Early this century, a box with a few wires sticking out and flickering
needles that jumped across a chart recorder would probably have impressed the
average person. Electric gadgets were newish inventions back then. An uncritical
enthusiasm for science was in the air. These are the most charitable
explanations for the origin of the lie detector.
No doubt William Moulton Marston would have disagreed. For it was Marston who
in 1915 claimed to have discovered that blood pressure often goes up when people
lie. That finding paved the way for the invention of the polygraph in the 1920s.
As for Marston, he went on to invent the comic strip character Wonderwoman,
giving her a lasso with special powers that forced villains to tell the
truth.
And aren鈥檛 comic strips where the entire pseudoscience of lie detection
properly belongs, too? Probably鈥攕o why is the US Department of Energy to
begin carrying out routine lie detection tests on scientists who have access to
nuclear weapons information?
Advertisement
The news becomes less perplexing when you realise that every year more than
10 000 federal employees and applicants for government jobs in the US are
already subjected to polygraph tests, and that calling for more of the same is
an ancient custom among American security chiefs following spy scandals and
leaks.
Native habitat
Less perplexing, but no less dismaying. In Britain and many other European
nations, the polygraph is (as far as anyone can tell) seldom used by governments
or courts. For years, it looked like the polygraph was headed for extinction
even in its native habitat of the US. Federal courts consistently ruled, on
scientific grounds, that polygraphs were inadmissible as evidence. And in 1988,
Congress banned most private companies from using polygraphs to screen
prospective employees.
But then came Ronald Reagan (鈥淚鈥檝e had it up to my keister鈥� with leaks),
Aldrich Ames, the CIA agent turned KGB spy, and now the loss of nuclear weapons
secrets to China from the US nuclear laboratories鈥攁ll of which led to more
polygraph tests for federal employees. The FBI, the CIA and the National
Security Agency all now require polygraphs for job applicants and test employees
at regular intervals. All admit that the test is intrusive. It鈥檚 simply a matter
of balancing individual rights with national security, they say.
Now, in the US you can get away with a lot in the name of national security,
but isn鈥檛 the Department of Energy supposed to be an advocate of sound science?
And hasn鈥檛 the polygraph always been an affront to science as well as
justice?
This is not to say that lying doesn鈥檛 cause hearts to thump, palms to sweat
and blood pressure to jump, as polygraphs demand. It鈥檚 just that honest people
often have the same reactions when merely accused of wrongdoing. Nobody has so
far even come close to discovering a physiological or behavioural 鈥渢ick鈥�
unequivocally linked to mendacity. The latest review of the scientific
literature by two Chicago-based researchers claims that there are 23 鈥減ractical,
objective鈥� signs of lying. Which sounds impressive until you see what鈥檚 on the
list: throat clearing, nose touching, speech errors, shoulder shrugs.
In most polygraph tests, the subject is asked relevant questions (鈥渉ave you
stolen classified documents?鈥�) and 鈥渃ontrol鈥� questions that are intended to
arouse greater anxiety in an innocent subject (鈥渉ave you ever stolen anything in
your life?鈥�). But interpretation of the results is highly subjective. It鈥檚 also
not hard to beat the test by biting your tongue during the control question. If
anything, the test is biased against the innocent鈥攁n icy, pathological
liar will usually pass with flying colours.
False positives
Worse, any test used in a 鈥渄ragnet鈥� mode is almost sure to produce many false
positives. To see why, assume polygraphs really can tell if someone is lying.
Assume, too, that they have a fantastic accuracy rate of 99 per cent. And assume
that as many as one in a thousand weapons-lab employees is a spy. In such a
case, the test would finger ten innocent employees for every actual spy caught.
But even the most ardent proponents of the polygraph don鈥檛 claim an accuracy
rate above 95 per cent.
In one sense, though, the polygraph does work. As an unnamed intelligence
officer recently explained to The Washington Post: 鈥淎 good polygraph
examiner can convince you he can see into your soul. You cannot believe the
stuff you can get people to confess to on these machines.鈥�
Well, it鈥檚 hard to feel sorry for people who are (a) guilty and (b) gullible.
If the polygraph is only there as a stage prop that鈥檚 one thing; the trouble is
that polygraph examiners, like all witch doctors, seem to believe it themselves.
They don鈥檛 just screw out confessions, they really do pore over the charts that
pop out of the machines before pronouncing judgment.
And meanwhile it鈥檚 mainly totally innocent people who are being hurt.