The following entry is divided into two sections: (i) an Introduction for the non-specialist and (ii) a detailed survey.
(I) HEBREW GRAMMAR: AN INTRODUCTION
There are four main phases in the history of the Hebrew language: the biblical or classical, the post-biblical or neo-classical and rabbinic (which includes medieval scholarly writings and continued until the latter part of the 19th century), and the modern. In biblical times Hebrew was a living, spoken language but, from the centuries immediately preceding the Christian era, it ceased to be the vernacular. Nevertheless, biblical Hebrew persisted as the language of the Scriptures and as a model for compositions of a devotional nature. Because it was transmitted from one generation to the next, over many centuries, as a written language which found oral expression only in pious recital, its structure became artificially fixed.
It is remarkable, however, that the basic structure of the language has remained constant throughout all its stages of development. In the post-biblical and modern phases there was a progressive accretion in vocabulary by the creation of new words in accordance with the inherent laws of the language and by borrowing. Yet, divergencies in grammar were, for the most part, not fundamental, but peripheral. Thus a general introduction to the Hebrew language would best be served by confining it to the biblical phase and, where relevant, by pointing out divergencies which appeared in the later stages. The scheme of biblical Hebrew grammar is derived from the literature of the Hebrew Bible, known as Masoretic (from מַסֹּרָה (massorâ) "tradition").
Understanding the Patterns of Biblical Hebrew Grammar
An understanding of the patterns of biblical Hebrew grammar, as opposed to the mechanical learning of a catalogue of seemingly irrational rules, may be achieved by recognizing that the formulation of these rules rests on three main principles:
1. the adoption of agreed conventional signs in writing to represent the spoken word, as it was traditionally transmitted and articulated;
2. deliberate adjustments in spelling, in conformity with any spontaneous modification in the articulation of the spoken word, due to natural fluctuations caused by inflection;
3. statements, in concise but adequate terms, of forms of Hebrew thinking, as expressed in speech.
The aim of this article is not to present a comprehensive scheme of Hebrew grammar, but to demonstrate that there is a rationality underlying it. To achieve this end, items of grammar will be selected to illustrate how the above three principles are translated into formal Hebrew grammar. Occasional analogies from other languages will be cited to show that, in other languages also, elements of grammar reflect articulated speech and thought processes.
The Hebrew alphabet consists of consonants only. The reader of a Hebrew consonantal text – if he was proficient in the language – automatically supplied the appropriate vowels, as determined by the context. Anyone familiar with English would know automatically whether the context of a sentence requires him to read the consonantal word r-d as "red," "rid," or "rod." Similarly, anyone who knows Hebrew well would immediately recognize from the context whether אם (ʾm) is to be read as אִם (ʾim, "if") or אֵם (ʾēm, "a mother"), or whether רבד(dbr) is to be read דָּבָר (dāḇār, "a word"), דֶּבֶר (deber, "a plague"), דִבֵּר (dibbēr, "he spoke"), or דֹבֵר (doḇēr, "speaking").
When Hebrew ceased to be a spoken language, the uninitiated were unable to supply the relevant vowels to a consonantal text. It was realized that some apparatus had to be devised to indicate vowel-sounds as aids to reading. There were two stages in the development of such an apparatus. The first was the employment of the weak letters י, ו, ה (h, w, y) to serve also as vowel indicators. ו (w) represented both the o and u sounds, so that מות (mwt) could be read as either מוֹת (môt, "death of") or מוּת (mût, "dying"). י (y) indicated both the i and e sounds; מי (my) could be read as either מִי (mî, "who?") or מֵי (mê, "waters of"). The terminal silent ה (h) generally indicated the a vowel-sound, but sometimes also the e sound, so that מה (mh) could be read as either מָה (mâ) or מֶה (meh), both meaning "what." This device was employed in the texts of the so-called Dead Sea Scrolls and is maintained even today in modern Hebrew writing.
Though this system reduced the area of possible error, it was clearly unsatisfactory, because it was not exact. In the eighth century c.e. the Tiberian system of vowel-points was devised to represent all the vowel-sounds, as traditionally held. This apparatus was generally adopted and is still in use.
Among Jews of European origin, the influence of their vernaculars on the articulation of Hebrew consonants led to the coalescing of several pairs of consonants. א and ע are both silent, כּ and ק are both pronounced as the same k sound, ב and ו are both articulated as v, ס, שׂ, and also ת among Ashkenazi Jews are pronounced as s, while in Israel the תּ and ת are both pronounced as t. However, it should be realized that, in biblical times, each Hebrew consonant had its own particular phonetic value, as is still the case among Oriental Jews in the recital of their sacred Hebrew texts. In ancient Israel there was a clear difference in sound between, for example, the word אִם (ʾim, "if"), and עִם (ʿim, "with"), the latter being articulated with a back-throated guttural sound. Similarly one could distinguish between אַתָּה (ʾattâ, "you," masc. sing.) and עַהָּה (ʿattâ, "now"). There was a clear distinction in sound between the words כֹּל (kōl, "all") and קוֹל (ḳôl, "a voice"), for the latter was articulated as a distinctive back-throated k.
The operation of the three above principles in the formulation of rules of Hebrew grammar can now be dealt with.
1a. When a vowelless letter stands at the beginning or in the middle of a word, the convention is to place two dots, vertically arranged (:), under it, as שְמוּאֵל (šmûʾêl, "Samuel") and יִצְחָק (yiṣḥāk, "Isaac"). This sign is known as שְׁוָא (šwāʾ, probably meaning "speed"), which for convenience is spelled shewa. It indicates that the letter under which it appears has no full vowel. It was found that, when articulating a syllable beginning with a vowelless letter, a quick, vowel-like sound was involuntarily induced (something like the quick e in the word "because"). The shewa under such a letter is known as vocal and is represented in transcription by a diminutive e; thus the first example is transcribed šεmûʾêl. At the end of a syllable in the middle of a word, as in יִצְ/חָק (yiṣ/ḥāk), the shewa under the vowelless letter is a silent ("quiescent") one.
The peculiarity of speech indicated by the vocal shewa suggests that the ancient Israelite could not articulate a word beginning with two consonants without involuntarily giving the first (i.e., the vowelless) letter a quick vowel-like sound. He would have pronounced the word black as bεlack. This peculiarity is shared by the Arabs, who would pronounce this word as either balack or iblack. In modern Hebrew, however, owing to the influence of European languages, there is no difficulty in articulating a word beginning with two consonants. The first example would be articulated as shmuel.
b. In the Hebrew alphabet there are six letters which under certain conditions are pronounced hard (b, g, d, k, p, t) and, in other situations, are pronounced soft (ḇ, g̠, ḏ, ḵ, p̠, ṯ). When such a letter is hard in speech, the convention in writing is to place a dot in it, called by grammarians דָּגֵשׁ (dageš, "piercing") lene, or light dageš (תּ, פּ, כּ, דּ, גּ, בּ). When such a letter is soft in the spoken word, it is left without any dot in it (ת, פ, כ, ד, ג, ב).
To the question as to when these letters are pronounced hard and when soft, the answer may be given that it seems that the vocal organs of the ancient Israelites were so conditioned that, when one of these letters began a syllable, and no vowel immediately preceded, they pronounced it hard, as פָּרַשׂ (pāraš, "he spread") and יִסְ/פֹּר (yis/pōr, "he will count"). It follows, then, that when one of these letters was at the end of a syllable, as יִפְ/רֹשׂ (yip̠rōś, "he will spread"), or in the middle of a syllable, as בְּכֹל (bεḵōl), or at the beginning of a syllable but with a vowel immediately preceding, as סָ/פַר (sā/p̠r, "he counted"), the letter was pronounced soft. When the word פָּרָה (pārâ, "cow"), whose initial letter is hard, receives the prefixed conjunction, it becomes וּפָרָה (ûp̠ārâ, "a cow"); the speaker automatically softens the letter after the vowel sound. In modern Hebrew, however, this rule is not always observed in fluent speech.
c. If a letter in the middle of a word is doubled in articulation, the convention is to write a single letter only, but with a dot in it. The word for "thief" is גַּנָּב and represents the spoken גַּנְנָב (gannāḇ). This dot, indicating a doubled letter, is known as dageš forte, or strong dageš, to distinguish it from the other dageš, the light one, which indicates the hard letter, as the גּ in this example. However, the six letters (b, g, d, k, p, t), when hard, may be doubled in the middle of a word, as שַׁבָּת, representing שַׁבְּבָּת (šabbāt), in which case the dageš is theoretically both lene and forte. It should be noted here that the guttural letters (ע, ח, ה, א), by their very nature of being either weak (inaudible) or throat letters, cannot be doubled in articulation, so that, in writing, they do not receive a dageš forte. Strangely enough, this also applies to the letter ר (r).
The doubling of a letter is by no means arbitrary; it is usually due to some natural phenomenon in speech. In anticipation an English composite word may be quoted, derived from Latin, with the negative prefix in-. The combination in-legal is articulated illegal and actually spelt phonetically. This is exactly what happens in Hebrew. When, for example, the two words מִן (min, "from") and שָׁם (šām, "there") are spoken together in the natural flow of speech, they become a composite word. The combination מִנשָׁם (minšām) is articulated as םִשׁשָׁמ (miššām) but written מִשָּם, with the dageš forte indicating the double letter. This phenomenon of speech is described in the following rule of grammar: "When a vowelless נ (n) stands between two vowelled consonants, it is assimilated to the consonant immediately after it, giving rise to the doubling of the latter."
2. In many English words the spelling has not kept pace with changes in pronunciation. The word daughter is pronounced dauter, but the persistence of the medial gh in the spelling suggests that originally it must have been articulated with a guttural-like sound, as in the parent German word tochter. The American spelling of some words shows an arrested attempt towards writing words phonetically: plough is spelled plow and though is sometimes written as tho, but the trend did not develop consistently. In Hebrew, however, every modification in the articulation of words, usually due to the effect of inflection or the presence of a peculiar letter, is faithfully reproduced in writing by corresponding adjustments in spelling.
a. The first example is the two-syllabled word דָבָר (dāḇār, "word, matter, thing"), which begins with an open syllable (i.e., one ending in a vowel) and has the stress on the second syllable. The plural is formed by attaching the ending ים–̣ (-îm) and the stress moves on to this new syllable at the end. The speaker, hurrying on to the stressed syllable at the end, quite naturally elides the vowel (ā) in the first (open) syllable. It becomes דְבָרִים (dεḇārîm), and not דָּבָרִים (dāḇārîm); the first syllable, now vowelless, is written with the vocal shewa. The modification in articulation is paralleled by the corresponding adjustment in the spelling. (An analogy of this phenomenon in speech is the English word médicine, which, with an accession at the end and the moving forward of the stress, becomes medicinal – almost mdícinal.)
b. The next example introduces a characteristic usage in Hebrew. In an expression such as "(the) word of Moses," דָּבָר (dāḇār, "word") and מֹשֶׁה (mōšeh, "Moses") are so closely associated that they become one compound idea; in fluent speech they are virtually one composite word and the stress is mainly on the second half of the composite word. The effect is similar to that in the example just quoted. That is, דָּבָר(dāḇār) has received an accession at the end and the stress has moved forward, so that the vowel in the first (open) syllable is elided and a secondary effect is that the vowel in the second (closed) syllable is shortened. The combination is pronounced דְּבַר־מֹשֶׁה (dεḇar-mōšeh) and written as such. The first noun is so dependent upon the second one that it is said to be in the construct state. This natural shortening of the vowel in the closed syllable of a word in the construct state is seen also in the combination of יָד (yāḏ, "hand") with מֹשֶׁה(mōšeh) in the expression "(the) hand of Moses" – יַד־מֹשֶׁה(yaḏ-mōšeh).
c. It was noted earlier that a vowelless נ (n) between two vowelled consonants is assimilated to the following letter which, in consequence, is doubled and that, in writing, this doubling is represented by a dageš forte – מִנשָׁם (minšām) becoming מִשָּם (miššām). When, however, the letter following the vowelless נ (n) is a guttural (ע, ח, ה, א) or ר which cannot be doubled in articulation, a natural adjustment is made. When the phrase "from a man" – מִן אָדָם (min ʾāḏām) – becomes a composite word in the flow of speech (hypothetically מִנְאָדָם, min'ādām), the vowelless נ (n) is assimilated, but the following letter cannot be doubled. The resultant form מִאָדָם, (miʾāḏām) leaves the first syllable open, i.e., ending in a vowel. Since the natural tendency was to pronounce an open syllable with a long vowel (unless that syllable was stressed, in which case the effect was the same), the short vowel (i) in the first syllable is automatically prolonged by the speaker to …(ē) and the combination becomes מֵאָדָם (mēʾāḏām), the spelling being adjusted to conform with the modification in speech.
d. The following example of a rule of grammar appears superficially to be irrational and yet, on examination, it reflects a normal fluctuation in speech which is represented phonetically by the written word. The possessives are expressed by particles suffixed to the noun as "a house is old," "my house is old." The feminine singular noun, such as תּוֹרָחִי (tôrâ, "Torah") with a suffix ("my Torah") becomes תּוֹרָתִי (tôrāṯi). How can one account for the apparent insertion of the letter ת (t) before the suffix? Arabic provides the perfect analogy.
In classical (literary) Arabic, nouns have three case endings. The feminine singular noun "city" is madīnatun (nominative), madīnatin (genitive), and madīnatan (accusative). In colloquial Arabic, however, the case-endings are dropped, leaving the form madīnat for all cases, but it is actually pronounced madīna (or medīna). The final t is not articulated (like the tendency in America to pronounce the word breakfast as breakfas). However, when this t is in the middle of a word and it has a vowel, as with a possessive suffix attached ("my city" being madīnatī) it is, of course, clearly articulated.
Scholars have pointed to the same phenomenon appearing in the transition from pre-biblical to biblical Hebrew. There are indications that, originally, Hebrew nouns had case-endings, like Arabic: the word for Torah was tôrāṯu (nom.), tôrāṯi (gen.), and tôrāṯa (acc.). The case-endings were dropped and the resultant form תּוֹרָת (tôrāṯ) was pronounced תּוֹרָה (tôrâ) and spelled that way. As with Arabic, it was the final ת (t) which was not articulated but, when it is medial with a vowel, it is, of course, clearly articulated – "my Torah" could only be תּוֹרָתִי (tôrāṯî). Because the final ת (t) was not articulated, it was dropped in spelling; when it is medial and audible, it is present in the spelling.
The tendency to drop a final t sound is present in other languages. In Ireland the well-known surname McGrath is actually pronounced McGra. The French say il est ("he is") pronounced il-ē but as a question it is est-il? ("is he?" – pronounced ēt-il), the medial vowelled t being quite naturally articulated. More striking still, and akin to the situation in Hebrew, is the French il a ("he has"), which, as a question, is a-t-il? ("has he?"), with the medial vowelled t articulated and reappearing in the spelling.
3. In biblical Hebrew the main idea of an expression is stated first and it is then qualified, limited in application or modified by what immediately follows. "My word" is, in Hebrew thinking, something like "word, mine" – דְּבָרִי (dεḇārî), the main idea being "word," which is first expressed and then limited to "mine" by the following particle י (î). "A big house" is, in the Hebrew order of thought, "a house (the main idea), a big (one)" (qualifying it) – בַּיִת גָּדוֹל (bayiṯ gāḏôl). It follows logically that the phrase "the big house" is, in Hebrew thinking, "the house, the big (one)"– הַבַּיִת הַגָּדול (habbayiṯ haggāḏôl) and "my big house" will be "house, mine, the big (one)" בֵּיתִי הַגָּדול(bêṯî haggāḏôl). The rule which is then enunciated is "the adjective follows the noun it qualifies (and agrees with it in gender and number), and if the noun is specific (as indicated by the definite article or the possessive suffix) the adjective following the noun has the definite article." In the same pattern of thought the adverb follows the noun it modifies; "exceedingly good" is "good, exceedingly" – טוֹב מְאֹד (tōḇ mεʾōd).
In the structure of the Hebrew verbal system one again detects characteristic thought processes. Whereas medieval and modern Hebrew adopted the European concept of past, present, and future tenses, in biblical Hebrew no such notion was formulated. Instead acts or states of being were viewed as either completed or incompleted. The completed state, referring to something finished or done, generally corresponds to the notion of the past tense but, with certain verbs, it may indicate a European present tense. The form זָכַרְתָּ (zāḵartā) could mean "you remembered" but, since the basic sense is "our state of remembering is completed," it could imply "you remember." The incompleted state, indicating something not yet finished or not yet done, generally refers to the future but with a few verbs it could imply a continuing present. תִּזְכֹּר (tizkōr) could mean "you will remember" but, since the basic notion is "our state of remembering is incompleted," i.e., it is still going on, the derived sense could be "you keep on remembering," that is "you are mindful of." Unhappily, grammarians have adopted the terms perfect and imperfect – as used for Latin and Greek conjugations of verbs – but these do not accurately represent the biblical Hebrew concept.
It is to be noted that in the completed state, it is the act or state of being which is regarded as the main idea. Thus the verbal element is expressed first and is limited or applied to the person (in the example given the particle תָּ – (tā) for אַתָּה – ʾattâ, "you," masc. sing.) which immediately follows. In the incompleted state, however, what seems to be more prominent in the mind of the speaker is the person who is about to do, or is in the process of doing, something. Thus the element representing the person (in the example the תּ, t) is stated first and the verbal element follows.
The Western notion of the present tense is represented in Hebrew by the participle, e.g., זֹכֵר (zōḵēr, "remembering"), preceded by the appropriate personal pronoun. "You remember" is, in Hebrew, אַתָּה זֹכֵר (ʾattâ zōkēr). Since the Hebrew participle is virtually a verbal adjective, the thought underlying that expression is "you (are) a remembering (person)."
There are seven forms in the Hebrew verbal system. The first may be regarded as basic and the other six as derived forms, as in the following scheme:
i. simple active: שָׁבַר (šāḇar, "he broke"). This is the completed state. It has also the incompleted state, participle, imperative, etc., and all these are conjugated with persons, numbers, and genders.
ii. simple passive, but, with some verbs, the passive has also something of a reflexive effect. Its form is נִשְׁבַּר(nišbar), meaning "he (or "it") was broken." However, from the simple active רָאָה (rāʾâ, "he saw") the derived form of this category נִרְאָה (nirʾâ) means "he was seen," but this produces the extended sense "he showed himself," that is, "he appeared."
iii. intensive active. Derived from the simple active שָבַר (šāḇar, "he broke") is the intensive form שִׁבֵּר (šibbēr) (with the middle root-letter doubled to express intensity) and the derived sense is "he smashed," "he shattered." It will be realized that a derived form in this category must produce a new idea by extension, so that only such verbs which lend themselves to such an extension, by which a new idea is derived, can be included in it.
iv. intensive passive. This is simply the passive of iii and its form is שֻׁבַּר (šubbar, "he (or "it") was smashed/shattered."
v. causative active. The notion of causative is present in a few English verbs. "To seat" is the causative of "to sit" and "to fell" is the causative of "to fall." Since only a limited number of simple active verbs can be extended with a causative effect which produces a new idea, the verb גָדַל (gāḏal, "he was great") is selected, of which the derived causative is הִגְדִּיל(higdîl, "he caused to be great," i.e., "he enlarged"). Of the verb רָאָה (rāʾâ, "he saw") the derived causative is הֶרְאָה (herʾâ, "he caused to see," "he let one see," "he showed"). Of the (weak) verb בָּא (bāʾ, "he came") the derived causative הֵבִיא (hēḇiʾ, "he caused to come") produces the sense "he brought."
vi. causative passive. This is the passive of v and its form is הָגְדַּל (hogdal), meaning "he (or "it") was made great," i.e., was enlarged.
vii. reflexive. Again, for the sake of clarity, another verb is taken as the parent of this derived form. From the simple active נָשָׂא (nāśā', "he lifted up," "he raised up"), the derived reflexive is הִתְנַשֵׂא (hiṯnassēʾ, "he raised himself up") and this, in turn, produces the sense "he boasted."
