Hollywood Cinema and Baroque Companionship: An Analytic Critique of David Bordwell’s “Classical Paradigm”*[1]
Concordia University, Montréal
Citation | Author | pdf version
1 In a key scene from Orson Welles’s The Lady from Shanghai (1948), the putative femme fatale Elsa Bannister (Rita Hayworth) is shown pressing a button, after which a dying man collapses in her nearby kitchen. She presses the button again, as a result of which it would seem a vehicle carrying Michael O’Hara (Orson Welles) and George Grisby (Glenn Anders) crashes. It is inconceivable that Elsa has provoked any of these events – not by the push of a button at any rate – which leads Edward Branigan to contend that the montage illustrates an “impossible causality”:[2]

[One] does not finally believe that the woman [Elsa] has directly caused the violent actions. The spectator’s perception of the distances among the spaces, coupled with culturally-acquired knowledge of how cars operate and how rich people summon servants at the push of a button, suggests that there is no immediate causal connection among the events.[3]

The scene invites numerous considerations, some hinging on notions of class and gender. But it is Branigan’s identification of its hermeneutic difficulty, even impossibility, as best summed up by the term “causality” that merits scrutiny. To the extent that what David Bordwell has termed character-centred causality is an inextricable feature of classical narration,[4] the question becomes: how should this aberration in Welles’s film be characterized, and what are its implications, if any, for how we define the cinema of the classical Hollywood era? What follows may not be the first effort to dismantle the presumed uniformity of Hollywood filmmaking, understood by Bordwell as operating under a single aesthetic paradigm, but I hope to make clear how the current endeavour is still crucial.

2 The purpose of this paper is, firstly, to develop a critique of David Bordwell’s characterization of the classical paradigm, as enunciated in his sections of The Classical Hollywood Cinema. In so doing, I aim to improve upon past efforts to undermine Bordwell’s paradigm. Not enough attention has been granted to the precise details of Bordwell’s theoretical model: his three-tiered system for dissecting and analysing narration and style. While an exhaustive review of the debates engendered by Classical Hollywood Cinema would prove impossible, or more likely tedious, I invoke two emblematic efforts, by respectively Elizabeth Cowie and Rick Altman, whose critiques of Bordwell misapprehend pivotal details of his project. By contrast, I meet Bordwell with greater proximity on his own terms. This mandates placing his model within the history of philosophical and scientific thought. In turning to the logical positivist Rudolf Carnap and the analytic philosopher Nelson Goodman, the former whom I position as an antecedent to Bordwell, and the latter whose critique of Carnap is applicable to Bordwell, I address the inferential missteps in his work. The ultimate question becomes, on what basis do we classify a text by one art-historical lineage and not another? The turn to Carnap-Goodman will elucidate how Bordwell’s methodology is wanting. Furthermore, this assessment of Bordwell, in light of Carnap-Goodman, occupies the majority of my analysis, while leading directly to my second objective.
3 Hence, my second aim is to take steps towards defining an alternate art-historical lineage within Hollywood’s nominally classical paradigm, which I dub “Hollywood Baroque.” This requires that I offer some feasible definitions of both classicism and Baroque, terms whose very adequacy and application are contested. I defend a definition of both style-classes, by appealing to Art History – a discipline from which Film Studies has frequently drawn its aesthetic predicates.[5] I thereafter show how the concept of “Baroque” may be productively applied to select instances from Orson Welles’s Lady from Shanghai (1948), though I should hasten to add that the current study is not film criticism. My brief analysis of the Welles film is illustrative of the theoretical difficulties located in Bordwell and the alternatives which I posit to Bordwell’s model. In so doing, I gesture to further research which might combine philosophy and art historiography to delineate, in conclusion, just how the Baroque (and perhaps other artistic practices) are to be theorized in Hollywood cinema.

