterizes the genre as a whole.
* Melodrama, as Peter Brooks has theorized it, is characterized by a "rhetorical excess," a hyperbole which exceeds verbal language.62 Like the nineteenth-century novel, melodrama tries to convey its meanings through the conventions at hand. But the realist aesthetic is never sufficient to melodrama's project, a project which requires vehicles that can express the grandiose and the profuse. In film melodrama, as Thomas Elsaesser discusses it, the vehicles of melodramatic rhetoric are those aspects of mise-en-scene which verge on the non-representational—gesture, lighting, camera movement, decor, and costume. Likening the excesses of visual style in silent film and Hollywood 1950s melodramas to the musical counterpoint of the original organ grinder who played a "commentary" on theatrical action, Elsaesser maps out the register of the melodramatic for film theory, suggesting that what is inexpressible in the narrative overflows into the more absorbant, purely aesthetic vehicles where it assumes an antithetical relation to the action. As Elsaesser describes the 1950s Technicolor melodramas, emotion is exteriorized in the lush mise-en-scene, almost as though the characters are turned inside out and their interiority displayed (in coded form) in the decor. Here, his analogy between the text and the psychoanalytic patient emphasizes the involuntary, unauthorized, and disjunctive aspect of melodramatic rhetoric: It is always a symptom of something elsewhere. Elsaesser's discussion of the mise-en-
Jane Gaines I 205
scene of Written On The Wind (1956) provides one starting point for seeing costume as melodramatic rhetoric, particularly in the women's film which, in the period I am concerned with here, tends toward an aesthetic luxuriance which sometimes matches the emotional opulence. Like the passions in these films, the costuming is unrestrained and relatively indulgent. Elsaesser singles out the black satin bow the wind detatches from the funereal wreath at the end of Written On The Wind, which, it seems to him, carries emotional qualities.63 What is significant about this moment is that the bow, in its detachment from the wreath, is relieved of its representational obligations, and in this form it seems to carry the strongest charge. It is this kind of emotional charge which may be carried by the lavish costuming in the women's film.
However, designers as well as directors in this period also adhered to a code which stipulated an inverse relationship between costume design and emotionality; in other words, the actress should be dressed down for the high emotional scenes and dressed up for the less significant moments. And yet.in the women's film, the costume for the "big" scene, in terms used by designers, had to be "important," especially if the occasion called for formality. In my analysis, we need to make a distinction here between textural extravagance (and, later, color) and design extravagance, often referred to as "style." As we shall see, each of these aesthetics could be troublesome in its own way, but of the two, textural density could be seen as working somewhat like music in its function as affective supplement to a scene. For on the bodies of the female heroines, such fabrics as lamé, silk velvet, duchesse satin and chiffon, simulate skin and thus seem to render tangible an emotional hypersensitivity.
My distinction, however, is still imperfect since texture and style also work together in the more lavish melodramas to stimulate the viewer's visual appetite for a crescendo of opulence as well as emotion. Part of the the pleasure of watching these films is in the prolongation of this excitation in the surprise aspect of the designing which carries over the fascination into transitional scenes. In these superfluous scenes the heroine may do nothing more than answer the telephone or pen a note, but she carries out this mundane task in the most visually stunning and complex costume featured in the entire film. And here lies the danger. The costume plot organizes an idiolect with its own motifs, variations, surprises, anticipations, and resolutions which unfold in a temporality which does not correspond with narrative developments, whose climaxes occur in alternation with key dramatic scenes, in the undramatic moments.
