Transliteration and the problem of the many names

A Russian or Kazakh name has one correct form in its own script and many possible forms in Latin letters. The same surname can appear in company registers, court records, sanctions lists, press reports and foreign corporate filings under different spellings. Research that searches for a single English spelling will miss records about the subject and may pick up records about someone else. Both errors are avoidable with a disciplined method.

Why spellings multiply

Several romanisation systems are in common use. The BGN/PCGN system, developed by the American and British bodies responsible for geographical names and widely followed in English, renders Щ as shch and Ж as zh. ISO 9 maps each Cyrillic letter to a single Latin character, often with diacritics, so the same letters become ŝ and ž. Library and academic systems differ again. Other European languages bring their own conventions, so a name romanised in a French, German or Polish document can look quite different from its English form.

Official practice has also changed. The rules for romanising names in Russian international passports have been revised over the years, so one person may hold documents, visas and foreign records carrying different spellings of the same name. Individuals may also adopt a preferred spelling for business or personal use abroad.

Naming conventions add a further layer. A Russian full name consists of a given name, a patronymic derived from the father's given name and a surname, and official documents often place the surname first. A patronymic may be reduced to an initial, omitted from foreign documents or mistaken for a surname. Familiar forms of given names, which can differ markedly from the formal name, appear in press reports and correspondence.

Central Asian names

Kazakh and other Central Asian names vary further. During the Soviet period many surnames took Russified endings such as -ov and -ova. Since independence some people have adopted national forms, using elements such as -uly (son of) and -kyzy (daughter of), or have dropped the Russified ending. The same family may use different forms across generations, and one individual may appear under both in different records.

Kazakh Cyrillic includes letters that Russian does not have. In Russian-language documents these are often replaced by their nearest Russian equivalents, so Қ may become К. The planned move of the Kazakh alphabet to Latin script, whose timetable has been revised more than once, introduces further spellings. In the version adopted in recent years, Қ is rendered as Q. A single name can therefore exist in Kazakh Cyrillic, in Russian Cyrillic, in an English transliteration of the Russian form and in Kazakh Latin. Uzbekistan, where Latin and Cyrillic scripts are both in use, presents similar issues.

Companies and their names

Company names raise a separate problem. A hypothetical Russian company called ООО «Северная звезда» may appear in English as Severnaya Zvezda LLC, as OOO Severnaya Zvezda or in translation as Northern Star. The legal form may be transliterated, translated or dropped. Many unrelated companies share common names, so a name match on its own proves very little. Registration and tax numbers are the reliable anchor.

The main variant types to account for are set out below.

• Romanisation system. Shch, sch, šč and ŝ can all stand for the same Cyrillic letter.
• Name order and patronymics. Surname-first order, initials and omitted patronymics change how a name is indexed.
• Russified and national forms. A Central Asian surname may appear with or without -ov or -ova, or with -uly or -kyzy.
• Script reform. Kazakh names may exist in Cyrillic and in one or more Latin versions.
• Translated company names. A company may be known by a transliterated name, a translated name or a brand.
• Familiar forms. Diminutives and anglicised given names appear in informal sources.

A practical method

We start in the original script. Searching domestic registers, courts and press in Cyrillic avoids most transliteration problems at source. From the original form we generate Latin variants systematically, covering the main romanisation systems, passport conventions and likely errors, and run each variant through foreign registers, sanctions lists and press archives.

Every match is then tested. Identity is anchored with lawful identifiers such as registration and tax numbers, dates of birth where they are lawfully available, addresses and documented associations. A similar name with no supporting identifier is recorded as unconfirmed. This discipline matters most in sanctions and adverse media screening, where a false match can harm an unconnected person and a missed match can leave a client exposed.

A matching name is a lead, and only an identifier confirms who it belongs to.

Personal data gathered to confirm identity should be limited to what the purpose requires and handled under the data protection rules that apply. Our working files record each spelling found, where it was found and how it was linked to the subject. That record lets a reader see why a match was accepted or rejected, and gives the conclusion its credibility.