Should Yu. A. Chibrikova read Yu. as a given-name initial? Two-letter initials at the start of a name read as titles #1136

The gap

Some initials are written with two letters: a transliterated Cyrillic letter (Yu. for Ю, Zh. for Ж, Kh. for Х, Sh. for Ш, Ch. for Ч, Ts. for Ц, Ya. for Я), or a Latin digraph (Th. for Thomas or Theodor, Ch. for Christian, Ph. for Philipp). Russian and older European scholarly bylines use them constantly. At the start of a name, rules/H2 (the leading-abbreviation title) reads any unlisted two-letter word ending in a period as a title, so the initial becomes a title. When the name has no other given word, given comes back empty. Measured on master 13e7314:

Yu. A. Chibrikova    → title 'Yu.', given 'A.', family 'Chibrikova'
Yu. Baryshnikov      → title 'Yu.', given '',   family 'Baryshnikov'
Zh. I. Alferov       → title 'Zh.', given 'I.'
Ya. B. Zeldovich     → title 'Ya.', given 'B.'
Th. Lyras            → title 'Th.', given ''
Chibrikova, Yu. A.   → title 'Yu.', given 'A.'   (H2 also fires at the head of the part after a family comma)
A. Yu. Petrov        → given 'A.', middle 'Yu.'  (right: H2 only fires in first position)
Ю. А. Чибрикова      → given 'Ю.', middle 'А.'   (right: the Cyrillic original is one letter, so H2 doesn't fire)

So the native-script name reads correctly, and only its romanization misreads.

How often

• The OpenAlex byline sample in ourresearch/human-name-parser-gold-standard (15,309 names) contains Yu. 11 times. In 5 it opens the name and reads as a title, alongside Th. (2), Sh. (1) and Ch. (1).
• ORCID credit names (orcid-researchers, the 356,956-entry agreement record) contain 21 digraph initials. 4 open the name and read as a title. The ORCID sample has few because ORCID holders tend to write out their given names, while journal metadata keeps the initials.

Not every two-letter word is an initial

The same shape includes real titles, so a shape-only fix would trade one misreading for another. Unlisted two-letter period words opening a name in the local corpora, measured 2026-11-06:

Ma. already reads as a given name through the given-name-abbreviation route (#789). Md. is the shape #853 left as a title ("the period-marked Md. is still a title by rules/H2's leading-abbreviation shape", decisions/indic-honorifics). ORCID's boxes are new evidence on that word: 140 of 158 credit names opening with Md. have it in the given box.

Options

1. Vocabulary, word by word. Give each initial a given reading with same_as, as fil.toml's ma has. The catch is that same_as names one full name, and Yu. stands for Yuri, Yulia, Yuliy or Yury. The property would need to accept "an initial" as a value, or a new property would need to mark one. List the real titles (Er., Ku., Pt., Vd.) as titles where they are missing.
2. A shape exception in H2. A two-letter word that is a transliteration digraph (Yu Ya Ye Yo Zh Kh Ch Sh Ts, and Shch, Dzh) would not be a title by shape. This is closed and script-based, but Ts. and Sh. collide (table above), so they would still need vocabulary or a pack.
3. Under the RU pack only. The Cyrillic original already reads correctly. Scoping the fix to a pack covers romanized Russian, but not Th. and Ch., and not a default parse of a Russian byline, which is where these appear.

Open

• Does a fix also cover Md. and reopen that part of Should a title word written after another title be the given name (Mr Sultan Khan, Mr Md Karim Uddin)? #853? The derived SUFFIX_ACRONYMS_AMBIGUOUS marker is what Should a title word written after another title be the given name (Mr Sultan Khan, Mr Md Karim Uddin)? #853 declined, and that question stays separate from the H2 shape question.
• Does a fix also cover the head of the part after a family comma (Chibrikova, Yu. A.)? H2 fires there too, and the citation form is common.

Found by scoring master against ourresearch/human-name-parser-gold-standard, an LLM-labelled sample of OpenAlex bylines.

🤖 Generated with Claude Code