A Tortured Logic
By Keir Liddle
Dubya has reignited the torture debate of late, and questions about the efficacy of torture have raised their ugly head again. Once again, I am posting my summary of some of the evidence that suggests that it doesn’t work as a means of extracting evidence. It probably works as a means of extracting confessions, but unfortunately, it works just as well extracting them from the innocent as from the guilty…
A report published by the Intelligence Science Board in 2007 found that no research existed to support the use of enhanced interrogation. O’Mara’s review in Trends in Cognitive Science describes a wealth of science that supports ending the practice.
O’Mara derides that the belief that extreme stress produces reliable memory as “folk neurobiology” that “is utterly unsupported by scientific evidence”. The hippocampus and prefrontal cortex — the brain’s centers of memory processing, storage and retrieval — are profoundly altered by stress hormones. Keep the stress up long enough, and it will “result in compromised cognitive function and even tissue loss,” warping the minds that interrogators want to read.
What’s more, tortured suspects might not even realize when they’re lying. Frontal lobe damage can produce false memories: as torture is maintained for weeks or months or years, suspects may incorporate their captors’ allegations into their own version of reality.
I’d further argue that the principle of convergent evidence strongly suggests that torture does not in fact work. Indeed, just by using the information contained in this report, the idea that torture is effective in extracting accurate and reliable evidence seems to be built on shaky ground.
The strongest argument in favour of torture is the so called ‘ticking bomb’ scenaro. Alan Dershowitz gave a good summary of it in the San Francisco Chronicle back in 2001:
“Everybody says they’re opposed to torture. But everyone would do it personally if they knew it could save the life of a kidnapped child who had only two hours of oxygen left before death. And it would be the right thing to do.”
It’s a compelling argument, until you start to look at the assumptions that you have to make to accept it. This argument assumes that you have the right person in custody, it assumes that this person actually has the information you need, it assumes that there isn’t a better way of getting hold of the evidence, and above all, it assumes that torture is an effective way of getting that information.
Yet this misses one crucial problem with the ticking timebomb scenario… In that there is no method of extracting information that falls under the umbrella of torture that doesn’t require quite some time to carry out. Indeed, torture methods rely on keeping people in stress positions or isolation for hours on end… so unless the bomb has quite a long fuse, then torture, as it is currently practiced under the auspices of KUBARK and it’s successors, is useless.
There is a further reason to assume that, even despite evidence that seems to suggest the contrary, if torture does work, it is not of benefit to a democratic country to make us of its techniques, as shown by the French battle for Algiers in the 1950’s, and a lesson we should have learnt from the Northen Ireland conflict in the 1970’s. Torture simply serves to further alienate the local population and turn them against you with more venom and more certainty in your status as enemy. In short torture plays directly into the hands of terrorist recruiters.
“It’s extremely ineffective, and it’s counterproductive to what we’re trying to accomplish,” he told reporters. “When we torture somebody, it hardens their resolve,” Alexander explained. “The information that you get is unreliable … And even if you do get reliable information, you’re able to stop a terrorist attack, Al-Qaeda’s then going to use the fact that we torture people to recruit new members.”
Alexander says torture techniques used in Iraq consistently failed to produce actionable intelligence, and that methods outlined in the US Army Field Manual, which rest on confidence building, consistently worked and gave the interrogators access to critical information.
The report on the news seemed to suggest that the torture carried out was based on the CIA’s classic text of torture the KUBARK manual, stress positions, disorientation and the like all present and correct. In the interests of attempting to debunk the idea (sometimes presented as the common sense notion) that torture works, it seems appropriate to look at the “research” which led to the techniques of modern torture contained within KUBARK being created and used.
To do that, we have to look at the research of a number of psychologists and other scientists conducted during the Cold War from 1950 to 1962 under the heading MKUltra. Headed by Sidney Gottileb the project funded experiments into LSD and other drugs, hypnosis and “psychic driving“. With the goal of finding a way to control or influence the mind in Manchurian candidate stylee…
Psychic driving was the precursor to Donald Ewan Cameron’s depatterning techniques on which much of the basic principles of KUBARK are founded. It was of particular interest to the CIA as it’s central concept was that it is possible to erase a person’s mind and start again from scratch: essentially by means of excessive use of electroconvulsive therapy and the drugging of “patients” to make them more susceptible to suggestion, then planting a suggestion in their minds that counteracted what was though to be the root of their neurosis. This was an arbitrary and highly damaging synergy of behaviorist ideas and Freudian flights of fancy which has now, unsurprisingly, been discredited. Much of these experiments were conducted without proper informed consent and damaged many, many lives without providing any evidence whatsoever that they worked.
