Another Visit to BE Supposed to from a Diachronic Constructionist Perspective

The BE supposed to string has meanings ranging from expressions of assumptions, hearsay, expectations, weak obligation, predestination and, in the negative, prohibition. Several problems have been discussed previously with the historical development of the deontic senses, most notably that it appears to break the deontic to epistemic unidirectionality constraint in grammaticalisation. Previous researchers have argued that the preceding accounts lacked support from the data and that SUPPOSE must have had an original intend sense when it was borrowed. They illustrated the paper with examples from the OED. However, the author shows how the OED interpretations are incorrect; none of the active or early passive examples cited actually have deontic senses. He presents a usage-based construction grammar account of the development, arguing that the deontic sense for BE supposed to arose when the form was re-interpreted as instantiating an existing passive deontic construction, licensed by analogy to the existing synonymy with BE expected to. This was possibly due to a process of metonymy, whereby part of the original meaning of the string, in particular usage-event contexts, becomes the whole meaning. The development may not in fact contravene unidirectionality constraints.