Turn footnote text in a converted Arabic book into Word footnotes

A printed Arabic edition carries two texts on most pages: the text of the work and, below a rule, the editor's notes. Variant readings, sources for hadith and verses, short biographies, corrections. In a muhaqqaq edition the notes are often the part a translator, a researcher or the next editor needs most.

After OCR, those notes usually arrive in one of three states. They may already be real Word footnotes. They may be ordinary paragraphs at the foot of each page, with the reference numbers in the body as plain characters such as (١) or (1). Or they may be merged into the running text. This guide is about the second case, which is also where the third ends up once you have cut the merged notes out. It goes deeper than the general QA checklist before delivery, which only asks whether footnotes are separated and markers match.

Why plain-text notes are a problem

Plain-text notes look fine on screen and break the moment anyone works with the file:

• They drift. Change the font, page size or margins and the note paragraphs flow into the middle of the next page's text, because nothing ties them to the foot of the page.
• Markers don't follow edits. Delete or add a note and every later number is wrong.
• Translators and typesetters lose them. Translation and layout tools generally handle real footnotes as notes. A paragraph is just more body text to translate or set in the wrong place.
• They pollute search. A search for a phrase in the work also hits the editor's notes.

Real footnotes keep each note attached to its reference mark, wherever the page breaks fall.

Step 1: decide the numbering before you touch anything

Most Arabic editions restart note numbers on every page: the first note on each page is (١). Once the notes are real footnotes, Word lays out its own pages, and they won't match the printed ones. That leaves three honest options:

Option	Word setting	Good for	Cost
Restart each page	Numbering: Restart each page	Matching the look of the print	Numbers no longer match the printed page, so "note 3 on p. 42" can't be found by number
Restart each section	Numbering: Restart each section, one section per chapter	Translators and long books	Different numbers from the print
Continuous	Numbering: Continuous	Short works, articles	Very large numbers in a long book

Microsoft documents all three numbering rules for Word footnotes (continuous, reset for each section, reset for each page). If readers must be able to cite the printed edition, keep the printed page numbers in the body as markers such as [ص ٤٢] whichever numbering you choose. Agree this with whoever receives the file before you start. Changing it afterwards is easy in Word, but redoing the cross-references in a translation is not.

Step 2: mark the job out

Before converting anything, work out how big it is:

1. Search for the opening bracket the edition uses before its note numbers, for example "(". Count roughly how many notes there are and whether they use Arabic-Indic digits (١٢٣) or Western digits (123). OCR can mix the two in one file.
2. Look for the separator. Many converted files keep a line of underscores or a short rule above the notes, which makes the note blocks easy to find.
3. Find the hard cases now: a note that continues onto the next page (it starts at the foot of one page and finishes on the next), two notes run together in one paragraph, and markers OCR read as ordinary numbers in the text (a year, a verse number, a page reference).

Step 3: convert each note

Work page by page, against the scan:

1. Select the text of note (١) at the foot of the page, without its number, and cut it (Ctrl+X).
2. In the body, select the plain marker (١) including its brackets and delete it. Leave the cursor where it was.
3. Select References > Insert Footnote. Microsoft's instructions: Word inserts a reference mark in the text and adds the matching mark at the bottom of the page.
4. Paste the note text after the footnote mark. Paste as text only (Keep Text Only) if the pasted note brings odd formatting with it.
5. Repeat for each note on the page, then delete the leftover separator line and any empty paragraphs.

For a note that continues onto the next page, join its two halves into one footnote. In a real footnote, Word itself decides where to split a long note across pages.

Step 4: set the number format and position

Open the Footnote and Endnote dialog (the small arrow in the corner of the Footnotes group on the References tab). There you choose the location, number format, starting number and numbering rule, and whether to apply the change to the whole document or the current section.

For Arabic digits, look for the style shown as ١، ٢، ٣ in the Number format list. Word's object model names it "Hindi Arabic", while "Arabic" there means Western digits (1, 2, 3). It also lists two Arabic letter styles. Which styles appear in the dialog can depend on the editing languages you have enabled, so check yours rather than relying on any screenshot.

Note from Microsoft's own support pages: in 2024 a Word update could change footnote numbers when numbering was set to restart each page, used an enclosed-number style and placed notes at the bottom of the page. Microsoft says it fixed this by a service change in May 2024. If you use that combination, spot-check the numbers after saving and reopening.

Step 5: make the notes read right to left

Real footnotes get their own paragraph style (Footnote Text). Set that style, not each note, to right-to-left with the Arabic font the recipient expects, and check the footnote reference style too. Do the same for the separator if it sits on the wrong side of the page: in Draft view you can edit footnote separators from the notes pane, though the exact menu path varies by Word version.

Illustrative example

This example shows how one page might look before and after. It is not taken from a real file.

Before: the body ends with "... ورواه البخاري (١) ومسلم (٢)." and two paragraphs follow at the foot of the page: "(١) صحيح البخاري، كتاب العلم." and "(٢) صحيح مسلم، المقدمة.", separated by a row of underscores.

After: the body reads "... ورواه البخاري¹ ومسلم²." with Word's own reference marks (shown here in Western digits for clarity). The two notes sit in Word's footnote area with the brackets and numbers removed, because Word now supplies the numbers. The underscores are gone.

Checklist before delivery

• Numbering choice (page, section or continuous) agreed with the recipient
• Printed page markers kept in the body, if citations to the print are needed
• Number of footnotes in Word equals the number of notes counted in the scan
• No plain (١)-style markers left in the body text
• No orphan note paragraphs left at page feet or after separators
• Notes that ran across two printed pages joined into one footnote
• Footnote Text style is right-to-left with the agreed font
• Number format matches the brief (١، ٢، ٣ or 1, 2, 3)
• A sample of pages read against the scan: each note is attached to the right word

Checking the result

Numbers alone don't prove anything: Word renumbers automatically, so a note attached to the wrong word still gets a perfect sequence. Check where each mark sits. Pick the densest pages, read the body word before each reference mark against the scan, and read each note against the printed note. Then count: Word's footnote total for a chapter should equal the count you made in Step 2. A difference means a merged note, a missing note or a stray number that OCR read as a marker.

Also search the body for "(" followed by a digit to find markers you missed, and look through the notes for text that belongs to the work itself. OCR sometimes takes the last line of the body into the note block, or the first line of a note into the body.

Limitations

• This is manual work. On a long, note-heavy book it takes time, and nothing in this guide checks that the editor's notes were read correctly. That needs a proofread against the scan.
• Matching Word's pages to the printed edition isn't possible in general. If page-level fidelity matters, keep page markers or deliver the scan alongside.
• Notes on notes, marginal glosses and multi-level apparatus (variant readings keyed by letter, references by number) don't fit Word's single footnote series well. Word can run footnotes and endnotes side by side, but a full critical apparatus may need layout software.
• Menu names and dialog options differ between Word for Windows, Mac and the web. Microsoft notes that converting between footnotes and endnotes isn't done in Word for the web.

ScribeTools converts scanned Arabic and Urdu books into editable Word files. See how it handles digitizing Arabic books and Arabic PDF to Word. Review the output, including the notes, before you rely on it.