Transcribing archival and handwritten sources means turning a manuscript into typed text a reader can trust, then citing it so that anyone can find the original. Two decisions shape the work: how closely the transcription follows the page (diplomatic, semi-diplomatic or normalised), and which version you cite. This guide is published by EdCitation, whose tools look references up in the publisher's record and never write them.
Every rule below was read at its source on 3 October 2026, from The National Archives, the Library of Congress, the Folger Shakespeare Library, the University of Cambridge, two scholarly editing bodies and the three style publishers. The last section quotes what EdCitation's free Cite a source and Verify references returned, faults included. For your own interviews, see transcription conventions in qualitative research.
What is the difference between diplomatic and normalised transcription?
A diplomatic transcription records the page as it stands, abbreviations, deletions and line breaks included; a normalised one keeps the words but tidies their presentation. Cambridge's handwriting course teaches the middle way, semi-diplomatic: original spelling, contractions expanded, so wᶜʰ becomes which with the supplied letters in italics (Zurcher, n.d.). Projects use the labels loosely:
| Kind | Keeps | Changes | Named in |
|---|---|---|---|
| Diplomatic | Spelling, abbreviations, deletions, insertions, line ends | Nothing, except through symbols | Cambridge; Folger; Glasgow |
| Semi-diplomatic | Spelling, capitals, u/v and i/j | Expands contractions, showing supplied letters | Cambridge; Folger |
| Normalised (Glasgow) | Spelling | Expands contractions, drops struck-out words, runs lines on | Wiggins (2013) |
| Regularised (Folger) | The words | Modern spelling: "Est horse lee" becomes "East Horsley" | Dingman (2017) |
The National Archives calls modernising the spelling "translating" and tells beginners to transcribe instead (The National Archives, n.d.-c).
One letter in two views
Glasgow's edition, Bess of Hardwick's Letters, shows its 234 letters in a normalised view by default, or a diplomatic one (Wiggins, 2013). Her letter to her servant James Crompe, dated 8 March [1560?] and held at the Folger as X.d.428 (83), shows what moves:
| Feature | Diplomatic view | Normalised view |
|---|---|---|
| Words she struck out | behend ‹of an olde› for work | behend for work |
| A raised-letter abbreviation | mʳ cauendyssh | master cauendyssh |
| Words added above the line | haue ^hym^ bundon yn ^a^ noblygacyon | haue hym bundon yn a noblygacyon |
| Spelling | sur Iames ys muche myslyked | sur Iames ys muche myslyked |
The edition strikes them through; we use Cambridge's ‹ ›. Neither view modernises spelling: only the edition's summary does, as "Sir James is much misliked for his religion".
How do you transcribe a handwritten document, step by step?
Copy letter by letter, mark every change, and write your rules down:
- Look for an edition first. Search the writer or document in EdCitation's Find sources; if a scholarly edition exists, check it against the original.
- Record the reference. Copy the catalogue reference with its own spacing and punctuation (The National Archives, n.d.-a), and keep an image beside your transcript, as Buchanan (2026) advises.
- Read letter by letter. Leave a letter you cannot read, or write your guess with a question mark, and return once you meet it elsewhere (The National Archives, n.d.-c). Count minims, single downstrokes, to tell i, m, n and u apart.
- Keep spelling, capitals and punctuation. Type what you see; never correct or modernise (Buchanan, 2026; Library of Congress, n.d.).
- Show every letter you supply, and mark what you cannot read, struck-out words and additions, with one set of symbols.
- Keep line breaks; Cambridge keeps lineation or marks line ends with a vertical line.
- Check against the original, then have someone else check. The MLA Committee on Scholarly Editions (2022) asks whether each transcription was compared with the original, not a photocopy, and checked in full by a second person.
- Write a statement of conventions.
Which symbols do archives and editors use?
