In short

Lessons 43 and 44 showed you two machines that take a verbless sentence and re-case it: كانَ leaves the first word مَرْفوع and pushes the second into the accusative, إِنَّ does the mirror image. This lesson adds the last two machines of that kind. ظَنَّ وأَخَواتُها swallow the whole nominal sentence and turn both of its parts into accusatives - الجَوُّ جَميلٌ becomes ظَنَنْتُ الجَوَّ جَميلًا. كادَ وأَخَواتُها keep the first part مَرْفوع and demand that the predicate be a مُضارِع verb rather than a noun - كِدْتُ أَسْقُطُ, “I almost fell”.

They share a page because they share a habit: each takes a sentence you can already build and bends its endings. The work is in three places - remembering that ظَنَّ wants two accusatives and not one, remembering which كادَ-sister insists on أَنْ and which forbids it, and surviving two awkward stems (ظَنَّ is doubled, كادَ is hollow) that reshape themselves the moment a suffix begins with a consonant.

The rule

1. ظَنَّ eats a nominal sentence and outputs two objects. Start from any مُبْتَدَأ plus خَبَر, put one of these verbs in front, and both words go مَنْصوب:

الجَوُّ جَميلٌ. → ظَنَنْتُ الجَوَّ جَميلًا. - “I thought the weather was nice.”

Traditional grammar calls the two words مَفْعولانِ أَصْلُهُما المُبْتَدَأُ والخَبَر - “two objects whose origin is subject and predicate”. Keep that name, because it hands you the test: if the two words are not a sentence on their own, they are not a ظَنَّ pair. الجَوُّ جَميلٌ is a sentence, so ظَنَنْتُ الجَوَّ جَميلًا works; الكِتابُ الجَديدُ is only a noun plus its adjective, so *ظَنَنْتُ الكِتابَ الجَديدَ says nothing.

2. The members you actually need. The traditional list is long and half of it is dead. These eight carry modern Arabic:

ظَنَّ، حَسِبَ، وَجَدَ، رَأى، عَلِمَ، اِعْتَبَرَ، جَعَلَ، اِتَّخَذَ

They fall into three meaning groups, and grouping them is easier than listing them:

• thinking it likely - ظَنَّ “think, suppose”، حَسِبَ “reckon, take to be”. The speaker may be wrong, and very often is.
• knowing or judging it to be so - عَلِمَ “know (to be)”، وَجَدَ “find (to be)” report a fact; رَأى “deem, regard”، اِعْتَبَرَ “consider, count as” deliver a verdict.
• turning X into Y - جَعَلَ “make into”، اِتَّخَذَ “take as”. Here the second accusative is a result, not an opinion.

3. Both objects are مَنْصوب, and the order is fixed. The first accusative is the thing being judged, the second is the judgement: اِعْتَبَرْتُ القَرارَ خَطَأً. Swap them and you swap the meaning, exactly as in English “I consider the decision a mistake” versus “I consider the mistake a decision”. The second object may also be a phrase or a clause, and then there is no ending to change: ظَنَنْتُهُ في المَكْتَبِ. / ظَنَنْتُ الطَّبيبَ يَنْتَظِرُني.

The first object is frequently an attached pronoun - ظَنَنْتُهُ نائِمًا “I thought he was asleep”، وَجَدْتُها مُغْلَقَةً “I found it (fem.) closed”. The pronoun is doing accusative duty just as in lesson 29; only the second word still shows a visible fatḥa.

4. Half of these verbs lead a double life. رَأى، وَجَدَ، عَلِمَ and جَعَلَ are ordinary transitive verbs too, and with one object they mean something quite different:

Nothing marks the difference except the number of accusatives. Count them before you translate.

5. اِعْتَبَرَ is the workhorse of the press. It is a Form VIII verb from the root عبر (اِعْتَبَرَ، يَعْتَبِرُ), and modern journalism runs on it: اِعْتَبَرَتِ الصَّحيفَةُ القَرارَ خَطَأً فادِحًا. Learn its passive at the same time, because the passive is even commoner than the active. When a two-object verb goes passive, the first object becomes نائِبُ الفاعِل and goes مَرْفوع, while the second stays مَنْصوب exactly where it was (lesson 52):

اِعْتَبَرَ الخُبَراءُ الاتِّفاقَ ناجِحًا. → يُعْتَبَرُ الاتِّفاقُ ناجِحًا.

