Accessibility links

A while ago Matthew Taylor posted on some research suggesting that crude financial incentives can make you stupid - they make you think in narrow linear terms instead of creatively and holistically. He was commenting on an experiment involving the 'candle problem' popularised by Dan Park:

A while ago Matthew Taylor posted on some research suggesting that crude financial incentives can make you stupid - they make you think in narrow linear terms instead of creatively and holistically. He was commenting on an experiment involving the 'candle problem' popularised by Dan Park:

In this exercise subjects are shown a picture of a table next to a wall. On the table is a candle, a book of matches and some drawing pins in a box. The task is to attach the candle to the wall over the table, light it, but not let it drip wax on the table.
On average it takes people about ten minutes to identify the solution. This is to take the drawing pins out of the box, pin the box to the wall, then stand the candle on the box so the wax drips on to it rather than the table. This requires the subjects to make the lateral leap of seeing that the box holding the pins is not just a receptacle for another object but an object in itself.
In this test those who are offered a cash prize for completion perform less well than those who are simply asked to solve it as quickly as they can. Fascinatingly, if the test is made easier - by taking the pins out of the box so it can be seen from the start as an object in play - then those offered incentives perform better than those not.
What has puzzled me about this is the neurology. What part of the brain is dominant when people are motivated by the cash incentive and are consequently worse at solving the complex problem (and better at solving the simple problem)?

The brain is a very, very complicated thing. But it can be understood in terms of roughly two systems. The first is the conscious system of deliberation, thought and decision-making. This bit is centred around the pre-frontal cortex, an area where information can be 'restructured' creatively when one is solving a problem. In the candle problem one has to see the box in a different way than it first appears, one cannot necessarily just follow one's past experience (doing this commonly results in trying to fix the candle to the box by heating its standing end). So it seems the pre-frontal cortex has to be engaged to solve this problem.

The unconscious brain has many functions but one of them is to predict future events based on past ones through learning and memory. This can happen automatically, as Antonio Damasio's famous gambling experiment showed: our behaviour can start changing on orders from our unconscious brains before we are consciously aware of a reason. This occurs because the unconscious brain can run cognitive processes in parallel - it can process millions of bits of data at once, so is much faster than the conscious brain which can only hold a small number of pieces of information in mind (between five and seven seems to be the limit). Much of our behaviour is driven unconsciously even when it appears conscious - Benjamin Libet's famous experiment showed that when we press a button our unconscious brain has started the action before we are aware.

When someone solves the candle problem what seems likely is that the conscious brain 'takes over' cognitive work and the pre-frontal cortex 'restructures' the problem. In a comment on Matthew Taylor's blog post a member of the Social Brain steering group, Ben Seymour, noted that in an experiment he had run, in responding to crude incentives the parts of subjects' unconscious brains concerned with basic reward-responding were highly active, crowding out the higher cognitive functions of the pre-frontal cortex. The inference to draw seems to be that the 'emotional' unconscious brain needs to be offline so the conscious brain can do its thing.

That would seem right at first blush. The candle problem is quite simple even in its complexity - it only involves a few elements. So the pre-frontal cortex is capable of the requisite 'restructuring' work that crude reward circuits would interfere with. But is this right? A couple of experiments have shown that insight into complex problems that require lateral thinking (such as solving word puzzles) actually results from unconscious cognition. This can be shown by distracting someone solving a problem so that their conscious brain is out of action. They then get struck by a 'feeling' of insight that must come from their unconscious brains (and which fMRI scans confirm does).

It seems more likely to me that the pre-frontal cortex and the unconscious brain work together, rather like an airline pilot works with an autopilot. The pilot directs the autopilot to make certain computations when he is required to solve a problem. It then does these computations and the pilot checks if they are right or not. Or, think of making a move in chess: a good player doesn't hold all the possible patterns in his/her head (he/she couldn't, there are far too many), rather she or he directs the unconscious brain to whir through thousands of possibilities with certain general instructions in mind (like 'trap his queen but don't leave my bishop exposed'). Various possible moves shoot up into consciousness with that 'oh yeah' feeling, and are checked by the rational brain (which often results in an 'oh no' feeling because the move isn't right).

That co-operation between conscious and unconscious brain processes seems likely to be how complex problem solving works. We are trapped into talking dichotomously about conscious processes operating at the exclusion of unconscious ones (and vice a versa) by a residual dualism. But from Damasio's work on emotion we know that the brain is not dualistic - conscious and unconscious systems are not separate, they are in fact beautifully woven together.

On this non-dualist understanding, what happens with simple incentives is that crude reward systems kick in and the cooperation between pre-frontal cortex and unconscious brain is momentarily backgrounded.

Why is this important? Because if you thought that only the conscious brain worked on complex problems you would think that what is required is practice at deliberative sequential problem-solving (think of someone solving a problem and giving a running commentary on the sequence of steps - 'first I do this, then that, and then...'). But what is required is that plus the intuitive thinking of the unconscious brain. In other words, sequential thinking integrated with rumination. And that is important to know if you are involved in organising how people learn and work. After all, if practice makes perfect, we want to practice the right things.


Be the first to write a comment

Please login to post a comment or reply.

Don't have an account? Click here to register.