The Italian Cannibal-Film Canon (the defensible ones)

The parenthesis in the title is doing heavy lifting, and I want to be honest about what it can and cannot carry. The Italian cannibal cycle ran from 1972 to about 1985, produced perhaps a dozen features, and did something almost no other film movement has done: it killed real animals on camera, deliberately, for effect. Censors have exaggerated plenty of things about plenty of films; this is on the negative and anyone can watch it. A turtle is dismembered in Cannibal Holocaust while alive. A coati is stabbed. A monkey’s skull is opened. Six or seven creatures die in that film alone, and the practice recurs across the cycle.

There is no critical frame that fixes this. The films were made in countries where the crews could do it without consequence, by directors who assumed nobody would care, and the footage is in the negative forever. Ruggero Deodato spent the last thirty years of his life apologising for it and saying he would cut it if he could; the rights holders have released versions with the animal footage removed, and those versions are the only ones I will recommend to anyone.

So what is left, and why bother? Because two or three of these films are doing something genuinely difficult, and one of them invented a form that took over horror cinema thirty years later. The cycle’s full sordid history is in the nasty business of the Italian cannibal cycle. This is the shorter question: which ones defend themselves.

Where it came from

The ancestor is the shockumentary. Mondo Cane (1962) toured the world’s alleged barbarities for Italian audiences and made a fortune, establishing a national appetite for footage sold as anthropology and consumed as atrocity — a trade covered in Mondo Cane, the shockumentary that named a genre. The cannibal film is that trade with a script attached, and its first entry is a knockoff of a Hollywood western.

Man from Deep River (Ruggero Deodato, 1972). A British photographer is captured by a tribe in Thailand, tortured, adopted, married in. Deodato lifts the structure wholesale from A Man Called Horse (1970) and transplants it to Southeast Asia, and for two-thirds of its length it is a competent adventure film with real curiosity about its setting. Then it turns, and the turn is where the cycle’s genetics announce themselves: cannibals arrive from upriver, and the film starts trading in the tourism of horror. The animal killings begin here too. It is the least gruelling of the six and the one where you can see a genuine director still deciding what he wants to be.

The two with an argument

Jungle Holocaust (Deodato, 1977). The most defensible film in the cycle, and the least seen. Massimo Foschi plays an oil-company man stranded in Mindanao and captured by a tribe who cage him, strip him, examine him and — crucially — treat him as an animal. The whole middle hour is a survival film with almost no dialogue, in which a Western executive is systematically reduced to the state he assumed everyone else occupied. Deodato shoots the tribe with a stubborn refusal to explain them, and the effect is genuine dislocation. The film’s final movement, in which Foschi escapes and does something unforgivable on his way out, is the cycle’s cleanest piece of moral engineering. Strip the animal footage and you have a hard, spare, functioning film about the assumption of superiority.

Cannibal Holocaust (Deodato, 1980). The famous one, the ambitious one, the one you have to argue about. A rescue team retrieves the film cans of four missing American documentarians; a New York university screens the footage; the footage shows the documentarians burning a village, assaulting its inhabitants and staging atrocities for the camera before the tribe finally does to them what the film’s audience arrived to see. The structure is a trap and it closes.

Two things make it work. The first is Riz Ortolani’s score, which is the most perverse decision in the cycle: over images of mutilation he plays a lush, sincere, almost romantic theme, and the mismatch does more moral work than any line of dialogue. The second is the found-footage conceit itself, deployed here with a rigour nobody would attempt again for twenty years — the grain changes, the camera panics, the operator’s own death is in the frame. Deodato was so convincing that Italian prosecutors charged him with obscenity and pressed the question of whether his cast were actually dead; he had signed the actors to a contract keeping them out of public view for a year to preserve the illusion, and had to produce them in court to clear it. The lineage from that courtroom to The Blair Witch Project is direct, and it runs through cannibal cinema and the ethics of the fake documentary and out into found footage after Paranormal Activity.

The film’s thesis — that the documentary crew are the savages — is delivered with a sledgehammer and it is still correct. The problem is that Deodato proves his case about exploitation by exploiting, and the turtle is in the film because he wanted the audience to feel something real. He got what he wanted. That is precisely the indictment.