That lone accusative trailing a مَرْفوع subject - يُعْتَبَرُ المَشْروعُ ضَرورِيًّا - is one of the most recognisable shapes on a news page.

6. The escape hatch: أَنَّ. The verbs of thinking and judging - ظَنَّ، حَسِبَ، عَلِمَ، وَجَدَ، رَأى، اِعْتَبَرَ - also accept a whole أَنَّ clause instead of two objects, and in modern prose that is the commoner construction:

ظَنَنْتُ الجَوَّ جَميلًا. = ظَنَنْتُ أَنَّ الجَوَّ جَميلٌ.

Watch the endings flip. Inside أَنَّ the rule of lesson 44 applies: the noun after أَنَّ is مَنْصوب (الجَوَّ) and its predicate is مَرْفوع (جَميلٌ). So the second word is accusative in the two-object version and nominative in the أَنَّ version - اِعْتَبَرَتِ الوِزارَةُ أَنَّ الخُطَّةَ ناجِحَةٌ. The two verbs of turning X into Y have no such route: جَعَلَ and اِتَّخَذَ need two real objects, and *جَعَلْتُ أَنَّ الغُرْفَةَ مَكْتَبٌ is not Arabic.

7. ظَنَّ is مُضَعَّف, and the doubling breaks. The root is ظنن, so the two نs sit under a shadda only while the ending begins with a vowel: ظَنَّ، ظَنَّتْ، ظَنّوا، يَظُنُّ. The moment an ending begins with a consonant the letters split and the stem vowel comes back - a fatḥa, because ظَنَّ is a فَعَلَ verb (lesson 49):

ظَنَّ → ظَنَنْتُ، ظَنَنّا، ظَنَنْتُمْ، ظَنَنَّ - never *ظَنَّتُ

The same split runs through the present with نونُ النِّسْوَة: يَظُنُّ but هُنَّ يَظْنُنَّ. This is the ظَلَّ / ظَلِلْتُ behaviour of lesson 43 with a different vowel.

8. The second family: كادَ plus a bare مُضارِع. كادَ وأَخَواتُها - traditionally أَفْعالُ المُقارَبَةِ والرَّجاءِ والشُّروع, “verbs of nearness, hope and beginning” - leave their subject مَرْفوع and take a verb as their predicate, always a مُضارِع:

كادَ الوَلَدُ يَسْقُطُ. - The boy almost fell. كِدْتُ أَسْقُطُ عَلى الدَّرَجِ. - I almost fell on the stairs.

كادَ means “was on the point of, nearly did” - and in the affirmative it means the thing did not happen. كادَ الوَلَدُ يَسْقُطُ says the boy stayed on his feet.

9. Who takes أَنْ and who refuses it. This is the whole practical difficulty of the family, and it splits three ways:

• كادَ - no أَنْ. كادَ يَسْقُطُ، كِدْنا نَصِلُ. The مُضارِع keeps its ordinary مَرْفوع ḍamma.
• أَوْشَكَ and عَسى - with أَنْ. أَوْشَكَ القِطارُ أَنْ يَتَحَرَّكَ. / عَسى المَطَرُ أَنْ يَتَوَقَّفَ. Because أَنْ is there, the verb goes مَنْصوب (lesson 34): يَتَحَرَّكَ, not يَتَحَرَّكُ.
• شَرَعَ، بَدَأَ، أَخَذَ - never أَنْ. شَرَعَ الطُّلّابُ يَكْتُبونَ. / بَدَأَ المُديرُ يَشْرَحُ الخُطَّةَ. / أَخَذَ الطِّفْلُ يَبْكي. *شَرَعَ أَنْ يَكْتُبَ is simply wrong.

The verbs of beginning have one extra restriction: in this construction they appear only in the past. You say شَرَعَ يَكْتُبُ, never *يَشْرَعُ يَكْتُبُ - for a present “he is beginning to” use بَدَأَ with a مَصْدَر: يَبْدَأُ العَمَلَ. Note too that جَعَلَ turns up here in the sense “he set about”: جَعَلَ يَضْحَكُ “he burst out laughing”. The same verb, two families.

