Site Logo

Spanish · c. 700–1200

Read Visigothic script from the Iberian scriptoria

Visigothic minuscule writes a like u, hides the sound of ti inside a ligature, and dates every charter by an era that runs thirty-eight years ahead of the Christian year. Evidano transcribes the codices of San Millán, León and Toledo and the cursive charters of the northern kingdoms line by line, expands only what you tell it to expand, and lets you check each reading against the folio.

Sample pages

Real pages from public collections, shown beside their transcriptions. Pages whose transcription is still being checked are marked.

Parchment page in Visigothic minuscule with many small interlinear glosses and a twelve-line marginal gloss in the lower right margin.
Sample 1. Folio 72r of the Códice Emilianense 60, Madrid, Real Academia de la Historia, a homiliary written at San Millán de la Cogolla in Visigothic minuscule around the year 1000. Interlinear glosses in a smaller hand sit above the Latin, the rubric of a homily attributed to St Augustine is set in capitals, and the long Romance gloss in the lower right margin is the best-known early text in a Hispanic vernacular.

Real Academia de la Historia, Madrid, Cod. 60, f. 72r, c. 1000 · Source · Public domain

Transcription

The twelve-line Romance gloss in the lower right margin only, in the normalised reading printed on Spanish Wikisource: word division and capitals are the editor’s, the interlinear Latin and Romance glosses and the Latin homily text are not included.

Con o aiutorio de nuestro
dueno Christo, dueno
salbatore, qual dueno
get ena honore et qual
duenno tienet ela
mandatione con o
patre con o spiritu sancto
en os sieculos de lo siecu
los. Facanos Deus Omnipotes
tal serbitio fere ke
denante ela sua face
gaudioso segamus. Amen.

Transcription from Glosas Emilianenses, via Wikisource (Spanish) (Public domain).

Black-and-white photograph of a narrow column of early Visigothic minuscule with a hole near the top and tiny marginal notes to the right.
Sample 2. Right-hand column of London, British Library, Egerton MS 1934, f. 2r, the oldest fragment of the Chronicle of 754, copied in the eighth or ninth century in an early, upright Visigothic minuscule. A hole in the parchment has taken several lines; the passage records the death of the caliph al-Walid and the rule of ʿAbd al-ʿAzīz over Spain, and a later reader has added tiny notes in the margin.

British Library, Egerton MS 1934, f. 2r, c. 750–900 · Source · Public domain

Transcription

The right-hand column only, which carries the end of chapter 78 and the start of chapter 79 in Mommsen’s numbering, from the death of Walid to the rule of ʿAbd al-ʿAzīz. Given as Mommsen’s edited text with his sentence punctuation rather than the manuscript line breaks; the square brackets he uses for words lost in the hole of this leaf (his manuscript A) have been removed and the OCR errors of the digitised volume corrected against the folio. The column ends at “seditione”; the rest of the chapter is supplied in the edition from other manuscripts. The left-hand column and the marginal notes are not included.

Ulit mortuo Zuleiman sanguine frater honorifice secundum expositum fratri succedit in regnum. Regnat annos tres.
hic infestus Romanie fratrem non dissimili matre progenitum Muzilima nomine cum centum milia armatorum ad delendum Romaniam mittit.
hic Asiam bello impeditam gladio simul cum igne finibit deceptam.
deinde Constantinopolim properans, dum periclitari se diversis necessitatibus Muzilima prespicit, ulterius principis iussu non nimium feliciter repedavit.
Per idem tempus in era DCCLIII, anno imperii eius VIIII, Arabum LXLVII, Abdellazis omnem Spaniam per annos tres sub censuario iugo pacificans, cum Spalim divitiis et honorum fascibus cum reginam Spanie in coniugio copulatam vel filias regum hac principum pelicatas et inprudenter distractas extuaret, seditione

Transcription from Continuatio Hispana a. DCCLIV, ed. Theodor Mommsen, MGH Auctores antiquissimi XI, Chronica minora II (Berlin, 1894), pp. 355–356, via Internet Archive (Public domain).

