When the 2002 poll came out I revisited the old ones to see how critical opinion has changed and make comments (some doubtless uninformed or inappropriately snide) For ease of discussion, I've compiled the various lists below. (I should note that all materials here are copyright whoever owns the copyright.)

The first thing that strikes me about the new poll is the cautious gerrymandering. The great national film cultures are now so evenly represented that it's as if there was a top ten slot reserved for Japan, Germany, and so on. Averaging hundreds of international ballots should produce some geographic diversity but even looking at the individual ballots it seems that most voters tried be geographically even-handed all on their own.

If you're one of hundreds of voters you really should vote your idiosyncrasies, not try to guess what the average of everyone else's votes will be.

This relativist effect afflicts most "best of" rankings. Top 10 rock album polls always include one Beatles album, one Rolling Stones album, one Springsteen album... what are the odds of that? I happen to think Elvis Costello was operating at a higher level than any other recording artist in the 1980s. If I vote in a top ten 1980s albums poll I should submit a ballot that's at least half Costello albums. Otherwise my strong view wouldn't be properly represented in the broad average. Ideally some individual Sight and Sound voters should pick all Japanese films while other pick all Hollywood films. Instead it's as if everyone was envisioning their own film history survey course.

I don't have a statistical breakdown of this, but looking over the raw data it appears that American voters were more likely to include foreign films while international voters picked more American films. What a grim indictment of film criticism that is! If I think THE SEVEN SAMURAI is super-profound but Japanese voters disagree then I am probably WRONG. I don't speak Japanese or understand Japanese cultural subtleties. I cannot assess the nuances of script or performance beyond a rudimentary silent-film level.

There are valid examples of critics finding gold in another country's unappreciated native cinema, like the French discovery of American film noir. There are also examples of people striking artistic fools-gold because they lack a full cultural understanding of a work, like the French discovery of Jerry Lewis. Any American who likes Jerry Lewis movies much is a frigging idiot.

Since so many Sight and Sound voters see the grass as greener in other countries it suggests that much of film appreciation is a Rorshach effect.

Dilution by region is matched by dilution by director, with no one person being responsible for more than one of the 10 greatest films. That's highly suspect. Wouldn't you think a director capable of making one of the ten greatest films might be able to make two or three of them? If any Alfred Hitchcock movie belongs in the top ten then two or three belong there because he made a lot of roughly comparable great films.

Another difficulty with the polling process is voters confusing historical importance and artistic quality. If there was such a thing as aviation criticism the Wright brothers' plane might be one of the "ten greatest planes of all time," despite barely even being able to fly. Every one of the actual ten greatest planes is, however, obviously all in service today if greatness involves capacity, lift, speed, durability, safety, or any other trait considered desirable in an airplane.

Art history as a catalogue of innovations is no more useful than European history as a list of kings. Demoiselles d'Avignon is considered the first cubist painting, and is thus one of Picasso's best known paintings and occupies a full page in most art survey texts. But it's an awful panting.  Picasso did many excellent cubist paintings later on and students would learn more about cubism looking at a more mature work. Picasso spent years exploring cubist ideas but instead of showing students the mature result of years of the great man's thinking we show an impulsive first draft.

BATTLESHIP POTEMKIN is a VERY historically important film because Eisenstein's approach to montage influenced almost all successive films. The approach to montage in BIRTH OF A NATION influenced ALL successive films without exception. No work is more important to its medium. BIRTH OF A NATION is more important to film than Shakespeare is to drama. That said, hundreds or thousands of films have been made since 1915 that are better films. If BIRTH OF A NATION doesn't belong an a modern 10 best list, which it doesn't, then does BATTLESHIP POTEMKIN really belong?

The most glaring anomaly in the 2002 Sight and Sound poll is that the most  recent films represented are over 30 years old. I suspect that half of all films ever made were probably made in the last thirty years so it's a notable omission. That said, the lack of recent films is probably for the best. Whenever critics reach consensus on the greatness of something recent it's almost invariably a trendy, gimmicky political work that will be forgotten in a few years.

Page 2: About the picks

Persona (Bergman)

L'avventura (Antonioni)

La Règle du jeu (Renoir)

Seven Samurai (Kurosawa)

La Règle du jeu (Renoir)

Le Million (Clair)

Brief Encounter (Lean)

Ugetsu Monogatari (Mizoguchi)

Wild Strawberries (Bergman)

The Searchers (Ford)

Rashomon (Kurosawa)

Seven Samurai (Kurosawa)