The terminology devised by the early Jewish grammarians to designate the above seven verbal forms has been universally accepted. They firstly considered the simple active to be קַל (ḳal, "light"), while the other six were said to be כְּבֵדִים(kεḇēdīm, "heavy"), since each one of them received additional letters or syllables. These six heavy forms were subdivided into the following categories. Taking as their basis the verb פָּעַל (pāʿal, "he did"), they designated the simple passive as a נִפְעַל (nip̠ʿal), i.e., a "was done" form. The intensive active was called a פִּעֵל (piʿēl), i.e, a "did intensively" form and its passive a פֻּעַל (puʿal) i.e, "was done intensively" form. The causative was designated a הִפְעיל (hip̠ʿil)i, i.e., a "caused (one) to do" form and its passive a הָפְעַל (hop̠ʿēl), i.e., a "was caused to be done" form. The reflexive became the הִתְפַּעֵל (hiṯpaʿēl), i.e., a "did to oneself" form. Though this system may appear to be clumsy and cumbersome, its adoption as the standard, universal terminology has avoided the multiplicity of terminologies in different languages.
the weak verb
Hebrew has very few examples of irregular verbs, in which a complete conjugation is made up of two different roots. One such example is the verb שָׁתָה (šāṯâ, "he drank"), of which the derived causative form, with the meaning "he caused to drink," "he gave to drink," "he watered, irrigated," is not הִשְׁתָּה (hištâ) but הִשְׁקָה (hišḳâ), from a different root, namely שָׁקָה (šāḳâ). There are, however, many verbs which, because of a peculiar letter in the stem, diverge from the normal or regular and these are known as weak. The reader will now be familiar with the categories of peculiar letters which bring about modifications in the articulated word. Examples of these as they affect the verb are given below:
a. The letter נ (n). The incompleted state of the normal verb שָׁמַר (šāmar, "he kept, watched") is יִשְׁמֹר (yišmōr, "he will keep, watch"). However, of the verb נָטַר (nāṯar) (which has the same meaning) the incompleted state is not יִנְטֹר (yinṯōr), for the medial, vowelless נ (n) standing between two vowelled consonants is assimilated to the next letter, which is thereby doubled, so that the resultant form of this word is יִטֹּר (yittōr) – a divergence from the normal.
b. Guttural letters (ע, ח, ה, א) and ר. It was noted above that the characteristic of the piʿēl (intensive) form is the doubling of the middle root-letter, as שִׁבֵּר (šibbēr, "he smashed"). When this medial letter is a guttural or ר, which cannot be doubled in articulation, the preceding vowel is prolonged. "He glorified" is not פִּאֵר (piʾēr) but פֵּאֵר (pēʾēr), "he refused" is not מִאֵן (miʾēn) but מֵאֵן (mēʾēn), thus producing a deviation from the normal. (Note: This always happens with the weak gutturals א and ע. With the harsh guttural ה and ח, no doubling takes place but, because of their harsh nature, the preceding vowel seems to merge with the letter (נִחַם – niḥam, "he comforted") and so the prolongation of that vowel is arrested.)
c. The weak letters (י, ו, ה, א). Taking again as the standard the verb שָׁמַר (šāmar, "he kept"), one with a medial weak letter ו (w) will deviate from the normal in the following way. "He arose," which one might have expected to be קָוַם (ḳāwam) deteriorates into קָם (ḳām); the weak ו melts into the vowel-sounds in which it is placed. The spelling is adjusted to the modified form. A verbal root with a terminal silent א also induces a deviation from the normal. "He found" is not מָצַא(māṣaʾ), for the terminal א is silent, so that, in actual sound, the syllable is open, i.e., ending in a vowel. Since the open syllable was usually pronounced with a long vowel, the speaker automatically prolonged it and the resultant form became מָצָא (māṣāʾ).
In this instance also the early Jewish grammarians devised a rather cumbersome terminology to denote categories of weak verbs, which has been universally adopted. It was based on the word פֹּעַל (pōʿal) which was their term for "verb." If the first root-letter was weak, they referred to it as the פּ (pe) of the root, the second root-letter as the ע (ʿayin) of the verb and the third root-letter as its ל (lameḏ). For example, the verb נָטַר (nāṯar) was designated as a פּ״ן (pe nun), i.e., initial נ, verb; the verb פֵּאֵר (pēʾēr) was described as an ע״א (ʿayin ʾalep̠), i.e., medial guttural, verb; the verb קָם (ḳām), whose root-letters are קום, was designated as an ע״ו (ʿayin waw), i.e., medial ו, verb; the verb מָצָא (māṣā) became known as a ל״א (lameḏalep̠), i.e., terminal א, verb.
The structure of Hebrew grammar, of which a partial sketch has been given here, has not changed appreciably through the centuries, from biblical times to the present day. It manifests itself even in the highly evolved spoken and written Hebrew of contemporary Israel. The realization that Hebrew grammar reflects natural phenomena in speech and characteristic forms of thought leads to an understanding and appreciation of the genius of the language.
For a different view of Hebrew morphology in general and of the verbal system in particular, see U. Ornan, Ha-Millah ha-Aḥaronah – Mangenon ha-Ẓurah shel ha-Millah ha-Ivrit (2003).
(II) HEBREW GRAMMAR: DETAILED SURVEY
1. definition of topic
2. writing and spelling
Phonology and Morphophonology
4. consonants as pronounced by various communities
7. other consonants
8. gemination and clusters
10. vowel quantity
11. vowels as part of the syllable
12. the phonological status of the vowels
13. the Šewa and ḤaṬefs
14. the accent
15. the diphthongs
16. interchange and elision of vowels
17. interchanges due to sound combinations
The Hebrew language is very old; but even in the oldest portions of the Bible, written more than three thousand years ago, it is a fully formed literary vehicle. No language, however, can remain unchanged over so long a period. Hebrew was subject to change, though for almost half its existence it was preserved only in writing, as a literary language. Nevertheless, an intelligent Hebrew speaker, of the level of an Israeli high school graduate, can read and understand literature written in Hebrew from the very earliest times to the most modern. Furthermore he will find in the Hebrew language of each of the periods many elements actively used in modern speech. Many highly civilized societies from time to time rewrite the literary products of past generations to make them more comprehensible to the modern reader, sometimes wholly translating, at others, merely replacing the obsolete by more modern terms. By contrast, Hebrew literature such as the Bible, Mishnah, Midrashim, medieval poetry, etc., are today taught in schools in their original form, and may be read by the community at large. Hebrew thus appears to have maintained its original uniformity; yet linguistic analysis reveals that this uniformity is in fact limited and in part imaginary. The relative uniformity of Hebrew results primarily from two factors: the graphic representation of the language, and its morphology. Hebrew today is written, as it was in the past, in an alphabet in which all the letters represent consonants. Although several of these letters may, in specific circumstances, also indicate vowels (and in modern times even more so than in the past), this does not alter the general character of Hebrew writing, in which only a part of the total sounds are fully expressed: the part represented consists of those sounds which are naturally more stable and not given to radical change, while those sounds given more naturally to change are not represented by the writing. The employment of vowel points to express these sounds is very limited, and even nowadays does not determine the form of the written Hebrew word. This retention of the relatively constant element, and elimination of the element given to change, are great aids in understanding the written Hebrew word regardless of its date of origin. However, the factor primarily responsible for the uniformity of the language is Hebrew morphology. In the accepted vocalization of the Bible, the morphology of biblical Hebrew was fashioned in accordance with one tradition, the Tiberian. This tradition was not only accepted as the norm, but in the eyes of many even became sanctified. The morphology of biblical Hebrew in the Tiberian tradition determines the nature of the grammar of written Hebrew in general, and even new elements, whether taken from other periods or traditions of the language, or originally foreign elements, are adjusted to comply with the regulations of Hebrew morphology. In contrast to the relative stability of morphology and writing, the meaning and the function of grammatical forms and vocabulary and pronunciation have been variable. The particular pronunciation reflected by the accepted vocalization tradition of the Bible (even now mandatory for a vocalized word), is still heard today, but not all of its details are together represented in any one tradition. Closest to the pronunciation of the original vocalizers is that of the Yemenite and Babylonian communities (cf. below, Table 3: Consonants as Pronounced by Various Communities). The grammatical description given below refers to the language as it is written in modern times; a description of the various periods of the Hebrew language is given in the entry *Hebrew Language; and the distinctive traits of the various living traditions are discussed in the article *Pronunciations of Hebrew.
Hebrew is written in two sets of symbols, letters and vowel points; the first is, of course, mandatory, while the latter is reserved, primarily for the areas of education, poetry, prayer books and, to a limited extent, publications intended to reach a very wide audience. The basic alphabet consists of 22 letters; to these one must add five (ץ, ף, ן, ם, ך) used in final word position only, another seven (שׂ/שׁ,ת,פ,כ,ד,ג,ב), whose individual pronunciation is reflected
|* Generally this Hebrew letter is not represented and the sign is used only in special circumstances.|
|כ ךְ||ḵ (kh)||ḵ||kh|
|פ ף||f ; p̄||f||f|
|ק||ḳ ; q||q||q|
|1) Some of the vowel signs are accompanied by letters (matres lectionis) which are not part of the system of vocalization e.g., ־ָא, ־ֵי, ־ִי.|
|2) In A the length (or absence of it) in the vowels is represented in accordance with linguistic considerations.|
|3) The term šewa indicates both a reduced vowel and the absence of any vowel after the consonant; in the latter case it is not transliterated.|
|4) The dageš is also a double valued sign. In the letters בג˝ד כפ˝ת it indicates a different phonetic character and not necessarily gemination. Whether these letters, when pointed with dageš, also indicate geminated consonants depends on syllable structure; thus the correct transliteration of בּ – by b or by bb – for example demands a knowledge of Hebrew grammar. The dageš is placed in the letter ה at the end of a word when the letter is pronounced consonantal h, and does not indicate gemination. In accordance with its function the dageš is variously called forte (for gemination), lene, and mappiq. Similarly a dot above ש indicates a pronunciation difference between š and ś.|
|segol||ֶ||ae (ä, ḝ||e||e|
|Qameṣ (long)||ָ||ạ̄ ā||a||a|
|Qameṣ (short)||ָ||ạ ; o (ọ, ɔ||o||o|
|Dageš as מּ4||mm||mm||mm|
only in vocalized writing, and three (צ׳, ז׳, ג׳) used in words of foreign origin only. All but four of the letters represent consonants only, while ו, ה, א, and י are used at times to indicate the presence of specific vowels. These letters, when not representing consonants, are called matres lectionis. Of the 13 vocalization signs, 12 are intended to represent vowels, and one represents changes in consonants. The various systems used to transliterate Hebrew into Roman script generally reflect two different approaches. The first transliterates the Hebrew alphabet into graphemes of the particular language in accordance with the spelling conventions of that language. This gives rise to a plethora of conventions according to the languages concerned. The other system attempts to transliterate Hebrew so that the letters and vowels will be perfectly or nearly perfectly represented. This system demands the addition of diacritic signs to the Roman alphabet. Table 1: Hebrew Letters and Punctuation is a synoptic table of three methods of transliteration.
The "a" system is used largely in linguistics or when an early Hebrew text is being transliterated. "b" and "c" were established by the Academy for Hebrew Language in 1957 in order to represent living Hebrew. "b" is used in catalogs, title pages, and maps. The sole difference between "b" and "a" is in the letters ב and צ where "B" represents the official modern pronunciation; this is also true with regard to the long qameṣ which is not differentiated from the pathaḥ. "c," on the other hand, intended for popular use, as in road and street signs, eliminates most of the diacritic signs, and simplifies the vocalization. The appearance of the unvocalized Hebrew word is likely to be different from the vocalized word even in its letters. This is due to the tendency to add matres lectionis in non-vocalized writing, in positions where they would be unacceptable in vocalized writing. For this reason it is also called "full spelling."
Table 2: Point and Way of Articulation summarizes the inventory of consonants in modern Hebrew, but they are not all at once present in any one of the varied pronunciation systems. The consonants are listed in accord with their general phonetic value, though some have more than one realization (ר, ל, כּ). This variety is of a wider scale in Hebrew than in most other languages because the speakers of Hebrew are of diverse language backgrounds, and this background is apparent even in the generation whose mother tongue is Hebrew and who are, themselves, monolingual. The range of differences within each consonant is not represented in the table. Those consonants in squares are part of the Tiberian pronunciation (see *Masorah), and are therefore represented in modern writing; they are, however, normally not differentiated in the general or official pronunciation. They can still be heard in liturgy and worship (on שׂ see below) in some communities, and even in the speech of the older generation. The consonants in circles are not pronounced by a large part of the general populace and have been assimilated to תּ, כּ, בֿ. Their independence is maintained in the language in so far as their influence is felt in inflection and declension of words. Those consonants in dotted squares can be heard at times in originally Hebrew words in special situations, or in careless speech, as ḥežbon (חשבון), but are not independent except in words of foreign origin, as in žargòn (jargon), žurnal (journal); they do not determine the character of the Hebrew language in any area of grammar. Therefore, they will be referred to as "foreign," as distinct from "inherited."
If the foreign consonants and those which appear in squares are subtracted, it will be found that the total of inherited consonants in official modern Hebrew is 25, and the incidence of three, circled in the table, is very limited. In other words, the total varies between 22 and 25. This state of affairs changes in regard to the inventory of the consonantal phonemes. The phoneme, by definition, is that unit which only in distinction to another in the same position, performs the function of distinguishing between two words. The phoneme can be of more than one sound (ideally this is always the case), and all these sounds are related to each other as allophones (= variants). Not everyone is agreed as to the number of phonemes in the language, a problem which exists not only in Hebrew. It depends on how we evaluate sounds in foreign words, personal names, or rare words. For example, פ, פּ hardly ever occur in identical surroundings in Hebrew; where the one is
|Point of Articulation||Way of Articulation|
|alveolo-palatal (tongue blade and teeth ridge)||1|
|velarized alveolar (tongue blade and back, and teeth ridge)||2|
|velar||k כּ||g גּ||k(χ) כֿ||4|
found the other is not (as וּפָנִים/פָּנִים and not וּפָּנִים/פָּנִים); they are, therefore, allophones of the same phoneme. When, on certain rare occasions each sound may be heard in the same environment, there will be no difference in the meanings of the words. Generally, one of the forms of the word is considered incorrect in proper speech, as קִצְבָֿה opposed to קִצְבָּה. However, if foreign words are considered, the difference between p and f proves to be distinctive as in the words פּונקציה (punkcya = puncture) opposed to פונקציה (funkcya = function). To the extent to which such words enter the language of less educated persons there is more likely to be created a split between the p (פּ) and f (פ), and each is liable to appear in the same conditions. With regard to the written language, these rare phenomena are to be ignored, and the count will include 22 consonantal phonemes of the 25 consonants mentioned above, since כֿ בֿ פֿ are allophones of כּ בּ פּ and together are three phonemes. The consonantal phonemes can be divided into nine groups in accordance with their point of articulation:
|bilabials||ו, מ, ב, פ||4|
|dental||זסצ רלנ דתּ||8|
|glottal||א,ה||2 (= 22)|
A pronunciation different from the one described above would show some slight differences in the localization of the phonemes. Ignoring the "foreign" consonants, the difference between the consonants in Hebrew and those (estimated) in proto-Semitic is in the number of fricative and affricative sounds only. Some of these sounds originate in Hebrew and did not exist in proto-Semitic (labio and dental fricatives), some palatals existed but were phonemes in proto-Semitic, while they are allophones in Hebrew (as כ). Similarly, all those proto-Semitic consonants which have disappeared from Hebrew are fricatives and were assimilated to other Hebrew fricatives, while in Aramaic, for example, the proto-Semitic fricatives were assimilated to the plosives (דֿכּר in proto-Semitic became זכר in Hebrew and דכר in Aramaic). There are three types of articulation which determine phonemic contrasts in the consonants: voice, emphasis, and nasality. The voiced/unvoiced distinction has five pairs: ע, ז, ג, ד, ב opposed to פ, ת, כּ, ס, ח; the emphatic/non-emphatic pairs are ק, ט op-posed to כּ, תּ; and nasality affects six consonants: פּ, בּ (together an archiphoneme) opposed to מ; and נ (dental and not alveolar), תּ opposed to ד. Voice and nasality are phonetic qualities, whereas emphasis does not denote a clearly defined common phonetic quality of the relevant phonemes. The most that can be said is that these sounds are produced partially with the back portion of the mouth. The plosive/fricative quality which in our arrangement does not determine the phonematicity of פ, כ, ב,, is the determining difference between א and ה. There are those who are of the opinion that א is a voiced consonant, but this results from a desire to produce a symmetrical phonological scheme and has no foundation in actual articulation. Only if it is assumed that the ה is voiced (as in Arabic) is the difference between א and ה not one of plosive/fricative, but one of voice. The determining features do not necessarily have to rule all consonantal oppositions; for example, voice is a determining quality in ten (or at the most twelve) sounds only, and there are voiced consonants (ר, נ, ל) which do not have unvoiced pairs, and unvoiced consonants (שׁ, צ, ק, ט) which do not have voiced pairs.
The official, model, most careful pronunciation of Hebrew, used, for example, by radio announcers, especially in reading selections from the Bible (and sometimes called "Semitic," or "Eastern"), is the result of a mixture of different systems of pronunciation used for generations in the various communities. This mixture is not precisely the result hoped for by those who were instrumental in the rejuvenation of Hebrew speech; generally, it can be said that, of those sounds which were distinctive to only one community, more were omitted from the official and general pronunciations than were accepted. The consonants of the official pronunciation are reviewed in the section entitled Consonants above according to their general phonetic value; therefore, it will be instructive to review synoptically the consonants as they were articulated in the past in the different Jewish communities (still occasionally used in prayer and liturgy), from which the sounds of modern Hebrew were extracted. Such a review will indicate what shades of pronunciation are likely to be heard, primarily in the speech of the older generation in official Hebrew.
Table 3: Hebrew Consonants indicates that some consonants were retained in all pronunciations while others were lost or assimilated to other similar sounds. If we arrange them next to the 29 consonants of the vocalizers of the Bible, the following picture emerges: 11 stable consonants (תּ, פֿ, פּ, נ,מ, ל, כֿ, כּ, י, דּ, בּ) are preserved in all traditions (nine, if we omit כֿ and פּ which developed in a unique manner in the Samaritan tradition); eight consonants שׂ, שׁ, ר, ס, ז, ה, גּ, א are retained by most of the pronunciations and changed in a few; and ten consonants תֿ, ק, צ, ע, ט, ח, ו, דֿ, גֿ, בֿ are changed in most of the pronunciations. It emerges that differences in pronunciation are mainly in two groups: five consonants ק, צ, ע, ט, ח which are "semitic" (gutturals and emphatics) and five others תֿ, ו, דֿ, גֿ, בֿ; of which at least four תֿ, דֿ, גֿ, בֿ; tend to lose their fricativeness.