The patients in Cameron’s experiments, much like torture victims, did not become more malleable or prone to revealing information. They became confused, experienced memory problems and in many cases, seemed to “regress” to a childlike state. None of this sounds like it would be particularly conducive to gathering reliable information. Even if they were not confused or didn’t suffer memory problems, the very basis of techniques based on Cameron’s depatterning involves increasing suggestibility. Again, something I contend that’s not conducive to gathering effective and reliable intelligence.
It is my contention that torture hasn’t moved on all that much since KUBARK first came into existence. For one thing, it would be very difficult to fund research on any great scale into methods of torture and its efficacy in the modern age. I find it hard to imagine that someone wouldn’t find out in a manner similar to that Abu Ghraib exposed. Scientists contend that the use of torture as understood by “folk psychology” (if I hit you and keep hitting you, you are more likely to tell me what I want to know to make me stop) is unlikely to work.
Neurobiologist Shane O’Mara of the Trinity College Institute of Neuroscience explains in his paper that:
“the use of such techniques appears motivated by a folk psychology that is demonstrably incorrect. Solid scientific evidence on how repeated and extreme stress and pain affect memory and executive functions (such as planning or forming intentions) suggests these techniques are unlikely to do anything other than the opposite of that intended by coercive or ‘enhanced’ interrogation.”
Extreme pain and stress has been shown to have the opposite effect to that which interrogators wish. Yes, they may “soften up” the victim, but also leave them in such a state that they cannot effectively respond to questioning accurately. As such, I would suggest that the interrogator will probably extract the information they wish, but would really have no means of knowing if this information is actually true, and could therefore not provide actionable intelligence.
Further to this, looking at how information is recalled from the brain shows another reason why torture probably doesn’t work.
To recall information stored in the brain, you must activate a number of areas, especially the prefrontal cortex (site of intentionality) and hippocampus (the door to long-term memory storage). Fact Two: Stress such as that caused by torture releases the hormone cortisol, which can impair cognitive function, including that of the prefrontal cortex and hippocampus. Studies in which soldiers were subjected to stress in the form of food and sleep deprivation have found that it impaired their ability to recall personal memories and information, as this 2006 study reported. “Studies of extreme stress with Special Forces Soldiers have found that recall of previously-learned information was impaired after stress occurred,” notes O’Mara. “Water-boarding in particular is an extreme stressor and has the potential to elicit widespread stress-induced changes in the brain.”
In fact, it would probably be prudent to link to the source of this information and allow you to read it for yourself: an excellent Newsweek blog from Sharon Begley. I also recommend this Wired article which covers similar ground.
Pain and Physical Discomfort
I contend the info from the research presented above suggests that physical torture may not have an effect. With regard to this, Biderman notes:
Reports about the treatment of POWs and foreign prisoners in China documented the use of physical abuse, but studies of the role of assault in promoting attitude change and in eliciting false confessions (even from U.S. servicemen) revealed that it was ineffective. Belief change and compliance was more likely when physical abuse was minimal or absent (Biderman 1960).
Sleep Loss/Deprivation is associated with:
• General cognitive slowing (Dinges and Kribbs, 1991)
• Impaired attention (Hockey, 1970; Norton, 1970)
• Diminished concentration (Williams, Lubin, and Goodnow, 1959)
• Impairment in cognitive functions associated with right anterior
hemisphere or subcortical areas such as motor, rhythm, receptive and
expressive speech, memory, and complex verbal arithmetic functions
(Kim et al., 2001)
• Impaired decisionmaking involving the unexpected, innovation,
revising plans, competing distractions, and effective communication
(Harrison and Horne, 2000)
• Reduced capacity for logical and sequential thought (Blagrove,
Alexander, and Horne, 1995; Horne, 1988b; Williams and Lubin, 1967)
• Decreased accuracy in time estimation, and both immediate and
delayed recall (Taylor and McFatter, 2003)
• Negative effects on mood (Lieberman et al., 2002)
• Alteration of the body’s immune system (Everson, 1997)
• Increased perception of physical pain (hyperalgesia) (Kundermann et
• Decreased motivation (Wilkinson, 1961, 1964; Horne and Pettitt, 1985;
• Increased suggestibility (Blagrove, Cole-Morgan, and Lambe, 1994;
Many of which imply that sleep deprivation is not much use as a means of obtaining information (being prone to suggestibility in particular, as well as reduced capacity for logical and sequential thought).
