The “Sight & Sound” of Canons

The question of film canons

by Donato Totaro Volume 7, Issue 1 / January 2003 11 minutes (2606 words)

With Peter Rist’s first of a two part look back at the best of the WFF, and the recent publication of Sight & Sound‘s decennial survey of Top Ten films of all-time, I thought it was a opportune time to take pen to paper on the question of film canons. No doubt pushed along by the recent millennium change, there has been an outpouring of published “great” or top lists. Outside of the list in the September 2002 Sight & Sound magazine (also online), there was the notorious AFI list of the 100 Greatest American Films, the BFI’s Top 100 British films, the ongoing “Top Ten” compilation at the wonderful online journal Senses of Cinema, "Top Tens or: Dare to a Cinephile," and Tim Dirk’s (also ongoing, also online) American-only database “The Great Films”. The exercise is harmless enough, as long as the context of who is doing the voting is always kept in sight. When the lists are left to stand on their own that is where misconceptions occur and this is what I’d like to concentrate on in the next few pages.

To begin, in most cases these lists are solicited without any onus on what a ‘great’ film is or what constitutes a ‘great’ film. There are no criteria or guidelines. That, of course, is part of the appeal and popularity of the list making: it is the great democratic leveler, since every list is ‘worth’ the same. But if there were criteria set in place, they would read something as follows: how well does the film stand the test of time (is it dated in any way)? has the film been influential on other films/filmmakers? was it seen at the time or in hindsight to have any groundbreaking or innovative features? has the film won many awards? is the director a recognized ‘auteur’? did it have any particular cultural significance? In fact the AFI list of the 100 Greatest American films set out the following voting criteria for its voting members: “Feature-Length Fiction Film,” “American Film,” “Critical Recognition Popularity Over Time,” “Historical Significance,” “Cultural Impact,” and “Major Award Winner.” But you have to wonder just how high historical awareness was on their minds when they limit the voting to ‘film professionals’ and restrict the voting by giving its members a pre-selection of 400 titles.

Although these and other criteria are no doubt used, consciously or not, by voters, there are other factors, some less noble, behind the reality of what films get chosen. After all, there are many great films, but why is it that certain films appear and others do not? An important factor is what I’d call the “omnipotence level” of a film, which refers to a film’s accessibility and visibility. Films which appear regularly on television, have had a long shelf life on home video formats (video, laserdisc, DVD), and screen regularly at repertory theatres, art houses, and school classrooms have a huge advantage over those that do not. When a film is easily available for viewing it opens itself up to critical exposure, which in turn leads to conference papers, essays, and books. A well-rounded exposure from both the academic and popular-commercial front is key to canonization. A case in point is Citizen Kane, incumbent number one on the Sight & Sound poll since 1962, in part because it has always ranked high among film critics/historians, filmmakers, while maintaining a high profile among the average filmgoer (it’s much more likely for an average filmgoer to at least have heard of Citizen Kane, than Battleship Potemkin or The Magnificent Ambersons). It may seem tautological and obvious, but a film can not be canonized unless it is seen frequently, continuously, and over a sustained period of time.

More unsettling is that the canon can be quite the self-serving monster. The recent AFI Top 100 American films is a case in point. As many critics were quick to point out, the survey was nothing more than a concerted ‘industry’ push to promote its products. Nearly every Academy Award Best Picture between its voting years (1896-1996) makes the list, and every one of the 100 films is available on home video (what a coincidence!). As the AFI states on its website, “And the 100 films will be available in video stores across the nation as part of this special celebration, with labels designating these films as part of AFI’s 100 greatest list.” 1 There are only four silent films on the whole list (no Buster Keaton, no Laurel and Hardy, no Harold Lloyd). As far as I am concerned, any list that purports to represent the Great American films that does not include Sunrise or The Magnificent Ambersons holds as much value as a lira.

