When a stimulus set is vertically, and a response set is horizontally arranged, an above-to-right/below-to-left mapping has an advantage over the opposite mapping. This is termed an orthogonal stimulus-response compatibility (SRC) effect. According to the categorical coding hypothesis, a categorical spatial representation that codes the general categorical relation between an object position and a referent position has salient features, but a coordinate spatial representation that codes continuous, metric relations does not. The S-R translation is efficient when the salient stimulus and response alternatives (above and right in horizontal and vertical dimensions, respectively) correspond in an experimental task. So the orthogonal SRC effect would appear with the categorical coding of the stimuli. There is another possibility that the mental representation of the vertical lower-to-higher position is spatially organized from left to right as digits. We tested these hypotheses in the experiments with four vertically arranged stimuli (two above and two below) and two horizontally arranged responses (one in each side). The categorical coding hypothesis predicts that the orthogonal SRC effect would appear regardless of the number of the stimuli in each side if only the categorical codes are formed. If the spatial representation of the vertical dimension is directly corresponded to the horizontal dimension, the prediction is that the orthogonal SRC effect appears regardless of the manipulation of the stimulus arrangement. We obtained the orthogonal SRC effect when we facilitated the categorical coding by serving the reference point for a sharp distinction between each side and by grouping of the stimuli in each side. However, the orthogonal SRC effect disappeared when the reference point was removed and four stimuli separated by equal spaces. The orthogonal SRC effect appeared only when the stimuli were coded categorically. The categorical coding hypothesis was supported.