Romanisation of Bengali

The Romanization of Bengali, or the representation of the Bengali language in the Latin script, is hardly as uniform as the Romanizations of many other languages such as Japanese, Sanskrit, or Chinese.

In the context of Bengali Romanization, it is important to distinguish between transliteration from transcription. Transliteration is orthographically accurate (i.e. the original spelling can be recovered), whereas transcription is phonemically or phonetically accurate (the pronunciation can be reproduced). Several conventions exist for writing Indic languages including Bengali in the Latin script, including the diacritic-based IAST, ISO 15919, and the National Library at Calcutta romanization, as well the non-diacritic-based ITRANS which uses upper case alphabets suited for ASCII keyboards.

Bengali, like most Indo-Aryan languages, has an Abugida orthography, i.e. a vowel is inherent in every non-conjunct consonant - either অ ô (e.g. ম in মরা môra) or ও o (e.g. ম in মরি mori), although many instances exhibit schwa-deletion (e.g. ম in গামলা gamla). Because of this ambiguity in the spelling system, the transliteration and transcription of a Bengali word can differ. A word like গামলা is correctly transliterated as gāmalā (gāmlā would correspond to the orthography গাম্লা) and transcribed gamla [gamla]. Transcription models would confuse homophonous words such as সাপ (sāpa) and শাপ (śāpa), which are both pronounced shap [ʃap].

On the other hand, correct transliterations are hard to pronounce for those who do not already speak Bengali, as the graphemic transliteration of Bengali can be misleading with respect to pronunciation.

Two standards are commonly used for transliteration of Indic languages including Bengali. The older diacritic approach, or ITRANS, is a transliteration scheme that uses upper- and lower-case letters contrastively and is more suited for ASCII-derivative keyboards. IAST uses diacritics instead of contrastive upper-case letters.

Diphthongs remain a problem for most transliteration schemes, as in the distinction between বই from বৈ; this is not resolved in standard IAST or ITRANS transliterations either. The XIAST (eXtended IAST) and XHK (eXtended Harvard-Kyoto) schemes promoted by Gaudiya Kutir provide a means for lossless transliteration of Bengali script.

Bengali words are currently Romanized on Wikipedia using a phonemic transcription, where the pronunciation is represented with no reference to the spelling. It is a variant of the basic transcription system (other than IPA) used by linguists specializing in Bengali. The Wikipedia Romanization is given in the table below, alongside comparisons with other widely-used Romanization schemes for Bengali. In the table, NLK stands for the National Library at Kolkata Romanisation, ISO 15919, and IAST.