When Hollywood defines the limits of good cinema

Posted On By Blair Fix

When Hollywood defines the limits of good cinema

May 10, 2022

Originally published at notes on cinema

James McMahon

On the question of who judges the quality of a film, it is easy to start with a notion that the ultimate judge of a film’s quality is the individual moviegoer. As individual moviegoers, this is often what we think we are doing: we have the autonomy “decide for ourselves” if a film is good or bad. Picturing each moviegoer as having a separated, independent subjective experience with a film makes sense on a physical level — a film’s images and sounds are being perceived with my eyes and ears. The reality of film criticism, however, is misunderstood if we imagine the social world of cinema to be simply composed of individuals exercising independent judgment. Hollywood is actively shaping this social world. What effect does it have on our conceptions of a good film?

As a social institution with considerable investment in the financial future of cinema, Hollywood studios are adept at promoting one-dimensional definitions of good cinema. A one-dimensional definition of good cinema, which I am conceptualizing with the critical theory of Herbert Marcuse, does not include the aesthetic and technological potentials of cinema — i.e., what cinema could be. Rather, a one-dimensional definition is more immediate, according to what is actually being watched. The actual realities of film consumption are important to Hollywood because inequalities in the distribution of films limit the existing reality of the average moviegoer, whose perspective on cinema is much narrower than what is technically available in film libraries, universities and the collections of film lovers. Thus, the potential of cinema does not simply live in the mind, as pure theory; it includes films that are neglected in common ideas of what people think the film medium can achieve.

The starkest examples of a one-dimensional definition of good cinema are found in Hollywood’s advertisements. With a perverse hope that the audience has cultural amnesia, Hollywood will tell you that its next film is promised to be a cinematic experience of the highest quality. Yet the one-dimensional definition of good quality comes from more than the superlatives of film advertising. Hollywood’s definition of good cinema is institutional at its core, which has effects on such factors as the level of public access, availability of cinematic alternatives and the strength of cultural education.

For example, Jonathan Rosenbaum‘s experience as a film critic enabled him to see the institutional barriers that protect Hollywood’s position in definitions of good cinema. His book Movie Wars enumerates how the Hollywood film business affects the way film journalism can explore the histories and geographies of film art. Often excluded from the category of “good” cinema are films that editors think will be too alien to their readership. The independent, avant-garde or just plain weird might find some amount of journalistic coverage, especially in the age of the internet, but Rosenbaum’s experience at film festivals, The Chicago Tribune and his guest television appearances on Chicago Tonight all reveal there is an underlying inequality in attention. Not only does mainstream journalism choose to stay within the boundaries that Hollywood helps define, but film critics also turn into pseudo-marketers:

Consider what might happen if Roger Ebert couldn’t find a single movie to recommend on one of his weekly shows. Or let’s assume that this has already happened once or twice. How much freedom would he have to assign a thumbs-down to everything three or four weeks in a row without getting his show canceled? And for all the unusual amount of freedom I enjoy at the Chicago Reader, how long could I keep my job if I had nothing to recommend week after week? For just as Communist film critics were “free” to write whatever they wanted as long as they supported the Communist state, most capitalist film critics today are “free” to write anything as long as it promotes the products of multicorporations; the minute they decide to step beyond this agreed-upon canon of “correct” items, they’re likely to get into trouble with their editors and publishers.

Rosenbaum, 2000, p.54

Rosenbaum’s criticism is similar to what Adorno and Horkheimer said about the power of scale in modern advertising: large capitalist firms can advertise their products to such a degree that we come to associate the quality of an object with the amount of advertising or publicity it gets. This power of scale also binds the profession of film criticism to a business that is not in the habit of admitting to the quality of what lies beyond its own boundaries. For instance, the journalistic conspiracy of silence regarding Bela Tarr’s seven-hour-long Sátántangó was, according to Rosenbaum, a means of ignoring hard truths about the institutional repression of aesthetic potential. If other film critics, like Rosenbaum, thought the film’s long shots and extremely slow pace excellently captured the philosophical themes of nihilism and authoritarianism, it would challenge, even in some small way, the rationale of the mainstream film business (Rosenbaum, 2004, p.48). If its length of seven hours is not excessive in terms of art, the high quality of Sátántangó shakes the illusion that Hollywood’s shorter films are the reflection of universal laws about the duration of good films. Additionally, the film reveals that there is an implicit business risk to artists making great films: “If great films invent their own rules,” writes Rosenbaum, Bela Tarr demonstrated that one can create a type of masterpiece that cannot be covered in the national media (Rosenbaum, 2004, p.48).

