Track: Learning to Read Unpointed Hebrew

This is the single highest-leverage thing a Hebrew learner can do, and the reason is arithmetic rather than pedagogy: real Hebrew is printed without vowels. Newspapers, signs, menus, contracts, novels, WhatsApp messages and street names all omit them. Until unpointed text is readable, every source of authentic input is closed to you and your Hebrew has to come out of textbooks. Start this track in parallel with the A1 path, not after it.

The track builds one skill only. The sound system and the accent belong to the speaking track; pointed classical text belongs to the classical track; and the grammar of whatever you have decoded belongs to the level paths — with the verb morphology that stage 4's roots feed into handed to the binyanim track.

Budget 60 to 100 hours to the point where an unpointed newspaper sentence is decodable at reading speed. Almost every page below is already written, so this is a route through existing material rather than a promise.

Stage 1 — The script and its traps

Goal: convert any printed letter into a sound without hesitating. Eleven pages, and the traps matter more than the alphabet: a beginner who confuses ד with ר loses a root, not a letter.

• ☐ pronunciation/alphabet-overview — all 22 letters, their modern sounds and their final forms. The starting inventory; everything else refines it.
• ☐ pronunciation/right-to-left — lines, books, menus and app buttons all mirror; numbers do not. A page about direction, read before anything long.
• ☐ pronunciation/final-forms — the five letters that change shape at the end of a word. Five extra shapes, learned with the alphabet rather than after it.
• ☐ pronunciation/letter-names — the names, and the variants Israelis actually use. You need them to ask what a letter is.
• ☐ pronunciation/lookalike-letters — ב against כ, ד against ר, ה against ח against ת. The highest-value page in the stage: these confusions destroy roots.
• ☐ pronunciation/print-vs-cursive — Israelis do not handwrite the letters you learned to read. Two alphabets, and nobody warns you about the second.
• ☐ pronunciation/shin-sin — one letter spelling two consonants, told apart by a dot print omits. Your first genuine unpointed ambiguity.
• ☐ pronunciation/begadkefat — b/v, k/kh, p/f, and which one you get. The rule is predictable, which is why it is safe to omit in print.
• ☐ pronunciation/vav-yod-glides — ו and י as consonants and as vowel letters. The hinge of the whole track: inside a word, these two letters are the only vowels print gives you.
• ☐ spelling/final-form-rule — when a letter switches, and when a suffix switches it back. Explains why the same word looks different with and without an ending.
• ☐ spelling/no-capitals — one case only, and what Hebrew does instead. Names are not marked; you have to notice them yourself.

You can now: read any Hebrew word aloud once someone has told you the vowels — and you can tell the look-alike letters apart at printed size.

Stage 2 — What niqqud does, and why print drops it

Goal: understand what the dots encode, so you can predict what is missing. Five pages. Niqqud is not a beginner's crutch that adults outgrow; it is a separate layer with a different job, and knowing what it would have told you is exactly what lets you supply it.

• ☐ pronunciation/niqqud-overview — the system, and what each sign fixes. First: you cannot predict a missing layer you have never seen.
• ☐ spelling/ktiv-male-vs-haser — two coexisting spelling systems, not one with errors. The central fact of the stage, and of Hebrew orthography.
• ☐ spelling/ktiv-male-rules — only i, o and u get a letter; a and e get nothing. One line that predicts most unpointed spellings.
• ☐ spelling/double-vav-yod — תקווה, בניין: a doubled letter marks a consonant. Immediately after the rules, because doubling is what the rules force.
• ☐ spelling/alef-hey-final — two silent finals, and the choice is not arbitrary. Ends the stage on the one position where you must know, not guess.

You can now: look at an unpointed word and say which vowels the writing has told you about and which it has left for you.

Stage 3 — The letters glued to the front

Goal: strip a word down to the word you can actually look up. Five pages, and they come before the root work for a hard reason: a reader who cannot peel ומה־ off the front of a word extracts the wrong root every single time, and then concludes the dictionary is missing entries.

• ☐ spelling/prefix-letters — the seven letters that attach with no space. The master page of the stage; everything else here is a detail of it.
• ☐ determiners/article-fusion — ב, ל and כ swallow the article and keep only its vowel. Explains a definite noun that shows no ה at all.
• ☐ determiners/article-vocalization — הַ, הָ, הֶ: three spellings, one rule. Recognition, not production — but it confirms that a ה really is the article.
• ☐ conjunctions/vav-vocalization — one letter, five readings, and nothing in print to choose between them. One letter doing more work than any other deserves its own page.
• ☐ spelling/maqaf — Hebrew's own hyphen, which is not the Latin one. Closes the stage: another join to recognise and cut.

You can now: take a word carrying two or three prefix letters, remove them in order, and search for what is left.

