Why we still don't need/want variables: Two SALTy case studies

Abstract

This paper defends the claim that variables and assignment functions are not needed as part of the semantic machinery against two apparent challenges. The relevant domains at first glance appear to show the need for variable names - but only if the variable-ful theory is supplemented with a stipulation preventing 'meaningless coindexation'. I argue that not only are both domains amenable to variable free analyses, but that indeed the variable free analyses have advnatages over the analyses using variables.