The Romani language has for most of its history been an entirely oral language, with no written form in common use. Although the first example of written Romani dates from 1542, it is not until the twentieth century that vernacular writing by native Romani people arose.
Printed anthologies of Romani folktales and poems began in the 20th century in Eastern Europe, using the respective national scripts (Latin or Cyrillic). Written Romani in the 20th century used the writing systems of their respective host societies, mostly Latin alphabets (Romanian, Czech, Croatian, etc.).
Standardization
Currently, there is no single standard orthography used by both scholars and native speakers. Efforts of language planners have been hampered by the significant dialectal divisions in Romani: the absence of standard phonology, in turn, makes the selection of a single written form problematic.
In an effort to overcome this, during the 1980s and 1990s Marcel Courthiade proposed a model for orthographic unification based on the adoption of a meta-phonological orthography, which "would allow dialectal variation to be accommodated at the phonological and morpho-phonological level". This system was presented to the International Romani Union in 1990, who adopted it as the organization's "official alphabet". This recognition by the International Romani Union allowed Courthiade's system to qualify for funding from the European Commission.
Despite being used in several publications, such as the grammar of Romani compiled by Gheorghe Sarău and the Polish publication Informaciaqo lil, the IRU standard has yet to find a broad base of support from Romani writers. One reason for the reluctance to adopt this standard, according to Canadian Rom Ronald Lee, is that the proposed orthography contains a number of specialised characters not regularly found on European keyboards, such as θ and ʒ.
Instead, the most common pattern among native speakers is for individual authors to use an orthography based on the writing system of the dominant contact language: thus Romanian in Romania, Hungarian in Hungary and so on. A currently observable trend, however, appears to be the adoption of a loosely English-oriented orthography, developed spontaneously by native speakers for use online and through email.
Descriptive linguistics has, however, a long and established tradition of transcription. Despite small differences between individual linguists in the representation of certain phonemes, most adhere to a system which Hancock terms Pan-Vlax.
Latin script
The overwhelming majority of academic and non-academic literature produced currently in Romani is written using a Latin-based orthography. There are three main systems that are likely to be encountered: the Pan-Vlax system, the International Standard and various Anglicised systems.
Pan-Vlax
In most recent descriptive literature, a variety of orthography which Ian Hancock terms Pan-Vlax will likely be used. This orthography is not a single standardised form, but rather a set of orthographical practices which exhibit a basic "core" of shared graphemes and a small amount of divergence in several areas. The Pan-Vlax script is based on the Latin script, augmented by the addition of several diacritics common to the languages of eastern Europe, such as the caron. Sometimes stress is indicated with an acute accent.
In the following table, the most common variants of the graphemes are shown. The phonemes used in the table are somewhat arbitrary and are not specifically based on any one dialect (for example, the phoneme denoted in the table can be realised as , or , depending on dialect):
The use of the above graphemes is relatively stable and universal, taking into account dialectal mergers and so on. However, in certain areas there is somewhat more variation. A typically diverse area is in the representation of sounds not present in most varieties of Romani. For example, the centralised vowel phonemes of several varieties of Vlax and Xaladitka, when they are indicated separately from the non-centralised vowels, can be represented using ə, ъ or ă. Another particularly variant area is the representation of palatalised consonants, which are absent from a number of dialects. Some variant graphemes for include tj, ty, ć, čj and t᾿. Finally, the representation of the second rhotic, which in several dialects has been merged with , tends to vary between ř, rr, and rh, and sometimes even gh, with the first two being the most frequently found variants.