Hollywood Classicism According to Bordwell

4 As readers of CHC know, David Bordwell posits the overwhelming majority of Hollywood filmmaking of the classical era, delimited within the years 1917-1960, as aesthetically uniform.[6] The thesis of course is not that all the films are self-identical, but that as individual films they all belong to a dominant category, understood alternately as a “style-class,” “historical mode of narration,” and/or “paradigm.” For the purposes of brevity, I will treat these terms synonymously: the semantic discrimination between them does not undermine my position.[7] What matters is that all Hollywood films exhibit a high degree of stylistic similarity, which is antithetical or incommensurate to alternate artistic practices: post-war art-cinema, the Soviet Montagists or so-called “parametric narration.”[8] This similarity corresponds to the aesthetic canons of classicism.
5 In the opening pages of CHC, Bordwell maintains: “the principles which Hollywood claims as its own rely on notions of decorum, proportion, formal harmony, respect for tradition, mimesis, self-effacing craftsmanship, and cool control of the perceiver’s response – canons which critics in any medium usually call ‘classical.’”[9] In film, Bordwell correlates his account to André Bazin’s designation of Hollywood cinema as classical.[10] Here, one preferred technique of Hollywood classicism is continuity editing, which for Bazin, seeks to desensitize audiences to the cutting from one shot to the next[11] via the evocation of an integrated and coherent spatial whole: all the shots interlock in depicting a stable, delimited environment, whereby one could mnemonically reconstitute without contest who is where, and in relationship to what. Bordwell’s notion of analytic cutting follows Bazin’s lead: space is atomized, from establishing shots to close-ups, in a manner whereby all the pieces fit together.
6 Moreover, the movement from long shot to close-up follows the movement of the narrative.[12] As seen in Bordwell’s and Kristin Thompson’s dissection of The Maltese Falcon (1941), the opening scene unfolds towards greater narrative specificity. One comes to better apprehend the relationship of the characters (Sam Spade/Humphrey Bogart, Brigid O’Shaughnessy/Mary Astor) to one another and their roles within the plot (private eye, client and possible love interest). Likewise, the shots become progressively more restrictive in framing, singling out the details of pertinent information (e.g. a facial close-up to accentuate a line and/or expression bearing directly on the plot). Such an analogy has its limits, as Bordwell counsels that one separate form from function: close-ups and establishing shots do not guarantee a coherent constitution of space, any more than the latter requires analytic editing. The fluency of Bordwell’s approach is that diverse techniques achieve similar ends, thus ensuring innovation, change and/or variety throughout Hollywood history. It also commits him to a more conceptually dexterous understanding of the “classical.”
7 The more complicated definition in CHC comes with Bordwell’s proposed means of envisioning narration. Partly to escape the essentialism found in Bazin, Bordwell develops an at once constructional and systemic model. Such a model entails that the technique be defined by its position within a larger network. No single technique has an exclusively defined expressive purpose, or even relation to some concomitant aspect of external reality. However, in contradistinction to Barry Salt’s style-theory, there is no statistical ordering of specific techniques, nor is there, as in Structuralism, the singular demarcation of one technique from another.[13] “Constructional” implies that no technique enjoys any exclusive role within the signifying process of the medium, such that itemizing a range of techniques does not afford any insight into the range of a director’s artistic choices and decisions. Instead, Bordwell’s three-tiered design is such that the content of the lower level informs that of the middle-level, which in turn helps determine the upper-level, via an apparently infinite array of combinations and inter-relations. Each technique functions differently as per its place within a combinatory organization, from one level to the next.
8 For example, Christian Metz’s Grande Syntagmatique[14] would be seen as circumventing the range of stylistic constructs discernible within a film, as well as their significance. Any number of filmic passages falls off the Metzian grid,[15] although that might corroborate the heuristic valence of Metz’s system. Unfortunately, the syntagmas are defined in such generic language that it becomes impossible to infer how the same syntagma might have different applications depending on the film. The same sub-type of a-chronological syntagma, dubbed “en accolade,”[16] appears in Une femme mariée (1964) and in The Scarlet Empress (1934).[17] This does little to position Jean-Luc Godard vs. Josef von Sternberg within their respective filmmaking contexts. The sequence referenced in Godard is stylistically (and thematically) consistent with the many descriptive, essayistic passages throughout his cinema; the corresponding syntagma in the von Sternberg, while structurally isomorphic, plays a different role within the film’s narration: despite being “a-chronological,” it presages the heroine’s later introduction into a world of political violence. One cannot verily take these as normative practices,[18] unless one construes Une femme mariée as exemplifying normative filmmaking. Metz may be aligned with theoretical poetics and not historical poetics[19] – evolving stylistic models that have greater universal applicability than, for example, Bordwell’s measurements of Average Shot Length (ASL)[20] – but this returns us to the problem that the syntagmas cover part, and not all, of film poetics.