As already pointed out, in the official language only פ, כ, ב are pronounced in two ways – hard and soft (spiranted) – but in vocalized spelling ת, ד, ג are similar to them in every way. The distribution of hard and soft allophones is not given to simple phonologic definition, and there are exceptions in both directions. However, since the distribution of the soft allophone is greater than the hard one, finding it in positions which contradict the rules is not a radical deviation. On the other hand, a hard allophone where it would be expected to find the soft constitutes a clear exception and, in proper modern speech, is even less common than in biblical Hebew. The following are the rules for their distribution:
a) The hard consonant always appears when (1) there is gemination, as סִפֵּר, סֻכָּה, סִכָּה, שַׁבָּת and (2) even when ungeminated, if it is not preceded directly by a vowel, as:מִסְפָּר, פֶּה, מִכְתָּב, כֶּסֶף ;הַשְׁבָּתָה, בַּיִת.
b) The spiranted consonant appears (1) when ungeminated and preceded directly by a vowel, as: סוֹפֵר, סִיכָה, שָׁבַת; (2) when it is the second part of a cluster in one syllable, either at the beginning or the middle but not at the end of the word, as עִזְ־זְבוֹנוֹת =) עִזְּבוֹנוֹת ;מַרְ־כְּבוֹתָיו, רָ־כְבָה, צָ־רְפָת ;כְּפָר, כְּבִישׁ, דְּבַשׁ)). Opposed to these are: יִשְׁבְּ, יֵבְךְּ (rare forms likely to appear in poetry) שָׁמַרְתְּ, נֵרְדְּ and (3) when it is the second part of a cluster which divides into two syllables (opposed to rule a 2) in the following morphological types:
(i) Nouns: (a) In the plural forms of nouns of the type פְּעָלוֹת, פְּעָלִים as: מַלְ־כוֹת ;אָסְ־פֵיהֶם, דַּרְ־כֵיהֶם, מַלְ־כֵי (construct) נְדְ־בוֹתֵיכֶם, כַּנְ־פוֹתֵיהֶם, בִּרְ־כוֹתָיו. (b) In declension of פְּעָלָה forms even in the singular, as: עַרְ־בַת, נִדְ־בָתוֹ. (c) In the declension of פְּעֵלָה, פָּעֵל forms in singular and plural even in those words where the ṣere is not retained, as: חֶשִׁ־כַת, חַנְ־פֵי. (d) In פַּעְלָן and derived forms, as צַרְ־כָנוּת, עַגְ־בָנִיָּה, חַנְ־פָן; the same is true for פַּעַלְתָן forms, where the spirant is expressed in writing only, as רַעַבְתָן, גַּאַוְתָן. (e) In פַּעִלוּת forms, such as עַצְ־בוּת, מַלְ־כוּת. (ii) Verbs: (f) In the conjugation of the infinitive qal, as: בְּמָלְ־כוֹ, כָּתְ־בִי. (g) In the conjugation of the imperative qal, as כָּתְ־בָה, כִּתְ־בוּ, כִּתְ־בִי.
(h) In all forms of verbs and nouns when the syllable preceding בגדכפ״ת is closed in declension, as: נֶאֶסְ־פוּ – נֶאֱסַף, יַעַרְ־בוּ – יֶעֱרַב, מַעַרְ־כוֹת – מַעֲרָכָה, מַאַרְ־בֵי – מַאֲרָב. (i) In all forms of words after a closed syllable to which is prefixed one of the servile letters בוכ״ל with the exception of the qal infinitive with ל, as: וּרְ־כַב, וּרְ־כַבְתֶּם, וּלְ־בִנְךָ, בִּנְ־פֹל, בִּלְ־בָבוֹ, לִזְ־בוּב, לִלְ־בָבוֹ, כִּסְ־פֹר, כִּזְ־בוּב, וּנְ־פֹל, but לִשְׁ־בֹּת, לִסְ־פֹּר, לִשְׁ־כַּב. Two exceptions worthy of note due to their frequency in the language are: שְׁתַּיִם, and the second person fem. perfect ending תְּ – as in שָׁמַעַתְּ, לָקַחַתְּ, in which a hard תּ is retained contrary to the above rules.
The distribution of the plosive and spiranted allophones of בגדכפ״ת is quite complicated, but can be ordered in accordance with the above rules with regard to the official language and the speech of intellectuals. In other social strata, and in the speech of children, these rules are not maintained, however; at times the plosive allophone dominates and at times the spirant (for example: יִתְפֿס by analogy to the perfect כַבֵּס ;תָּפֿס from the imperfect יְכַבֵּס). A further weakening of these rules is due to the foreign words in Hebrew which contain a p or f which is not in accordance with the aforementioned rules (cf. section on Consonants above).
These consonants are similar in that (1) they cannot be geminated; (2) they do not usually close a syllable;
|Persia||Western Persia||"Little" Persia||Bukhara||Dagestan (Mountains)||Babylonian||Yemenite||Moroccan||Sephardi||Portuguese||South||Central||North-west||North-east||Lithuania||Poland||Germany||SAMARITAN|
|Notes: (1) Research into the various traditional pronunciations is still far from able to present all the information clearly tabulated; in fact it is only just beginning. Therefore, the table contains only those sounds described in the scientific literature, and it must be noted that the quality of these descriptions is not uniform, nor do they all attain the same level of precision. Still the table reflects the situation generally. It should be noted that the term "Sephardi" does not represent a homogenous community but one spread over Europe and Asia. This accounts for the high number of variants. (2) The horizontal line next to 'alef, heh, ḥet, 'ayin, indicates that this consonant is not pronounced at all; when this line appears with a consonant it indicates that in certain environments it is not articulated, or that part of the community does not pronounce this consonant. (3) A consonant in parentheses is a less common variant or is limited to certain instances (as f, bb in the Samaritan tradition). (4) The consonants are listed in accordance with the rules of the International Phonetic Association. (5) For further details of pronunciation of the various communities see *Pronunciations of Hebrew.|
|א||ʔ, –||ʔ, –||ʔ, –||ʔ, –||ʔ, –||ʔ||ʔ||ʔ||ʔ||ʔ||–||–||–||–||ʔ, –||ʔ, (h), –||ʔ, –||ʔ, –|
|בֿ||v, (β), w||v, (β), w||v, (β), w||v, (β), w||v, (β), w||(β), b||v, (β), b||w, b||b, v||b||v||v, (b)||v||v, (w)||v||v||v||b, (f)|
|דֿ||d||d||d||d||d||d, (δ)||δ||d||d, δ||d||d||d||d||d||d||d||d, t||d|
|ה||h||h||h||h||h||h||h||h||h, ʔ||h||–||–||–||–||h||h||h||ʔ, –|
|ו||v, (β), w||v, (β), w||v, (β), w||v, (β), w||v, (β), w||w||w||w||w, v||w||v||v, (w)||v, w||v, (w)||v||v||v||w, b|
|ח||h||ḥ||h||ḥ, (h)||ḥ||ḥ||ḥ||ḥ||ḥ, χ||χ, (ʁ)||χ,||χ,||χ,||χ,||χ,||χ,||χ,||ʔ, ʕ, –|
|ס||s||s||s||s||s||s||s||ś||s||s||s||s, (z)||s||s||s, ś, (š)||s||s||s|
|ע||ʔ, –||ʕ||ʔ||ʔ, ʕ||ʕ||ʕ||ʕ||ʕ||ʔ, ʕ||ʔ, ŋ||ŋ||ŋ||ŋ||ŋ||ʔ, ŋ y||ʔ, y||ʔ||ʔ, ʕ, –|
|צ||s||s||s||s||s||ṣ||ṣ||ṣ||ṣ, c||c||c||c||c, (s)||c, (s)||c||c||c||ṣ|
|ק||G, (k), ʁ||G, (k)||G||G, (k), ʁ||G, ʁ||q||q||q||q, k||k||k||k||k||k||k||k||k||q, (ʔ)|
|ר||r||r||r||r||r||r||r||r||r||r||r||r||r||r||r, γ||r, γ||r, γ||r|
|שׁ||š||š||š||š||š||š||š||ś||š||š||s (š)||s, (ž)(š)||s, (š)||s, (š)||s, š, ś||š, ś||š||š|
|שׂ||s||s||s||s||s||s||s||ś||s||s||s||s||s||s||s, š, (ś)||s||s||š|
|תֿ||t||t||t||t||t||θ||θ||c, t||t, θ||th||d||d||d||d||s, š, (ś)||s||s||t|
(3) they generally influence the vowel which precedes them either in quality, or by creating a vowel similar to that with which the syllable is or should have been, closed; (4) they are vocalized with ḥaṭefs when they are not in syllable final position as: (1 and 3) (בֵּאֵר) instead of ‡ מְטֹהָר ;בִּאְאֵר instead of ‡ מְטֻהְהָר (2, 3, and 4); נֶאְדָּר instead of ‡ נֶחְמָד ;נִאְדָּר instead of ‡ מַעְבָּרָה ;נִחְמָד opposed to רוּחַ ;מִשְׁטָרָה instead of ‡ רֵיחַ ;רוּחְ instead of גָּבֿוֹהַּ ;רֵיחְ instead of ‡ נִצַּח ;גָּבֿוֹהְּ opposed to מַעֲלִית ;נִסֵּךְ opposed to פָּעֳלוֹ ;מַרְבִּית opposed to קָדְ־שׁוֹ. As a result of these qualities, they are classified under the heading of 'gutturals,' a name which has been accepted even though it does not accurately describe all of them from the point of view of their articulation.א.
In modern speech, this consonant never closes a syllable. Still, there are some who, in certain words such as יַאְדִּים, נֶאְדָּר, מַאְפֵּלְיָה, are careful to close the syllable with the א because they are so vocalized in the Bible, a fact which is also exploited in writing poetry. On the other hand, (1) א always disappears at the end of a word, as וַיַרְא, שָׁוְא, חֵטְא, נָשׁוּא, מָבוֹא, מָלֵא, צֵא, מָצָא (compare נֵרְדְּ, וַיַשְׁקְ); (2) within the word it sometimes disappears, and at other times receives a ḥaṭef vowel. This situation is not the result of precise phonological conditioning, but differs in different morphological situations, as תֹּאמַר, תָּבוֹאנָה, לֵאלוֹהִים, מָצָאתִי; but תָּאֲךָ, נְשׂוּאֲכֶם, בּוֹאֲכֶם, מְצָאֲךָ, and not בּוֹאכֶם, תָּאךָ, מְצָאךָ, etc.; (3) unlike the biblical norm the א rarely disappears between a šewa and a vowel, but in several forms this is always the case: רָאשִׁים (from ‡ מָאתַים, (רְאָשִׁים opposed to מְאַת. (When not pronounced it sometimes does not appear in the written word, and this is the rule in the imperfect of פ״א verbs in first person singular, as: אֹמַר from כְּלוֹמַר, לוֹמַר, אֹאמַר (from the root אמר). In the Bible this elision is more common.ה.
(1) Feature number (3) of א is the rule with respect to ה: (a) when used as the definite article and coming after ל, כ, ב. For example: לַדָּבָר, בַּדָּבָר, בַּדָּבָר (but וְהַדָּבָר!);
(b) In the imperfect and participial forms of hifʿil and hufʿal, as מוּשָׁב, נוֹדִיעַ, מַפְקִיד, יַפְקִיד from ‡מְהוּשָׁב‡, נְהוֹדִיעַ‡, מְהַפְקִיד‡, יְהַפְקִיד but in the infinitive the ה is retained: לְהַפְקִיד, כְּהוֹדִיעוֹ, etc. It is also retained in a number of personal names, as: יְהוֹרָם, יְהוֹנָתָן alongside יוֹרָם, יוֹנָתָן and יהודה alongside יוּדה.
Some generations ago there was a tendency to retain the definite article and forms like להדבר, כהדבר were common. Today this is maintained in the words לְהַבָּא כְּהַיּוֹם, but the form כַּיּוֹם is also used.
(2) The ה tends to be assimilated to the preceding consonant when it is part of a pronomial suffix שְׁמָרַתּוּ ;–הָ, –הוּ (rare form יִשְׁמְרֶנּוּ, (שְׁמָרַתְהוּ (rare form יִשְׁמְרֶנָּה, שְׁמָרַתָּה, (יִשְׁמְרֶנְהוּ (rare form יִשְׁמְרֶנְהָ).
(3) It tends to be elided in the pronomial suffix הוּ after י–ִ as שְׁמַרְתִּיו, אָבִיו, פִּיו (the rare forms שְׁמַרְתִּיהוּ, אָבִיהוּ, פִּיהוּ, etc., are found mainly in poetry) and almost always is elided after שְׁמָרוֹ :־ָ (from שְׁמָרָהוּ).ח.
Even in pronunciations in which there is no difference between this consonant and the spiranted כ, it maintains its independence in that it acts as a guttural and not as one of the בגדכפ״ת. It does not become a כּ and prefers those vowels preferred by gutturals, as: נִצַּח opposed to נִסֵּךְ (only uneducated speakers, if they pronounce the כ as a ח are likely to equate them, for example saying סוֹמַכַת instead of סוֹמֶכֶת and vice versa).ע.
Even in the pronunciation which identifies this sound with the א, the ע is kept separate in different phonetic contexts. In this way it is similar to that pronunciation which maintains the ע as an independent sound. For example, it tends to be pronounced with the furtive pathaḥ (נוֹסֵעַ opposed to נוֹשֵׂא) and often demands a vowel different from the vowel used for א as: יַעֲבֹר opposed to יֶאֱסֹף, and לַעֲבֹד opposed to לֶאֱבֹד. This independence is further realized in various morphological situations.
נ followed directly by another consonant is usually assimilated to it; only the נ which is part of the root is not usually assimilated to אהח״ע; in some roots and other forms this is also the case (perfect נָפַל, imperfect יִפֹּל; perfect נָהַג, imperfect יְנְהַג; perfect נִדּוֹן, imperfect יִדּוֹן; perfect נֵעוֹר, imperfect יֵעוֹר; infinitive לִפֹּל but לִנְגֹעַ, etc.).
צ pronounced as ẓ (= c) is a compound sound and can be heard in speech when ס and ת are contiguous, as, for example, in בית־ספר (a common spelling mistake among children is ביצפר!). It still acts in Hebrew as one sound (monophonematic), and it is impossible to demonstrate oppositions in meaning which depend on the opposition (ẓ/ts); the fact that ẓ can be only one of the elements of a root is itself proof of this assumption.שצסז.
The clusters דז, טצ, תש, תס where the ד) ת or ט) is the t of the hitpaʿel, are impossible and the order is reversed to זד, צט, שת, סת. Also rare are the clusters תס etc., when the ת (or ד or ט) are elements of the root. In fact, in the words or forms derived from these roots a vowel usually appears between them – תסס, תשש. However, since the Middle Ages these combinations have appeared in a few words where the t is part of the root, as מַתְסִיס, הִתְשִׂיר.
It is possible to prove that the שׂ was an independent sound in biblical times and was so considered by the vocalizers of the Bible. Still, here and there, there are examples of the merging of this consonant with the ס in the Bible as in כעש alongside כעס, and in rabbinic Hebrew many of the words and roots with שׂ in the Bible appear with ס, as: סִיחָה instead of: שִׂיחָה etc. Hundreds of years have passed since שׂ ceased to exist as an independent sound and became ס in all Jewish pronunciations and שׁ in the Samaritan pronunciation. Only the spelling recognizes the differences between שׂ and ס, and in certain instances this is an aid in differentiating homonyms, as: שָׂרַר = rule / סָרַר = rebel, שָׂכַר = hired / סָכַר = closed, סְמִיכָה= diploma / שְׂמִיכָה = a blanket. The medieval paytanim used to put a word beginning with שׂ between נ and ע in their acrostics even though the biblical poets put it together with שׂ (Psalms 118); similarly there were lexicographers in the Middle Ages who placed words beginning with שׂ with those beginning with ס that is, according to their pronunciation rather than their graphic form.ו״י.
These two sounds (if the ו is pronounced bilabially) are to be classified as being between a consonant and a vowel, also in their function in the language. Unlike other consonants, they interchange with vowels in specific instances and they are geminated like consonants. The ו at the beginning of a word or as the first element in a root was rare and was generally replaced by י; thus even in modern Hebrew there are very few words which begin with ו. [In the pronunciations of most of the communities which contain the spiranted ב, there is a tendency to pronounce the ב and ו identically, the ב at times being pronounced as a ו (bilabially) and vice versa. But even in this case only the ו retains the above relationship to the vowels, never geminated into a בּ. Only in the Samaritan pronunciation did the ו become a בּ in most positions, after the soft ב was entirely lost.]
Besides the gemination of a consonant caused, in Hebrew as in other languages, by the occasional immediate sequence of that consonant namely when a root consonant comes into contact with an affixed formative (for example הִתָּמֵּם < הִתְּ־תַּמֵּם, כָּרַתָּ < כָּרַתְּתָּ, נָתַנּוּ < נָתַנְ־נוּ), gemination is very common in Hebrew and serves to create nominal and verbal forms. Some call a geminated consonant in Hebrew a "long consonant" parallel to a "long vowel," but this comparison is justified neither by the phonetic process which takes place (the syllable boundary is within the consonant; a difference is felt, for example, between lengthened ז in תִּזְכְּרוּ and the regular geminated pronunciation of the ז in תִּזָּכְרוּ), nor by the function in the language which it fulfills.
(1) Every form which has a geminated consonant, can have that consonant replaced by a cluster of two consonants but never by one alone, and thus פִּלְפֵּל, פִּרְנֵס, פִּקֵּד are all of the same verbal type, and דַּיָּן and פַּרְנָס are considered to be of the same nominal type; (2) A cluster of two consonants at times becomes a geminated consonant, and the two forms may even exist side by side (לִפֹּל, לִנְפֹּל). (3) A geminated consonant may split into a cluster of two consonants (common in Aramaic, rare in Hebrew), and in some of the words which have a two consonant cluster in Hebrew today there was originally a geminated consonant, as: גַּלְמוּד (from שַׁרְבִיט, (גַּמּוּד (from ‡שַׁבִּיט). (4) The limitations regarding the vowel before the last consonant in the syllable, are also in force for the vowels which precede a geminated consonant. Therefore, a geminated consonant functions just like a cluster of different consonants, being more limited only in that it cannot come at the end of the word (compare צֵלְלְ >) צֵל) opposed to כָּרַתְּתְּ >) כָּרַתְּ, נֵפְטְ, opposed to שָׁמַרְתְּ)). At the beginning of a word a geminated consonant can be found only when the first consonant is not part of the root but a formative element, as: תְּתַרְגֵּם, מְמֻכָּן (see below); or when the word comes after the interrogative מַה־זֶּה) מַה), thus constituting a phonetic unit (in the Bible this is common in various combinations of words and vowels). In addition to geminated consonants, there are times in modern Hebrew, especially in the speech of young girls, when, as an expression of emotion, a consonant is lengthened, as yof-f-f-fi (יוֹפִי), but this has no grammatical function.
There is a tendency to eliminate gemination, especially when the geminated consonant is vocalized with a šewa; and thus the gemination is usually eliminated in י with šewa and in the מ of the participle after the definite article, as הַמְסִבּוֹת, הַיְקָבִים (= they are causing; but הַמְּסִבּוֹת (= parties)). However, this tendency is overruled by morphological considerations (הַמְּסַפְּרים and not הַמְּסַפְּרִים; but in the Bible הַמְבַקְשִׁים). The tendency to eliminate becomes the rule when the consonant involved is either אהחע״ר or the spiranted allophones of בּגּדּכּפּ״תּ which are never geminated. In other words, five phonemes cannot be geminated. (In the Samaritan pronunciation the ר is geminated just as any other consonant, and this was also the case in the Hebrew of Septuagint times.) Many Hebrew speakers today do not commonly geminate consonants in their speech, but they maintain those conditions which derive from gemination and determine the form of the word, as the hard פּ in סַפָּר (a barber) and the spirant in סָפַר (counted) etc. According to the rule, a cluster of more than two consonants is impossible in Hebrew. A cluster at the beginning of a syllable is realized generally by a šewa inserted between them. Clusters of more than two consonants are found only in international words used in Hebrew, for example סטראטגיה (alongside אס־טראטגיה). In the middle of a word a cluster of three consonants is conceivable. This is especially true when the first element in the cluster is a geminated consonant as, שִׁמְמְרוּ ‡>) שְׁמְּרוּ, (שַׁבְּבְּלוּל ‡>) שַׁבְּלוּל), which is usually articulated with the addition of a šewa between the geminated consonant and the one following it. In "inherited" words a cluster at the end of a word is possible if it ends in a plosive (for example קֹשְׁטְ, נֵרְדְּ, יֵבְךְּ, but תֵּפֶן). In international words this rule does not always apply (פִילְםְ, סוֹצְיאַלִזְםְ), but modern pronunciation tends to insert a šewa in such cases.
