Lady Be Good
Do dogs dance?

by Jane Gaines

from Jump Cut, no. 31, March 1986, pp. 19-23
copyright Jump Cut: A Review of Contemporary Media, 1986, 2006

"One can hardly conceive of a world in which certain beings communicate without verbal language, restricting themselves to gestures, objects, unshaped sounds, tunes, or tap dancing; but it is equally hard to conceive of a world in which certain beings only utter words."[1]
— Umberto Eco, A Theory of Semiotics

[open notes in new window]

Dancer Marilyn Marsh (Eleanor Powell) kicks back the rugs in her high rise apartment living room, switches on the record player, and runs through a routine with her pet dog Buttons to a jazz rendition of the film title song, "Lady Be Good." In the culmination of this highly acrobatic interplay, Marilyn cartwheels over the dog [Figure 1]. Buttons jumps through Marilyn's arms [Figure 2], over her head, and then over her entire body as she drops into the splits [Figure 3]. The dog's last trick is a leap from the couch across the room into Marilyn's arms [Figure 4]. As she catches the dog, she falls backward onto a sofa bed [Figure 5], holding it as it licks her face [Figure 6]. Then, with her back to the camera, she rolls over. [see images #1-6]

A few years ago, I was part of a group which undertook a collective analysis of LADY BE GOOD (MGM, 1941) in an advanced film theory course. The Marilyn-Buttons novelty act described above attracted interest among the students who generally agreed that something additional was going on in the sequence; in fun, some suggested far-fetched readings. Then I was intrigued with the possibility that I could locate the elements which were creating the hint of perversity if I looked at the sequence enough times on the horizontal editor. Later I became more interested in the tools available to critics who want to demystify the way meaning is produced in aesthetic objects. The following is an introduction to semiotics, which both illustrates the capabilities of Umberto Eco's theory of codes and tests its limits. As my example I have taken this very short scene from LADY BE GOOD, a musical which imagines the euphoria of its title song's own popular and financial success through a vicarious vehicle — a story about a song that gets to be a hit tune.

The film's narrative concerns a man and woman songwriting team, Dixie Donegan (Ann Sothern) and Eddie Crane (Robert Young), whose first marriage coincides with the beginning of their career as collaborators. Donegan and Crane's complementary songwriting talents, however, translate into marriage incompatibility. (He aspires to the high art concert stage and she wants to write popular lyrics.) They are divorced. Together again for an evening, they write the song "Lady Be Good," which becomes their biggest success. On the wave of the acclaim, they marry for a second time, then separate. Dixie seeks another divorce, but a fatherly judge refuses it; his determination is that they still love each other after all. Eventually I will show how the Marilyn-Buttons sequence fits into this Arthur Freed-produced musical celebrating love, marriage, success, and hit tunes. First, I want to indicate how some basic semiotic terms might be used in film criticism while demonstrating how semiotics forces us to think more exactly about the viewer's cultural competence. The succeeding sections are meant to suggest that much more cultural information is relayed by the cinema than critics ordinarily acknowledge.


As defined by Umberto Eco in A Theory of Semiotics, the code is the system which makes it possible" for cultural units to mean.[2] Benjamin Whorf's "prisoners of culture," who are busy using one thing to mean another in social intercourse, would be surprised to learn that there was anything systematic about telling a smile from a smirk, a jazz tune from the blues, or an outmoded handbag from a stylish purse. To the user, the code is always invisible. So the question arises, "Why would semioticians want to study what people already know?" The answer has to do with the function of the semiotician who begins from a message or symbolic act and works backward to the code. Sign-system or code, as introduced here, is a tool used by the semiotician to reconstruct this social competence.[3]

Adult moviegoers, the culturally competent who take their complex communication skills for granted, would, for instance, already know codes of gesture, music, and dress, which they would have internalized years before arriving at the theater for an evening of entertainment. How viewers "know" the various codes found in combination in the Hollywood musical film, is, then, somewhat like the way in which speakers "know" their native language. The linguistic code/cultural code comparison may seem forced to the student new to general semiotics because cultural codes are relatively loose as compared to linguistic codes.[4]

How do audiences use cultural codes to "read" the image "dog circling woman" within the context of a 1940s musical?[5] Viewers familiar with U.S. vernacular dance in amateur or professional musical revues and variety shows, or musical comedies on stage or screen, can make meaningful sense of the Marilyn-Buttons interaction because they know a code which, for our purposes, I will call "tap dance." The tap dance "system" is comprised of steps organized according to a principle: Marilyn's steps begin as basic tap (Falling Off a Log, Shuffle Off to Buffalo) and move through leg lifts, splits, spins, and finally cartwheels. Buttons' corresponding movements progress from small jumps to twirling (on his hind legs) to running leaps.

If audiences know the code {tap dance}, they may also see that the partners are not exactly "dancing" together and that whatever they are doing is something beyond basic tap. Their movements are organized according to other principles, or, semiotically, their movements are governed by other codes, as I will show. A young child, familiar with this cultural phenomenon, might be able to decode this sequence as {circus dog act} while knowing that this is not truly a circus performance since the dog and woman are pretending in her living room.

If the woman and dog are "pretending" in her living room, a third code, {pet play}, is operating here. You may be saying that many people play games with their pets without ever having had instruction, and that this is the sort of recreation one would expect to find with pets in the U.S. and Europe. Those things we do which "go without saying," however, are among the most coded aspects of daily life. People do not pretend that animals are little humans in all societies. In those cultures where pets are not taught to do complex tricks for the amusement of the owner, /dog circling woman on its hind legs/ might have other meanings, such as "foolish bourgeois affluence." The point in understanding such a wide ranging and loose activity as code is just this: to point out its cultural specificity.