On this last point it is worth noting that suggestibility increases specifically under conditions simulating an interrogation. At least one study has found that “the effect on suggestibility of one or two night’s sleep loss is comparable to the difference in suggestibility between true and false confessors.” (Blagrove, 1996)
Sensory deprivation is associated with:
• Impairment in higher mental functions and complex intellectual tasks
(Myers, Murphy, Smith, and Goffard, 1966; Kitamura, 1967)
• Increased susceptibility to infl uence (under some conditions) (Myers et
• Heightened hypnotic susceptibility (Sanders and Reyher, 1969)
• Diminished EEG activity correlated with apathetic, lethargic behavior,
and a reduction in stimulation seeking behavior (Scott and Gendreau,
• Behaving in a way that is more boring and unlikable (Zuckerman et al.,
Increased anxiety and depression (Zuckerman et al., 1970)
• Greater instability of beliefs and of both peripheral and central attitudes
(Tetlock and Seudfeld, 1976; Seudfeld and Borrie, 1978)
• Cognitive disorganization (Seudfeld and Borrie, 1978)
• Increased persuadability (Seudfeld and Borrie, 1978)
• Increased compliance behavior (beyond usual social infl uence
conditions) (Moscovici and Doms, 1982)
Two papers from the journal of Biological Psychiatry also seem to back up the views that battlefield stress (which I contend makes a good proxy for torture in the sense that it is stress caused by the fear that death could arise at any minute, with the disadvantage as a proxy that there is a greater degree of control than found in a torture scenario):
1) Stress-Induced Deficits in Working Memory and Visuo-Constructive Abilities in Special Operations Soldiers
2) Severe decrements in cognition function and mood induced by sleep loss, heat, dehydration, and undernutrition during simulated combat
Interestingly, the second study deals with the effects of high levels of stress experienced during simulated combat. I would suggest this shows that just because it is unethical to study something directly does not imply that you cannot study a lesser or similar experience to gain insight.
Finally, the research suggests that the act of torturing someone even makes the innocent torture victim look guilty.
Participants in the study met a woman suspected of cheating to win money. The woman was then “tortured” by having her hand immersed in ice water while study participants listened to the session over an intercom. She never confessed to anything, but the more she suffered during the torture, the guiltier she was perceived to be.
The research, published in the “Journal of Experimental Social Psychology,” was conducted by Kurt Gray, graduate student in psychology, and Daniel M. Wegner, professor of psychology, both in Harvard’s Faculty of Arts and Sciences.
This research suggests that, far from uncover guilt, torture actually leads to convincing the torturer, and observers of torture, of the victim’s guilt regardless. It is as though “those that know of the victim’s pain must somehow convince themselves that it was a good idea” in order to explain why such pain might be being visited on an individual under torture. Their captors will dehumanise them and come to believe them guilty, rather than accept that they have done wrong. The findings of this study also perhaps shed some light on the reprehensible goings on at Abu Ghraib and the public reaction. It was found that whereas those closely involved in the torture were more likely to classify the victim as deserving of the pain and see them as guilty, those observing at a distance were able to distance themselves from the act and feel sympathy for the victim.
This finding is enough to my mind to suggest that there is an inherent bias in “extreme interrogations” that leads to prolonged and self-perpetuating interrogations, ultimately proving fruitless. Arguably, this process is similar to the dehumanization that we associate with the everyday business of war and war atrocities, spousal abuse and various other nasty things. In a strange way, it’s heartening that people feel the need to make those that they are about to inflict pain on seem less human and less like themselves, but on the other hand, it’s horrifying how easily people can slip into the language of dehumanization. It’s worrying how effective dehumanization is at allowing people to commit extraordinary atrocities.
In summation, the answer to the question “does torture work?” is not currently a definite “no”, but is a very strong “probably not”.