Right-to-left scripts in a reference list: Hebrew, Urdu, Persian

Hebrew, Arabic, Persian and Urdu sources go into an English-language reference list romanised, with an English translation of the title in square brackets: that is APA's rule, while MLA prefers the original script with a translation. All four scripts run right to left, so the trouble starts when the script stays in an English line. This guide, published by EdCitation, covers both halves: where right-to-left scripts sit in a reference list, and how to keep each line in order in Word.

We read every rule on 3 October 2026 at its source, APA's in an archived copy; the romanisations and translations are ours. For the styles used inside each language, see Arabic and Persian citation styles; for Arabic in English papers, citing Arabic sources.

The last section quotes what EdCitation's free Find sources, Cite a source and Verify references returned on real Hebrew, Persian and Urdu articles, faults included.

Do APA, MLA and Chicago keep Hebrew, Arabic, Persian or Urdu script?

APA romanises by default, MLA keeps the script first, and Chicago leaves the choice to the field.

APA 7

Adams (2021), on APA's blog, names Arabic, Farsi and Hebrew among languages with non-Roman alphabets and gives the seventh edition's rule, section 9.38: transliterate the title into the Roman alphabet, then add a translation in the paper's language in square brackets, never italicised. Where transliterating is not practical, the title may stay in its own script. And where an entry's first element, the author or a title standing in for one, is in the script, you file it by your own judgement, guided by published articles, or put it at the end of the list.

MLA 9

The Modern Language Association (2017) prefers the original characters with a translation, because the script "shows a respect for the foreign language". In the works-cited list the order is script, then a transliteration if you give one, then the translation. If every entry under one author is in the other language, alphabetise by that language's rules; in a mixed list, add transliterations and alphabetise by them. People's and publishers' names are transliterated only, and so, almost always, are institutions in Hebrew and other non-Roman scripts (Modern Language Association, 2018). Both posts predate the ninth edition, which we could not read.

Chicago 18

The manual's chapter 11 has sections on Arabic (11.83 to 11.88), on Hebrew (11.100 to 11.106, among them "Unromanized Hebrew phrases") and two on South Asian languages, but none named for Persian or Urdu, and its text needs a subscription (University of Chicago Press, 2024). Asked whether to write Shelomoh ben Yizḥaq or Solomon ben Isaac, Chicago's Q&A declines to rule and leaves it to the discipline, the author and the readers (University of Chicago Press, n.d.). EdCitation builds Chicago's author-date system only.

Which Library of Congress table do Hebrew, Persian and Urdu use?

Each language has its own ALA-LC romanisation table, approved by the Library of Congress and the American Library Association: Hebrew and Yiddish (2011), Arabic (2012), Persian (2012) and Urdu (2013), all on one index page (Library of Congress, 2026). Use the table for the language of the work, not for the script it is printed in.

Hebrew: you supply the vowels

The Hebrew table follows modern Israeli, mainly Sephardic, pronunciation and uses dots to tell letters apart: ḥ for ח, ṭ for ט, ḳ for ק, ṿ for a consonantal ו, and ś for שׂ where שׁ is sh (Library of Congress, 2011). For the vowels, it says, a dictionary is often needed, and it names Even-Shoshan's ha-Milon he-ḥadash. A 2023 article title, מורות ומורים אקטיביסטים, becomes Morot u-morim aḳṭivisṭim [Activist teachers, women and men]: ḳ and ṭ from the table, the vowels from us, and the prefix hyphenated as ha- is in the table's own example.

Persian and Urdu: one script, different letters

The Arabic, Persian and Urdu tables romanise the same letter differently, because the language decides. Urdu keeps ṭ, ḍ and ṛ for its retroflex sounds and kh, ch and gh for its aspirates, so خ, چ and ط take other forms (Library of Congress, 2013).

The Persian and Urdu tables both romanise Arabic words in their own texts by their own rules, and both take missing vowels from a named dictionary. An Arabic name in an Urdu book is therefore ʻAbdurraḥmān, the Urdu table's own example, where the Arabic table would keep al-. A Persian half-space becomes a prime, as in khānahʹhā.

Where do right-to-left entries go in an English reference list?

Among the other entries, by the romanised author or title, unless the entry begins in the script. A romanised title standing in for an author files by its first significant word (Adams, 2021). In APA, an entry that begins in the script goes where comparable published articles put such entries, or at the end; MLA wants a transliteration to alphabetise by. A separate Arabic list placed first belongs to theses written in Arabic, not to an English paper.

Write names as the authors do

Use the author's own Latin spelling. The Hebrew article above is by קרן קטקו-איילי, whose ORCID record gives the name as Keren Ketko-Ayali; the Urdu one below is by Ali Kavousi Nejad, as the journal prints it. Where there is no Latin spelling, romanise the name by the table and keep that form. Our guide to names in references covers order and particles.

