Systematicity between Forms and Meanings across Languages Supports Efficient Communication

Doreen Osmelak, Yang Xu, Michael Hahn, Kate McCurdy · 2026

Read Systematicity between Forms and Meanings across Languages Supports Efficient Communication by Doreen Osmelak, Yang Xu, Michael Hahn, Kate McCurdy (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-17181.txt.

Introduction

Doreen Osmelak, Yang Xu, Michael Hahn, and Kate McCurdy test if tongue forms map to sense in set ways. Tongues map sense to word forms in wild, unlike ways. Past work says those maps aid swift talk. But form-to-form links in a tongue got less heed. Their 2026 preprint checks how verb and pronoun forms mark set sense traits. Traits span person and count, in a wide tongue set.

Set marks that recur across tongues hint at shared rules. Such rules would mean talk shapes form, not chance. The finds back the swift-talk tale in a new way. Tongue science gains cross-land proof here.

Catalog record

CREATORDoreen Osmelak, Yang Xu, Michael Hahn, Kate McCurdy
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE81 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-17181

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Systematicity between Forms and Meanings across Languages Supports Efficient Communication Doreen Osmelak, Yang Xu, Michael Hahn, Kate McCurdy Abstract Languages vary widely in how meanings map to word forms. These mappings have been found to support efficient communication; however, this theory does not account for systematic relations within word forms. We examine how a restricted set of grammatical meanings (e.g. person, number) are expressed on verbs and pronouns across typologically diverse languages. Consistent with prior work, we find that verb and pronoun forms are shaped by competing communicative pressures for simplicity (minimizing the inventory of grammatical distinctions) and accuracy (enabling recovery of intended meanings). Crucially, our proposed model uses a novel measure of complexity (inverse of simplicity) based on the learnability of meaning-to-form mappings. This innovation captures fine-grained regularities in linguistic form, allowing better discrimination between attested and unattested systems, and establishes a new connection from efficient communication theory to systematicity in natural language. 1 Introduction Languages express a vast array of meanings with a limited set of forms. A fundamental goal of natural language research is to understand how these forms map onto meanings, and why this mapping differs across languages. For example, standard English uses the pronoun you for any second-person addressee, while other languages use distinct forms based on features like gender or number (e.g., sen “you (singular)” vs. siz “you (plural)” in Turkish; Figure 1 ). What drives this cross-linguistic variation in form-meaning mappings? Figure 1: Turkish pronouns show systematic form-meaning mappings: person is consistently marked by prefixes (e.g., s- for second person), number by suffixes. Language evolution research demonstrates that such systematicity supports learnability . Our model connects these findings, proposing that learnable, systematic mappings contribute to communicative efficiency . We posit that this variation arises because languages evolve under two competing communicative pressures. On one hand, a general bias toward simplicity encourages languages to minimize formal distinctions. This criterion favors the English pronoun system, with only one second-person form, as less cognitively demanding than Turkish, which requires the speaker to distinguish addresses by number. On the other hand, a competing bias toward accuracy encourages languages to maintain formal distinctions, enabling precise decoding of meaning from form. For instance, a Turkish listener can instantly understand whether the speaker intends to address an individual ( sen ) or a group ( siz ), while an English listener must infer between these two meanings of the word you . English and Turkish pronoun systems illustrate two different design solutions to the trade-off between simplicity and accuracy ( Zaslavsky et al. 2021b ) . This trade-off has been explored through two distinct research traditions. An influential line of work on efficient communication ( Kemp et al. 2018 ; Gibson et al. 2019 ; Hahn et al. 2020 , e.g.) analyzes how natural languages balance simplicity and accuracy in their lexicons. This research finds that lexicons achieve near-optimal simplicity-accuracy trade-offs across semantic domains including color ( Zaslavsky