Abstract
Casasanto and Boroditsky (2008) published an influential paper arguing for non-linguistic evidence of using space to think about time (asymmetrical dependence). In their task, a spatio-temporal event, such as a moving dot or growing line was used to present temporal durations and spatial extents. The two properties were uncorrelated across trials (though perfectly correlated, moment-to-moment within each trial). Across six experiments, they reported an asymmetrical effect of spatial extent on participants’ reproductions of temporal duration. Their analysis of the effect of space on time involved collapsing their time estimation data by distance and reporting evidence of a trend for longer distances to produce longer times, but not vice-versa. The inappropriateness of this style of analysis is evident in their report of R2 values for time and distance predictors that summed to greater than 1. A series of carefully conducted replications revealed contamination of time judgments by distance in only a subset of our participants, resulting in an overall “reliable” trend when collapsed across participants (probably due to inattention rather than metaphor). When spatial parameters were modified so as to make the distances slightly less saliently different, symmetrical effects of time on distance were observed using these same measures, though none of these effects were reliable using more traditional mixed-effects models. Humans can perceive the magnitudes of short intervals of time perfectly well without recourse to perceived distance, so the observed contamination is more likely due to magnitude representation, and unlikely due to metaphor. The space/time asymmetries reported by Casasanto and Boroditsy were (1) likely due to inattention/magnitude interference, (2) may have been due to a subset of participants, and (3) were consistently produced in their experiments using exactly the same spatial and temporal parameters over and over. Small changes to those parameters are sufficient to remove or reverse the asymmetry.