Abstract: The spatial representation of numbers has been intensively investigated in the field of number processing mechanism. To better explore the spatial representation, the mental number line hypothesis has been proposed, which gets support from research on distance effect and SNARC effect. Although much has been done to investigate the spatial representation of positive numbers, studies focused on negative numbers are rather scarce. Moreover, the results on negative numbers are essentially inconsistent, in that there are two major opposite hypotheses on the spatial representation: the phylogenetic hypothesis and the ontogenetic hypothesis. The phylogenetic hypothesis suggests that negative numbers are mapped onto the mental number line according to their absolute numerical value: the small absolute value on the left, and the large absolute value on the right. Conversely, the ontogenetic hypothesis asserts that negative numbers are mapped onto the mental number line according to the numerical value: the small value on the left, and the large value on the right. Further analysis of the two hypotheses reveals that two different paradigms were adopted in previous research, namely the odd-even judgment paradigm and the magnitude judgment paradigm, which might be explicable for the contrasting conclusions. In addition, both paradigms are essentially improper for the research on the representation of negative numbers. Considering the two existing problems, this study adopted the speeded magnitude classification paradigm to further investigate the spatial representation of negative number processing.
Two experiments were conducted with the numbers ranging from -9 to 9, but excluding 0 as materials. On each trial, a rectangle-frame was firstly displayed in the center of the screen as fixation, then a number was displayed horizontally in the center of fixation. The participant should classify the magnitude as quickly and accurately as possible. The first experiment, which contained half of positive numbers judgment and half of negative numbers judgment, was aimed at investigating the difference between two types of number processing and the direction of spatial representation. The second experiment was aimed at further investigating the direction of negative numbers’ spatial representation through a mixed display method, in which both positive numbers and negative numbers were displayed serially but only negative numbers were targets of magnitude judgment. We varied the consistency between the response key and digit magnitude as the independent variable in Experiment 1 and 2. Each experiment included eighteen participants. ANOVA analyses and regression analyses were performed in both experiments.
The data analyses showed that SNARC effect appeared in both experiments. Specifically, participants responded faster as to the left-side decision for negative numbers with small absolute value and faster as to the right-side decisions for negative numbers with large absolute value.
The pattern of decision latencies suggested that negative numbers are mapped onto the mental number line according to their absolute value instead of numerical value, thus supporting the phylogenetic hypothesis

Key words: negative number, mental number line, SNARC effect, phylogenetic hypothesis, ontogenetic hypothesis