Tagalog
Read Tagalog across four centuries of script and spelling
Baybayin, the pre-colonial syllabary, could not show a consonant at the end of a syllable until Spanish printers added a cancelling mark in the 1620s, and three centuries later Tagalog print still spelled words the Spanish way — c, qu, a tilde over ng̃ — until the 1940 Abakada reform. Evidano reads baybayin as the syllabic script it is, keeps pre-reform spelling as printed rather than modernising it, and reads a present-day Filipino form or sign with the same care.
Sample pages
Real pages from public collections, shown beside their transcriptions. Pages whose transcription is still being checked are marked.
Source · CC BY-SA 3.0 · Gubernatoria
Transcription being verified
A page-keyed facsimile and transcription of the complete 1593 edition was published by Edwin Wolf II, Doctrina Christiana: The First Book Printed in the Philippines, Manila, 1593 (Washington: Library of Congress, 1947); no openly licensed transcription matched to this exact opening has been located.
What the model is told to watch for: Produce a diplomatic transcription in Tagalog.
Biblioteca Nacional de España, 1794 · Source · Public domain
Transcription being verified
This is the title leaf; the dictionary’s paired Spanish-Tagalog and Tagalog-Spanish entries begin further into the volume. The complete work is digitised by the Biblioteca Nacional de España, but no separate transcription has been located.
What the model is told to watch for: Produce a diplomatic transcription in Tagalog.
Source · CC BY-SA 4.0 · Andre Ryan Maceda
Transcription being verified
Signage text with no separately published transcription; the wording has been checked directly against the photographed marker.
What the model is told to watch for: Produce a diplomatic transcription in Tagalog.
Philippine Statistics Authority, c. 2018 · Source · Public domain · Philippine Information Agency
Transcription being verified
An official specimen with placeholder data; no transcription applies beyond the printed field labels visible on the card.
What the model is told to watch for: Produce a diplomatic transcription in Tagalog.
Bartlett Collection on the Philippines and the East Indies, American Philosophical Society Library; microfilm digitised by the University of Michigan, 1914 · Source · Public domain
Transcription
The whole leaf — the display title and the three opening quatrains — as proofread and validated on Tagalog Wikisource (page quality 4). Line breaks, the accented ú of búhay and the chapbook’s habit of writing an enclitic with a comma on either side (Caya,t, cahima,t,) follow the print. The n-with-tilde that the 1914 type sets in libang̃an, cayong̃, susunung̃in, mang̃a and the title’s at ng̃ has been restored from this scan, which shows it plainly; Wikisource prints those words with a plain ng.
SALITA AT BÚHAY NA PINAGDAANAN NI Sigesmundo at ng̃ Duquesa Policena AT NANG PRÍNCIPE ADONAIS AT NANG PRINCESA CLORIANA AT PRÍNCIPE TANG-CREDO AT NI LAVEÑA SA CAHARIAN NANG ALBANIA. Ang binalangcas co na naguing libang̃an Corona nang pula na di naigtaran cayong̃ bumabasa i, bahalang mag atang sa susunung̃in cong mang̃a cahihian. Caya,t, ang samo co daluyan nang habag ang pagca aba co sa macacaharap houag ibulusoc sa pagcapahamac sa alin mang isip nitong mamimintas. Mangyari caya na di abuluyan sa ilalagda co na cahima,t, culang na matuclas naua at aquing maalaman ang cay Sigesmundo buhay na dinaanan.
Transcription from Pahina:Salita at buhay na pinagdaanan ni Sigesmundo at ng duquesa Policena.pdf/4, via Tagalog Wikisource (Public domain).
From baybayin to Abakada: four spellings of one language
Baybayin is an Indic-descended abugida of three vowel signs and around fourteen consonant symbols, each of which stands for a consonant plus the vowel a unless a kudlit — a dot or a short line placed above or below — shifts it to i/e or o/u. In its earliest form the script had no way to write a bare consonant at the end of a syllable at all; Fray Francisco López introduced a cross-shaped kudlit around 1620 to cancel the vowel and mark a closed syllable, so a page can belong to either convention depending on when and by whom it was cut or printed. The first dated specimen of baybayin in print is also the first book printed in the Philippines: Doctrina Christiana en lengua española y tagala, produced at Manila in 1593, which gives its catechism in Spanish, in romanised Tagalog set in Gothic type, and in baybayin, so the same prayer can be checked across three forms on one opening.
Spanish missionaries used baybayin for a time — colonial records describe Tagalog notaries signing legal documents in it well into the seventeenth century — but printing settled on romanised Tagalog spelled by Spanish convention: c and qu for the sound now written k, a tilde over ng̃ for the velar nasal, doubled vowels, and an unaspirated h. Franciscan and Augustinian friars fixed this spelling for two centuries of vocabularios, catecismos and corridos verse romances, from Pedro de San Buenaventura’s 1613 dictionary through Fray Domingo de los Santos’s Vocabulario de la lengua tagala, reprinted at Sampaloc in 1794 and still following the same conventions.