Let us say, then, that moviegoers in this culture read /dog circling woman on its hind legs/ by means of the codes {tap dance}, {circus act}, and {pet play}. Did the filmmakers who produced the sequence have in mind one or more of these codes, or, what did the filmmakers mean to represent? Semiotic theory structures this problem in terms borrowed from communication theory: messages are encoded and decoded by senders and receivers. One of the problems in adapting this communication model for cultural studies is that it still evokes the image of a single channel and a wireless transmitter.

In cinema, there is no immediate two-way exchange between stations or parties. In fact, few send, and hundreds of thousands receive, but these receivers do not return messages in the same way.[6] Furthermore, Hollywood cinema images are industrially produced rather than individually created. Studio personnel (director, set designer, hair stylist, director of cinematography, producer, and choreographer) all serve as encoders. If we see meaning in commercial cinema as in flux, we should also consider the fact that studio publicists help produce meaning — from the shooting through exhibition. To promote LADY BE GOOD, the MGM publicity department sent suggestions inspired by the Marilyn-Buttons number to exhibitors to use in picture exploitation. As a lead into possible exhibitor tie-ups with local pet stores, Buttons' training was featured in press book material which recommended encouraging children to teach their dogs to do tricks.[7]

From this evidence, we can conclude that of the creative personnel who worked on LADY BE GOOD, the publicists, at least, did intend the sequence to be read as pet play in addition to tap dance. How can it be both at once? Again, the channel/ signal/ transmission model does not adequately represent the density of meaning in visual arts or literature. Properly, these forms have multiple meanings, both concurrent and sequential. A cinematic, literary, or pictoral message, following Eco, is best considered as a text.[8] Text, as a concept, accommodates levels of meaning and overlapping codes. I will show that it explains the coexistence of tap dance, pet play, circus act, and even additional codes. The complexity of communication on the encoding side is theorized, then, by the concept of text, but on the decoding side, how do we consider the access points at which text and viewers come together? Also, what are the limits to the numbers of meanings audiences might "read out"?

Eco deals with multiple readings in his concept of "connotative paths." One message may be decoded several different ways although the sender may have intended a single meaning. Denotatively, the message may carry the sender's meaning, but receivers may also take various directions in decoding. The work of art is semiotically characterized by the number of possible directions the reader may take. Eco says here:

"Like a large labyrinthine garden, a work of art permits one to take many different routes, whose number is increased by the criss-cross of paths."[9]

Further, this special semiotic case, the aesthetic text, whose codes are organized ambiguously, gives the impression of semiotic motion in the way it "continually transforms its denotations into new connotations."[10]


An example of this shift from denotation to new connotation is the way Marilyn's leg signifies <dog hurdle>. The denotative meaning of her extended appendage is certainly <leg>, but /leg/ also stands for hurdle, or, semiotically, /leg/ is a sign connoting <hurdle> [Figure 7]. Eco's definition of a sign which includes "everything that, on the grounds of a previously established social convention, can be taken as something standing for something else," incorporates this "standing-for" function and the notion of social agreement originally from Saussure.[11] If culture members agree, unrelated cultural units (woman's leg and dog hurdle) can be brought together in equation and held there by means of a signification system. The leg=hurdle equation poses no comprehension problems to viewers here because the circus dog act code "correlates" these units.[12] If leg=dance step in addition, this is because, as Eco shows, more than one code will often "rule" the same sign.[13]

The codes ruling the sign /leg/ within the sequence also fall into two large categories in Eco's theory. Those identifiable objects and figures commonly called images are iconic signs, which he distinguishes from kinesic signs, denoting movement and gesture. For instance, as it connotes <hurdle>, /leg/ is part of a system of circus props which might include couches and a coffee table standing for <risers>. Marilyn's extended leg is not only an iconic prop, but within kinesic systems ranging from the disciplined tap dance to casual /kicking back the rug,/ the moving leg signifies <kick>. I will return to the special problems of both iconic and kinesic signs and the relationship between the two in a later section.

Closer analysis of the signs /kick/, /cartwheel/ and /splits/, suggests that they belong to yet another system which might be represented on an imaginary semiotic grid cutting diagonally across the first three codes: tap dance, pet play, circus act. Note that {work out}, here in the sense of rehearsal session, depends on the other codes, {tap dance} and {circus act}. Still other codes depend on the {work out} code, and by reference to them, the viewer would know this is informal practice and not formal rehearsal. For instance, {improvisation}, or {ad lib}, as system, organizes and makes one particular sense of /hand clap/ as well as /cartwheel/, /splits/, and Marilyn's /hula/. These dependent codes, which Eco designates as connotative codes or subcodes, are those descriptive clusters which are brought out in conventional criticism and reviews. For instance, critics' recognition of period, authorial, or visual style is at the connotative level.[14] One of the uses of semiotics is to show that meanings are not mysteriously drawn from the aesthetic text by an insightful reader as "out of thin air." They are concretely carried out by sign-vehicles carrying connotations which in turn are sign-vehicles for other connotations.

