Mechanisms of Forgetting: Which Is Most Functional? Plus What Animals Teach Us About Learning (Salwiczek et al., 2012)

Part 1: The three mechanisms of forgetting

Forgetting is usually seen as a failure of memory. In the psychology of learning, however, forgetting is better understood as a normal process that can be adaptive: it helps organisms keep their behavior matched to a changing environment (Bouton, 2016). Learning researchers describe three main mechanisms of forgetting:

1. Decay (trace decay)

Decay theory holds that memories fade simply with the passage of time if they are not used, as the physical trace in the brain weakens. Ebbinghaus’s forgetting curve shows rapid loss shortly after learning, then slower loss. Decay is hard to test in isolation, because time is always filled with other experiences that could cause interference. Neuroscience does show that the brain actively weakens some memories, for example through synaptic pruning and the birth of new neurons in the hippocampus, which can degrade older memories (Richards & Frankland, 2017).

Function: decay clears out information that is no longer used, saving storage and processing effort. Information that is used often is rehearsed and stays strong; information that hasn’t been needed for a long time is probably less likely to be needed again (Anderson & Schooler, 1991).

2. Interference

Interference occurs when other learning disrupts a memory:

• Proactive interference: old learning interferes with new learning. For example, you keep typing your old password.
• Retroactive interference: new learning interferes with recall of old learning. For example, after learning Spanish, you struggle to recall French vocabulary.

In conditioning, extinction and counterconditioning are examples: a tone that once predicted food now predicts nothing, and the new learning competes with the old.

Function: interference lets new, more relevant information take priority. When the environment changes, such as a food source moving or a predator appearing in a new place, updating behavior is more useful than holding on to outdated information.

3. Retrieval failure (cue-dependent forgetting)

Retrieval failure means the memory is still stored but cannot be accessed because the cues or context present at learning are missing at test. Examples include the tip-of-the-tongue experience and context-dependent memory (recalling more in the same room where you studied). Learning research shows that extinguished responses often return through spontaneous recovery (after time passes) and renewal (when the context changes), proving the original learning was not erased. Bouton (1993) argued that time itself acts as a changing context, so much “forgetting” over time is actually retrieval failure.

Function: retrieval failure allows the brain to keep multiple memories and retrieve the one that fits the current context. What was true in one setting may not be true in another, and the old memory can be retrieved again if circumstances return.

The debate: which mechanism is most functional?

Sample position. Your instructor may ask you to defend one mechanism; any of the three can be argued with evidence.

I argue that retrieval failure is the most functional mechanism of forgetting. Unlike decay, which permanently loses information that might be useful again, retrieval failure keeps the memory stored and makes it available only when the right cues return. This is efficient in a world where conditions change but often repeat: an animal that learned a food patch was depleted in winter does not need to relearn its location from scratch in spring, because the returning context cues the old memory. Research on extinction supports this view: extinguished fear or appetitive responses return through spontaneous recovery and renewal, showing that new learning does not erase old learning but makes its retrieval context-dependent (Bouton, 1993, 2016). Retrieval failure also includes the benefits of interference, since new learning can dominate in the current context, without its cost of losing older knowledge. Decay offers simplicity, and interference supports rapid updating, but only retrieval failure lets organisms flexibly use the memory that best fits the present situation while preserving the rest.

Part 2: What can we learn from animals that solve some problems better than humans?

Much of what we know about learning comes from rats, pigeons and other animals, based on the assumption that basic learning processes, such as classical and operant conditioning, are shared across species. That assumption has been productive, but evolution has also shaped each species to solve the problems most important for its survival. Some animals therefore outperform humans on specific tasks.

Problems other animals solve better than humans

Salwiczek et al. (2012): fish versus primates

Salwiczek and colleagues (2012) tested a striking prediction. Cleaner wrasse are small reef fish that eat parasites off “client” fish. Some clients are residents that will wait for service, while others are visitors that leave and go to another cleaner if they are not served quickly. A cleaner that serves the visitor first gets to eat from both clients, so in nature, cleaners learn to prioritize visitors.

The researchers turned this into a laboratory task. Subjects chose between two plates offering the same food. One plate stayed if it was not chosen first (like a resident), while the other was removed if not chosen first (like a visitor). Choosing the “ephemeral” plate first earned both rewards. The immediate rewards were identical, so the only way to succeed was to learn the delayed benefit of the correct choice.

