Part I · Chapters 1–3

Our Mental Machinery

Part I claims that the binding constraint on analysis sits inside the analyst: perception constructs rather than records, memory decides what is available to think with, and you have almost no conscious access to either process while it runs.

Thinking about thinking

Efforts to improve analysis usually target the product — writing, formats, the relationship with the consumer — or the input, by collecting more. Almost none target the reasoning between, which is largely invisible to the person doing it:

“What appears spontaneously in consciousness is the result of thinking, not the process of thinking.”

His frame is Herbert Simon’s bounded rationality (p. 2): unable to hold the world’s complexity, the mind builds a simplified model and reasons inside it. You behave rationally within that model; whether it fits the world is a separate question, and not one the model can answer. Heuer uses mental model and mind-set more or less interchangeably (p. 4). The consequence is uncomfortable for a profession built on expertise: when a situation turns, the specialist has the most to unlearn. Some German specialists had to be prodded by generalist supervisors to accept the significance of the changes running toward reunification (p. 5).

Perception is an inference, not a recording

Chapter 2 is the one most people remember, because it makes its case with pictures you fail in real time. Its three principles are blunt enough to serve as headings. We tend to perceive what we expect to perceive (p. 8) — shown with playing cards gimmicked so spades were red and hearts black, which subjects identified more slowly and less accurately even after being told the anomaly existed. Mind-sets tend to be quick to form but resistant to change (p. 10). New information is assimilated to existing images (p. 11), which is why gradual change goes unnoticed and why a newcomer sometimes sees what a ten-year veteran has stopped seeing.

A fourth finding reads operationally: initial exposure to blurred or ambiguous material interferes with accurate perception even after better information arrives (p. 13). Subjects who first saw a badly out-of-focus image had more trouble identifying it once it sharpened. Early, thin reporting sets an expectation that later, better reporting has to overcome.

Memory decides what you can think with

Chapter 3 separates sensory information storage, short-term memory and long-term memory, and locates the limit in the traffic between them. Long-term memory is effectively unbounded; the difficulty is getting things in and retrieving them (p. 17). The pinch is in working memory, which Chapter 7 puts at seven items, plus or minus two (p. 85).

Heuer’s claim is that the difference between stronger and weaker analysts is attributable in large measure to how data and experience are organized in long-term memory (p. 17) — expertise as a retrieval structure, not a stock of facts. Once well worn, that structure is what Chapter 6 calls a mental rut.