9 Bordwell’s approach, by contrast, notes that any number of technical units – a shot, an edit, staging, lighting, music etc. – composing the lower level is granted different valence based on their combinations within systems of space, time and/or causality, i.e. the second level. The upper level is organized around the hierarchical organization of time, space and causality.[21] Therefore, between the lower and middle levels, shot length may be subordinated to character movement throughout the frame, as Bordwell and Thompson have remarked in specifying Howard Hawks’s use of duration.[22] This character movement might in turn have narrative import, as pertaining to the upper-level. There are diagrammatically precise, yet supple, means of charting the applications of a technique in different contexts which, in turn, allow one to identify a filmmaker’s art-historical lineage and affiliation. Returning to Hawks, the subordination of shot length to staging, which in itself exemplifies narrative purpose, means his evocation of time is narrative time, rather than the temps mort of a Michelangelo Antonioni or a Béla Tarr. But Bordwell’s language is simpler and more exacting. In Hollywood cinema – unlike art-cinema – time and space are subordinated to causality: the viewer is presented with any given number of locations, and offered some indication as to how much time has elapsed within and between each location, to inscribe the whole within a chain of cause-and-effect, in short, a goal-oriented plot.
10 This simplifies Bordwell’s analysis, as he distinguishes between types of causality. But as a provisional insight, it sheds light on how a film such as The Lady from Shanghai might be analysed. It should be noted that Branigan characterized the aforementioned sequence as denoting an “impossible causality,” albeit a causality nonetheless. This corresponds to Bordwell’s bounds of difference: Welles is beholden to the causal logic of classical narration, even as he mocks it.
11 Indeed, it might seem tempting to argue that Welles’s ironic treatment of causality in narration is itself proof that he is working in a discursive mode radically apart from Hollywood classicism. However, this only corroborates Bordwell’s historiography: the very terms according to which Welles defies conventional story-telling are the very terms dictating conventional story-telling, i.e. character-centred causality, or goal-oriented narration. Orson Welles is thus a “deviant,” but his films, in Hollywood, do not represent “thorough-going alternatives.” The Hollywood system cannot allow Welles to dispense with narrative causality altogether as in Un chien andalou (1929), À bout de souffle (1960), or Wavelength (1967).[23]
12 This is a powerful argument that withstands the criticisms that have been directed at Bordwell over the years; I also take it to be misleading in important regards. Let us attend to Bordwell’s strengths first. There is a potential hazard in how Bordwell collapses classical plot-structure with classicism tout court, as not all classicism in the arts is understood in narrative terms: this partly informs Elizabeth Cowie’s critique[24] of the CHC, as she notes that The Big Sleep (1946) is classically shot and edited, but defies narrative logic. Conversely, as Rick Altman has suggested, classical narration is not itself classical, if the latter is understood in the aesthetic terms which Bordwell invokes: “decorum, proportion, formal harmony [etc.].”[25] This leads Altman to consider melodrama, whose aesthetics of spectacle are understood as non-classical or anti-classical, as forming one half of the dialectical whole within Hollywood narration.
13 Neither critique carries much force. Bordwell thoroughly emphasises that the classical paradigm allows for some subversion. In The Big Sleep, the narrative gaps in the mystery story-line are papered over by the reigning stylistic devices (e.g. continuity editing) that convey the romance plot, itself causally transparent. “Who killed Owen Taylor?” hardly looms over the film with the same ambiguity as “what happened last year at Marienbad” in Alain Resnais’s Last Year at Marienbad (1961). Quite the contrary: its ambiguity is contained within not only the Bogart-Bacall romance, but also the larger mystery that is ultimately solved. Eddie Mars (John Ridgely) is exposed as the chief villain; his motives, actions and relations to the other characters revealed; the narration culminates shortly after his defeat, and the main couple is reunited. In underscoring how this classically constructed film momentarily accommodates a storytelling aberration, Cowie unwittingly corroborates Bordwell’s thesis. It is a problem moreover shared by Altman’s critique,[26] which I pass over only to mention that elsewhere in CHC particular attention is granted to melodrama’s constitutive function in the Hollywood model of storytelling:[27] to mix metaphors, Altman’s smoking gun had already been detected by the very authors he targets.

