Why Memories Need to Be Lost

This is a draft invited blog post for The Memory Network.  Update: the revised & edited blog version is here:  Memories are for Action, Not Just for Keeping.  A slightly longer version will also appear in the Memory Network's forthcoming book.

People have the wrong idea about archives, like they have the wrong idea about saving money.  If you bury money in your garden, you won't get any interest on it nor help build any new homes or companies.  You might think the risk of losing it is lower, but someone or something else might dig it up, or it might stop being a valid form of currency, or you might forget where it is.

Memory is like that, only more so.  Just as the purpose of money is transferring wealth, the purpose of memory is action.  As children we develop our ability to plan at roughly the same time we develop our ability to rememberWe use the same parts of our brain to remember as to imagine or plan.

Why do we need memory?  Because the world at any instant does not provide us with enough information to make the best decision about what to do next.  Even simple animals and plants have a kind of memory.  A tree drops leaves that are in the shade and grows towards leaves that are in the sun, because it's quite likely that the shade & sun tell it where other trees are, and the other trees will be there tomorrow as well as today.

Some species use their own memory less than others.  Trout can pretty much figure out where they need to be by what they smell.  They can find food, mating partners, fresher water that way.  So a fish brain takes in smell, translates it into goals, then pushes that into movements of its spine.  Since we all evolved from fish, our brains are shaped that way too – our goals are selected by our forebrain which is right up near the olfactory lobe and the nose.

But some species work in niches where the immediate environment doesn't give that many clues about where the next meal is coming from, or how they can successfully collaborate to survive. The reason species can work on such hard problems is that they've co-evolved bigger brains with their adaptive behavioural strategies.  So primates can spread out through the forest and figure out which trees are fruiting right now.  Birds have smaller brains but travel faster, and similarly exploit fruit that trees have evolved to hide until the day that all their seeds are likely to be mature enough to produce new trees after their fruit is spread.  The point is that even a bird brain takes much more information from memory than a fish brain, and directs it up to the goal-choosing forebrain.  In fact if you look at a human brain, the information from our ears, touch and eyes have to go via the locations we use to associate information with memory before they come back forward to the planning.  The olfactory lobe is still in its evolved location, but humans have worse sense of smell than most other species.  We call species that integrate memory into plans to choose their behaviour cognitive.

When humans are born, we cannot perceive or understand most of what happens around us.  In fact, that's true for adults as well, but adults have learned an amazing range of things that makes us the most cognitive of all species.  Not only do we learn from our own experience, but we learn from the experience of others.  This is also true of other animals, not just chimpanzees, but we're coming to realise that many species mine information from the memories of others.  But humans do more than this.  We have evolved the capacity to share stories.

Of course, stories are more than just memory about useful ways to operate in the world.  Stories are also a form of art, and as such, they contribute to our idea of identity.  Group-level identity allows us to achieve both great and terrible things, but overall the effect of our incredible capacities has been our ability to vastly increase our numbers and to dominate almost every ecology on the planet.  Before the evolution of contemporary levels of art, language, writing and agriculture there were more macaques on the planet than humans.

All of these advantages to memory might make you think that holding onto it would be the best plan.  But the fact is, if memory is for acting, then there is no reason to remember things that no longer affect our behaviour.  In fact, it's worse than "no reason".  There is a computational and therefore cognitive cost to remembering too many things.  It takes time to search through memory, and we don't want to accidentally apply a principle that no longer holds.  Forgetting is our brains automatically lose access to that which is too long ago. Some people think the same is true about reproduction: generally cognitive animals live longer than non-cognitive ones, perhaps because our memory is such a resource not only to ourselves but to our children.  Nevertheless, we age and die, and this seems not to be an accident but a regulated biological trait, just like how tall we grow or how many fingers we have.

Of course, just like fingers and height there is variation in how good people's memory is.  Because we have been a social and a cognitive species, in fact genus, for millions of years – this is something we share with chimpanzees! we may be able to rely on the fact that some people will have long memories to help us if an old problem comes back, while others will always be ready to try new things.  The evolution of cooperation is not a mystery (as it is sometimes portrayed in the press), but it is one of the most exciting and challenging areas of research with more understanding emerging every day.  My group is one of many working to understand where our society comes from and how it works.  If you would like to read more academic papers about the evolution of human society and language, I have a list there.