On a less sinister note, canons are bound by subjectivity, and hence the formation of the voting majority will dictate what shape the canon will take. A study of the individual critic/director lists in the Sight & Sound poll clearly reveals this subjectivity. For example, there were 21 people in the 1992 poll and 34 in the 2002 poll who voted for three or more Asian films among their top ten. 2 Of this grouping, 13 out of 21 and 20 out of 34 were Asian. As a rule, this is evidence of a general national bias. Except where the UK is concerned, since the British seem to be notoriously frigid when it comes to voting for their own films. On the contrary, the British over compensate by tending to favor US films, and this Anglo-Saxon bias only helps to increase the already high presence of US members in the voting. Which helps explain the Anglo-Saxon bias in not only the Sight & Sound poll, but most Greatest Film lists. The most international of the Sight & Sound polls was in 1962, when only two of the ten films were Anglo-Saxon. This can be explained by the explosion of the European and Asian (namely Japan and India) new wave and art cinema directors that burst onto North American screens in the 1950’s and early 1960’s. Since then, however, there has been a not too surprising steady trend toward increased Anglo-Saxon presence: from 27% in 1972 (3 out of 11), to 55% in 1982 (6 out of 11), to 44% in 1992 (8 out of 18), to 59% in 2002 (10 out of 17).

A bias which is far more interesting and in keeping with the passionate nature of list and canon making is when critics and filmmakers vote according to their particular critical interest or personal taste. So, for example a scholar such as Ed Buscombe (1992 poll), who has done extensive writing on the western, lists three westerns among his top ten films; and John Carpenter does not hide his love of American machismo classicism by stacking his list with seven films directed by either John Ford, Howard Hawks, or Alfred Hitchcock; while horror director’s Wes Craven, Stuart Gordon and George Romero feature many horror/fantasy titles among their top ten.

The almost complete disregard for the documentary, animation, short, and experimental cinemas, leads to the “idiosyncratic” or “polemical” list, where voters stack their films to one extreme to make a political point. So in the 1992 Sight & Sound list Jonas Mekas includes 9 silent films and five experimental works, Stan Brakhage 9 experimental films, and Michael Snow 10 experimental films. Sometimes it is hard to tell when sarcasm is used, but a good bet for such an occasion is the 2002 list from Indian director Anurag Mehta, which comprises exclusively of American Big Box-Office hits, including such mediocre stuff as Forrest Gump, Superman, Back to the Future, Jerry Maguire, and Rocky (I hope he’s kidding!).

And then there is the “saintly” list, where the voter makes a valiant attempt to accommodate as many cinematic ‘territories’ as possible. Jim Hoberman’s 1992 (and his slightly revamped 2002) top ten is exemplary here, as it includes silent films (Fantômas, Man with a Movie Camera, The Lonedale Operator), a film by an African-American (Oscar Micheaux’s God’s Stepchildren), a woman (Chantal Akerman’s Jeanne Dielman, 23 Quai du commerce, 1080 Bruxelles), an Asian film (Pather Panchali), experimental films (Scenes from under Childhood), a short film (Rose Hobart), a European art house film (Two or three Things I Know About Her), and an American classic (Vertigo).

In “Light my Fire: The Geology and Geography of Film Canons” Australian film critic Adrian Martin distinguishes three types of canons: the “Star Wars” canon, which constitutes the populist, commercially successful films. The AFI canon is an example of this. The “Citizen Kane” canon, which is the old guard canon that clings to the 1950’s, 1960’s flourishing of the art house and New Wave periods. The Sight & Sound decennial poll is an example of this type of canon. The third canon is what Martin proposes as an antidote to the biases and omissions of the “Citizen Kane” canon (Martin simply disregards the "Star Wars" canon), most of which I noted earlier: dearth of experimental/avant-garde films, short films, films by women, popular genres, documentary, flawed films, and non-auteurist films. These are the films which gain minor representation only in the odd ‘dissident’ voter, such as Snow, Mekas, Brakhage, Hoberman, etc. Martin calls this third hypothetical alternative canon the “Kiarostami Canon” (an odd choice to represent underground cinema, exploitation, low budget, etc.). Martin’s own “Kiarostami Canon” stands as follows:

