Looking for an honest machine

A papyrus dating back to 900 BC provides the ground rules for an early lie detector test: ''He gives evasive answers, he speaks nonsense, he rubs the great toe along the ground and shivers, he scratches the roots of his hair.''

In these words a reader can still sense the power of the community as it brings to bear upon an individual its scrutiny - its will to know what the individual, rightly or wrongly, does not wish it to know.

One would not go so far as to call this the Age of the Lie Detector - not while the computer is around to claim honors. But the government certainly seems to be feeling the ancient need to peel away our secrets in order to protect its own.

Once the only argument for the use of the lie detector was ''national security,'' the civil servant's favorite excuse for all extreme measures. With the assistance of a polygraph, it was a little simplistically assumed, the CIA, the FBI, and the National Security Agency would detect spies, rather as a Geiger counter detects radioactive material.

But thanks to the overclassification of ''secret'' and ''confidential'' documents - an old form of bureaucratic self-flattery that is getting worse - the most trivial data can be connected to ''national security,'' and thus to the rationale for lie detector tests.

As the habit of the lie detector test becomes established - and such things do tend to become habits - difficult questions arise.

Who will determine when and to whom a polygraph test should be given? At the time of the Watergate investigation, would a polygraph test have been deemed appropriate for President Nixon, who favored polygraph tests for other government workers?

How can the lie detector avoid abuse as an electronic inquisitor? Much is made of the positive aspects of a polygraph test, as an opportunity for the trustworthy to prove themselves. This is the thrust of F. Lee Bailey's television show, ''Lie Detector.'' But still, nobody calls the polygraph a ''truth detector.'' And in the wrong hands, it seems perfectly designed to intimidate.

How reliable is the lie detector? It is a phenomenon of any test that the people who administer it tend to believe their results. Salem witch-hunters believed that witches sank in the water. The chief of the polygraph division of the National Security Agency claims virtual infallibility - a 97.7 percent record of accuracy. Less involved students of the machine give it credit for as little as 70 percent accuracy. Some academic skeptics find the lie detector scarcely an improvement on random chance.

What does this mean to the 30 percent - or even 3 percent - who are electronically misjudged? One's reputation, one's livelihood hang in the balance - a situation likely to produce the very symptoms the polygraph is looking for.

Given the grave consequences of failing - or even refusing - a lie detector test, how can a test be conducted that is not, per se, prejudicial?

These questions make it obvious that the lie detector is a delicate issue for civil liberties. Nor is the government the only player. The fragile balance between the right of the individual to privacy and the right of the community to information is further complicated by the readiness of more and more businesses to see lie detector tests as the quick solution to measuring the character of their employees and prospective employees.

One need not cry ''Police state!'' at the first abuse of a polygraph, or even a wire-tap. But before lie detector tests become more general and more casual, perhaps we should ask ourselves two further questions.

What does the procedure do to a climate of mutual trust, without which no real community can exist?

And, above all, what does it do to our vision of ourselves when human beings are strapped to a machine that draws a moral judment on them, based upon physiological symptoms of sweating, heartbeat, and blood pressure?

We are back almost 3,000 years with the humiliation scene described on papyrus.