José Rizal argued in the 1880s for a spelling closer to Tagalog’s own phonology, replacing c and qu with k and marking the language’s own sounds directly; Lope K. Santos’s 1940 Balarila ng Wikang Pambansa then fixed the twenty-letter Abakada as the basis of the national language, later widened to today’s twenty-eight-letter Filipino alphabet to admit loan letters such as f and j. Baybayin, meanwhile, survived as a ceremonial script rather than a working one: it names government seals, appears on the peso banknote, and was declared the country’s national writing system by law in 2018, so a single research file can now hold a 1593 catechism, an eighteenth-century missionary dictionary and a modern plaque cut in the same syllabary.
Why ordinary OCR struggles here
A syllabary, not an alphabet
Each baybayin symbol stands for a whole consonant-vowel syllable rather than a single sound, and in its earliest form the script marked no consonant at the end of a syllable at all. A recogniser built for alphabetic scripts has no way to represent that, and the missing final consonant has to be supplied from the word it belongs to.
Kudlit marks that move the vowel
A single dot or line placed above or below a baybayin consonant switches its vowel from a to i/e or o/u, and a small cross beneath cancels the vowel altogether. At the resolution of a photographed page these marks are easy to lose or to misplace against the wrong symbol.
Two centuries of Spanish-based spelling
Printed Tagalog from 1593 to the 1940s used c and qu for k, a tilde over ng̃, and doubled vowels — spelling a model trained on modern Filipino reads as error and silently corrects, changing a name or a word rather than transcribing it.
Worn letterpress and crowded woodcut borders
Early Manila printing, including the 1593 Doctrina Christiana, shows uneven inking, worn type and dense ornamental borders pressed close against the text block, which confuses layout detection built for clean modern scans into reading ornament as text or text as ornament.
Who works with this material
Baybayin and early-print researchers
Scholars working on the 1593 Doctrina Christiana and other rare survivals need the syllabic text kept as a syllabary, not silently converted to Latin letters, with each kudlit recorded so a question about the pre- or post-1620 convention can be checked directly against the image.
Church and colonial-record historians
Parish sacramental books, missionary vocabularios and gobernadorcillo correspondence sit across diocesan and Spanish archives. Reading them means reproducing Spanish-era Tagalog spelling as printed, so a name or place is not quietly changed to its modern form.
Philippine heritage and revolutionary-history projects
Cataloguing historical markers, monument inscriptions and revolutionary-era ephemera such as the Kartilya at scale needs modern Filipino capital-letter signage read accurately, abbreviations and all, for a heritage register or a local-history database.
Civil registries and identification agencies
PhilSys enrolment, PSA civil registration and similar bodies process bilingual Filipino-English forms by the thousand and need each labelled field kept attached to its value, so Apelyido and Last Name are not flattened into one column of text.
Getting the best transcription
Say whether the page is baybayin or romanised
State the script so the model reads baybayin symbols as syllables, with kudlit marks changing or cancelling the vowel, rather than trying to map them onto Latin letters one by one.
Name the period for Spanish-based spelling
Telling the model the material is, for example, "1790s missionary Tagalog" keeps c, qu and the tilde over ng̃ in the transcription instead of silently correcting them to k and ng as modern Filipino would write them.
Start a worn early printed book on its clearest opening
The Doctrina Christiana’s dense woodcut borders and worn type make it the hardest page in a run of similar material; confirm the model separates ornament from text there before working through cleaner, later pages.
Keep bilingual form fields paired
For a modern Filipino-English form or card, ask for each label kept with its value, so a bilingual document is returned as matched pairs rather than collapsed into a single run of text.
Frequently asked questions
- Can it read baybayin, the pre-colonial syllabary?
- Yes. Each symbol is read as a consonant-plus-vowel syllable, and the kudlit marks placed above, below or through a symbol are read as changing or cancelling that vowel rather than being treated as stray marks.
- Does baybayin need the cross-shaped cancelling mark to be read correctly?
- No. A bare consonant symbol is read correctly whether or not the cross-shaped kudlit that Spanish printers added around 1620 is present, since both the earlier and later conventions appear in the historical record.
- Will old Spanish-based Tagalog spelling be modernised automatically?
- No, unless you ask for that. Spellings such as c, qu, ll and the tilde over ng̃ are reproduced as printed, and a modernised Filipino version can be produced as a separate output.
- Can it tell baybayin apart from similar Southeast Asian scripts?
- Yes. Baybayin’s smaller set of consonant shapes and its own kudlit conventions are recognised as the Tagalog script rather than folded into a related but distinct script such as Javanese or Balinese.
- How does it handle a bilingual Filipino-English form?
- Each field is read together with its label, so a Filipino heading and its English translation stay paired with the same value instead of being merged or duplicated.
- Does it cope with worn letterpress and crowded woodcut borders?
- It separates ornamental borders and printer’s marks from the running text, and flags letters obscured by worn or blotted type as uncertain rather than guessing at them.
Related scripts and pages
Transcribe your Tagalog material
Upload a baybayin manuscript, a Spanish-era catechism or dictionary, a historical marker or a bilingual Filipino form, and check the reading line by line before you export it.