Structural priming implies that speakers/listeners unknowingly re-use syntactic structure over subsequent utterances. Previous research found that structural priming is reliably enhanced when lexical content is repeated (lexical boost effect). A widely held assumption is that structure-licensing heads enjoy a privileged role in lexically boosting structural priming. The present comprehension-to-production priming experiments investigated whether head-constituents (verbs) versus non-head constituents (argument nouns) contribute differently to boosting ditransitive structure priming in English. Experiment 1 showed that lexical boosts from repeated agent or recipient nouns (and to a lesser extent, repeated theme nouns) were comparable to those from repeated verbs. Experiments 2 and 3 found that increasing numbers of content words shared between primes and targets led to increasing magnitudes of structural priming (again, with no ‘special’ contribution of verb-repetition). We conclude that lexical boost effects are not diagnostic of lexically-specific syntactic representations, even though such representations are supported by other types of evidence.