Bordwell’s Antecedents: Logical Positivism, Carnap and Goodman

14 Despite the voluminous commentary which the CHC has elicited, scant attention has been granted to Bordwell’s theoretical model taken as an outgrowth of logical positivism. Bordwell himself does not make the connection, although I argue it is unavoidable, and leads to the drawbacks of the CHC. In considering Rudolf Carnap’s 1928 book The Logical Structure of the World,[28] I limit my analysis to its intellectual proximity to the CHC. Nonetheless, some words on Carnap’s project are in order.
15 In his book, Carnap aims for nothing less than a theoretical model for analysing the whole of phenomenal reality. For this he employs a constructional system,[29] structurally identical to Bordwell’s, albeit erected on the basis of “elementary experiences”: perceptions given in pre-linguistic/pre-cognitive form.[30] Examples could be what one sees in the opening shot of The Fearless Vampire Killers (1967), or Omar Sharif’s entrance in Lawrence of Arabia (1962). In either case, what is respectively revealed to be the surface of the moon and a character riding forward on a horse, are first given to spectators as variegating textures of grey, or a mobile black patch on the horizon. It is only upon inspection, and the progression of the scene, that these are recognizable as physically identifiable objects, which constitutes the second level of Carnap’s structure. The upper-level is comprised of cultural objects, and pertains to the conceptual difficulty in later analytic philosophy as to how one might differentiate Marcel Duchamps’s Fountain (i.e. a cultural object) from everyday urinals or even fountains (i.e. both merely physical objects).[31]
16 In short, Carnap aims for a comprehensive and internally coherent theoretical model for understanding the world. The problem arises when Nelson Goodman is able to show that its logic is compromised. Part of Goodman’s critique hinges on a thought-experiment and a problem named the “companionship difficulty.” Goodman’s analysis is detailed and unusually technical.[32] For the current purposes, consider the following. He imagines a hypothetical situation (derived from a close reading of Carnap’s own theorizing) whereby one is informed of the existence of a series of multi-coloured objects. One does not know what individual colours each object has, but only the following: which objects share the same colours. Hence, we may know that objects 1 and 2 share a certain colour, designated k1.[33] We do not know if object 1 or 2 possess other colours, other than what colours (e.g. k2, k3) they might share with other objects (e.g. 3, 4). The epistemological gamble is to infer each existing colour (e.g. k1, k2, k3 etc.) within a delimited set of objects (e.g. 1, 2, 3 etc.) knowing only which objects share the same colour, and not the full palette of colours that any one object may possess on its own. More importantly, what is at stake in the current enquiry is not the physiology and/or mechanics of colour perception per se, but the problem of how we ascribe properties (whether this be colours, or anything else) to objects, and on what basis. Meaning: how do we ascribe a property to an entity, and in so doing, are there any other properties that escape our attention?
17 In Goodman’s example, the prospect of inferring individual properties from the shared properties of groups of objects is alarmingly dim. He lists five objects whose colours are identified thus: 1. k1-k3; 2. k1; 3. k1-k2; 4. k2; 5. k1-k2-k3.[34] Again: the numbers, 1 through 5, refer to the objects; the lettered numbers (k1, k2, k3) to the colours. The challenge is, can we infer each colour (k1, k2, k3) if all we have is the following information: objects one, two, three, and five share at least one colour; and so do objects three, four, five.
18 Hence, we know that object one shares a colour with objects two, three and five, and hence infer the existence of one colour (e.g. k1). Likewise, we know that objects three, four and five share a colour, so we can infer a second category (k2). Unfortunately, the possibility that objects one and five might also share yet another colour (i.e. k3) goes unnoticed, because “k3” only appears alongside other colours, and never alone. In other words, knowing that objects one and five share a colour, we will have inferred k1, after which there is no reason to believe that further colours might also be deduced – whether this is k3, or anything else. As a colour k3 appears only alongside other colours: its companionship disguises its identity, hence “companionship difficulty.”
19 The above thought-experiment, while apparently esoteric and eccentric, is not as far removed from Bordwellian historiography as might be assumed. I contend that due to the companionship difficulty, Bordwell misinterprets the flamboyant passages in films such as The Lady from Shanghai as deviations. How does this inferential mishap occur? I offer only a simplified, albeit defensible account.
20 Bordwell’s opening seven chapters in the CHC are intended to detail the poetics of the “typical film” as produced in Hollywood, which equally serves as a norm against which deviant or stylistically anomalous films are evaluated. Within the Goodmanian language above, he has his objects – a corpus of films, the listing of which occupies several pages in appendix of CHC – and he has his k1s: the uninterrupted recurrence of character-centred causality. Now what does he do with character-centred, Wellesian “impossible causality?” Is it regarded as a variation of the classical paradigm, or is it to be labeled as something else, the aesthetic-stylistic equivalent to a k3? Bordwell’s diagnosis is the former. My contention is that Goodman’s thought-experiment helps us to recognize this is the wrong answer, by explicating the flawed logic whereby one overlooks properties (e.g. Baroque aesthetics) insofar as they recur in the same entities (e.g. films) in companionship of a nominally separate set of characteristics (e.g. character-centred causality).
21 Instead of the style-classification entailed by Bordwell’s account of the classical paradigm, I propose that Welles is part of a distinct tradition, which I dub Hollywood Baroque, due to the proximity of its discernible tone and effects to what art historians understand as Baroque aesthetics in the fine arts. Although it cannot be exhaustively examined as a style-class within the current context, I would postulate that Hollywood Baroque surfaces throughout the history of the studio system, but cannot be encapsulated by the style-class “classicism.” Instead, in Goodmanian terms, a film’s Baroque features subsist in companionship to its otherwise classical poetics. Consequently, in considering the Hollywood Baroque, it is necessary to equally review the term (“Baroque”) as defined by past art historians.