There is no essential difference between the two types of sounds – vowel and consonant – the difference being rather one of degree; in fact, the semi-vowels ו and י are proof that it is possible to pass from one category to the other. The consonants are classified above according to three criteria: point of articulation, method of articulation (open or closed), and the action of the vocal chords (voiced or unvoiced). Since the vowels are all voiced and articulated in the open position, we are left only with the criterion of point of articulation, that is, the relative closeness of the organs of articulation to each other. In articulating the vowels, actual contact is not conceivable; the basis for classification is, therefore, the movement of the tongue and the working of the lips (the traditional Hebrew names for the vowels, pathaḥ, qameṣ, etc., are an attempt to describe the opening of the mouth and the activity of the lips). Following the direction in which the tongue is raised during the articulation of the vowels, we may divide them into "front," "central," and "back." Generally, a rounding of the lips is an accompanying feature of back vowels, and spreading the lips an accompanying feature of front vowels, but rounding and spreading of the lips are possible in all the types of vowels. The section on consonants above mentioned the great variety of consonants heard from the Hebrew speaker and the reasons for them. The vowels, by their very nature, are even more variegated. For the consonants it is at least possible to establish ideal standards for official speech, but there are no set standards applicable to the vowels. There have been no studies to date which can supply information as to the exact phonetic makeup of the vowels in modern Hebrew, but this deficiency does not prevent a phonemic description, since the variations in the vowels are not distinctive and do not affect the meanings of words. Therefore, it will be sufficient to classify the vowels generally into: front: i, é, e, represented in writing by ־ִ ḥireq, ־ֵ ṣere, ־ֶ segol, and ־ֱ ḥatef segol; central; a, represented by ־ַ pathaḥ, ־ָ qameṣ (gadol), and ־ֲ ḥatef pathaḥ, and ě, represented by the šewa; back: o represented by ־ֹ ḥolem, ־ָ qameṣ qaṭan, and ־ֳ ḥatef qameṣ, and u represented by וּ šureq and ־ֻ qibbuṣ. By using average measurements the relationship among the vowels can be graphically described in the following way:
Since only the general values of the vowels have been mentioned, it emerges that the number of vowels is close to the number of vowel phonemes, which are all the above, save the two circled, i.e., the segol and the šewa. Regarding the phonemic value of a sound there are likely to be differences of opinion (see section 3, Consonants, above), and the doubts which might be raised about this classification will, therefore, be discussed in sections 12 and 13, on The Phonological Status of the Vowels and The Šewa and Ḥaṭefs, below. Many Hebrew speakers do not differentiate between e and é, and even those who do differentiate do not always apply the ṣere and the segol respectively where demanded by the rules of vocalization. Even among those who do differentiate, there are some who articulate the sere almost as a dip-thong ey. Note that ḥaṭefs differ phonetically from the šewa, and do not represent independent qualities but are identical to full vowels.
In the common pronunciation the vowels are not differentiated as to length, only the šewa being of shorter duration than the other vowels. Experimental methods have proved that the vowel in an accented syllable is slightly longer than the vowel in an unaccented syllable, but this difference is not discernible by the ear, since no semantic difference depends on a vowel length. (In English, by contrast, this is a distinguishing feature, cf. [it] (it), opposed to [i:t] (eat).) In fact it may be said that the vowels in Hebrew are isochrons (of equal length). However, since there is long standing tradition in Hebrew of dividing the vowels into "long" (lit. big) and "short" (lit. small), an aspect which is also relevant to the rules of punctuation, this division will now be considered. The differentiation originated in an attempt to divide the vowels according to length (compare O-mega and O-mikron in Greek), that is: long and short. The Spanish grammarians of the Middle Ages felt that the qameṣ (pronounced "a"), the ṣere, ḥolem, šureq, and full ḥireq (with yod) were long vowels, while the pathaḥ, segol, qameṣ (pronounced "o"), qibbuṣ and ḥireq (without yod) were short vowels. They considered that there were five qualities of vowels (a, i, e, o, u) and that these were either long or short. This division is a reflection of the "Sephardi" pronunciation, but it must be recognized that in that pronunciation the accent also caused a lengthening of the vowel, and so the "short" vowels in accented syllables were long (for example פֶּה is pronounced with a long vowel even though it is a segol; the same is true for the pathaḥ of שָׁמַרְתִּי). The opposite is true for "long" vowels which are next to the accented syllable and are pronounced short (an exact description was given by R. Joseph Kimḥi in the 12th century). In the 19th century, which was interested in comparative historical study, this principle of division was accepted by Hebrew linguistics with one change: the symbol representing a vowel (excluding the šewa and ḥaṭefs) represents only quality, and that quality can be either long or short. The length is not determined by the symbol or by its place in the word (contrary to Kimḥi, above), but rather in accordance with comparative grammar. As a result of these considerations long vowels are those which generally remained unchanged in the declension of the word, while those which change are either short or lengthened in special phonetic conditions. Thus, for example, the ṣere of מֵת (dead) is always long while that of שֵׁן (tooth) is short (except for biblical pausal forms); similarly, the ḥolem of חוֹל is always long, while that in חֹל is short (except in pausal forms). This differentiation between originally long vowels and secondarily lengthened vowels gave rise to a threefold distinction: "long" vowels (that is originally long), "middle" (that is lengthened), and "short." However, the term "middle" never achieved wide acceptance. This division, unlike that proposed by Kimḥi, is not rooted in any real tradition of Hebrew pronunciation, but is entirely based on theoretical considerations, which assume a Hebrew pronunciation among the Masoretes, when they determined the vocalization of the Bible. Early evidence, such as Greek transliterations and well-based considerations, tends to justify the assumption that in early Hebrew there was a difference in the length of vowels, and that the behavior of the vowels as it appears in the vocalization of the Bible reflects the ancient division as to length. There is, however, no proof that these differences of length existed at the time of the vocalization of the text. In any case the vowel signs are indicative of seven qualities only (excluding the šewa and ḥaṭefs). Phonematic investigation does not indicate any semantic differences dependent on vowel length. Such proofs as חוֹל (ḥōl) = sand opposed to חֹל (ḥol) = secular or profane, or עֵד (ʿēd) = witness, opposed to עֵד (ʿéd) = a piece of cloth, are misleading, since according to the traditional division the vowels are "long" in both pairs of words and the assumption that the vocalizers differentiated is completely unproven. According to the traditional pronunciation two pairs of homonyms can be seen, each of which arose in specific situations. The difference between the vocalization of official Hebrew today and that which is the assumed basis of the rules of vocalization – which are obligatory even today – lies not (at least from the phonemic point of view) in quantity, but in the change of quality of several vowels. Though it is claimed that Hebrew vowels are isochronic phonemes, this does not mean that in Hebrew speech all a vowels, for example, are of the same length in all situations. But on the other hand, the difference between מִן (min) = from, and מִין (mīn) = type, does not indicate that the vowel system is to be divided into long and short. The truth is that confronting such pairs is artificial since the two words differ in their syntax and linguistic position.
From the discussion above it is clear that a description of vowel distribution should be on two levels: (1) the vowels as pronounced today; (2) the use of the vowel signs. In describing the written language it is impossible to ignore the distribution of the vowel signs, since it is not only an important part of Hebrew spelling, but helps to understand the morphophonemic relationships. Table 4: Syllables reflects the distribution of the vowels which are listed according to type of syllable and place of accent in the word.
From the above it is clear that there are altogether five phonemic vowels – i, e, a, o, u. This can be ascertained from a consideration of the possible oppositions, ten in all . We also see that all possible oppositions are utilized only in (a) open syllable, whether accented or not and in a (b) closed accented syllable; in an (c) unaccented closed syllable many oppositions neutralized, as the examples below will show.
In syllables of the types (a) and (b) i is defined by its opposition:
- to é in שְׁכֵנָה/שְׁכִינָה ;נֵר/נִיר;
- to a in רַחֵם/רִחֵם ;שָׁרָה/שִׁירָה ;פָּעַל/פָּעִיל ;רַק/רִיק ;דָן/דִין;
- to o in בְּכוֹרָה/בְּכִירָה ;שׁוֹר/שִׁיר;
- to u in נֻחַמוּ/נִחֲמוּ ;אֲמוּרָה/אֲמִירָה ;צוּר/צִיר; é is defined by its opposition to i (see above);
- to a in צָדָה/צֵדָה ;עַד/עֵד ;גָּר/גֵּר
- to o in גְּדוֹלָה/גְּדֵלָה ;כָּבוֹד/כָּבֵד ;קוֹץ/קֵץ
- to u in שְׁכוּנָה/שְׁכֵנָה ;גּוּר/גֵּר; a is defined by its oppositions to i and é (see above);
- to o in קוֹרָה/קָרָה ;אָמוֹר/אָמַר ;שׁוֹר/שָׁר;
- to u in צוּרָה/צָרָה ;אָמוּר/אָמַר ;שׁוּר/שָׁר; o is defined by its opposition to i, é, a (see above);
|open syllable||closed syllable|
|* Accented syllables are marked by a single quote. Notes to Table 4:|
|(1) The examples show the vowels (a) in an opening syllable, (b) in a medial syllable (c) in the word final syllable; between each type there is a semicolon. Although there is no real difference between the first two categories, examples are cited for the sake of completeness. This enables us to see whether a certain vowel cannot appear in one of the syllables of the word, which situation is indicated by a dash.|
|(2) In each type examples are given for orthographical reasons. A comma is placed between them. This gives us a view of the distribution of the vowel signs. It may be concluded that each of the six vowels, except for é, can appear in all positions of the word. é alone does not appear in a closed unaccented syllable. The šewa, is – phonemically – the realization of a cluster, but from the phonetic point of view it is a syllable peak, and appears only in an unaccented syllable. The ḥatefs are to be classed with the šewa, and even those who pronounce the ḥatefs as full vowels do not produce any change in the above distribution picture. However, if the uses of the vowel signs and their relationship to the vowels which they represent are investigated, it will be seen that, unlike the vowels themselves, their use is limited and conditioned by the type of syllable in which they appear, and the rule is as follows: (1) šere, ḥolem, qameṣ gadol, and šureq do not appear in unaccented closed syllables; (2) qameṣ qaṭan and qibbuṣ do not appear in closed or open accented syllables. This shows that there is a connection between the distribution of vowel signs and accentuation. This relationship cannot be explained if only the accepted Hebrew pronunciation is considered; but must be seen against the phonetic values of the signs in the pronunciation of the Tiberian vocalizers. Similarly there is apparently a relationship between the type of syllable and the use of the full vowel signs (with ו and י), cf. the use of the ḥolem without a ו in קָטֹנְתִּי. This relationship is also not comprehensible given the modern pronunciation, but only on the basis of Hebrew pronunciation in the ancient past, when two important factors were in force: quantitative differences of vowels and a tendency to express that quantitative difference in the writing, as, for example, עוֹלָם opposed to אֹמֶר. This presents one of the greatest problems in teaching proper vocalization, for without the clear presentation of the historic background it is, of necessity, a mechanical process.|
|יִۧפֶן, שִۧימוּ;||רִחֵם, כִּישׁוֹר;||מִן, מִין;||גְּבִרְתִּי, שִׂמְחָה;|
|i – ḥireq||חָנִۧינוּ;||אַבִּירוּת;||–;||אוֹיִבְךָ;|
|é – ṣere||הוֹצֵۧאנוּ;||מַקְהֵלָה;||הֵסֵۧבּוּ;||–;|
|צֵא, הַקְנֵה||–||לֵב, לֵיל, מֵסֵב||–|
|e – segol||סוּסֶۧיךָ;||–;||–;||מַקֶּלְכֶם;|
|מִרְעֶה||דֶּۧשֶׁא, הֶۧגֶה||אֱמֶת, בַּרְזֶל||קֹۧדֶשׁ|
|pataḥ||קָۧמוּ, נַۧחַל;||גַּג;שָׁמַר, יַעַבְרוּ, נַחֲלָה;||מַלְכּוֹ;|
|a – and||עָשָׂۧהוּ, שְׁמָרַۧנִי;||מִשְׁמָרוֹ;||הוֹרַדְנוּ;||הוֹרַדְתֶּם;|
|qameṣ||שִׁמְךָ, מַלְכָּה||–||עוֹלָם, דּוֹנַג||רۧוּחַ, בֹּۧחַל|
|ḥolem||קֹۧדֶשׁ;||צָהֳרַۧיִם, גֹּרַשׁ, קוֹנֶה;||קֹמְנָה, תָּשׁוֹבְנָה;||כְּתָבְתִּי, קָרְבַּנְכֶם;|
|o – and||נָדוֹۧנוּ;||אַרְצוֹתֵיהֶם||קָטֹۧנְתִּי;||כֻּתָּנְתָּם;|
|qameṣ||פֹּה, מַלְכּוׁ||אָדֹם, כָּבוֹד||–|
|šureq||קۧוּמוּ;||רֻחַם, סוּגַר;||–;||וּלְקַחְתֶּם, חֻלְצָה;|
|u – and||תָּקۧוּמוּ;||מַלְבּוּשִׁים;||–;||מַשְׂכֻּרְתּוֹ;|
|šewa||–;||אֳנִיָּה, חֲמוֹר, גְּבוּל;||–;||–;|
- 10) to u in אָמוּר/אָמוֹר ;שׁוּק/שׁוֹק; u is defined by its oppositions to the other vowels (see above).
In syllables of type (c) only the following oppositions are found:
i (1) to a in דַּבֵּר/דִּבֵּר; (2) to o as הָפְקַדְנוּ/הִפְקַדְנוּ; (3) to u in פֻּנָּה/פִּנָּה ;חֻיַּבְנוּ/חִיַּבְנוּ; a opposed to i (see above); (4) to o in צָצְמוֹ/צַצְמוֹ; (5) to u as הֻקֵּשׁ/הַקֵּשׁ ;חֻפָּה/חַפָּה. It will be seen, therefore, that é is not opposed to any vowel in this position and that o/u are not opposed; however, variants like אֻמְנָם/אָמְנָם and הֻפְקַד/הָפְקַד are found in Hebrew. In fact, most of the oppositions are derived from the conjugations of the verb, where o/u indicates the passive and a/i the active forms, and even in this area the oppositions are limited. This is the result of a process in ancient Hebrew, during which oppositions of short vowels in closed syllables were eliminated; and when the use of the internal passive in post-biblical Hebrew was minimized the scope of these oppositions was, automatically, greatly reduced.
In the above description the e (segol) was not included in the phonemes although there are cases where a difference of meaning between a pair of words is reflected in the relationship between the segol and some other vowel, as: לַחַם/לֶחֶם ;עֵרֶב/עֶרֶב ;אַרְאֶה/אֶרְאֶה (= solder). Though the segol is a very common vowel, such cases are quite rare and we cannot therefore assume the opposition segol to another vowel in the structure of the language. It is correct to see the segol as an allophone of the ṣere. If the forms אַרְאֶה/אֶרְאֶה through the whole of their paradigms, are investigated, it will be seen that the opposition in all conjugated forms is a/i, as, for example תַּרְאֶה/תִּרְאֶה; only in the first person singular is this i represented by a segol, since the i does not appear after an alef in a closed syllable (when not geminated). The é should have taken the place of the i but é cannot appear in an unaccented closed syllable. This complimentary relationship between é and e is common in all forms of the conjugation and declension of ל״י roots, as: מִקְנֵה/מִקְנֶה (construct state); לֵב/לֵב ;מִקְנֶיהָ/מִקְנֵהוּ, מִקְנֵינוּ (possible in construct state); בֶּן/בֵּן (construct); and others. On the other hand e is found at times in the same environment as é in פֵּתַח/פֶּתַח ;בָּהֵן/בָּהֶן ;שֵׂכֶל/שֶׂכֶל ;יֵתֶּר/יֶתֶר ;נֵצַח/נֶצַח ;נֵדֶר/נֶדֶר and others without any distinction. It may, therefore, be concluded that at times e is a conditioned allophone which becomes mandatory in certain environments, and at times – under different conditions – it is (in very limited scope), an optional allophone (cf. N.S. Trubetzkoy, Grundzuege der Phonologie, p. 46, regarding d/t). Examples such as עֶרֶב (evening)/עֵרֶב (wasp) which both contain the phoneme é should be considered homonyms as are their plural forms עֲרָבִים and as is עָרֵב (pleasant/guarantor) etc. (The fact that the segol appears only with the א and is a conditioned vowel, and not independent, is also seen clearly in the forms אַעֲלֶה/אֶעֱלֶה; the segol does not appear in the rest of the qal paradigm, as ‡תַּעֲלֶה/תֶּעֱלֶה, but all the forms save the first remain homonymous.)
Although in modern Hebrew the segol must be considered to be an allophone of the ṣere, it is possible to prove that in ancient Hebrew the segol is a reflex, in all cases, of an original a and was an allophone of a.
a) The grapheme known as the šewa (–) represents two independent phonetic values: the absence of a vowel, and a very short vowel, which can be described as central and vague. In grammatical terminology the former is called šewa quiescens (שווא נח) and the latter šewa mobile (שווא נע). Only the second type interests us in our study of the behavior of the language. Indeed this term, šewa, has become accepted in general linguistics as describing a vowel of this quality. In Hebrew the šewa cannot rightfully be listed with the phonemes, since no difference in meaning depends on the šewa (the same applies to ḥaṭefs). The šewa must be regarded as a conditioned vowel which appears in clusters of consonants, whether, historically, the šewa comes in place of a full vowel (גָּדוֹל>גְּדוֹל) or the absence of a vowel (מַעְבָּר>מַעֲבָר compare מַעְבָּרָה). Therefore, phonetic differences such as פְּסוּל–פָּסוּל or זְרוֹעַ–זָרוֹעַ have the following phonologic makeup: /pasul/:/psul/, /zaroʿ/:/zroʿ/. In Hebrew the šewa is never accented and is always found between two consonants.
b) Since there is one grapheme, the šewa, for two values, the grammarians established rules indicating when the šewa was to be pronounced as mobile (נע); however, the pronunciation of the vocalizers who instituted the šewa sign differs from the Sephardic pronunciation, whose rules are accepted in the pronunciation of official Hebrew today. In common speech, different groups of speakers pronounce the šewa differently, or do not sound it at all. The following are the rules for educated speech: the šewa is pronounced (a) at the beginning of a word (קְנֵה, בְּדִיל, גְּבוּל); (b) when it is the second of two šewas (יִשְׁמְרוּ); c) when it comes with a dagešed consonant (יִפְּלוּ); d) after a vowel in an open syllable (יִי־רְאוּ, שִׁי־רְכֶם). Since the ability to distinguish the length of vowels has been largely lost, only a trained ear can determine which is an open syllable followed by a consonant plus šewa mobile, and which a syllable closed by the first of a cluster of consonants. Only when the cluster consists of a geminated consonant is the difference clear: גָּדְלוּ, סָבְרוּ are at times pronounced (sav-ru) (gadlu), but not גָּלְלוּ, סָבְבְוּ (savĕvu) (ga-lĕlu). This differentiation was passed by analogy to words like קִ־לְלַת, קִ־לִלוּ, whose first syllable is – historically – closed. In this way a differentiation developed between רִנַּת and רִ־נְנַת which was apparently unknown to early Hebrew, where both were pronounced simply (rin-nat); only in certain prosodic situations could רננת be pronounced (ri-nĕnat).
c) There are those who, in addition to the two types of šewa mentioned, find in Hebrew a third type which they call šewa medium בֵּינוֹנִי or מְרַחֵף. This is a šewa which comes after a "small" vowel (see above Vowel Quantity): (1) if the following consonant is a spirantized בגדכפ״ת, as מַרְבָד, מַלְכֵי (see section on בגדכפ״ת above); (2) with an originally geminated consonant, as אִלְמִים (compare singular הַמְקִימִים, (אִלֵּם (sing. הַמֵּקִים); (3) with the first consonant of a cluster as רִנְנַת (see above). This šewa medium is not a separate phonetic entity, but in types (1) and (2) the šewa is quiescent, that is, it does not represent any vowel in pronunciation, and in type (3) may be pronounced as a šewa mobile, as pointed out above. Phonetically there is no such thing as a šewa medium which is between the quiescent and mobile; it is a fiction created to explain certain phonetic developments which are not uniform. There are some grammarians (Bergstraesser and others) who tend to see in the šewa medium an independent historical entity, i.e., a šewa quiescens where once there was a vowel, while others see in it an independent phoneme (Birkeland, p. 55). However, a sound which is not an independent part of the order of sounds (since it is either quiescens or mobile) cannot be considered a phoneme. It must be remembered that both the šewa and ḥaṭefs, as well as the absence of any vowel, are likely to appear in Hebrew in a position where once there was a vowel (שָׁ־מְרוּ, in pause שִׁמְ־רוּ ;שָׁמָרוּ, in pause, שְׁמׁרוּ) or where there was no vowel at all (נֶ־אֱמָן opposed to כַּעְ־סוֹ, זַעְ־פּוֹ, נִשְׁ־מָר but נַ־עֲרוֹ).
