דיבור ודיון: spoken Hebrew in a qualification that does not examine speaking

Candidates on qualifications with an oral component spend a large share of their preparation on it. You have those hours free. Allocate them against the actual mark weightings rather than by habit.

1. 01Do not simply write more essays

   The instinct is to add essay practice. But Paper 2 is 20 per cent of the qualification and Papers 1 and 3 together are 80, and within those, reading and listening comprehension carry 35 marks each. Adding essays to a plan that is already essay-heavy improves the smallest component.

2. 02Work out the real weights first

   Of 400 scaled marks: Paper 1 is 170, of which the research essay is 40 raw and therefore 80 scaled; Paper 3 is 150; Paper 2 is 80. So the research-project essay alone is worth as much scaled as the whole of Paper 2. That single arithmetic fact should reshape most revision plans, and almost nobody does the arithmetic.

3. 03Allocate a week accordingly

   A defensible weekly shape for the second year: a third of the time on listening, because AO1 is 20 per cent and listening is the slowest skill to build; a third on wide reading, because AO2 is 30 per cent and reading feeds everything; and a third on the research project, because it is the largest single task. The two literary essays get their own concentrated block nearer the exam, because they depend on knowing two works rather than on general fluency.

4. 04Insert the spoken activities as tools, not as a subject

   Fifteen minutes of reading aloud folds into the reading block. Shadowing folds into the listening block. Saying a paragraph before writing it folds into the writing block. None of these is a separate item on the timetable, because none of them is separately assessed - they are the way the assessed work is done better.

5. 05Build in the check

   Once a month, mark your own week against the weights: did listening actually get a third, or did it get the twenty minutes left at the end? Revision plans drift towards whatever is most comfortable, which for most candidates is reading. The check takes five minutes and it is the only thing that keeps the plan honest.

Result: A weekly allocation that follows the mark weightings rather than habit - a third to listening, a third to wide reading, a third to the research project, with the literary essays in a concentrated block - and the spoken activities folded in as tools inside those blocks rather than treated as a fourth subject.

You have five minutes a day and a Hebrew article. Design the session so that it produces information rather than just practice.

1. 01Read it silently first, and mark the words you cannot resolve

   Before you say anything, read the paragraph and underline every word whose vowels you are not sure of. That list is the whole point of the exercise: reading aloud will force you to commit on each of them, and the commitments you get wrong are your syllabus.

2. 02Read aloud without stopping

   Do not correct yourself mid-sentence. Stopping to fix a word destroys the phrasing, and phrasing is half of what reading aloud is training. Record the whole paragraph in one pass.

3. 03Listen back with the text and mark three categories

   Words you mispronounced; words you hesitated on; words you read fluently but wrongly stressed. The three categories have three different cures: the first is vocabulary, the second is recognition speed, and the third is a pattern you have learned wrong and will keep repeating until you notice it.

4. 04Fix the third category first

   A wrong stress pattern is invisible to you and self-reinforcing, so it will not fix itself with more exposure. Look the word up, say it correctly ten times, and put it on a list you re-read weekly. The first two categories improve with ordinary reading; the third does not.

5. 05Re-record the same paragraph a week later

   The comparison is the diagnostic. If the same words are still wrong, the fix did not take; if new words are wrong, you have moved on. Five minutes a day plus one comparison a week is a complete pronunciation programme for a qualification that does not examine pronunciation - and it will show up in your Paper 3 marks.

Result: A five-minute daily routine that produces a weekly list sorted into vocabulary problems, recognition-speed problems and learned-wrong stress patterns, with the third treated first because it is the only one that does not fix itself.

You are about to write four sentences. Say each aloud first and diagnose what the ear catches.

1. 01Sentence one: the English possessive

   You want the school's library. The English structure pushes you towards הספרייה של בית הספר, which is grammatical, or towards a word-for-word attempt with the possessive in the wrong place. Say ספריית בית הספר aloud and it sounds like Hebrew, because the construct chain is the natural form. The ear chose correctly before the analysis did.

2. 02Sentence two: the missing et

   You want I read the book. Say קראתי הספר aloud. Something is missing, and with enough exposure you will hear it as clearly as an English speaker hears I read book. Say קראתי את הספר and it settles. The et before a definite direct object is one of the most-lost AO3 marks and one of the most audible errors.

3. 03Sentence three: the English present continuous

   You want I am writing an essay. The English pushes towards a two-part construction with an auxiliary. Say אני כותב חיבור and hear that Hebrew needs no auxiliary at all, because the participle is the present. Attempting an auxiliary produces a sentence that does not sound like anything.

