Expertise and attention
So, my last few posts started to describe a theory of expertise as something like: the set of factors acquired over time that help someone achieve success in a domain (see this and this for some background), and specifically the (picture my hands waving madly now) “mental” ones. I think it’s time to say (just a bit) more about that.
As of right now I think of expertise as some mix of: representational knowledge (facts, etc.), neuromuscular patterns and changes (the underpinnings of domain-specific movement — for instance, the increased sensitivity and dexterity of a right-handed guitarist’s left hand), and attentional control and discrimination within the domain (sometimes called perceptual expertise or learning, but I want to expand it beyond what that normally implies). Different domains require different mixes (theoretical physics and chess are light on the neuromuscular aspects, for instance).
For this post, I want to focus on the last element: attention.
Expert performers attend to different things than novices do; that’s a well-established finding across many fields, and it helps to explain several other facets of high-level performance — like domain memory, where chess masters can remember full boards after brief exposure. In some cases, that attention is on things that novices don’t even know exist, like the differences between one LBB and the next. In others, the attention is on higher-level patterns — expert performers are seeing the forest instead of focusing on the individual trees.
Expert performers also neglect to attend to some aspects that novices will carefully watch. For instance, a study on burglars in the UK compared how burglars and (presumably law-abiding) homeowners judged whether a house seemed like a tempting target; experienced burglars ignored a number of factors that the homeowners naively thought relevant.
The burglar studies also have interesting implications for automaticity as a core part of expertise. Automaticity is developed when actions take progressively fewer attentional resources — for many, reading is automatic in that you don’t have to think about it. You see a word, you immediately grasp what it says without having to think about or attend to translating the marks into meanings.
Some theories of expertise — the Dreyfus model, prominently — hold that expertise just is the maximal set of automatic action. You see a tennis star and they seemingly don’t think — they just play. This matches with a lot of people’s experiences as they get better in a domain — they stop thinking so carefully about, for instance, which fingers to press to produce a C# on the flute, and can concentrate on just playing the C#.
I think that last sentence is the key: as you develop expertise, you develop more automaticity at progressively-higher-levels of analysis. As a novice, you have to be conscious of pretty much everything (and, accordingly, you can’t actually do that much — it’s too hard to focus on everything). As you develop expertise, you’re able to automate away the most mechanical skills — you don’t have to spend your attention on “how am I holding the bat?,” and can instead devote more of it to “where is the ball?” Continue practice, and you start to automatically do more and more.
But! It’s not just a matter of changing what you focus on — expert performers are also able to step back (or closer) and go back to focusing on the things that they’re doing automatically. An Olympic volleyball player can go back to thinking about the most minute positioning of her elbow and shoulder as she serves, if she thinks there’s something wrong. The best performers can choose what they attend to, where novices don’t have that freedom.
And what’s more, expert performers can be broken out of their automatic actions (or, what is the same thing, out of their focus on higher-level actions). Back to the burglars: practiced criminals report that their searches within a house are basically automated: they check the same places in the same order, because people are predictable. If they hear an unexpected sound, though, the search pattern is broken and they have the chance to consciously decide what to do next.
There’s more here — there are really interesting theories about choking based around misplaced attention, for instance, but I’ll save that for another post.