In this respect, the difference between the ordinary moviegoer and a critic or connoisseur is that the latter have access to rare or specialized codes which enable them to notice correspondences; to "appreciate" an art form on more levels.[15] An expert in the history of Afro-American dance, for instance, would know that the Marilyn-Buttons routine contains elements of the flash act, a variant of tap which combines jazz dance with ad libbed acrobatics. This specialist might quickly see Marilyn's act as echoing the Berry brothers who were featured in two specialty numbers in LADY BE GOOD [Figure 8]. Because of their acrobatic ingenuity and skill, typified by their famous splits, the Black trio became known as an act with more flash than tap.[16] Flash here is not only a rare code but also a subcode depending upon the {tap dance} code.


Although one might see the Marilyn-Buttons virtuouso flash act as connoting the circus side show and vaudeville theatre origins of the film institution, this sophisticated interpretation would not be possible if the viewer did not at least read the smaller figure on the screen as denoting <dog>. How does the viewer know that the figure is not hyena, or monkey, or seal? /Dog/ seems the most elementary of all the images in the sequence. For that reason I will use it to raise some of the more difficult issues, which the theorist confronts every time it is necessary to say what is meant by the iconic sign. Earlier work in semiotics used iconic sign in opposition to verbal sign to emphasize the conventionalization basic to verbal language, in which words stand for things so arbitrarily. In contrast, iconic signs, as seen in the visual arts, were said to "share some of the same properties" with the real world objects to which they refer. Subsequently, this special analogical relationship between sign and referent has been challenged by the evidence that recognizing photographic "likeness" requires learning. A new possibility has arisen — that the iconic sign, sharing none of the properties of its object — is also arbitrary, completely conventional. Although this view of the iconic sign would strengthen Eco's argument that what appears to be nature is actually culture, he has rejected this extreme position because it leaves no opening for any account of perceived similarity.[17]

Following Eco's argument, the motion picture image of a spotted dog and the dog Spot tied up in your "real world" backyard, share no physical properties. In the real dog, spots are the effect of dark patches of hair that absorb light against the white of the areas that reflect light. Photographically, spots are the result of concentrations of silver nitrate crystals. To continue this reasoning, one can further compare the properties of the celluloid strip and the familiar damp, lumpy furriness of any dog.[18] If the motion picture image of the dog and the real-world dog share no physical properties, what explains the perception of sameness? Eco says that the same conditions or mechanisms of perception function here, but that they rely on two different "perceptual supports."[19] Cultural learning, or the repeated experience of seeing a dog as compared to the image of a dog, develops expectations; these constitute another kind of code governing perception. While Eco does not deny that there may be a "natural link" of some sort between image and real world referent, that linkage is still subject to conventionalization.[20] As he emphasizes perceptual learning, he can retain a theory of the cultural coding of the iconic sign.


Semiotics counters the idea that stars are real people, showing instead how star image and popular type, quickly comprehensible public selves, are constructed. Recent film theory considers star image as a text, with its fluctuating meaning distributed across a diverse range of media from the star performance itself to studio publicity, newspaper coverage, radio interviews, public appearances, and Hollywood gossip.[21] In 1941, Eleanor Powell was considered the most accomplished tap dancer on stage or in motion pictures, and producers consistently cast her in films to feature her specialty. An Eleanor Powell fan would know that the star was so proficient that MGM had trouble finding her a dancing partner. She was featured with a horse who raced to the opera Figaro in BROADWAY MELODY OF 1938, then matched with Fred Astaire in BROADWAY MELODY OF 1940, but producers decided that Powell and Astaire did not make a team.[22] Was she too singular in her style to meld with Astaire's? Critics in the 1940s attributed Powell's tap virtuosity to a technique more male than female, and Astaire recalled in his biography: "She put 'em down like a man, no ricky-ticky-sissy stuff with Ellie."[23] Her publicly constructed "go-it-alone" image intersects with the Marilyn Marsh career-girl-dancer role in LADY BE GOOD and contributes to the ambiguities which invite such imaginative decoding of the dance sequence.

The <masculine> connotations in Powell's dance style combine with her long-legged physique to signify the <swell gal> or <sister> type. How would the 1940s audience understand Powell as more <boyish> than <girlish>? One way they would understand any star is by reference to a code of types, which classifies popular heroes and heroines according to personality traits and physical attributes.[24] But body type and facial features carrying type connotations proclaim themselves to be natural, which puts them beyond the reach of culture, it would seem. How, then, can we say that a star is "constructed?" Culture has "borrowed" the natural. Culture constructs physiological ideals out of the racial characteristics exhibited by the beings who people a society. In North American culture, for instance, the preference for "long-stemmed" women is based on available stock. Translated into movie stardom, this preference corresponds with the enthusiasm for leggy stars such as Powell, Ann Miller, and Cyd Charisse.

Interest in female star iconography has brought needed critical attention to the decorative detail. Costume as well as interior design are often overlooked as connotative signifiers because the same signs signify <actual setting> or <appropriate dress> and establish that the fiction stands for specific person and place. Background detail is usually organized by a code of realism. As such, detail clusters help to signify <this is reality>.[25]

In the LADY BE GOOD apartment sequence, for example, such details include furnishings and clothing which connote <real life>. Raised to another level, these become signifiers of <forties contemporary style>. Eleanor Powell wears paisley print lounging pajamas which blouse in the leg like the harem pant. In 1941, lounging pajamas provided fashionable "at home wear." Combined with the sculpted sectional sofa, blonde wood coffee table, and floral print above the fireplace, the pajamas and set signify <contemporary apartment living>. Powell's pompadour hair style, worn in an earlier scene; Ann Sothern's veiled black hat, tilted over one eye; and set designer Cedric Gibbons' white pianos coordinated with white interiors — these also signify <forties modern>. Still another code of bourgeois tastefulness organizes stylistic connotations into connotations such as <lavish>, <plush>, or <overdone>. As I understand Eco, these codes of taste and sensibility are those vacillating rules which coordinate and differentiate styles from one decade to the next. Thus the same signs which signify <ultra-modern> in one era may mean <authentic period piece> or <camp> forty years later.