D) In the official language the grapheme šewa reflects one sound (ě) in all phonetic situations where it is pronounced. In ancient Hebrew this sound varied between ă and ě, and tended to assume the sound of the neighboring vowels. This fact is reflected in personal names which have been transmitted in their Greek or Latin pronunciations as: סְדוֹם Sodom, שְׁלֹמׁה Solomon, נְתַנְאֵל Nathanael, גְּדֵרָה Gedera. In the pronunciation of the vocalizers the šewa was generally ă (this caused it to be interchanged with the ḥaṭef-pathaḥ), and it changed, according to ancient rules, toward the sound of the neighboring vowel when the šewa is next to אהח״ע. This feature can still be heard in the pronunciations of several communities where כְּמוֹ and נְקִיָּה are pronounced kămo, năqiyya and not kěmo, něqiyya. The different nuances of the šewa in official Hebrew are the ḥaṭefs (if pronounced quickly), and these sounds are interrelated with the consonants ח, ה, א, and ע; only rarely is a ḥatef found with a different consonant, as דֳּמִי, כֻּתֳּנוֹת, צִפֳּרִים. In the Hebrew of the vocalizers the šewa and ḥaṭef-pathaḥ represented one quality; in the official pronunciation today the šewa and ḥatef-segol are equal, the latter with a guttural.
As opposed to what has been said above, on the relation-ship between the šewa and the ḥaṭefs, it is possible to claim, that semantic differences are dependent on the ḥaṭefs, unlike the šewa, and as a result the ḥaṭefs are phonemes. The differences are of the type אֳנִי/אֲנִי, חֳדָשִׁים/חֲדָשִׁים, עֱלִי/עֲלִי, חֳלִי/חֲלִי. However, as pointed out above regarding the segol (in section 12, The Phonological Status of the Vowels, above), it may be said that the ḥaṭefs are reflexes of the šewa – the result of a cluster of consonants – while the choice as to which nuance of the šewa is used in a particular instance depends, to a great extent (at least with regard to the ḥatef-qames), on the quality of the full vowel found in the same position in other forms of the word; thus חֲדָשִׁים follows חָדָשׁ and חֳדָשִׁים follows חׁדֶשׁ. The oppositions of the ḥaṭefs are only apparent; the real oppositions being between the full vowels; there is thus the possibility of ḥatef interchanges within the very same word, as in שִׁבֲּלֵי– שִׁבֳּלִים, דְּמִי–דֳּמִי, אֲמָרְכֶם–אֱמָרְכֶם–אֱמׁר, אֲכָלְךָ–אֱכׁל.
Diachronically, and also from the phonologic synchronic point of view, the furtive pathaḥ must be considered the same as a ḥatef which is pronounced before the guttural. However, unlike the ḥatef, the furtive pathaḥ is always pronounced as a full a vowel.
The accent in Hebrew falls on one of the last two syllables of the word. On the ultimate syllable it is called מלרע (milraʿ) and on the penultimate מלעיל (milʿel). It is impossible to determine clear phonological rules for each type of accentuation since the situation in Hebrew is the result of a complicated development, not all of which is clear today. One may say that in Hebrew the ultimate accentuation is dominant, while the penultimate is found:
A) in the noun:
1) in segolate forms, i.e., when preceding the final consonant there is
a) a segol (תִּלְבּׁשֶׁת, מִסְפֶּרֶת, כּׁתֶבֶת, שׁוֹמֶרֶת, כּׁתֶל, סֵפֶר, מֶלֶךְ);
b) a pathaḥ – if the last consonant is ח, ה, or קֶבַע, מֶלַח, גּׁבַהּ, תֵּמַהּ) ע) or if the consonant before the last is ח, ה, א, or נַעַל, נַחַל, רַהַב, שֹׁהַם, תּׁאַר) ע);
c) ḥireq if the consonant before the last is לַיִל, בַּיִת) י). In all of these forms the common feature is that the unaccented vowel is lost in declension. To this category belong also those nouns ending in an open syllable with segol, hiriq, or šureq, which when declined place a consonant for this vowel, as אַחְוֵי–אָחוּ, יָפְיוֹ–יׁפִי, דִּשְׁאוֹ–דֶּשֶׁא; this is also the case for the demonstrative אֵלֶּה;
2) nouns (and other words) to which are suffixed locative ה–ָ or the dual ending מִסְפָּרַיִם, מַיִם ;שָׁמָּה, מִצְרַיְמָה, אַרְצָה) ַ–יִם similarly הֵן = הֵנָּה, הֵמָּה, לַיְלָה);
3) nouns (and other words) to which are suffixed the possessive pronouns: אוֹתָנוּ, אָחִינוּ, מַלְכֵּנוּ, חָזֶהָ, אָבִיהָ, מַטֵהוּ, עֵינַיִךְ, עָלֶיךָ, סוּסֶיךָ) –נוּ, –הָ, –הוּ, ַ–יִךְ, ֶ–ךָ); similarly מוֹ–, which today is only found in poetry (עָלֵימוֹ, בָּתֵּימוֹ).
B) in the verb:
4) in the perfect: before the suffixes הִשְׁכַּמְנוּ, שִׁבַּרְתָּ, שָׁמַרְתִּי) –נוּ, –תָּ, –תִּי);
5) in the imperative and the imperfect: before תְּשַׁבֵּרְנָה, שְׁמׁרְנָה) –נָה);
6) in addition to those instances mentioned in (4) and (5): in hif ʿil of all forms, excluding ל״י forms, and in עו״י and ע״ע forms in qal and nif ʿal, and in ע״ע forms even in huf ʿal, in forms ending הוּסַבָּה, יָסֵבּוּ, טוֹבוּ, בּׁאִי, קָמוּ, הַגִּידוּ, הַגִּידָה) –וּ, ִ–י, ָ–ה);
7) in verbs with suffixed object יִמְצָאֵנוּ, יִשְׁמְרֶנָּה, יִשְׁמְרוּהָ, יׁאכְלֶנּוּ, יׁאכְלֵהוּ, קָנָהוּ, יׁאחֲזֵנִי, שְׁמָרַנִי) –נוּ, –נָּה, –הָ, –נּוּ, –הוּ, –נִי).
c) in certain forms which today appear only in poetry, such as the pausal forms (יִשְׁמׁרוּ, שָׁמָרוּ, אָנׁכִי, אָָנִי) or the inversive tenses (וַתָּגֵד, וַיָּשֶׂם, וַיָּקָם).
There are exceptions to the above rules: the accent is on the penultimate in the third person fem. sing. perfect with pronominal suffixes שְׁמָרָתַן, שְׁמָרָתַם, אֲהֵבָתֶךְ, אֲהֵבָתֶךָ, the accent is on the ultima in the inversive perfect tense וְשָׁמַרְתָּ, וְשָׁמַרְתִּי.
If for the moment the segolate nouns and the forms appearing in inverted tenses are excluded, it may be concluded that the penultimate accent appears in Hebrew only when the word ends in an open syllable; this, to a certain degree, reflects the early division of ultimate and penultimate accents in Hebrew.
In the Bible the accent is given to change from penultimate to ultimate (as in עוּۧרִי in Judg. 5:2) and especially from ultimate to penultimate in word groups, but there are no cases of enclisis, only proclisis. In modern Hebrew, however, enclitic forms are common not only in speech but also in poetry; as opposed to אֱמָר־ۧלִי, בּוֹא־ۧנָא, שְׁמַע־ۧנָא we find אֱۧמוֹר־לִי, ۧבּוֹא־נָא, שְׁۧמַע־נָא; however, this is not considered the norm. Proclitic forms are common in construct. In speech there are certain tendencies to penultimate accentuation which differ from the rules given here, but they are not considered correct (as שָׁۧמַרְתֶּם). The accentuation of foreign and borrowed words and personal names must be considered separately. The rules of accentuation in the original languages have affected the accentuation of the borrowed words, so that at times even syllables before the penultimate are accented. In personal names the emotional factor cannot be ignored, and feelings such as love and indulgence affect the accentuation. However, the standard which is demanded by the Academy for Hebrew Language and used in broadcasting is the ultimate accentuation, and so the norm is אוּנִיבֶרְסִיۧטָה and not אוניברۧסיטה or אוניۧברסיטה, and similarly ۧרָחֵל and not ۧרָחֵל. On the other hand, foreign words with Hebrew suffixes are usually accented in keeping with the norm, as ליגאליוۧתה, אוניברסיטۧאות. The tendency to penultimate accentuation is found in the pronunciations of many communities, and in Samaritan Hebrew has become the rule.
Accentuation has phonemic value in Hebrew, since it is a distinguishing feature between certain pairs of words – ۧבָּנוּ (in us) / בָּۧנוּ (they built), ۧבָּאָה (she is coming) / בָּۧאָה (she came), ۧקוּמִי (imp. fem. sing.) / קוּۧמִי (infinitive with pron. suf.). Hebrew accentuation is "stress" type (dynamic accent) and by its nature is likely to affect the vocalization. In fact it causes changes in the vowels of a word depending on their nearness to the accent (see section 16. Interchange and Elision of Vowels, below).
The diphthong is – by definition – the combination of two vowels within one syllable. The diphthong is created when the point of articulation glides from place to place within the one breath. More precisely it is a combination of a series of vowels, but it is sufficient to indicate the extreme vowels, i.e., the opening and the concluding (or the intended conclusion). Thus instead of the sign a..e..é..i, the sign ai (or ay, ai̭) is used. By its very nature one of the parts of the diphthong is primary, and the other secondary, or accompanying. If the diphthong begins with the accompanying element it is called rising (as ṷa); if it ends with this element it is called falling (aṷ).
In Hebrew grammar, combinations such as וּי, ַ–י, יַ, ִ–יו, יָו, וַ are called diphthongs. However, in Hebrew it is the falling diphthongs that are of significance because of the morphological changes they cause. The use of the term diphthong in Hebrew is therefore limited to cases of a vowel with "semi-vowel," ו or י (see section 7. Other Consonants, above). A combination of two vowels in one syllable, whether caused by the splitting of one vowel into two (as in the frequent pronunciation of the ṣere, ey), or by the proximity of a certain vowel to ח, ה, or ע (furtive pathaḥ) – which is realized phonetically by being split into two syllables – concerns Hebrew grammar only as far as establishing the phonetic facts. In fact the term diphthong does not accurately describe the combination of a vowel and ו in the regular pronunciation of modern Hebrew, where it is a combination of a full consonant (v) and vowel. However, this combination can be regarded as a diphthong because in the pronunciation of some communities it is actually articulated as such, and because it interchanges morpho-phonemically with vowels, unlike the combination of vowel plus spiranted ב which is phonetically its equal. In the interchanges the early history of this combination is still apparent. (The fate of the early aw, ew, etc., in modern Hebrew is similar to that of the same diphthongs in late Greek.) From the point of view of function it is clear that the Hebrew diphthongs are syllables with a consonant; not a phoneme in the precise meaning of the word, but a bi-phonematic element. The only interchange within the provenance of the rising diphthong is the וי״ו החיבור (waw copulative) becoming וּ (u) before a cluster of consonants and before מ, ו, ב and פ as וְשֵׁם but וְנָתַן ;וּשְׁ־מִי but וְיֶלֶד ;וּמָשַׁךְ but וּוֶרֶד. In all other changes of vowel the waw copulative acts in the same way as the consonants כ, ב, and ל. The total of diphthongs is eight, that is ו with i,e,a,u, and י with i,a,o,u. The falling diphthongs which are always maintained and do not interchange are iw, ew, uw (וּוְ־לָדוֹת ;גֵּו, כִּסְלֵו ;עִוְרִים, אָחִיו); oy, uy (גָּלוּי, נוֹי) while the other diphthongs are replaced in the paradigm of the word by vowels, or are split into two syllables. Even though there are no absolute rules, there is great consistency, especially with regard to the diphthongs aw, ay. The main points are as follows:
A) iy. iy is used only when the accompanying element is geminated, as קִיֵּם, עִבְרִיִּים, עִבְרִיָּה, סִיּוּם. The combination iyyi is at times interchanged with i as עִבְרִיִּים = עִבְרִים, צִיִּים = צִים.
i appears in all other situations, such as: לִיהוּדִים, מִימִינוֹ, בִּירוּשָׁלַיִם (in poetry there is sometimes to be found בִּיְרוּשָׁלַיִם, לִיְהוּדִים, מִימִיְנוֹ). The morphological variant יָּה–(feminine ending) / ית–ִ is connected with this phenomenon.
B) aw. aw is found (a) within the word (1) when the accompanying element is geminated, and (2) in the syllable preceding the heh locale; and (b) at the end of a word when it concludes that word.
o is found within the word in an unaccented syllable; as far as this rule is concerned "within the word" includes the end of the word in construct, provided that the syllable is closed. awe is found at the end of a word, when, after the diphthong that should rightfully appear, there is a consonant and an accented syllable, that is always in the absolute state.
Note: "End of the word" includes monosyllabic words. Examples for the interchange awe/aw/o are: /מוֹת־/מוֹתוֹ/הַמָוְתָה בָּנָיו, עָנָו, סְתָיו ;צַו/צַוֵּה ;כְּגוֹן/גּוֹנוֹ/גָּוֶן ;חֲצַרְמוֹתִי/חֲצַרְמָוְתָה/חֲצַרְמָוֶת ;מָוֶת.
C) ay. ay is found (a) within the word (1) when the accompanying element is geminated, and (2) in the syllable preceding the heh locale, and (b) at the end of a word when the syllable ends with it and is not a construct form.
é is found within the word in an unaccented syllable (and here "within the word" includes the end of the word in construct forms); e is found in the same positions as é, and in place of it before the pronomial suffixes הָ, –ךָ–;; ayi comes at the end of a word, when, after the diphthong that should be there, there is a consonant and the syllable is accented. Examples for the interchange ayi/ay/é/e are מָתַי ;חֵי־/חַי ;/דַּיּוֹ/דַּי דֵּי־ ;לֵיל־/לַיְלָה/לַיִל ;מַיִם/מֵי־/הַמַּיְמָה ;בֵּית־ ;בּיתוֹ/הַבַּיתָה/בַּיִת ;סוּסֶיהָ, סוּסֶיךָ/סוּסֵיכֶם, סוּסֵינוּ/סוּסַי סוּסַיִךְ. Exceptions to the above rules are found in both directions, and in modern Hebrew more so than in the past. The important exceptions are the following:
The diphthong is found (a) where gemination has disappeared (see 8. Gemination and Clusters) as: וַיְחִי (see הַיְהוּדִים, (וַיֶּחִי; (b) in words where the accompanying element is felt to be essential for maintaining the paradigmatic connection, as in רַוְחָה, שַׁוְעָה, עַוְלָה (which are connected etymologically to מַוְרִידִים, הַיְשַׁר, מַיְמִינִים, שָׁלַוְתִּי, (רֶוַח, שֶׁוַע, עָוֶל (but מוֹרִידִים from יָרַד!). A vowel is at times found alongside a split diphthong, as: לֵיל alongside אֵין, לַיִל alongside אוֹן, אַיִן alongside אָוֶן (with a semantic difference), צִים alongside עִבְרִים, צִיִּים alongside עִבְרִיִּים. Today the tendency (apparently also in other areas) is to exploit the phonetic variants יִּים–ים– for semantic distinctions. Comparative grammar teaches that in various words which today have a vowel, there was originally a diphthong (for example סוֹף, יוֹם) and in several instances this fact is reflected in the plural forms, as שְׁוָרִים from שְׁוָקִים, שׁוֹר from דְּוָדִים, שׁוּק from דּוּד (also לְוָחִים, (דּוּדִים from לוּחַ (alongside לוּחוֹת). At times – completely exceptional – a vowel other than the ones listed in the rules above interchanges with a diphthong; for example i in the already archaic form, in the Bible, עִירׁה (ass; Gen. 49:11) and in the common word עִיר (town), where not only the archaic plural עֲיָרִים (Judg. 10:4) hints at an original diphthong, but even the derivative word עֲיָרָה (small town) (through עֲיָרוֹת) retains this connection; e is found alongside é in גֶּיא.
Generally it can be said:
1) finding a vowel in a position where a diphthong is expected is part of a general tendency of early Hebrew, and in many words and forms there is no remnant of this original diphthong;
2) finding a diphthong in a position where a vowel is expected is an increasing tendency of later and modern Hebrew, due to the morphological considerations stated above, and thus there is to be heard not only שָׁמַיְמָה but also שָׁמַיְמִי (not considered a literary form), דַּוְקָא, כַּוְנֵן, דַּיְסָה (and cf. Bialik in Ha-Berékhah; גַּוְנֵי גְּוָנִים). Variants such as עַוְלָה/עוֹלָה (cf. Job. 5:16; עַוְלָתָה/עׁלָתָה Ps. 92:16) שׁוֹעַ/שֶׁוַע (Isa. 22:5) are not produced in modern Hebrew.
The expansion of the diphthong and even its splitting into two syllables occurs in late Hebrew even more than in modern official speech: it is found in medieval Mss. and in modern Samaritan Hebrew. The special relationship between the diphthong and vowel affects the spirantization of the בגדכפ״ת (see section 5. בגדכפ״ת above) when near a diphthong. Within the word the diphthong usually acts as a vowel and causes the spirantization of בגדכפ״ת, as in הַיְבָמָה, הַיְבוּל, בַּיְתָה, מָוְֿתָה but שָׁלַוְתִּי.
Examining the diphthong and its interchanges, we learn that:
a) ay, aw display almost (cf. above aw (a) 2: ay (a) 2) identical traits: this is not the case in, for example, biblical Aramaic and later Jewish Aramaic, where ay is still found without gemination, whereas aw is always interchanged with a vowel. In an effort to limit the occurrences of the diphthong (which is common in Canaanite) by interchanging it with a vowel, Samaritan Hebrew went much further than classical Hebrew (although there are parallels in Jewish traditions); the original aw diphthong has disappeared completely while the ay diphthong appears only rarely;
b) the type of syllable – open or closed – and the place of accentuation affect the diphthongs and their interchanges. Still, it is impossible to establish pure phonological criteria for the above rules, since the phonological rules which governed biblical Hebrew with great regularity have long ceased to operate with respect to the quantity of the vowels and the structure of the syllable, and a new situation has developed in which the morphological factor has become dominant. This new situation, reflected in the vocalization of the Bible, continues to spread. Mention should be made not only of interchanges of the diphthong but to all vowel changes, to be discussed below, and accentuation discussed above. In these it is possible to discern morphophonemic phenomena and in fact morphophonemics plays an important role in Hebrew grammar. This may explain the large number of exceptions to phonological rules, especially in matters of the vowel system. These exceptions can be grouped into morphological rules (cf. above aw b); ay b); cf. also section 5. בגדכפ״ת above and the Historical Note in section 16 on Interchange and Elision of Vowels, below).
The dominance of morphological principles over phonological is one of the features of a literary language which continues to be used many generations after certain phonological rules have ceased to be operative and are exchanged for other rules.
In the light of the above-mentioned assumptions that vowel quantity is not a distinguishing feature in Hebrew (see section 10. Vowel Quantity, above), and that almost any vowel quality can appear in every type of syllable (see section 11. Vowels as Part of the Syllable, above), the fact that a given vowel is maintained in all the forms of a paradigm but interchanged with another vowel in a different paradigm is rather surprising. This is a complicated aspect of Hebrew grammar which cannot be understood without resort to its historical background. Consideration will be given first to the phenomena themselves, excluding those interchanges caused by ח, ה, א, and ע (see section 6. אהח״ע above).
A) Interchange of Vowel with Vowel.
