The Adult Math Project

In the AI Story

The study's design was deceptively simple but methodologically innovative. Rather than relying on laboratory experiments or self-report, Lave and her team conducted ethnographic observation of shoppers in their actual grocery stores, documenting the mathematical reasoning embedded in real shopping decisions. They then administered formal tests containing the same mathematical operations — comparing unit prices, calculating discounts, determining best values — stripped of the contextual features that made the in-store mathematics meaningful.

The findings challenged a foundational assumption of Western educational theory: the assumption of transfer. Transfer is the idea that knowledge acquired in one context — a classroom, a textbook, a training program — can be applied directly and without significant transformation to another context. Transfer is the foundational justification for formal education. If knowledge does not transfer, the classroom has no obvious relationship to the world it claims to prepare students for, and the billions of hours spent teaching abstract mathematics become difficult to defend.

What Lave observed was that the shoppers were not applying school-learned rules to practical settings. They were performing a different kind of mathematics entirely — improvisational, contextually driven, shaped by the physical arrangement of products on shelves, the specific goals of that day's shopping trip, and the embodied experience of hundreds of previous trips. A shopper comparing two bottles of ketchup did not set up a formal proportion and solve for unit price. She picked up both bottles, felt their weight, estimated — not as a fallback from real mathematics but as a sophisticated cognitive operation embedded in a specific physical and social context.

For the AI era, the Adult Math Project reframes the central question of what large language models produce. The models are, in a precise sense, the purest expression of decontextualized knowledge ever constructed — the shopper taking the paper test, scaled to civilizational dimensions. Their outputs are often correct. The question Lave's data forces is whether correct outputs in the absence of context are equivalent to the contextually situated understanding that situated practice produces. The data says no. The question is what that means for a professional culture that increasingly relies on decontextualized outputs.

Origin

The study was conducted in the early 1980s and results published in journal articles beginning in 1984, with fuller treatment in Cognition in Practice (1988). The work built on earlier ethnographic studies Lave had conducted of Vai and Gola tailoring apprentices in Liberia in the early 1970s, which had produced similar findings about the situated character of mathematical reasoning in non-school settings.

Key Ideas

Identical mathematics, different results. The 98/59 gap cannot be explained by differences in ability, education, or training. The variable that changed was context.

Practical arithmetic is sophisticated cognition. What looked like rough estimation was, in fact, a complex cognitive operation drawing on sensory data, spatial reasoning, and contextual goals — often more accurate than formal calculation.

The location of reliability matters. In-context performance was not merely faster than out-of-context performance. It was substantially more reliable.

Schooling does not transfer well. The shoppers' school mathematics was largely irrelevant to their supermarket performance, which drew on different cognitive resources altogether.

Debates & Critiques

The study has been criticized on methodological grounds — sample size, generalizability, the possibility that the testing environment produced artificial anxiety. Later replications and extensions have largely confirmed the core finding, though the magnitude of the gap varies across populations and domains. What remains contested is the interpretation: whether the gap reveals a fundamental property of cognition (Lave's position) or a remediable feature of how mathematics is taught (the position of more optimistic transfer theorists).