Here, also, is where the comparison with music must end. Much more, of course, needs to be done with the contrast between costume idiolects and musical structures as carriers of affect, and between the eye and the ear as receivers, perhaps following E.H. Gombrich's suggestion that while
206 I Costume and Narrative
the "eardrum vibrates sympathetically" with instrumental music, "light energies [impinge] on the rods and cones" of the eye.64 Our clue to the hyperactivity of the eye is the costumer's postulate that design extravagance (style) is in competition with emotional content since such artistry encourages a kind of visual intellection at odds with delicate nuances of feeling. The basic principles of dress design adhered to in motion picture costuming during this period relate style to the roving eye which could be rerouted on the body. The eye may be directed inward, for instance, so that it is "entrapped," or upward so that it "travels" beyond the figure.65 In the heyday of the studio designer, from about 1929 to 1940, one can see numerous examples of activity for the eye which was routed around the bodies of actresses in unpredictable directions. The mark of "style" for the virtuoso designers literalizes Barthes's understanding of the "aberrant message which 'surprises' the code," and as exemplified by Adrian's work during this period, it is both startling and deviant.66 By stylistic deviance I mean oversized buttons, misplaced pockets, asymmetrical collars, and unconventional use of fabrics. Adrian himself described his characteristic exaggeration as "tabloid fashioning," the design which involved "keeping to the point, yet embroidering the facts."6 This, then, is the kind of "eye-catching" surprise which George Cukor and Alfred Hitchcock saw as endangering subtle moments in a scene.68
Whereas stylistic eccentricity is thought to be the potential ruin of a scene, texture and color may be more easily shaped to fit narrative priorities. Design which maps out the proairetic in visual terms as, for instance, Orry-Kelly's costuming for Bette Davis in Juarez (1939) (in which the gradation from white to black parallels the mental disintegration of the Empress Carlotta), achieves the kind of harmony with narrative goals for which directors strive. Costuming, however, cannot anticipate narrative developments so closely that it gives away the plot. The heroine cannot wear all-black before the tragedy.
The questions of narrative suspense and affectivity merge in the problem of how to costume the heroine for the scene in which she kills her lover. Some of the costuming solutions to this problem that I have considered suggest that the intensity of the scene requires the banishment of style but the presence of some other strong costuming statement that would reiterate the intensity of the passion and echo the fear. One solution is to strip the heroine down todark shades with no accessories or detailing (as seen in the simple dress Joan Crawford wears when she shoots her ex-lover [Van Heflin] in Possessed [1947]). In Mandalay (1934), Orry-Kelly has Kay Francis wear a thin pale-colored shimmering satin robe to watch her ex-lover drink from the glass she has filled with poison. Orry-Kelly solves a more difficult problem in The Letter (1940) in which Bette Davis/Leslie Crosbie shoots her lover a few seconds after the opening credits. Here,
Jane Gaines I 207
Leslie's guilt or innocence hinges upon the question of her relationship to the dead man. If the viewer is watching closely she may note that Davis kills the man who has betrayed her in a full-length black-checked chiffon with diaphonous sleeves and a strapless black underdress (a cross between an evening gown and a lounging robe), and that she quickly changes into a white blouse and tailored skirt before her husband and the investigators arrive. Considered retrospectively, the first costume confirms the character's guilt, establishes the murder motive as jealous passion, and even hints (in the wide white collar) at the prudishness Leslie hides behind (Figure 10.10).
What is similar in these costumes is that, in comparison with the other costumes the heroines wear in the same films, they have relatively simple lines and virtually no decoration, but do feature what we might call empathetic textures. But then, there are significant exceptions to the conventions of costuming the scene of passion, ones which suggest that melodrama encourages rule-breaking. Banished as style, the overblown comes back as texture. Here I am thinking of the scene in Letty Lynton (1932) where Adrian puts Joan Crawford in a two-piece silver lame suit dress to poison the Nils Ascher character (Figure 10.11). The costume works to divert our expectations, since the enigma posed is whether Letty will drink the poison herself or whether she will leave it for him. The costume accommodates both the daring stubbornness of the sophisticated "society girl" determined to have her own way with the villain and the stunning ruthlessness of the murderess into which she must be transformed without the aid of a costume change. Relative to the other costumes Crawford wears in this film, the lame suit is not the most spectacular design—indeed, it does not compete with the puffy-sleeved white mouse-lline de soi which the film made famous.6" But the textural rigidity (heart-lessness) of the silver lame fabric, overwhelming even the design features (the asymmetrical cut of the cape-like collar as well as the peplum) is a visual "knockout" in its own way. While some directors might see such visual brilliance as undercutting the scene, to me it is one of a few cases in which the connotative charge in the one system is at least equal to that of the other.
