Is the text Latin or early Portuguese?

Almost always Latin, but a notarial Latin full of Romance spellings, place-names and personal names. The transcription keeps those forms: it does not correct mediadade to medietatem or turn an iben patronymic into a Latin genitive, because those are the details linguists and genealogists want.

How is the Era date handled?

The numerals stand as written, superscript ordinal letters included, and the converted year can be added in a note: Era 920 is AD 882, Era 1036 is AD 998. The day stays in Roman form, so vi kalendas abriles is kept with 27 March supplied beside it.

Can it tell the two ti ligatures apart?

Yes, where the resolution allows. The assibilated form has a distinct upper loop, and the transcription can mark it so that an editor studying the pronunciation of gratia or sententia can find every instance.

What happens with the chrismon, signa and witness columns?

The chrismon is recorded as a symbol rather than as letters, drawn signa are noted where they interrupt the text, and each witness column is read top to bottom with test. and conf. kept, so the order and grouping of subscribers survives.

Does it work on cartulary copies such as the Livro Preto of Coimbra or the Liber Fidei of Braga?

Yes. Those copies are in Caroline or early Gothic minuscule rather than Visigothic, and the change of script is handled; what no model can do is restore what the copyist misread. Where a copy stands beside an original, transcribe both and compare.

How does it cope with the mixed hands of the twelfth century?

Many charters between about 1100 and 1200 combine Caroline letterforms with Visigothic abbreviations or the old looped t. Each letter is read on its own merits, and the scope note can record which system the abbreviations follow, which is itself evidence for dating.

Portuguese · c. 800–1200

Read the Visigothic charters of early Portugal, letter by letter

Between the ninth century and about 1200 the scribes of Braga, Guimarães, Coimbra and Lorvão wrote a script in which a looks like u, t carries a loop that turns it into a small a, and the letter that became ç was still a z. Evidano transcribes these sheets line by line, keeps the Era date and the abbreviation signs as written, and separates the witness columns and signa from the body of the act.

Upload a page See the samples

Sample pages

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

Upper left part of the 882 foundation charter of the church of Lardosa, tall Visigothic cursive in long lines on a creased parchment sheet — Sample 1. Foundation charter of the church of Lardosa, near Paredes, dated 27 March 882 (Era 920), the earliest dated document in Visigothic script from Portugal. Lisbon, Arquivo Nacional da Torre do Tombo, PT/TT/MSPC/01/1. The detail shows the chrismon, the invocation and the address to Saints Peter, Paul and Michael in a tall documentary cursive.

Opening lines of a 998 sale charter to the monastery of Lorvão, a large sloping Visigothic cursive with a chrismon and drawn signa in the roboratio line — Sample 2. Sale by Ioacino, his wife Guntili and their sons to Abbot Beniamin and the monks of Lorvão of a vineyard at Vila Cova on the Mondego, 21 March 998 (Era 1036). Torre do Tombo, PT/TT/MSML/CE1/14, f. 1r. A looser, larger cursive than the Lardosa sheet; the drawn signa of the sellers interrupt the word rouorauimus in the roboratio line at the foot of the crop.

First lines of a 1053 charter of the monastery of Paço de Sousa in a small, precise Visigothic cursive with very long ascenders — Sample 3. Judicial record from the Benedictine monastery of São Salvador de Paço de Sousa, 1053, catalogued as a “sentença muito antiga”. Torre do Tombo, PT/TT/MSPS/005/0001. The opening address to the saints of the house names Abbot Gomize; the hand is small and disciplined, its ascenders reaching well up into the line above.

Text column of the Lorvão Apocalypse folio 49 beside a miniature of Christ holding seven stars, in a rounded late twelfth-century minuscule — Sample 4. The message to the church of Ephesus (Revelation 2) in the Lorvão Apocalypse, a Beatus commentary copied at Lorvão by the scribe Egeas in 1189. Torre do Tombo, Ordem de Cister, Mosteiro de Lorvão, livro 43, fl. 49. Lorvão had by then abandoned Visigothic script: the column is a rounded Caroline-derived minuscule beside a miniature in the older Beatus iconography.

