פ״י is a root whose first letter is י: י־שׁ־ב sit, י־ר־ד go down, י־ד־ע know, י־צ־א go out, י־ל־ד give birth, י־ס־ד found. Against the undeformed grid on שלמים, this class is the one where the same radical shows up as three different things depending on which binyan you are in — and the reason is a piece of prehistory that is still fully readable in the paradigm.
In the large subclass the first radical was originally ו, and Hebrew turned every word-initial w into y. Word-initial is the whole rule. So:
- Nothing stands in front of the root (pa'al past and present) → the ו was word-initial and became י: יָשַׁב, יוֹשֵׁב.
- A one-letter prefix stands in front (pa'al future, nif'al past, hif'il, huf'al) → the ו was protected, and it contracts with the prefix vowel into a long e or o: יֵשֵׁב, נוֹלַד, הוֹשִׁיב, הוּשַׁב.
- A prefix plus a full vowel stands in front (pi'el, pu'al, hitpa'el) → the radical is an ordinary consonant with a vowel of its own and simply stays: לְיַישֵּׁב, הִתְיַישֵּׁב.
The grid, on י־שׁ־ב
| בניין | past (he) | present m.sg. | infinitive | The radical |
|---|---|---|---|---|
| pa'al | יָשַׁב yasháv | יוֹשֵׁב yoshév | לָשֶׁבֶת lashévet | י · but gone in the future יֵשֵׁב and the infinitive |
| nif'al | gap on this root — see the note below | |||
| pi'el | יִישֵּׁב yishév | מְיַישֵּׁב meyashév | לְיַישֵּׁב leyashév | י as a plain consonant |
| pu'al | יוּשַּׁב yusháv | מְיוּשָּׁב meyusháv | — none exists | י as a plain consonant |
| hif'il | הוֹשִׁיב hoshív | מוֹשִׁיב moshív | לְהוֹשִׁיב lehoshív | ו, as a long o |
| huf'al | הוּשַׁב husháv | מוּשָׁב musháv | — none exists | ו, as a long u |
| hitpa'el | הִתְיַישֵּׁב hityashév | מִתְיַישֵּׁב mityashév | לְהִתְיַישֵּׁב lehityashév | י as a plain consonant |
The nif'al cell needs care, because it is where reference tables go wrong. Pealim lists no nif'al on י־שׁ־ב — there is no לְהִיוָּשֵׁב, and the string להיוושב returns nothing on Hebrew Wikipedia either (insource:/…/, 2026-08-07). The class does have nif'als, on other roots, and they show the same נוֹ־ shape: נוֹלַד "was born" (י־ל־ד), נוֹדַע "became known" (י־ד־ע), נוֹסַד "was founded" (י־ס־ד). Their futures and infinitives put the ו back as a doubled consonant, and ktiv male duly writes it twice: יִיוָּולֵד yivaléd (ייוולד) and לְהִיוָּולֵד lehivaléd (להיוולד), both checked on Pealim (2026-08-10). The doubled ו is the radical; the second י in the future is a separate matter, the prefix's chirik written full, which is why the infinitive has only one. That paradigm is set out on nif'al of weak roots.
ישבנו על המדרגות כי כל השולחנות היו תפוסים.
yashávnu al hamadregót ki kol hashulkhanót hayú tfusím
We sat on the steps because all the tables were taken.
התיישבנו בשורה האחרונה כדי לצאת מהר בסוף.
hityashávnu bashurá ha'akharoná kdei latsét mahér basóf
We sat down in the back row so we could get out quickly at the end.
בסוף יישבו את המחלוקת מחוץ לבית המשפט.
basóf yishvú et hamakhlóket mikhúts levéit hamishpát
In the end they settled the dispute out of court. (formal / journalistic)
בית הספר נוסד לפני שמונים שנה ומאז לא שיפצו אותו.
beit haséfer nosád lifnéi shmoním shaná ume'áz lo shiptsú otó
The school was founded eighty years ago and hasn't been renovated since.
Two spellings that collapse, and both are on one root
Because the ו and the י both end up written as ordinary vowel letters, this class produces some of the tidiest homographs in the language — and they sit inside the grid you just read.
| Written | Reading | What it is |
|---|---|---|
| יושב | יוֹשֵׁב yoshév | pa'al participle "sits" — and the noun "resident" |
| יוּשַּׁב yusháv | pu'al "was settled" (formal) | |
| מושב | מוֹשָׁב mosháv | noun: a seat; a moshav, a cooperative village |
| מוּשָׁב musháv | huf'al participle "is seated" (formal) |
האזור הזה יושב רק בשנות החמישים.
ha'ezór hazé yusháv rak bishnót hakhamishím
This area was only settled in the fifties. (formal / written)
The second hif'il route, and why it matters here
Most פ״י roots build their hif'il on הוֹ־. A small closed set does something else entirely: it assimilates the yod into the second radical, exactly the way פ״נ assimilates its נ.
