Remembering Bartlett

Almost a century after its publication, "Remembering" is still worth reading—for its clarity, empirical rigour, and, like all good psychological research, its revelations of what makes us human.

In The Psychopathology of Everyday Life (1901), Sigmund Freud describes a conversation with a young man during a train journey. While discussing rising anti-Semitism, the young man attempts to quote in Latin a verse from Virgil’s Aeneid. The exact quote is “Exoriare aliquis nostris ex ossibus ultor, (“Let someone arise from my bones as an avenger.”). Freud notes that the young man gets the quote wrong; he omits the word aliquis (“someone”). 

Freud, who believed that memory errors and slips of the tongue are not accidents but motivated by unconscious forces, askes the young man what comes to his mind when he hears the word aliquis. After producing associations such as liquid, the man moves on to Saint Simon, who had been murdered as a child, and Saint Januarius, whose dried blood mysteriously turns to liquid on the same day each year. The man then hesitates but, with Freud’s encouragement, admits that he has received a worrying message from a woman. Freud immediately guessed the message was that she had missed her menstrual period. All the man’s associations were connected to this theme: the calendar-like name Januarius, the Saint murdered as a child, and the blood that “miraculously” liquefies. According to Freud, the man’s memory lapse can now be understood. The quotation concerns posterity—a subject weighing heavily on the young man’s mind.

Freud’s “dynamic” view of memory was out of step with the experimental psychology of his time. The work of the German psychologist Hermann Ebbinghaus dominated memory research.  Ebbinghaus’s technique was to memorise lists of nonsense syllables (CAZ, for example), wait a while, and then test his recall.  In this way, he could measure the effects of practice (the more initial practice, the stronger the memory) and chart the natural history of forgetting (rapid initially, then slow).

Ebbinghaus’ approach was deliberately artificial; he wanted to study pure memory untainted by experience or meaning. Simplification is often necessary in science, but Ebbinghaus’s memories had almost nothing in common with those described by Freud. Ebbinghaus’s memories were formed by repetition and stored in a mental warehouse until they faded away. Outside the laboratory, most memories have little to do with repetition and don’t gradually degrade. On the contrary, some memories last a lifetime, changing with time and affecting everything we perceive, do, and say.

When Frederick Bartlett came to the University of Cambridge in the early 1900s, Freud’s view of memory had little influence on academic psychology, and psychology as a science was underdeveloped. There was no Professorial Chair of Psychology at Cambridge or elsewhere in England. Fortunately, CS Myers established a psychology laboratory at Cambridge using his own money and what he could get from relatives. In 1914, Bartlett became the laboratory’s (unsalaried) assistant director. Except for war service, Bartlett spent the rest of his life at Cambridge. In 1922, when Myers retired, Bartlett became Director of the Psychology Laboratory and Reader in Psychology. At that time, no higher position in academic psychology existed in England. When Cambridge finally established a Professorial Chair in 1931, Bartlett became Cambridge’s first Professor of Experimental Psychology. At the time, the only other psychology professor in England was Cyril Burt in London. However, establishing these two professorial chairs should not be taken as signifying a trend. Oxford did not get around to appointing a psychology professor until 1947.

The 1920s and 1930s were productive years for Bartlett. He published books on various topics, including his famous work, Remembering. In this book, Bartlett aimed to demonstrate that memories are not passively stored on mental shelves but actively constructed. He gathered empirical evidence for this thesis using meaningful materials recalled in controlled settings. For example, he asked people to memorise stories such as the Native American tale, The War of the Ghosts. Bartlett chose this story because it was meaningful (unlike nonsense syllables) but foreign to the young British volunteers who participated in his research. The story concerns Native Americans who go out hunting and come upon members of another tribe canoeing upriver to make war on the local people. The warriors ask the hunters to go with them, but one of the hunters protests that he has no arrows, while another complains that his relatives will miss him. The story contains several supernatural elements as well. 

When people were asked to recall this story, Bartlett found that their memory lapses were not random. On the contrary, there were distinctive patterns to what was recalled and what was forgotten. For example, men omitted the “no arrows” excuse, whereas females did not. As many of the male participants were about to leave for war, Bartlett reasoned that males “repressed” this potentially frightening detail. Bartlett also found that males, more often than females, elaborated the “unhappy relatives” excuse for not fighting (adding details that did not exist in the original). Again, grieving relatives may have been high in the minds of men about to go to war.