Lady from Shanghai and Baroque Aesthetics

22 Although Alois Riegl’s historicist theory of the origins of artistic styles may not survive Ernst Gombrich’s trenchant criticisms,[35] his preliminary description of Baroque art is apt: “The meaning we associate with [Baroque art] is clear: peculiar, unfamiliar, extraordinary. […However] we do not understand the extraordinary in Baroque art and […] we are appalled by it, and we perceive it as disturbing or as a troublesome confusion.”[36] Heinrich Wölfflin identifies the Baroque with an art based on excess, provoking revulsion and unease. In the introduction of Renaissance and Baroque he cites the following disparaging comments, which he mistakenly attributes to Denis Diderot, whereby baroque is “a nuance of the bizarre” and: “It is, if you like, its refinement, or, if it were possible, to say so, its abuse.” Further still, it “is the superlative of the bizarre, the excess of the ridiculous.”[37] A few pages earlier, Michelangelo is singled out as having earned the reputation of “father of the baroque,” to which Wölfflin adds: “Michelangelo treats his forms with sovereign inconsiderateness, purely as elements in a composition of plastic contrasts and overall effects of light and shade, with no thought for their structural purpose.”[38] Later, he describes Baroque architecture: “The freedom of line and the interplay of light and shade are satisfying to the painterly taste in direct proportion to the degree to which they transgress the rules of architecture.”[39]
23 While these might seem like approximate designations, they point to more sustained insights. One can outline these claims, with equal appeals to art historiography and the “companionship” principle, whereby a film’s Baroque style accompanies aspects of its classical construction. Hence, films within the Hollywood Baroque periodically undermine the principles of “decorum, proportion, formal harmony [etc.]” already invoked by Bordwell in his designation of the classical.[40] Instead, they create variously jarring effects by emphasising degrees of discontinuity and/or disproportion between the graphic elements of the images, or the sequencing from one image to the next. However, they display such “troublesome confusion[,]”[41] or “sovereign inconsiderateness,”[42] while also respecting other key principles of classical poetics: for example, as we shall see with Lady from Shanghai, maintaining a goal-oriented narration, while interrupting this narration and other vestiges of the Hollywood tradition with “impossible causality”[43] and other such means.[44]
24 Indeed, the inconsistent deferrals to continuity editing, already cited and to be further itemized below, is why it would not be enough to posit that Lady from Shanghai is a synthesis of classical storytelling and a Baroque visual style (i.e. the inverse of Cowie’s reading of The Big Sleep). Quite the contrary: the film draws on character-centred causality (e.g. the plan to frame O’Hara) while also mocking it (e.g. the cross-cutting between Elsa’s button-pushing and the car accident); it makes use of over-the-shoulder editing and centred framing, while also devising various asymmetries from one shot to the next, or within a shot. The whole point of “companionship” is that the Baroque aesthetics sit alongside nominally classical features, and the two are intertwined.
25 In the following four examples, one can better appraise how Welles creates such incongruous effects. The first two undermine the degree of continuity habitually conveyed via shot/reverse shot editing, with unexpected variations on respectively how characters are placed before the camera, or in relation to the background. The latter point – the relationship of character to background – can be observed in a third example, while a fourth example is a more vertiginous medley of varied shot scale, editing and composition.
26 The first two examples hinge on encounters between O’Hara and the crooked lawyer George Grisby (Glenn Anders). In the first, comparatively less flamboyant scene, the two meet in a restaurant. Welles films Grisby so that he is directly facing the lens. However, the reverse shots of O’Hara are framed differently, with the latter shot in the more standard three-quarter view that is the mainstay of continuity editing. Hence, the scene both follows the conventions of continuity editing (the three quarter view) and strays from them (the frontal view), such that the juxtaposition of either approach proves all the more unsettling.
27 A similar effect surfaces during the pivotal exchange when Grisby hires O’Hara to kill him. Welles again films the conversation in alternating close-ups, but creates various contrasts via the different backgrounds. Each character is framed in a standard three-quarter view, albeit placed in visually separate spaces. Grisby is shot against the background of hilltops and trees, thus placing him within a fairly tangible environment. In contrast, O’Hara is filmed against a dark clouded sky, which imbues the image with an ethereal quality: O’Hara appears dislocated from any concrete setting. There is a firm contrast between his corporeal presence in the foreground, and the gaseous, ever-receding expansion of the cloud and sky behind him. This contrast is reinforced by the visual anomaly of first seeing a landscape in the background, and then, in the next shot, clouds and sky. The characters are spatially proximate, yet seem to stand in physically discontinuous locations, which heightens the sense of disorientation.
28 This tactic of creating a tension between background and foreground is maintained in a later encounter between O’Hara and Elsa, when they meet to discuss the details of Grisby’s presumed murder. The scene is set in an underground aquarium, such that throughout the course of several shots, Welles frames the characters in front of glass partitions behind which swim assorted fish and other sea animals. The interplay of movement, as well as light and shadow, between background and foreground is continually varied. Hence sometimes the characters are silhouetted in front of the comparatively luminous aquarium; in other instances, Rita Hayworth is presented in more brightly lit close-ups. Likewise, the characters sometimes adopt stilted poses, or move slowly, in front of the more hurried pace of the animals behind them. At other times, Welles stages the action such that their walking across the frame closely parallels or follows the motion of an animal behind them. The sum effect is a sequence which proves almost impossible to pin down, in terms of consistently maintained lighting, staging, or background/foreground interplay.
29 The fourth and final example – an early confrontation between Arthur Bannister (Everett Sloane) and O’Hara on the former’s yacht – may be among the most atypical passages in the film.[45] This sequence begins as an unnerving, off-centred ensemble of Bannister, his wife Elsa, O’Hara and Grisby. With his back turned to the camera, O’Hara stands before a seated Bannister, who is on the left edge of the frame in the middle-ground. Bannister taunts O’Hara with the possibility that he may be in the know of the latter’s mutual attraction to his wife (Elsa and O’Hara have shared a kiss, to which Grisby was privy). Bannister thus appears to be in command of the scene, a privilege reinforced by the lines running throughout the composition. Perspective is created by Elsa’s diagonally placed legs, lying upon a set of rectangular cushions, all pointing to Bannister. That O’Hara is seen with his back to the camera places him in a position of vulnerability.
30 At the same time, the shot is systematically disordered. While Grisby’s participation in the scene is at best peripheral, he sits closer to the centre of the frame. His presence throughout the evolving scene is contrapuntal. As Bannister and O’Hara launch into a taut exchange about whether or not wealth buys happiness, George at first sings off key while playing the piano, poorly, and then offers some dumb rejoinders to the conversation. These are conveyed by close-ups of Grisby, which interrupt the series of shot-reverse shots between O’Hara and Bannister, and appear dislocated: Grisby, while purporting to speak to Bannister, in effect is staring away from any of the other characters in the scene. Additionally, the rocking of the boat off-sets the apparently static nature of the shots, in which everyone is largely immobile. The same sense of off-centred action occurs when Elsa asks Grisby for a light: it overlaps with O’Hara’s and Bannister’s dialogue, and moreover is at best incidental and banal, in contradistinction to the more heated conversation between the two antagonistic men. One could hardly find a better expression to summarize the jarring compositions and editing in this sequence than the aforementioned descriptions of Baroque aesthetics as either a refinement, or superlative, of the bizarre.[46]
31 To the above examples, I would equally advance that similar poetics, and concomitantly jarring effects, are maintained throughout the majority of the film. If this remains fair characterization of Lady from Shanghai, then the next step would be to locate other works that exhibit similar features. Understanding how one might identify such art-works, and on what basis, is a question I now address in the conclusion.


