Attention

Attention, considered as a scarce resource rather than as a faculty of the soul, is the bottleneck through which all thinking must pass. The observation sounds austere, and it is meant to. Psychology inherited from philosophy a picture of attention as a power the mind exercises — a lantern, a spotlight, an act of will — and spent a century refining the metaphor. The picture is not so much wrong as unhelpful. It tells us what attention feels like from the inside; it does not tell us what work attention does, why it is limited, or what follows from the limitation. Those are the questions worth asking, and they turn out to be questions of economics and of design.

Begin with a fact about the world rather than a fact about the mind. The environments in which human beings now think are rich in information — richer by many orders of magnitude than the environments in which the species learned to think. Whenever a resource becomes abundant, some other resource, complementary to it, becomes the scarce factor. What information consumes is the attention of the person who receives it. A wealth of information therefore creates a scarcity of attention, and it imposes a need to allocate attention efficiently among the many sources that might consume it. This single proposition, plain as it is, carries most of what follows: the design problem of the mind, of the organization, and of the school is not to gather more information but to spend attention well.

The bottleneck. The human central nervous system, whatever its peripheral capacities, operates in thought as a fundamentally serial device. We attend to one thing at a time, or at most to a few; tasks that appear simultaneous are, on inspection, interleaved. The elementary operations of thought take times measured in tenths of seconds, and the store in which thought holds its working materials accommodates only a handful of items — some seven, or fewer, familiar units. These are not incidental engineering details. They are the constraints under which every deliberation, every judgment, and every act of comprehension is carried out. A theory of thinking that ignores them describes an ideal reasoner who does not exist.

Bounded rationality. Classical theories of choice imagined a decision-maker who surveys all alternatives, computes all consequences, and selects the best. No such being is found in nature or in administration. The alternatives are not given but must be searched for; the search consumes attention; and attention exhausted on one branch of the problem is unavailable to the others. Real deliberation therefore stops not when the best has been found but when something good enough has been found — when the alternative in hand satisfies the aspirations the chooser brought to the problem. To satisfice rather than to maximize is not a moral failing or a lapse of intellect. It is the rational response of a bounded mind to an unbounded world, and attention is the bound.

It follows that what a person notices determines, far more than logic does, what the person decides. Two administrators with identical values and identical files will decide differently if their attention is captured by different aspects of the situation — and it will be, for the situation always presents more aspects than can be attended. The genuinely scarce commodity in deliberation is not information, not intelligence, and not time in the calendar sense, but the capacity for attention; and whoever or whatever directs that capacity — habit, salience, interruption, the order of the morning’s mail — is exercising the real control over conduct.

Selection is not sovereign. It would be convenient to suppose that we allocate attention as a prudent householder allocates a budget, deliberately and from a standpoint outside the spending. Observation does not support the supposition. A large part of attention is captured rather than directed: the sudden sound, the bright motion, the urgent request arrive with a priority the agenda never voted them. The interrupt system is not a defect; an organism whose attention could not be seized by the unexpected would be eaten by it. But the consequence is that the allocation of attention is only partly governed by the goals of the moment, and any account of thinking — or any design of an institution — that assumes sovereign, voluntary selection will be wrong in practice.

Expertness as economy. How, under so narrow a bottleneck, is skilled performance possible at all? The answer that emerges from the study of experts — chess masters serve as the laboratory case — is that expertness is chiefly a vast economy of attention. The master does not examine more possibilities than the novice; the evidence is that he examines rather fewer. What the master possesses is a repertoire, built by years of practice, of familiar patterns — tens of thousands of them — each of which is recognized in a glance and each of which carries with it, already assembled, the information and the responses relevant to it. Recognition substitutes for search. The unit of attention, the chunk, has been enlarged by learning until a single act of noticing takes in what would cost the beginner a laborious inspection. Intuition, on this view, is not a separate and mysterious power set over against analysis; it is analysis frozen into habit — recognition, earned earlier and spent now.

The ant on the beach. An ant makes its way across a wind-sculpted beach, and its path is tortuous: it climbs, doubles back, skirts an obstacle too large to see over. A watcher who plotted the path might credit the ant with a complicated program. The complication, of course, is mostly in the beach. Viewed as a behaving system, a human being is quite simple as well; the apparent complexity of our behavior over time is largely a reflection of the complexity of the environment in which we find ourselves. The bearing on attention is direct. If the mind’s operations are simple and few, then the texture of a life of thought is set by what the environment offers to be noticed — and an environment can be designed. Change what is put before the mind, and in what order, and you have changed most of what there is to change.

The design of institutions. An organization is, among other things, a machine for the allocation of attention. Its division of labor is a division of attention: each position is an arrangement for noticing some things reliably at the price of not noticing others. Its hierarchy, its agenda, its channels and reports are devices that determine which matters will reach whose eyes, in what form, and in competition with what else. When such a system fails, the failure is seldom a shortage of information; it is nearly always information without an audience — a warning filed, a signal buried in noise, a report addressed to an office too busy to read it. The scarce resource was misallocated. The remedy is likewise a matter of design rather than exhortation: not to demand that officials be more attentive, which is to demand a larger budget from a fixed income, but to arrange matters so that what is important arrives where attention is available, and so that the technology of the institution — filing, summarizing, filtering, and now computing — conserves attention instead of consuming it. An information system that leaves its users knowing more and attending worse has subtracted value, however impressive its throughput.

Education as the training of attention. The same accounting applies to the school. A curriculum is a proposal for spending several thousand hours of a child’s attention, and it should be audited as such. Since recognition is the foundation of skill, the hours are well spent when they build the learner’s stock of patterns — which is to say, when they are spent in practice on material that the learner attends to, at a difficulty that holds attention without exhausting it. Length of exposure is a poor measure; what was actually attended is the whole of what was actually learned. And since the interrupt system does not mature early, the environment of learning matters more for children than for administrators: a classroom, like an office, can be designed so that the valuable is also the salient, or it can leave salience to whatever machinery of distraction has been placed within reach.

The account given here will strike some readers as a diminishment — attention reduced from a spiritual exercise to a scheduling problem. The reduction is the point, and it is not a diminishment. To say that attention is scarce is to say that every act of noticing is purchased by all the noticing it displaces, and that a mind, a family, an institution, or a civilization reveals its actual values not in what it praises but in what it attends to. Scarcity is what makes allocation meaningful, and allocation is what makes attention a subject for deliberate care. We cannot enlarge the budget much. We can spend it very much better, and the spending can be designed.