ARE THE WORST paranoid fantasies of the Woodstock generation true? Do the giant corporations crumple up lives and toss them away like Kleenex, all the while protected by armies of faceless spies and spooks?

Or are the worst paranoid fantasies of the establishment true? Do activists long to destroy the machinery of industrial progress, taking drug-crazed pleasure in chaos?

Perhaps more than any other event of the 1970s, the Karen Silkwood case has inflamed those who harbor paranoias on either extreme. In my experience, mere mention of Silkwood's name can spark vicious, irrational arguments among otherwise reasonable people. The angry arguments (and the paranoia) continue because, after six years of stories, hearings and investigations of Silkwood's death, almost nothing is known with certainty. If anything, the mystery is even more perplexing than before.

An impressive and vital new book. The Killing of Karen Silkwood , goes a long way toward changing that. Shunning the fist-shaking hysteria brought to the case by writers on both sides, author Richard Rashke has produced a chronicle that meets a demanding test of objectivity. You know in your heart Rashke is rooting for Silkwood, but he seldom lets this prejudice intrude -- faithfully presenting, for instance, all the unpleasant information about Silkwood's personal life. He even takes the unusual but welcome step of listing his sources (most of which, in the custom of modern "investigative reporting," turn out to be court documents and congressional reports, not late-night sleuthing).

In brief outline, here are the undisputed facts. Silkwood was a 28-year-old lab technician in Kerr-McGee's plutonium processing plant in Crescent, Oklahoma. She became worried about plant safety, and started gathering evidence to hand over to her union. Meanwhile, things were not going well for the union; a decertification vote to expell it from the plant, was gearing up. Neither were things going well for Silkwood; her marriage had broken up, she had attempted suicide, and she was using several drugs, including Quaalude.

In short order the affair became very strange. After a series of minor but troubling radiation exposures Silkwood suffered at the plant, she found herself contaminated with a perhaps-lethal dose. The contamination was traced to a large (by nuclear standards) amount of plutonium in her apartment refrigerator. Who put it there? No one knew. The idea that somebody, anybody, could sneak so much plutonium out of what was supposed to be a high-security installation suggested, however, that something was deeply wrong at the plant.

Under pressure from her union (which saw Silkwood as a catalyst to winning a new contract) she arranged to meet a New York Times reporter and present documents she claimed would prove Kerr-McGee's safety and security failings. On the way to that meeting in November 1974, her car went off the road, slammed into a concrete wingwall, and she was dead. No documents were found in the wreck. The FBI and other federal agencies investigated extensively, but have refused to reveal anything meaningful about their findings, even under congressional subpoena.

From here Rashke goes on to present a balanced, detailed account of what may have happened that night, and the many things that have happened since. His approach was the right choice - both because drawing your own conclusions is usually more persuasive, and because Rashke's skills as a reporter seem to outweigh his insight as an analyst. In the end, I think, Rashke marshals enough evidence to answer once and for all two of the major questions of the Silkwood case, and point the way toward the answer to the third. I will try to summarize his evidence on the three questions as best as is possible in the given space:

Was Silkwood killed? Officially her death is just a strangely-timed random accident -- she fell asleep at the wheel either from exhaustion or drugs. Rashke presents convincing evidence that she was killed. The key element in his argument is this: her car ran along an inclined road shoulder for 240 feet before striking the wingwall. Only a conscious driver could have kept the car on course.

Rashke speculates a chase car frightened Silkwood off the raod onto the embankment, then began racing along parallel to her to keep her from getting back on the raod. Whoever was in the chase car, Rashke suggests, might not have been planning to kill Silkwood -- only to flag her down and win her silence with money or threats. Silkwood would not stop, however, and while she was looking back over her shoulder at the chase car, she hit the wingwall, which was not visible until it was too late.

Was her death part of a plot? Again Rashke's answer is yes, and he demonstrates it convincingly. The most telling piece of evidence is that nearly every law enforcement agency you can name turns out to be wrapped up (some intimately) in an investigation of the affair. It stretches from the Oklahoma City police and Oklahoma Highway Patrol to the FBI, the CIA, and the National Security Agency to a handful of exotic CIA front groups to private agencies like Pinkerton, Wackenhut and Intertel and even to SAVAK. They are bugging, tailing, and monitoring everyone in sight. Some of the monitoring began before her death, and much of it went on years later, and involved only minor characters.

The FBI is especially active, throwing up smokescreens with everything from Keystone Kop-variety slander of congressional investigators to sophisticated diversionary tactics. The Bureau tricks Congress and the Silkwood estate's lawyers into concentrating their energies on investigating a decoy who ultimately turns out to know nothing of value.