32 To recapitulate: the above analysis is not an exhaustive scrutiny of Welles’s signal work. Nonetheless, the arguments remain consistent with the research of prior scholars who seek to dislodge Hollywood filmmaking from any single aesthetic paradigm. In addition to The Lady from Shanghai, an assessment of Hollywood Baroque would have to consider the majority of Welles’s other Hollywood efforts,[47] as well as films by Josef von Sternberg,[48] Alfred Hitchcock,[49] and Samuel Fuller,[50] works which are frequently characterized by excessive foregrounding of style, blatant shifts in tone, and/or loosely organized if not episodic narrative construction. At the same time, I have endeavoured to point to the logical deficiencies in Bordwell’s approach, as too often critiques of the CHC do not sufficiently respect his original arguments: i.e. as has been suggested in the case of Rick Altman, a designated feature of Hollywood cinema is arbitrarily selected and accorded special prominence by fiat. Instead, I would suggest that film scholarship (re)turn to the philosophical canon in developing its arguments. While continental philosophy has proved seminal in generating new concepts and nomenclature,[51] my debt to analytic philosophy serves a different purpose: subordinating existing concepts and nomenclature to greater scrutiny, and exigencies of clarity and consistency.[52]
33 As such, the term Baroque ought to be envisaged from a novel perspective. Film scholars interested in the Baroque aesthetics of Hollywood cinema have succumbed to the teleological and/or programmatic thinking of Riegl and Wölfflin. Just as Baroque painting grew out of Renaissance classicism, so it is assumed that the Hollywood Baroque is situated after the end of the classical studio system, post-1960.[53] Just as Wölfflin organized the scope of classical and Baroque art around five sets of antithetical polarities, so are the same polarities applied to commercial Hollywood.[54]
34 My approach is less literal-minded in indicating that Baroque aesthetics in Hollywood were contemporaneous to their classical counter-part. The two styles do not adhere to a pre-determined pattern that replicates their historical antecedents in the fine arts. Equally, one must look beyond the application of pictorial techniques associated with Baroque painting, and attend more to the function of those techniques: whether they can be seen as violating the classical norms identifiable with the “ordinary film.”[55] The shower scene in Psycho can stand as a key instance of Hollywood Baroque, even if pictorially its concatenation of geometric motifs is closer to the iconography of the twentieth century avant-garde. Yet images such as the tracking in onto Janet Leigh’s eye, in extreme close-up, help convey the sense of imbalance and of the appalling that art historians have likened to Baroque painting, albeit by means foreign to the fine arts of the seventeenth century. This ironically pays homage to Bordwell’s distinction between technique and function.[56] At the same time, if the same techniques can be said to have alternately classical, Baroque and other style-related functions, ultimately, categories of style may have to be defined not as an ensemble of conventions and structures deployed by artists, but in terms of their effects as experienced by audiences.
[1] An abbreviated version of this paper was first presented at Deviate! The 2nd International Alphaville: Journal of Film and Media Studies Conference at University College Cork, on September 5, 2014. It has since been considerably revised and expanded.

[2] Edward Branigan, Narrative Comprehension and Film (London/New York: Routledge, 1992), 51.

[3] Ibid., 47.

[4] David Bordwell, Janet Staiger and Kristin Thompson, The Classical Hollywood Cinema: Film Style and Mode of Production to 1960 (New York: Columbia University Press, 1985), 13. Hereafter CHC.

[5] As we shall later see in this chapter, Bordwell tacitly acknowledges the prior example of art historians in positing his conception of classicism (see note 9); Sean Cubitt makes comparable mention of Heinrich Wölfflin in his account of baroque film aesthetics (see note 54). More generally, Bordwell has periodically drawn on Ernst Gombrich in developing his historiography of style, including in the CHC. A fuller excavation of how the relationship of film aesthetics to fine arts and especially painting has both influenced and been studied by film critics and scholars includes authors like Bazin, Lotte Eisner and more recently Jacques Aumont.

[6] CHC, 3-84.

[7] On “mode,” see David Bordwell, Narration in the Fiction Film (Madison: Wisconsin University Press, 1985), 150; on “paradigm,” see Bordwell, ibid., 151; Bordwell writes of “group style” instead of “style-class” in CHC, 1. Discriminating between either does not sway the argument.

[8] Ibid., 156-310.

[9] CHC, 3-4. Note that Bordwell writes in any medium (see note 5).

[10] For an English translation see: André Bazin, What is Cinema?: Volume 1, trans. Hugh Gray (Berkeley, Los Angeles, London: University of California Press), 23-40. See also: Bazin, What is Cinema?, trans. Timothy Barnard (Montreal: Caboose), 87-106. For the original French : Bazin, Qu’est-ce que le cinéma? I: Ontologie et langage (Paris: les Éditions du Cerf, 1958), 131-148.

[11] In the original French, Bazin uses the term “insensible” when discussing the so-called invisible style of continuity editing, which I take as a corollary to: one does not sense the latent disjunction of cutting from one shot to the next: Bazin, Qu’est-ce que le cinéma? I: Ontologie et langage, 132. By contrast, in Hugh Gray’s translation there is mention instead of “conceal[ment]”: Bazin, What is Cinema? Volume 1, 24. In turn, Timothy Barnard elects “imperceptible”: Bazin, What is Cinema?, 88.

[12] David Bordwell and Kristin Thompson, Film Art: An Introduction, 10th ed. (New York: McGraw-Hill, 2012), 235-238. For Bordwell’s analysis of “continuity editing,” see: CHC, 56-59.

[13] See David Bordwell and Kristin Thompson, “Toward a Scientific Film History?,” Quarterly Review of Film Studies 10: 3 (Summer 1985), 224-237 for an ambivalent appraisal of Barry Salt, Film Style and Technology: History and Analysis. (London: Starwood, 1983). See David Bordwell, “Neo-Structuralist Narratology and the Functions of Filmic Storytelling,” in Narrative Across Media: The Languages of Storytelling, Mary-Laure Ryan ed. (Lincoln: University of Nebraska Press, 2004), 203-219 for his critique of Structuralism in film and narrative theory.

[14] Christian Metz, Essais sur la signification au cinéma. Tome 1 (Paris: Éditions Klincksieck, 1968/1971) 111-146. For the English translation, see: Christian Metz, Film Language: A Semiotics of the Cinema, trans. Michael Taylor (Chicago: The University of Chicago Press, 1974/1991), 108-146.