• Wild-caught adult cleaner wrasse learned the task quickly, and even relearned it when the rule was reversed.
• Juvenile cleaner wrasse did not show this ability, suggesting the skill depends on experience in the wild.
• Primates struggled: within the first 100 trials, no capuchins or orangutans and only two of four chimpanzees learned the task, and many developed a side bias instead. Most primates eventually learned only after the procedure was modified.

The authors concluded that adult cleaners succeeded because repeated exposure to this exact problem in nature produced specific learned decision rules, not because fish are “smarter” than apes. Primates, which had previously learned to choose actions giving immediate double rewards easily, struggled when the benefit was delayed and the task did not match their ecology.

Can we still learn about learning from these problems?

Yes, and these cases may teach us some of the most important lessons:

• Learning is shaped by ecology and evolution. Performance depends on how well a task matches a species’ natural problems. This is the “ecological” or “adaptive specialization” view of learning (Shettleworth, 2010).
• Intelligence is not a single ladder. Ranking species from “smart” to “less smart” is misleading; each is specialized. A fish can outperform apes on one problem and fail on others.
• Experience interacts with biology. Juvenile wrasse failed where adults succeeded, showing that natural experience shapes learning abilities.
• Shared principles still exist. Across species, animals learn from reinforcement, delay and prediction error. Comparing species reveals which processes are general and which are specialized.
• Better research design. Tests should be ecologically valid; a species that “fails” may simply face a task that doesn’t fit its sensory or ecological world.
• Applications to humans. Studying specialized abilities, such as cache memory in birds, has advanced understanding of the hippocampus and spatial memory, and taste aversion research has informed care for patients undergoing chemotherapy.

Sample discussion post

Although humans are often seen as the most intelligent species, evolution has equipped many animals to solve specific problems better than we can. Bats use echolocation, food-caching birds remember thousands of hiding places, and rats learn to avoid poisoned food after a single experience. Salwiczek et al. (2012) showed this clearly: wild-caught adult cleaner wrasse quickly learned to choose an “ephemeral” food plate first to gain two rewards, mirroring how they serve visiting clients before residents on the reef, while most capuchins, chimpanzees and orangutans did not learn the task within 100 trials. Juvenile wrasse also failed, suggesting that ecological experience shapes this ability. These findings do not mean fish are smarter than apes; they show that learning is tuned to the problems each species faces. We can still learn a great deal from such problems. They reveal which learning processes are shared across species, such as sensitivity to reinforcement and delay, and which are specialized adaptations. They also remind researchers to design ecologically valid tasks and to avoid assuming a single scale of intelligence. Studying how other animals solve problems we cannot helps us understand the evolution and diversity of learning, including our own.

Related reading: psychology theorists study guide and Person-Centered Therapy.

Conclusion

Forgetting is not simply a flaw. Decay clears out unused information, interference allows new learning to take priority, and retrieval failure keeps memories available for the contexts where they matter, which makes it a strong candidate for the most functional mechanism. Comparative research shows that learning is also shaped by each species’ evolutionary history: cleaner wrasse can outperform great apes on a task that matches their reef ecology. Studying such specialized abilities does not weaken the study of learning; it deepens it, showing both the general principles we share with other animals and the unique adaptations that make each species, including humans, distinctive.

References

Anderson, J. R., & Schooler, L. J. (1991). Reflections of the environment in memory. Psychological Science, 2(6), 396–408. https://doi.org/10.1111/j.1467-9280.1991.tb00174.x

Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114(1), 80–99. https://doi.org/10.1037/0033-2909.114.1.80

Bouton, M. E. (2016). Learning and behavior: A contemporary synthesis (2nd ed.). Sinauer Associates.

Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Science, 4(3), 123–124. https://doi.org/10.3758/BF03342209

Richards, B. A., & Frankland, P. W. (2017). The persistence and transience of memory. Neuron, 94(6), 1071–1084. https://doi.org/10.1016/j.neuron.2017.04.037

Salwiczek, L. H., Prétôt, L., Demarta, L., Proctor, D., Essler, J., Pinto, A. I., Wismer, S., Stoinski, T., Brosnan, S. F., & Bshary, R. (2012). Adult cleaner wrasse outperform capuchin monkeys, chimpanzees and orang-utans in a complex foraging task derived from cleaner–client reef fish cooperation. PLoS ONE, 7(11), e49068. https://doi.org/10.1371/journal.pone.0049068

Shettleworth, S. J. (2010). Cognition, evolution, and behavior (2nd ed.). Oxford University Press.