4. 04Sentence four: the English preposition

   You want I am waiting for him. English attaches for to wait; Hebrew attaches ל to מחכה. Say אני מחכה לו aloud and it settles; say a form with a separate pronoun and it does not. Preposition choice is a lexical property of the verb in both languages and it does not transfer, so the ear is a better guide than the logic.

5. 05Notice what the ear could not do

   In none of these four did the ear tell you whether a word was spelled with tet or tav, or with alef or ayin, because those pairs are homophones. That is exactly the boundary between the two checks: the audible pass catches structure, the mechanical pass catches spelling, and you need both.

Result: Four sentences in which the ear corrected a possessive structure, a missing object marker, a spurious auxiliary and a wrong preposition - and one clear statement of what the audible check cannot do, which is why it never replaces the six-point mechanical check.

Your intended essay claim is that a change in Israeli society happened because of a particular cause. Run the challenge aloud.

1. 01State the claim in one sentence, aloud

   Not a paragraph, a sentence. If it takes more than one sentence you do not yet have a claim, you have a topic - and an essay built on a topic rather than a claim cannot reach the top AO4 band, which asks for arguments and conclusions rather than coverage.

2. 02Question one: what is your evidence

   Answer aloud, and count the pieces. If the answer is a general impression, or a single example, the claim is not yet supportable and the fix is research, not rewriting. If the answer is a body that publishes relevant data plus a dated institutional fact, you have the two kinds of evidence this specification most rewards.

3. 03Question two: what would someone who disagreed say

   This is the question candidates skip and examiners notice. Answer it in the other person's own terms, at its strongest, aloud. If you cannot produce a serious version of the counter-position, you do not understand the argument well enough to write about it, and the essay will read as one-sided however well it is written.

4. 04Question three: what would change your mind

   Answer with a specific kind of evidence, not with never. This is the question that converts an opinion into a judgement, and its answer usually becomes a sentence in the essay - כדי להכריע בשאלה דרושים נתונים על, to decide the question, data on such-and-such would be needed - which is a top-band evaluation move.

5. 05Listen back and mark the hesitations

   Play the recording and mark every pause longer than a second. Those are the joints. In a written draft they would be invisible, because writing lets you go back and smooth them; spoken, they are exactly where the argument does not hold. Fix them before drafting and the essay will be better than any amount of redrafting could have made it.

Result: A one-sentence claim tested against three questions that map onto the AO4 grid, with the recording used to locate the weak joints - and with the answer to the third question converted directly into a top-band evaluation sentence for the essay.

You speak Hebrew at home. Run the four-list audit on a timed 300-word piece and turn the result into a revision plan.

1. 01List one: the merged pairs

   Go through word by word and check every alef, ayin, het, khaf, tet, tav, kuf, kaf, samekh and sin. Count the errors. If there are more than five in 300 words, spelling is your largest single AO3 loss and the fix is a word list, not more writing - because more writing at the same accuracy just produces more errors.

2. 02List two: register

   Mark anything you would say to a friend but would not read in a newspaper: כאילו, בקיצור, a dropped אשר where the sentence wanted formality, a colloquial verb where a written one exists. Register errors cluster, so if you find three you will find ten, and the fix is reading rather than drilling - specifically, reading the register you need to produce.

3. 03List three: formal structures

   Does the piece contain at least one construct chain, one passive or reflexive binyan, one relative clause and one formal connective? If not, it is a transcript of speech in written characters, and the AO3 range descriptor will not be satisfied however accurate it is. The fix is the six-structure drill, deliberately placed.

4. 04List four: claim and evidence

   Underline every sentence that makes a claim and every sentence that supports one. If the second set is empty, the piece describes rather than argues, and the AO4 grid will place it in the middle bands regardless of the language. The fix is the three-question challenge, not more vocabulary.

5. 05Convert the audit into a plan, honestly

   Rank the four lists by number of failures and spend your time in that order. The uncomfortable result, for most heritage speakers, is that vocabulary - the thing they are strongest at and enjoy most - should get the least time, and spelling and argument the most. A plan that does not feel slightly wrong is usually a plan that avoided the audit's finding.

Result: A four-list audit of one timed 300-word piece producing a ranked list of failures, with the specific fix attached to each - a word list for spelling, targeted reading for register, the six-structure drill for range, and the three-question challenge for argument - and the deliberate discipline of spending most time on the weakest list rather than the most enjoyable one.