The history of mentalities (histoire des mentalités) is the Annales School's most ambitious methodological extension — the attempt to reconstruct not what historical actors believed but what they were equipped to believe. Mentalities are the inherited cognitive frameworks — the categories, the analogies, the unquestioned assumptions — that operate beneath explicit belief, shaping what can be thought, what can be said, what can be perceived as a problem at all. Developed by Febvre in his study of sixteenth-century unbelief and extended by Le Roy Ladurie, Duby, and Ariès, the concept applies with special force to the AI transition, which is reshaping mentalities faster than explicit beliefs can keep pace.
Mentalities operate at the deepest timescale — the longue durée of cognitive inheritance. A mentality can persist for centuries, transmitted through language, institutions, child-rearing, and daily practice. Individuals within a civilization typically cannot see their own mentalities; these are, in Edo Segal's image, the water the fish breathes. Mentalities become visible only from outside — through historical comparison, cultural encounter, or the shock of their sudden failure.
The AI transition is producing precisely such a shock. The mentalities of professional expertise, craft mastery, and linear career trajectory — deeply inherited from nineteenth-century industrial modernity — are failing visibly. Workers discover that the categories by which they understood their own work no longer correspond to what the market values. The shock is not that the categories have changed; it is that they had been visible as categories at all.
Febvre's methodological insight was that mentalities cannot be recovered through the straightforward reading of sources. The sources themselves are products of the mentality; they reveal it only when read against the grain, attending to what they presuppose rather than what they state. Applied to AI, this means that the most consequential evidence about how the technology is reshaping human cognition lies not in what users say about AI but in how they reach for it, what they stop doing without noticing, and which questions they no longer form.
The framework also names a specific danger. Mentalities change slowly under normal conditions — faster than events, slower than technologies. When technology changes at conjunctural speed (AI), mentalities lag. The lag produces the characteristic confusion of the present: people operating with nineteenth-century mentalities in a twenty-first-century technical environment, unable to see why their inherited frameworks no longer produce the outcomes they once did.
Developed by Febvre in the 1940s (especially The Problem of Unbelief in the Sixteenth Century, 1942), extended by the third Annales generation (Le Roy Ladurie, Duby, Ariès, Le Goff) in the 1960s and 1970s into the history of death, childhood, fear, and religion.
Beneath explicit belief. Mentalities are the frameworks that make specific beliefs possible, not the beliefs themselves.
Invisible from within. Members of a civilization typically cannot perceive the mentalities they inhabit; visibility requires outside perspective or sudden failure.
Longue durée cognition. Mentalities persist for centuries, transmitted through language, institutions, and daily practice.
The AI test case. Current mentalities of work, expertise, and cognition are visibly failing under technological pressure, exposing what was previously invisible.
Critics argued that mentalities were too diffuse to be operationalized rigorously, too elusive to be falsified, and too prone to anachronistic projection. Defenders responded that the concept names something real even if it resists precision: the background conditions of thought that structure foreground beliefs.