10. أَوْشَكَ and عَسى. أَوْشَكَ (Form IV, root وشك) conjugates normally: أَوْشَكَ، أَوْشَكَتْ، أَوْشَكْتُ، يُوشِكُ. It is the neutral “be about to”, slightly more formal than كادَ and preferred when the event is expected rather than narrowly avoided.

عَسى is جامِد - frozen. It has no مُضارِع, no أَمْر and no participle, and in modern writing it is effectively invariable: عَسى أَنْ يَصِلَ الوَفْدُ غَدًا “the delegation may arrive tomorrow”. It expresses hope, and you will meet it far more often in reading than in your own writing - لَعَلَّ (lesson 44) does the same job more naturally. It also takes a glued pronoun, and in that one shape the أَنْ is usually dropped: عَساهُ يَعودُ.

11. لَمْ يَكَدْ … حَتّى - “no sooner … than”. Negated, كادَ normally reads the other way round: affirmative كادَ = it nearly happened but did not; negated كادَ followed by حَتّى = it barely did happen. Hence the fixed frame:

لَمْ يَكَدْ يَجْلِسُ حَتّى رَنَّ الهاتِفُ. - He had hardly sat down when the phone rang. ما كادَ يَبْدَأُ الحَديثَ حَتّى قاطَعَهُ أَحَدُهُمْ. - He had barely begun to speak when someone interrupted him.

ما كادَ and لَمْ يَكَدْ are interchangeable here. The same reading gives you the very useful لا أَكادُ أَسْمَعُكَ “I can hardly hear you” - I do hear you, only just. Outside these two frames a negated كادَ can equally be a flat denial (“was nowhere near doing it”), which is why grammarians have argued about it for centuries; read it from the context, not from a formula. Note that لَمْ puts the jussive only on يَكَدْ; the second verb keeps its ḍamma, because لَمْ reaches no further than the verb it precedes, and that مُضارِع is a clause of its own - the خَبَر of كادَ.

12. كادَ is أَجْوَف, and its stem shortens. The root is كود, and a long alif cannot stand in front of a sukūn (lesson 50), so before a consonant-initial ending it collapses to a short vowel: كادَ → كِدْتُ، كِدْنا، كِدْتُمْ. Note that the vowel is a kasra: a و root normally leaves a ḍamma (قالَ → قُلْتُ), but كادَ is underlyingly كَوِدَ / يَكادُ, the خافَ / خِفْتُ class, and that class keeps a kasra. The مُضارِع is يَكادُ with a fatḥa, and the jussive shortens it again: لَمْ يَكَدْ.

13. كادَ أَنْ - the honest note. You will constantly see كادَ أَنْ يَسْقُطَ in newspapers and novels. Classical grammar disapproves - أَنْ with كادَ is described as قَليل, rare, and the bare مُضارِع is the standard - but modern usage has largely ignored that. So recognise كادَ أَنْ without flinching and write كادَ يَسْقُطُ, which nobody will ever mark wrong.

14. Not in this lesson. When a particle such as ما، لا or إِنَّ comes between ظَنَّ and its objects, the verb’s case work is suspended and the clause keeps its own endings. Grammarians call it التَّعْليق, and it belongs with the other faces of ما in lesson 68. The rarer classical members of both families (كَرَبَ، طَفِقَ، حَرى، اِخْلَوْلَقَ، خالَ، زَعَمَ) belong to lesson 70’s register.

Forms & patterns

ظَنَّ - the doubled paradigm, where the split is visible

(The dual forms are left out; they keep the shadda - ظَنّا، ظَنَّتا، يَظُنّانِ.)

