Accidence and syntax

The grammar list: what 3.3 contains and how it is examined#

FoundationAQA 7677 section 3.3 Grammar — the general principle of this list is to cover all forms which may be examined in Paper 1AQA 7677 section 3.3.2 Grammatical terminology — Term, Acceptable alternative/s, Examples of unacceptable alternative/sAQA 7677 section 3.2 — as well as all the grammatical forms defined in Grammar, students should study more complex grammatical forms where they appear in the set text

The architecture of section 3.3

Fig. 1The two halves of section 3.3 mirror each other sub-section for sub-section: everything you must be able to form has a matching list of what you must be able to call it. The emphasised branch is 3.3.1.4, Pointing, the five bullets that generate the 5-mark pointing question in Paper 1 Section A and underwrite the pointing column of the composition mark scheme.

Section 3.3 opens by telling you exactly what it is for: the general principle of this list is to cover all forms which may be examined in Paper 1. Everything in it is therefore assessable in the language paper, and the list is the only closed thing in the qualification, because there is no defined vocabulary list and the passages are unseen. The list has two halves doing two different jobs. Section 3.3.1, Accidence, is what you must be able to form and to recognise: five sub-sections and thirty-two bullets in all, six on nouns, adjectives and pronouns, nine on verbs, eight on words formed with prefixed and suffixed morphemes, five on pointing and four on syntax, followed by a further list of ten syntactic constructions that a Paper 1 passage may contain for translation. Section 3.3.2, Grammatical terminology, is what you must be able to name: five sub-sections and seventy-three terms. One hundred and fifteen items in all, and every one of them countable.

The terminology half is laid out as a three-column contract and repays reading as one. The first column, Term, is the vocabulary the examination itself will use, either inside the question asked or as part of the answer required. The second column lists acceptable alternatives, and the specification states that students will not be penalised for correctly using any of them. The third column lists unacceptable alternatives, and it is almost entirely the Hebrew grammatical vocabulary that a student taught in Hebrew already owns: shoresh for root, semichut for construct form, avar and atid for perfect and imperfect, tsivui for imperative, niqqud for pointing, pa'al and nif'al for qal and niphal, dagesh kal for weak dagesh. The specification adds a specific warning that phonetic spellings of pa'al are likely to be confused with other conjugations and that candidates should learn the recommended spelling or use qal or kal. One concession runs the other way, and it is generous: phonetic spellings in English letters will be accepted as long as the intended meaning is clear, so kamats and qamets both pass, and so do hifil and hiphil.

Two things sit outside 3.3 and govern how you use it. First, section 3.2 extends the requirement for the set texts. As well as everything in 3.3 you must study more complex grammatical forms where they appear in the set text; be able to offer up to two explanations of the meaning of any complex, ambiguous or uncommon word or phrase, supported by contextual and grammatical justification; know the derivations of nouns which are derived from verbal forms and of proper nouns; know background information about incidents, characters, places and people where it is relevant to a complete understanding of the storyline; and know the approximate locations of important places on maps of the region. Second, the grammar is examined in two directions. Paper 1 Sections A and B ask you to recognise and explain; Section C asks you to produce, and to produce pointed. A form you can parse but cannot write is worth half of what the specification actually wants from you.

A word about how Hebrew is printed in these notes, because 7677 makes pointing an examined skill twice over and the decision cannot be left implicit. Wherever pointing is the teaching point the Hebrew here is written fully pointed: every paradigm form, every vowel pattern, every composition answer, and every word whose pointing carries the argument. Verb roots are written unpointed, following the convention the Core Vocabulary List uses for itself, where roots are marked with a root sign and are not pointed. Consonantal skeletons in the pointing exercises are unpointed on purpose and are labelled so each time. The accents, the te'amim, are described in English but never printed: the specification examines their effect on pointing, in the rules about milra and milei stress and in pausal forms, but nowhere asks a candidate to write an accent, so printing them here would teach a skill that is not assessed and add a class of error that no reader could check.

The specification also settles the presentation of Divine names, and these notes follow its own convention, because it is what will be printed on the question paper. Section 3 states that the specification follows the Orthodox Jewish tradition: the four-letter Divine name is written יי, and other Divine names are written as they appear in the Hebrew Bible except that a dash is added between the first and second letters, so that the specification's own examples are אֱ-לֹהִים, אֲ-דֹנָי and צְ-בָאוֹת. It then adds, in its own words, that you may use this approach but are also free to approach the qualification from a Christian or other perspective. Nothing grammatical changes either way. A dashed name is pointed, inflected and construed exactly as an undashed one, and the two yods stand for a word whose reading aloud is a matter of tradition rather than of grammar.

The terminology contract: what to write and what not to write

Fig. 2A sample of the three-column contract in section 3.3.2, in transliteration; the specification prints the unacceptable column in Hebrew letters as well. The emphasised cell is the one that catches most candidates: dagesh lene and dagesh qal are both accepted for weak dagesh, but the phrase beged kefet dagesh is not.

Worked example

Parsing four forms into the specification's own terminology

Constructed teaching Hebrew, not a quotation. Parse each of the following as an examiner would want it named, using only terms from section 3.3.2: (a) וַיֹּאמֶר (b) הַדְּבָרִים (c) לִשְׁמֹר (d) נִשְׁמַר

(a) the form that carries biblical narrative
Segment before you name. The word divides as וַ plus יֹּאמֶר. The verb is the imperfect of the root אמר, third person masculine singular, in the qal; the prefixed וַ with its patah and the strong dagesh in the following yod is the vav consecutive. Name it in that order: vav consecutive with the imperfect, qal, third person masculine singular of אמר. The specification accepts vav conversive as an alternative for vav consecutive but not vav hahipuch; and note that although the form is an imperfect, the vav consecutive gives it past reference, which is why calling the imperfect a future tense is only an acceptable shorthand and never a translation instruction.
(b) an article, a plural and a dagesh that means something
The word divides as הַ plus דְּבָרִים. The noun is דָּבָר, masculine, in the plural absolute; the prefix is the definite article, pointed with patah, and the strong dagesh in the dalet is the doubling that the article always produces where the letter can take it. Note what changed in the noun itself: the qamats under the dalet of דָּבָר has reduced to a vocal sheva because the stress has moved to the plural ending. Name it: definite article plus the masculine plural absolute of דָּבָר, the words or the things.
(c) an infinitive, and which infinitive
The word divides as לִ plus שְׁמֹר. The base is the qal infinitive construct of שמר; the prefix is the inseparable preposition לְ, which takes hiriq here because the following letter carries a vocal sheva. Two infinitives exist in Biblical Hebrew and the specification lists them separately, so the answer must choose: this is the infinitive construct, the one that takes prepositions and pronominal suffixes, not the infinitive absolute, which does not. The specification lists both mekor and the bare word infinitive as unacceptable, so write infinitive construct in full.
(d) a prefixed nun that is not a preposition
There is no prefix to strip here: the nun is part of the verb pattern. נִשְׁמַר is the niphal perfect, third person masculine singular of שמר, he was kept or he was guarded. The niphal is marked in the perfect by a prefixed nun with hiriq, and its usual force is passive or reflexive; the specification also lists interactive, with reciprocal as an acceptable alternative, among the terms you may need. Nifal is an acceptable spelling; the Hebrew word is not. Compare שָׁמַר, qal perfect third masculine singular, he kept, and the difference between the two answers is one binyan and one voice.
The shape of a full-mark answer
A grammar question will not ask you to feel a form; it will ask you to state it. The reliable format is three clauses and no prose: the form, its full grammatical category, and where relevant the word that governs it or the word it agrees with. For (a) that reads: vav consecutive with the imperfect, qal, third masculine singular of אמר, the main verb of its clause. Adding a translation costs nothing and often secures the mark when the terminology is imperfect; adding an explanation of why you chose it costs time you do not have in a three-hour paper.