What a Visigothic page from Portugal looks like

Visigothic script was the ordinary writing of the western Iberian peninsula from the eighth century, and in the land that became Portugal it outlasted its use elsewhere: the foundation charter of the church of Lardosa of 882, the earliest dated Portuguese document, is written in it, and northern charters keep Visigothic letterforms after 1200. Two registers coexist. The book minuscule is round, upright and evenly spaced, the hand of cartularies and of the few surviving liturgical codices. The charter cursive is tall and narrow, its ascenders stretching far above the line, its letters leaning into each other in long chains of ligatures; it is this cursive that fills the monastic fonds of the Torre do Tombo.

The letters that matter are the ones that do not resemble their modern descendants. The a is open at the top and reads as u; the t carries a loop to the left of its stem and is taken for a; the g hangs below the line like a q with a curled tail; the z is a hooked shape that later scribes reinterpreted as c with a cedilla, which is where ç comes from. Two ligatures for ti separate a plain syllable from an assibilated one, s and r are nearly identical, and the sign for 40 is a barred X. Abbreviation runs on superscript letters, a stroke through the descender of p for per, contractions such as nsr for noster, and the nasal bar.

Layout is formulaic. A charter opens with a chrismon or the word Christus, an invocation and an address to the saints of the church concerned; the dispositive text runs in long lines across the sheet; a sanction clause threatens anathema; the dating clause gives a Roman calendar day and a year in the Era Hispanica, thirty-eight years ahead of the Christian year; and the roboratio is followed by drawn signa and by columns of witnesses marked test. or conf. The Latin is full of Romance intrusions, and Lorvão and Coimbra acts carry Mozarabic names of the type iben Zaton. By the later twelfth century Caroline minuscule had taken over in the cathedral chapters, and the Lorvão Apocalypse of 1189 is written in the rounded book hand that replaced the old script.

Why ordinary OCR struggles here

a, u, t and the ti ligatures

Optical models trained on print read the open Visigothic a as u and the looped t as a, so that testamento comes back as aesaamenau. The two ti ligatures, which record whether the syllable was pronounced ti or tsi, look like decorative variants and are flattened into one. A model that has learnt the script reads them as letters and can note which form the scribe used.

Signs that are not letters

The superscript -is and -us marks, the stroke through the p, the abbreviation for que and the flourish on a final s all sit at letter height, so they are read as extra letters or dropped altogether. The same happens to the barred X for 40 in a date, and a whole Era hangs on that one stroke.

Dates, numerals and formulae

The year is given as an Era with superscript ordinal letters, the day is counted backwards from the kalends, nones or ides, and sums are in Roman numerals inside running text. A transcription that reproduces the numerals as written and adds the converted year in a note is far more useful than one that silently modernises either.

Signa, columns and later hands

Below the text come drawn signa, sometimes splitting the word roboramus itself, and several narrow columns of witness names in more than one hand. Archival stamps, dorsal notes in Gothic or modern cursive and the holes left by seals lie on the same sheet, and a line-by-line reading has to keep them apart.

Who works with this material

Editors of cartularies and monastic fonds

The Lorvão, Pedroso, Paço de Sousa and Guimarães collections were printed in the nineteenth century, often from copies and with gaps. Re-editing them from the originals starts with a transcription that keeps the scribe’s spelling, marks every expansion and records where signa and witness columns fall, so the new text can be collated against Portugaliae Monumenta Historica rather than retyped.

Historians of settlement and lordship

Presúria clauses, boundary descriptions and lists of appurtenances are where the history of the Entre-Douro-e-Minho and the Mondego frontier is written. A searchable transcription that leaves place-names in their charter spelling lets a researcher follow one villa or one family across dozens of acts.

