The Question of Reliability

Some discussions of the “torture question” suggest – or
simply declare – that torture is never
justifiable in part because it is an unreliable source of information. In my view, such claims need to be
“unpacked” a bit, and are not quite obviously true enough to allow us to skirt
the ethical challenges presented by this quest

It
certainly seems likely that “torture” can
be an unreliable source of information. The question is whether it is necessarily so. To
be sure, asking a prisoner leading questions – making clear that a particular
response is desired and that if this response is given, the abuse will stop –
is presumably a highly unreliable way to get to the actual truth.   Enough people would agree to anything under such circumstances that an interrogator
probably could not rely much upon the results. From this perspective, torture would seem a particularly
unhelpful way to approach criminal justice matters: one might obtain lots of
confessions, but they would be essentially worthless as indicators of the
suspects’ actual responsibility. The problem in such “confession” scenarios is that torture may well decrease , rather than increase, the likelihood of
truthfulness.

But
what if the interrogator asked only carefully-phrased open-ended questions about matters potentially subject to
empirical verification? A
terrorist suspect simply asked about how decision-making works within his
organization, for instance, would not so easily know what information would satisfy his torturers. Moreover, he would seem to have
little incentive to say just anything , because details that did not pan out when measured against other
prisoners’ stories or other evidence could result in the prolongation or
intensification, rather than the cessation, of his mistreatment. It might still be that the prisoner
would lie, of course, but the use of harsh techniques might plausibly increase
the likelihood of his truthfulness when compared to “normal” interrogations. In this sense, torture would seem to be
more potentially useful in intelligence-gathering scenarios than in criminal
justice cases.

The
balance of interrogator incentives would also seem likely to be different in
the two types of case. In the law
enforcement arena, where success is generally measured by the obtaining of
convictions after crimes have occurred, there might be some temptation to
torture techniques – if they were available – notwithstanding their
unreliability. In the intelligence
arena, however, the confession of a terrorist has less intrinsic value (though
surely not none ).   Instead, the focus of
information-gathering is more upon forward-looking operational detail, upon
discovering such things as the salient details of planned operations, or the
identity and location of cell members. Where an unscrupulous prosecutor might find some value in a false
confession, an intelligence interrogator has much less use for a fabricated
response. (Indeed, a false answer
in the intelligence context could lead to catastrophe, sending agents off on a
wild goose chase and leaving the real plan for the next terrorist attack to proceed uninterrupted.) The odds that an intelligence officer
would seek to employ carefully-structured questioning designed to minimize the
chance of false, “say anything” answers would thus seem to be higher than in
the case of a prosecutor.

All
this is not to say that torture is a reliable source of information. I am happy not to be any sort of
authority on the subject, and hope there are very few people alive who could
claim to be. My point is
merely that torture is not obviously unreliable enough that we can avoid
struggling with the moral dilemmas presented by “ticking bomb”-type
questions. If information obtained
by the use of torture were always
untrustworthy, things would certainly be easier. It is precisely because it is at least possible that torture might in some cases yield sound data,
however, that we have no choice but to wrestle with the ethical dilemma.