The problem of costuming scenes of passion recalls Edith Head's commentary in The Costume Designer quoted at the beginning of this article. As the sign of emotion most eloquent and subtle at the same time, tears are easily overwhelmed by sequins, dismissed by chintz, or upstaged by white fox, hence Head's decision to soften her character's outfit. In the woman's films I am concerned with here, however, this principle works somewhat differently, since such melodrama stars as Bette Davis and Joan Crawford are not known for displaying tears at all, hence the institution of the "surrogate sufferer," the friend or relative who assumes the burden
208 I Costume and Narrative
Figure 10.10   The Letter (1940). Bette Davis shoots her lover in Orry-Kelly's black-checked chiffon. Warner Brothers, courtesy of BFI Stills Archives.
of feeling.70 The empathetic costuming of the woman's film heroine in the depths of despondency works somewhat like the surrogate sufferer device. Richness of feeling deserves enriched texture, and velvet, wool jersey, chiffon, satin, bugle-beading, or sable are often used on the bodies of these heroines. These fabrics seem to capture and hold the pathos before our eyes.
But the crucial point here is that in order to capture and hold, the camera has to linger. And yet, as I have said, it will not be waylaid by costume
Figure 10.11 Lelty Lynton (1932). Joan Crawford poisons her lover wearing Adrian's silver lame suit. MGM, courtesy of the Museum of Modern Art Stills Archives.
detail. The moments when the costuming eludes meaning, when affect becomes ascendant, then, are those moments in which the camera is held on the heroine for some other dramatic reason. The costume detail is trapped in the frame! Two examples come to mind,both of them cases of costuming which is connotatively rich enough to match the volume of suffering. One recalls Orry-Kelly's fox hat and matching bolero-length jacket worn by Bette Davis as Judith Traherne in Dark Victory (1939)
T
210 I Costume and Narrative
Jane Gaines I 211
black velvet hat, cocked over one eye as seen in the last frames of Anna Karenina. The artificial cherries, magnified in the close-up over Garbo's lowered eyes (stunned by the train wheels in her contemplation of suicide), are empathetic in their drooping shape, and disconcertingly sensual. The presence of these cherries is strange and inexplicable, the more strange and inexplicable the longer one thinks about it. Is this what Barfhes means by "inexpressibility" and "obtuseness" in his essay "The Third Meaning" (which has clearly informed my analysis of these scenes)?72
Yes and no. One needs to recall that Barfhes derives his theory of the third meaning from scrutinizing frame enlargements from Eisenstein's Ivan The Terrible, not from watching women's pictures. The notion of cinematic excess derived from the Barthes essay and the notion of melodramatic excess have yet to meet and establish some ground rules. While the costume idiolect does threaten to cohere into a different temporality and a different film, this other (more wonderful) film is not completely "indifferent" to story (as is Barthes's third meaning), although it may appear at an obtuse angle to the narrative.73 My preliminary findings suggest that in terms of costume surplus, film melodrama can absorb and motivate beyond the capacity of other genres.74 This may call into question the use of the concept of narrative unity without reference to genre in relation to a notion of cinematic excess. Here, then, is the question: Does the woman's film narrative so thoroughly motivate costume that all apparent excesses (even style extravagance) are not excessive within the film, but only in relation to films in other genres? Is this why, as we watch these films, we are so often left unsatisfied in our craving for costuming which crescendos to meet the heights of the passions the films strive to dramatize?
Figure 10.12 Dark Victory (1939). Bette Davis's suffering exteriorized in Orry-Kelley's fox hat and matching jacket. Warner Brothers, courtesy of Wisconsin State Historical Society Stills Archives.
(Figure 10.12). This is the outfit she is wearing for the close-ups in three crucial scenes, the first of which is the scene in which she reads her fate in the words "Prognosis Negative," highlighted on the pages of Dr. Steele's medical file. But what is it about the fur fringe creeping into Bette Davis's eyes? And the incongruous mass of it on her shoulders? How can an image be horrible and beautiful at once, unless what we are looking at is the image of her glorious suffering intermingled with the disease itself?71 One also thinks of the cherries on Greta Garbo's Empress Eugenie-style