The loss of high-frequency function words is puzzling. Although they form part of core grammar—and, in some cases, have done so for thousands of years—some function words seem to just suddenly disappear. While the grammaticalization of content words into function words correlates with increase in usage, the loss of high-frequency function words cannot simply be explained by decrease in usage because of the indispensable function of these words. This article deals with the loss of the Germanic question particle, of the Germanic coordinating sentence conjunction, and of the Germanic negation particle. It describes their gradual decline as a result of language-specific interactions between phonology, syntax, and information structure: Function words occupy a fixed syntactic position, where they are systematically unstressed. Instead of being strengthened in their old position, they were lost. Instead of linking the loss of elements of core grammar to frequency-based semantic bleaching, it is attributed here to the interaction of linguistic subsystems. It is suggested that this development was unavoidable as the non-Proto-Indo-European structure of Germanic subsystems was eroding old Indo-European lexical material. Germanic prosody was not in harmony with the substance of the inherited Proto-Indo-European lexicon.*