Result: (a) Vav consecutive with the qal imperfect, third person masculine singular of אמר: and he said. (b) Definite article plus the masculine plural absolute of דָּבָר: the words, or the things; the dagesh in the dalet is strong. (c) Qal infinitive construct of שמר with the inseparable preposition לְ: to keep. (d) Niphal perfect, third person masculine singular of שמר: he was kept.

Typical mistakes

Answering a grammar question in the Hebrew metalanguage learned at school — shoresh, avar, atid, tsivui, semichut — every one of which 7677 prints in its unacceptable column.
Treating 3.3 as a reading list. It is a production list too: Section C is worth 50 marks and asks you to write these forms, pointed, without a dictionary.
Assuming the set texts contain only what 3.3 lists. Section 3.2 says the opposite and asks you to explain the more complex forms as well.

Active revision

Print section 3.3 and mark it up. Against each of the thirty-two accidence bullets write R if you can only recognise the form and P if you can also produce it pointed. Against each of the seventy-three terms write the English term, the acceptable alternative you would actually use, and the word from the unacceptable column that you currently use instead. Then count the R marks and the unacceptable words. Those two lists, in that order, are your revision programme, and unlike the vocabulary they are finite.

Active recall

Recall the key points — then reveal.

Sources: AQA A-level Biblical Hebrew 7677 specification, section 3.3 Grammar (AQA) · GCE AS and A level subject content for ancient languages (Department for Education)

Nouns, adjectives and pronouns: gender, number and state#

FoundationAQA 7677 section 3.3.1.1 — gender, number and state of regular and irregular nouns; nouns formed with preformative letters and afformative nun; agreement of adjectives; comparatives and superlatives; cardinal and ordinal numbers; pronominal suffixesAQA 7677 section 3.3.2.1 — construct form, preformative, cardinal, ordinal, dual form, mappiq, pronominal suffix, singular, plural, comparative, superlative

A Biblical Hebrew noun carries three pieces of information and the specification names all three: gender, number and state. Gender is masculine or feminine and is not always visible, because a feminine noun need not end in ה or ת: אֶרֶץ, land, and עִיר, city, are feminine without any marker, and their agreement gives them away. Number is singular, plural or dual, and the dual is a listed term: it is used for natural pairs and for some measures and times, יָדַיִם, hands, עֵינַיִם, eyes, שְׁנָתַיִם, two years. State is the distinction with no English equivalent and therefore the one that costs marks. A noun in the absolute state stands alone; a noun in the construct state is bound to the noun that follows it and means of it. דָּבָר is a word; דְּבַר הַמֶּלֶךְ is the word of the king. The construct form is usually shorter and always unstressed, and only the last noun in a chain can take the article.

The construct is worth a paragraph of its own because it does the work that English does with of and with the apostrophe, and because the specification names it twice, once in the accidence list and once in the terminology list, where the Hebrew term semichut is marked unacceptable. Learn the visible changes. In the singular, a long vowel usually shortens: בַּיִת becomes בֵּית, as in בֵּית הַמֶּלֶךְ, the house of the king. In the masculine plural the ending ־ִים becomes ־ֵי: דְּבָרִים becomes דִּבְרֵי, and בָּנִים becomes בְּנֵי, so that בְּנֵי יִשְׂרָאֵל is the sons, or children, of Israel. In the feminine singular the ending ־ָה becomes ־ַת: תּוֹרָה becomes תּוֹרַת. A chain can run to three or four members, and definiteness is carried by the last: since only the last noun may take the article, בֵּית הַמֶּלֶךְ is definite throughout, and there is no way to say a house of the king without rephrasing.

Nouns are also formed by adding letters to a root, and the specification lists exactly which: preformative מ, א and ת, and afformative נ. The preformative mem is the commonest and the most useful to recognise, because it generates whole classes of meaning from a known root: מִזְבֵּחַ, altar, from a root meaning to slaughter for sacrifice; מִשְׁפָּט, judgement, from שׁפט, to judge; מַלְאָךְ, messenger. The term to write is preformative, with noun prefix as the accepted alternative and introductory marked unacceptable. Adjectives agree with their noun in gender, number and definiteness, and the position tells you the syntax: הַמֶּלֶךְ הַגָּדוֹל, with the article on both words, is the great king, an attributive phrase; גָּדוֹל הַמֶּלֶךְ, with the article on the noun only and the adjective first, is a verbless clause, the king is great. That difference is a whole clause, and it is decided by two dots and a word order.

Comparison has no separate adjective form, so Hebrew builds it out of syntax. The comparative uses the preposition מִן: an adjective followed by מִן plus a second noun means more so than that noun, and the same construction with טוֹב, good, gives better than. The superlative is made either by the definite article, so that הַגָּדוֹל in the right context is the greatest, or by a construct chain in which a noun governs its own plural, the pattern behind שִׁיר הַשִּׁירִים, the song of songs, that is, the finest of songs. Numbers are the other place where the system surprises. Cardinal and ordinal are both listed terms. The numbers three to ten reverse the expected agreement: the form that looks masculine is used with feminine nouns and the form that looks feminine with masculine nouns, so שְׁלֹשָׁה אֲנָשִׁים is three men and שָׁלֹשׁ נָשִׁים is three women. One and two agree normally: אֶחָד and אַחַת, שְׁנַיִם and שְׁתַּיִם.

Pronominal suffixes are the last item in 3.3.1.1 and the one that appears on every page of narrative, because the specification says they attach to nouns, verbs and prepositions alike. On a noun the suffix is possessive: דְּבָרוֹ, his word, דְּבָרִי, my word, דְּבָרֵנוּ, our word, and on a plural noun the linking vowel changes, so דְּבָרָיו is his words and דְּבָרַי is my words. On a preposition the suffix is the object: לִי to me, לוֹ to him, לָהּ to her, לָהֶם to them. On a verb the suffix is the direct object, though the same job is more often done by the object marker אֵת with its own suffixes, אֹתִי, אֹתוֹ, אֹתָם. One dot decides one of the commonest ambiguities on the page: the mappiq, a dot inside a final ה, marks that letter as a consonant rather than a vowel, so אַרְצָהּ with mappiq is her land while אַרְצָה without it is to the land. Mappiq is a listed term and mapik is its accepted spelling.

Worked example

Reading a noun phrase four ways

Constructed teaching Hebrew, not a quotation. Translate each phrase and state the gender, number and state of each noun, and for the adjectives whether they are attributive or predicative: (a) דִּבְרֵי הַמֶּלֶךְ (b) הָעִיר הַגְּדוֹלָה (c) גְּדוֹלָה הָעִיר (d) שְׁלֹשָׁה בָנִים וְשָׁלֹשׁ בָּנוֹת

(a) a construct chain
דִּבְרֵי is the masculine plural construct of דָּבָר: the absolute plural is דְּבָרִים, and the construct replaces the ending ־ִים with ־ֵי and re-points the stem. הַמֶּלֶךְ is masculine singular absolute with the definite article. Because only the last noun of a chain may carry the article, the article on הַמֶּלֶךְ makes the whole phrase definite. Translate: the words of the king, or in the right context the acts of the king, since דָּבָר covers both word and matter.
(b) the adjective inside the phrase
עִיר is feminine singular, and it is one of the nouns with no feminine ending, so the agreement is the only evidence: the adjective is גְּדוֹלָה, feminine singular, not גָּדוֹל. Both noun and adjective carry the article, which is the mark of an attributive phrase. Translate: the great city. Note that the article before the ayin is הָ with qamats and no dagesh, because gutturals cannot be doubled.
(c) the same two words as a clause
Here the adjective comes first and carries no article, while the noun carries one. That is the predicative pattern, and Biblical Hebrew makes a complete clause out of it with no verb at all. Translate: the city is great. The specification lists verbless clauses among the constructions a Paper 1 passage may contain, and this is the commonest kind. Supply the English verb to be; do not look for a Hebrew word to translate it from.
(d) the numbers that reverse
בָּנִים, sons, is masculine, and it takes שְׁלֹשָׁה, the form with the feminine-looking ending ־ָה. בָּנוֹת, daughters, is feminine, and it takes שָׁלֹשׁ, the shorter form. This chiastic agreement applies to the numbers three to ten and to no others; one and two agree normally. Note also that בָנִים here has lost its dagesh after the number, which stands in construct to it, and that the vav conjunctive before שָׁלֹשׁ keeps its plain sheva, since shin is not a labial letter and carries a full vowel of its own. Translate: three sons and three daughters.
What to write in the answer
Name the state explicitly, because that is the term the specification is testing: masculine plural construct, not simply plural. Name the agreement evidence when you identify gender on a noun that has no feminine ending, because that is what shows the examiner you have read rather than guessed. And when you meet a predicative adjective, say the words verbless clause, which is the specification's own term.