1) é is interchanged
(a) with i, as הִנָּם/הֵן, קִנִּים/קֵן, לִבּוֹ/לֵב, סִפְרִי/סֵפֶר, עִנְבֵי־/עֵנָב, יְקִלֵּנִי/יָקֵל, מָגִנִּים/מָגֵן, מְסִבִּים/מֵסֵב;
(b) with e, as בֶּן/בֵּן (construct), יוֹצֶרְכֶם/יוֹצֵר, יִתֶּנְךָ/יִתֵּן, מַקֶּלְךָ/מַקֵּל; and in biblical forms used today only in poetry, as: וַ֫יֵּלֶךְ/יֵלֵךְ, וַ֫יָּקֶם/יָקֵם;
(c) with a, as: זְקַן־/זָקֵן (construct), תֵּלַדְנָה/תֵּלֵד, קַן/קֵן, תִּשָּׁמַרְנָה/תִּשָּׁמֵר (construct), יָשַׁנְתִּי/יָשֵׁן, כַּנּוֹ/כֵּן, מַקַּל־/מַקֵּל. But: é is retained in חֵיק :חֵיק, עֵדְךָ :עֵד, מֵתוֹ :מֵת (construct), הֵיכְלִי :הֵיכָל, אֵינְכֶם :אֵין, תְּשַׁמֵּרְנָה :תְּשַׁמֵּר; in certain construct forms é is retained alongside e as: אֶשׁ :אֵשׁ, לֶב :לֵב; i is not interchanged in: נִחַמְנוּ :נִחֵם, בִּיתָנֵי :בִּיתָן, נִיצוֹצֵי :נִיצוֹץ, שִׁיר :שִׁירוֹ.
2) e is interchanged
(a) with i, as in גַּרְזִנּוֹ/גַּרְזֶן, בַּרְזִלָּם/בַּרְזֶל, כִּבְשָׂה/כֶּבֶשׂ, צִדְקוֹ/צֶדֶק, יְפֵהפִיָּה/יְפֵהפֶה, פִּיּוֹת/פֶּה, כַּרְמִלּוֹ/כַּרְמֶל;
(b) with a, as in מַפְסַלְתְּךָ/מַפְסֶלֶת, שׁוֹמַרְתּוֹ/שׁוֹמֶרֶת, מַלְכּוֹ/מֶלֶךְ; in biblical forms such as: וַיּׁאמַר/וַיּׁאמֶר (pausal), וַיֵּלַךְ/וַיֵּלֶךְ;
(c) with e, as in: יִקְנֵהוּ/יִקְנֶה, שְׂדֵה־/שָׂדֶה, קוֹנֵהוּ/קוֹנֶה.
3) a is interchanged with i, as in: פִּתָּה/פַּת, גִּתִּים/גַּת, בִּתָּם/בַּת, סִפִּים/סַף, מִסִּים, מַס, but maintained in גַּלֵּי/גַּל, גַּגּוֹת/גַּג, חַגִּי/חַג, עַמִּים/עַם.
There are forms where both possibilities exist: e.g., זַלְזַלִּים, סַנְסַנִּים, but even in modern Hebrew סַנְסִנָּיו is also found, and the same is true for גַּלְגַּלֵּי, גַּלְגַּלִּים and גַּלְגִּלַּיִם, גַּלְגִּלֵּי.
4) o is interchanged with u, as in: נְסוּגוֹתִי/נָסוֹג, מְנוּסִי, מְנוּסָה/מָנוֹס, מְתוּקָה/מָתוֹק, יְסֻבֵּנִי/יָסׁב, זָּם/מָעׁז מָעֻ־, חֻלּוֹ/חׁל, רֻבָּם/רׁב but is retained in: מְלוֹנוֹת :מָלוֹן, חוֹלֹו :חוֹל, שׁוֹרָם :שׁוֹר. But there is also עָזִי :עׁז.
5) u is interchanged with o, as in: קוֹם, תָּקׁמְנָה, וַיָּקָם/יָקוּם (infinitive).
B) Elision of Vowels.
Here we refer both to šewa quiescens, which is phonetically a zero vowel, and to the šewa mobile, which is a zero vowel from the phonemic aspect, but realized as an ě vowel in a cluster of consonants. In fact the distribution of the šewas depends solely on the type of syllable, and each can replace the other.
1) é / šewa as in: בְּרֵכָתָם :בְּרֵכָה ;מַקְ־דְּחוֹ/מַקְדֵּחַ ;לִשְׂ־מֵחֵי, שְׂמֵחֵי :שָׂמֵחַ ;חֲפֵצֵי :חָפֵץ ;הֵיכָלוֹת :הֵיכָל ;(יִתְּ־תְּנוּ=) יִתִּנוּ/יִתֵּן ;יוֹ־צְרוֹ/יוֹצֵר ;אוֹת/כִּסֵּאכִּסְ־ ;מַקְ־דֵּחוֹ/מַקְדֵּחַ ;נִבְ־לָתוֹ/נְבֵלָה ;חַבְ־רֵי/חָבֵר ;עֲנָ־בִים/עֵנָב ;סִבִּים בִּמְ־, מְסִ־בִּים/מֵסֵב ;לִסְ־פָרִים, סְפָ־רִים/סֵפֶר.
2) e / šewa, as in: יוֹ־צְרוֹת/יוֹצֶרֶת ;לִמְלָ־כִים, מְלָ־כִים/מֶלֶךְ.
3) a / šewa, as in: יִלְ־בְּשׁוּ/יִלְבַּשׁ ;בִּרְ־כוֹתֵיהֶם, בִּרְ־כַּת/בְּרָכָה ;דִּבְ־שָׁהּ/דבַשׁ ;כִּתְ־בֵי /כְּתָב ;שְׂמֵ־חֵי/שָׂמֵחַ ;זִקְ־נֵיכֶם, םלִזְ־קֵנִי־, זְקֵ־נִים/זָקֵן ;דִּבְ־רֵיהֶם, דְּבָ־רִי/דָּבָר; and others but: יִלְבָּשׁוּם ;עֲיָרוֹתֶיהָ :עֲיָרָה ;חֲטָאִים חֲטָאֵיהֶם : ;חָרָשֵׁי :חָרָשׁ and others.
4) o / šewa: יָ־כְלָה/יָכׁל ;יִשְׁ־מְרוּ, יִשְׁ־מְרֵם/יִשִׁ־מׁר ;מַחְ־לְקוֹת/מַחְלֹקֶת ;(צִפְּ־פֳּרים =) צִפֳּרִים/צִפׁר ;(שִׁבְּ־בֳּלִים =) שִׁבֳּלִים/שִׁבּׁלֶת ;קָדְ־קֳדִים/קָדְקׁד but מַשְׂכּוֹרוֹת :מַשְׂכּׁרֶת ;כְּתוֹבוֹתֵיהֶם, כְּתוֹבוֹת :כְּתׁבֶת and many others; some people retain the o even in the nouns צִפּׁר, שִׁבּׁלֶת and מַחְלֹקֶת.
The phenomena listed above present a many-faceted and complicated picture of vowel changes. The main points are as follows:
A) The vowels e, a, o in an open syllable or in an accented closed syllable tend to be interchanged or elided when the word is declined (the same is true when the word serves as the basis for a derivation; as סִפְרוֹן :סֵפֶר). However, in a closed unaccented syllable, they are always retained in all declensions and conjugations of the word: (n.b. In this regard the imperfect verbal forms are independent "words," and we are not to treat their vowels in connection with the perfect).
B) In e vowels the phenomena is limited to that morphological type called "segolate nouns." Since e is an allophone of /é/ (see sections 2. Writing and Spelling; 12. The Phonological Status of the Vowels, above), the interchange e / é is not of the type under discussion, and in fact its conditions and results are different from the other changes. Phonemically, there are no e interchanges, but these are part of é changes. The interchanges of u vowels are common only in a few forms and this vowel is not elided. In fact the historic basis for this change differs from that of the other vowel changes, and the only factor they have in common is that of accent.
C) The conditions for the interchange of e, a, o, and their elisions are clear: the place of the accent and the structure of the syllable. When an accented closed syllable becomes unaccented the above vowels tend to interchange; in an open syllable whose accent has been removed the vowel tends to be elided.
D) With regard to elision the noun acts differently from the verb, while the verb itself acts differently according to whether or not it has a pronomial suffix. A verb without pronomial suffixes elides the vowel next to the accent; a noun elides the vowel penultimate to the accent, the verb with suffixes acts at times like the noun (especially in the qal perfect) and at times like a verb without suffixes.
E) As regards the very consistency of elision, the verb differs from the noun, and the behavior of the o, é, vowels differ from that of the a vowel.
1) In a verb without suffixes the elision is consistent;
2) In the noun, é and o tend to be maintained while a tends to be elided.
3) In the verb with suffixes e and o are elided while in the same position a is maintained as: yišmor: yišměréni, yittén yittěnéni: but yilbaš, yilbašéni.
F) Although the phonetic conditions – accentuation and syllable structure – are determining factors, it is impossible to classify – phonologically – all the various vowel changes without involving the morphological factor. In fact, it is only possible to depict the vowel system by listing the various morphological types in which one or the other situation occurs. Even within the morphological types there are variations which confuse the Hebrew speakers. As far as the official language is concerned, the Academy for Hebrew Language determines what is correct in each type.
G) Historical Note: This complicated situation, full of phonological inconsistencies, is the result of the change which took place in early Hebrew when there were still quantitative differences in vowels (see section Vowel Quantity). When quantity ceased to be free and phonemic – that is, when the language no longer accepted short vowels in open syllables – it lengthened or elided such vowels (depending on their relationship to the accent). In this way there developed either long syllables (that is, consonant plus long vowel, or consonant plus short vowel plus consonant) or syllables with šewas (or ḥaṭefs). As a result, in the paradigm of a given word which contained an original short vowel there were forms with long vowels (that is the vowel was lengthened because of the accent and this in turn caused a change in quality) alongside forms with short vowels (in a closed syllable), or elided vowels. At the same time, there are words with original long vowels which are maintained in all the declensions of the word, for example:
1) עֵדָה (= assembly): ʿēḏā (<ʿidat): ʿăḏāṯi: ʿēḏot: ʿāḏōṯēḵā: עֵדָה (= feminine witness): ʿēḏā: ʿēḏātī: ʿēḏōṯ: ʿēḏōṯēḵā;
2) גִּבּוֹר: gibbōr (<gibbār): gibbōrěḵem: gibbōrīm: צפּוֹר ṣippor (<ṣippur): ṣippurḵem: ṣippǒrīm.
Since a short vowel in Hebrew has become mechanically lengthened or elided by the stress, the basis for interchange and elision of originally short vowels has been lost. As a result of reciprocal influences, originally short vowels have begun to act as long vowels and vice versa. é and o were originally both long and short, and as long vowels played a key role in maintaining what were originally short vowels in declension and conjugation (compare כְּתוֹבוֹת, צִפּׁרָה, לֵדַת, זֵעַת). a, which was originally a short vowel (the original long vowel was ō), affected originally long ā which in certain instances (as in the words כְּתָב, קָרְבָּן) did not become ō, and so was later elided in positions where other a vowels would normally be elided; therefore, one can find כִּתְבֵי, קָרְבְּנֵיהֶם, etc. Beginning in mishnaic Hebrew, many words entered Hebrew from Aramaic with what was originally a long ā, and these maintained the ā in their declension; in this manner the relative symmetry of biblical Hebrew, which maintained ā only in few morphological types, was disturbed. Another factor disturbed the vowel šewa relationship: the existence of vowels (originally short) in syllables that became open after the cancellation of gemination (see section on Gemination and Clusters above) in all the declensions of the word. So, for example, פַּרְרָשׁ>) פָּרָשׁ) meaning "horseman," caused פָּרָשׁ meaning "horse" (ii Sam. 1:6, Ez. 27:14) to retain the a vowel after the פ, and not elide it in the manner that the a vowel was elided after the ג of גָּמָל. The verb (without suffixes) in Hebrew is less given to change than the noun and to a greater extent reflects the early relationships (and not only in this regard). The noun, however, is subject to the influence of analogy; there are thus, from the phonologic point of view, many more contradictions in the paradigm of the noun. The great confusion with regard to interchanges made it necessary for the Academy for the Hebrew Language to establish the rules for maintaining or eliding a vowel. The Academy based these rules on morphological and semantic principles, i.e., principles which are at variance with precise phonetic processes.
In addition to the changes already discussed, there are changes in Hebrew caused by the chance sequence of sounds in the word. The natural tendency of the speaker is to conserve effort in his speech, and to try to minimize sharp changes in the use of one or the other of the organs of speech. In fact, speech is full of the assimilation of one sound to the other. Only when this assimilation is particularly sharp is the change felt. Since the consonants are more stable than the vowels, they tend to change less; cf. the changes due to assimilation, e.g., the נ to the neighboring consonant, the exchange of ת with ט or ד when close to צ or ז (see section 7. Other Consonants, above). At other times, a sequence of similar sounds demands a greater effort from the speaker and he tends to dissimilate them, as: displacement of ו by י in the plural form עֲרָיוֹת (from עֶרְוָה, but קְצָווֹת!) or י by א in עַרְבִיאִים alongside עַרְבִיִּים (but there is no יְהוּדִיאִים!) or הֲגָאִים alongside הֲגָיִים. Many variants in Hebrew comparable to הַשְׁקָאָה/הַשְׁקָיָה are occasionally used to distinguish differences of meaning, as הַלְוָאָה/הַלְוָיָה, הוֹרָאָה/הוֹרָיָה. In this class of variants is the plural ending יוֹת/ָ–אוֹת–ָ found commonly in loan words from the period of mishnaic Hebrew. Another type of change, called metathesis, is found in Hebrew in words like שַׂלְמָה/שִׂמְלָה, כֶּשֶׂב/כֶּבֶשׁ. More common are the vowel changes due to environment. This causes the šewa/ḥatef change after ח, ה, א and ע; a is preferred over é in שִׂמַּח (opposed to שִׁמֵּר) and over i in יַעֲמׁד opposed to אֶשְׁמׁר, יִשְׁמׁר against יִשְׁמׁר (but אִשָּׁמֵר is found alongside אֶשָּׁמֵר). Medieval manuscripts contain many more changes than are common in the official language, and some of the common forms in modern Hebrew can be explained as a result of this practice. So the common plural of מֵסֵב is מְסִבִּים, but the plural מְסֻבִּין) מְסֻבִּים) is simply a variant of מְסִבִּים, which assumed one of the meanings of the word. Among the changes whose origin is the desire to dissimilate the following are noteworthy:
1) A change which is active to a certain extent and noticeable in modern Hebrew: – u or o in a syllable next to a syllable with u or o is interchanged with i or é. This is found not so much in the inflection of words as in the derived forms. In inflection: נִכְחוֹ from נׁכַח; in derived forms חִיצוֹן, תִּיכוֹן, רִאשׁוֹן from חוּץ, תּוֹךְ, רׁאשׁ, and חִלּוֹנִי instead of ‡חֻלּוֹנִי (from חׁל); in this way the i of שִׁלְטוֹן can be understood as opposed to שָׁלְטָן; similarly in all paʿul participles with the suffix וּת– the u tends to be changed to i, פְּעִילוּת; in a combination of words לוּלֵא from לוּלֹא.
2) The exchange of a by e is found in the definite article (also in the word מַה), given certain conditions in the word; this change is not properly maintained in the spoken language.
1. definition of the subject
2. the root and the stem
3. the "basic element"
4. the formative and the morpheme
5. parts of speech
7. principles of pattern analysis
8. determining the patterns
10. prefixes and compounds
11. the declension of the noun
15. what is to be classed as the verb
16. the conjugations
17. a consideration of earlier views on the relationship between the conjugations
18. the extent to which a predictable relationship exists between the conjugations
19. the inflection of the verb
20. changes in the base
21. the inflection of objective pronouns
22. inflections of weak verbs
23. paradigms of assimilated forms
24. paradigms of mute forms
There is considerable disagreement as to which linguistic features are to be included in the area of morphology. Generally, semiticists have commonly included the discussion of parts of speech and the changes which they undergo as a result of their declension, as well as word-formation, in discussions of morphology. This is usually from the "form" aspect alone, without entering into investigations of the uses of these forms in speech. This latter problem is included in the study of syntax. The exceptions to this rule are Gesenius' Hebrew Grammar (29th edition) which was written by Bergsträsser (incomplete) and the Mishnaic Grammar by M.Z. Segal, which include the functions of the forms in their discussion of morphology. Another problem is, what is to be included in the term "form": a part of a word which does not have an independent existence? A word? A combination of words with a specific meaning or a particular structure (com-pound)? Or a feature such as the word order in the sentence which in some languages is morphological? Those who take morphology in its simplest, most straightforward meaning, include the linguistic form and exclude meaning, and distinguish between Lexical Morphology and Syntactic Morphology on the one hand, and, parallel to these, between Lexical Semantics and Syntactic Semantics, on the other hand (cf. S. Ullman, Principles of Semantics (1957), 33ff.). The preceding statement has not discussed all the methods of systematization but has merely alluded to the wide gulf which separates the different systems. The following description includes in the term "form": a form which is not in itself an independent word; an independent word; and to a limited extent the unit which supersedes a single word, if it is a lexical unit. This discussion of the structure of a linguistic form also includes its functions in the expression.
A study of series of Hebrew words which are related semantically, such as:
a) נִשְׁמַר, שִׁמֵּר, שָׁמַר ;שִׁמּוּרִים, שְׁמוּרָה, מִשְׁמֶרֶת, מִשְׁמָר, שְׁמִירָה, שׁוֹמֵר, etc., and
b) הִתְפַּקֵּד, הִפְקִיד, פִּקֵּד, פָּקַד ;פִּקָּדוֹן, תַּפְקִיד, מִפְקָדָה, פָּקַד, פִּקּוּד, פְּקִידָה, פּוֹקֵד, etc., will immediately demonstrate that there are a number of consonants in each of the words, which contain the common semantic element (even if only in a general way) and a number of vowels or vowels plus consonants which serve to qualify the meaning which is common to the entire family, to the particular, specific meanings of the various words or forms. The group of consonants found in each word of the above examples (a) שמר (b) פקד is called the root while the rest is called the formative (see section Phonology: 4. Consonants as Pronounced by Various Communities above). In Hebrew as in the other Semitic languages the root is always made up of a group of consonants. This is not the case in English, for example (and other Indo-European languages), where the roots also include vowels, as: "cut," "boy," "love." Only in certain cases are there those who call the consonants common to a group of words the root, as: s-ng, in the words, song, sing, sung, while others will choose one of these words and refer to it as the root, the other forms which differ from it being called the derivatives of that root. Clearly, the Hebrew root is only the abstract basis of a family of words used in the language, and does not denote the origin from which these words are derived, as it is hard to assume any level of the language in which the speaker was able to pronounce consonants alone as words. However, the fact that it is an abstraction is not to say that it is a grammatical fiction and merely a technical tool for the analysis of linguistic forms; it is in fact a living reality, an integral part of the structure of the language, which every Hebrew speaker feels. The root is not simply a prehistoric residual or an inherited element but a reality which is continually being produced in the language, and to a certain extent modern Hebrew suffers from a hypertrophy of root production. The reality of the root in Hebrew is seen from the modern roots נטרל, אכלס, טלפן, דוח which are derived from the abbreviation דין וחשבון) דו״ח) through the verb דִּוַּח, and from the words "telephone," אוכלוס ("population"), and "neutral," by eliding the vowels. This is done even though the vowels in these foreign words are an intrinsic part of the word, without which they would not retain the meaning they have as a Hebrew root. In slang and children's talk new roots are formed more easily – whether from Hebrew words such as צברח in the word מְצֻבְרָח from the phrase מצב רוח, or from foreign words such as בלף from bluff. However, only a small number of these words enter more educated speech. This is adequate proof of the reality and vitality of this phenomenon.