Cataloguing every item in a film's mise-en-scene may seem an exhaustive task, bearing no relation to the way audiences experience motion pictures as entertainment. Every detail does, however, come into play semiotically, as viewers "take in everything" to make sense of the narrative. However, some signs, for instance those denoting decor, audiences would read only at the level of connotation. Studies of cultural coding demonstrate that some signs carry a thicker cultural overlay than others, which indicates the frequency of their use. On another scale, such density of coding indicates a national preoccupation or ideological emphasis. For instance, relative to other areas, the communicative areas of fashionable adornment and sexuality in North American culture are more "built up" culturally.[26]

We can measure such an ideological concentration in the proliferation of subcodes, which shows up as inflection and ambiguity, whether in iconic caricature or verbal word play. The sexual image of woman, and here I mean screen star, centerfold, or cover girl, evidences a dense cultural build up. As an area of semiotic concentration, the sexual image of woman marks a social zone of combat over who makes meaning. Feminists who have surveyed this field argue that mass media images articulate women's social identities without women's consent. In the 1970s, feminist film critics rediscovered star images from the Golden 30s and 40s with the idea that they could reinvest these images with meaning and reclaim for those images connotations of women's <power>, <command>, and <strength>.

These connotations have been associated with the independent stars and in particular those who had at least one time in a career crossed the sex barrier to dress as a man. Like Marlene Dietrich in BLONDE VENUS and MOROCCO, Katherine Hepburn in SYLVIA SCARLETT, and Greta Garbo in QUEEN CHRISTINA, Eleanor Powell played the woman masquerading as a man. She was often costumed in top hat and tuxedo with tails for her tap dance specials, as seen in the LADY BE GOOD "Fascinating Rhythm" finale, where she dominates an extravaganza featuring, in addition to a singer and the Berry brothers, eighty male dancers, five pianos, and an orchestra [Figure 9].[27] Feminist film critics have claimed this kind of strong female star (coded as masculine to bring out the <tough> and <resilient> connotations in her star image) as their lost lesbian heritage.[28] Yet, if current critical practice considers lesbian and gay messages in film or image as subtext, does this mean that there is a "hidden" meaning encoded by gays working in the industry forty years ago? Is this retrospective reading of a film or a star icon a wishful misreading or does the sexual ambiguity allow and even encourage the lesbian connotations?

Eco's semiotic theory leads away from any idea that meaning is embalmed in an artistic text. His understanding of signification as social allows any textual interpretation to shift with social and historical change. The work of art, he says, can be "adapted" to meet the expressive needs of those who receive its messages, and this adaptation need not ignore or misconstrue the rules governing meaning.[29] In semiotic terms, interpretive reading also involves testing codes and rejecting them, much as we have done here with the proposal that Marilyn and Buttons may be dancing or playing. The rule which the sender used is "extrapolated" from the codes the viewer already knows. Also, the viewer might either try closely to follow the sender's intended message or to investigate other possible messages.[30] This theory of interpretation with its equal emphasis on sender and receiver seems to allow so much "play" with the text that it might accommodate the interpretive license of impressionistic criticism. For example, several contemporary viewers who saw the sequence in LADY BE GOOD told me that they suspected woman and dog were mating. To make such a reading would be to take interpretive license with the text if, in fact, there were no basis for seeing the interaction as a mating ritual.


How do we determine whether a dance is "saying" that a woman is mating with a dog? Could the woman and dog's strutting and preening movements be based on a primal courtship ritual still retained here in a residual shadow? Common sense suggests that human societies have elaborate sex differentiation rituals because male and female homo sapiens are so distinctively and visibly biologically different. Birdwhistell's work in kinesics, the study of motion and gesture, shows that, although a patterned set of movements facilitates the gender identification preliminary to and requisite for species reproduction in humans and animals, there is an inverse relationship between male and female secondary sex distinctions and ritualized gender display. Because humans are not that well-differentiated anatomically, this space creates an "opportunity" for the construction of tertiary, or learned, sexual behavior. This is how we should consider gender display.[31] Costume, cosmetics, gesture, and ritual elaborate upon and magnify masculine-feminine difference.[32] Even more important for our purposes, this space makes room for invention, which in turn can foster social myth. For instance, as the choreography "arranges" a dance which imitates animal gender display, it says in a coy and charming way how much humans seem like animals in this regard, and how different in all species the male is from the female.

The film develops the human/animal analogy primarily by the movements which emphasize the similarity of the partners. I have identified four patterns in the sequence which stress similarity:

  • paralleling, in which the dancers' bodies conform to each other positionally;
  • mimicry, in which the dancers exchange movements and roles, one copying the other;
  • the split, in which the two move apart and back together;
  • and, finally, a beckoning and advancing exchange where the partners alternately signal each other, "Come here."

These patterns set up woman and dog as matched partners. I would argue that the patterns themselves convey this meaning.

A fifth pattern, which I will call the manipulation pattern, accentuates the differences between the partners. In the conventional male-female duo, this pattern becomes visible in the way partners exchange active and passive roles, with the male partner, usually the more active of the two, directing the woman's body movements. An example here would be Fred Astaire's style, described as an initiation and seduction of the female partner."[33] Astaire conveys instruction and his control over his partner by turning her at the waist, raising an arm or leg, and lifting her body over his head.

