United Nations Group of Experts on Geographical Names

Report of the UNGEGN Working Group on Romanization Systems

20th March 2018

The UNGEGN Working Group on Romanization Systems (WG) met at the Institut Géographique National in Paris on 20th March 2018 and discussed items currently on the WGRS agenda and also considered other languages and scripts that might need inclusion in the Group’s work.


The most substantial item on the agenda related to the romanization of Arabic. It was recalled that UNGEGN had approved a new system for the romanization of Arabic during the 11th UN Conference on the Standardization of Geographical Names in August 2017. This system was acknowledged to have been accompanied by controversy, and was not liked by all members of the WG. Additionally, it was reported that a recent communication had been received by the Chairman of the Arabic Division containing a number of alterations to the approved system1. The WG discussed the alterations, noting that, having approved the system in a specific form, substantive alterations could not easily be made to this. The Convenors would write to the Arabic Division expressing concern at the suggestion to alter the system having been approved, whilst also commenting on the specific proposals being made.

WG member Bogusław R. Zagórski presented a paper expressing his views on the preferred treatment for the romanization of Arabic, noting that standardisation could best be achieved through application to a standard written form, though noting that this was unusually difficult given the wide variety of dialects spoken. Mr Zagórski invited input to his paper and would submit it for inclusion to the next edition of the UNGEGN Bulletin.

The WG discussed the content of the romanization system that had been approved by Resolution XI/3 noting that there were certain areas of incompleteness in the system, as well as errors in the paper’s formulation, resulting in ambiguity in application. The Convenors had prepared an initial draft of a summary of the system for the WG website. The WG had reviewed this draft and discussed suggested alterations to the introductory text, expressing that the system’s lack of widespread implementation might be incorporated. A number of items in the system’s content, such as on nunation, were also discussed, and the Convenors would update the draft accordingly for circulation and further review.


The WG had been invited to review the current draft of the Lao Romanization System prepared by the National Geographic Department of the Lao People's Democratic Republic. It was noted that the current draft included a significant list of vowel+consonant and consonant+vowel+consonant combinations that were not strictly necessary for the successful application of the system (the same results being derived through simple application of the component consonant and vowel tables). It was suggested that though these might perhaps be useful for Lao users, for the WG‘s representation of the system these would not need to be captured. In preparing a draft for the WG’s website, the Convenors would capture this simplification and also propose to the authors that it might be suitable for the Lao version of the system.

Other points that would be noted to the authors included the possibly erroneous recording of the romanization éum with acute accent.

The WG also agreed to comment of the use of ‘y’ to mark a lengthened ‘i’. It was noted that this was not thoroughly implemented in the examples given in the system and it would be verified whether this was an error or intentional, whilst also proposing that it might be simplified. It was also noted that the syllable-final consonants for a number of syllables containing long vowels had been captured as a voiced counterpart of the consonant (e.g. where a syllable followed a short vowel the consonant final was /p/ or /t/ but after a long vowel these were replaced with /b/ and /d/ respectively). It was notable, however, that syllable-final /k/ did not transform to /g/ as might be expected. The WG agreed to put this point to the authors to verify the intention.


The WG considered the most recent draft of the proposed Sinhala romanization system. It was noted that the scope of the system’s implementation had not yet been verified (i.e. whether it was planned that the Sinhala system would be applied also in Tamil-speaking areas), and the ramification for the previously adopted Sri Lankan practice to produce mapping in 3 parallel editions (English, Sinhala and Tamil) was similarly not well understood. It was, however, noted that with the approval of a system for Sinhala, the UN and international community would have a recognised and agreed method of representing Sinhala geographical names. Until now, the unapproved 1972 ‘Sharma’ system had been applied in such instances, in the absence of anything more official.

The Convenors noted that, with the WG’s input, they had made several suggestions to the authors of the system proposing alternative romanizations; some of these had been reflected in the current draft and others not, and it was acknowledged, as with some other agreed romanization systems, that there was an element of compromise in the agreed system. The WG had agreed with the Sri Lankan experts that a letter from the WG would be sent to Sri Lankan officials, with an official endorsement of the system. The WG agreed that the draft of the letter, which had already been circulated for input from the WG, would be sent to Mr Hettiarachchi.


The WG reported on the current status of the introduction of a Roman-script alphabet for the Kazakh language. This proposed alphabet had seen some significant change since the previous meeting of the WG which had been held in the course of the 11th UN Conference on the Standardization of Geographical Names in August 2017. It was recalled that at that meeting the expert from Kazakhstan, Mr Aimenov, had reported on the planned introduction of the Roman script. However since that meeting, in October 2017, an alternative alphabet had been decreed. This alphabet had inspired significant negative response with the result that in February 2018 the proposed alphabet had been changed again. The Convenors agreed to circulate information on the latest proposed alphabet and the WG agreed that it was too early to implement the newly decreed alphabet. There were a number of areas of linguistic anomaly in the newest alphabet and it was possible that the alphabet would again be subject to alteration. The WG agreed to monitor the situation noting that the phased introduction was planned to be complete by 2025.

Other scripts for which the WG agreed to collect information were:


With the approval of an ISO 639-2 language code for Montenegrin (cnr), it was agreed that a summary of the script as used in Montenegro should be prepared for the WG’s website. It was notable that Montenegro used the Roman script alongside Montenegrin Cyrillic and therefore it was hoped that, if further evidence were collected, a simple ‘table of correspondences’ (i.e. a table for the conversions used between the two writing systems) could be prepared.


The official status of Kurdish in Iraq was noted and it was agreed that therefore the script should be subject to the WG’s review. Mr Zagórski would research information on the system he had seen in use for the language. The BGN/PCGN system, derived from the Roman-script Hawar alphabet, was also noted.


Given the official status of the Amazigh language in a number of countries in North Africa the WG would include Tifinagh in its agenda. It was noted that it had not yet been decided in Algeria which would be the official script for the language, but that in Morocco the Tifinagh script was used officially. The WG agreed to collect further information in the hope of including information on Tifinagh romanization on the website.


It was agreed, given that the structure and organisation of UNGEGN had changed, that the previous practice of reissuing all the romanization systems after an UNGEGN Conference would be revised, and that only substantive changes to a system would result in a new version of the system being posted. Substantive changes would be recorded as a new version, e.g. V4 becoming V5, where minor editorial changes would be captured simply as e.g. V4.1.


The Convenors warmly thanked the participants and their IGN hosts, and particularly Élisabeth Calvarin, and noted that the meeting had been fruitful as well as enjoyable.


Brahim Atoui (Algeria), Peeter Päll (Estonia), Ulla Onkamo (Finland), Élisabeth Calvarin (France), Jean-Sebastien Majka (France), Sungjae Choo (Rep. of Korea), Bogusław R. Zagórski (Poland), Catherine Cheetham (United Kingdom), Leo Dillon (United States).

  1. Specifically the digraphs in the Roman-script results appeared in the same case as one another, so that if e.g. خ were word initial that this would be represented as KH rather than Kh.
    The optional inclusion of a dot to disambiguate between a digraph and two adjoining consonants (e.g. /sh/ as opposed to /s/ followed by /h/ was presented as . (low dot) rather than · (middle dot).