The norm for a root is three consonants, and this is in any event the minimum needed for formation of a verb. Hebrew, however, still recognizes roots of one consonant (especially in particles and pronouns) such as ז (in the words פ, (וזּ, זוֹ, זֶה (in the words כּ, (פֹּה, פֶּה (in the word כֹּה); two letter roots, such as: חָם, אָב, צֵץ, בֵּן, יָד, but these are doubtless vestiges from an ancient period, when Hebrew had not yet separated from other Semitic languages. However, this is no longer a productive method for producing new roots in Hebrew. The standardization of the root to three consonants took place in the proto-Semitic period. Hebrew also contains a number of roots with four or more consonants. In the earlier stages of the language they are few, but today they have been greatly expanded, as for example תפקד from the noun תַּפְקִיד, which is in turn derived from פקד, or מספר from מִסְפָּר originally from ספר. The verb, however, needs a minimum of three consonants and whenever a verb is created, a root (at least of three consonants) is implied, even from a one-consonant root like זהי (in זִהוּי, זֵהֶה) from ז, or from a two-consonant root like אִחוּי, אִחָה) אחי) from דִּמּוּם, דֶּמֶם) דמם, אח) from דם. There is still no conclusive proof whether עו״י (see section 23. Paradigm of Mute Forms, below) verbs and ע״ע verbs are derived primarily from original two-consonant roots or from three-consonant roots one of whose consonants is elided under specific phonetic situations (cf. Biblical Hebrew). However, it is important to note that structurally in the historical period these verbs are integrated into the three consonantal root system and follow its rules, so that verbs of the form דָּן, קָם in the perfect, generate nouns and other forms like דַּיָּן, קִיוּם. Theoretically, in all words which can be analyzed to a root, all the consonants of the root are present. These consonants of the root usually appear in all the forms, but as a result of phonetic processes, some of which took place in the earliest stage of Hebrew, there are cases where one (and occasionally even two) of the root consonants was weakened and does not appear in all of the derived forms of the root. For example, as a result of the tendency of the nun to be assimilated to the following consonant (cf. section Phonology: 7. Other Consonants) one finds forms such as מַפָּלָה, יִפֹּל, from נפל, and as a result of the elimination of the diphthong (see section Phonology: 15. Diphthongs) we find in Hebrew יִבְנֶה from the root בני against בַּנַאי, בִּנְיָן and others. In systematization of the Hebrew forms, and in categorizing the words grammatically, one must consider this feature which affects the external forms of words without necessarily weakening its association with the other forms and words derived from the same root. In accordance with the structure, it is customary to divide the Hebrew roots into two main groups (stems) called Strong Verbs and Weak (Hollow) Verbs. The three consonants פעל are used as the symbol of the root and in accord with the place of elision the weak verbs are divided into the following stems: פ״א (that is, the weakening takes place in the first consonant of the root and is the consonant alef, and similarly) ל״י, ל״א, ע״י, ע״ו, פ״נ, פ״י, פ״ו (usually called ל״ה because of the spelling of the perfect form like קָנָה). A separate category is assigned to the roots with duplication or the ע״עwhich stand between the strong and weak verbs.
There are, however, a considerable number of Hebrew words – excluding verbs – in which morphological analysis does not yield a root in the form described above but a combination of consonants and vowels, whether (a) the word is an independent form (a "free" form, as in the Indo-European root discussed above) or (b) it does not appear as an independent form (a "bound" form) or (c) it is a loan word or (d) it is an old inherited part of the language. Thus a group of related words such as פִּנְקְסוֹן, פִּנְקְסָנוּת, פִּנְקְסָן, פִּנְקָס, does not yield a Hebrew root פנקס, just as the words טֵלֵפוֹנָאוּת, טֵלֵפוֹנַאי, טֵלֵפוֹן, although there is a Hebrew root טלפן, cannot be derived from that root since the formatives é-é-o, é–é-o-ay do not exist in Hebrew. In both these groups of words the "elements" פִּנְקָס and טֵלֵפוֹן, being independent words in the language, are not susceptible to further morphological analysis. The same is true for original Hebrew words such as אֶצְבְעוֹנִי and מְחִירוֹן, which semantically have no connection to צבע, and מחר, but retain their connection to מְחִיר, אֶצְבַּע, which are independent words. This phenomenon is especially noticeable in those cases where the "elements" do not serve, or, because of their makeup, are unable to serve, as independent words, such as מדיניות (the independent word is צורני, (מדינה (the independent word is ירושלמי, (צורה (the word is רשימון, (ירושלים (the word is רשימה and not תברואן, (רשים (the word is בַּקָשֹׁנֶת, (תברואה (the word is בקשה: familiar usage). A "basic element" of this type is parallel to a root insofar as the derived forms and their semantic content are directly related to it and not its root, even if it can be analyzed further into a root. For the concept "basic element" which we have introduced into the morphological analysis of Hebrew we use, in Hebrew, the term נטע (plant) which is found in a grammatical text of the Middle Ages.
Note: Attention should be paid to the difference between the concepts "basic element" as used here and "base" which is used by some scholars in morphological analysis. They refer to a specific form of a noun or verb which is itself a combination of a root plus pattern. It is the base to which other morphemes or suffixes are added, thus שָׁמַר is the base of שָׁמַרְתִּי etc. or כַּלִבּ is the base for כַּלְבּוֹן, כַּלְבָּם, כַּלִבּוֹ. The base is an historical genetic concept (cf. Brockelman, Grundriss, 1, 287; Bauer & Leander, 246). The term "basic element," however, refers to a structure parallel to the root and of the same level, in that it cannot be further analyzed without losing its semantic relation to the word which is based upon it.
In biblical Hebrew the formation basic element + formative was rare and found mainly in nouns with the ending ־ִי (denoting belonging in the widest sense; cf. section 9. Suffixes); in modern Hebrew this formation is much more common, and in fact it was already in wider use in tannaitic Hebrew, with the adaptation of many foreign words. At times, when there is no clear-cut morphological analysis as with רְשִׁימוֹן, or as with words such as אֻמְדָּן, סֻלְחָן, קָרְבָּן there is some doubt as to which type of formation it is. Words like קרבן can be analyzed qorb+an or QRB+.o..an (QoRBan). In such instances one must fall back on semantic analysis. If there is a semantic connection between the word and the root, and this formative is found in the Hebrew language and forms are built up in this way – then one is dealing with a root, but if there is no semantic connection with the root or such a morpheme is not known in the language (cf. above טֵלֵפוֹן) then this form is the result of a "basic element" plus a formative. So רַגְזָן is easily analyzed: a root plus the morpheme a.an., while מרגיזן (in children's language) must be analyzed as a basic element (מרגיז) plus the formative an, for if this were not the case the causative quality, which is not expressed in the root but in the hifʿil formative included in the "basic element" which is the hifʿil participle, would be lost.
The formative is that element – a phoneme or group of phonemes – with which a word is created, whether from a root or a "basic element." In Hebrew morphology it is possible to speak of two types of formatives: one which is combined with the root and in Hebrew linguistics is traditionally called the משקל (pattern); and one added to the "basic element" and is either a prefix or a suffix. The formative called משקל is always an infix, because it comes within the root, but can also be a prefix and infix at one and the same time as in: mišMaR, or an infix and a suffix as: KišRon; or a prefix, infix and suffix as with mišBeẓet. This formant is always discontinuous while the one which comes with the "basic element" is always continuous. The group of patterns which make up one verbal paradigm is called a conjugation (בנין) as: šaMarti, šaMaRta, šaMaR. (In the early days of Hebrew linguistics in the Middle Ages the terms משקל and בנין were used interchangeably.) In the common spoken language there is also a "minus formative" discernible, where a word is developed by removing a part of the word which serves the basis of the derivation (this is called a back-formation), as the elision of the ־ִי in the words אנטיפאטי, פסיכי: resulting in אנטיפאט, פסיך. In this way a differentiation is achieved between the description of the quality (adjective) and its subject (substantive). Literary language includes non-accentuation (the accentuation which has phonemic value in Hebrew, cf. Phonology: 14. The Accent) of one of the elements of a compound and so a difference is achieved between the preposition ע֫ל יד and על י֫ד when both words are given their full meaning. Formations such as מגדל אור>מגדלור, יכול אני>יכולני (in this last word the plural מִגְדָּלוֹרִים as against מִגְדְּלֵי אוֹר is proof for that formation) come into being by eliding part of the compound. In the section on "Basic Element" it was stated that the "basic element" could be a word in itself used in Hebrew or something different or less than a word. In the formation of "basic element" plus formative, attention should be paid to the third category, where the formative plays a special role. While Hebrew can absorb foreign words easily, Hebrew grammar does not absorb elements which are foreign to its structure. It has already been seen that it is impossible to form a Hebrew verbfrom a foreign word if a root is not abstracted from its consonants (cf. טלפן). Similarly, Hebrew has difficulty in absorbing words which are adjectives or adverbs without first giving them a Hebrew form. This is not the case with other nouns. Words like טנק, ריאליזם, אידיאליזם, אידיאליסט, כלור, בנק and many others were assimilated into Hebrew without any serious attempt to exchange them for Hebrew innovations. This is not the case with words such as banker, chloric, realistic, psychic, and clerical; if they appear in Hebrew whether in their English, French, German, or other form, they remain foreign. In order to derive Hebrew forms two methods are used: either the Hebrew formative is added to the foreign word like קלריקלי (clerical+i), ריאליסטי, etc., or the corresponding foreign formative is exchanged for a Hebrew one, like: אַי) בנקאי in place of -er), ־ִי) כלורי in place of -ic), טרגיקון (Tragik + er), היסטוריון (histor + ion), etc. In each instance the grammatical element which determines the category of the word in the original language is replaced by a Hebrew element. Were it not for this process the word could not be assimilated into Hebrew, and would certainly not be able to serve as the basis for other derived forms. A similar situation is the addition of the feminine ending ־ָה to words which are borrowed from a language in which they are feminine even though they do not have a special feminine ending or are not used as the feminine at all. The Hebrew feminine form lends to the borrowed word a Hebrew form which makes its declension simpler, as אוניברסיטה (plural פונימה, (אוניברסיטאות (phoneme). A special function of the Hebrew formative is, therefore, to adapt foreign words to a Hebrew form. This type of formation has not yet been thoroughly investigated, nor has it been described.
Unlike the roots and "basic elements" which develop in modern times, the formatives (not only the pattern formatives) are mostly inherited from earlier times and fixed. They change as to their function and semantic value, which at times differ in modern Hebrew from what they were in biblical or earlier Hebrew. Still, it cannot be said absolutely that no new formatives are being created in Hebrew. The history of different languages shows that formatives generally originate from independent words whose meaning has become blurred as a result of the wide use of a particular compound, or by transferring an element from a word which already exists in the language (the so-called metanalysis). Thus in post-biblical and modern Hebrew the compound היה + participle is used to express continuous (durative) and repeated (iterative) action, and so there is a difference between הוא היה אומר and הוא אמר.
In biblical Hebrew היה, even when found in a similar syntactical frame, is not the formative element (auxiliary word) it is in modern Hebrew. Still, even in modern Hebrew היה has not become fixed as a formative only in the strict sense. However, in literary Arabic, for example, of the compound sawfa yaḫruǧu (סופו לצאת in rabbinic Hebrew, "He will go out") only sa remains from the first word (sayahḫruǧu) and this becomes the formative for expressing the future. Spoken Arabic knows of other such formatives. An investigation of modern Hebrew is likely to reveal several other candidates for new formatives, one of which is ש with the imperfect verbal form to express desire: שילך = (נא) ילך. In the literary language this formative developed from a certain syntactic combination. By analogy a new formative was created in modern Hebrew (probably as a result of a sarcastic expression), טרון– from the word תיאטרון, and it serves to refer to a place where performances are presented, as: זירהטרון (circus), צ׳יזבאטרון (a satirical theater), and בובהטרון (puppet theater). In literary Hebrew נֶת– is used as a diminutive form (as טִפֹּנֶת, קְטַנְטֹנֶת). This is a new formative. Some centuries ago וֹנִית– was used (cf. צַלְמוֹנִית, נַצֲרוֹנִית) coined in accordance with אֲדוֹנִית (fem. counterpart of אדון) or חִיצוֹנִית (fem. of חִיצוֹן). The attempts to reestablish the šafel (שקם, שכפל, שכתב, שכנע), which in ancient Hebrew had no position and certainly no fixed function (words of this form are always borrowed), are all parts of this process. To summarize: not even the area of the formative is completely closed to new addition or limited to its original complement. The root, "basic element," and formative are the three components to which the Hebrew word can be analyzed, and each, individually, is the minimal meaningful morphologic unit which cannot be divided further and which has semantic content, in other words: morphemes.
Note: Some modern linguists use the term morpheme as the basis of their morphologic analysis of the Hebrew language and do not see any need for the term formative. But even they cannot ignore the traditional concept, the root, completely (nor do they eliminate the concept of pattern); instead they speak of a root morpheme. But it seems that in a language such as Hebrew, where the root is a vital and living element (see section 2. The Root and the Stem), one must relate to the fundamental difference between the abstract "root" and the "formative," which is the element that generates a real word, a noteworthy stage in any morphologic analysis, even though both elements are similar in that they are minimal units – morphemes. The morpheme is a concept in morphology which includes inflection and derivation, while the formative is reserved for the process of derivation. Inflection is generally an automatic process, depending on the type of word (noun, verb, etc.); derivation is always a new process.
The elements discussed above combine to make words. It is common to sort the words into categories called "parts of speech." Traditionally, Hebrew grammar differentiates (as did Aristotle) between three types only: noun, verb, participle. However, for several generations, under the influence of the grammar of various European languages, the division into nine parts of speech has become part of Hebrew grammar. There is no area of modern linguistics where the differences between scholars are more pronounced than in the division into parts of speech. It has been correctly claimed that the criterion for the accepted division is not consistent; at times it is the form and at times the content, or a mixture of the two. The logical demand to categorize the words based on differences in form leads to the conclusion: "every language has its own scheme. Everything depends on the formal demarcations which it recognizes" (E. Sapir, Language, 1949, 119). It must be admitted that it is not easy to fulfill this prerequisite. In our opinion it is better to analyze Hebrew in accordance with the traditional division into three parts of speech since it is thus possible to include the formal criterion more precisely. A sharp distinction exists between the noun and the verb. The verbal nouns, the participle, and the infinitive belong morphologically to the category of nouns, although syntactically there are features common to them and the verb. Regarding particles, there is not always a sharp distinction between them and the noun; some are inflected like the noun (עִמְךָ like יָדְךָ) or have other qualities which are like the noun, while only the conjunctions and the interjections are entirely different from the noun and the verb. But there are particles (prepositions) which can, in accord with their morphological behavior, be classified as "nouns" (תַּחַת, אחוֹר־אחוֹרֵי, אֵין־אַיִן, בֵּין). Also from a syntactical point of view the only clear division is between the noun and the verb (a "verbal sentence" has a verb as a predicate; a "nominal sentence" has a noun or particle as a predicate). The division into nine parts of speech confuses, since it confounds meaning (substantive, adjective, number) with the criterion of form and does not necessarily follow from an analysis of Hebrew speech.
To the problem of how many patterns there are in Hebrew and what they are, there is apparently a simple answer; if the word is analyzed to its root then the pattern is left after eliding the root consonants. But, surprisingly enough, there is a great divergence between what is commonly presented in the grammars and scientific literature (particularly that not written in Hebrew during the last generations) and the practical grammars (especially those written in Hebrew) based on the long internal Jewish tradition. Suffice it to point out that a standard work such as Bauer and Leander's Hebrew Grammar lists about 80 patterns for the noun while the "traditional" count (since David *Kimḥi) is about 290. It is not in the nature of the language observed that the difference between these two systems lies, since few new patterns have been added to biblical Hebrew. The critical difference is the method of observation. The traditional method depends on a descriptive approach, in which each form is considered to be another pattern, while the accepted scientific system is based on a historical-diachronic approach, in which are classified together all nouns even if they appear in different forms if they were the same in the early (sometimes even prehistoric) stages of the language; that is, the criteria come from outside the linguistic stage being described. Following are some examples of different types of classification.
a) The nouns: גַּן, גְּדִי, עִיר, לֵיל, לַיִל, תַּיִשׁ, נַחַל, מֶלֶךְ
|"Traditional" system||"Scientific" system|
|these are all part of the qatl pattern because their original forms were: ‡malk, ‡naḥl, ‡tayš, ‡layl, ‡ʿayr, ‡gady, ‡gann|
b) The nouns: קְרָב, כְּתָב, כָּרוֹז, עָשׁוֹק, גָּב(וֹ)הַּ, עָבֹת, אָדֹם, גָּדוֹל, רָחוֹק, כָּבוֹד, שָׁלוֹם are classified:
|"Traditional" system||"Scientific" system|
|1. פָּעוֹל pattern: All nouns till כתב(but some point to the differences in declension) a) elimination of the qameṣ שלום, etc., b) maintaining the qames כרוזetc., c) gemination of the 3rd radical אדֹם.|
|2. פְּעָל pat: קרב, כתב.|
c) The nouns: אַיָּל, רַכָּב, קָשּׁת, קַנָּא, צִפֹּר, גִּבּוֹר, שִׁכּוֹר, קַנּוֹא, רַתּוֹק are classified:
|"Traditional" system||"Scientific" system|
|(Another less acceptable division in the scientific system:|
|qattāl: קַנוֹא, רתוק||qattal: צפור, שכור|
|qattal: איל, קנא).|
A further difference: the nouns with the addition ־ָה as: אַיָּלָה, נֶחָמָה, גַּנָּה are independent patterns in the "traditional" system, qatl, qattāl, qattal in the "scientific" grammar. The examples show that none of the methods is entirely consistent. Traditional grammar does not distinguish between מֶלֶךְ with two segols and נַחַל with two pathahs and מֶלַח with a pathah and a segol, while historical grammar is at times confused about the original form, and forced to establish patterns such as קְטִיל, קְטוּל (with šewa!). The criteria upon which we should base the different patterns in Hebrew from a descriptive structural point of view, will be given below with examples.
Note: In confronting the "traditional" and "scientific" methods of classification two typical forms of classification have been used, but it should be remembered that in each there are differences (especially in the "traditional") which affect the final count of patterns. While David Kimḥi (12th cent.) counted 290 patterns, Jonah ibn Janāḥ (11th cent.) counted about 80 (if one eliminates about 60 which are patterns of personal names). He already classified diverse forms such as נֵרְדְּ, פֶּתִי, עֲיָרִים, עִיר, אֶרֶץ under the same pattern.
It is clear from the above that the "scientific" methods describe the way the Hebrew word was created from its proto-semitic form and are essentially interested in prehistory and not the historical reality. On the other hand, the "traditional" system is found to describe the external appearance of the noun, even if that appearance is unique among the forms which make up the paradigm of that noun, and due entirely to chance as a result of the coincidence of certain sounds in the word. For example, בַּיִת, נַחַל, מֶלֶךְ have two syllables in these forms only, while in other forms of the declension of the singular, the base is of one syllable: בֵּית־, נַחְלְ־, מַלְכְּ־ as in the nouns עם, ליל, יום. The analysis of patterns must be done on two levels: first the nouns must be analyzed as they appear: that is the root and the formative element must be distinguished; then the common features in appearance must be investigated in relation to the structure of the root ("stem") and each group will yield its pattern. Just as every group of words with a basic common meaning will yield a "root," so too from a group having common formation features the "pattern" will emerge. On both levels of the analysis the process is only descriptive and refers to the language in the given circumstances. It may be said that the relationship of the "appearance" to the "pattern" is as the phone to the phoneme or the morph to the morpheme. In the process of the analysis the following rules will be carefully considered:
1) The need to distinguish between nouns derived from roots by patterns and nouns derived from a basic element +formative. For example: לַהֲטוּטָן, כַּרְטִיסָן, though there is a root כרטס (note: כַּרְטֶסֶת) להט, are to be analyzed ־ָן + להטוט, ־ָן + כרטיס as ־ָר + סמרטוט = סמרטוטר (see section 3. The Basic Element). These are not of concern here. One must be especially wary of nouns ending in the feminine as: אילה : יַבָּשָׁה, אַיָּלָה is to be analyzed יבשה, ־ָה + אַיָּל however, is to be analyzed יבש+ a..a-a.
