In short

Some of the most common verbs in Hebrew — קנה (kana, “bought”), ראה (ra’a, “saw”), עשה (asa, “did/made”), רצה (ratza, “wanted”) — belong to the ל”ה class (gzarat lamed-he): roots whose third slot holds a weak ה. You already met one as a modal, רוצה (rotze, “want”), and the whole pa’al paradigm of these verbs was built back in the gzarot lesson (present, past) and in the pa’al tense lessons. What is new here is that the ל”ה tail is not a pa’al quirk — it overrides the strong paradigm in every binyan.

So instead of learning a fresh irregular verb in pi’el, hif’il, hitpa’el and nif’al, you learn one set of endings and graft it onto whichever binyan stem you already know. This lesson isolates exactly that deviation: the shared present, past, future and infinitive tails that a ל”ה root imposes wherever it appears.

The rule

• The whole class is defined by one thing: a ה in the third root slot. In the template פ־ע־ל, the “ל” position is a ה — hence ל”ה. In practice that ה never surfaces as a consonant; it dissolves into a vowel ending, and a י shows up in part of the past. The four surface signatures below are the same across all five binyanim.

• Present: -e (m.sg) / -a (f.sg) — one written word — plus -ים / -ות. קונה is both קונה (kone, m.sg) and קונה (kona, f.sg): unvocalized they are the identical spelling, and only the vowel you say tells them apart. The plurals are קונים (konim) and קונות (konot). There is no strong-verb ־ת feminine here (never *קונת).

• Past: bare 3ms, ־תה 3fs, ־ו 3pl, and a yod through 1st/2nd person. The third-person forms are bare/short — קנה (kana, “he bought”), קנתה (kanta, “she bought”), קנו (kanu, “they bought”) — while everywhere else a י threads in: קניתי (kaniti), קנית (kanit / kanita), קנינו (kaninu), קניתם (knitem). This yod is the class fingerprint: קניתי, never *קנתי.

• Future: -ה (sg) / -ו (pl). יקנה (yikne, “he will buy”), יקנו (yiknu, “they will buy”); likewise אקנה (ekne, “I will buy”), נקנה (nikne, “we will buy”). The strong final consonant is replaced by that open -e / -u ending.

• Infinitive: -ות. לקנות (liknot), לראות (lir’ot), לעשות (la’asot), לרצות (lirtzot), לשתות (lishtot). The infinitive is the fastest class-detector — a ל”ה verb’s shem ha-po’al always ends in ־ות, never a strong *לקנה (likne).

• The same tail rides on every binyan stem. Build the pi’el / hif’il / hitpa’el / nif’al stem exactly as usual, then attach the ל”ה endings instead of the strong ones: pi’el ניסה (nisa, “tried”), hif’il הראה (her’a, “showed”), hitpa’el השתנה (hishtana, “changed”), nif’al נראה (nir’a, “looked/seemed”). The infinitives — לנסות, להראות, להשתנות, להיראות — all end in ־ות, proving the shared class.

• One divergence to flag: nif’al feminine present is ־ית, not -a. In pa’al, pi’el, hif’il and hitpa’el the f.sg present is the -a homograph (מנסה menase/menasa, מראה mar’e/mar’a). Nif’al breaks this: its f.sg is נראית (nir’et) — “she looks tired” is היא נראית עייפה (hi nir’et ayefa), not *נראה (nir’a).

Forms & patterns

The class signature — pa’al רצה (want), all four surface tails:

Past person paradigm — root קנה (buy). The yod threads through, except 3rd person:

Frequent pa’al ל”ה verbs — the anchor forms:

The same ל”ה tail across the derived binyanim — one verb each:

Notice each derived past 3ms already ends in the bare ל”ה vowel (ניסה, הראה, השתנה, נראה), and every infinitive ends in ־ות — the class overriding the binyan.

Examples

Common mistakes

• קנתי ספר -> קניתי ספר (kaniti sefer) - never drop the yod in the 1st/2nd-person past; the ל”ה fingerprint is ־יתי, so it is קניתי.
• קריתי מכתב -> קראתי מכתב (karati michtav) - קרא is ל”א (final א), not ל”ה: keep the alef and take no yod (also מצא→מצאתי matzati, נשא→נשאתי nasati).
• אני רוצה לקנה -> אני רוצה לקנות (ani rotze liknot) - don’t over-regularize the infinitive; the ל”ה shem ha-po’al ends in ־ות.
• היא קנה שמלה -> היא קנתה שמלה (hi kanta simla) - the 3fs past ends in ־תה (kanta); bare קנה (kana) is the masculine “he bought.”
• היא רוצת קפה -> היא רוצה קפה (hi rotza kafe) - there is no strong ־ת feminine in the present; m and f singular are the one spelling רוצה.
• הוא הראתה לי -> הוא הראה לי (hu her’a li) - the hif’il 3ms past is הראה (her’a); the ־תה form הראתה (her’ata) is feminine “she showed.”

Tips

• Learn each ל”ה verb by two anchors — present m.sg and past 3ms (רוצה/רצה, קונה/קנה, עושה/עשה). From those two the whole class falls out.
• Memorize the four shared tails once, and they carry across every binyan: present -e / -a (homograph) + ־ים/־ות, past ־יתי/־ית/־ינו/־יתם with bare 3ms, ־תה 3fs, ־ו 3pl, future ־ה/־ו, infinitive ־ות.
• Unvocalized, the singular present is one word for both genders: רוצה = rotze or rotza, קונה = kone or kona. Only the spoken vowel and context decide.
• The infinitive is your class-detector: לקנות, לראות, לעשות, לנסות, להראות, להשתנות all end in ־ות — see ־ות, think ל”ה.
• Keep ל”ה apart from ל”א: לקרוא/קראתי, למצוא/מצאתי keep the alef and take no yod. That look-alike is exactly what gets tested.
• In the derived binyanim the deviation is only the tail grafted onto the derived stem: מנסה → ניסה → ינסה → לנסות differs from a strong pi’el like מדבר → דיבר only in its final letter. Build the binyan once, then swap on the ל”ה ending.
• Watch the one exception: nif’al f.sg present is נראית (nir’et), not the -a homograph — “she looks tired” is היא נראית עייפה (hi nir’et ayefa).