et al. 2018 ) , kinship ( Kemp and Regier 2012 ) , and numerals ( Xu et al. 2020 ) . In parallel, research on language evolution ( Smith et al. 2017 ; Culbertson and Schuler 2019 , e.g.) investigates how artificial languages evolve in laboratory settings. Through experimental manipulation, this work shows that interacting biases toward simplicity and accuracy drive the emergence of core “design features” of natural language ( Hockett 1960 ; Christiansen and Chater 2008 ; Smith 2022 ) . One such core feature is systematicity , 1 1 1 Systematicity enables another core design feature: compositionality , the capacity to build complex expressions by combining discrete parts ( Frege 1914 ; McCurdy et al. 2024 ) . in which discrete parts of linguistic forms reliably indicate specific meanings; consider the Turkish pronouns in Figure 1 , where prefixes indicate person (e.g. second person pronouns begin with s- ) and suffixes indicate number. Systematic mappings emerge in artificial languages as a design solution to the competing pressures of simplicity and accuracy ( Smith et al. 2013 ; Kirby et al. 2015 ; Smith 2022 ; Smith and Culbertson 2025 ) : decomposing complex expressions into parts yields a smaller, simpler lexicon, while recombining these parts enables fine-grained distinctions and accurate decoding. The language evolution literature thus identifies systematic form-meaning mappings as a core mechanism of efficient communication. However, this insight is not yet reflected in research on communicative efficiency in natural languages, which lacks a general framework for modeling forms’ internal structure. We propose a unified information-theoretic framework to integrate these two research traditions and assess how systematic form-meaning mapping supports efficiency in natural language. Our key contribution is a novel complexity measure based on the learnability of meaning-to-form mappings. This measure captures systematicity in linguistic forms within the theoretical framework of efficient communication. We evaluate our framework in two domains: verbs and pronouns. Verbs offer rich variation in how structured meanings map to forms across languages, and in many languages exhibit overtly systematic form-meaning covariation through inflectional morphology ( Haspelmath and Sims 2010 ) . Pronouns provide a natural comparison case to an earlier analysis under the well-known Information Bottleneck model ( Zaslavsky et al. 2021b ) , allowing us to directly compare approaches. We find that our framework successfully represents structured form-meaning mappings in both domains, and consistently discriminates attested from unattested systems. Crucially, our comparison finds that explicitly modeling systematicity yields a more precise account of efficiency in language, better capturing the fine-grained regularities that characterize natural linguistic systems. singular (sg) dual (du) plural (pl) 1 m a- 12u3 -u na- 12u3 -u na- 12u3 -u f a- 12u3 -u na- 12u3 -u na- 12u3 -u 2 m ta- 12u3 -u ta- 12u3 -āni ta- 12u3 -ūna f ta- 12u3 -īna ta- 12u3 -āni ta- 12u3 -na 3 m ya- 12u3 -u ya- 12u3 -āni ya- 12u3 -ūna f ta- 12u3 -u ta- 12u3 -āni ya- 12u3 -na Table 1: Basic imperfective paradigm of a stem I Classical Arabic verb for feature values gender { m asculine, f eminine}, person {1, 2, 3}, and number {sg, du, pl}. Numbers (e.g. 12u3) represent root consonants. 2 Background 2.1 Paradigms and Syncretism A paradigm comprises systematic relations between word forms which are structured by grammatical categories such as person, gender, number, tense, or case. These relations can hold between distinct words, such as the Turkish pronouns shown in Figure 1 , or parts of words, such as the prefixes and suffixes shown in Table 1 . Specific feature combinations are realized by the form in the corresponding cell of the paradigm table (e.g. first person singular \rightarrow a- ). Syncretism occurs when two or more cells of a paradigm with distinct grammatical functions share the same surface form. For instance, ta-,-u realizes multiple feature combinations in Arabic ( Figure 1 , Table 1 ). Within highly structured paradigmatic meaning spaces, syncretism directly corresponds to the form-meaning partitions relevant to communicative efficiency; for instance, Zaslavsky et al. 2021b show that pronoun syncretism patterns are optimized for efficient communication. By definition, syncretism reflects only identity relations between forms ( Haspelmath and Sims 2010 ) . Our proposed model extends beyond syncretism to capture partial overlap in forms, such as

Read the complete text