[15] As Warren Buckland has remarked, Metz acknowledges this: not surprisingly, his example is a sequence in a Godard film, namely Pierrot le fou (1965). See Warren Buckland, The Cognitive Semiotics of Film (Cambridge, UK: Cambridge University Press, 2000), 124-131. Metz addresses the stylistic challenges of modernist cinema with nuance: Metz, Essais sur la signification au cinéma. Tome 1, 211-212. Nonetheless, according to Robin Wood, even a nominally Hollywood film may defy Metz’s categories: see his comments on Raging Bull (1980) in Robin Wood, Hollywood from Vietnam to Reagan… And Beyond. Expanded and Revised ed. (New York: Columbia University Press, 2003), 223-225.

[16] Translated as “bracket syntagma” in Metz, Film Language, 126.

[17] Metz, Essais sur la signification au cinéma. Tome 1, 127-128.

[18] As Warren Buckland claims, reiterating Metz’s own characterization. See Warren Buckland, The Cognitive Semiotics of Film, 121 and Metz, Essais sur la signification au cinéma,104. Conversely, Buckland draws attention to a richly theorized effort by Michel Colin to think past the limitations of la grande syntagmatique. See Warren Buckland, The Cognitive Semiotics of Film, 113-119 and Michel Colin, Cinéma, Télévision, Cognition (Nancy: Presses universitaires de Nancy, 1992), 47-85.

[19] On the different types of poetics, see David Bordwell, Poetics of Cinema (New York: Routledge, 2008), 12-13.

[20] David Bordwell, The Way Hollywood Tells It: Story and Style in Modern Movies (Berkeley: University of California Press, 2006).

[21] CHC, 6-7.

[22] Bordwell and Thompson, “Toward a Scientific Film History?”, 227.

[23] See especially CHC, 81. However, throughout CHC, 3-84, Bordwell periodically contradistinguishes between the poetics of Hollywood cinema and those of non-Hollywood, ostensibly non/anti-classical texts.

[24] See Elizabeth Cowie, “Storytelling: Classical Hollywood Cinema and Classical Narrative,” in Contemporary Hollywood Cinema, Steve Neale and Murray Smith eds. (London, New York: Routledge, 1998), 178-190 for the article in full.

[25] CHC, 4.

[26] See Rick Altman “Dickens, Griffith, and Film Theory Today,” in Classical Hollywood Narrative: The Paradigm Wars, Jane Gaines ed. (Durham: Duke University Press, 1992), 9-47.

[27] For example: “Certainly the Hollywood style seeks effects that owe a good deal to, say, romantic music or nineteenth-century melodrama.” (CHC, 4).

[28] Rudolf Carnap, The Logical Structure of the World and Pseudoproblems in Philosophy (Der logische Aufbau der Welt), trans. Rolf A. George, 3rd ed. (Chicago: Open Court, 1928/2003). Hereafter Aufbau. Indeed, what follows is an especially abbreviated account of Carnap’s detailed tome, with the sole purpose of re-contextualizing Bordwell’s theoretical model. A more developed reading of the Aufbau – assessing its relative merits and flaws; detailing its place within the history of logical positivism, analytic philosophy, and more precisely the mathematical logic of Frege-Russell – would take us too far astray of our current endeavours.

[29] See especially Aufbau, 175-246.

[30] To reiterate, the structure of either system is identical, but not its ostensible content. Carnap’s system starts with “elementary experiences,” whereas Bordwell’s system begins with cinema’s technical devices. What is identical is that in each case one passes from a lower-level (technical devices, in Bordwell) to a mid-level (systems of space, time and causality, in Bordwell), with the latter premised on the former, and so on for the upper-level. That Carnap’s system contains entirely different objects (e.g. “elementary experiences,” and not identifiable cinematic devices) does not detract from the arguments at hand.

[31] A problem considered (via a different set of arguments) in Arthur Danto, The Transfiguration of the Commonplace: A Philosophy of Art (Cambridge, MA: Harvard University Press, 1981).

[32] In what follows, I offer a summary adequate to the purposes at hand. For the analysis in full, see: Nelson Goodman, The Structure of Appearance, 3rd ed. (Dordrecht: D. Reidel Publishing Company, 1951/1977), 114-117. Hereafter SA. I also pass mention on the considerable range of debates and disagreements engendered by Goodman’s critique of Carnap, though as should be clear, I remain sufficiently convinced by Goodman’s approach.

[33] That Goodman is deploying such apparently cryptic symbols – object 1, colour k1 – is consistent with his adherence to analytic and/or ideal language philosophy. K1 can stand for blue, or red, or yellow, or whatever one likes. What matters is determining conclusive insights on the basis of systematically applied terms.

[34] I paraphrase. See SA, 116.

[35] See Ernst H. Gombrich, Art and Illusion: A Study in the Psychology of Pictorial Representation, Millenium ed. (Princeton: Princeton University Press, 2000), 17-22. Gombrich has elsewhere reminded us that “many of the stylistic terms with which the art historian operates began their career in the vocabulary of critical abuse.” See Gombrich, Norm and Form: Studies in the Art of the Renaissance, 3rd ed. (London: Phaidon Press, 1978), 81. This includes the Baroque, and might cast doubt on whether it is even a legitimate aesthetic predicate, especially when contrasted to classicism. However, in the same text Gombrich must also concede that “Man is a classifying animal,” and so “the art historian has rather to admit that classification is a necessary tool, even though it may be a necessary evil.” (Ibid., 82). As such, we should “open our minds to an appreciation of non-classical solutions representing entirely fresh discoveries.” (Ibid., 98). From “non-classical solutions” and “fresh discoveries,” it can be but a small step to envisage more consistent and comprehensive generalisations, many of which arm us with the necessary vocabulary for placing artistic achievements within an interpretive context, even as historians debate whether such vocabulary is rightly applied. It is precisely what I do with Lady from Shanghai and the Hollywood Baroque.

