“Compacted” procedures for adults’ simple addition: A review and critique of the evidence

Abstract

We review recent empirical findings and arguments proffered as evidence that educated adults solve elementary addition problems (3 + 2, 4 + 1) using so-called compacted procedures (e.g., unconscious, automatic counting); a conclusion that could have significant pedagogical implications. We begin with the large-sample experiment reported by Uittenhove, Thevenot and Barrouillet (2016, Cognition, 146, 289–303), which tested 90 adults on the 81 single-digit addition problems from 1 + 1 to 9 + 9. They identified the 12 very-small addition problems with different operands both ≤ 4 (e.g., 4 + 3) as a distinct subgroup of problems solved by unconscious, automatic counting: These items yielded a near-perfectly linear increase in answer response time (RT) yoked to the sum of the operands. Using the data reported in the article, however, we show that there are clear violations of the sum-counting model’s predictions among the very-small addition problems, and that there is no real RT boundary associated with addends ≤4. Furthermore, we show that a well-known associative retrieval model of addition facts—the network interference theory (Campbell, 1995)—predicts the results observed for these problems with high precision. We also review the other types of evidence adduced for the compacted procedure theory of simple addition and conclude that these findings are unconvincing in their own right and only distantly consistent with automatic counting. We conclude that the cumulative evidence for fast compacted procedures for adults’ simple addition does not justify revision of the long-standing assumption that direct memory retrieval is ultimately the most efficient process of simple addition for nonzero problems, let alone sufficient to recommend significant changes to basic addition pedagogy.