Result: (a) The words of the king: דִּבְרֵי masculine plural construct, הַמֶּלֶךְ masculine singular absolute and definite. (b) The great city, attributive: הָעִיר feminine singular absolute, הַגְּדוֹלָה feminine singular, both definite. (c) The city is great, a verbless clause with the adjective predicative. (d) Three sons and three daughters, with the chiastic number agreement of three to ten.

Typical mistakes

Translating a construct chain word by word and losing the of, or trying to put the article on the first noun of a chain, which Hebrew does not allow.
Assuming a noun without a feminine ending is masculine. אֶרֶץ, עִיר, יָד and נֶפֶשׁ are all feminine and their adjectives and verbs show it.
Reading a final ה with a mappiq as the locative ending, so that her land becomes to the land and the possessor disappears from the sentence.

Active revision

Take twenty lines of narrative from set text 1 and rule four columns: noun, gender, number, state. Enter every noun. For each construct form, write the absolute form beside it and mark what changed — a shortened vowel, a changed ending, a lost article. Then find every adjective and decide from its article and its position whether it is attributive or predicative, and translate the phrase or clause accordingly. Finally, circle every final ה on the page and say for each whether it is a vowel letter, a locative ending, a feminine ending or a consonant carrying a mappiq.

Active recall

Recall the key points — then reveal.

Sources: AQA A-level Biblical Hebrew 7677 specification, sections 3.3.1.1 and 3.3.2.1 (AQA)

The verb: aspect, mood and the seven binyanim#

StandardAQA 7677 section 3.3.1.2 — all persons and numbers of the perfect and imperfect tenses/aspect including the imperative; cohortative, jussive and polite imperative; participles in all conjugations; formation and function of qal, niphal, piel, hiphil and hithpael; commonly occurring formations of hophal and pual; commonly occurring formations of polel, hithpolel, pilpel and hithpalpel; infinitive construct; infinitive absoluteAQA 7677 section 3.3.2.2 — thirty terms including root, aspect/tense, perfect, imperfect, cohortative, jussive, imperative, emphatic imperative, binyan, the seven binyan names, active, passive, causative, reflexive, interactive, participle, qal passive participle

The seven binyanim, in transliteration

Fig. 3The specification requires qal, niphal, piel, hiphil and hithpael in full and hophal and pual only in their commonly occurring formations; the four minor patterns polel, hithpolel, pilpel and hithpalpel are likewise required only where they commonly occur. Transliteration is used here because the figure engine has no right-to-left support; the pointed Hebrew templates are in the prose above. The emphasised cell is the doubled second root letter, the single most reliable visual mark of the piel and pual.

The specification hedges its own terminology in a way that tells you what the system is: it writes perfect (aspect/tense) and imperfect (aspect/tense), and it lists past tense and future tense as acceptable alternatives while marking the Hebrew avar and atid unacceptable. The hedge matters. The perfect, קָטַל, presents an action as a completed whole and usually corresponds to an English past; the imperfect, יִקְטֹל, presents it as incomplete, and corresponds to an English future, a present habitual, or a modal may and should, depending on context. Neither is a tense in the Latin sense, which is why the vav consecutive can attach to an imperfect and produce past narrative without contradiction. The perfect is inflected by suffixes throughout: קָטַלְתִּי I killed, קָטַלְתָּ you killed, קָטְלוּ they killed. The imperfect is inflected by prefixes and, in some persons, by suffixes as well: אֶקְטֹל I shall kill, תִּקְטֹל you or she will kill, יִקְטְלוּ they will kill, נִקְטֹל we shall kill.

Around the imperfect stand three short forms that the specification names separately and examines as a group. The imperative, קְטֹל, is the plain command and exists only in the second person; the specification accepts command as an alternative and rejects tsivui. The cohortative is a first-person form, usually lengthened with ־ָה, expressing let me or let us, as in אֶשְׁמְרָה, let me keep. The jussive is a third-person form, usually shortened, expressing let him or may he. And the emphatic imperative, listed with the alternatives softened imperative, softening ה, extended imperative and polite imperative, is a second-person imperative with a ־ָה added, as in שָׁמְרָה beside the plain שְׁמֹר: the added he does not change the person or number but softens the demand, which is why English needs a please or a do. Negative commands do not use the imperative at all: לֹא with the imperfect states a permanent prohibition, of the kind used in law, while אַל with the jussive states an immediate one, as in the recurring formula אַל־תִּירָא, do not be afraid.

The binyanim are the heart of the system and the specification grades them precisely. Five are required in full — qal, niphal, piel, hiphil and hithpael, whose formation and function must be known in the prescribed aspects. Two more, hophal and pual, are required only in their commonly occurring formations. Four minor patterns — polel, hithpolel, pilpel and hithpalpel — are likewise required only where they commonly occur, and they exist chiefly because hollow and geminate roots cannot form a normal piel and hithpael. The pattern of the whole is regular: qal is simple and active, niphal is its passive and reflexive, piel is intensive or factitive with the second root letter doubled, pual is the passive of piel, hiphil is causative, hophal is the passive of hiphil, and hithpael is reflexive or interactive, the specification's own term, with reciprocal as an accepted alternative. The 3ms perfect templates are קָטַל, נִקְטַל, קִטֵּל, קֻטַּל, הִקְטִיל, הָקְטַל and הִתְקַטֵּל.

Participles and infinitives are listed separately and are examined separately. The specification requires participles in all conjugations, used as verbs, nouns or adjectives, and singles out the qal passive participle, which has its own entry in the terminology list with qal passive as the accepted alternative and passive participle marked unacceptable. The qal active participle is קֹטֵל, killing or one who kills, and it is one of the commonest forms in narrative and in prophecy; the qal passive participle is קָטוּל, killed. Participles have no tense: their time comes from the clause around them. The two infinitives do different jobs. The infinitive construct, קְטֹל, takes prepositions and pronominal suffixes and behaves like a verbal noun: לִשְׁמֹר to keep, בְּשָׁמְרוֹ in his keeping, that is, when he kept. The infinitive absolute, קָטוֹל, takes nothing and is used to reinforce a finite verb of the same root or to stand in for one, and it is a form a candidate should learn to recognise rather than expect to build.

Two habits make the verb system usable at speed. The first is to read the prefix and the vowel before you read the root, because those two things fix the binyan and the aspect: a prefixed ה with hiriq and a long hiriq in the stem is a hiphil perfect, a prefixed נ with hiriq is a niphal perfect, a prefixed הִתְ is a hithpael, a doubled second root letter with hiriq under the first is a piel. The second is to name the function alongside the form, because grammar questions ask for both. Hiphil is causative, so הִקְטִיל is not simply a variant of killed but he caused to kill; niphal is passive or reflexive, so a niphal answers the question what happened to the subject; and hithpael is reflexive or interactive, so it answers what the subject did to itself or what two parties did to each other. A parsing that names the binyan without naming its force has answered half the question.

Worked example

Producing four named forms of one root, pointed

Constructed teaching Hebrew, not a quotation. From the root שמר, to keep or guard, write and point: (a) the qal perfect third person masculine plural; (b) the qal imperfect first person common plural; (c) the niphal perfect third person masculine singular; (d) the hiphil perfect third person masculine singular of a root you may substitute if שמר is awkward. Then say what each means.