  1. Killing of a Chinese Bookie (John Cassavetes, 1978)
  2. Alone. Life Wastes Andy Hardy (Martin Arnold, 1998)
  3. L’Atalante (Jean Vigo, 1934)
  4. The River (Tsai Ming-liang, 1997)
  5. Haut bas fragile (Jacques Rivette, 1995)
  6. The Night of the Hunter (Charles Laughton, 1955)
  7. Phenomena (aka Creepers, Dario Argento, 1984)
  8. The Emperor’s Naked Army Marches On (Kazuo Hara, 1987)
  9. The Spirit of the Beehive (Victor Erice, 1973)
  10. F For Fake (Welles, 1974)

Will the Sight & Sound poll ever shake its “Citizen Kane” canon status? What will the results look like in 2012? Ten years seems like a long time, but the canon shifted very little in between the last two decennial polls; or even in the last twenty years if you consider that 8 of the 11 films from the 1982 list appear on either the Critics or Directors list of 2002. I imagine it will take several decades before the Sight & Sound canon shifts dramatically, but if I were forced to predict where such a shift may occur I would say with the appearance of one or perhaps two of the many excellent Asian films which have been gaining critical success and international recognition throughout the 1980’s and 1990’s. One of the most enlightened of the many millennium inspired list makings is the “Best of the Nineties” issue of the Canadian film journal Cinéma Scope (Winter 2000, Issue 2). In a poll of over 150 film critics, programmers, and directors from 15 countries, four Asian directors appeared among the top ten: Abbas Kiarostami, Wong Kar-wai, Hou Hsaio-hsien, and Takeshi Kitano (and two more among the next top ten, Tsai Ming-liang and Edward Yang).

Although the number of Asian films appearing on both the Critics and Directors Lists in 1992 and 2002 dropped from 4 to 3 (Pather Panchali being dropped with Tokyo Story, Rashomon, and Seven Samurai carried over), there may be an indication of a contrary trend if one scratches below the surface. Disregarding the two Top Ten lists, there was actually a considerable increase in the number of Asian films that were cited across the Critic and Director lists from 1992 to 2002. Here is a graph which reveals this subtle below-the-surface increase:

1992 Sight & Sound Poll Results 2002 Sight & Sound Poll Results

Number of Critics Who Voted: 132

Number of Critics Who Voted:145

Number of Times an Asian Film is Cited: 142 Number of Times an Asian Film is Cited: 197
Overall Percentage of Asian Films: 11% (142 out of 1320) Overall Percentage of Asian Films: 14% (197 out of 1450)
Number of Directors Who Voted: 101 Number of Directors Who Voted: 107

Number of Times an Asian Film is Cited: 74

Number of Times an Asian Film is Cited: 102

Overall Percentage of Asian Films: 7.3% (74 out of 1010) Overall Percentage of Asian Films: 9.5% (102 out of 1070)

Total Overall Percentage of Asian Films: 9%

Total Overall Percentage of Asian Films: 12%

Hence the overall number of Asian films cited went up from 216 in 1992 to 299 in 2002 (9% to 12%). Although a three percentage increase may not appear considerable, it is, I think, a direct result of two interrelated factors that have taken hold over the last 10-15 years: a) an increasing critical and academic acknowledgement and appreciation of Asian cinemas and b) the increased visibility of Asian films in film festivals, specialty cable television stations, and, perhaps most importantly, the DVD market. The DVD format has far surpassed laserdisc in reaching mainstream market places, aided considerably by online mail order sales. Even something as innocuous as the smaller, lighter format of DVD makes a considerable difference in mail order purchasing. (I missed the laserdisc bandwagon, but I can only imagine what the shipping costs must have been for large orders of those bulky laserdiscs!). Esoteric films, including horror, eurocult, art house, Asian, and foreign films are now more available than ever, especially for the consumer willing to do some legwork on the internet.