The Marilyn-Buttons act transgressively uses these conventionalized patterns or choreographic codes of similarity and difference. Here the patterns intended to create a symmetrical effect instead call attention to the partners' asymmetry. The manipulation pattern, finally, is impossible to adapt gracefully because of the anatomical difference between woman and dog. Here it's "spoofed" in the way Marilyn lifts the dog onto the record player and catches him in the air. In parallel steps, Marilyn, posterior slightly raised, "conforms" her body position to that of the dog, and the dog "walks" on his hind legs [Figure 10]. Mimicry demands less precision than paralleling here, and Buttons' tail wagging becomes equated with Marilyn's hula. (We know this because Marilyn says, "Oh, so you can do the hula too?") Anatomical incongruity is used to transgress choreographic codes.

If we were to locate the operations which produce humor in the sequence, we might also look at the way the various codes are transgressed and rules broken. Eco theorizes this kind of code abuse in an aesthetic message. He says it creates potential to "bring the code into question" in such a way that we understand the "possibilities" of the code in new ways."[34] "Calling a code into question" to bring new concepts to light certainly will have an important significance when a message has profound social implications to begin with. If not so "significant," this whimsical dance affords an example of the same phenomenon. The substitution of a dog for a male partner calls our attention to the rules and assumptions we make about couples dancing together — that their bodies should empathetically meld together and that their climactic clinch prefaces the sex act. Finally, for the characters to "dance" acrobatic sideshow tricks, and for the film to athleticize dance is to draw attention to the rules of both circus act and tap dance, if only fragmentarily and fleetingly.

As further argument against looking for echoes of innately known (and hence uncoded) primary gender display in the strutting and preening of woman and dog, I would stress the distance between elemental ritual and satirical mimicry within an entertainment form such as cinema. Even if gesture in the entire body of U.S. cinema were successfully classified according to movement, we would still have recorded only a secondary system, dependant on the primary gestural system. The cinema, says Eco, is a highly stylized "artificial kinesics," best thought of as a "language speaking another preexisting language." When we consider that we are analyzing filmed dance, this becomes yet another generation removed from the primary.[35] Dance is a code, like game, play, and mime, "derived" from the primary system of communication. And Birdwhistell cautions that these derived codes are relatively resistant to systematic analysis of the sort he has undertaken with primary gesture.[36]

While we can't hope to understand primary gestural language by studying cinematic theatrical gesture, we can easily see broad patterns of cultural emphasis in the secondary system's embellishments and condensations. Theories of social construction must analyze the derived and artificial as well as map the linguistic features of the primary. Eco recommends borrowing Birdwhistell's scheme, developed from systematic research on primary gestural codes. I introduce that scheme here as a model of how to see movement as communication.

The gesture which people consider "innate," like the image which they think "resembles" something, presents us with an area seemingly resistant to being analyzed as communication and therefore seemingly impossible to reduce to culture.[37] As with the iconic sign, we are faced with the question of how to break up the object of study, which is in this case the field of filmed action. What shall we designate as the units of meaning? How do we isolate them? Of what size are the signifying chunks of action which cinema viewers "take in"? Although human bodies produce gesture and posture in a continuum rather than in action increments, in perceiving gesture, every society slices the movement continuum in a different way.

Birdwhistell has reduced movement to the kine, that fraction which cultural "informants" recognize. It might be as small as the degree of eyelid closure.[38] Kines in themselves, however, have importance only insofar as they combine meaningfully on another level. This level Birdwhistell calls the kinemorph, the equivalent to what is meant here by the kinesic sign.[39] Kinesics, then, measures action minutely, showing that meaning is communicated through body shift or facial twitch. From kinesic research we know how every slice of action carries some portion of meaning in relation to other bits of action.

This theory of the segmentation of the action continuum is useful to explain why viewers might not see Marilyn's /knee slapping/ or /kicking/ as commands to the dog. Dance is read largely as abstract pattern. Here, the viewers combine as rhythmic design what the dog might respond to as discrete obedience signals. Furthermore, whereas a choreographer would see the basic tap steps (Falling Off a Log and Shuffle Off to Buffalo) or the flash steps (Over the Top and Through the Trenches),[40] an unschooled viewer would pick up patterns connoting "energy" or syntagmatically, "acceleration."

The choreographic steps could be productively studied as kinemorphs, combinable units in a kind of choreographic language understood by dancers. But what about the arcs and spheres and multidimensional textural effects that dance movement creates? Do viewers read dance movement like they read gesture? In some cases, to insist on a distinction between iconic and kinesic signs is not so useful. As Eco describes it, icons must always generate kines.[41] In addition, dancing bodies, their costume style, and physiological type contribute modifying connotations to movement sensations. Richard Dyer, in Entertainment and Utopia, holds that the nonrepresentational forms carrying the expressive contents in cinematic musical numbers are predominantly iconic. These iconic figures do not combine into meaningful units that refer to real world objects. We "recognize" them in another way. The best analogy here is to the language of music whose abstract melodic and rhythmic forms signify definite sensibilities, which can be translated into variations of human emotion.[42]

Movement patterns, then, systematically communicate feelings, although these waves or splashes seem, on first consideration, illusive and erratic, picking us up emotionally in a quite unsystematic manner. Critical studies of dance as design show that there are abstract shapes recurrent in every dance and that forms in themselves have expressive correlations. In this vein, Dyer, in the British Film Institute Study Guide on the Musical, draws from choreographer Doris Humphrey's dance theory to demonstrate the way in which movement design is culturally coded. Every dance, says Humphrey, has alternating symmetrical and asymmetrical design, patterned either into a solo dancers body position or into the partners' interaction. The similar and different body arrangements in the dance analyzed here correspond with this symmetrical and asymmetrical division. They bear out Humphrey's theory that every dance will contain both of these patterns in alternation.

