The Situationist receives many good comments, but some are too good to not be posts. In response to our recent post, “Opening Black Boxes – Abstract,” Tamara Piety wrote an excellent comment, which we’ve posted below. Thanks Tamara.
* * *
I look forward to reading this article in detail, but I think the prospect of accurate fMRI lie detection is (as I understand it) still relatively remote, while the dangers of courts and the public leaping to adopt the latest technology that purports to offer reliable lie detection is fairly proximate and of grave concern. As the author notes, all one has to do is to look at the history of other lie detection technologies to feel some caution about embracing a new technology too enthusiastically, particularly when offered by the prosecution. The National Research Council’s recent report on the state of the forensic sciences contains some cautionary tales on this issue with respect to all manner of forensic testimony purporting to be “expert” (with some corresponding danger of deference from the jury) that has regularly (and continues to be) admitted by the courts but which lack empirical foundation. It is extremely important to not confuse or conflate what courts decide to admit with real evidence of the technology’s validity or reliability. A court’s decision to admit something is legal proof but not empirical proof, any more than a court’s decision to admit the testimony of an “expert” astrologer would be proof of the external validity of astrology as a practice. Compared to the rather remote and largely hypothetical question of whether, assuming fMRI lie detection achieves some reliability, it would do away with juries, the danger of admitting it without sufficient grounding seems to me of much greater concern. Many of those subsequently exonerated by DNA were initially convicted with dubious forensic science testimony. Conversely, there has been for some time fairly good evidence that juries cannot do many of the things we ask them to do. (Ex: disregard something they have already heard; use evidence for one purpose but not another, as when jurors are asked to use evidence of a witness’ prior conviction as evidence for assessing his credibility or character for truthfulness, but not for propensity; not discuss issue of insurance; and a host of other incredible constructs). And courts have been fairly resistant to permitting expert testimony about things that challenge existing assumptions, say the frailties of eyewitness testimony or memory or fingerprint comparisons, etc. Were accuracy of factual determinations the main concern in trials one might have predicted that all this evidence concerning unreliable forensics would be admitted and that the Rules of Evidence would have long since been reformed to bear some resemblance to how people really think. Indeed we might have predicted a system of evidence-based evidence as Simon Cole has suggested, and relied less on juries at that point, were it not clear that something else is going on. I think something else is going on and that the experience with forensic science generally to some extent answers the question posed here.