Abstract Rewards and Other Minds

(Yesterday I posted about a fascinating article in Seed Magazine on dopamine, artificial intelligence and social learning.  This is the second post inspired by that article.  It may make more sense if you read yesterday’s post first.)

Humans are odd creatures. We often prefer the fulfillment of abstract ideas to concrete rewards. In the words of Read Montague as quoted by Jonah Lehrer of Seed Magazine:

“What other animal goes on hunger strike? Or abstains from sex? Or blows itself up in a cafe in the name of God?”

We find abstract ideals highly rewarding because our brains use the same kinds of processes to evaluate implicit / abstract ideas and explicit / concrete factors.  The mind uses neuronal firing rates to compare the relative merits of the options we are faced with.  If the abstract idea excites our neurons more than the concrete one, we prefer it.  According to Montague:

That’s what makes ideas so powerful: No matter how esoteric or ethereal they get, they are ultimately fed back into the same system that makes us want sex and sugar.

Montague believes that human beings are the only animals that chase ideas instead of primary rewards.  That we are the only species that will go so far as to  reject a concrete reward in lieu of pursueing an abstract goal.  I’m not sure I agree.

Montague bases this idea of the intrinsic value of chasing abstract rewards partly on a simple economic game called the “Prisoner’s Dilemma” where receiving the maximum reward for all players requires strangers to trust each other.  The rules of the game allow one selfish player to reap more than his “fair” share of the reward and break the bond of trust.  Montague was able to predict several seconds in advance when a player would defect by monitoring his caudate nucleus because activity there would decrease.

The caudate nucleus consists of a pair of C-shaped structures of the striatum.  They contain input neurons for voluntary control of movement and are vital in learning, trust, memory and pleasure systems — and have been referred to as one of the reward centers of the brain.  There is some evidence that the caudate is also important in developing self-control and falling in love.  But it’s the caudate’s importance in trust-feedback loops and abstract reward systems that interest us here.

We’ve wondered for a long time why cooperation exists in man and other animals.  Especially in cases where it doesn’t result in concrete or immediate rewards. Trust is the basis of altruistic cooperation – and cooperation is something that dogs excel at.

Species like dogs, wolves and humans who cooperate to survive need to have mental processes that allow them to tell the difference between cheaters and cooperators.  Understanding the rules that govern the expression, repayment and betrayal of trust are vital when you live a social, cooperative life because they form the basis of things like reciprocity, reconciliation, consolation, conflict intervention, and social signaling.

Trust is the foundation of a healthy social fabric — but how do we learn to trust?  Are social species like humans and dogs born with an innate capacity to trust, or do we have to learn how to do it?  According to Montague, we are born with the capacity to trust, but we need to learn who to trust and how much to trust them:

But what exactly is the caudate computing? How do we decide whom to trust with our money? And why do we sometimes decide to stop trusting those people? It turned out that the caudate worked just like the reward cells in the monkey brain. At first the caudate didn’t get excited until the subjects actually trusted one another and garnered their separate rewards. But over time this brain area started to expect trust, so that it fired long before the reward actually arrived. Of course, if the bond was broken — if someone cheated and stole money — then the neurons stopped firing; social assumptions were proven wrong. 

It appears that once our brains learn to trust, we find the reinforcement of that trust highly rewarding.  It feels good when our beliefs are validated.  It feels bad when our expectations of trust or reward are violated.  And that’s a good thing because it’s the contrast between these two situations that lets us learn complex things.  So, yesterday we discussed the fact that mistakes are a vital part of learning and now today we find that having our trust and expectations violated is an important component as well.  Now it makes perfect sense that moderate levels of stress can be conducive to learning.

Trust and empathy give us insights into the minds of others.  These capacities not only allow us to live with other animals in a rich social fabric; learning how to accurately predict how others will interact with us may also have sent us the first tentative steps down the road to having a theory of mind.