Clarifying orthography: Orthographic transparency as compressibility
This provides a script-agnostic tool for linguists and researchers to measure orthographic transparency, though it is incremental as it formalizes existing intuitions.
The authors tackled the lack of a unified metric for orthographic transparency by quantifying it as mutual compressibility between orthographic and phonological strings, using algorithmic information theory and neural sequence models, and validated it across 22 languages with diverse scripts, confirming intuitive rankings.
Orthographic transparency -- how directly spelling is related to sound -- lacks a unified, script-agnostic metric. Using ideas from algorithmic information theory, we quantify orthographic transparency in terms of the mutual compressibility between orthographic and phonological strings. Our measure provides a principled way to combine two factors that decrease orthographic transparency, capturing both irregular spellings and rule complexity in one quantity. We estimate our transparency measure using prequential code-lengths derived from neural sequence models. Evaluating 22 languages across a broad range of script types (alphabetic, abjad, abugida, syllabic, logographic) confirms common intuitions about relative transparency of scripts. Mutual compressibility offers a simple, principled, and general yardstick for orthographic transparency.