Different ones, so choose a set, state it and keep to it: the one rule they share (Zurcher, n.d.).
| What you meet | The National Archives | Library of Congress | Cambridge course | Documentary editing guide |
|---|---|---|---|---|
| Letters you supply | Square brackets: th[a]t | Not on the pages we read | Italics: which | [roman] |
| A word you cannot read | A gap, or a guess with ? | [?], or s[????] for letters | {...}, a dot for each letter | [. . .], varying by edition |
| Words struck out | Not on the pages we read | Legible words in brackets | ‹xxx› or [xxx] | Angle brackets, or struck through |
| Words added between lines | Not on the pages we read | Typed in place, no caret | ┌xxx┐ or \xxx/ | ↑xxx↓, or ^xxx^ with a caret |
Sources: The National Archives (n.d.-c); Library of Congress (n.d.); Zurcher (n.d.); Kline and Perdue (2008).
The thorn, u and v, i and j
The guides disagree. The National Archives' exercises write ye as "the" and yt as "th[a]t". Buchanan (2026) reads "ye olde shoppe" as the old shop but transcribes it with a thorn. Cambridge replaces it with "th", saying so in its conventions. The National Archives warns that u and v, i and j interchange ("Iohn" is John, "vnto" is unto), and Cambridge says not to regularise them at first. Our view: a diplomatic transcription keeps them; a normalised one changes them only where its statement says so.
Money, places and dates
The National Archives (n.d.-b) writes the pound abbreviation "li" as £, and leaves ob (a halfpenny) and county abbreviations such as Northants unexpanded. Before 1752 the English year began on 25 March, so February 1750 is 1751 by modern reckoning, written 1750/51. Regnal years count from accession: 1 Elizabeth I runs from 17 November 1558 to 16 November 1559. For [sic] and old spelling in quotations, see quoting dialect and historical spelling.
How do you cite a manuscript in APA, MLA and Chicago?
Cite what you read: the manuscript, with its repository and reference, if you saw it in the archive; the edition or website if you read a transcription there.
| Style | Its own example | Uncertain date | Source |
|---|---|---|---|
| APA 7 | Frank, L. K. (1935, February 4). [Letter to Robert M. Ogden]. Rockefeller Archive Center (GEB Series 1.3, Box 371, Folder 3877), Tarrytown, NY, United States. | ca. 1937 | American Psychological Association (2019) |
| MLA 9 | Benton, Thomas Hart. Letter to Charles Fremont. 22 June 1847. John Charles Fremont Papers, Southwest Museum Library, Los Angeles. Manuscript. | 1955–60? | Rappaport (2018) |
| Chicago 18 | Deborah Dorman, “Psychoanalyzing the Penguin,” manuscript, Misc. Papers File, Chicago College Library, n.d. | n.d. | University of Chicago Press (n.d.-a) |
| The National Archives | The National Archives of the UK (TNA): CO 5/690, pp 143-149 | Not set out | The National Archives (n.d.-a) |
| Any of six styles | An edition with an ISBN, or an article with a DOI | From the record | Cite a source |
APA: archives live on the website
APA's archival guidance is not in the seventh-edition Publication Manual, only on the APA Style website (American Psychological Association, 2019). Give box and folder where there is no finding aid, "ca." for an estimated date, "(presumed)" after an author who is reasonably certain but not named, and square brackets for anything not on the document. A document on the open web takes the ordinary templates; a private letter in no archive is a personal communication, cited in the text only.
MLA, Chicago and The National Archives
In an MLA works-cited entry, Rappaport (2018) puts the collection in the container and the library, box and folder in the location; "Manuscript" at the end marks a handwritten letter. For a digitised manuscript, Gibson (2018) adds the website after the library and shelfmark, and leaves a name such as Thomas de Cantimpré uninverted.
Chicago's Q&A asks for author, title, kind of document, place, date and where you found it; its example is a note, so in author-date the year follows the name (University of Chicago Press, n.d.-a). Our Chicago history essay guide covers collections. Handwritten notes in a printed book are not a separate source: cite the book and add where the annotated copy is (University of Chicago Press, n.d.-b). The National Archives (n.d.-a) wants its reference, then the place within it: "SP 52/40, f 21r", for folio 21 recto.
The 1560 letter, cited three ways
These entries are ours, for the transcript read online. The letter is signed "E Seyntelo", her name as Lady St Loe; we use Bess of Hardwick because the edition files every letter under it.