(a) qal perfect, they kept
The 3cp perfect ending is ־וּ and it is stressed, which pulls the stress off the stem and reduces the first stem vowel to a vocal sheva. From שָׁמַר, he kept, the plural is שָׁמְרוּ. Note what happened: the patah under the mem has gone entirely, because the syllable is now open and unstressed, and the qamats under the shin survives because it stands in a syllable that is still pretonic. Writing שָׁמַרוּ would be a spelling error and, in a composition, a pointing error too.
(b) qal imperfect, we shall keep
The imperfect is built with prefixes. The first person common plural prefix is נִ, and the qal imperfect stem vowel of a strong verb is holam: נִשְׁמֹר. Two checks. The shin carries a silent sheva because it closes the first syllable, and the holam sits under the mem as the stem vowel. Do not confuse this form with the niphal perfect נִשְׁמַר: the prefixed nun is the same letter but the stem vowels differ, holam against patah, and the sheva against a full vowel.
(c) niphal perfect, he was kept
The niphal perfect prefixes נִ to the stem and points the stem with patah: נִשְׁמַר. Its force is passive here, he was kept or he was guarded, but the niphal is also the reflexive of many roots, so the answer to a grammar question should name the binyan and then state the force in this clause rather than reciting passive by rule. Where the root begins with a guttural the prefix vowel lengthens to tsere, because a guttural cannot carry the doubling the pattern implies in the imperfect.
(d) hiphil perfect, he caused to keep
The hiphil prefixes הִ and points the stem with a long hiriq written with yod: הִשְׁמִיר. With שמר that verb is not common, which is exactly why a substitution is honest teaching: take מלך, to reign, whose hiphil הִמְלִיךְ means he made king, and the causative force is unmistakable. Note the pattern in both: prefixed he with hiriq, silent sheva on the first root letter, long hiriq with yod on the second. In the imperfect the he disappears and leaves its vowel behind, which is why יַמְלִיךְ, he will make king, must be told apart from יִמְלֹךְ, he will reign, by the patah and the hiriq rather than by any visible he.
The habit that makes this fast
Read prefix and vowel before root. A prefixed he with hiriq plus long hiriq in the stem is hiphil; a prefixed nun with hiriq plus patah is niphal perfect; a prefixed הִתְ is hithpael; a doubled second root letter with hiriq under the first is piel. Four visual tests cover the great majority of finite forms you will meet in an unseen, and they work before you have identified the root, which is the order in which you actually need them.

Result: (a) שָׁמְרוּ, qal perfect 3cp, they kept. (b) נִשְׁמֹר, qal imperfect 1cp, we shall keep. (c) נִשְׁמַר, niphal perfect 3ms, he was kept. (d) הִשְׁמִיר in pattern, and for a common root הִמְלִיךְ, hiphil perfect 3ms of מלך, he made someone king.

Typical mistakes

Translating every perfect as a past and every imperfect as a future, which produces nonsense the moment a vav consecutive or a conditional clause appears.
Confusing the two infinitives. Only the infinitive construct takes prepositions and suffixes, and the specification lists them as separate items with separate terms.
Calling a piel an intensive and stopping there. Many piel verbs are simply factitive or denominative, and the safer answer names the binyan and gives the verb's actual force in the clause.

Active revision

Build a single sheet for the root שמר. Write out, pointed, the qal perfect and imperfect in all persons and numbers, the imperative in all four forms, the cohortative and jussive, both participles, both infinitives, and the 3ms perfect of the other six binyanim. Then take a chapter of set text 2 and, for the first forty finite verbs, write the binyan, the aspect, the person, number and gender, and a three-word gloss. Every form you had to look up goes on a second sheet, which is your revision list.

Active recall

Recall the key points — then reveal.

Sources: AQA A-level Biblical Hebrew 7677 specification, sections 3.3.1.2 and 3.3.2.2 (AQA)

Weak roots: what happens when a radical will not behave#

AdvancedAQA 7677 section 3.3.1.2 — weaknesses in verb roots; the following contracted or irregular forms of verbs: pe-yod, pe-nun, ayin-vav, lamed-he, pe and ayin guttural, geminate, irregularities of first root letters in hithpaelAQA 7677 section 3.3.2.2 — root, first root letter, second root letter, third root letter, geminate

The eight weak classes the specification names

Fig. 4These eight classes are the whole of what section 3.3.1.2 requires under weaknesses in verb roots and contracted or irregular forms. The diagnostics in the last column are what you can see on the page before you have identified the root, which is the order in which they are actually useful. The emphasised cell is the dagesh that a vanished nun leaves behind, the single commonest weak-root signal in narrative prose.

Almost every verb a candidate meets in Samuel, Kings, Jeremiah or the Psalms has a weak root, so the weak classes are not an appendix to the verb system but the bulk of it. The specification names eight, using the traditional labels built on the paradigm root פעל: pe-yod and pe-nun, where the first root letter misbehaves; ayin-vav, the hollow verbs; lamed-he, where the third root letter is a he; pe-guttural and ayin-guttural; geminate, where the second and third root letters are the same; and irregularities of first root letters in hithpael. The terminology list gives the English names to use — first root letter, second root letter and third root letter — with pe-, ayin-, lamed-, first radical, second radical and third radical, and the Roman numerals I, II and III, all accepted as alternatives. The principle underneath is single: a weak letter cannot carry what the pattern demands of it, so something else in the word absorbs the change, and that change is what you see.

The first-letter classes are the easiest to diagnose because the damage happens at the front of the word. In pe-nun verbs the nun assimilates into the following letter whenever it would close a syllable, and leaves a strong dagesh behind as its receipt: the qal imperfect of נפל, to fall, is יִפֹּל, and the dagesh in the pe is the vanished nun. In pe-yod verbs the yod either becomes a vowel or drops: the qal imperfect of ישׁב, to sit or dwell, is יֵשֵׁב, with a tsere where the yod stood, and the imperative and infinitive construct lose the yod altogether, giving שֵׁב and שֶׁבֶת. The most consequential member of this class is הלך, to walk or go, which behaves like a pe-yod verb although its first letter is a he: its imperfect is יֵלֵךְ and its vav consecutive form וַיֵּלֶךְ, which opens more sentences in the historical books than almost any other word.

The hollow verbs, ayin-vav, have no stable middle radical at all, so their surface forms often show only two letters and their vowel is the whole of their morphology. From קום, to arise, the qal perfect is קָם, the imperfect יָקוּם, the participle קָם, the infinitive construct קוּם and the vav consecutive form וַיָּקָם. From בוא, to come or go in, the perfect is בָּא and the imperfect יָבוֹא. Two traps follow. First, the perfect and the participle can look identical, and only the clause decides between he arose and arising. Second, in the hiphil the hollow verbs take a long tsere or hiriq and can be mistaken for a qal, so הֵקִים, he set up, must be told from קָם, he arose, by the prefixed he. Lamed-he verbs behave in the opposite way: the he is not a real third radical but a vowel letter, so it disappears the moment anything is added. From עשׂה, to do or make, the perfect is עָשָׂה, the plural עָשׂוּ, the imperfect יַעֲשֶׂה and the vav consecutive form וַיַּעַשׂ, in which the third radical has gone entirely.

The guttural classes are the ones the pointing chapter of the specification cares about most, because gutturals refuse doubling and refuse a simple vocal sheva. A guttural as first root letter takes a hataf vowel instead of a vocal sheva, which is why the qal imperfect of עמד, to stand, is יַעֲמֹד rather than the expected form with a plain sheva, and why the prefix vowel changes from hiriq to patah to match. A guttural as second root letter cannot take the strong dagesh that piel, pual and hithpael require, so either the preceding vowel lengthens to compensate — the compensatory lengthening the terminology list implies when it names the guttural letters — or the doubling is simply left virtual. The guttural letters the specification names are alef, he, het and ayin, together with resh, which shares their refusal of the dagesh. Geminate roots, where the second and third letters are identical, collapse the pair into one letter with a strong dagesh whenever a vowel follows: from סבב, to go around, the qal imperfect is יָסֹב and the plural יָסֹבּוּ, where the dagesh in the bet is the second bet.

