Identifying a Verb's גזרה from Its Infinitive

The גזרות page names the classes. This one is the decision procedure: you have met לְהַגִּיעַ in a sentence, you need its root, its binyan and its class, and you need them in the next four seconds.

The infinitive is the right form to start from, because it is the form Israelis actually hand you when you ask how to say something, and because it carries the binyan's whole signature at the front. Run it in two steps: the front of the word gives you the binyan; what is left gives you the גזרה.

Step 1 — the binyan, from the front

Strip the ל and look at the first letter or two of what remains. There are only five possible answers, not seven, because pu'al and huf'al have no infinitive at all.

Three footnotes make the table reliable. hitpa'el hides when the first radical is a sibilant: the ת swaps places with it, so להשתמש, להסתדר, להצטלם and להזדקן all begin with something other than הת. That is metathesis, and the ת is still there, just one letter later. Second, the pa'al/pi'el split is decided for you by ktiv male: pa'al's o is written with a vav — לכתוב, לסגור, לשמוע — while pi'el never has one. Among full-stem infinitives, then: no vav, no pa'al. Only full-stem, because pa'al's short infinitives have no o to write either — לָרֶדֶת, לָצֵאת, לָתֵת, לָלֶכֶת — and step 2's ־ֶת row is what catches them. Third, a ־ות ending swallows the stem vowel: לְהַרְאוֹת shows no long i and לְנַסּוֹת no a–e, so neither of those rows fires on a ל״ה infinitive. Read it by elimination instead — a לְהַ־ that is neither הת־ nor הי־ is hif'il.

Step 2 — the גזרה, from what is left

The two rows that say "or" are the ones to slow down on, because guessing there is how people end up hunting for roots that do not exist. The ־ֶת row also has two lexical exceptions: לָקַחַת is ל־ק־ח, whose ל assimilates as though it were a נ, and לָלֶכֶת is ה־ל־ך, whose ה drops. The ending promises a missing first radical, not that it was a י or a נ.

Eight infinitives, end to end

Four of those eight need a word of reasoning.

לָרֶדֶת. The ־ֶת ending and the two consonants ר and ד say a first radical is missing, and the class could be either. Check the past tense: יָרַד. There is the letter, and it is a י.

לְהַגִּיעַ. The dagesh in the ג has no source in the visible letters, so a first radical assimilated into it. Was it a נ or a י? The yod route, גזרת חפי״צ, is a closed set of eight roots whose second radical is צ or ז — and here the second radical is a ג. So the letter is a נ: נ־ג־ע. The same test tells you that הִצִּיג and הִצִּיעַ are not פ״נ.

לְהָבִין. After the hif'il's ה only ב and ן remain, and hif'il's own long i fills the gap. Two consonants plus a long vowel means a hollow root — and hollow roots hide either a ו or a י, with nothing in hif'il to say which. Here it is ב־י־ן, and no amount of staring will tell you; look it up.

לָצֵאת. Two signals at once — the ־ֵאת shape of a lost first radical, and an א sitting at the end of the stem. That is two classes on one root, פ״י and ל״א, which is exactly why the infinitive is shorter than its neighbours לָשֶׁבֶת and לָרֶדֶת.

Running it backwards

Given a root and a binyan, can you predict the infinitive? In four binyanim out of five, yes — the shapes are fixed. כ־ת־ב happens to fill all five, so it can show the whole answer at once.

pa'al is where prediction stops, and it stops completely. Look at what the same binyan produces: לִכְתּוֹב, לִשְׁמוֹעַ, לָשֶׁבֶת, לָקַחַת, לָתֵת, לִיפּוֹל, לָצֵאת, לָלֶכֶת. Within פ״נ alone the outcomes are לִיפּוֹל, לִנְסוֹעַ, לָגַעַת, לָתֵת and לָגֶשֶׁת — five different answers from one class, and no rule generates them. Learn every pa'al verb as a pair, citation form plus infinitive, and treat the infinitive as a fact about the verb rather than something you derive.

Common Mistakes

1. Looking the verb up as you met it. This is the single biggest obstacle English speakers hit with a Hebrew dictionary, and it makes the language look as though it is missing its commonest words.

2. Counting a prefix as a root letter. Stopping at the letters you can see gives you a real root that is the wrong root.

3. Assuming an assimilated first radical is a נ. Both פ״נ and פ״י assimilate, and the surface result is identical.

4. Filing a bare-ל infinitive as pa'al. pa'al and pi'el are the two binyanim with no ה, and only the vowels separate them — but ktiv male prints the difference for you.

5. Filing a hif'il as pi'el. Four letters and a מ in the present tense make הבין look exactly like דיבר.

Key Takeaways

• Work in two steps: the front of the infinitive gives the binyan, the stem gives the גזרה.
• Only five binyanim have an infinitive — pu'al and huf'al do not — so step 1 has five answers: הת־ hitpa'el, הי־ nif'al, ה + long i hif'il, bare a–e with a doubled middle radical pi'el, bare stem with an o pa'al.
• ktiv male decides pa'al against pi'el for you among full-stem infinitives: pa'al's o takes a vav, pi'el never has one. The short pa'al set (לרדת, לצאת, לתת) has no vav either and belongs to step 2.
• hitpa'el disguises itself before a sibilant — להשתמש, להסתדר, להצטלם — because the ת moves. And a ־ות infinitive shows no stem vowel at all, so step 1 reads it by elimination.
• The two "or" answers need a second check. For a ־ֶת infinitive, look at the past tense (יָרַד → י), and remember the two that hide neither a י nor a נ: לָקַחַת is ל־ק־ח, לָלֶכֶת is ה־ל־ך. For an unexplained dagesh, ask whether the second radical is צ or ז: if not, the letter is a נ.
• Backwards works in four binyanim and fails in pa'al. לִכְתּוֹב, לָשֶׁבֶת, לָקַחַת, לָתֵת, לִיפּוֹל, לָגֶשֶׁת are not derivable — learn each pa'al verb as a citation-form-plus-infinitive pair.
• The habit to break: searching for a verb under its prefix letter. ל, מ, נ, ה and ת at the front of a verb belong to the pattern, never to the root.