Humphrey further identifies a second basic choreographic division, the oppositional and successional. He links these patterns with emotional expressions. According to Dyer, the oppositional pattern, in which lines are opposed at right angles, contrasts with the curved and fluid successional pattern, which conveys a more gentle, yielding, and usually romantic sense. Forceful oppoitional lines become vehicles for "aggressive energy" and "vitality," according to Humphrey, and this kind of energetic happiness the viewer feels as "exuberant joyousness' and "exultant hope."[43]

Such dichotomies also represent some of the broader cultural patterns or emphases to which I have referred. For instance, Dyer relates the oppositional and successional division to concepts of male and female qualities. He notes that the aggressive oppositional forms are danced by males while the yielding successional forms are reserved for female dancers. Even more pertinently, Dyer finds a happiness/romance dichotomy solidly implanated in the show business tradition to which the U.S. musical belongs.[44] If happiness stands as opposed to romance, this tells us that the culture conceives of some aspects of "happiness" that are distinctly different from those commonly associated with romance, i.e., bliss, rapture, serenity, contentment. These other aspects of happiness are associated with "aggressive energy" and "vitality." They correlate with the oppositional forms in dance, and they refer to, on an ideological level, some particularly North American aspirations, which I will discuss in the concluding sections.

The audience's empathizing with characters and feeling up or down with the rhythm of an entertainment film indicate still another kind of cultural expertise which moviegoers have without awareness — emotion. Response to film's visual and aural rhythms, which we may have experienced as spontaneous actually comes as the result of learned analogy, as I have already argued.[45] The connection between rhythmic intensity and eroticism, for instance, is culturally inculcated.

How does this help us to understand the sensation audiences would have watching Marilyn tap out rhythms and the dog jump to beats corresponding with a snappy orchestral arrangement of the George Gershwin melody? The musical film employs a number of different vehicles to represent rhythm. These include cutting, camera movement, dance, narrative pace, and music, which is both melodic and rhythmic.[46] These vehicles all share in conveying rhythmic pulses, but not always equally. Since Marilyn's tap dancing has the strongest rhythmic pulse in this scene, I have concentrated on it here.

Movements also involve kinetic release, which choreographers measure as the amount of energy compressed in the dynamics of dance. Dyer refers to Margaret D'Houbler's work on dynamics. She classifies dance according to four binds of energy release: swinging, sustained, collapsing, and percussive.[47] Because of its bursting and leaping, and in the case of flash, springing movements, tap dance offers predominantly a percussive release, but the percussive drill-like tapping can be punctuated with more languid shuffles. The loose body and "wound-up" feet of the tap dancer, then, have their own peculiar energy release and this kinetic quality is another dance component which audiences relate to familiar emotions.


LADY BE GOOD is not an adaptation of the Broadway musical, the source of the two Ira and George Gershwin songs the film features. The film uses an original story by Jack McGowan, previously entitled "Feeling Like a Million."[48] This title would have worked as well on almost any Freed Unit musical of the 40s and 50s, usually in that tradition of optimistic self-confidence which Michael Wood has described as offering audiences less a feeling of 'how to succeed" than "how it feels to be succeeding."[49] LADY BE GOOD conveys the very U.S. sense of "feeling like a million dollars" in two specific ways. First, the narrative concerning the Donegan and Crane team makes connections by loose analogy between "a million dollars" and other cultural units such as "hit tune." "Feeling like a million" is like "writing a tune that becomes a popular hit," or "breaking record and sheet music sales," and also like "falling in love and marrying." Second, the film conveys the sensation of "feeling like a million" through music and dance.

The Marilyn-Buttons routine contributes "spontaneity" and "exhilaration" to the overall sense of "feeling like a million." As signifiers, both the dance design and its dynamics are culturally linked to an extremely energetic but self-assured sort of happiness, which U.S. audiences might associate with "being on top" {mastery} while also "feeling at ease with oneself." I would suggest that executing something very difficult with such ease that it looks like child's play is especially a U.S. ideal; tap dancing summarizes this near impossibility. We can add to this the extreme joy felt in doing those things that "take one away" (from the ordinary), which Powell once expressed in her declaration: "I would rather dance than eat!"[50]


Why a woman can dance with a dog in this film has to do with both the spirit of "feeling like a million" and the conditions the musical has established within which extraordinary things can and do happen. Recall that the narrative concerns the way in which a song becomes a national craze. The film offers one explanation for this, and it is not that the tune has been promoted excessively or "plugged to the hilt" in the publicist's jargon. The explanation offered is that the song is irresistible. It just "catches on" by itself. Window washers and "shoe shines" can't help whistling, humming and dancing a few steps to it. It takes people over. They can't help doing things they might not ordinarily do. Taking this a step further, the finale — which was added to feature the second Gershwin song around which LADY BE GOOD was conceived — offers a seemingly broader explanation for why we can't help doing things: "Fascinating Rhythm." In the words of the song,

"Fascinating Rhythm is a pickin' on me … it's got me on the go …"

Of course, the song "Fascinating Rhythm" does not offer any logical explanation as to why people are "carried away." Instead it is a celebrated bromide that ties up the loose ends of the film.