The last class the specification names is a small one with a large payoff: irregularities of first root letters in hithpael. The hithpael prefix הִתְ places a tav immediately before the first root letter, and when that letter is a sibilant the two swap places, a change called metathesis. From שׁמר the hithpael is הִשְׁתַּמֵּר rather than the expected order; from צדק it becomes הִצְטַדֵּק, where the tav has both moved and hardened to tet under the influence of the emphatic tsade. Where the first root letter is ד, ט or צ the tav may instead assimilate entirely, leaving a dagesh. The practical rule for reading is short: if you see a tav that seems to be in the wrong place near the front of a verb, suspect a hithpael and look for the prefixed he. The practical rule for writing, which matters in Section C, is shorter still: if you are composing a hithpael of a root beginning with a sibilant, put the sibilant first.

Worked example

Diagnosing four surface forms back to their roots

Constructed teaching Hebrew, not a quotation. For each form, give the three root letters, name the weak class, say what the weakness did, and translate: (a) וַיִּפֹּל (b) וַיֵּלֶךְ (c) וַיָּקָם (d) וַיַּעַשׂ

(a) the dagesh that used to be a nun
Strip the vav consecutive and you are left with an imperfect יִפֹּל with a strong dagesh in the pe. Two letters of the root are visible, pe and lamed, and the dagesh accounts for the third: a nun that closed the first syllable and assimilated into the following consonant. The root is נפל, to fall, a pe-nun verb. Translate: and he fell. The same mechanism explains יִתֵּן from נתן, to give, where the dagesh in the tav is the assimilated nun, and where the third root letter, also a nun, assimilates at the end of the word in other forms.
(b) the verb that pretends to be pe-yod
The visible letters are lamed and kaf, with a tsere in the stem and a segol before the final kaf. The root is הלך, to walk or go, which is treated as a pe-yod verb throughout: its imperfect is יֵלֵךְ, its imperative לֵךְ, its infinitive construct לֶכֶת. The initial he has gone. Translate: and he went. This is worth memorising rather than deriving, because הלך is one of the most frequent verbs in the historical books and its forms will not yield to the pe-he rules a candidate might expect.
(c) the hollow verb in narrative
Only two root letters are visible, qof and mem, and the qamats between them is doing the work of the missing middle radical. The root is קום, to arise, an ayin-vav verb. The vav consecutive form shortens the imperfect יָקוּם to וַיָּקָם, with the stress drawn back to the first syllable and the long shuruk replaced by a qamats. Translate: and he arose. Note the contrast with the hiphil, where the same root gives וַיָּקֶם, and he set up or he raised, and where the difference in English is transitive against intransitive.
(d) the lamed-he verb that loses its third letter
The visible letters are ayin, sin and nothing else. The root is עשׂה, to do or make, a lamed-he verb, and in the vav consecutive form the final he has disappeared entirely, as it always does when the form is shortened. The ayin, being a guttural, has attracted a patah both before and after it, which is the pe-guttural rule doing its work in the same word. Translate: and he did, or and he made. The plural is וַיַּעֲשׂוּ, where the hataf patah reappears because a vowel now follows the ayin.
Turning the diagnosis into an answer
A grammar question on any of these will ask for the root and often for an explanation. The three-clause format works here too: root letters in dictionary order, then the class, then the change. For (a): the root is נפל, a pe-nun verb, and the nun has assimilated into the pe, leaving a strong dagesh. That is one sentence and it earns everything the question can pay.

Result: (a) נפל, pe-nun, the nun assimilated leaving a strong dagesh: and he fell. (b) הלך, treated as pe-yod, the initial he lost: and he went. (c) קום, ayin-vav hollow, the middle vav vocalised then shortened: and he arose. (d) עשׂה, lamed-he with a guttural first radical, the final he lost: and he did, or and he made.

Typical mistakes

Failing to reconstruct an assimilated nun, so that יִפֹּל is filed under an unknown root beginning with pe instead of under נפל.
Reading a hollow verb's participle as its perfect, or the reverse, and getting the time of the whole clause wrong.
Writing a hithpael of a sibilant root in the expected letter order in the composition, which produces a form that does not exist.

Active revision

Take the first fifty finite verbs of set text 1 and sort them into the eight classes the specification names, with a ninth column for strong verbs. Count each column. For every form in a weak column, write the three root letters in their dictionary order and say in one clause what the weakness did to the form: assimilated, vocalised, dropped, doubled, lengthened, metathesised. Then reverse the exercise: from each root, produce the vav consecutive third person masculine singular, pointed, without looking back at the text.

Active recall

Recall the key points — then reveal.

Sources: AQA A-level Biblical Hebrew 7677 specification, section 3.3.1.2 (AQA)

Prefixed and suffixed morphemes: the letters that change a word#

StandardAQA 7677 section 3.3.1.3 — inseparable prepositions and the preposition from; definite article; interrogative he; vav conjunctive; vav conversive/consecutive; gentilic yod; locative he; paragogic lettersAQA 7677 section 3.3.2.3 — definite article, interrogative he, locative he, vav conjunctive, vav consecutive, gentilic yod

Biblical Hebrew stacks up to four morphemes in front of a word and one or two behind it, and the specification lists eight kinds of them. Because they are letters rather than separate words, a single graphic word can contain a conjunction, a preposition, an article, a noun and a suffix at once, and reading such a word means peeling it in order. The order is fixed and worth learning as an order: vav conjunctive first, then an inseparable preposition, then the article, then the word, then any suffix. So וּבְעָרֵי, in the cities of, contains vav plus bet plus the plural construct of עִיר, and וּבָעִיר contains vav plus bet plus an article the preposition has swallowed plus the noun. Segmentation is the first skill and it is testable directly: a grammar question that asks you to explain a form is very often asking you to name the morphemes in it.

The three inseparable prepositions are בְּ in, at or with, לְ to or for, and כְּ like or as, and they change shape predictably. Before a word beginning with a vocal sheva they take hiriq, so לְ plus שְׁמֹר gives לִשְׁמֹר. Before a hataf vowel they take the corresponding short vowel, so לְ plus עֲשׂוֹת gives לַעֲשׂוֹת, to do. Before the definite article they swallow the he and take its vowel, so בְּ plus הַמֶּלֶךְ gives בַּמֶּלֶךְ and לְ plus הָאָרֶץ gives לָאָרֶץ. The specification lists a fourth, the preposition from, separately because it behaves differently: מִן normally assimilates its nun into the following letter and leaves a strong dagesh, giving מִמֶּלֶךְ from a king, but before a guttural, which cannot take the dagesh, the vowel lengthens instead, giving מֵאִישׁ from a man. Written in full it is joined to its noun by a maqaf, מִן־הָעִיר.

The definite article is a single he with patah and a strong dagesh in the following letter, הַמֶּלֶךְ, and its variations are pure pointing rules. Before alef, ayin and resh, which cannot be doubled, the article's vowel lengthens to qamats, giving הָאִישׁ, הָעִיר and הָרֹאשׁ. Before he and het the vowel stays patah and the doubling is simply left virtual, giving הַהוּא and הַחֹשֶׁךְ. Before an unaccented guttural pointed with qamats the article becomes segol, giving הֶהָרִים, the mountains. Beside the article stands its lookalike, the interrogative he, which turns a statement into a yes-or-no question and is normally pointed with hataf patah, as in the question at Genesis 4:9, הֲשֹׁמֵר אָחִי אָנֹכִי, am I my brother's keeper. Telling the two apart is a pointing decision and nothing else: hataf patah with no dagesh asks a question, patah with a dagesh makes a noun definite.

