Social biases can lead to less communicatively efficient languages

ABSTRACT

Language is subject to a variety of pressures. Recent work has documented that many aspects of language structure have properties that appear to be shaped by biases for the efficient communication of semantic meaning. Other work has investigated the role of social pressures, whereby linguistic variants can acquire positive or negative evaluation based on who is perceived to be using them. While the influence of these two sets of biases on language change has been well documented, they have typically been treated separately, in distinct lines of research. We used a miniature language paradigm to test how these biases interact in language change. Specifically, we asked whether pressures to mark social meaning can lead linguistic systems to become less efficient at communicating semantic meaning. We exposed participants to a miniature language with uninformative constituent order and two dialects, one that employed case and one that did not. In the instructions, we socially biased participants toward users of the case dialect, users of the no-case dialect, or neither. Learners biased toward the no-case dialect dropped informative case, thus creating a linguistic system with high message uncertainty. They failed to compensate for this increased message uncertainty even after additional exposure to the novel language. Case was retained in all other conditions. These findings suggest that social biases not only interact with biases for efficient communication in language change but also can lead to linguistic systems that are less efficient at communicating semantic meaning.

Acknowledgments

We thank Aja Altenhof for help with participant recruitment, as well as Charlie Torres and Vanessa Nieto for help with stimuli creation. The second experiment was supported by a grant from the University of Pennsylvania University Research Fund. The third author was also supported by the National Science Foundation (grant number 1946882).

Data availability

The data that support the findings of this study are openly available in an Open Science Foundation repository at https://osf.io/hb9dc/

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹ Here we are assuming that case marking involves extra morphemes. This is not strictly speaking necessary; case can also be marked by phonological alternations on the root (as, for instance, in Irish). However, this is a less common pattern, and the alternations involved often arise historically under the influence of case-marking morphemes that were subsequently lost. There is also a cost associated with maintaining complex case paradigms, however marked (Ackerman & Malouf, 2013).

² The Japanese morphemes in question are more typically referred to as discourse particles, rather than case markers, owing to the nature of what they mark; our point here is that similar post-nominal morphemes in Japanese are used to mark both case and information structure.

³ This is not to say that redundant morphology carries no information; on the contrary, it plays an important role in combating noise (Frank & Jaeger, 2008; Stevens & Roberts, 2019). For the rather constrained communicative contexts in our setup, however, noise levels are low enough for a single grammatical cue to be generally sufficient.

⁴ Post-hoc power analysis conducted to determine participant numbers for Experiment 2 (see Experiment 2 Participants section) further confirmed that this final sample had appropriate power to detect all effects considered in Experiment 1 (80% as recommended by Brysbaert & Stevens, 2018).

⁵ Sliding difference coding compares the mean of the dependent variable for one level of the categorical variable to the mean of the dependent variable for the prior adjacent level. This coding scheme is appropriate for this analysis since the conditions in our experiment are ordered in terms of the expected likelihood of case use: bias-for-case > no-bias > bias-for-no-case.

⁶ See Appendix A for full results of this model and for all other models reported.

⁷ We include graphs of the equivalent data from Experiment 1 (Figures B1 and B2 in Appendix B), but do not provide statistical analysis because there was not adequate power for models investigating specific strategies, as explained in the Participants section.