If Silkwood's death was just an unfortunate mistake -- say, the work of an overzealous corporate goon squad doing things Kerr-McGee's management did not condone -- why would the CIA and NSA care? Why would an endless string of spooks be tied to a mere traffic accident?

Who was behind the plot? Rashke takes us two-thirds of the way home by answering the first two key questions, but then never even addresses the third and most important. His objectivity gets in his way; because he cannot prove who did it, he does not even speculate.

Could Kerr-mcGee be to blame? It's possible, but not logical. The Crescent plant already had such a low reputation for safety and security (Dr. Karl Morgan, director of health and safety at the Oak Ridge National Laboratory, said he'd never seen a nuclear plant "so poorly operated") that it's hard to imagine how Silkwood could have damaged it any further -- except by getting publicity. Her weird contaminations and suspicious death generated far more publicity than the AEC investigation the union wanted ever could have, so killing her would have made no sense for Kerr-McGee. That doesn't prove the company didn't do it -- the nuclear industry has a longstanding tradition of irrational behavior. But still, a Kerr-McGee action wouldn't account for the frenzied interest of the spy agencies.

The book may contain the third answer, although the author himself doesn't seem aware of it. Rashke makes passing reference to the fact that 40 pounds of plutonium was missing from the Crescent plant before Silkwood's death. Federal agencies were investigating.

This reference flashes like a fireball from the book's drone of information. Forty pounds of plutonium is enough to make at least two atomic bombs.

Since Rashke refuses to speculate on the significance of the missing plutonium, let me play out a scenario that may, finally, explain the mystery.

Security was distressingly lax at the Crescent plant. A former Kerr-McGee official, James V. Smith, told investigators that in early years the plant had "zero" security -- not even a fence. A Kerr-McGee attorney later said in court (amazingly, in defense of his company's position) that Silkwood might have been to blame for contaminating herself, because it was easy to take plutonium out of the plant. This would have made the Crescent plant an ideal place from which to "divert," as is said in intelligence parlance, the deadly metal.

Rashke doesn't mention it, but it is now widely believed that in the mid-1960s, up to 206 pounds of weapons-grade uranium was "diverted" from a plant in Apollo, Pennsylvania, owned by a since-defunct company called 'NUMEC. The uranium found its way to Israel, and was used to make that country's first atomic bombs. Inside the intelligence community it is generally believed the CIA either helped Israeli agents get the material, or at least did not interfere. There is further speculation that the agency has aided or condoned the "diversion" of nuclear materials to other "friendly" nations.

If plutonium was being stolen from the Crescent plant, that would have been a matter of extreme concern to federal agencies -- especially the FBI What if, in the course of gathering plant documents for the union, Silkwood had inadvertently found evidence of plutonium theft? Or at least someone thought Silkwood stumbled onto evidence? Someone would have had to get that evidence back, and since Silkwood had hidden it, her drive to meet the Times reporter was the only time that someone could have been certain of finding it.

This offers the only logical explanation I can imagine of why the Silkwood case caught the attention of intelligence agencies all over the world. It accounts for the multiple actors and uncoordinated strategies that run through the bizarre case. It also explains the extraordinary secrecy. After the NUMEC escapade, the FBI and other agencies clamped down years of ultra-secrecy, refusing, according to one report, to tell even the Secretary of Defense what had happened. (The "diversion" was ultimately revealed only because of a CIA official's slip of the tongue at a dinner party).

Rashke unveils one important new piece of evidence in his book, and it supports the plutonium smuggling hypothesis. He says -- unfortunately only third-hand through a source who has a source -- that the FBI was following Silkwood and the chase car on the night of her death. The FBI may know who did it.

Rashke seems to conclude that the FBI was there in order to tail Silkwood. But what if the bureau was tailing Silkwood's tail. That could make sense if Silkwood accidentally ran afoul of plutonium smugglers the FBI was trying to track.

Publication of The Killing of Karen Silkwood , and the author's allegation that the FBI knows significantly more than it is telling, make good reason to reopen the case. There has been a lawsuit in which Silkwood's estate won a $10.5 million judgment from Kerr-McGee, but that suit concerned only the company's negligence in letting plutonium out of its plant. The question of who killed Silkwood and why has yet to be raised in a court. Maybe Congress and the press couldn't pry anything meaningful out of the FBI, but the president can -- and should. There can be no doubt something much larger and more significant than the troublemaking of a confused junior-level Oklahoma technician was involved in the Silkwood case. It's time we knew for sure what that something was.