Since canons are notoriously backward looking and slow to shift, this recent popularisation of Asian cinema may just reap canon reward in twenty or so years time. So here it goes, I’ll stick my neck out and predict that in the Sight & Sound poll of 2012 at least one film from one of the follow directors will appear on the celebrated Sight & Sound Top Ten list: Abbas Kiarostami, Hou Hsiao-hsien, Wong Kar-wai, Sang-soo Hong, Takeshi Kitano, Mohsen Makhmalbaf, Jafar Panahi, Edward Yang, Tsai Ming-liang, or Jia Zhang-ke.

Ten years is a long wait. In the interim I grudgingly offer my own list as a “conservative” assault on the old guard “Citizen Kane” canon (not that I have anything against Citizen Kane, I think it is a ‘great’ film), in no particular order:

  • A Moment of Innocence (Mohsen Makhmalbaf, 1996)
  • Stalker (Andrei Tarkovsky, 1979)
  • Black Sunday (Mario Bava, 1960)
  • Sátántangó (Bela Tarr, 1996)
  • Stranger than Paradise (Jim Jarmusch, 1984)
  • Earth (Alexander Dovzhenko, 1930)
  • Kwaidan (M. Kobayashi, 1964)
  • Wavelength (Michael Snow, 1967)
  • The Battle of San Pietro (1945, John Huston)
  • The Music Box (1932, Laurel and Hardy, dir. James Parrott)

As an antidote to the Anglo centric bias of so many of these recent favorite/great lists I have compiled the top ten Asian films as voted on by 34 mainly Asian film critics, programmers, and academics in the 10th anniversary issue of the important Asian film magazine, Cinemaya. The issue, published in 1998, listed each voters top ten list but did not tabulate the results, which are as follows (extended to allow for ties):

  • Tokyo Story (Yasujiro Ozu, 1953, Japan, 19 votes)
  • Pather Panchali (Satyajit Ray, 1955, India, 13 votes)
  • Ugetsu (Kenji Mizoguchi, 1954, Japan, 13 votes)
  • Ikiru (Akira Kurosawa, 1952, Japan, 10 votes)
  • Seven Samurai (Akira Kurosawa, 1954, Japan 10 votes)
  • Where is My Friend’s Home? (Abbas Kiarostami, 1987, Iran, 9 votes)
  • The Apu Trilogy (Satyajit Ray, 1955-59, India, 8 votes)
  • Yellow Earth (Chen Kaige, China, 1984, 8 votes)
  • A Time to Live and a Time to Die (Hou Hsiao-hsien, Taiwan, 1986, 6 votes)
  • A City of Sadness (Hou Hsiao-hsien, Taiwan, 1989, 6 votes)
  • (tied with 4 votes)
  • Charulata (Satyajit Ray, 1964, India)
  • Floating Clouds (Mikio Naruse, 1955, Japan)
  • Mandala (Im Kwon-Taek, South Korea, 1981)
  • The Music Room (Satyajit Ray, 1958, India)
  • Spring in a Small Town (Fei Mu, China, 1948)
  • Subarnarekha (Ritwik Ghatak, India, 1962-65)


  1. The 100 Greatest American Films
  2. For the purpose of designating what constitutes an Asian film I was guided by the nation profile outlined by the online site “AsiaSource”: East Asia (China, Hong Kong S.A.R., Japan, Mongolia, North Korea, South Korea, Taiwan), South Asia (Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, Sri Lanka), Central Asia(Afghanistan, Iran, Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, Uzbekistan), Southeast Asia (Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, Philippines Singapore, Thailand, Vietnam), and Australasia (Pacific Island Countries, Papua New Guinea).

The “Sight & Sound” of Canons

Donato Totaro has been the editor of the online film journal Offscreen since its inception in 1997. Totaro received his PhD in Film & Television from the University of Warwick (UK), is a part-time professor in Film Studies at Concordia University (Montreal, Canada) and a longstanding member of AQCC (Association québécoise des critiques de cinéma).

Volume 7, Issue 1 / January 2003 Essays   film canon   film history