Rosenbaum gives another relevant example (Rosenbaum, 2000, pp.91-16). In his opinion, the American Film Institute (AFI) betrays its mandate to honour “the most outstanding motion pictures” because its acts of honouring American cinema hardly ever stray from the films of major Hollywood studios. This produces an ideological echo in the AFI’s lists of “top” films. Some of the listed films are outstanding in their own right, but the AFI uses its institutional power to tell people what they already think — that Casablanca, The Wizard of Oz, Gone with the Wind, E.T., The Godfather and Star Wars are outstanding films in the history of cinema. As should be the case in film journalism, so should it be the case with the AFI: neglected films in independent, alternative or foreign cinema should be pulled out from under the shadows of Hollywood, rather than get pushed deeper into the darkness. Even though the AFI is only concerned with American cinema, Rosenbaum claims that it is not difficult to produce much more representative lists of what has been outstanding in all of its history.

Rosenbaum’s critical views of Hollywood’s dominance rely on inter-connected arguments: that the aesthetic qualities of mainstream Hollywood are overvalued and that alternatives like Sátántangó are undervalued. I agree with Rosenbaum but I also recognize that the argument is dependent on how one rates his ability to judge film quality. Consequently, I think we can corroborate a part of Rosenbaum’s argument in way that relies less on his likes and dislikes of specific films in the history of cinema.

Rosenbaum’s argument presents Hollywood as an institution that can influence the chances of critics and consumers wandering outside the boundaries of Hollywood’s familiar territory. It is unlikely we can clearly delineate this territory at the scale of individual films; at this small scale, subjective predilections and differences in personal experience create many opportunities to move the line that defines what is included in “mainstream Hollywood”. We can, however, investigate whether a larger set data on Hollywood cinema spreads across the sample space of a measurement. In probability theory, a sample space is the set of all possible outcomes of an experiment. For example, the sample space of drawing a single card from a 52-card playing deck is the set of all single cards in the deck: {A♡, 2♡, 3♡, …}. In our case, we can use probability to randomly draw a film and use the trial to investigate if the inequality of theatrical film distribution is affecting the likelihood of someone seeing the film.

As Ian Hacking shows, the same physical object (e.g., deck of cards, coin, film) can be the starting point of a broad range of experiments in probability, ranging from simple trials to complex ones that incorporate conditionals into the results. Thus, the definition of a sample space is determined by the “chance set-up”, which is Hacking’s phrasing to explain how a “device or part of the world” is used to conduct trials:

A piece of radium together with a recording mechanism might constitute a chance set-up. One possible trial consists in observing whether or not the radium emits radiation in a small time interval. Possible results are “radiation” and “none”. A pair of mice may provide a chance set-up, the trial being mating and the possible results the possible genetic make-ups of the offspring. The notion of a chance set-up is as old as the study of frequency.

Hacking, 1965, p. 12

Our chance set-up starts with the following question: if I randomly select a film from a group of films, what is the expected Motion Picture Association (MPA) film rating (G, PG, PG-13, R)?