In addition to such individual concerns, Bartlett found that cultural experiences also affected people's memory. For example, although the Indians in the story “paddled canoes”, the British volunteers recalled them as “rowing boats”. Similarly, British participants omitted or rationalised the culturally alien supernatural events in the story, including the appearance of ghosts, to produce stories that fit their everyday experiences. For example, in the story, when someone died, “something black came out of his mouth”. This was remembered as “he foamed at the mouth”. In this way, Bartlett demonstrated that memories are not passively recorded and retrieved. They are constructed from an amalgam of story details and a person’s experience, vocabulary, and unconscious fears.

One of Bartlett’s most influential ideas was the concept of a schema, a body of organised knowledge in memory. For example, we have schemas for dining out (waiting to be seated, receiving a menu, ordering wine and food, and paying the bill). When faced with new information to learn, we try to incorporate it into an already existing schema whenever possible. For example, in an experiment requiring people to recall a story about someone going to a restaurant, some remembered a description of the person eating and paying the bill even though neither event is mentioned in the story. 

Although schemas may distort recall, they can also improve memory by permitting us to use information already in memory to help organise new knowledge. Adriaan de Groot, a Dutch psychologist, demonstrated this point in his study of chess masters. After a brief exposure of 5 seconds, chess masters could reproduce a chessboard with the pieces in their correct position 90 per cent of the time. Ordinary players achieved only 40 per cent accuracy under similar conditions. But this difference was only noticeable when the boards represented real chess games. If the pieces were randomly distributed around the board, the masters were no more accurate than average players. De Groot concluded that chess masters have developed elaborate schemas for chess games. These help them assimilate board information rapidly and reproduce it accurately. When they are faced with a random board, however, their schemas do not work, and their memory is no better than average.  

The constructive nature of memory has important implications for criminal cases in which credence is placed on eyewitness testimony. American psychologist, Elizabeth Loftus, highlighted these implications in an experiment in which people were shown a film of a motorcar accident. Immediately after the film, viewers were asked one of two questions: How fast were the cars going when they crashed into one another? Or, how fast were the cars going when they hit one another? A week later people returned to answer questions about the film. On this occasion, everyone was asked: Did you see any broken glass? Although there was no broken glass in the film, those who received the “crashed” question a week earlier remembered seeing the broken glass twice as often as those given the “hit” question.  The questions activated a schema for crashes, which included broken glass. Asking the questions determined what was recalled. It is easy to see how certain questions could persuade eyewitnesses that they had seen something different from what they had seen. 

The 20th century saw psychology evolve from a fledgling field into a robust scientific discipline. Bartlett's contributions were pivotal during this transformation. By emphasising the active construction of memory, he moved psychology beyond the mechanical views of memory as mere storage. This shift paved the way for the cognitive revolution in the mid-20th century, where understanding mental processes like memory, perception, and problem-solving became central to psychological research.

Bartlett’s work highlighted the interplay between individual experiences and cultural contexts in shaping memory, anticipating later social and cultural psychology developments. His concept of schemas has been foundational in various areas of psychology, including developmental psychology, which helps explain how children learn and assimilate new information, and clinical psychology, which informs cognitive-behavioural therapies.

Bartlett’s insights into memory are also profoundly relevant to developing artificial intelligence (AI) and machine learning. AI systems today are built on the premise of learning from data, much like humans learn from experiences. Concepts akin to Bartlett’s schemas can be seen in how AI models use prior knowledge to make sense of new information.

The idea that memories (or data) are actively constructed and can be distorted by biases aligns with challenges faced in AI, such as the need for algorithms to handle noisy, incomplete, or biased data. Moreover, understanding how human memory works can inspire more sophisticated and human-like AI systems that better mimic human learning and memory processes.

As AI advances, incorporating more nuanced and human-like memory systems, Bartlett’s pioneering work remains a touchstone. His emphasis on the dynamic, constructed nature of memory and the role of individual and cultural factors offers valuable lessons for cognitive scientists and AI researchers striving to create machines that can learn and adapt in increasingly human ways.

Bartlett was elected a Fellow of the Royal Society and knighted in 1948. Today, almost a century after its publication, Remembering is still worth reading—for its clarity, empirical rigour, and, like all good psychological research, its revelations of what makes us human.

References:

Bartlett, F. C. Remembering. An Experimental and Social Study. (1932). Cambridge University Press.