HackerPapers

arXiv:1912.13321cs.CL

OTEANN: Estimating the Transparency of Orthographies with an Artificial Neural Network

Xavier Marjou

Abstract

To transcribe spoken language to written medium, most alphabets enable an unambiguous sound-to-letter rule. However, some writing systems have distanced themselves from this simple concept and little work exists in Natural Language Processing (NLP) on measuring such distance. In this study, we use an Artificial Neural Network (ANN) model to evaluate the transparency between written words and their pronunciation, hence its name Orthographic Transparency Estimation with an ANN (OTEANN). Based on datasets derived from Wikimedia dictionaries, we trained and tested this model to score the percentage of correct predictions in phoneme-to-grapheme and grapheme-to-phoneme translation tasks. The scores obtained on 17 orthographies were in line with the estimations of other studies. Interestingly, the model also provided insight into typical mistakes made by learners who only consider the phonemic rule in reading and writing.

arXiv comments
9 pages, 9 figures, 3 tables
Journal reference
DOI
doi:10.18653/v1/2021.sigtyp-1.1
Subjects
cs.CLComputation and Language

On Hacker News

Linked by 3 identifiable commenters in 5 comments across 3 threads, from Nov 2020 to Jul 2025.

Comments linking this paper, per year.

Discussed in comments3 threads

  1. Why English doesn't use accents 3 comments

    • According to this paper [0] and my own experience, it's way easier to pronounce a word in French given the spelled word than in English. It's slightly harder to spell French than English for the model of the study, but it's really close. Now, in my personal experience, I feel…

      dadoumView on HNParent context

    • The complexity of these rules, and the number of exceptions that you need to learn notwithstanding the rules, can be roughly estimated for any given language by training a language model on word <-> IPA correspondence for that language (using a subset of the vocabulary as a…

      int_19hView on HNParent context

    • English is singled out because its orthography really is atrociously bad even when compared with other languages. https://arxiv.org/pdf/1912.13321

      int_19hView on HNParent context

  2. Deseret Alphabet 1 comment

    • There was an amusing study where they tried to measure how phonemic various language orthographies are by training an ANN model on some data to convert back and forth between spelling and pronunciation, and then checking the error rate on more inputs…

      int_19hView on HNParent context

  3. What language do people speak in the Balkans, anyway? (2019) 1 comment

    • This is more generally the concept of Orthographic depth [1] A recent study attempted to categorize languages and found English and French to be among the most opaque, Finnish and Korean to be the most transparent, and German in the middle. [2] [1]…

      gokView on HNParent context

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.