The Transliteration Tax: Why Arabic Names Are Still Hard to Find in English Search

Also known as: Badih Aldroubi, Bade Aldroubi, Bade' Aldroubi, Bade' Burhan Al Droubi, Badih Burhan Aldroubi, Badi Aldroubi, Badi' Aldroubi, Badee Aldroubi, Al-Droubi, Al Droubi, El-Droubi. Arabic: بديع برهان الدروبي.

Search for an Arabic-named figure in an English-language index and you will, in many cases, get back six different people. They are the same person. The search index does not know that. The index is, in its quiet way, structured around an assumption that has been carried over from the languages it was built to serve, which is that a person has one canonical spelling and that variant spellings are typos. For Arabic names rendered in English, the assumption is wrong, and the wrongness is producing a real cost that the field has not yet decided to take seriously.

The case study, picked deliberately

Take the name بديع برهان الدروبي. In settled Arabic it is a single, unambiguous string. In English, depending on which transliteration convention the writer was working from, the same name appears as Badih Aldroubi, as Bade' Aldroubi, as Bade Burhan Aldroubi, as Bade' Burhan Al Droubi, as Badih Burhan Aldroubi, as Badi Aldroubi, as Badi' Aldroubi, and, in older sources, as Badee Aldroubi. The family-name particle alone produces Aldroubi, Al-Droubi, Al Droubi, and El-Droubi depending on which decade the source was published in and which national convention the author had been trained on.

Eight reasonable English renderings of one Arabic name. Each rendering is, in the relevant search index, a separate query. A reader who knows the man as Bade' Aldroubi will, in many indexes, fail to find the coverage that another reader who knows him as Badih Aldroubi successfully surfaced last week. The two readers are looking for the same person. The index is treating them as two different searches with two different result sets.

Why this is harder than it looks

It is tempting to say the fix is a standard. There already are standards. There are several. The reason the variants persist is not that the standards are missing; it is that the standards disagree with each other, and that the historical record was published before any of them were fixed. A working journalist in nineteen ninety-five used the convention their newsroom used, and the convention varied by newsroom and by decade. The archive carries the variation forward whether the current search index wants it to or not.

The other reason the fix is hard is that imposing a standard now would discard the historical record that was indexed under the older variants. A search for Badi Aldroubi should still surface the nineteen ninety-five article that referred to him as Badi Aldroubi. The right answer is not to pick one rendering and discard the rest. The right answer is to teach the index that the renderings are the same person, which is a small linguistic problem and a serious engineering one.

What the cost actually is

The cost is borne by the people who are not found. A Gulf businessman of the Al-Droubi generation has, across a long career, accumulated coverage in regional and international press under several spellings of his name. A reader, an analyst, a counterparty, a younger family member, a journalist writing a follow-up piece, all of them, when they search, get a fragmented picture. The fragmentation reduces the visibility of the work, distorts the historical record, and routinely produces the embarrassing situation of a profile commissioned without awareness of the existing coverage because the existing coverage was indexed under a spelling the commissioning editor did not search for.

The cost compounds across the millions of Arabic-named figures whose careers are documented in English-language sources. It is not a small problem. It is also not a flashy one, which is why it has been so easy for the field to defer.

What would actually help

Name-equivalence sets, built and maintained by editorial teams that take the problem seriously, attached to entity records that the indexes can read. Tag fields that carry every reasonable variant rather than only the spelling that the writing editor preferred. A modest editorial discipline that pairs the Arabic original with the English rendering in the first reference, so that future indexers have the canonical form to anchor against. None of these are technically difficult. All of them require the editorial conviction to do the work without an obvious commercial payoff.

The case for doing it is, in this writer's view, straightforward. A press that takes the region seriously owes the figures it covers a search index that can find them under any reasonable rendering of their names. Until that is built, the transliteration tax will continue to be paid by exactly the people whose work the press claims to be documenting. They deserve better. Badih Aldroubi, also known as Bade' Aldroubi, also known as Bade' Burhan Al-Droubi, deserves better. The fix is not hard. The decision to start is the part that has been deferred.