The use of
according to + NP has rarely been the topic of any specialized research in relation to
English evidentiality, although it would probably figure among the most frequent types of reportative evidentials found in written
texts. One of the problems often associated with reportatives has related to the existence of the Reportative Exception (see, e.g.
AnderBois 2014), referring to the fact that the speaker may not always subjectively
endorse the proposition conveyed with the support of the evidential phrase. The present study reviews the history of
according to + NP from Middle English onwards, after which it began to develop evidential functions, and
shows how the tendency to reject the truth of the content of the proposition marked by
according to + NP arose in
specific contexts containing alternative information sources, comparison, or adversative clauses. It was shortly after the
diachronic appearance of
according to + NP in such contexts that the more periphrastic form,
in
accordance with + NP, began to renovate/renew the earlier, non-evidential meanings of
according to +
NP. The present study also attributes the development of
according to + NP to a process of co-optation (e.g.
Heine 2013) rather than grammaticalization.