Research project USER-GRAM: Uncovering how Speaker Efficiency Reshapes Grammar

Language is in constant flux. Language change is driven by speakers who continually invent new ways of expressing themselves and reshape a language's grammar through use. One of the primary forces behind linguistic innovation is efficiency. In everyday conversation, spontaneous simplifications frequently arise that make communication more efficient in the moment. Many speakers, for instance, use the reduced form massa instead of en massa in Swedish, eles casam instead of eles se casam in Portuguese, or sais pas instead of je ne sais pas in French.

But efficiency is not merely a matter of shortening. According to the Principle of Negative Correlation between Accessibility and Costs (Levshina 2022), speakers expend little effort on meanings that are predictable and more effort on meanings that are less so: shorter forms for what the addressee can infer anyway, longer forms for what is less accessible. Efficiency can therefore expand grammatical structures just as readily as it reduces them — it depends on how predictable the intended meaning is.

A clear example of the former is found in Spanish oblique relative clauses, where the definite article is optional. In the examples below, the antecedent is in italic text and the varying sequence in the relative clause appears in bold. Historically, the point of departure is the shorter variant of the relative, without the article.

(1) Compró una casa en que pensaba instalarse al jubilarse.
'She bought a house that she planned to move into when she retired.'

In contexts where the antecedent is hard to anticipate — when it is distant and competes with other constituents for the addressee's attention — a longer variant with the article emerges, easing the cognitive processing of the utterance:

(2) Compró una casa en las afueras, cerca de donde habían vivido sus padres durante muchos años, en la que pensaba instalarse al jubilarse.
'She bought a house on the outskirts of town, near the place where her parents had lived for many years, that she planned to move into when she retired.'

Up to this point, the change is precisely what the efficiency principle predicts: more material exactly where the addressee needs it (Levshina 2022). Over time, however, the longer variant also spreads to the readily accessible contexts where the shorter one once sufficed:

(3) Compró una casa en la que pensaba instalarse al jubilarse.
'She bought a house that she planned to move into when she retired.'

(3) differs from (1) in the article alone, and here the article does no work for interpretation. What emerged as an efficient solution in (2) has thus become a pure cost in (3), which appears paradoxical from the standpoint of efficiency.

While some innovations remain transient and confined to individual conversations, others become a permanent part of the language over time — and once they have, they continue to spread, including into contexts where they no longer serve the function they arose from. The structure in (3) has outlived its own efficiency motivation. What keeps it there is an open question, and one of the questions this project asks.

The project takes this observation as its point of departure and poses two questions: how can something that arises as an efficient solution at a particular moment develop into a lasting change that spreads throughout an entire language? And what becomes of the change once it has left the contexts in which it was efficient?

Why is this important to investigate? Although we know that linguistic efficiency drives language change, we still do not fully understand how transient efficient innovations become permanent parts of grammar. Still less do we know what happens afterward. There is considerable evidence that such changes eventually reach contexts where they no longer yield any gain in efficiency, and may in fact work against it: the speaker expends effort on meanings the addressee could have anticipated anyway. Grammar, in other words, carries traces of solutions that were once efficient but no longer are, and efficiency can therefore explain how a change arises without explaining where it ultimately ends up. This is a central question for understanding how people in different historical periods process language cognitively, and how speakers reshape linguistic systems over time through use.

By using corpus-based and quantitative methods to trace efficiency-driven changes across five languages (English, French, Portuguese, Spanish, and Swedish) over several centuries, the project will produce important new knowledge about how people adapt their language to communicate more efficiently in different settings, and about the mechanisms by which languages change. This can help us understand why some linguistic changes endure while others fade, why they spread further than efficiency warrants, and how the particular properties of individual languages shape that process. In doing so, we gain deeper insight into how human communication works and develops over time.