Seijun Suzuki: The Pop-Art Provocateur

In 1968 the president of Nikkatsu sacked one of his contract directors and gave a reason for the record: the man made films that made no sense and no money. It is the most accurate insult ever aimed at a great filmmaker, and Seijun Suzuki spent the following half-century turning it into a job description.

He had been at the studio for twelve years, grinding out roughly forty pictures on the B-programme — the second half of a double bill, shot in three or four weeks on scripts he had no say over and largely despised. The Nikkatsu system existed to feed cinemas. It handed a director a genre, a star under contract, a schedule and a budget, and expected a competent ninety minutes. Suzuki delivered the ninety minutes every time. He also, gradually and then with increasing brazenness, began doing something else inside them.

The war, the studio, the trade

He was born in Tokyo in 1923, was drafted, and served in the Pacific; his troop transport was torpedoed, twice, and he survived both. He came home, failed the university entrance exams, saw an advertisement for Shochiku’s assistant-director programme in 1948 and took it because it was there. He moved to Nikkatsu in 1954 and started directing in 1956.

That origin matters for the same reason Fukasaku’s factory matters. A man who has been sunk twice and washed up on a beach does not treat a studio contract as a calling. Suzuki was, by every account including his own, entirely without artistic pretension in conversation. He turned up, he was handed a script about a hitman or a brothel or a wandering tough, he found it dreadful, and he made it interesting for his own amusement.

Sabotage as a method

Youth of the Beast (1963) is where the thing announces itself. Jo Shishido — the actor who had cosmetic implants put in his cheeks to give himself a face nobody could forget — plays a man infiltrating two gangs at once. The plot is standard programme material, and Suzuki treats it as an excuse to build things. A gangster’s office has a one-way window onto a cinema screen, so a beating plays out in silhouette against a projected film. A sandstorm turns the frame yellow for a scene, for no reason connected to anything. The blood is too red.

Gate of Flesh (1964) puts prostitutes in the ruins of occupied Tokyo and colour-codes them — each woman assigned a dress in a primary shade, so that the compositions read as abstraction while the content is the most brutal thing he ever filmed. Story of a Prostitute (1965) and Fighting Elegy (1966) push further.

Then Tokyo Drifter (1966), where the sabotage becomes total. The hero wanders a Japan of pure design — a nightclub that is a white void with a yellow staircase, a snowfield with a single colour in it, walls that fly out of shot mid-scene to reveal the soundstage. The theme song is sung by the hero, whistled by the hero, played on the soundtrack over the hero, until the film seems to be about the song. Nikkatsu had asked for a yakuza picture with a pop star in it and got a delirium.

The system he was sabotaging

The context makes the sabotage legible, so it is worth setting out. Nikkatsu in the sixties was selling mukokuseki akushon — “borderless action”, a genre of gangster pictures deliberately stripped of any recognisable Japan. The heroes wore Stetsons and drove American cars, the towns had no names, the geography was nowhere. It was a commercial strategy aimed at a young audience who wanted Hollywood and could be sold a domestic imitation for less.

Suzuki took that instruction at its word. If the world of the film is officially nowhere, then nowhere can look like anything, and a director who has been told that realism is off the table cannot be accused of abandoning it. His flying walls and impossible corridors are the studio’s own premise followed to the end by the one man on the lot willing to do so, and Hori’s fury when he finally watched Branded to Kill has the particular heat of a man discovering that his employee has been obeying him.

The writing arrangement deserves a mention too. Branded to Kill is credited to Hachiro Guryu, a pen name covering a collective of writers who worked on it together, rewriting through the night as the shoot ran. That collaborative churn is why the film’s plot behaves like a dream someone is having while being interrupted, and it is also why no single intention can be extracted from it. Suzuki’s crews were small, permanent and loyal, and the pictures were made at a speed that left no room for a producer to intervene until the negative was cut. The B-programme’s contempt for its own product was the freedom: nobody senior watched the second half of a double bill until it was too late.