Without going into a detailed analysis of the Busby Berkeley finale, I will just note how it integrates some of the elements introduced in the film. First, there is a resolution of the popular and the classical, represented by the elegant black grand piano on which the curtain rises [Figure 11]. This gives way to the "pop" white pianos and Powell's tap dancing. Second, the finale traces rhythm to its African origins; here Black cultural forms assume a popular, colonized disguise. As signified by dark projections, Black rhythms connote a racist "spooky" effect, rather than rhythmic eroticism. The projections are actually shadows cast against the back stage wall by orchestra members representing the instrumental sounds which carry "that rhythm." These shadows "answer" the images of the Berry brothers, whose silhouettes loom large on the stage curtains. In turn, these Black tap connotations are passed along to Powell, whose tapping feet are seen over the top of the sheet music for "Fascinating Rhythm," a visual refrain of the popular music sales and irresistible rhythm elements [Figure 12].

The finale cuts to the courtroom scene with Dixie explaining to the judge why she wants a divorce; this reveals that the entire film has been in flashback. In the courtroom present, the story of Dixie and Eddie's collaboration and the immense popularity of their efforts then stands as evidence to the judge. He refuses Dixie the divorce and proclaims they should remain married. Because the film makes an analogy between the hit tune's unification of popular and classical and Dixie and Eddie's marriage, each "match" provides evidence for the validity of the other.[54] The love between man and woman is proof of the success of the song and the success of the song "Lady Be Good" proves the rightness of their marriage.

For Marxists, one of the lures of semiotic studies has been the hope of understanding the way ideology is built at the level of daily life, from the bedrock of social knowledge up. Eco's theory promises a systematic approach to "… everything which can be used in order to lie." This is to say that Eco looks at how this bedrock of social wisdom is laid.[52] Some of this potential for seeing how the ideological is structured into culture comes from Eco's theory of code-switching. Here he shows how signs accumulate new equivalences and these come to have the status of "truth."[53]

My study has looked at how a musical film has made rhythm interchangeable with success. Certainly the rhythm = success equation has not been proposed for the first time in this film; it "rings true" from our cultural experience. The idea that "feeling like a million dollars" describes the best state of things gets a "boost" from popular wisdom about money, success, and the inexplicable drive leading to accomplishment. Strangely, "rhythm" in this film's finale is both active and fixating, an impetus for achievement ("You got me on the go …") and a mesmeric fascination ("You can get it from the slap of the big strong bass/Or the moan of the saxophone"). Characteristic of ideology, such "truths" restated in film are limited and contradictory. Surely they cannot account for very much if we try to use them to make sense of the world.

The notion of cultural coding as it has evolved in mass culture studies still provides only an imprecise critical tool. Because we need to establish general operational communicational rules applying to such various systems, we may sometimes force similarities between systems.[54] Some systems — for example, rituals such as play — are so very loosely coded that to consider them linguistically, we must stretch a comparison with verbal language, which is so strictly coded. Another critic has complained that the titles assigned the various systems seem arbitrary.[55] For example, Eco himself has this problem in his references to codes of perception and codes of recognition. Further, because the difference between an iconic sign and an iconic seme is relative, an important distinction is left as an open-ended choice when it comes to applying the principle to a specific example. Finally, the relation between the iconic and the kinesic has not, in my opinion, been adequately theorized, especially as this pertains to images of the human body represented in cinema, an art ruled by both iconic and kinesic codes. The notion of code is most useful as an orientation. It forces us to see how existing knowledge is used to build new semantic systems (or languages) which construct "realities" parallel to a real world, and how in this adjacent "reality" we can "say" anything, even that dogs do "dance."


1. Umberto Eco, A Theory of Semiotics (Bloomington: Indiana University Press, 1976), p. 174.

2. Eco, p. 8.

3. Eco, p. 72.

4. Ferdinand de Saussure, Course in General Linguistics, trans. Wade Baskin, eds. Charles Bally and Albert Sechehaye (1957); rpt. New York: McGraw-Hill, 1966), p. 16. Saussure here first suggested a general semiology, of which linguistics would be a branch.

5. The graphic conventions I am using are based on those Eco has used in A Theory of Semiotics. To make a point that images, gestures, and objects are signs and that I must use words as signs to refer to these, I have set off these special signs with slant lines //. Thus, /dog/ is meant to designate the image rather than the word dog. Connotation and denotation, or that content which is signified, is set off by arrows <>. For instance, /leg/ is a sign connoting <hurdle>. Codes or systems are set off with brackets {} instead of quotation marks, hence the {tap dance} code. Double quotation marks are reserved for emphasis and quotation.

6. Stuart Hall, in "Encoding and Decoding the Television Discourse," Centre for Contemporary Cultural Studies Occasional Paper, 1973, p. 2, describes transmission and reception as a complete circuit because received messages are translated into social systems which then become sources. Reprinted in Culture, Media, Language, eds. Stuart Hall, Dorothy Hobson, Andrew Lowe, and Paul Willis (London: Hutchinson, 1980).

7. Metro-Goldwyn-Mayer publicity copy for LADY BE GOOD, Arthur Freed Collection (University of Southern California Special Collections).

