Constitutional Amendment Versus Constitutional Replacement: An Empirical Comparison

Abstract

The field of comparative constitutional law has been transformed in recent years by groundbreaking research – often of a quantitative and empirical variety – on the creation, content, and longevity of constitutions. A common and convenient approach is simply to look to whether these changes are formally labeled as ‘amendments’ or ‘replacements’. The amendment–replacement distinction is thus central to their analysis: one cannot study the effect of constitutional amendment on the likelihood of constitutional replacement unless one can distinguish between amended constitutions and replacement constitutions. Measurements of semantic similarity, by contrast, can differentiate at the linguistic level between different ways of addressing the same topic or even reaching the same outcome. Natural language-processing techniques excel at quantifying semantic similarity between documents and are thus well suited to exploring the difference in magnitude between so-called ‘amendments’ and ‘replacements’. Flexible judicial interpretation and constitutional amendment might be thought to function as substitutes for one another.