Towards a quantifiable measure of orthographic congruence between two languages

The advantage of having orthographic information available during vocabulary acquisition has been consistently observed in an L1 learning context. However, studies are inconclusive about whether this "orthographic facilitation effect" also extends to the acquisition of L2 vocabulary. One of the main reasons for this discrepancy is the language-specific nature of grapheme-to-phoneme correspondences (GPCs), as L1 GPCs are not necessarily identical to L2 GPCs even though the graphemes themselves are shared between both languages. Indeed, recent studies suggest that whenever L1 and L2 GPCs are different from each other, i.e., incongruent, the resulting interference might be detrimental to the integration of novel lexical representations. The dominant approach for determining congruence in the prevailing literature consists in a subjective speaker-based evaluation, meaning that non-native L2 word pronunciations are deemed congruent or not by L2-proficient examiners. Although this method has yielded promising results, the lack of an alternative procedure underlines the necessity for a quantifiable measure of congruence. Through the use of infralexical statistics, we have established a framework for calculating the degree of L1-L2 GPC congruence for any given L2 word. This entails the placement of L2 pronunciations on a percentage-based continuum of congruence, instead of a dichotomic categorization. The grapheme-to-phoneme probability (G2P%) framework is simple to use, open-source, and readily available on  https://osf.io/5h2mn/ for immediate employment, catering to a wide range of investigative goals surrounding the concept of orthographic congruence.