Estudillo, A. J. . 1 , Casado, N. 2 & Bermudo, E. . 2
1 University of Edinburgh
2 Universidad de Málaga
According to Dehaene’s triple code model (1992; Dehaene & Cohen, 1995), subtractions and quotients are solved using an analogue-magnitude code (semantic), which is not precise. On the other hand, simple additions (5+2) and multiplications (4x3) are solved automatically, by memory retrieval using verbal representations. Other models of mathematical cognition (McCloskey, 1992; Cipolotti & Butterworth, 1995) consider that automaticity in different operations will depend of the frequency of solving. In this sense, the more frequent the more automatic. In our daily life, we solve small subtraction quite often, so it reasonable that this kind of subtraction may be solved automatically. We use a verification task (is the next operation correct? 4+2 =6) to shed light on this question. Participants had to determine whether the sum displayed were correct or not. Incorrect sums could be the correct result of a subtraction (interference condition, e.g., 7+3=4) or not (control e.g., 7+3=5). Problem-size was also manipulated, considering small subtraction as single digit subtraction, and large ones as a two digit number in the minuend and one digit in the subtrahend. Results showed longer responses in the incongruent condition in small subtractions but not in large ones. These results showed that small subtractions are solved automatically and they challenge Dehaene’s account about subtraction.