In praise of high-level cognitive control when performing complex tasks

In Going from Good to Great with Complex Tasks, Ozgun Atasoy explains that the belief that consciously thinking about what we are doing, when performing complex tasks, by definition harms our performance, is wrong. It is true that some type of conscious thinking can harm our functioning. For example, when we are typing on a keyboard, we run largely on auto-pilot. If we would try to consciously control the typing of each separate letter, this would slow us down a great deal and probably cause us to make many mistakes.

But, as Atasoy explains, when we are performing complex tasks, running on autopilot (performing it with little cognitive control/engagement) leads to sub-optimal performance in the sense that our performance becomes rigid. We lose the ability respond to unexpected events. Also, the task is likely to become boring, this way.

While low level conscious control is impossible (as explained above), high level control is possible and is likely to improve our performance. For example, a pianist cannot control low level activity (which fingering shall I use for this chord?), and should not run on auto-pilot (without cognitive control). The pianist can play with high level cognitive engagement by focusing on higher level mental events or concepts such as the type of emotion the piece should express. For example, an experiment by Ellen Langer demonstrated that an orchestra performed better when increasing their high-level engagement when performing a symphony. So, the key is, as Atasoy says, to stay consciously engaged, but in the right way, by focusing on high-level concepts.


Yes and. . . what does 'high level' mean to a parent struggling with a five year old or a manager struggling with her first meetings at a new job? The problem is bridging from a theoretical appreciation of 'high level concepts' to a pragmatic approach to improving at activities much less defined and much more dynamic than one pianist practicing a solo.
Coert Visser said…
I agree, that may be more difficult to imagine and I agree that this is precisely what the challenge is. In our conversations with clients we are also performing highly complex tasks. through to our training and years of experience we have automated many aspects of doing these conversations.

For example, we may have learned to ask certain types of questions and to apply certain types of interventions. At one time, when we were learning them they may have beeen hard for us, but not anymore. Through are years of training and experience we have automated many of the skills required to talk with clients.

But it would be unwise to do such conversations on 'autopilot'. How can we do these conversations with high-level cognitive engagement? I deiberately use many ways to try to do this. For example, I ask myself many types of questions when having a conversation with clients, in order to try understand well what the clients are saying and what they want to accomplish. This helps me to remain very conscious during the conversation and I think this improves the quality of my contribution.

I am curious what high-level cognitive engagement you use or can think of, Linda (and other readers, of course).

Thanks for your comment!