[36] Alois Riegl, The Origins of Baroque Art in Rome, ed. and trans. Andrew Hopkins and Arnold Witte (Los Angeles: Getty Publications, 2010), 94.

[37] For the text as it appears in Wölfflin (untranslated and in French), see: Heinrich Wölfflin, Renaissance and Baroque, trans. Kathrin Simon (London: Collins, 1964), 23. The first two quotes are in fact from Quatremère de Quincey, for which I rely in part on Helen Hills’s translation. See: Helen Hills, “The Baroque: The Grit in the Oyster of Art History,” in Rethinking the Baroque, ed. Helen Hills (Burlington: Ashgate, 2011), 13 and 32, n. 9. The second is by Francesco Malizia. See: Liane Lefaivre and Alexander Tzonis (eds.), The Emergence of Modern Architecture: A Documentary History from 1000 to 1810 (London; New York: Routledge, 2004), 461. Regardless of Wölfflin’s misattribution, both these eighteenth century authors are rightly understood as exemplifying their era’s pejorative views against the Baroque.

[38] Ibid., 19.

[39] Ibid., 29.

[40] CHC, 4. Admittedly, he later refines this analysis with more exact terminology, but neither does he reject this initial characterization. It is therefore argumentatively fair that he be held to these terms.

[41] Riegl, The Origins of Baroque Art in Rome, 94.

[42] Wölfflin, Renaissance and Baroque, 19.

[43] Branigan, Narrative Comprehension and Film, 51.

[44] Bordwell might still choose to disavow the Hollywood Baroque and insist that Lady from Shanghai is merely an extreme variation within the single paradigm of Hollywood classicism. I hope my appeal to Goodman makes clear why such reasoning arises from a flawed theoretical model (i.e. the constructional system and its blindness to “companionship”). I would also add that for Bordwell to argue that such Hollywood features do not represent a “thorough-going alternative” (CHC, 3-84) in the manner of Godard’s or Robert Bresson’s films is as unproductive as insisting that, in light of Cubism, Abstract Expressionism and Pop Art, Baroque painting was little more than Renaissance classicism by other means.

[45] Here I pass commentary on the most famous scene in the film – the hall-of-mirrors finale – as this sequence has already been widely commented upon. Regardless, it could otherwise be considered the most obvious candidate for exemplifying the film’s Baroque aesthetics.

[46] To paraphrase Quatremère de Quincey: Hills “The Baroque,” 13.

[47] One must allow for the fact that most were reedited and/or re-shot at the studios’ behest: James Naremore, The Magic World of Orson Welles (Dallas: Southern Methodist University Press, 1989). On collective authorship, see Paisley Livingston, Art and Intention: A Philosophical Study (Oxford: Clarendon Press, 2005), 62-90. On implied authorship as pertaining to film, see Seymour Chatman, Coming to Terms: The Rhetoric of Narrative in Fiction and Film (Ithaca: Columbia University Press, 1990), 74-108.

[48] On von Sternberg’s uncertain tone, oscillating between earnestness and parody, see George M. Wilson, Narration in Light: Studies in Cinematic Point of View (Baltimore: The Johns Hopkins University Press, 1986), 145-165.

[49] On Hitchcock’s stylized motifs, see François Truffaut, “Un trousseau de fausses clés,” Les Cahiers du cinéma 39 (October 1954), 45-52. See also Eric Rohmer, Le goût de la beauté (Paris: Cahiers du cinéma, 2004), 266-273.

[50] On Fuller’s defiance of classical narration, see Lisa Dombrowski, The Films of Samuel Fuller: If You Die, I’ll Kill You (Middleton: Wesleyan University Press, 2008), 13-14.

[51] See for example in: David N. Rodowick, Reading the Figural, or, Philosophy after the New Media (Durham: Duke University Press, 2001).

[52] See also Richard Allen and Murray Smith, eds., Film Theory and Philosophy (Oxford: Clarendon Press, 1997).

[53] In Angela Ndalianis, Neo-Baroque Aesthetics and Contemporary Entertainment (Cambridge, MA: MIT Press, 2004). Colin Burnett helpfully sums up the problems with this position: Colin Burnett, “A New Look at the Concept of Style in Film: The Origins and Development of the Problem-Solution Model,” New Review of Film and Television Studies 6:2 (2008): 131-132.

[54] Sean Cubitt, The Cinema Effect (Cambridge, MA: The MIT Press, 2004) 228-230.

[55] CHC, 10-11.

[56] See especially Bordwell, “Neo-Structuralist Narratology and the Functions of Filmic Storytelling,” in Narrative Across Media: The Languages of Storytelling, Mary-Laure Ryan ed. (Lincoln: University of Nebraska Press, 2004).

To cite this article

Julien Lapointe, “Hollywood Cinema and Baroque Companionship: An Analytic Critique of David Bordwell’s ‘Classical Paradigm’”, Film Journal, 3 (2016). URL:
Back to top


Julien Lapointe is a doctoral candidate in Film and Moving Image Studies at Concordia University. His thesis combines cognitive theory and analytic philosophy to delineate a universal theory of the aesthetic experience of film texts. He has previously published in the journal Mise au Point (MAP) and the proceedings for Film Forum’s A History of Cinema without Names. He also has a chapter in The Legacies of Jean-Luc Godard (WLU Press, 2014) and a forthcoming publication in an anthology with Amsterdam University Press.
Back to top