Introduction We investigated the emergence of indexicality, whereby linguistic features acquire social or contextual associations. Any linguistic feature can acquire “constellations” of indexical meanings, with speaker-group associations referred to as first-order indices and higher-order indices being derived from associations with those speakers [1-2]. Much work has been conducted in this area, but little experimental work has focused on the emergence of indexicality [3-4]. We present miniature artificial-language experiments designed to break ground on this question. Results show ready formation of first-order indexicality, with higher-order indexicality emerging in extension to new speaker groups, modulated by the social importance of the indexed trait.

Method 300 participants learned a miniature alien language with ten nouns and two plural endings (-gok and -dem). Two alien species (Nulus and Gilis) wearing two different ceremonial outfits were presented as users of this language. The experiment consisted of four phases: Familiarization (to familiarize participants with the aliens and the outfits and ensure equal attendance to both); Training (to expose participants to the language); Memory test (to maintain attendance to both species and outfits); and Association test (in which participants paired plural words with aliens and vice versa).

Experiment 1 investigated to what extent participants acquired an equal association of plural variants with both species and outfits. In training, each outfit and each plural ending was paired 100% reliably with one alien species. In the Association test, we manipulated whether the outfit–alien correspondence was the same as in training or flipped, allowing us to determine if participants associated plural endings with aliens, outfits, or both. Results from mixed-effects logistic regression models revealed that participants strongly associated suffixes with the aliens who used them, regardless of outfit. However, the interaction effect suggested a slight tendency to make secondary associations with outfit, which we hypothesized might be generalizable to new aliens.

Experiment 2 tested this hypothesis by introducing new alien species in the test phase. Participants again strongly associated suffixes with established aliens but now extended associations to new aliens via clothing. We hypothesized that associations would be strengthened by attaching practical social importance to the outfits.

Experiment 3 investigated this question. In the familiarization phase, we varied instructions to manipulate participants’ beliefs about outfits being socially important (social condition), aesthetically important (nonsocial condition), or neither (baseline). Compared to the baseline condition, the tendency for participants to associate different variants with different species became more pronounced in the nonsocial and social conditions. Participants also treated established aliens and new aliens differently, with greater variability for the latter.

Conclusion Participants acquired strong associations between linguistic variants and speaker groups, with only slight secondary association with contingent traits of those speakers. However, they generalized these secondary associations to new speakers, and this effect was modulated by the perceived practical importance of the trait in question. Taken together these results shed new experimental light on the emergence of sociolinguistic indexicality.