Manuscript page with a column of short glossary entries in Visigothic minuscule, the green and red interlaced initial cut by the left edge of the frame at the foot.
Sample 3. Folio 72v of the Glossarium latinum, Madrid, Real Academia de la Historia, Cod. 46, written at San Millán and dated to June 964. The interlaced initial that opens the letter h sits at the foot of the left column, cut by the edge of the frame, and the column beside it lists lemmata beginning hab-, hac- and hec-, each followed by its definition, in a compact Riojan Visigothic minuscule with a wavy nasal stroke on nearly every line.

Real Academia de la Historia, Madrid, Cod. 46, f. 72v, 964 · Source · Public domain

Transcription being verified

The glossary has not been edited; the manuscript is described in Elisa Ruiz García, Catálogo de la sección de códices de la Real Academia de la Historia (1997), no. 46, and the entries on this folio are being read against the Corpus glossariorum latinorum.

What the model is told to watch for: Expect Visigothic minuscule and cursive, whose a, g, t, z, and ligatures differ sharply from Caroline or later Gothic forms. Distinguish the Visigothic z associated with the development of cedilla (ç), regional forms of ti, and abbreviations whose strokes may resemble letters. Note superscript characters, contraction marks, punctus systems, chrismons, and formulaic signs in royal, monastic, and legal documents.

Manuscript page opening with four lines of red and blue nested capitals, a large interlaced initial on the left and Visigothic minuscule text below.
Sample 4. Folio 178r of the Códice de Roda, Madrid, Real Academia de la Historia, Cod. 78, copied c. 990, the opening of the Chronicle of Alfonso III in its Rotensis version. The title runs across four lines of nested red and blue capitals, an interlace initial frames the first paragraph, and the first era date, DCCX, stands in red inside the text.

Real Academia de la Historia, Madrid, Cod. 78, f. 178r, c. 990 · Source · Public domain

Transcription being verified

The Rotensis text was first printed by Manuel Gómez-Moreno in the Boletín de la Real Academia de la Historia 100 (1932) and re-edited in Crónicas asturianas (Oviedo, 1985); both editions remain in copyright, so the folio is being transcribed for checking against them.

What the model is told to watch for: Expect Visigothic minuscule and cursive, whose a, g, t, z, and ligatures differ sharply from Caroline or later Gothic forms. Distinguish the Visigothic z associated with the development of cedilla (ç), regional forms of ti, and abbreviations whose strokes may resemble letters. Note superscript characters, contraction marks, punctus systems, chrismons, and formulaic signs in royal, monastic, and legal documents.

Lithographed facsimile of fifteen lines of sloping Visigothic cursive from a twelfth-century royal charter, ending with an era date.
Sample 5. Donation by Alfonso VII and the empress Berenguela of the church and coto of Santiago de Castro to the monastery of San Martín Pinario in Compostela, April 1134, reproduced from plate XXXIX of Muñoz y Rivero’s Paleografía visigoda (1919). The royal chancery still wrote Visigothic cursive here: sloping, linked letters, the ꝸ sign for -us in cuius and omnibus, and the era at the foot.

Muñoz y Rivero, Paleografía visigoda, plate XXXIX (facsimile), April 1134 (facsimile 1919) · Source · Public domain · Facsimile plate from Jesús Muñoz y Rivero, Paleografía visigoda (Madrid, 1919)

Transcription

The whole plate, in Muñoz y Rivero’s own reading of the fragment he reproduced: a normalised transcription that expands every abbreviation silently, prints j for consonantal i (cujus, Jacobi) and adds modern punctuation. The plate breaks off after “que eidem in” and resumes with the dating clause; that gap belongs to the facsimile, not to the parchment.

