Tuesday, September 20, 2011

Daniel Schacter, The Seven Sins of Memory (2001)

I started this blog, in small part, because I could not always recall when I read a book and what I got out of it. I thought about making notes in a notebook, like I did in college, and thought why not a digital notebook? The subject memory has continuously popped up in my postings, as the previous post recalls. In The Seven Sins of Memory, Harvard psychologist Daniel Schacter explains why I can't recall when I read a particular book and other failings of the human memory. And it's not all bad. In fact, Schacter says it's OK to forget. It's common. It's natural.

A clever title. Like the "seven deadly sins," the title is not an allusion to sinful behavior. It is a reference to our vulnerabilities and imperfections. Seven imperfections of memory: transience (memory that fades with the passage of time); absent-mindedness (rapid forgetting due to attentional lapses); blocking (hopefully a merely temporal phenomenon, like the name on the tip of your tongue that you can't remember); misattribution (mistaken identification); suggestibility (tendency to incorporate misleading information from external sources by suggestion into personal recollections); bias (five tendencies in which we generalize to reduce dissonance, reconstruct the past to fit the present, organize and regulate our mental life, and categorize; and persistence (remembering things we really want to forget).

All of us are familiar with and have experienced most, if not all of these imperfections of our memory. Some memories are encoded and "stick." Eric Kandel and others have described to us much of the biological basis about why we understand certain memories stay around longer than others --- long term potentiation. But not all memories stick. Some memories disappear immediately (absent-mindedness, because encoding fails due to divided attention); other memories stick, but are not retrievable --- sometimes for reasons we still do not fully understand --- but sometimes because we have little opportunity ever to recall a personal experience or piece of knowledge again and we forget(transience). And some memories are simply not accurate --- misidentification due to the inability to have specific recall of people or events that we are generally familiar with, or revised because external sources of information suggest to us that something different occurred, or biasing influences.

The bias influence is something we are already familiar with from Michael Shermer's book, The Believing Brain. (See June 12, 2011 post). What Shermer called the "confirmatory bias," Schacter refers to as the "consistency bias," where we infer past beliefs from our current state. This bias promotes psychological stability, perhaps avoiding cognitive dissonance. Similarly, hindsight bias, Schacter says, is ubiquitous, and reconstructs past to fit the present. What Shermer called the "self justification" bias, Schacter labels the "egocentric bias" that gives more credence to our own recollection of events --- reflecting the role of "self" in organizing and regulating our mental life. Studies show that when the mind encodes new information by relating it to the self, subsequent memory improves. Finally, the "stereotypical bias" reflects the minds tendency to categorize. What Shermer failed to address is the neurobiological basis for these biases. Schacter tries to address this, and, not surprisingly, calls on research by Michael Gazzaniga (see September 27, 2009 post) to explain it. The source of bias probably begins in the left hemisphere of the brain, which is adept at coming up with explanations and rationalizations for situations. This is the region where Gazzaniga's "interpreter" resides, the area responsible for language and symbols. The left hemisphere is responsible for inferences and generalizations to relate past and present. It draws on general knowledge and past experience to bring order to our psychological world. In contrast, the right hemisphere is where images and spatial location are prominent. The right hemisphere has a proclivity to respond on a literal basis to our environment, and acts as a check on the generalization of the left hemisphere. The left hemisphere probably contributes to consistency bias, hindsight bias, and egocentric bias. The left hemisphere is where our storytelling capacity resides. It is also where memories may be strengthened. Thinking and talking about our experiences changes the likelihood of subsequent remembering, reducing transience: elaborate encoding occurs to create memory.

I am reminded of an article by Joseph LeDoux of New York University in The Scientist back in 2009, entitled "Manipulating Memory." LeDoux reports on work that supports a theory of "reconsolidation" whereby memories are updated and influenced by new information. In other words, once a memory is recalled it is in a fragile state, and iit s susceptible to disruption by the very act of remembering so that what is "remembered" is not the same as what was experienced at the time of the event recalled. Several of Schacter's "sins" seem to support this view of human memory that it is neither fixed nor permanent.

These vulnerabilities of memory are not design flaws. They are likely by-products of of adaptive features of memory that have, in the long-run, served human memory well. Schacter reminds me of Antonio Damasio's reliance on the concept of homeostasis (see April 8, 2011 post) when he refers to the brain's "trade-off" between reducing the need to access information that is not needed or has not been needed and the cost of forgetting. The brain has adapted to an equilibrium condition. Transience occurs because there is little utility in storing unimportant, noncurrent information. Retrieval of "too much information" (data overload) has negative consequences and sacrifices abstract thinking. The brain does go on to autopilot --- repetitive events are handled by automatic processes, freeing up the brain's ability to devote attention to things with consequences; as a consequence occasional absent-minded errors occur when the mind is focused on something else, a small cost for the larger benefit of being able to pay attention. Memory sometimes operates because cues trigger recall. Contrary to some popular philosophers, David Hume's associationism is not dead. (See February 27, 2011 post).

The ability to record (encode) the gist of what happens and not every detail is a strength of memory, says Schacter. This capacity is fundamental to categorizing and comprehending and allows us to generalize across experiences without dwelling on details. Misattribution (false recognition) is probably a price we pay for the benefit of generalization. Generalizing, as discussed above, also leads to biases. Some features of memory are adaptations in evolutionary terms -- larger hippocampuses that facilitate finding stored food, and the amygdala that plays a role in emotional conditioning (contributing to the sin of "persistence"). But several features of our memory are unintended by-products of an existing feature or functionality of the brain. Schacter postulates that aspects of memory are what Stephen J. Gould referred to as "exaptations," features co-opted from an existing function that enhance fitness, but were not built by natural selection. Biases, Schacter concludes, are an incidental by-product of general knowledge and beliefs; blocking, absent-mindedness, misattribution, and suggestibility are exaptations of a memory system that does not routinely preserve all details required to specify the source of an experience.

There are practical concerns that arise from the learning in this book. The "sins" of misattribution, suggestibility, and bias all have implications for our justice system -- particularly criminal justice. Eyewitness misidentification, false confessions, witnesses influenced by suggestive questioning that retrieves a memory that does not match the actual experience are particularly worrisome. Law enforcement has taken notice and is learning from this research.

That we recognize that human memory is vulnerable, and that it is neither fixed nor permanent, informs us with respect to our history as well. While there are histories that are literally revisionist histories --- Schacter cites George Orwell's 1984 where a totalitarian government deliberately revises the past to suit the present --- our oral and written histories are known to have been revised to suit a current agenda. Knowing that "gospels" are the product of human writers, editors and redacters, it is difficult to understand why some have difficulty in acknowledging that the stories in these gospels suffer from the same vulnerabilities that human memory suffers. There are undoubtedly other psychological states that produce such difficulty.