Stage 4 — Root and pattern: the strategy that resolves the ambiguity

Goal: recognise a word by its consonant skeleton and the mould it sits in. Eight pages. This is the stage that has to come before the core method, because supplying missing vowels means supplying them from the root, the pattern and the sentence at once — and two of those three are learned here.

• ☐ roots/root-concept — vocabulary filed under three-consonant skeletons. The idea the rest of the stage operationalises.
• ☐ roots/triliteral-roots — extracting a root, and the eleven servile letters. Turns the idea into a procedure you can run on an unknown word.
• ☐ roots/root-families — sixteen words on כ־ת־ב, each labelled with its mould. One root worked exhaustively, which is worth more than ten described.
• ☐ roots/pattern-concept — משקל for nouns, בניין for verbs. The second of the three inputs to a good guess.
• ☐ roots/mishkal-overview — which shape makes a profession, a place, a tool, an act. Lets you guess a meaning class before you know the word.
• ☐ roots/weak-roots-overview — why some words look as though a letter is missing. Because sometimes one is, and you must not go looking for it.
• ☐ roots/root-and-dictionary — why a lookup fails in a way English never prepares you for. The practical payoff of the whole stage.
• ☐ verbs/fundamentals/citation-form — verbs are listed under the past-tense הוא. Last: it is the one lookup convention the root pages do not cover.

You can now: meet an unknown word, extract its root, guess its mould, and narrow it to two or three possible readings before you reach for a dictionary.

Stage 5 — The method, and the conventions around it

Goal: read unpointed text the way it is meant to be read. Seven pages, built around one core page and six sets of conventions that turn up in real print and nowhere in a textbook.

• ☐ spelling/reading-unpointed — supplying the vowels yourself, as a method rather than a knack. The page this whole track exists to make readable.
• ☐ spelling/geresh-gershayim — the two marks that are not an apostrophe and a quotation mark. You will hit one in the first paragraph of anything.
• ☐ spelling/acronyms — צה״ל, בג״ץ, עו״ד, and how they are pronounced. Israeli print is saturated with them, and they are not read like English acronyms.
• ☐ spelling/letter-numerals — letters used as digits, in dates and lists. Dates, chapter numbers and school years all use it.
• ☐ spelling/punctuation-rtl — the same marks, at the other end of the line. Brackets and quotation marks that look reversed and are not.
• ☐ spelling/foreign-words — names and loans transcribed by sound, not by source spelling. A quarter of a news page is transcribed names.
• ☐ spelling/typing-hebrew — the layout, the niqqud, and the bidi traps. Closes the stage: reading and writing hit the same joins.

You can now: work through an unfamiliar unpointed sentence deliberately — skeleton, mould, context — and get the right reading most of the time.

Stage 6 — Practice on things short enough to fail at cheaply

Goal: build speed on real text where a wrong guess costs nothing. Four annotated texts, deliberately short. A sign you misread has one line; a novel you misread has four hundred pages.

• ☐ texts/signs-and-notices — doors, platforms, stairwells, shopfronts. Shortest authentic Hebrew there is, and it is everywhere.
• ☐ texts/news-headlines — twelve headlines in the four shapes a front page uses. Compressed to the limit, and the standard test of unpointed reading.
• ☐ texts/whatsapp-thread — fourteen messages between two friends. Informal spelling, which behaves differently from print.
• ☐ texts/recipe — infinitive instructions, line by line. Longest of the four, and predictable enough to read fast.

You can now: read a sign, a headline, a message thread and a recipe without pointing — which is the threshold at which authentic input becomes available and the rest of your Hebrew starts accelerating.

Common Mistakes

Treating niqqud as a stage you graduate from. Learners often plan to "move to unpointed text later". Later never arrives on its own, because pointed material is a closed set written for learners. Start decoding in week one, on signs.

Sounding a word out letter by letter. That is how you read a language whose spelling encodes vowels. Hebrew asks you to recognise a skeleton and then choose a reading, which is a different operation and has to be practised as one.

Looking a word up exactly as it appears. ומהמשרד is not in any dictionary. Peel the prefixes, undo the final-form switch, find the root — stage 3 and stage 4 exist for precisely this.

Committing to the first reading that works. ספר, שם, אמר and hundreds like them have several legal readings, and the sentence often settles it two or three words later than you would like. Hold the ambiguity open.

Learning print and skipping cursive. Everything handwritten — a note, a form, a whiteboard — is in a second alphabet you have not been taught. print vs cursive is not optional.

Key Takeaways

• Real Hebrew is unpointed. Until you can read it, your input is limited to material written for learners.
• Guess from consonants plus context, never from a vowel you supplied first.
• Prefix letters come before root work: strip the word, then extract the root.
• Root and mould together narrow an unknown word to a handful of readings; the sentence picks one.
• Practise on signs, headlines, messages and recipes — short texts where a wrong guess costs nothing.
• Sixty to a hundred hours, run in parallel with a level path, and the whole rest of the language opens up.