| Root | hif'il | Route |
|---|---|---|
| י־ר־ד | הוֹרִיד horíd lower, download | ו restored → long o |
| י־צ־א | הוֹצִיא hotsí take out, spend | ו restored → long o |
| י־ד־ע | הוֹדִיעַ hodía inform | ו restored → long o |
| י־צ־ע | הִצִּיעַ hitsía suggest | assimilated → dagesh |
| י־צ־ג | הִצִּיג hitsíg present, show | assimilated → dagesh |
| י־צ־ת | הִצִּית hitsít set alight | assimilated → dagesh |
The second group is גזרת חפי״צ, a closed set of eight roots with י first and צ or ז second: י־צ־ב, י־צ־ג, י־צ־ע, י־צ־ק, י־צ־ר, י־צ־ת, י־ז־ם, י־ז־ע. The practical consequence runs in the other direction and it is the single most useful thing on this page: a dagesh in a hif'il does not prove a hidden נ. הִצִּיעַ and הִצִּיל are the same shape and their roots are י־צ־ע and נ־צ־ל.
הוא הציג את עצמו כמומחה ואף אחד לא בדק.
hu hitsíg et atsmó kemumkhé ve'áf ekhád lo badák
He introduced himself as an expert and nobody checked.
תוריד את התמונה מהקבוצה, היא לא מחמיאה לאף אחד.
toríd et hatmuná mehakvutsá, hi lo makhmi'á le'áf ekhád
Take that photo down from the group — it doesn't flatter anyone. (to a man)
The infinitives that make beginners give up
Take the pa'al infinitives of this class on their own and they look like six unrelated words:
לָשֶׁבֶת · לָרֶדֶת · לָדַעַת · לָלֶדֶת · לָצֵאת, plus לָלֶכֶת, which joins irregularly from ה־ל־ך.
Now say what they have in common: a לָ־ prefix rather than the לִ־ of לִכְתּוֹב, no first radical at all, and a ־ֶת ending welded on to replace it. Five of them are one shape — with לָדַעַת showing the ordinary guttural adjustment, ־ַעַת rather than ־ֶת, because a ע pulls the vowels on both sides of it to patach. לָצֵאת is the odd one, short because י־צ־א is doubly weak, פ״י and ל״א, and the silent א swallows the ending. That is the entire mystery.
אני אף פעם לא זוכר לצאת מוקדם מספיק.
aní af pá'am lo zokhér latsét mukdám maspík
I never remember to leave early enough. (m. speaker)
אני רוצה לדעת מה קורה לפני שאני מסכימה למשהו.
aní rotsá ladá'at ma koré lifnéi she'aní maskimá lemáshehu
I want to know what's going on before I agree to anything. (f. speaker)
The class, and where each member is regular
| Root | pa'al | hif'il | Also fills |
|---|---|---|---|
| י־שׁ־ב sit | יָשַׁב · לָשֶׁבֶת | הוֹשִׁיב · huf'al הוּשַׁב | pi'el, pu'al, hitpa'el |
| י־ר־ד go down | יָרַד · לָרֶדֶת | הוֹרִיד · huf'al הוּרַד | — nothing else |
| י־ד־ע know | יָדַע · לָדַעַת | הוֹדִיעַ · huf'al הוּדַע | nif'al נוֹדַע, pi'el לְיַידֵּעַ |
| י־צ־א go out | יָצָא · לָצֵאת (also ל״א) | הוֹצִיא | huf'al הוּצָא |
| י־ל־ד give birth | יָלַד · לָלֶדֶת | הוֹלִיד (literary) | nif'al נוֹלַד, pi'el לְיַילֵּד |
| י־שׁ־נ sleep | יָשַׁן · לִישׁוֹן — keeps its yod: יִישַׁן | — | — |
That last row is a different animal. י־שׁ־נ has a genuine yod, not a converted ו, so it never disappears — the future is יִישַׁן yishán, written with two yods, and the present is the stative יָשֵׁן rather than yoshén. The two subclasses are separated in full on pa'al פ״י.
Reading it in both directions
From a form to a root. Meet מוריד: strip the hif'il's מ־, and the וֹ that follows is not a root letter but the class's marker, so it stands for a first radical you cannot see. Two consonants remain, ר and ד. Try the yod: י־ר־ד. There it is.
From a root to a form. Given י־ר־ד and asked for "take down", do not build on the pa'al stem ־רד — that stem exists only because the ו was word-initial. Go back to the ו: הוֹ + רִיד.
נודע לנו רק אתמול שהם עוברים דירה.
nodá lánu rak etmól shehém ovrím dirá
We only found out yesterday that they're moving house.
Common Mistakes
1. Sending a חפי״צ root down the הוֹ־ route. The class has two hif'il patterns and the small one is invisible from the outside.
❌ הוא הוציע רעיון חדש בישיבה
Incorrect — there is no such form. י־צ־ע assimilates its yod instead of restoring it: הוא הציע רעיון חדש בישיבה, hitsía. Compare הוציא, from י־צ־א, which does take the הוֹ־ route.
✅ הוא הציע רעיון חדש בישיבה וכולם התעלמו.
hu hitsía ra'yón khadásh bayeshivá vekhulám hit'almú
He put a new idea to the meeting and everyone ignored it.
