Universal Transliteration Model and a Reversible System for Korean

Peter Kang - CIA

Intended Audience:
Session Level: Beginner

Currently two different Korean transliteration systems are used in the USG. The most widely used is the McCune Reischauer (M R) system, adopted by the Board on Geographic Names as the USG standard. The M-R system is built on the phonetic assimilation rules, and entails considerable spelling ambiguity. For such a pronunciation based system, search and retrieval from computer databases is difficult and automating transliteration process is impractical. To achieve consistency and allow information to be flawlessly converted from Hangeul to Latin and back to Hangeul, some linguists use the Morse system which provides the one to one correspondence between the two alphabets and is thus completely reversible. Unfortunately Morse expressions are unreadable and " meaningless," e.g., Lum Beu Phk for Kim Dae Jung.

The paper will discuss the Universal Transliteration Model (UTM), and the KORDA (Korean Romanization for Data Applications) system built on that model. UTM is an idealized approach to transliteration for an alphabetic language such as Korean and Arabic (not applicable to ideographic language like Chinese), guaranteeing reversibility while at the same time optimizing readability. UTM would thus enable seamless automation and greatly simplify data search. Using KORDA as an exemplar, I will discuss some of the successful applications of UTM that have been made, and how this approach can improve the way the USG does business in many areas, including interagency sharing of intelligence data critical to the war on terrorism, unambiguous spellings of geographic names in military operations, and keeping track of native data during immigration processing.