The process of changing Arabic script into a romanized (or latinized) equivalent would seem to be a straightforward or even trivial one, but it is not. Because it requires complete accuracy, “scientific transcription” is a painstaking effort demanding high levels of phonological and morphological knowledge, and transliteration errors and inconsistencies tend to creep into even the best-edited publications. For a number of reasons, romanization (conversion from a non-roman alphabet to a roman one) from Arabic can be problematic, especially because of the invisibility of short vowels in Arabic script, which are crucial to accurate pronunciation. Other common areas of romanization problems include the misrepresentation of gemination, short versus long vowel notation, representation of epenthetic vowels, morpheme boundaries (if noted), case-marking, and word boundaries.

Traditionally, a distinction is drawn between transcription and transliteration. I have long relied on Charles Ferguson’s definitions of these processes, and reproduce them here.