- APA 7, listing the edition and naming the letter in the citation, as APA does for a collection: Wiggins, A., Bryson, A., Starza Smith, D., Timmermann, A., & Williams, G. (Eds.). (2013). Bess of Hardwick's letters: The complete correspondence, c.1550–1608. University of Glasgow. https://www.bessofhardwick.org. In the text: (Wiggins et al., 2013, Bess of Hardwick to J. Crompe, March 8, ca. 1560).
- MLA 9: Bess of Hardwick. Letter to James Crompe. 8 Mar. 1560? Cavendish-Talbot Manuscripts, Folger Shakespeare Library, Washington, DC, X.d.428 (83). Bess of Hardwick's Letters: The Complete Correspondence, c.1550–1608, edited by Alison Wiggins et al., U of Glasgow, 2013, www.bessofhardwick.org/letter.jsp?letter=100.
- Chicago 18 author-date: Wiggins, Alison, Alan Bryson, Daniel Starza Smith, Anke Timmermann, and Graham Williams, eds. 2013. Bess of Hardwick's Letters: The Complete Correspondence, c.1550–1608. University of Glasgow. https://www.bessofhardwick.org. In the text: (Wiggins et al. 2013, letter 100).
Can software transcribe handwriting for you?
It can draft, not finish. Schoen and Saretto (2022) trained the Kraken OCR engine on one early-fifteenth-century scribe's Middle English hand; their abstract reports 97 percent accuracy in training but 27 to 86 percent on manuscripts it had not seen, and the software needed diplomatic transcriptions where medievalists usually make semi-diplomatic ones. Check every line of a machine transcript against the image. Consent and data questions are in AI transcription tools and research ethics.
What does EdCitation do for a transcription project?
EdCitation does the published half of the job, and does it best, because every reference it gives comes from a publisher's or library's record and none is written, which no chatbot can promise. EdCitation never writes any part of anyone's work: the reading, the brackets and the argument stay yours. Cite a source builds APA 7, MLA 9, Chicago 18 author-date, Harvard, IEEE and Vancouver entries from a DOI, title, PubMed ID, ISBN or web address, screened for retraction; Verify references checks a whole list. Both are free with no account. Our runs of 3 October 2026, quoted exactly:
Books and articles
For ISBN 9780813927275, Cite a source returned in APA: "Kline, M.-J., & Perdue, S. H. (2008). A guide to documentary editing (3rd ed.). University of Virginia Press." Right as it stands. For DOI 10.1353/hms.2003.a383339 its APA title ended "Political whigs, religious whigs, and Jacobites": Whigs, a party's name, keeps its capital. For Schoen and Saretto (2022) it kept "Medieval" capitalised, which APA's sentence case lowers.
A page in a digital edition
For the letter's address, the page gave the site's title, no author and no date: "Bess of Hardwick’s Letters | The complete correspondence, 1550-1608. (n.d.)." Cite a source lets you correct those before the entry is built. With the letter as title and Bess of Hardwick as author, the APA entry began "Hardwick, B. of. (n.d.). Letter to James Crompe." The name was inverted as if Hardwick were a surname, and a 1560 date did not carry into the entry, though a 2013 date did. Fix both by hand, or cite the edition as above.
Verify references on manuscripts
We pasted seven entries, among them APA's Frank letter and our APA entry for Bess's original letter. Five came back verified, the Kline and Perdue book with "Title, author, edition and year match a library catalogue record". Both manuscripts came back "could not check", never "not found", which is right for letters with no publisher's record. The reason given, a title "in romanised form or in translation", did not fit them; act on the verdict and check such entries in the archive's catalogue, as in how to check a reference is real.
Find sources
"diplomatic transcription manuscripts" returned 5,770 results, led by published diplomatic transcriptions such as a 2003 Hume Studies edition of a Hume pamphlet; one of the first eleven was off topic, a brain-tumour paper with "diplomatic" in its title. For a dissertation, Pro at $8 a month adds References from a file; see pricing.
Quick questions
What is a diplomatic transcription?
A diplomatic transcription reproduces a manuscript as closely as type allows: original spelling, abbreviations, deletions, insertions and line breaks.
What is the difference between semi-diplomatic and normalised transcription?
A semi-diplomatic transcription keeps the spelling but expands abbreviations, showing the supplied letters. Normalised varies by project: some keep the spelling and tidy the layout, others modernise it, so say which you mean.
