Reporting with clausal embedding and without: Another look at the Pirahã controversy

This chapter explores the relation between the syntax of clausal embedding and
the ability to represent what others are saying, thinking, dreaming etc.. I’m using the Pirahã controversy as a lens through which to study this relationship because, supposedly, the Pirahã language has no clausal embedding and hence no analogue of English indirect discourse (Katy said/thought/dreamed that she was rich). I frst show how hearsay evidentiality and direct quotation, both of which are attested in Pirahã, difer semantically from each other and from indirect discourse. However, together, these two arguably embedding-free report strategies could cover two of the most common uses of indirect discourse in English, viz. (i) efcient communication that keeps track of speaker’s evidential sources through a not-at-issue information channel, and (ii) vivid description of speech and thought in narratives. I also argue that reporting in general is best understood as a discourse phenomenon, only optionally encoded in the grammar. Spelling this out in a formally explicit and independently motivated general model of discourse structure and coherence relations (including a non-veridical relation of Attribution) we actually derive Dan Everett’s own diagnosis of the situation, viz. that “there can be recursive discourses in the absence of recursive sentences.