Preferred sound groups of vocal iconicity reflect evolutionary mechanisms of sound stability and first language acquisition: evidence from Eurasia

Abstract

In speech, the connection between sounds and word meanings is mostly arbitrary. However, among basic concepts of the vocabulary, several words can be shown to exhibit some degree of form-meaning resemblance, a feature labelled vocal iconicity. Vocal iconicity plays a role in first language acquisition and was likely prominent also in pre-historic language. However, an unsolved question is how vocal iconicity survives sound evolution, which is assumed to be inevitable and 'blind' to the meaning of words. We analyse the evolution of sound groups on 1016 basic vocabulary concepts in 107 Eurasian languages, building on automated homologue clustering and sound sequence alignment to infer relative stability of sound groups over time. We correlate this result with the occurrence of sound groups in iconic vocabulary, measured on a cross-linguistic dataset of 344 concepts across single-language samples from 245 families. We find that the sound stability of the Eurasian set correlates with iconic occurrence in the global set. Further, we find that sound stability and iconic occurrence of consonants are connected to acquisition order in the first language, indicating that children acquiring language play a role in maintaining vocal iconicity over time. This article is part of the theme issue 'Reconstructing prehistoric languages'.

Keywords: first language acquisition; language evolution; phonology; sound evolution; typology; vocal iconicity.

Figure 1.

Representation of our model of emergence of vocal iconicity in relation to articulatory mechanisms of speech production, sound stability and sound preference in first language acquisition.

Figure 2.

Barplot demonstrating the stability rates of sound groups, divided by ‘stable’ and ‘shift in group’ (=stability rate) versus ‘shift out of group’ and ‘loss or gain’ (=instability rate), organized from most stable sound group (top) to most unstable (down) (see electronic supplementary material, S1).

Figure 3.

(a) Density plot (i) and scatter plot with regression line (ii) with a corresponding density plot (iii), adapted to a logarithmic scale, comparing sound stability (x) and iconic value (y), separating the distribution of vowels (yellow) and consonants (grey) (see electronic supplementary material, S4, S5). (b) Density scatterplots, adapted to a logarithmic scale, comparing sound stability (x) and iconic value (y), separating the distribution of ‘earlier’ (red) and ‘later’ (blue) in first language acquisition (relative distinctions of acquisition order of features, following the babbling period), given by the feature classes ‘manner of articulation’ (i) and ‘place of articulation’ (ii) (see electronic supplementary material S4-S6). (Online version in colour.)