Every once in a while I come across work that conflates iconicity and indexicality, or lumps them together under a broad label of motivation (often in opposition to ‘arbitrariness’). Even if I tend to advocate for treating terminology lightly, I think there are many cases where it does pay off to maintain this distinction, and conflating it comes at a cost.

Not distinguishing iconicity and indexicality means losing the ability to explain how and why some linguistic resources differ in markedness & morphosyntactic behaviour, as I point out for the analogical issue of ideophones vs interjections here. A related case is transparent compounds, which naïve raters (under some instructions) also rate as highly iconic, yet for which it helps to be able to articulate how they differ from the kind of form-meaning resemblance usually targeted by the technical term iconicity.

There are also deeper evolutionary implications you’d lose sight of without the distinction. If an ancestral pain vocalization underlies interjections like ‘ow’, that makes for a different causal story than cross-linguistic similarities that can be ascribed to (possibly convergent) iconic mappings. So to explain why today’s languages are the way they are, a distinction like this comes in useful.

But for my money, the most interesting questions lie in where iconic vs indexical motivations overlap and where they diverge, and how this influences learning, processing, and cultural evolution. We can’t see those questions if we lump the notions together, nor when we dichotomize them.

This short post originated as a twitter thread.