The vav is the busiest letter in the language and the specification names two of its uses. The vav conjunctive means and, and is pointed with a plain sheva, וְ, except that before the labial letters bet, mem and pe and before any letter carrying a vocal sheva it becomes shuruk, giving וּמֶלֶךְ and a king and וּשְׁמוֹ and his name, and before a hataf vowel it takes the matching short vowel, giving וַאֲנִי and I. The vav consecutive is a different creature. Prefixed to an imperfect it is pointed with patah and doubles the following letter, וַיֹּאמֶר, and it converts the imperfect to past narrative; prefixed to a perfect it is pointed like an ordinary vav conjunctive, וְשָׁמַר, and converts the perfect to future or habitual sense, usually with the stress moved to the final syllable. The specification's term is vav consecutive, with vav conversive accepted and vav hahipuch unacceptable. Where the following letter is a yod with a sheva no dagesh is possible, which is why the commonest narrative opening of all is written וַיְהִי.

Three smaller morphemes complete the list and each is a single letter with a single job. The gentilic yod turns a place or a people into an adjective of belonging: עִבְרִי, a Hebrew, מוֹאָבִי, a Moabite, and with the feminine ending מוֹאָבִיָּה, a Moabitess. The locative he, which the specification also allows to be called directional he, is an unaccented ־ָה added to a noun to mean motion towards: אַרְצָה, to the land, הָעִירָה, to the city, מִצְרַיְמָה, to Egypt. It is not a feminine ending and it is not the third person feminine suffix, and the mappiq is what tells you which is which, as in the pair אַרְצָה, to the land, against אַרְצָהּ, her land. Paragogic letters, the last bullet, are letters added without changing the grammatical value of the form: a paragogic nun on some imperfect plurals, giving יִשְׁמְרוּן beside יִשְׁמְרוּ, and a paragogic he lengthening a cohortative or an imperative. They should be recognised and stripped, not translated.

Worked example

Peeling four words down to their parts

Constructed teaching Hebrew, not a quotation, except where a reference is given. Segment each word, naming every morpheme, and translate: (a) וּבְעִיר (b) לָאָרֶץ (c) וַיְהִי (d) הֲשֹׁמֵר, as at Genesis 4:9

(a) two prefixes and a noun
The word is וּ plus בְּ plus עִיר. The first morpheme is the vav conjunctive, written as shuruk because it stands before a bet, one of the labial letters; the second is the inseparable preposition בְּ, in; the third is the feminine noun עִיר, city, in the singular absolute with no article. Translate: and in a city. Note what the absence of the article does: had the article been there the preposition would have swallowed it and the word would read וּבָעִיר, and in the city.
(b) a preposition that has eaten an article
The visible form is לָ plus אָרֶץ. There is no he on the page, but the qamats under the lamed is the article's own vowel, lengthened because alef cannot be doubled. So the morphemes are the inseparable preposition לְ, the definite article, and the feminine noun אֶרֶץ, land, whose segol has lengthened to qamats in this environment. Translate: to the land. A candidate who reads this as to a land has lost the definiteness, which in a narrative usually means losing the reference to a specific place.
(c) the commonest opening in biblical narrative
The morphemes are the vav consecutive, pointed with patah, and the qal imperfect third masculine singular of היה, to be, giving and it was or and it came to pass. The expected strong dagesh in the yod is absent because the yod carries a vocal sheva and cannot be doubled in that position. This word opens set text 1 itself: I Samuel 1:1 begins וַיְהִי אִישׁ אֶחָד, there was a certain man. Recognising it instantly is worth several seconds on every page of an unseen, because it announces that a narrative is starting and that what follows is past.
(d) a question mark made of one letter
The morphemes are the interrogative he, pointed with hataf patah, and the qal active participle of שׁמר, keeping or one who keeps. It is the opening word of the reply at Genesis 4:9, הֲשֹׁמֵר אָחִי אָנֹכִי, am I my brother's keeper. Everything turns on the pointing: hataf patah with no dagesh in the following letter is the interrogative; patah with a dagesh would be the definite article and the phrase would become the keeper of my brother. There is no other signal, because Biblical Hebrew has no question mark and word order alone will not decide it.
The order in which to peel
Always work from the outside in and in the fixed order: vav, preposition, article, word, suffix. Doing it in that order means you never mistake an article's vowel for a preposition's, and it makes the pointing rules do the identifying for you. Write the segmentation out with plus signs the first hundred times; after that you will see it.

Result: (a) vav conjunctive plus preposition bet plus עִיר: and in a city. (b) preposition lamed plus the definite article plus אֶרֶץ: to the land. (c) vav consecutive plus qal imperfect 3ms of היה: and it came to pass, the opening of I Samuel 1:1. (d) interrogative he plus qal active participle of שׁמר: am I keeping, or in context am I the keeper, Genesis 4:9.

Typical mistakes

Reading a vav consecutive with the imperfect as a future, so that a narrative chain is translated as a series of predictions.
Missing that a preposition has absorbed the article, so that in the city is translated as in a city and the definiteness of the whole clause shifts.
Translating a locative he as a feminine ending, or a paragogic nun as a plural suffix that is already there.

Active revision

Take fifteen lines of set text 2 and box every word with more than one prefix. For each, write the morphemes in order with a plus sign between them and name each one: vav conjunctive, inseparable preposition, definite article, and so on. Then rebuild each word from its parts without looking, pointed, and check. Finally, list every vav on the page and classify it as conjunctive or consecutive, giving the pointing evidence in each case.

Active recall

Recall the key points — then reveal.

Sources: AQA A-level Biblical Hebrew 7677 specification, sections 3.3.1.3 and 3.3.2.3 (AQA)

Pointing: syllables, sheva, the dagesh and the pausal form#

AdvancedAQA 7677 section 3.3.1.4 — syllables and vowels including sheva and hataf vowels; weak dagesh and strong dagesh including euphonic dagesh; the effect of the guttural and labial letters on pointing; the effect of the accents milra/milei on pointing; pausal formAQA 7677 section 3.3.2.4 — twenty-one terms including pointing, patah, kamats, segol, hataf, hiriq, tsere, holam, kubuts, shuruk, vocal sheva, silent sheva, open and closed syllable, accented and unaccented syllable, guttural and labial letters, weak, strong and euphonic dagesh

Telling a weak dagesh from a strong one

Fig. 5The whole of the distinction between the specification's three named dageshes reduces to two questions: is the letter one of the six, and does a vowel precede it. The emphasised node is the outcome candidates most often get wrong — a dot in a begadkephat letter after a full vowel is strong and doubles, not weak. Labels are in transliteration because the figure engine has no right-to-left support.

Pointing is not decoration in this qualification: it is examined directly for 5 marks in Paper 1 Section A and indirectly for a great deal more in the 50-mark composition, where the mark grid states that an answer with no attempt at pointing cannot gain more than 2 marks out of 5 for any phrase. Section 3.3.2.4 lists twenty-one terms, more than any other pointing sub-section in any A-level ancient language, and section 3.3.1.4 gives the five rules that generate them. The system rests on the syllable. A syllable is open if it ends in its vowel and closed if it ends in a consonant; it is accented or unaccented, for which the specification also accepts tone syllable, stressed syllable and unstressed syllable. The governing generalisation, which will carry you a long way, is that an open unaccented syllable takes a long vowel and a closed unaccented syllable takes a short one. Once you have that, most of the vowel changes in the paradigms stop being arbitrary.

The vowels themselves are ten signs with a strong preference for their own environments, and the specification fixes the spellings it expects while accepting phonetic variants. Patah ־ַ and kamats ־ָ are the a vowels, and the specification accepts patach and qamets. Segol ־ֶ and tsere ־ֵ are the e vowels, tsere also spelt zeire. Hiriq ־ִ is the i vowel, also chirik, and holam ־ֹ the o, also cholam. Kubuts ־ֻ and shuruk, written with a vav carrying a dot, are the u vowels. Sheva ־ְ is the reduced vowel and comes in two kinds that the specification names separately: vocal sheva, which is pronounced and counts as the vowel of an open syllable, and silent sheva, which is not pronounced and merely closes a syllable. The hataf vowels are sheva combined with a short vowel, and the term to write is hataf, with reduced, chataf, composite and compound accepted. They belong to the gutturals, which is where the next rule begins.

