Researcher proposes formal adequacy index to score whether text formality fits the situation
An arXiv paper introduces a quantitative measure comparing a text's observed formality with the formality its communicative context expects, tested on Polish apology emails.
In plain words
A researcher has proposed a way to score not just how formal a text is, but whether that level of formality suits the situation it is used in. The method compares the formality measured from a text's language with the formality expected from the context, allowing for different tolerance when a text is too casual or too stiff. It was illustrated on Polish apology emails written by humans or generated by a large language model.
What was announced
The paper argues that measuring the level of linguistic formality alone cannot show whether a chosen mode of communication is appropriate for a given situation. Its proposed formal adequacy index instead models two quantities separately: observed formality, derived from the linguistic features of a text, and expected formality, derived from the characteristics of the communicative situation. Adequacy is then assessed by comparing the two, using contextual tolerance thresholds that may differ for a shortfall of formality versus an excess.
To illustrate the approach, the author analyses Polish-language emails containing apologies for being late, written either by human language users or generated by a large language model. The author frames the contribution as a general methodological framework for quantitatively studying formal adequacy across communicative contexts, while stressing that the implementation is preliminary and requires further calibration and empirical validation.
What it does not show
- The evidence is the paper's abstract only; the full article's methods, data and results are not covered, and the author states the implementation is preliminary and requires further calibration and empirical validation.
- The index is built and illustrated on Polish-language material, so nothing in the source establishes how it performs in other languages or contexts.
Sources
Source-support review completed 20 Nov 2026. This story was drafted with AI assistance and checked against retained source passages. Automated checking is not independent verification.
Coverage: arXiv: abstract.
Updates and corrections
No corrections so far.
First published 20 Nov 2026
New to the subject? AI from the Ground Up starts from the beginning.