An argument against the crush origin of pseudotachylytes based on the analysis of clast-size distribution