We contrast evidence of urban path dependence with efforts to analyze calibrated models of city sizes. Recent evidence of persistent city sizes following the obsolescence of historical advantages suggests that path dependence cannot be understood as the medium-run effect of legacy capital but...