One of the most intriguing problems in pragmatics and Speech Act theory is indirect speech acts like the classic example Can you pass the salt? Sadock (1972; 1974) named these utterances whimperatives for their ability to function both as questions and requests, and he considers the hearer’s primary interpretation of this construction to be its conventionalized meaning.  Using this definition of conventionalization, this paper explores whimperatives via their syntactic composition (modals, subjects, and verb phrases), and establishes clear criteria for what constitutes conventionalization.  Using this compositional description, the study investigates the accuracy of Sadock’s (1972) please-insertion test.  Sadock (1972; 1974; 2004), Levinson (1983), Stefanowitsch (2003), and others use the felicitous insertion of pre-verbal please as the litmus test for conventionalization in whimperatives.  For example, Can you (please) pass the salt? admits please grammatically and can therefore be considered a conventionalized request.  However, Should you *please pass the salt? does not felicitously admit please; therefore the utterance Should you pass the salt? is not a conventional request.  The present study challenges this test using clear criteria for conventionalization born in the syntactic units rather than an attempt to establish conventionalization rules from the construction as a whole.  To support this thesis, a survey was administered to 79 English speakers who judged whimperative constructions by context-appropriateness, i.e. participants read a whimperative and chose the most likely context (predetermined to represent either a question- or request-reading).  The whimperatives alternated their use of pre-verbal please and of verbs between dynamic and stative (with statives representing the non-conventionalized, e.g. Can you (??please) hear this song?).  The results showed definitively that the request/question interpretation patterned to the dynamic/stative dichotomy.  However, when please was present, all whimperatives were judged primarily as requests, proving not only inaccuracy with the please-insertion test, but suggesting that the coercive force of please presents an intra-sentential conflict that hearers must resolve while parsing, similar to what occurs in garden path sentences.  The key contribution of this paper is a clearly identified set of formal features that grant whimperatives the ability to function as indirect requests (including specific moods of modality, direct/indirect specification of an addressee, and the appropriate type of verb), with the larger argument that a compositional methodology should be present in any construction analysis.  Additionally, this paper raises questions for future research about coercion effects on sentence processing.