The guttural letters, which the specification names as alef, he, het and ayin together with resh, refuse two things: they refuse the dagesh, and they refuse a plain vocal sheva. Both refusals produce visible consequences. Because a guttural cannot be doubled, a pattern that requires doubling must either lengthen the preceding vowel to compensate or leave the doubling virtual: this is why the definite article is הָאִישׁ with qamats but הַחֹשֶׁךְ with patah, and why the niphal and piel of guttural roots look different from the paradigm. Because a guttural cannot take a plain vocal sheva, it takes a hataf vowel instead, and the vowel of the preceding syllable often shifts to match, which is why the qal imperfect of a pe-guttural root is יַעֲמֹד with patah and hataf patah rather than the expected hiriq and sheva. The labial letters, which the specification names as bet, vav, mem and pe, matter for one rule in particular: the vav conjunctive becomes shuruk before them.

The dagesh is a single dot with three jobs and the specification names all three. The weak dagesh, for which dagesh lene and dagesh qal are accepted but the phrase beged kefet dagesh is not, appears only in the six letters bet, gimel, dalet, kaf, pe and tav, and only where they are not preceded by a vowel: at the start of a word, or after a silent sheva. It hardens the sound and does not double the letter. The strong dagesh, dagesh forte or dagesh chazak, can appear in any letter except the gutturals and resh, and it does double: it is what you see after the definite article, in the vav consecutive, in the piel and pual and hithpael, and where a nun has assimilated. The euphonic dagesh, for which the specification allows the terms it prints as asei merachek and dechik, where correct, appears at the junction of two words to smooth the transition. Distinguishing weak from strong is therefore a two-question test: is the letter one of the six, and is it preceded by a vowel.

The last two bullets of 3.3.1.4 concern the accents, and they are the reason the accents are worth understanding without being worth writing. Hebrew stress falls either on the final syllable, called milra, or on the one before it, called milei, and the specification asks you to know the effect of that on pointing. It is considerable: stress attracts long vowels and reduces the unstressed syllables around it, so moving the stress rewrites the word. When a word stands at a major pause in a verse — under a strong disjunctive accent, typically at the middle or the end — it takes its pausal form, and the pausal form regularly lengthens the stressed vowel and may move the stress. So מַיִם becomes מָיִם in pause, and a segol may become qamats. The specification lists pausal form as a term, with in pause accepted. Two practical consequences follow: in an unseen, a pausal form tells you where the verse divides, which is free syntactic information; and in the pointing question, a word at a pause should be pointed in its pausal form.

Worked example

Justifying every point in one short clause

Constructed teaching Hebrew, not a quotation. The clause is וַיִּשְׁמֹר הַמֶּלֶךְ אֶת־הַדָּבָר, and the king kept the matter. Account for every vowel and every dagesh in it, naming each with the specification's own terminology.

וַיִּשְׁמֹר
The vav carries patah and the yod carries a strong dagesh: that pair is the vav consecutive, and the dagesh is strong because yod is not a begadkephat letter and because the pattern requires doubling. The yod then carries hiriq, the normal prefix vowel of the qal imperfect. The shin carries a silent sheva, closing the second syllable; the sheva is silent and not vocal because it stands at the end of a closed syllable after a short vowel. The mem carries holam, the qal imperfect stem vowel of a strong verb, and the syllable is closed by the resh and accented, which is why the holam does not reduce.
הַמֶּלֶךְ
The he carries patah and the mem a strong dagesh: that is the definite article, and the dagesh is the doubling the article always produces where the letter allows it. The mem then carries segol, and the lamed carries segol again, which is the segolate pattern: a noun stressed on the first syllable with a helping vowel in the second. The final kaf carries a silent sheva, closing the word, and it carries no dagesh, because it follows a vowel; had it stood at the beginning of a word it would have taken a weak dagesh, since kaf is one of the six.
אֶת־
The object marker is pointed with segol before a maqaf, where the full form אֵת with tsere appears when the word stands alone. This is a stress effect: joining two words with a maqaf makes the first lose its own accent, and a word without an accent shortens its vowel. That is the milra and milei rule of 3.3.1.4 in miniature, and it is why the maqaf is a pointing signal and not merely a hyphen.
הַדָּבָר
The he carries patah and the dalet a dagesh. Two claims are needed here and only one of them is obvious. The dagesh is in a begadkephat letter, so a candidate might call it weak; but it follows a full vowel, the article's patah, so it is strong, and it is doing the article's doubling. That single dot is the clearest illustration of why the two-question test exists. The dalet then carries qamats and the bet qamats again, which is the ordinary absolute-state pointing of דָּבָר, with the second qamats standing in the accented final syllable.
The count that the mark scheme would make
There are three dageshes in this clause, all strong: in the yod of the vav consecutive, in the mem of הַמֶּלֶךְ and in the dalet of הַדָּבָר. Beside them stand twelve vowels and shevas, of which two are silent shevas — fifteen separate points in four words. Under the Paper 1 pointing grid each wrong or missing sign is one error and each superfluous sign is one error, so a clause of this length carries roughly fifteen opportunities to be right. Two lines of narrative prose, the length the specification actually sets, carry something over a hundred.

Result: Three strong dageshes, none weak, and no euphonic dagesh: the vav consecutive doubling the yod, and the definite article doubling the mem of הַמֶּלֶךְ and the dalet of הַדָּבָר. The sheva under the shin is silent; the segol on אֶת is the shortened form produced by the maqaf; the holam of the imperfect stem stands in a closed accented syllable and therefore does not reduce.

Typical mistakes

Calling every dagesh in a begadkephat letter a weak dagesh. After a full vowel the same dot in the same letter is strong and doubles.
Putting a plain vocal sheva under a guttural, which is the single most frequent pointing error in composition and one that recurs across a whole answer.
Ignoring the pause, so that a pausal form is either mis-parsed as a different word in reading or mis-pointed as the contextual form in writing.

Active revision

Take three verses of narrative prose from set text 1 and mark every syllable as open or closed and as accented or unaccented, then check each vowel against the generalisation that an open unaccented syllable takes a long vowel and a closed unaccented syllable a short one, listing every apparent exception and explaining it. Next, circle every dagesh on the page and label it weak, strong or euphonic, giving the evidence. Finally, find the word at the main pause of each verse and write both its contextual and its pausal pointing.

Active recall

Recall the key points — then reveal.

Sources: AQA A-level Biblical Hebrew 7677 specification, sections 3.3.1.4 and 3.3.2.4 (AQA) · GCE AS and A level subject content for ancient languages (Department for Education)

Syntax: word order and the clauses a passage may contain#

AdvancedAQA 7677 section 3.3.1.5 — word order; relative clauses; oath clauses; interrogative clauses; and in addition ten syntactical constructions that examination passages may contain: relative, conditional, temporal, concessive and causal, purpose, result, verbless, asseverative, negative including legal prohibitions, oath and interrogative clausesAQA 7677 section 3.3.2.5 — parallelism, pausal form, subject, object, ellipsis

The clause types a Paper 1 passage may contain

Fig. 6The list is closed: section 3.3.1.5 states that passages selected for examination may contain any of these constructions, which makes them exhaustively preparable in a way the vocabulary is not. The emphasised cell is the one that most often goes wrong in translation, because the Hebrew oath idiom is inverted with respect to English.

The syntax bullets in 3.3.1.5 are short, but they are followed by a sentence that widens them considerably: in addition, passages selected for examination in the language component may contain any of ten named syntactical constructions. Those ten are relative clauses, common types of conditional sentence, common temporal, concessive and causal clauses, purpose clauses, result clauses, verbless clauses, asseverative clauses, common negative clauses including legal prohibitions, common oath clauses and common interrogative clauses. That list is a promise about what an unseen may contain, and it is short enough to prepare completely. The terminology list beside it adds only five terms, but they are the five you need to write an answer with: parallelism, pausal form, subject, object and ellipsis.