A worked example in APA 7

Kavousi Nejad (2022), an Urdu article in a Pakistani journal, romanised by the Urdu table, with the journal's own English title as the translation, in sentence case:

Kavousi Nejad, A. (2022). Iqbāl kī shāʻirī men̲ Afrang o K͟hāvar kī kashmakash [The conflict of East & West in Iqbal's poetry]. TAṢDĪQ, 4(1), 23–35. https://doi.org/10.56276/tasdiq.v4i01.96

The Urdu table gives k͟h for خ, n̲ for the nasal ں, and o for و between two closely paired words; we supplied the vowels. Where romanising is not practical, APA lets the script stay, and because this entry begins with a romanised name it files under K as usual:

Kavousi Nejad, A. (2022). اقبال کی شاعری میں افرنگ و خاور کی کشمکش [The conflict of East & West in Iqbal's poetry]. TAṢDĪQ, 4(1), 23–35. https://doi.org/10.56276/tasdiq.v4i01.96

How do you keep a right-to-left line in order in Word?

Give each entry the direction of its first element: an entry that starts with a romanised name stays left to right, and one that starts in the script needs a right-to-left paragraph.

Why a mixed line scrambles

Spaces and punctuation have no direction of their own: Ishida (2024) shows that a comma and space between two right-to-left runs inside a left-to-right line are read as part of the right-to-left text, so the runs swap places, and that a punctuation mark closing a right-to-left phrase joins the English that follows. Microsoft (n.d.-b) gives the rule for text whose direction comes from context: the first strong character, a letter with a direction of its own, sets the reading order.

We tested this in Chrome with the Persian entry Cite a source built. In a left-to-right paragraph the line began at the wrong end: the volume and issue sat at the left margin as (42)13, the author's name came mid-line and the DOI was pushed to the far right. Set as a right-to-left paragraph, the same text read correctly from the right margin to the DOI. The Urdu title inside the English-led entry above displayed in order.

Setting the direction, step by step

1. Add the language. In Word, choose File, Options, Language and add the right-to-left language as an authoring language, with a keyboard for it; the Right-to-left and Left-to-right paragraph buttons then appear in the Paragraph group of the Home tab (Microsoft, n.d.-b).
2. Leave entries that start with a romanised name left to right.
3. Click into an entry that starts in the script and choose Right-to-left. In Word for Mac the shortcut is Control+Command+Left arrow, once shortcuts are enabled for that input language (Microsoft, n.d.-a).
4. If a full stop or bracket lands on the wrong side of a script title in an English line, put a left-to-right mark, U+200E, straight after the title; the right-to-left mark is U+200F (Ishida, 2024). In Word for Windows, type the code and press Alt+X, which Microsoft documents for turning a code into its character.
5. Keep any Persian half-space, U+200C, that the publisher printed: it decides which letters join (Ishida, 2025).
6. Check names with Find: Word searches text in the order it was typed, not as displayed, and can match with or without kashida, alef hamza or Hebrew niqqud (Microsoft, n.d.-b).
7. Save as PDF and read each script line from the right.

What do EdCitation's tools return for Hebrew, Persian and Urdu sources?

The record's own title in its own script, with the names, volume and pages the record holds; romanising and translating stay with you. That makes EdCitation the place to start: it looks the work up in the publisher's or registry's record and never writes a reference, which no chatbot can promise, and it is free with no account. Cite a source builds six styles: APA 7, MLA 9, Chicago 18 author-date, Harvard, IEEE and Vancouver. EdCitation never writes any part of your paper: it finds and checks, and the writing is yours.

Find sources

Find sources handled all three scripts. "اردو ادب" (Urdu literature) returned 273 works, and all 96 titles shown were in Arabic script; library records gave names romanised as catalogues write them, such as Kishvar Nāhīd and Abūllais̲ Ṣiddīqī. "شاهنامه فردوسی" (Ferdowsi's Shahnameh), filtered to articles, returned 677, the first 100 all Persian; "חינוך מורים" (teacher education), also filtered, returned 27. Check each record: 24 of the Urdu results carried an unrelated journal name, "Medical Entomology and Zoology".

Cite a source

For DOI 10.22054/ltr.2010.6520, a Persian article registered with mEDRA rather than Crossref, Cite a source returned in APA:

تمیم داری, ا. (2010). تأثیر شاهنامه فردوسی بر آثار حماسی غرب. متن پژوهی ادبی, 13(42). https://doi.org/10.22054/ltr.2010.6520

with "(تمیم داری, 2010)" in the text. Romanise the name and title, add a translation, and add the pages, which the record lacks. The Hebrew DOI 10.54301/pouc2348 gave "קטקו-איילי, ק. (2023). מורות ומורים אקטיביסטים. לקסי-קיי, (19), 16–19." For the Urdu DOI the record holds the whole name in one field, so APA came out as "Ali Kavousi Nejad. (2022).", with the journal's English title joined to the Urdu by a colon: invert the name and move the English into brackets.

