P1.M.3Memory as an active process: Bartlett's War of the Ghosts study and theory of reconstructive memory

Frederic Bartlett (1932) rejected the idea that memory is a passive recording. He argued that we reconstruct memories every time we recall them, fitting fragments into our existing schemas — mental frameworks built from past experience.

The "War of the Ghosts" study

Bartlett asked British participants to read a Native-American folk tale ("The War of the Ghosts") and reproduce it after delays of varying lengths (minutes, days, even years), using two methods: repeated reproduction (the same person retelling it several times) and serial reproduction (a chain of people, each retelling to the next).

The story contained unfamiliar elements: hunting seals, mysterious ghosts, "something black" coming from the dying man's mouth. Bartlett observed three systematic distortions:

1. Levelling — losing details. Participants dropped the unfamiliar (e.g. the Native-American context, the ghosts).
2. Sharpening — adding emphasis to certain elements that fitted their own culture.
3. Rationalisation — changing the story to make sense within British schema. "Hunting seals" became "fishing"; the ghost element became background; the eerie ending became more conventional.

Schemas

A schema is an organised packet of information about a topic, situation or person stored in long-term memory. Schemas help us interpret new information rapidly — but they also distort memory. When the new material clashes with existing schema, the schema usually wins: people "remember" details that match their expectations even when those details weren't there.

Strengths and weaknesses of Bartlett's claim

Strengths:

• High ecological validity — uses meaningful prose rather than nonsense syllables.
• Real-world implications: eyewitness testimony is famously unreliable for the same reasons.
• Replicated across cultures.

Weaknesses:

• Bartlett didn't standardise his procedure (no fixed delay between exposures), making it hard to replicate.
• Some distortions may reflect demand characteristics — participants knew the story was being tested.
• Subsequent work (Wynn & Logie, 1998) showed memory is more accurate than Bartlett implied when participants are told to be exact.

Why this matters for the exam

This topic is the bedrock for understanding eyewitness testimony, false memory, and the difference between what we experienced and what we remember. When asked to evaluate accounts of events, mention reconstruction, schemas, and at least one of the three distortion types.