Branded to Kill, and the sacking

Branded to Kill (1967) is the one that ended it. Jo Shishido plays Hanada, the number three killer in Japan, who is sexually aroused by the smell of boiling rice, botches a hit when a butterfly lands on his rifle sight, and is then hunted by the mysterious Number One, who moves into his flat and will not let him urinate alone. It was shot in twenty-five days in monochrome on a script Suzuki and a collective of writers rewrote nightly.

Kyusaku Hori, Nikkatsu’s president, watched it, withdrew it from release, and fired Suzuki. The studio also pulled his back catalogue from circulation, which meant that a director could no longer be seen even in revival. Suzuki sued for breach of contract. Student film societies, critics and other directors mobilised behind him — the case became a genuine cultural flashpoint in a country already in ferment — and he won a settlement in 1971. He was then blacklisted from directing features for roughly a decade. Winning cost him ten years of work.

Why it works

The signature is easy to describe and very hard to imitate, and every account of it has to start with a name that is not Suzuki’s. Takeo Kimura was his production designer, and the two men built the style together in the gap between what the script required and what the studio could check. Kimura would construct a set that was frankly impossible — a room with no fourth wall, a corridor that ends in colour, a floor of water — and Suzuki would shoot it as though it were an office.

The mechanics underneath are these. He flattens. The camera sits square on, the lens is long, the depth is compressed, and the result reads as a poster in motion. He uses colour non-diegetically: a wash of red or yellow arrives because the scene needs it emotionally, with no light source, decades before that was ordinary. He cuts on impossibility — a character leaves a room and arrives in a landscape, and the film declines to explain the join. And he treats the plot as a rumour. Suzuki’s films remain legible only if you stop tracking the story, which is exactly the move his audiences in 1966 refused to make and his admirers since have made instantly.

The comparison worth drawing is with Mario Bava, working the same years on the other side of the world, under the same conditions: a contract, a genre, no money, and a designer’s eye. Danger: Diabolik and Tokyo Drifter are the same film made by men who never met — pop-art crime fantasias in which the design is the subject and the plot is a delivery mechanism. Neither director thought he was making art. Both were the only person on set who could see what the set could do.

Against Suzuki stands Kinji Fukasaku, who took the same yakuza material apart with a handheld camera and a documentary rage. Between them, over about a decade, they left the Japanese chivalry film with nothing to stand on.

The wilderness and the return

The blacklist years were real poverty. He wrote, he taught, he did television, he acted in other people’s films. He returned to features with A Tale of Sorrow and Sadness (1977) and then made the work his Japanese admirers rate highest: the Taisho trilogy of Zigeunerweisen (1980), Kagero-za (1981) and Yumeji (1991) — ghost stories of the 1920s, slow, gorgeous and structurally unhinged. No distributor would touch Zigeunerweisen, so its producer screened it in an inflatable dome pitched in city centres. It won the Japan Academy prize for best picture and was later voted the finest Japanese film of the decade by Kinema Junpo, which is a satisfying thing to happen to a man sacked for making no money.

He kept going. Pistol Opera (2001) revisits Branded to Kill in saturated colour with a female assassin. Princess Raccoon (2005), with Zhang Ziyi, is a musical fairy tale made at eighty-two. He died in February 2017, aged ninety-three, by then the most quoted Japanese genre director alive: Branded to Kill is watched on screen in Jarmusch’s Ghost Dog, and the flat colour, the impossible cuts and the pop-art gangster owe him from Wong Kar-wai to Park Chan-wook to Tarantino to Refn.

The case against, and where to start

The films are frequently incoherent, and the incoherence is not always productive — for every wall that flies away to reveal a soundstage, there is a plot turn that simply fails because Suzuki lost interest in it. The early forty are mostly what they were paid to be. He was indifferent to actors beyond their silhouettes, and Jo Shishido’s cheeks are doing a lot of characterisation that a script might otherwise have supplied. A viewer who needs a story to hold will find him maddening, and the maddening is the medium.

Start with Tokyo Drifter, which is eighty-three minutes and the best possible advertisement. Then Youth of the Beast, which is the most conventional and the most fun. Branded to Kill is the deep end and should be taken third, after the grammar is in your hand. Gate of Flesh is the one that will surprise anyone who thinks he was only a stylist. And Zigeunerweisen is for later, when you have accepted that the man never had any intention of making sense.