The Alphabet

Twenty-two letters and fifteen vowel points, every one of them nameable. Pick a letter to see it large, beside the shape it had before the square script, with what epigraphers think that shape was a picture of. Then open a verse and take it apart, letter by letter, and read it aloud.

Using it to read

Go to Read a verse and choose one of the five. Tap any letter and the panel below takes it apart: the consonant, then every mark sitting on it, each with its name and its job. A single letter routinely carries three at once — a vowel, a dot that changes how the consonant is pronounced, and a melody sign. Knowing which mark does what is half the work: it is what separates stumbling from reading.

The two reading modes are the practice itself:

Everything is the text with its vowel points, as the Masoretes left it. Consonants only is what a reader had before the seventh century, and what a Torah scroll still has today — the same text, with every vowel supplied from memory and grammar, and without the two dots at the end of the verse. Moving from the one to the other, a verse at a time, is the oldest way there is to learn to read Hebrew.

The letters are old, the vowel points are not

The consonants are ancient. The dots and dashes around them are not: they were worked out by the Masoretes of Tiberias between roughly the seventh and tenth centuries, and the system this instrument shows is theirs. Two other systems, Babylonian and Palestinian, wrote their vowels above the line and lost out.

That gap matters when you read. The vowels record how the text was being recited in Tiberias around 900 CE — a careful, conservative tradition, but a tradition, and eight centuries or more after the consonants were fixed. Where the Septuagint transliterates a Hebrew name differently from the way the points tell you to say it, that difference is worth noticing rather than explaining away.

A printed Hebrew Bible carries a third kind of mark above and below the words: the cantillation, which holds the chant and divides the verse into clauses. Those have a tab of their own here, Cantillation, split into the marks that divide the verse and the marks that bind words together — and below them, separately, the paragraph marks: the פ of an open section, the ס of a closed one, and the reversed nun that brackets a passage nine times in the whole Bible.

Out of the five verses the accents are left out, so that there the letters and the vowels remain — and that is where reading begins.

Except for three, because those three do not sing.

The sof pasuq ׃ closes the verse: two dots one above the other, after the last word. It belongs not to the melody but to the punctuation, the way a full stop does at the end of a sentence. Every verse in the Hebrew Bible ends with it, from Genesis 1:1 to the last verse of Chronicles, and without it a verse would simply stop.

The paseq ׀ is a bare vertical stroke between two words. It separates them without giving them a melody — often to keep you from reading them run together. In Deuteronomy 6:4 it stands between the Name and echad, exactly where you want to hold back for a moment.

The maqaf ־ joins two or more words into a single accent unit. The first word gives up its own stress, and the pair is read as though it were one word. In Deuteronomy 6:5 it stands three times: בְּכָל־לְבָבְךָ, וּבְכָל־נַפְשְׁךָ, וּבְכָל־מְאֹדֶךָ — with all your heart, with all your soul, with all your might. Turn the marks off and it falls away, leaving two separate words again, exactly as a Torah scroll has them.

Click any of the three in the verse and the panel below names it. All three also stand in the grid under Vowel points, under the heading The reading marks.

What the Paleo-Hebrew forms are, and are not

The letters at the left of each entry are drawn, not typeset. Unicode has a block for them, but almost nobody has a font that covers it, so a reader would get a row of empty boxes. Each drawing is a schematic of the Old Hebrew form of roughly the tenth to sixth century BCE — the script of the Siloam inscription and the Lachish letters — not a facsimile of any one hand.

Beside each one is the acrophonic reconstruction: the word the letter’s name seems to come from, and the thing the earliest shape seems to depict. That enterprise is real scholarship, running from Alan Gardiner’s reading of the Serabit el-Khadim inscriptions in 1916 through Albright, Naveh, Sass, Hamilton and Goldwasser. But it is not uniform in strength, so every entry says how firm its own reading is:

The one thing this will not do

It will not tell you what a word means from the letters in it.

There is a popular method that reads each letter as carrying a concept — alef strength, bet house, and so on — and then builds the sense of a Hebrew word out of the sense of its letters. It is attractive, but that is not how the language works. By the time a single line of the Bible was written, the letters had long been signs for sounds and nothing more. That a letter’s ancestor was once a picture of an ox tells you about the history of writing, not about the meaning of the words spelled with it — any more than the Latin A, which descends from the very same ox head, tells you anything about English words containing it.

Scholarship has a name for the move: the etymological fallacy — reading a word’s meaning out of where the word came from. James Barr settled the case against it in The Semantics of Biblical Language in 1961, and it has not needed making twice. Meaning is settled by usage — by how a word is actually used in the sentences we have — and Moria is the instrument on this site for looking at that.

So the origin column here answers one question only: where did this shape come from? It is a fact about the history of an alphabet, which is interesting on its own, and it stops there.

The numeric values

Each letter also carries a number, and the instrument gives it: alef 1 through tet 9, then the tens, then the hundreds. This is genuinely old — it appears on Hasmonean coins and throughout rabbinic literature — and it is the basis of gematria. Worth knowing when you meet it; not a method this site uses.

Moria, for the words themselves →  ·  Texts and datasets used →