A central part of communication through human language is the extraction of meaning from linguistic forms. This is harder to achieve in cases of mismatch, such as when the mapping from form to meaning is not a one-to-one mapping. Looking at how this lack of transparency plays out in different domains (i.e., phonology, syntax, semantics) has the potential to identify common patterns in the nature of such mismatches, as well as in how they are engaged with by speakers, and ultimately resolved by hearers. We investigate mismatches of different guises—namely one-to-many, many-to-one, and unpredicted one-to-one. Also integral to this line of inquiry are cases of disambiguation, where linguistic expressions do not exhibit mismatch despite their similarity/relatedness to those that do. That further illuminates the grammatical ingredients that give rise to mismatch.