The eight sisters of ظَنَّ

One sentence, three operations

كادَ - the hollow paradigm

The كادَ family and the أَنْ question

Examples

Common mistakes

• ظَنَنْتُ الجَوَّ جَميلٌ. -> ظَنَنْتُ الجَوَّ جَميلًا. - The old خَبَر does not keep its ḍamma. Both halves of the swallowed sentence go مَنْصوب, and the second is the half learners forget.
• ظَنَنْتُ الجَوُّ جَميلًا. -> ظَنَنْتُ الجَوَّ جَميلًا. - The first half too. A leftover nominative is the reflex of كانَ وأَخَوات, and this is the wrong family for it.
• ظَنَّتُ أَنَّ الخَبَرَ صَحيحٌ. -> ظَنَنْتُ أَنَّ الخَبَرَ صَحيحٌ. - ظَنَّ is doubled: a consonant-initial ending splits the shadda. (ظَنَّتْ, with a sukūn on the ت, is the perfectly good “she thought”.)
• ظَنَنْتُ أَنَّ الجَوَّ جَميلًا. -> ظَنَنْتُ أَنَّ الجَوَّ جَميلٌ. - Once أَنَّ is in the sentence, أَنَّ governs, not ظَنَّ: accusative noun, nominative predicate (lesson 44). You cannot use both systems at once.
• وَجَدْتُ البابَ مَفْتوحٌ. -> وَجَدْتُ البابَ مَفْتوحًا. - If وَجَدَ has two complements it is a ظَنَّ-sister, and the second one takes a fatḥa.
• شَرَعَ أَنْ يَكْتُبَ الرِّسالَةَ. -> شَرَعَ يَكْتُبُ الرِّسالَةَ. - The verbs of beginning refuse أَنْ, and with no أَنْ there is nothing to make the مُضارِع مَنْصوب.
• كادَ الوَلَدُ يَسْقُطَ. -> كادَ الوَلَدُ يَسْقُطُ. - The manṣūb ending arrived without its cause. كادَ takes a bare مُضارِع, and a bare مُضارِع is مَرْفوع.
• كادْتُ أَسْقُطُ. -> كِدْتُ أَسْقُطُ. - كادَ is hollow: the long alif collapses before the consonant-initial ـْتُ, and the vowel it leaves behind is a kasra, because كادَ conjugates like خافَ / خِفْتُ, not like قالَ / قُلْتُ.
• أَوْشَكَ القِطارُ يَتَحَرَّكُ. -> أَوْشَكَ القِطارُ أَنْ يَتَحَرَّكَ. - أَوْشَكَ with a bare مُضارِع is allowed and does occur, but أَنْ is the normal modern shape by a wide margin; keep أَنْ, and keep the subjunctive after it.
• يَعْسى الطَّقْسُ أَنْ يَتَحَسَّنَ. -> عَسى الطَّقْسُ أَنْ يَتَحَسَّنَ. - عَسى is جامِد: no *يَعْسى, no أَمْر, no participle. The past-shaped form carries every meaning.

Tips

• Use the sentence test before writing a double accusative: strip the verb off and see whether the two words stand alone as جُمْلَة اسْمِيَّة. الجَوُّ جَميلٌ passes, so ظَنَنْتُ الجَوَّ جَميلًا is safe.
• When the two accusatives make you hesitate after a verb of thinking or judging, take the أَنَّ route. It is more frequent in modern prose anyway, and its endings are the ones you drilled in lesson 44. (جَعَلَ and اِتَّخَذَ have no أَنَّ route - they need two real objects.)
• Memorise the أَنْ split as a three-beat chant - كادَ never, أَوْشَكَ and عَسى always, the beginning verbs never. Most errors in the second family are a wrong answer to that one question.
• One stem test covers both families, the one from lessons 49 and 50: does the ending begin with a consonant? If yes, ظَنَّ → ظَنَنْ and كادَ → كِدْ. If no, leave the stem alone.
• Learn اِعْتَبَرَ together with يُعْتَبَرُ - the passive is the shape you meet first in any newspaper.
• Keep the two families apart by what follows the subject: كانَ وأَخَوات take a noun predicate in the accusative, كادَ وأَخَوات take a verb. كانَ الوَلَدُ نائِمًا against كادَ الوَلَدُ يَنامُ.
• Ignore for now: التَّعْليق, the frozen relatives كَرَبَ and طَفِقَ, and the dual paradigms. Eight ظَنَّ-verbs and the seven in the كادَ table is the whole working set.