Linguists and onomasticians

These Latin texts are shot through with early Galician-Portuguese forms and with Mozarabic and Arabic personal names. Corpus builders need the Romance spellings and the iben patronymics exactly as written, with the ti ligature distinction preserved because it records pronunciation.

Teachers of Iberian palaeography

A transcription set beside the image makes a Visigothic alphabet exercise self-checking: students see where they read a as u or missed a superscript -is, and the Era conversion can be practised on real dating clauses.

Getting the best transcription

Step 1

Name the script and the dating system

Tell the model that the text is Latin in Visigothic cursive or minuscule, that years run in the Era Hispanica and that numerals may use a barred X for 40 and superscript ordinal letters. That instruction alone stops it modernising dates or reading the 40 sign as a ten.

Step 2

Fix the conventions for letters and expansions

Ask for u and v, i and j to be kept as the scribe wrote them, for the Visigothic z to be transcribed as z rather than ç, and for expansions to be shown in brackets. If you want the two ti ligatures distinguished, say so; otherwise both come back as ti.

Step 3

Try the foot of the sheet first

Run a sheet whose lower half is witness columns and signa. If the columns come back in order, each test. and conf. is kept and the drawn signa are treated as non-text, the running text above will be straightforward; if the columns are merged into one line, add an instruction on reading order.

Step 4

Keep the lineation and note the anomalies

Transcribe line by line so every reading can be located, and ask for doubtful letters to be flagged rather than guessed. Label cartulary copies and later dorsal notes as such, so that a twelfth-century Caroline copy of a tenth-century act is not mistaken for the original.

Frequently asked questions

Is the text Latin or early Portuguese?: Almost always Latin, but a notarial Latin full of Romance spellings, place-names and personal names. The transcription keeps those forms: it does not correct mediadade to medietatem or turn an iben patronymic into a Latin genitive, because those are the details linguists and genealogists want.
How is the Era date handled?: The numerals stand as written, superscript ordinal letters included, and the converted year can be added in a note: Era 920 is AD 882, Era 1036 is AD 998. The day stays in Roman form, so vi kalendas abriles is kept with 27 March supplied beside it.
Can it tell the two ti ligatures apart?: Yes, where the resolution allows. The assibilated form has a distinct upper loop, and the transcription can mark it so that an editor studying the pronunciation of gratia or sententia can find every instance.
What happens with the chrismon, signa and witness columns?: The chrismon is recorded as a symbol rather than as letters, drawn signa are noted where they interrupt the text, and each witness column is read top to bottom with test. and conf. kept, so the order and grouping of subscribers survives.
Does it work on cartulary copies such as the Livro Preto of Coimbra or the Liber Fidei of Braga?: Yes. Those copies are in Caroline or early Gothic minuscule rather than Visigothic, and the change of script is handled; what no model can do is restore what the copyist misread. Where a copy stands beside an original, transcribe both and compare.
How does it cope with the mixed hands of the twelfth century?: Many charters between about 1100 and 1200 combine Caroline letterforms with Visigothic abbreviations or the old looped t. Each letter is read on its own merits, and the scope note can record which system the abbreviations follow, which is itself evidence for dating.

Related scripts and pages

Portuguese Gothic Documentary Hands, c. 1200–1400The royal chancery, notarial and book hands that followed, including the first documents in Portuguese.Iberian Visigothic Manuscripts and Charters, c. 700–1200The same script in the Beatus codices, glosses and charters of León, Castile and Navarre.Insular, Visigothic and Beneventan Latin, c. 650–1100Visigothic minuscule alongside the other national scripts of the early medieval West.All Portuguese hands and document typesEvery Portuguese domain from Visigothic charters to colonial correspondence.Image to TextSearch all supported languages, scripts and document types.

Transcribe your Visigothic charters

Upload a scan, choose Portuguese Visigothic Manuscripts and Charters as the domain, and check the dating clause and witness columns on one sheet first.

Start transcribing Create a free account