8. Eco, p. 141.

9. Eco, p. 275.

10. Eco, p. 274.

11. Eco, p. 16; Saussure, p. 113.

12. Eco, p. 72.

13. Eco, pp. 268-275.

14. Eco, A Theory of Semiotics, p. 56; Umberto Eco, "Articulations of the Cinematic Code," in Bill Nichols, ed., Movies and Methods (Berkeley: University of California Press, 1976), p.,596.

15. Eco, A Theory of Semiotics, p. 274.

16. Marshall and Jean Stearns, Jazz Dance (New York: Macmillan, 1968), p. 281.

17. Eco, "Articulations of the Cinematic Code," p. 592.

18. Eco, A Theory of Semiotics, p. 193

19. Eco, A Theory of Semiotics, p. 193.

20. Eco, A Theory of Semiotics, P. 191.

21. Richard Dyer, Stars: A Study Guide (London: The British Film Institute, 1979), p. 7.

22. John Kobal, "I Would Rather Dance than Eat!" interview with Eleanor Powell, Focus on Film No. 19 (Autumn, 1974), p. 23.

23. Fred Astaire, Steps in Time (New York: Harper and Brothers, 1959), quoted in Stearns, p. 224.

24. For an idea of the scope and complexity of social typage see Orrin E. Klapp, Heroes, Villains, and Fools (Englewood Cliffs: Prentice-Hall, 1962).

25. See Roland Barthes, "The Realistic Effect," Film Reader 3 (February, 1978), for a discussion of the way detail signifies "reality" in literature (p. 134).

26. Peter Berger and Thomas Luckman, The Social Construction of Reality (New York: Doubleday, 1966) theorize this more fully as "collective sedimentation" (p. 39).

27. Hugh Fordin, The World of Entertainment (New York: Doubleday, 1975), p. 43.

28. These rediscoveries range from Molly Haskell's cautious suggestion that Garbo, Dietrich, Powell, and Hepburn "strayed" from heterosexuality but returned to it in From Reverence to Rape (New York: Holt, Rinehart and Winston, 1973), p. 132; to Janet Meyers' identification of the sexually "inscrutable" stars as certainly lesbian, quoted in Richard Dyer, Stars (London: British Film Institute, 1979), p. 67; and the "Introduction to the Lesbian Special Section" (Jump Cut Nos. 24/25, March, 1981), p. 18, by Edith Becker, Michelle Citron, Julia Lesage, and B. Ruby Rich, which acknowledges "viewing strategies" which find the lesbian subtext in some of the classic star performances.

29. Eco, A Theory of Semiotics, p. 310.

30. Eco, A Theory of Semiotics, pp. 275-276.

31. Ray L. Birdwhistell, Kinesics and Context (Philadelphia: University of Pennsylvania Press, 1970), p. 42. Birdwhistell's "rule of thumb" is that the more similar (unimorphic) the male and female of a species, the more necessary the gender display. Humans, although they are dimorphic (having both a male and a female form), rather than unimorphic, are, relative to other species, not strongly marked as different by secondary sex characteristics.

32. Birdwhistell, p. 46.

33. Dennis Giles in "Showmaking," Movie No. 24 (Spring, 1977), pp. 17-18, quotes a lecture by Paddy Whannel in which Whannel describes Fred Astaire's dance style as "seduction and initiation of the female partner."

34. Eco, "Articulations of the Cinematic Code," p. 592.

35. Eco, "Articulations of the Cinematic Code," p. 600.

36. Birdwhistell, p. 113.

37. The idea that gesture is innate rather than learned has had particular tenacity because Charles Darwin espoused the instinctual explanation. See, for example, Charles Darwin, The Expression of Emotions in Man and Animals (1887, rpt. New York: St. Martin's Press, 1979).

38. Birdwhistell, p. 193.

39. Birdwhistell, p. 197.

40. Stearns, p. 140.

41. Eco, "Articulations of the Cinematic Code," p. 602.

42. Richard Dyer, "Entertainment and Utopia," Movie, No. 24 (Spring 1977), p. 3. Reprinted in Genre: The Musical, ed. Rick Altman (London: Routledge and Kegan Paul, 1981).

43. Richard Dyer, The Musical, Study Unit No. 16 (London: British Film Institute Education Advisory Service, 1975), pp. 18-19.

44. Dyer, The Musical, p. 19.

45. Dyer, "Entertainment and Utopia," p. 3. Dyer makes the case for the cultural basis of feeling as seen in the relationship between tonal structures and the range of emotions endemic to every society.

46. Dyer, "Entertainment and Utopia," p. 3.

47. Dyer, The Musical, p. 19.

48. Fordin, p. 41.

49. Michael Wood, "When the Music Stopped," The Columbia Forum, 1972, reprinted in Dyer, The Musical, p. 25.

50. Kobal, p. 24.

51. Jane Feuer, The Hollywood Musical (Bloomington: Indiana University Press, 1982), p. 61. Feuer points out that the popular/classical polarization in the film comments on Gershwin's own classical aspirations.

52. Eco, A Theory of Semiotics, P. 7.

53. Eco, A Theory of Semiotics, pp. 286-298.

54. An example of this is Eco's idea, developed in "Articulations of the Cinematic Code" (originally a paper delivered in 1967) that cinema iconography is triply articulated into figures, iconic signs, and iconic semes. This paralleled Birdwhistell's kine and kinemorph, to which Eco added the kinesic figure in this early paper. Later, in A Theory of Semiotics, he dismisses the possibility that iconic figures combine to produce iconic signs (p. 215).

55. John Corner, "Codes and Cultural Analysis," Media, Culture, and Society, Vol. 2, No. 1 (January, 1980), p. 81.