How do I show a word I cannot read?
Use the symbol your statement of conventions names, such as [?] in the Library of Congress's rules, and put a guessed reading in square brackets with a question mark.
How do I cite a manuscript in APA?
Give the author, the date, a bracketed description such as [Letter to Robert M. Ogden], then the collection with box and folder, the repository and its location.
Can EdCitation cite a manuscript from an archive?
EdCitation's Cite a source cites the edition or digitised page from its ISBN, DOI or web address. A manuscript in a box has no record to build from, so write that entry by hand.
References
- American Psychological Association. (2019). Archival documents and collections. APA Style. https://web.archive.org/web/20260930141818/https://apastyle.apa.org/style-grammar-guidelines/references/archival
- Buchanan, C. (2026). Paleography: Interpreting handwriting in genealogical research: Transcription and spelling [Research guide]. Library of Congress. https://guides.loc.gov/paleography/transcription-spelling
- Dingman, P. (2017, May 2). Okay, but what does it mean, or how do you regularize an early modern transcription? The Collation. Folger Shakespeare Library. https://www.folger.edu/blogs/collation/okay-but-what-does-it-mean/
- Gibson, A. (2018, April 25). Citing artifacts in a digital archive. MLA Style Center. https://style.mla.org/citing-work-in-digital-archives/
- Kline, M.-J., & Perdue, S. H. (2008). A guide to documentary editing (3rd ed.). University of Virginia Press; Association for Documentary Editing. https://gde.upress.virginia.edu/05-gde.html
- Library of Congress. (n.d.). Transcription: Basic rules. By the People. https://web.archive.org/web/20260804215056/https://crowd.loc.gov/get-started/how-to-transcribe/
- MLA Committee on Scholarly Editions. (2022, May 4). Guidelines for editors of scholarly editions. Modern Language Association. https://www.mla.org/Resources/Guidelines-and-Data/Reports-and-Professional-Guidelines/Guidelines-for-Editors-of-Scholarly-Editions
- The National Archives. (n.d.-a). Citing records in The National Archives. https://www.nationalarchives.gov.uk/help-with-your-research/discovery-help/citing-records-national-archives/
- The National Archives. (n.d.-b). Palaeography: Quick reference. https://web.archive.org/web/20231208023852/https://www.nationalarchives.gov.uk/palaeography/quick_reference.htm
- The National Archives. (n.d.-c). Palaeography: Where to start. https://web.archive.org/web/20230307053204/https://www.nationalarchives.gov.uk/palaeography/where_to_start.htm
- Rappaport, J. (2018, April 25). A guide to citing materials from physical archives and collections. MLA Style Center. https://style.mla.org/citing-materials-physical-archives/
- Schoen, J., & Saretto, G. E. (2022). Optical character recognition (OCR) and medieval manuscripts: Reconsidering transcriptions in the digital age. Digital Philology: A Journal of Medieval Cultures, 11(1), 174–206. https://doi.org/10.1353/dph.2022.0010
- University of Chicago Press. (n.d.-a). Citation, documentation of sources (Q&A No. 140). The Chicago Manual of Style Online. https://www.chicagomanualofstyle.org/qanda/data/faq/topics/Documentation/faq0140.html
- University of Chicago Press. (n.d.-b). Citation, documentation of sources (Q&A No. 404). The Chicago Manual of Style Online. https://www.chicagomanualofstyle.org/qanda/data/faq/topics/Documentation/faq0404.html
- Wiggins, A. (2013, April). Transcription policy. In A. Wiggins, A. Bryson, D. Starza Smith, A. Timmermann, & G. Williams (Eds.), Bess of Hardwick's letters: The complete correspondence, c.1550–1608. University of Glasgow. https://www.bessofhardwick.org/background.jsp?id=188
- Wiggins, A., Bryson, A., Starza Smith, D., Timmermann, A., & Williams, G. (Eds.). (2013). Bess of Hardwick's letters: The complete correspondence, c.1550–1608. University of Glasgow. https://www.bessofhardwick.org
- Zurcher, A. (Ed.). (n.d.). Basic conventions for transcription. English Handwriting 1500–1700: An Online Course, University of Cambridge. https://www.english.cam.ac.uk/ceres/ehoc/conventions.html