The three that only sell the atrocity

Mountain of the Cannibal God (Sergio Martino, 1978). The glossiest, with Ursula Andress and Stacy Keach on a New Guinea expedition, and the one that shows what happens when the cycle gets a budget and no argument. Martino is a real craftsman — his giallo work is covered in Sergio Martino, the giallo craftsman — and the film is handsomely shot and paced. It is also a straightforward star vehicle in which the anthropology is set dressing, the tribe is a menace, and the film’s interest in Andress is entirely in what will be done to her. The animal cruelty here is gratuitous even by the cycle’s standard.

Eaten Alive! (Umberto Lenzi, 1980). Lenzi grafts a Jonestown plot onto the formula — a cult leader in the jungle, Ivan Rassimov playing him as a knock-off Jim Jones — and pads the film with footage recycled from earlier entries, including his own. The Jonestown angle could have been the cycle’s smartest idea; the massacre was two years old and the parallel between an American guru and a colonial expedition writes itself. Lenzi has no interest. He shoots the sect scenes as an excuse and returns to the jungle as fast as he can. His career’s other half is worth more, and it is in Umberto Lenzi, from poliziotteschi to cannibals.

Cannibal Ferox (Lenzi, 1981). Sold as Make Them Die Slowly, advertised as banned in thirty-one countries, and the crudest thing in the cycle. Three American students in Colombia, a sadist among them, a series of set-piece mutilations. Giovanni Lombardo Radice, who plays the sadist, has spent forty years publicly disowning the film and describing the shoot with open contempt — a rare and useful thing, an actor telling you the truth about what he made. Lenzi’s defence was that the film condemns the Americans. What the film actually does is queue its atrocities and photograph them, in order, with the running time apportioned according to severity.

Why the good ones work

The craft that separates Deodato from the rest is documentary discipline. He shot on real locations with real logistics, gave his camera operators licence to fail, and cut for the rhythm of an amateur rather than a professional. Watch the village-burning sequence in Cannibal Holocaust: the framing is wrong, the exposure blows out, a hand crosses the lens. Every one of those errors is deliberate and every one of them costs a take. The generation that made The Blair Witch Project and [REC] rediscovered this and got credit for it. It was worked out in a Colombian jungle in 1979 by a man who wanted to fool a prosecutor.

Ortolani’s score is the other lesson, and it generalises. A film’s moral position lives in the gap between what the image shows and what the soundtrack believes. Score an atrocity with dread and you have told the audience how to feel. Score it with beauty and you force them to notice that they are watching, and that the film is not going to help them. Almost nothing else in exploitation cinema is that sophisticated.

The audience problem

The cycle exists because of a specific commercial accident, and understanding it explains the films better than any auteur reading. Italian producers in the seventies worked to a rule: identify a hit, copy its shape, release within eighteen months, recoup from foreign territories before anyone notices. The cannibal film was an unusually good fit for that machine because it needed no stars, no sets and no sound recording worth the name. You could shoot in Colombia or the Philippines with a crew of fifteen, post-sync everything in Rome, and sell it in forty countries on a poster and a rumour.

That economy explains the animal killings more precisely than sadism does. Effects work costs money and takes time. A real death costs nothing and takes one take. Every producer in the cycle knew that the single scene the distributor would show the buyer was the one the audience could not dismiss as latex, and the cheapest route to that scene was through something living. The cruelty was a line item.

It also explains the marketing, which was frequently more dishonest than the films. Cannibal Ferox was sold as “banned in 31 countries” — a number nobody has ever substantiated and Lenzi’s own distributor appears to have invented. The cycle’s real censorship history is duller and stranger: most territories cut the animals and passed the human violence, which tells you what actually offended, and the British prosecutions of the early eighties targeted these titles above almost anything else on the shelf.

What to actually do

Watch Jungle Holocaust and the animal-cut Cannibal Holocaust, in that order, and stop. Both are on disc from boutique labels in editions that offer the excised versions; take them. The rest of the cycle exists to be read about. The films’ British history — the video-shop panic, the DPP list, the prosecutions — is its own story, told in the video nasties panic and what the BBFC was really afraid of.

And hold the line on the parenthesis. “Defensible” here means two films with real ideas made by a director who later understood what he had done. It does not extend to the cycle, which was a commercial reflex that discovered atrocity sold tickets and kept selling them until the market moved on. Deodato made an argument about exploitation that his own production disproved. That is a genuinely interesting failure, and it is still a failure.