Why should we investigate the expected MPA film rating? The MPA film-rating system is most frequently seen from the perspective of the consumer, who understands the rating system as information about much or how little “objectionable” content is in every theatrically released film. From a different perspective, however, one can understand the MPA film-rating system to be one of Hollywood’s non-governmental means to control the creation and distribution of artistic creativity. For example, artistic labour is channeled into ranges of subject matter, language and imagery by the requirement that a North American theatrical release must have a MPA film rating. Moreover, there is a ceiling that “adult-themed” films can bump against: the infamous “X” or NC-17 rating. This is technically the fifth rating, but it is virtually a taboo rating to all major theatre exhibitors (Lewis, 2002). Thus, the film directors in Hollywood are effectively tethered to four “acceptable” tiers before anything is shot. As Lewis notes, contracts oblige directors to “deliver their film as G, PG, PG-13 or R” (Lewis, 2013, p.43). Film directors must also appeal or agree to the changes the MPA’s Classification and Rating Administration says are needed to approve the final cut of the film for theatrical release.

Figure 1 presents the distribution of all of the films that are included in our investigation. In total there are over 11,000 theatrical releases in the dataset. To account for the presence of re-releases, only the first appearance of each unique IMDB film ID is kept in the dataset. For example, The Polar Express was originally released in 2004, but it had limited re-releases for the next fifteen years. We can use the distribution in Figure 1 to calculate the probabilities of randomly selecting a film rating from the entire population: G (0.028), PG (0.159), PG-13 (0.275), and R (0.536).

Figure 1: Film count by rating, 1983-2019
Source: IMDB

Figure 1 shows that R-rated films are the majority rating in the entire population. If the R rating is where creative filmmakers want to be — at least in principle — the distribution appears to show there is an abundance of R-rated films in our cinematic universe. With data from 1985 to 1996, De Vany and Walls (2002) investigated if an over-abundance of R-rated films has had any effects on the business of Hollywood. Their paper argues this output of so many R-rated films is a mistake for business interests. Hollywood might be pressured to cater to filmmakers who gravitate to the R rating out of desires for creativity and prestige, but De Vany and Walls believe the views of studio executives should change: “An executive seeking to trim the ‘down-side’ risk and increase the ‘upside’ possibilities in a studio’s film portfolio could do so by shifting production dollars out of R-rated movies into G-, PG-, and PG13-rated movies” (De Vany & Walls, 2002, p. 426). However, De Vany and Walls might be preaching to the converted. When we introduce the inequality of film distribution as a variable, we can see how the R-rated film does not dominate the “core” of Hollywood theatrical distribution. Rather, R-rated films, which are numerous as a group, are marginalized in the wide-release strategy, which is key to blockbuster cinema.

De Vany and Walls overlook this point because their analysis implies that Hollywood is responsible for all of the films in their dataset — they often describe the distribution of films as the set of films in Hollywood’s “portfolio” (De Vany & Walls, 2002, p. 438). It is very likely that all of their films are distributed by American firms; it is also likely that many of the films are “inside” Hollywood as they are distributed by a major studio or one of its subsidiaries. However, the general usage of “Hollywood” minimizes Rosenbaum’s critique and also obscures the realities of some films getting much more corporate attention than others. By using opening theatre rankings, we can group films by their relative places in the annual distributions of opening theatres (e.g., the film with a ranking of 4 was the 4th largest opening of its year). This ranking is very important in determining the distribution of promotional and advertising budgets across a film-release schedule. The low-ranking film, for instance, is released into the world on small advertising campaigns; they rely on word-of-mouth behaviour to make the audience grow. The high-ranking film in opening theatres has an entirely different birth. They enter the theatrical market across entire countries and receive aggressive levels of promotion and advertising.

Table 1 outlines how bins of films are created from the population. For each year, films are labeled according to the percentile of their opening theatre ranking. There are five labels: [Blockbuster, Major, Medium, Limited, Very limited]. The second row of the table indicates where, on average, the cut between bins is made. For example, the average cut between “Major” and “Medium” occurs at 89.8, which makes the last “Major” film 89 in rank and the first “Medium” film 90.