Ego Adefonsus Dei gratia Hispanie imperator comitis domni Raymundi regineque domne Urrace filius et uxor mea imperatrix domina Beringaria et filii mei et omnis vox mea pro remedio anime mee meorumque parentis. Vobis abbati domno Petro et ceteris monasterii sancti Martini de Pinu monachis, ad utilitatem et tuicionem ipsius monasterii cujus ecclesia sita est in urbe Compostella, cauto ecclesia sancti Jacobi apostoli de castro cum villa vocitata Busiana cum integra sua feleclesia et adjunctionibus suis omnibusque prestacionibus, exitibus, bonis appendiciis et directuris que eidem in
Facta scriptura cauti sub era millesima CLXXII et quot XII Kalendas Maii.

Transcription from Jesús Muñoz y Rivero, Paleografía visigoda (Madrid, 1919), p. 156, transcription of plate XXXIX, via Internet Archive (Public domain).

What Visigothic writing looks like

From the eighth century until the early twelfth, the Christian north of the Iberian Peninsula and the Mozarabic communities of the south wrote in a national script descended from late Roman cursive rather than from the Caroline minuscule of the Frankish world. Palaeographers call it Visigothic, although almost everything that survives postdates the Visigothic kingdom: the Beatus commentaries, the Bibles of León, the glossaries and homiliaries of San Millán and Silos, and thousands of charters kept in monastic archives. It exists as a calligraphic book hand with short, rounded bodies and as a cursive charter hand whose letters lean, stretch and link. The Council of Burgos in 1080 and the arrival of Cluniac monks brought in Caroline minuscule, the letra francesa, and by the 1130s only conservative Galician and Leonese scribes, and the Mozarabs of Toledo, still wrote the old way.

The letters that mark the script are few but decisive. The a is open at the top and easily read as u; the g looks like a q with a long tail, or like the figure 9; the t carries a bowl to the left of its stem, and when t stands before a vowel in an assibilated syllable the scribe uses a distinct ti ligature that a Caroline-trained eye reads as ci or gi. The z drops a tail below the line, and that tailed z, used for the sound of Romance ç, is the ancestor of the cedilla. Tall I opens words and stands for the consonant, e climbs into ligatures with r, s and t, and the round s of the book hand alternates with the long s of the cursive.

Abbreviation follows its own conventions. A wavy stroke over a vowel drops a nasal; per is a p with a bar through the descender; -us and -is are shown by hooks and by the sign ꝸ; the nomina sacra and words such as noster and omnis are contracted under a bar, and superscript vowels stand for whole syllables. Charters open with a chrismon, the monogram of Christ drawn as a large sign, and close with a dating clause in the Hispanic era, where the X may carry a tail that makes it worth forty, followed by confirmations and witness signa in crosses and personal monograms. Book pages run in two columns of thirty or so lines, with headings in mixed uncial and capital letters, often nested one inside another and alternating red and blue, and later readers added interlinear and marginal glosses in Latin, Romance and, at San Millán, Basque.

Why ordinary OCR struggles here

a, u and the ti ligature

Because the Visigothic a is open, sequences such as aut, au and ua become a run of near-identical strokes, and the assibilated ti ligature resembles no letter that a Latin OCR model has seen. Engines trained on Caroline hands or on print resolve them by guessing from a dictionary, which silently rewrites the Latin.

Abbreviations that look like letters

The ꝸ sign for -us, the hooked -is, the barred p for per and the superscript vowels of the cursive sit inside the word rather than above it, so a reader must recognise them before expanding them. Generic OCR treats them as stray letters or drops them, and the case endings that carry the syntax vanish.

Display capitals and glosses on the same page

A rubric in nested capitals across the width of the page, a lacework initial that fills a column, interlinear glosses in a smaller cursive and marginal notes by a later hand all compete with the main text. A transcription is only useful if it says which is which and keeps the reading order of each.

Cursive charters and their dates

In the documentary hand ascenders stretch, letters join in long ligatures and the whole line slopes. The dating clause mixes Roman calendar days, the era count and abbreviated numerals, and an error of one stroke moves a charter by decades. Witness lists with crosses and monograms add signs that are not text at all.

Who works with this material

Editors of cartularies and monastic archives

Diplomatic editions of the charters of Sahagún, Celanova, San Millán or Santo Toribio depend on a first transcription that keeps the scribe’s spelling, marks expansions and separates the notarial subscription from the body of the act. Line-true output lets the editor collate against the older printed collections.

