martin_seligman's picture
Professor and Director, Positive Psychology Center, University of Pennsylvania; Author, Flourish

If we could teach intuition, people would be smarter.

Most of real world “intelligent” performance is based on intuition, not on reasoning. The expert surgeon just “knows” where to cut. The experienced farmer just “knows” that it is going to rain. The expert firefighter just “knows” that the roof is about to collapse. The judge just “knows” the defendant is lying. These finely honed intuitions are fast, unconscious, multidimensional, inarticulate, and are made confidently. What separates intelligent from mechanically stupid action is wrapped up in this mysterious process. If we could only teach intuition, we could raise human intelligence substantially.

I believe that the teaching of intuition is on the horizon by computationally driven simulation.

Intuition is a species of recognition, formally akin to the way we recognize that a table is a table. We are now close to understanding how natural classes are recognized. Consider the universe of objects all people agree are tables. There are a great many features of tables that are potentially relevant (but neither necessary nor sufficient singly or jointly) to being a table: e.g., flatness of the surface, number of legs, capacity for supporting other objects, function, compatibility with chairs, etc., etc. Each of these features can be assigned some value, which could either be binary (present vs. absent) or continuous. Different instances of tables will have different values along several of the dimensions, e.g., some, like dining room tables are flat, whereas others, like pool tables, have pockets. This means that the process of categorization is stochastic in nature. Upon observing a new object one can decide whether it is a table by comparing its features with the features of stored tables in memory. If the sum of its similarity to all of the tables in memory is higher than the sum of its similarity to other objects (e.g., chairs, animals, etc.) then one “knows” that it too is a table.

Now consider an “eagle” lieutenant recognizing a likely ambush. Here, is what this eagle has stored in her brain. She has a list of the dimensions relevant to an ambush site versus a non-ambush site. She has values along each of these dimensions for each of the ambush and non-ambush sites that she has experienced or learned about. She has a mental model which assigns weights to each of these basic dimensions or features (and to higher order features, such as the interaction between two dimensions). Based on past experiences with similar sets of features and knowledge of the outcomes of those feature sets, she can predict the outcome of the present feature set and based on her predictive model, choose how to respond to the possibility of this being an ambush.

This strongly implies that intuition is teachable, perhaps massively teachable. One way is brute force: simple repeated experience with forced choice seems to build intuition, and chicken-sexing is an example of such brute force. Professional Japanese chicken-sexers can tell male from female chicks at a glance, but they cannot articulate how they do it. With many forced choice trials with feedback, naïve people can be trained to very high accuracy and they too cannot report how they do it.

A better way is virtual simulation. A sufficient number of simulations, with the right variations, to allow a buildup of the mental model will result in a commander or surgeon who when it happens in real life has “seen it before,” will recognize it, and take the life saving action at zero cost in blood. It would be a waste of training to simulate obvious decisions in which most commanders or surgeons would get it right without training. Computational modeling of the future can derive a decision contour, along which “close calls” occur. These are the scalpel-edge cases that yield the slowest response times and are most prone to error. One could also systematically morph material along the decision contour and thereby over-represent cases near the boundary.

By so simulating close decisions in almost every human domain, vastly better intuition becomes teachable. Hence many more intelligent surgeons, judges, commanders, investors, and scientists.