In generalized quantifier theories and many other contemporary approaches, it is standard to analyse the natural language determiner 'every' as the inclusion relation. Traditional observations and recent empirical results surrounding sentences with empty restrictors (e.g., 'Every French king lives in Paris') appear to contradict the predictions of the standard analysis. One fairly popular response is an analysis where 'every' yields undefinedness with an empty restrictor. Another response, which virtually no contemporary theorists advocate, is an analysis where 'every' yields falsity with an empty restrictor. This talk considers why there is often resistance to these alternative analyses. I will take the strongest justification for maintaining the standard analysis to be that the alternatives alter logical properties of the quantifier (positive strength, left downward monotonicity, etc.). However, I will argue that these altered logical properties do not appear to have any problematic consequences. Hence sufficient evidence for alternative analyses might motivate replacing the standard analysis. I will also consider broader concerns about how we can assign the correct semantics to natural language quantifier expressions.