ICONIC - The Emergence and Stabilisation of Iconicity in Human Speech

Linguists traditionally believe that a word's form doesn't relate to its meaning. For example, the Norwegian word “tre” or the German word “Baum” don’t sound like a tree. Iconicity, which refers to the perceived resemblance between the form and meaning of a linguistic form, has been thought as only marginal in human language. For instance, the word “boom” bears resemblance to the loud, explosive sound to which it refers, and the word “teeny-weeny” conveys a sense of smallness through their high-front vowels. Researchers believed that these forms of iconicity fade away due to the way language changes over time, leading to a purely arbitrary vocabulary. However, this view clashes with evidence showing that iconicity can be found in all aspects of modern languages and that people continue to have intuitions about how the form of a word relates to its meaning. The ICONIC project aims to resolve this contradiction by demonstrating that iconicity can emerge and stabilise over time, rather than disappear. ICONIC sets out to shed light on how the vocabularies of languages evolve over time by providing new ways to study this evolution in the laboratory. Ultimately ICONIC will help us better understand why human language looks the way it does.

Linguists often assume that the form of a word does not relate to its meaning. For example, nothing about the sound sequence of the Norwegian word tre is tree-like. Iconicity on the other hand, the perceived resemblance between the form of a word and its meaning, has been largely confined to marginal language phenomena. It is marginal, so linguist have argued, because it is assumed to erode over time by diachronic mechanisms, inevitably creating a purely non-iconic lexicon. This view, however, stands in stark contrast with two observations: First, iconicity can be widely found within and across the world’s languages. Second, language user’s have robust intuitions about form-meaning correspondences in their own and other languages. The present project will resolve this contradiction and will experimentally demonstrate that iconicity, instead of eroding over time, can emerge and stabilise within a sound system across repeated transmissions. Using an innovative combination of the iterated learning paradigm and state-of-the-art acoustic analyses, iconic will quantify when and how iconicity evolves. Acoustic recordings of individual speech transmission outcomes allow us to measure which (combinations of) acoustic features speakers use to express size-based iconic relationships (e.g. large vs. small referents), including features such as vowel quality, but also notoriously under researched features of the speech signal like voice quality and pitch. After defining the acoustic profile of size-based iconicity, iconic will explore and quantify the (in)stability of iconic form-meaning pairs across generations in order to, for the first time, experimentally probe the hypothesis that iconic form-meaning pairs are robust to change. This empirical approach will not only shed new light on the evolution of linguistic signs, it will also serve as the first proof of concept of how we can harness experimental innovations to explore questions related to sound change.