Verify references

Verify references gave a reason for every verdict. The Urdu entry with its script title and no DOI was verified: "Title, year and first author match a published record." So were the Hebrew one in script and the Persian article above with its DOI. Persian entries in script without a DOI came back could not check, with the advice to "add its DOI"; so did the romanised Urdu entry with its DOI, its record being "in another script or language from your entry".

A Persian title typed with its half-spaces, as its own DOI registry record has them, came back verified with its DOI: "The DOI resolves to this record and the title matches." One result needs care: our romanised Hebrew entry without a DOI came back not found, though the article exists; with its DOI it came back could not check. Keep every DOI and read every flag against the article, as checking a reference is real explains. With Pro, at $8 a month, References from a file checks a whole paper and matches its citations to the list.

Quick questions

How do I cite a Hebrew source in APA?

Romanise the title, for example by the Library of Congress Hebrew table, and add an English translation in square brackets. Write the author's name as the author spells it in Latin letters.

Should I keep Urdu or Persian script in my reference list?

In APA, only where romanising is not practical. MLA prefers the script first, followed by any transliteration and the translation.

Where does an entry in Arabic script go in an alphabetical list?

In APA, where published articles put such entries, or at the end of the list. In a mixed MLA list, alphabetise by its transliteration.

Why does my Hebrew or Arabic reference look backwards in Word?

The paragraph runs left to right while the entry starts in a right-to-left script. Set that paragraph to Right-to-left, or start the entry with a romanised name.

Can EdCitation romanise a Hebrew, Persian or Urdu title?

No. EdCitation's Cite a source returns the record's own script in six styles; the romanisation and translation are yours to add.

References

• Adams, A. (2021, August 23). When and how to transliterate titles in references. APA Style. https://web.archive.org/web/20260805230910/https://apastyle.apa.org/blog/transliterated-titles-references
• Ishida, R. (2024, April 25). How to use Unicode controls for bidi text. W3C Internationalization. https://www.w3.org/International/questions/qa-bidi-unicode-controls
• Ishida, R. (Ed.). (2025, October 2). Arabic & Persian layout requirements (W3C Group Draft Note). World Wide Web Consortium. https://www.w3.org/TR/alreq/
• Kavousi Nejad, A. (2022). Iqbāl kī shāʻirī men̲ Afrang o K͟hāvar kī kashmakash [The conflict of East & West in Iqbal's poetry]. TAṢDĪQ, 4(1), 23–35. https://doi.org/10.56276/tasdiq.v4i01.96
• Library of Congress. (2011). Hebrew and Yiddish [Romanization table, PDF]. ALA-LC Romanization Tables. https://www.loc.gov/catdir/cpso/romanization/hebrew.pdf
• Library of Congress. (2012a). Arabic romanization table [PDF]. ALA-LC Romanization Tables. https://www.loc.gov/catdir/cpso/romanization/arabic.pdf
• Library of Congress. (2012b). Persian romanization table [PDF]. ALA-LC Romanization Tables. https://www.loc.gov/catdir/cpso/romanization/persian.pdf
• Library of Congress. (2013). Urdu romanization table [PDF]. ALA-LC Romanization Tables. https://www.loc.gov/catdir/cpso/romanization/urdu.pdf
• Library of Congress. (2026). ALA-LC romanization tables. https://www.loc.gov/catdir/cpso/roman.html
• Microsoft. (n.d.-a). Keyboard shortcuts in Word. Microsoft Support. https://support.microsoft.com/en-us/office/keyboard-shortcuts-in-word-95ef89dd-7142-4b50-afb2-f762f663ceb2
• Microsoft. (n.d.-b). Using right-to-left languages in Office. Microsoft Support. https://support.microsoft.com/en-us/office/languages/using-right-to-left-languages-in-office
• Modern Language Association. (2017, February 8). When translating from a language that does not use roman characters, like Chinese, do I include the original characters or a transliteration? And how do I alphabetize titles of nonroman works? MLA Style Center. https://style.mla.org/nonroman-characters/
• Modern Language Association. (2018, March 13). Should I translate names of foreign institutions? MLA Style Center. https://style.mla.org/names-of-foreign-institutions/
• University of Chicago Press. (n.d.). Proper names (Q&A). The Chicago Manual of Style Online. https://www.chicagomanualofstyle.org/qanda/data/faq/topics/ProperNames/faq0033.html
• University of Chicago Press. (2024). Chapter 11 contents: Languages other than English. The Chicago Manual of Style Online (18th ed.). https://www.chicagomanualofstyle.org/book/ed18/part2/ch11/toc.html