Category: Blockbuster Major Medium Limited Very Limited
percentile 0.9-1 0.8-0.89 0.6-0.79 0.3-0.59 0-0.29
inner cut, opening theatre rank, mean 44.8 89.8 178.4 310.8
op. theatres, mean 3073.3 2027.8 577.1 77.9 2.1
Table 1: Binning films by opening theatre rank

Each category of films has its own sample distribution of ratings. A measure of proportional representation can be calculated with a ratio between the distribution of ratings in each sample and the distribution of ratings in the entire population. For example, 17.6 percent of the films in the “Major” category are PG rated. In the entire population, 15.9 percent of the films are PG rated (see Figure 1). This means there is a slight over-representation of PG-rated films in the “Major” category (17.6/15.9 = 1.1). Conversely, a ratio less than 1 signals an under-representation, proportional to the entire population of films in the dataset.

Figure 2 plots the proportional representation of each rating for each category. From this perspective, the abundance of R-rated films looks different. There is a visible under-representation of R-rated films in the “Blockbuster” and “Major” categories. Consequently, the more likely outcome for the R-rated film is to have a limited release of some type.

Figure 2: Proportional representation of MPA film ratings, by category, 1983-2019
Note: See Table 1 for method to bin films into five categories. Source: IMDB.

Figure 3 helps demonstrate that the influence of opening theatre rank on film rating has a historical trend. For the plot in Panel A, the average film rating was taken from a random sampling of 1 million draws. A numerical average was produced by making the G-to-R scale a numerical scale from 1 to 4. The random sampling occurred on each year (e.g., 1 million samples from the “Blockbuster” category in 1983, 1 million for 1984, …). The time series in Panel A show the “Blockbuster” category hovering around PG-13, where it is the other categories that have averages closer to 4, the R rating. Panel B of Figure 3 indicates that deviations from the averages are also reducing over time. This suggests that the MPA rating system is more entrenched, whereby there is less variance in the ratings-release relationship.

Figure 3: Mean and standard deviation of MPA film ratings, by category, 1983-2019
Note: See Table 1 for method to bin films into five categories. Source: IMDB.

There are some limits to how this experiment can be connected to Rosenbaum’s argument. For instance, our perspective is too broad to see the reasons why any film receives its MPA rating. Moreover, the approach risks suggesting that R-rated films are “outside” Hollywood, when it is also entirely possible that a radical alternative to a Hollywood film contains no R-rated content. Nevertheless, opening-theatre rank is a helpful method for seeing trends in the distribution of our social attention — the landscape is neither flat nor open to a moviegoer randomly finding a film like Sátántangó. If blockbuster cinema is Hollywood’s main territory, it has issues of over- and under-representation within. The film released nation-wide and with lots of advertising and promotion is, relative to other releases, more likely to be PG-13 or below. The R-rating, which includes many films that receive this rating for how they address social issues of race, gender and sexuality (e.g., 12 Years a Slave, Brokeback Mountain, Carol, Do the Right Thing, Moonlight, Revolutionary Road), are on the margins by virtue of their likelihood of not receiving wide-release distribution strategies and lots of corporate attention in advance of a theatrical run.

Further Reading

McMahon, J. (2022). The political economy of Hollywood: Capitalist power and cultural production. Routledge.


De Vany, A. S., & Walls, W. D. (2002). Does Hollywood Make Too Many R‐Rated Movies? Risk, Stochastic Dominance, and the Illusion of Expectation. The Journal of Business, 75(3), 425-451.

Hacking, I. (1965). Logic of statistical inference. Cambridge University Press.

Lewis, J. (2002). Hollywood v. Hard Core: How the Struggle over Censorship Saved the Modern Film Industry. New York University Press.

Lewis, J. (2013). “American morality is not to be trifled with”: Content Regulation in Hollywood after 1968. In D. Biltereyst & R. V. Winkel (Eds.), Silencing Cinema: Film Censorship Around the World. New York: Palgrave Macmillan.

Rosenbaum, J. (2000). Movie Wars: How Hollywood and the media conspire to limit what films we can see. Chicago, IL: A Cappella Books.

Rosenbaum, J. (2004). The Importance of Being Sarcastic: Sátántangó. In Essential Cinema: On the Necessity of Film Canons. Baltimore: Johns Hopkins University Press.