Is it possible for someone to know something but not know that they know it? If so, how would such a person be distinguished from someone who knows something but lies about it when asked? Interrogators and criminal investigators are faced with such conundrums all the time.
What is State-Dependent Learning?
In the 1960s researcher Donald Overton suggested that the likelihood of successful learning is greatest when there is a clear correlation between the circumstances which prevail while forming a memory, and during subsequent retrieval of that memory. Included here are both internal states, such as mood, alertness and attention, along with external conditions (temperature, nearby sounds or smells). Overton referred to this as “state-dependent learning”.
As an example, Overton tested rats learning to successfully wend their way through a maze while under the influence of pentobarbital. When later tested without the drug they failed to run the maze. Restoration of the pentobarbital led to a renewal of the rat’s successful performance.
Many of the findings on state dependent learning date from the 1960s, principally the work of Overton. His seminal 1964 paper “State-dependent or dissociated learning produced with pentobarbital” essentially defined the field. A person too may retain a memory, but the memory may be inaccessible, because conditions present when creating the memory were different during attempts at memory retrieval.
How State-Dependent Learning Affects Human Memory Retrieval
Thus memories are not stored like a video disc—an unalterable conveyor of established information—but as a relationship between the conditions that prevailed when the memory was encoded, compared to the conditions prevailing at the time of being asked about the memory. If the conditions differ sufficiently, the memory, if it exists, may be irretrievable.
As an example, people suffering from panic attacks often prove resistant to persuasion and even formal psychotherapy. That’s because the calmness of the discussion with the friend or therapist was totally different from the mentally tumultuous experience associated with the subsequent panic attack. In other words, both the external and internal cues present during the panic attack were totally opposite to the cues experienced during the previous help session.
A similar thing happens when a person is caught up in anger, stress or anxiety. When under the influence of these emotions, it’s hard to recall and put to use advice previously received from friends and therapists about remaining calm.
So the answer to our first question at the beginning of this essay is Yes. A failure of coming up with a memory (retrieval failure) doesn’t necessarily mean the interrogatee hasn’t retained some memory of it. But the only way of proving this is by testing for recognition memory, rather than retrieval memory. In other words, testing for whether they recognize the memory when they can’t retrieve it upon request. In students this takes the form of multiple-choice testing, whereby the student who cannot respond when asked for the correct answer nevertheless identifies it from a menu of possible choices only one of which is correct.
Wouldn’t it seem reasonable to assume that the student couldn’t recognize the correct answer if they hadn’t previously learned it? Inevitably in discussions about such situations, the issue of guessing arises. The student may even say “I was only guessing.” Although that claim can’t be disproven, an equally likely explanation holds that they did know the information (the test they had just taken demonstrated it!) but didn’t know that they knew it. An incorrect answer, in contrast, is a fairly reliable indicator that the information never was incorporated or encoded within the brain, and therefore cannot be retrieved because, simply, there is no memory to be retrieved.
State-Dependent Learning and Eyewitness Testimony
In the courtroom state-dependent learning provides an explanation for the unreliability of eye-witness testimony. Difficult emotional states may have characterized the events the eyewitness saw and is now recalling. But the internal and external cues may differ dramatically: witnessing, say, a violent assault versus later describing that crime while sitting in the comparatively calm atmosphere of the courtroom.
Typically, even the most cooperative of witnesses possesses more information than comes up during conventional questioning. The problem is not storage, but retrieval. The skillful interrogator is tasked with facilitating that retrieval without denigrating the witness.
The most important quality of a good interrogator is the ability to resist prematurely concluding the subject is lying, strictly on the basis of a “gut feeling” created by interview cues (the suspect appeared nervous, couldn’t maintain eye contact etc.). In the recent past interrogators have switched their emphasis from behavioral observations to the search for some physiological instrument for determining falsehood, starting with the lie detector test and extending to alterations in measurable brainwave forms elicited by physiological stimulation. None of these quasi objective tests have held up and, as a result, are not routinely used in courtroom determinations.
The Daubert Standard and Scientific Evidence for Detecting Lies
Phrased slightly differently, no method, technique, or instrument has passed the Daubert standard—established by the Supreme Court in 1993 in Daubert v. Merrell Dow Pharmaceuticals, Inc. to address whether scientific evidence when presented by an expert should be accepted by a court.
Currently no claimed “objective” tests for lying have been accepted under Daubert as a reliable indicator that someone is lying. So the answer to the second question raised at the beginning of this essay is: Currently there is no objective test that infallibly identifies someone who knows something but lies about it. The standard for detecting concealed knowledge is slightly less stringent but, in general, the answer is still No when it comes to introducing such findings into a courtroom.
Ironically, the accuracy of the oldest method “the Lie Detector Test (LDT)” is only slightly better than chance, according to the social psychologist Aldert A. Vrij (“Outsmarting the Liars: Towards a Cognitive Lie Detection Approach”). Nonetheless, it has been applied within the past two years in the White House, the FBI and in several large organizations.
Nor is it probable that the idea that lying can somehow be reliably identified by some scientific test likely to be accepted in the courts any time soon. Why? Since lengthy prison terms or even capital punishment remain as a potential consequence of “proving” somebody a liar in court, the tolerated margin of error must be near zero. No currently available “brain test” comes even close to meeting that criteria.
Richard M. Restak, M.D.