Step 7 of 11
Dispositions: what you bring to the conversation
Five durable tendencies that make wise engagement easier or harder. They are not skills, and no single conversation installs one.
A disposition is not what you can do. It is what you are inclined to do when nothing is forcing you. Two people with identical prompting skill will use the same model differently if one is inclined to check and the other is inclined to accept, and that inclination is the thing that persists after the course ends.
This is why dispositions sit where they do in the chain. Comparing two people at one moment, dispositions come first: what someone is already inclined to do shapes which principles they can hold onto and which behaviors they reach for. Following one person over time, the arrow reverses, and practising a behavior under a principle is how a disposition grows. Neither direction is the whole story, which is what the Becoming loop is for.
What follows is the prior work under each of the five. As on the theory step, each entry separates what the source established from what it licenses here.
The five, and what each rests on
None of these five was invented for this model. Each is a construct with its own measurement history, borrowed for a specific job and no more.
Supports Intellectual Humility
Intellectual humility as a measurable disposition
What it established. Intellectual humility — recognising that one's beliefs may be wrong — behaves as a measurable individual difference, distinct from general humility and from low confidence, and predicts how people handle evidence that contradicts them.
What it licenses here. Treating humility as a disposition that varies between people and can be measured, rather than as advice. It also supplies the boundary: intellectual humility is not diffidence, so the model must not score low confidence as if it were humility.
Leary, M. R., Diebels, K. J., Davisson, E. K., & Jongman-Sereno, K. P. (2017). Cognitive and interpersonal features of intellectual humility. Personality and Social Psychology Bulletin, 43(6), 793–813. Source
Supports Internal Locus of Evaluation
Where a person locates the standard
What it established. People differ stably in whether they attribute outcomes to their own action or to forces outside them, and that generalized expectancy shapes how they respond to feedback.
What it licenses here. The idea that where a person locates the standard of judgment is a durable trait-like tendency rather than a momentary mood. Rotter is about control over outcomes, not evaluation of quality; the model borrows the internal-versus-external structure, not the construct itself, and the disposition here is specifically about who decides whether an answer is good.
Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs: General and Applied, 80(1), 1–28. Source
Supports Investment in Process
Effortful retrieval beats fluent restudy
What it established. Retrieving material is more effective for long-term retention than restudying it, even though restudying feels more productive at the time and learners predict the opposite.
What it licenses here. The core of this disposition, and the reason it is hard: the feeling of learning and the fact of learning come apart, and AI is very good at producing the feeling. It does not license claiming that any particular AI workflow is a retrieval practice; that is an empirical question this model has not answered.
Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255. Source
Supports Tolerance for Ambiguity
Ambiguity as a stable personality variable
What it established. The tendency to perceive ambiguous situations as threatening rather than desirable is a stable individual difference, measurable and distinct from anxiety in general.
What it licenses here. Why some people close a question down at the first plausible answer and others keep it open. It is one of the oldest constructs in the set and its measures have been criticised since; the model uses the concept, and does not adopt Budner's scale.
Budner, S. (1962). Intolerance of ambiguity as a personality variable. Journal of Personality, 30(1), 29–50. Source
Supports Tinkering Disposition
Tinkering as a legitimate epistemology
What it established. Working by negotiation with materials — trying, watching, adjusting — is a coherent way of knowing rather than a deficient version of planning from the top down, and it is systematically undervalued in computing.
What it licenses here. Treating tinkering as a disposition worth cultivating rather than as impatience. It matters here because iterative and single-shot users extract very different value from the same model, and the difference is not skill.
Turkle, S., & Papert, S. (1990). Epistemological pluralism: Styles and voices within the computer culture. Signs: Journal of Women in Culture and Society, 16(1), 128–157. Source
Where would you put yourself?
A self-rating, with the caveat the next panel makes: this is your account of yourself, which is worth having and is not evidence about what you do.
Where are you right now?
Pick the stage that best describes your current AI engagement. Most people overrate themselves by one stage on first pass.
These stages are buildable. The next step is deliberate practice in the modes you don’t currently default to. The trade book walks through that progression mode by mode; the app provides guided workflows that target the modes specific to your stage.
Dispositions explain why two people behave differently. They do not tell either person what to do next. That is what the principles are for: the standing policies you hold to while the conversation is actually running.
Quick check
How is the dispositions layer measured?