Despite the personal trust we have invested in our own senses– even when presented with illusions or effects that prove to us our own fallibility– we continue to perceive them as objective lenses into a unidimensional world. But trusting our senses is only our first misstep; trusting our memory, which remembers– and misremembers– the erroneous details from our sensory experiences is where our shortcomings can become consequential.
Wrongly convicted individuals imprisoned by faulty eyewitness testimony are of a non-negligible number– a 2013 Stanford University research group estimated that approximately 4% of incarcerated individuals are innocent. 4% may seem like an acceptable margin of error, given that its converse of 96% is a high enough standard of subjective “success” for most Americans. However, those who believe 96% is “good enough” risk equating people to numbers in a context that, unlike other statistics-reliant situations like college admissions and demographic information, desperately needs humanization; 4% of incarcerated Americans constitutes almost a hundred thousand individuals who are spending decades to life sentences in wrongful imprisonment.
Eyewitness testimony has proved fallible in a number of high-profile cases, including that of Jennifer Thompson, a 22-year old rape victim whose line-up selection (even after her conscious attempts to memorize the physical attributes of her rapist) convicted an innocent man, Ronald Cotton, to 11 years of prison. After reporting the rape, Thompson was presented with a photo line-up, and instructed to select the rapist if he was present. Following Thompson’s selection (which, following DNA evidence and fingerprinting administered more than a decade in the future, proved inaccurate), a detective assured the shaken woman that she “did great”.
Forensic psychologists are now aware of how Thompson’s case can and should have been avoided, even if it has taken decades and dozens more similar cases– and untold numbers of un-exonerated innocent convicts– to become common practice in the judicial community. For one, line-ups are now discouraged. Instead, suspects are shown one by one to allow the victim to ruminate on each photograph without the potentially memory-altering “noise” generated by neighboring photographs. Line-ups also pressure victims to choose a suspect because they are subconsciously led to believe that the suspect must be one of the options presented. By going one by one, victims are unaware how many photographs are in the line-up, so they can save their definitive “yes” for a confident match.
Similarly, authorities that interact with the victim should take precautions to ensure that their actions, including their body language, and their speech should not bias the victim in any way towards making a selection. That means neither providing encouragement, nor suggesting that the victim “keep looking”.
But why was Thompson, and the other thousands of individuals who have provided false eyewitness testimony, susceptible to making such an egregious memory error? What is it about memory that makes it so faulty, and so vulnerable to mistakes?
Some psychologists and neuroscientists believe that it has to do with the way our memory is constructed; instead of being a bank of easily retrievable and undiminishing experiences, memories are constantly being updated to accommodate the newest incoming information, even if it is not wholly relevant. Evolutionarily, the natural world is rarely constant. Some natural artifacts, like components of a landscape, remain indelible, but many other features that deserve committing to memory are non-stationary. To account for the fact that past memories may not necessarily reflect the contemporary state of information, the brain is plastic. That is, it readily forges new synaptic connections between previously unrelated neurons, it strengthens existing associations, or it prunes away obsolete ones. By doing so, the brain can stay up to date with the most relevant information even when conditions change.
In many cases, the brain’s ability to edit memories is a boon. But the brain’s sensitivity to new information and its readiness to alter past memories often fails in the context of eyewitness testimony. Pointed questions by interviewers can subconsciously overwrite details from a memory by encouraging a person to think that the interviewer is, in some way, privy to information that they are not aware of. For example, asking “did the victim have dark hair?” can cause the victim to revise their memory and think “maybe he did have dark hair” based on suppositions that the interviewers have already selected suspects.
Questioning immediately after the event of the crime is rife with its own eyewitness contamination based on the constant renewal of memory, but the long-term nature of trials can render eyewitness testimony all but useless. With the checks and balances in place in the American judicial system– as well as the constant overburdening of the courts– trials can sometimes occur months to years after the crime. In that time, the memory can accumulate such a weight of biases and contaminations that it can deviate significantly from the actual events.
Numerous scientific studies have confirmed how easily memories are influenced by external factors. In the 1970’s, psychologist Elizabeth Loftus demonstrated how easily biased questions could tamper with a person’s memory. A subset of study participants were instructed to view a photograph of a car at instruction with a stop sign, whereas another subset viewed an identical photograph with the stop sign substituted for a yield sign. When the study subjects who had viewed the photograph with the stop sign were asked questions about the scene where “yield sign” was substituted for “stop sign”, as “How far from the yield sign was the vehicle?”, researchers found that the study subjects consciously remembered there being a yield sign as opposed to a stop sign. Those whose image originally depicted a yield sign demonstrated the same effect when misled by questions replacing “yield sign” with “stop sign”.
Subtle use of wording can have an equally powerful influence on memory retrieval. Another study pioneered by Loftus in 1974 asked study participants who viewed the video of a head-on automobile accident to estimate the speeds of the cars at the time of their collision, variably using the verbs smashed, collided, contacted, bumped, and hit. Loftus found that subjects who were interrogated using “smashed” and “collided” reported that the vehicles were moving at higher speed than reported by those who were interrogated with the verbs “hit”, “contacted”, or “bumped”. Similarly, subjects interrogated using “smashed” said “yes” to the question “did you see any broken glass?” more often than subjects interrogated with “hit”. Loftus’ work again suggests that subtle wording can influence the ways that a person remembers; even small contextual cues can reconstruct a qualitatively different memory from the one originally cemented.