2. Reaching for the pa'al when the hif'il is what the sentence needs. לדעת takes no addressee; "letting someone know" is a different cell of the grid.
❌ תדע לי מתי אתה מגיע
Incorrect — ידע cannot take ל־ for a person. Causing someone to know is the hif'il הודיע: תודיע לי מתי אתה מגיע.
✅ תודיע לי מתי אתה מגיע, אני אצא לאסוף אותך.
todía li matáy atá magía, aní etsé le'esóf otkhá
Let me know when you're getting in and I'll come out and collect you. (to a man)
3. Writing one yod where the consonant is doubled. In the heavy binyanim the radical is a real consonant, and ktiv male doubles a medial consonantal yod.
❌ מתישב
Incorrect as ktiv male for mityashév — the yod here is a root consonant, so full spelling writes it twice: מתיישב. The same goes for מיישב, ליישב, ליידע and לייצג. It stops at מיושב meyusháv, where the single yod is the consonant and the ו after it is the vowel.
4. Restoring the yod in the nif'al. The nif'al past is one of the cells where the original ו comes back, and the yod has no place in it.
❌ החברה נייסדה מיד אחרי המלחמה
Incorrect — the nif'al of י־ס־ד shows the ו as a cholam: החברה נוסדה מיד אחרי המלחמה, nosdá. Same shape as נולד and נודע.
✅ החברה נוסדה מיד אחרי המלחמה ועברה שלוש מדינות מאז.
hakhevrá nosdá miyád akharéi hamilkhamá ve'avrá shalósh medinót me'áz
The company was founded right after the war and has moved country three times since.
Key Takeaways
- The first radical of the large subclass was ו, and Hebrew changed word-initial w to y. Everything else in the class follows from whether anything stands in front of it.
- Three outcomes: י where the root is word-initial (יָשַׁב, יוֹשֵׁב); a long vowel where a one-letter prefix protects the ו (יֵשֵׁב, נוֹלַד, הוֹשִׁיב, הוּשַׁב); a plain consonant in the heavy binyanim (לְיַישֵּׁב, הִתְיַישֵּׁב, לְיַידֵּעַ).
- י־שׁ־ב has no nif'al. The class's nif'als sit on other roots and show the same נוֹ־: נוֹלַד, נוֹדַע, נוֹסַד, with futures and infinitives that restore the ו as a doubled consonant, written twice in ktiv male (יִיוָּולֵד, לְהִיוָּולֵד).
- Two hif'il routes: הוֹ־ for most of the class (הוֹרִיד, הוֹצִיא, הוֹדִיעַ), and assimilation into a dagesh for the eight roots of גזרת חפי״צ (הִצִּיעַ, הִצִּיג, הִצִּית). So a dagesh in a hif'il proves nothing about a נ.
- The pa'al infinitives לָשֶׁבֶת, לָרֶדֶת, לָדַעַת, לָלֶדֶת, לָצֵאת — plus לָלֶכֶת from ה־ל־ך — are one shape: לָ־ prefix, no first radical, ־ֶת ending — ־ַעַת in לָדַעַת, where the guttural ע forces a patach. לָצֵאת is short because י־צ־א is also ל״א.
- יושב is both yoshév and yusháv; מושב is both mosháv and musháv. The class's vowel letters are what make these collapse.
- י־שׁ־נ is the other subclass: a real yod that never leaves, giving יִישַׁן with two yods and the stative present יָשֵׁן.
Now practice Hebrew
Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.
Start learning Hebrew →Related Topics
- Pa'al פ״י: ישב, ידע, ירד, יצאB1 — Two different classes hide behind one name. The large one drops its yod and gives יֵשֵׁב, לָשֶׁבֶת, לָדַעַת, לָצֵאת; the small one keeps it and gives יִישַׁן. Split them and the paradigm stops looking random.
- Hif'il פ״י: הוריד, הושיב, הודיעB1 — A root that begins with י turns that letter into a long o in hif'il, written with a vav: י־ר־ד gives הוֹרִיד, י־שׁ־ב gives הוֹשִׁיב, י־ד־ע gives הוֹדִיעַ. This is the class where the causative relation to pa'al really does hold — and where מוריד can be mistaken for a huf'al.
- פ״נ Across the BinyanimB2 — One deformation — a vowelless נ swallowed by the next radical, leaving a dagesh — accounts for forms right across the verb system. The same page has to say what the dagesh does not prove: it names an assimilation, not a letter.
- Weak Roots (גזרות): Why Some Words Look BrokenA2 — קם has two consonants, יפול has lost its נ, and קנה ends in a letter nobody pronounces. None of these verbs is irregular — each belongs to a גזרה, a root class where one weak letter behaves predictably. Here are the classes, the reason a letter goes missing, and where each one is taught in full.
- שלמים: The Sound Root in Every BinyanB1 — A root with no weak letter behaves identically in all seven binyanim — which makes it the control case. Here is one grid, two roots, and the honest gaps in both, so that every weak class on the pages that follow can be read as a documented deviation from this.