We commend the target paper (henceforth S&N) for bringing reported speech to attention in the typological space, and for making a number of highly pertinent observations. We agree that reported speech deserves to be seen as a sui generis domain or topic, well deserving of typological attention and not reducible to an intersection of other phenomena. We would prefer to characterise it as a semantic or functional domain, rather than as a "syntactic" domain, given that key aspects of S&N's

more »

... on hinge on semantic notions, but this is not our main concern in this commentary. Instead, we would like to take issue with the target paper on more important theoretical and methodological matters. The most significant concerns S&N's reliance on complex, poorly-defined, English-bound terms, including both technical terms such as semiotic, 'demonstratedness', epistemic, modality, and representation, and ordinary, but equally Englishbound, words such as report(ed), message, discourse, and utterance. In this commentary we aim to demonstrate, so far as possible in the space available, that the use of such opaque and/or English-bound terminology is unnecessary and to outline an alternate approach to the same phenomena. The first step, in Section 1, is to show that the most explicit and unambiguous mode of reported speech (corresponding closely to the traditional idea of "direct speech") can be described in simple, cross-translatable words. This, we argue, is a universal of human languages and provides a universal prototype or conceptual anchor for "reported speech" broadly. In Section 2, we outline a strategy for characterising the various modes of "non direct speech", such as, inter alia, quotative particles, prosodic cueing, and subordinate constructions. Rather than trying to bring everything that may be counted as "reported speech" under a single, extremely abstract characterisation, we favour an approach that analyses these diverse constructions one at a time, so to speak, linking them all Cliff Goddard, [klɪf ˈgɔdɐ:d]