Historians of early Romance

The glosses of San Millán and Silos, the Nodicia de kesos and the vernacular words buried inside Latin charters are the earliest evidence for Castilian, Leonese and Navarro-Aragonese. Linguists need the exact letters, including the tailed z and the open a, rather than a normalised Latin.

Liturgists and Mozarabic studies

The Old Hispanic rite survives chiefly in Visigothic-script antiphoners, orationals and homiliaries. Reconstructing a formulary means matching hundreds of rubrics and incipits across manuscripts, which goes far faster once every page is searchable text.

Teachers of Spanish palaeography

Courses on letra visigótica still use the plates of Muñoz y Rivero. A transcription placed beside a modern photograph gives students an answer key for every ligature and lets a class compare a Riojan book hand with a Galician charter hand.

Getting the best transcription

    Step 1

    Name the script and the type of hand

    Say whether the page is a book minuscule or a documentary cursive, and where it was written if you know, because the Leonese, Riojan, Toledan and Galician forms of g, t and the ti ligature differ. The model then reads the strokes against the right alphabet instead of a generic medieval one.

    Step 2

    Fix the expansion and glyph rules

    Choose whether ꝸ, the barred p and the nasal stroke are expanded silently, expanded in brackets or left as signs, and whether the tailed z is to be printed as z or ç. State that u and v, i and j should follow the manuscript.

    Step 3

    Ask for the layout to be labelled

    Have rubrics, interlinear glosses and marginalia returned as separate blocks, in the order they occur, with the main text line by line. For a charter, ask for the chrismon, the dating clause and each subscription to be marked so the diplomatic structure stays visible.

    Step 4

    Check the dating clause and numerals first

    Before accepting a batch of charters, verify the era, the calendar day and the regnal notes on one document and compare them with the printed catalogue. If the model reads the crossed X as forty and the Kalends correctly there, the rest of the archive will follow.

Further reading on this hand

The palaeography guides the in-app selector points to for this domain, if you want to check a transcription against the standard references.

  • Spanish Handwriting History

    This Brigham Young University guide outlines the major historical handwriting systems used for Spanish and its Iberian predecessors. It describes Visigothic writing, its chronological range, and characteristic forms including the Visigothic z associated with the cedilla. The guide then traces the replacement of Visigothic forms by later Gothic and humanistic scripts. Its early section provides direct orientation for Spanish or Romance material written in Visigothic manuscripts and charters.

Frequently asked questions

Will the transcription keep the Visigothic letterforms or modernise them?
It gives a diplomatic transcription by default: letters are recorded as Latin letters, so an open a is a and the ti ligature is ti, but nothing is normalised. If you want Romance ç for the tailed z, or classical spellings, ask for them in the prompt and the rule is applied throughout.
How does it deal with the era dates?
The Hispanic era is transcribed as written, including the crossed X for forty, and can be converted to the Christian year on request with the thirty-eight-year offset applied. The conversion is reported separately so that the original numeral stays in the text.
Can it read the cursive charter hand as well as the book hand?
Yes. The cursive has longer, more slanted strokes and many more ligatures, and it is where most errors happen. Test on the least tidy charter in your set; a clean result there means the calligraphic pages will be straightforward.
What about the Romance and Basque glosses?
Glosses are returned as separate items keyed to the word they gloss, in the language in which they are written. They are not merged into the Latin text and they are not translated unless you ask.
Does it recognise the chrismon, signa and monograms?
It marks them as non-textual signs with a short description, for example a chrismon at the head of the charter or a cross before a witness name, instead of trying to read them as letters.
How are damaged or faded passages handled?
Lost text is marked with a gap of the estimated length and uncertain readings are flagged. If a printed edition exists, as for the Chronicle of 754 or the charters of Alfonso VII, you can paste its readings alongside for comparison.

Related scripts and pages

Transcribe your Visigothic folios and charters

Upload a scan, choose Iberian Visigothic Manuscripts and Charters as the domain, and test one charter or one glossed page before running the rest.