Word order comes first because it is the default against which everything else is a departure. Biblical Hebrew narrative puts the verb first, then the subject, then the object, and it links clause to clause with the vav consecutive: וַיֹּאמֶר הַמֶּלֶךְ, and the king said, not the king said. Departures from that order are meaningful. Fronting the subject before the verb usually marks a contrast or a change of topic, which in English is often best rendered with a but, an as for, or a change of paragraph; fronting the object marks emphasis. Because the verb comes first, the subject of a Hebrew clause is often not visible until the second or third word, so the discipline in an unseen is to read to the end of the clause before committing to a subject. Poetry, as section 3.3.2.5 hints by listing parallelism among the syntax terms, plays with this order far more freely than prose.

Four of the ten constructions have no verb of their own and are therefore the ones that catch a candidate reading word by word. The verbless clause simply juxtaposes subject and predicate and requires the English verb to be to be supplied: הָאִישׁ טוֹב, the man is good, and הַמֶּלֶךְ בַּבַּיִת, the king is in the house. The relative clause is normally introduced by אֲשֶׁר, which is invariable and means who, which or that regardless of the case it stands for, so the English relative pronoun has to be chosen from the sense: הָאִישׁ אֲשֶׁר שָׁמַר, the man who kept, but הַדָּבָר אֲשֶׁר שָׁמַר, the thing which he kept. Hebrew also forms relative clauses with no marker at all, especially in poetry, which is why the term the specification actually lists beside them is ellipsis: a word the reader is expected to supply. Recognising an ellipsis and naming it is a rewardable answer.

Conditional, temporal, causal, purpose and result clauses are all built from a small stock of particles and, in several cases, from the infinitive construct. The commonest conditional particle is אִם, if, with לוּ for the unreal condition; the apodosis is often introduced by a vav. Temporal clauses are frequently made with a preposition and the infinitive construct rather than with a conjunction, so that בְּשָׁמְרוֹ, literally in his keeping, means when he kept, and כְּשָׁמְעוֹ means when he heard, and this construction has no English equivalent and must be rebuilt as a clause. Causal clauses use כִּי, because, and phrases like יַעַן and עַל־כֵּן, therefore; purpose clauses use לְמַעַן or a vav with the jussive or cohortative; result clauses are often left to the vav consecutive chain, which does the work of and so. Concessive clauses, which the specification names alongside temporal and causal, use גַּם כִּי and similar phrases.

The last three constructions are the ones with a distinct biblical flavour and they are the ones a candidate is most likely to mistranslate. Negative clauses use לֹא with the imperfect for a permanent prohibition, which is the form used in the legal codes and which the specification signals when it says including legal prohibitions, and אַל with the jussive for an immediate one, as in the recurring formula אַל־תִּירָא, do not be afraid; verbless existence is negated with אֵין. Oath clauses are built on a formula in which the speaker calls on a life — the recurring pattern with חַי, as such-and-such lives, which runs through Samuel and Kings — followed by a clause introduced by אִם or אִם לֹא. The idiom is inverted relative to English: אִם in an oath means assuredly not, and אִם לֹא means assuredly. Asseverative clauses use כִּי to mean surely rather than because, and interrogative clauses use the interrogative he, or one of מִי, מָה and מַדּוּעַ. All of these are recognisable by particle, which is what makes the list preparable.

Worked example

Naming the clauses in a constructed period

Constructed teaching Hebrew, not a quotation. Read the following and, for each clause, name its type from the specification's list, name the word that introduces it, and translate. וַיְהִי בַּיָּמִים הָהֵם וַיֵּצֵא הַמֶּלֶךְ מִן־הָעִיר וַעֲבָדָיו עִמּוֹ׃ וַיִּרְאוּ אֶת־הָאֲנָשִׁים יֹשְׁבִים תַּחַת הָעֵץ וַיֹּאמֶר הַמֶּלֶךְ אֶל־עֲבָדָיו מִי הָאֲנָשִׁים הָאֵלֶּה׃

The opening formula
וַיְהִי בַּיָּמִים הָהֵם is the standard narrative opening: a vav consecutive with the imperfect of היה followed by a temporal phrase, in those days. Strictly this is not a clause type from the list but a temporal setting, and the English convention is either it came to pass in those days or, more economically, in those days. Note the pointing of בַּיָּמִים: the preposition has absorbed the article, which is why the bet has patah and the yod a strong dagesh.
The main narrative clause and its circumstantial tail
וַיֵּצֵא הַמֶּלֶךְ מִן־הָעִיר is a straightforward vav consecutive clause with the verb first, the subject second and a prepositional phrase third: and the king went out from the city. The verb is a pe-yod form of יצא. What follows, וַעֲבָדָיו עִמּוֹ, is a verbless clause used circumstantially: no verb, subject his servants, predicate the prepositional phrase with him. Translate as with his servants with him, or better, since English dislikes the repetition, and his servants were with him. Naming it a verbless clause is the answer the specification's terminology wants.
A participle doing the work of a subordinate clause
וַיִּרְאוּ אֶת־הָאֲנָשִׁים יֹשְׁבִים תַּחַת הָעֵץ has one finite verb, the vav consecutive imperfect of ראה, and then an object marked with אֶת followed by a masculine plural active participle agreeing with it. The participle is not a second main verb; it describes the state in which the object was seen. English can render it as a participle too, they saw the men sitting under the tree, or as a clause, they saw that the men were sitting under the tree. Either is accurate; the second is the one to use if the examiner has asked what the participle is doing.
Direct speech and an interrogative clause
וַיֹּאמֶר הַמֶּלֶךְ אֶל־עֲבָדָיו introduces direct speech in the usual way, with the verb of speaking and the addressee marked by אֶל. Biblical Hebrew has no quotation marks, so the direct speech simply begins. מִי הָאֲנָשִׁים הָאֵלֶּה is an interrogative clause of the verbless kind: the question word מִי, then a definite noun phrase with its demonstrative, and no verb at all. Translate: who are these men. Two of the specification's construction types therefore overlap in five words, which is normal and is worth saying in an answer.
Writing the analysis down
For a grammar question, name the type and the marker in the same clause: a verbless clause, with no verb, requiring the English to be; an interrogative clause introduced by mi; a circumstantial verbless clause attached to the preceding narrative clause. Three of the ten listed constructions appear in these two verses, which is a fair indication of the density an examination passage will have.

Result: Clause by clause: a temporal setting formula; a vav consecutive narrative clause; a circumstantial verbless clause; a narrative clause with an object complement participle; a speech-introducing clause; and a verbless interrogative clause. Translation: In those days the king went out from the city, and his servants were with him. And they saw the men sitting under the tree, and the king said to his servants, Who are these men?

Typical mistakes

Translating an oath clause literally, so that if I do this becomes a condition rather than the emphatic denial the idiom actually carries.
Rendering a preposition with the infinitive construct as a noun phrase, so that when he heard becomes something like in his hearing and the clause loses its verb.
Assuming אֲשֶׁר must be translated by who, when the case it stands for may require which, whose, where or a preposition plus which.

Active revision

Rule a page with the ten construction types the specification lists as column headings. Work through one chapter of set text 2 and enter every clause you meet under its heading, writing the Hebrew particle that introduced it and your English rendering. Any column still empty at the end of the chapter is a gap to fill from elsewhere in the set texts. Then reverse it: write one English sentence of your own for each of the ten types and translate it into Hebrew, pointed, which is exactly the Section C skill.

Active recall

Recall the key points — then reveal.

Sources: AQA A-level Biblical Hebrew 7677 specification, sections 3.3.1.5 and 3.3.2.5 (AQA)

Key points

Parsing four forms into the specification's own terminology

Key points

Reading a noun phrase four ways

Key points

Producing four named forms of one root, pointed

Key points

Diagnosing four surface forms back to their roots

Key points

Peeling four words down to their parts

Key points

Justifying every point in one short clause

Key points

Naming the clauses in a constructed period

Sources

Accidence and syntax

Key points

Parsing four forms into the specification's own terminology

Key points

Reading a noun phrase four ways

Key points

Producing four named forms of one root, pointed

Key points

Diagnosing four surface forms back to their roots

Key points

Peeling four words down to their parts

Key points

Justifying every point in one short clause

Key points

Naming the clauses in a constructed period

Sources