What it is, how it was made, and which parts of it have been checked.
A chrestomathy is a graded collection of real passages for learning a language; this one is Syriac, read first.
That last phrase is the design. A beginner's Syriac normally opens with the alphabet, spends some weeks on the noun and some more on the verb, and reaches continuous text somewhere near the end — by which point the reader has learned a good deal about Syriac and read almost none of it. This course inverts the order. Lesson 1 is Genesis 1:1–5, in Syriac, on the first page you turn to. Grammar arrives when the line in front of you needs it and not before, and every lesson ends with you reading a real sentence unaided.
The trade is deliberate and worth stating plainly. Eleven lessons will not make you a Syriacist. They are meant to leave you able to sit down with a page of the Peshitta and a dictionary and get through it — which is the thing most people who want Syriac actually want, and the thing a paradigm-first course reaches last.
There are eleven documents, Lesson 0 through Lesson 10. Lesson 0 is the twenty-two letters and nothing else — no meaning at all, on purpose, because learning shapes and learning sense at the same time is two problems, and beginners fail at the seam between them. From Lesson 1 the text is real throughout: the first four chapters of Genesis, the Beatitudes, the Lord's Prayer, and you are the light of the world, which is where the course ends.
There are two doors into a language, and they are not the same skill.
paleography.app teaches you to read the hand — ink on a manuscript page, in this case a pocket Gospel book written at Vienna in 1554 by Moses of Mardin, which is the very copy the compositors worked from when they set the first book ever printed in Syriac the following year. That track will not teach you a word of the language. It teaches your eye.
This site teaches the language, in a modern printed face where every letter is unambiguous. It will not teach you to read a manuscript.
They are built to meet. Lesson 10 ends by handing you the same line you have just learned, as ink, on the other track. Learn one and the other costs you a week rather than a year; learn neither and a manuscript is a wall.
Four public-domain books, all read from scans at the Internet Archive:
The spine of the whole thing — its grammar supplies the paradigms, its glossary the vocabulary, and its chrestomathy the text. The word on this site's masthead is Nestle's own: Chrestomathia stands at the head of p. 67, which is the page Lesson 1 is cut from.
glossary leaves a lesson short.
source but a control: an independent printed paradigm to check ours against, cell by cell.
Every keyed page of Genesis in this course has been diffed against it.
One finding from that reading is worth repeating here, because a reader meets it in Lesson 5 and would otherwise think the site had broken. Nestle does not print his Genesis in one script. Caput I and Caput III are set in vocalized Serto — vowel points and all — and Caput II and Caput IV in unvocalized Estrangela. The alternation is by chapter, and it is deliberate on his part. This project first recorded it as a gradual drift across the pages and had to correct itself. So Lesson 5 is the first lesson with no vowel points on the page not as a pedagogical stunt but because that is what the book does there — and reading unpointed text, supplying the vowels you already know, is the skill every manuscript will eventually demand of you anyway.
Nothing in a lesson is typed from memory. Each printed page is read off a high-resolution crop of the plate and captured as a record — a small structured file that says what the page prints, where it prints it, how confident the reading is, and what it was checked against. There are three kinds in play here:
| kind | what it holds | how many |
|---|---|---|
| R3 | a keyed page of continuous text, with its verse numbers, marginal line numbers and layout as printed | 66 |
| R1 | a grammatical paradigm — a verb's cells, a noun's states, the numerals | 14 |
| R4 | a lexicon entry — lemma, part of speech, gloss | 901 |
A lesson is then built backwards from the records: choose the destination passage first, see which forms and words it actually requires, and teach only those. That is why Lesson 2 covers eight verses and Lesson 5 covers three — the count that matters is new material, not verses.
Every record is published. The footer of each lesson names the records it was built from, and each one links to the reading itself, so a claim in a lesson can be traced to the plate it came off in two clicks. That is the point of the sources shelf: not decoration, but the ability to check us.
This is the section to read first if you know Syriac, and the reason the next page exists.
are fully keyed and independently diffed against a second printed edition. Those are the pages the lessons are built from. The remaining 44, the Vitae Prophetarum and the Historia inventionis, are partially or wholly unkeyed and no lesson touches them.
printed paradigms, and the collation is recorded in the record, disagreements included. Of the four without: two are calendar lists, for which Nöldeke has no section at all; one is the alphabet table; and one — the Peal perfect with object suffixes — is simply still owed.
edition.** The vowel points are legible as shapes at the resolutions used and are recorded. The finest layer — the rukkakha and qushshaya bars, the ܗ̇ point, the linea occultans — is recorded where seen and is not warranted point by point.
plates are pointed and the points are visible; there is no control for that typeface's pointing and no second reader. So the consonantal skeleton, the part of speech and the English gloss went in — all of them reliable — and the vocalisation stayed out rather than being guessed. 161 of the 874 records taken from Nestle's own glossary carry an uncertainty flag on their face for similar reasons.
file: expert = "". Everything here was read off the plates by one person who reads Latin and Greek, working with a language model, against standard reference grammar. Where a reading could not be resolved it is marked unresolved rather than smoothed over, on the same rule the hand track runs on: a confident wrong reading is worse than none.
It teaches reading, and only reading. It will not teach you to compose Syriac, and it does not try to make you fluent — eleven lessons cannot. It is not a substitute for a grammar; Nöldeke is free and is named above. It teaches West Syriac letterforms because that is what the sources print, so East Syriac and Estrangela will each cost you a little adjustment and not a new alphabet.
It is free, it has no account, no login and no paywall, and it never will.
Nestle, Payne Smith, Nöldeke and Barnes are in the public domain and were read from scans at the Internet Archive. No licence is claimed over their text: a transcription of a public-domain work creates no new copyright, and saying otherwise would be a claim we have no right to make. What is offered under CC BY-NC 4.0 is this project's own work — the lesson prose, the drills, the glosses and notes, and the encoding as a compilation. Syriac is set in Noto Sans Syriac (SIL Open Font License).
Made alongside Wroot Press, which publishes translations from the manuscripts these hands were written in. Corrections, and especially corrections to the Syriac, to wilson.pruitt@gmail.com.