2) Pattern formations are (primarily) the result of the relationship between the consonants of the root and the vowels and consonants which are not part of the root. The relationships between the root consonants themselves, such as the hollowness of certain roots (see the section: 2. The Root and the Stem) or the repetition of one or two of the consonants, do not affect the concept of pattern, nor the declensions which are connected to the structure. Pattern is an abstraction and the appearance of the word is its realization. This principle is not properly reflected even in scientific grammar books where the patterns qattīl and qatlīl, for example, are separated as are others, although from the point of view of the number of root consonants there is no difference between them (see Phonology: 8. Gemination and Clusters) nor is there a difference in the way they are declined. Repetition of a root consonant can have an expressive function, but is not a matter of the pattern. From this point of view, רככבת =) רַכֶּבֶת, כַּרְטֶסֶת, טַפְטֶפֶת, דַּפְדֶּפֶת) are of the same form, as are סדדור =) סִדּוּר, (סבבוב =) סִבּוּב, נִדְנוּד), and אללם =) אִלֵם, פִּלְפֵּל) and פַּרְנָס, (דיין =) דַּיָּן.
3) As a result of vowel interchanges (cf. Phonology: 15. The Diphtongs; 16. Interchange and Elision of Vowels) a differentiatioin ferentiation must be made between those changes which affect the meaning of a word and those conditioned changes which do not. For example the vowels which follow the formative מ in the nouns מַסְמֵר :מִזְבֵּחַ, מַצֲצוֹר :מִזְמוֹר do not indicate a pattern change (although they produce different appearances!) whereas the three respective vowels after the צ in מַעֲצֹר מַעֲצֵר מַעֲצָר determine three different patterns. The same is true for the vowels of the formative in מֻפְקָד :מִפְקָד or in מָעֳמָד :מַעֲמָד which determine different patterns, while the change מֻעֲמָד :מָעֳמָד does not determine different patterns.
4) As to the "appearance" one must consider its connection to the root, and in all the declensions of the noun in a specific paradigm, a form may be picked out and established as representative of the pattern, as long as the form chosen serves to clarify the others in accord with the rules of the language. The linguist is liable to discover that in Hebrew very often the declined form and not the dictionary form ("casus rectus" or absolute form) is the one which is most representative of the pattern. This is a result of certain developments in the language which caused change in the absolute form of the word. So from qullo, qullot, the form qol (lightness, easiness) could easily be understood according to the rules of vowel interchange (see section Phonology: 16. Interchange and Elision of Vowels) but not the opposite (the absolute qull is not possible!); there is, indeed, the homonym qol (voice) in the absolute which declines as qolo, qolot. Clearly one appearance (as qol) is liable to produce more than one pattern and vice versa. At times this method of analysis is likely to agree (although without intention) with the historical method, but very often it will yield different results. If the decision that qull is of the qutl pattern is in agreement with the historical position, the decision that qol is not of the qatl pattern (qawl in Arabic) or the qal pattern (qāla in Aramaic, and so in the Silwan inscription) but of the qol pattern, is opposed to the accepted historical point of view. Any agreement with the historical position is indicative of the fact that here and there the early state is still reflected in the modern makeup of the language.
5) From the above (4) it is clear that the patterns as they are determined by the structure of modern Hebrew must be arrived at not from the vocalized forms but from the pronunciation which does not recognize quantitative differences, recognizing instead a total of five vowel phonemes (see section Phonology: 12. The Phonological Status of the Vowels).
In accord with the above principles the various stages in determining the patterns of the noun can be described: Example i: (1) מֶלֶךְ, (2) צֶדֶק, (3) סֶגֶל, (4) סֵפֶר, (5) אֶבֶן, (6) עֵדֶר, (7) יֶרַח, (8) נֶצַח, (9) צַד, (10) גַּן, (11) עַיִר, (12) גָּוֶן, (13) בְּכִי, (14) חֲצִי, (15) חֵטְא, (16) דֶּשֶׁא, (17) בֶּכֶה, (18) עֵד, (19) לֵיל, (20) שׁוֹר, (21) אוֹן. In almost all of them the three root-consonants are immediately recognizable, and in some of them when they are declined in singular or plural (חֲטָאִים, צְדָדִים, בִּכְיוֹ). In many there is an obvious connection with other nouns (מְאֻבָּן, מְלוּכָה, צַדִּיק, etc.), or to a verb (בָּכָה, גִּוֵּן) – indicating that these forms are derived from roots and not from a basic element. They can be divided, based on appearance, as follows:
פֶּעֶל: 1, 2, 3, 5, 7, 8, 16, 17
פֵּעֶל: 4, 6
פַּל: 9, 10
פֵּל: 15, 18, 19
פָּוֵל: (or פָּעֶל): 12
פּוֹל: 20, 21
פְּעִי: 13, 14
In all there are at least eight different appearances (some differentiate between פֵּל and לֵיל] פֵּיל]]). Checking the structure of the roots (= "stems") we find five types: (a) strong (1–8), (b) ע״ו/ע״י (11, 12, 18, 19, 20, 21), (c) ל״א (15, 16), (d) ל״י (13, 14, 17), (e) ע״ע (9, 10). Since the types of consonants in the roots do not determine the pattern (cf. principle 2) it is fundamentally possible for all the words to be variations of one pattern, if it can be shown that with regard to the vocalizations attached to the root (the formative) there is no difference between them. In accordance with the declension of the following (a) אַבְנֵיכֶם, גּנּוֹ, יַרְחִי, מַלְכִּי, etc. (b) צִדּוֹ, בִּכְיוֹ, עֶדְרוֹ, סִפְרוֹ, צִדְקוֹ the diachronic approach is likely to distinguish two patterns qitl and qatl, and, based generally on comparative reasoning, these nouns will be included in one of the two patterns, especially recognizable in the ע״ו/ע״י stems since some of the nouns have a diphthong and others a simple vowel (cf. section on Phonology: 15. The Diphtongs). On the other hand, the descriptive grammar of Hebrew in its historical setting will abstract from these forms three patterns:
a) pi/aʾl 1–17, (20?)
b) pel 18, 19
c) pol 21, (20?)
Since for many generations (a fact which is already evident in the vocalization of the Bible) the form פעל is declined with an i after the first root-consonant, while a appears onlyif the first or second root-consonant is אהח״ע, and in a number of ancient words (as מַלְכּוֹ), therefore the interchange a/i is conditioned and this pattern may be called paʿl or piʿl; both have been absorbed in historical Hebrew to one pattern. On the other hand the ancient group which included 11, 18, 19, has been broken up and there is no longer any similarity in their behavior; this leads to the need for a pel pattern; the same is true for 12, and 20 is more properly placed in (a) because of its plural שְׁוָרים.
Example ii: (1) מִשְׁמָר, (2) מַעֲמָד, (3) מַחְמָד, (4) מַסָּע, (5) מַדָּע, (6) מֵידָע, (7) מוֹדָע, (8) מָקוֹם, (9) מָדוֹן, (10) מִמְצָא, (11) מִבְנֶה, (12) מַעֲשֶׂה, (13) מָסָךְ, (14) מֵסַב, (15) מַשָּׁק.
They are commonly divided into:
מפְעָל: 1, 2, 3, 10
מִפְעֶה: 11, 12
מַפָּל: 4, 5, 15
מָפוֹל: 8, 9
An investigation of the structures yields seven types: (a) strong 1–3; (b) פ״נ 4, 15 (מַשָּׁק meaning touch, is related to the root נשק); (c) פ״ו/פ״י 5–7 (there is no root נדע to which מַדָּע may be related); (d) ע״ו/ע״י 8–9; (e) ל״א 10; (f) ל״י 11–12; 13 ע״ע (compare the verb סכך), 14, 15 (מַשָּׁק meaning noise is related to the root שקק). The diachronic approach will establish for these 15 nouns of eight appearances two patterns maqtal and miqtal. The last is indicated only in 14, just as the first is derived from the rest except for 1–3, 10–12; these cannot be historically included with certainty in either of the two patterns. But in accordance with principles no. 2 and 3 above one pattern, maf ʿal, can be determined whose first vowel is easily distinguished in all the nouns except for 14 and for that reason a better symbol would be ma/if ʿal.
It must be noted that there is no permanence in the language; relationships to roots are constantly being eliminated and new relationships develop. A word which cannot be analyzed into a root and pattern is a basic element. The two words שִׁיר, אִישׁ are similar in their appearance, the second is related to the root שָׁר, מְשׁוֹרֵר, שִׁירָה) שיר) and so is analyzed according to the pattern pi/aʿl while the first was not related until modern times to any family of words, and when the root איש was created (מְאֻיָּשׁ, אִיֵּשׁ) its morphological status was changed. Consistency demands that we analyze words with fewer than three root-consonants such as פֶּה, שָׂפָה, יָד (which are actually relics of a period which preceded the three-root system!) as "basic elements," until there will be a family of words which will relate them to a root. In this way the 290 appearances in the "traditional" grammar will be reduced to about 90. About 50 of these are infix types, and the rest are prefix-infix-suffix types. The consonantal elements which precede the root are (in alphabetical order) תַּלְמוּד) ת, (נַפְתּוּלִים) נ, (מוֹסָד, מַלְבּוּשׁ, מַאֲכָל) מ, (יַלְקוּט) י, (הֶבְדֵּל) ה, (אַזְכָּרָה) א). The elements of the pattern which follow the root are: ה– אֻמְדָּן) ־ָן, (עִקָרוֹן) וֹן, (מִשְׁקֹלֶת, מִסְגֶּרֶת) ־ֶת, (קַבָּלָה, פְּשָׁרָה) ־ָה). From this great number of possible patterns in Hebrew there are today no more than 25 which are productive (this is a general impression and not the result of an exact statistical investigation). If one ignores the patterns used for participles and infinitives (nomina actionis), which are automatically formed with the verb, then the infixed pi/aʿl is especially productive for concrete and abstract nouns, such as: מֶנַע, קֶצֶר, פֶּסֶל, מֶרֶק, כֶּבֶל; po/uʿl mainly for abstract nouns: אֹדֶם, קֹצֶר, אֹרֶךְ but also, טֹפֶס, עֹתֶק; while paʿal is used to indicate profession or permanent character: נַגָּח, שַׁדָּר, נַגָּן, סַפָּן. The prefixed and suffixed patterns which are productive are the ones with מ before the root and a "feminine" ending such as מִכְפֶּלֶת, מִקְטֶרֶת, מִנְהָלָה, מִסְפָּרָה, מַזְמֵרָה, etc.
Although the creation of forms through patterns is dominant in Hebrew (verb patterns should be added to noun patterns), ancient Hebrew once created forms from the basic element + suffix. The changeover from one type of word formation to the other can be illustrated by two examples. (1) basic element + formative > pattern. The nouns רְעָבוֹן, רָעָב are differentiated by the suffix ־וֹן and it is likely that this was added to the basic element רָעָב, just as ־וֹן was added to ראש (there is no such root) and ראשון was formed. However, in various nouns which are formed in this way, a syllable + an open vowel which in antiquity must have become lengthened, are changed to a closed syllable by gemination of the next consonant. According to the accepted rule: לְבָנִים :לָבָן is equivalent to קְטַנִּים :קָטָן and which was qatal+on could develop into qattalon leading to פִּעָלוֹן. In historical Hebrew while רעבון can still be analyzed as basic element רָעָב + suffix צִמָּאוֹן, ־וֹן is a piʿʿalon pattern, and cannot be analyzed as ־וֹן + צָמָא. The same word in the Samaritan pronunciation, however, samamʾon enables us to see the older stage and can be analyzed as ־וֹן + צָמָא parallel to רעבון.
(2) Noun pattern > basic element + formative. The words יַהֲדוּת, גַּדְלוּת, עַמְקוּת, גַּבְהוּת, מַלְכוּת are each made up of two morphemes which are obtained through the analysis: roots עמק, מלכ etc. and the formative. a..ut, and cannot be analyzed into the stems מַלְכּ (note מַלְכֿוּת and not עָמְק, (!מַלְכּות or יהודי, עמוק and the suffix וּת, but in אֱנוֹשׁוּת the analysis is necessarily וּת + אֱנוֹשׁ. At times both types of creation function alongside each other as in the synonyms אדנ) אַדְנוּת + the pattern. a..ut), אֲדוֹנוּת (the basic element אֲדוֹן + the suffix ut). Though it is possible that the pattern containing וּת was formed from a basic element and a formative as in example (1) (since the element וּת was probably created as a result of metanalysis from ל״ו), historically one first finds in Hebrew words with וּת which are analyzed by patterns and only later those which are analyzed "basic element" + formative. Note that this type of formative is common in names of people and places.
The suffixes can be divided into two groups: those which have fallen into disuse (obsolete) and those still used. The suffixes which are no longer in use and are not productive, are at times not recognized even by the expert as formative elements. Still, they should not be ignored, and should be included in a descriptive grammar, since there is still a relationship between a noun with such a suffix and a noun without it. Furthermore, the availability of the early classical sources creates new forms through formatives which have been considered unproductive and long dead (cf. (נִּי(ת = below). Of the suffixes in common use, some are very productive.
a) vowels: וֹ, –ֹה–, are found mainly in personal names, as מגדו, עדו, שלמה, but also בן =) בנו, (חית =) חַיתו; perhaps the ו of בזמנו, בשעתו which was originally the third person suffix whose value was weakened because of the weakening of the syntactic connection, also belongs to this category). אִשֶׁה :̵ֵה, ̵ֶה (fire-offering), לִבְנֶה (in this way the singular of בָּטְנֶה :בָּטְנִים was formed several decades ago. Today this form is considered incorrect.), אריה (from ארי cf. אֲריוֹת). ה–ָ ≤ : so-called "he locale," found originally and mainly in adverbs denoting places as in לְמַטָּה, הָלְאָה, שָׁמָּה. See also תָה–ָ below.
b) vowels and consonants: ב– in the place name שעלבים which includes תֿעלב) שעלב in Arabic) = fox. ג–ַ : the name אבישג, in use even today as opposed to אבישי, the ג is considered to be a formative element regardless of its origin. כְּנוֹפְיָה, מֶרְחַבְיָה, מַאְפַּלְיָה :יָה; also to be analyzed in this way אַסְפַּקְלַרְיָה (from specularium), as there is the basic element עֲבַרְיָן :–יָן. אַסְפַּקְלָר the basic element עֲבַר, connected with the noun עֲבֵרָה (cf. the noun, ירושלמי, no 3). סֵפֶל, עֲרָפֶל, כַּרְמֶל :ל (the biblical word סף is found in poetry, as Bialik's עַרְסָל, (לבי סף דמעה (basic element פִּתְאוֹם, סָלְעָם, חִנָּם, אָמְנָם, רֵיקָם :–ם. גבעול, (עֶרֶס and in personal names like זְבוּלוּן, גַּרְזֶן, כְּנַעַן, יְשׁוּרוּן, צִפֹּרֶן :–ן. מִרְיָם (this analysis is arrived at from comparative grammar).
(אֲחוֹרַנִית, קְדוֹרַנִּית :–נִּית(ת. The new words: קְדוֹמַנִּי, קְדוֹמַנִּית, קְדוֹרַנִּי have been created analogically. ת–ַ : Primarily in personal names (including those recently created) such as רִנַּת, יוֹנַת, אָסְנַת, בָּשְׂמַת. The nouns מנה =) מְנָת) and שָׁנָה =) שְׁנָת)) can be so analyzed. This suffix is also found in place names: תה. שָׁמְרַת, דּוֹבְרַת, נָצְרַת–ָ≤ָ: Originally this is the so-called "helocale " suffix which is added to nouns with feminine endings, but when this function disappeared (as in the noun ליל = לַיְלָה) it became a suffix which is used particularly in poetic language ישועה =) ישועתָה, (עֶזְרָה =) עֶזְרָתָה) and others.
ה–ָ: Common feminine suffix in declension, and used to indicate: (1) collectives such as גּוֹלָה, "all those exiled," דָּגָה = "all the fish." As to the nouns שִׁקְמָה :שִׁקְמִים, חִטָּה :חִטִּים, etc. the plural always indicates the collective, while the form with the ה–ָ suffix indicates the collectives or one of the items in it, according to context; (2) an artificial as opposed to the natural limb בִּטְנָה, מִצְחָה.
י: also called the "relation suffix," since it relates the noun with this suffix to another by attributing to the new noun some quality of the noun serving as its "basic element." יְרוּשַׁלְמִי = of Jerusalem, רַגְלִי = on foot, יְמָנִי = on the right side, רָאשִׁי of the head (in the concrete and borrowed senses; see above, section 4. The Formative and the Morpheme).
b) Vowels and consonants:
א)י)–ַ: It may be that this was originally two suffixes which were consolidated. One, which is used in personal names such as זכריה >) זַכַּאי, (יוחנן >) יַנַּאי), etc., and one parallel in function to י–ִ (and used in Aramaic parallel to Hebrew י–ִ). Usually it indicates a professional such as חַקְלַאי, חַשְׁמַלַּאי, עִתּוֹנַאי. (This form should not be confused with the ל״י verbal form זַכַּאי, "innocent" derived from the root זכי in the פַּעָל pattern; the name זַכַּאי is זכ(ריה) + אי!)
וֹנִי) –וֹן–): Its modern use is preeminent (1) to create diminutives: גַּנּוֹן, דֻּבּוֹן, סִפְלוֹן; (2) to indicate publications which appear at regular intervals such as יַרְחוֹן, שְׁנָתוֹן, שְׁבוּעוֹן, עִתּוֹן and lists of similar items such as תַּקָּנוֹן, חִידוֹן, שִׁירוֹן, מִלּוֹן. But there are other nouns derived in this way and the formative fulfills other functions. The combination ונ + י– becomes וֹנִי– in words like צִמְחוֹנִי, צַהֲבוֹנִי.
וּת–: Used primarily for abstraction (see above in this section).
וֹת–: Common suffix for feminine plural used (1) adverbially ישירות, רַבּוֹת, and (2) for collective and abstract nouns such as: נסתרות, מפורסמות, מֻסְכָּמוֹת.
ית, יָּה–ִ: Combinations of the relation suffix י– and the feminine suffix, are used (1) for diminutive: עוּגִיָּה, מַצִּיָּה, יָדִית, כַּפִּית; (2) a workshop or gathering place כְּנֵסִיָּה, פְּנִימִיָּה, כְּרִיכִיָּה, סַנְדְלָרְיָּה; (3) a collection of things סִפְרִיָּה, תַּקְלִיטִיָּה, צִמְחִיָּה.
ין, ִ–ים–ִ: Originally the plural suffix, they are used (1) with adverbs (plus the preposition) such as לחלוטין, במישרין, לסירוגין; (2) for abstract nouns such as ניחומים, נעורים.
ן–ָ: Today mainly used to indicate the subject of an action as לַהֲטוּטָן, תַּכְסִיסָן, רַפְתָּן (see above section 3. "The Basic Element").
תן–ָ: Functions as ן–ָ and originates in nouns which end in ה–ָ as גאותן (conceited) and through metanalysis תָן as opposed to ן–ָ became more expressive: רַעַבְתָן a very hungry person, כרסתן = one having a large stomach.
תן–: Loaned from Latin-arius and is found in original Hebrew words such as עוּגָבָר, נַחְתּוֹמָר, סְמַרְטוּטָר serves the same function as ־ָן.
נֶת–: For diminution (see section 4. The Formative and the Morpheme).
The above survey indicates that the suffixes, like the noun patterns, are morphemes, every one of which has more than one semantic function, and at times these functions are quite dissimilar and it is difficult to find a logical connection between them (cf. for example וֹן–). The reason is that Hebrew is a very old language and in the course of time the formatives changed their functions, or new functions were added to them, without eliminating the words which were derived from them when their prime function was different.
Words formed by the addition of prefixes ("secondary derivatives") as are common in Indo-European languages such as the English print, offprint, reprint, imprint and come, income, outcome, overcome, become, are unknown in Hebrew, which expresses these different notions by different noun patterns or by compounding words, or by different roots, as in (1) מִטְּבָע, הדפסה, חוזרת, תדפיס, דפוס; (2) היות, התגברות, הוצאה, הכנסה, בוא or היעשות. Nevertheless, there are already compounds in biblical Hebrew which might be taken to be a prefix + a "basic element" when the prefix is a word of negation, as in: לא־עם, לא־אל (Deut. 32:21), אל־מות (Prov. 12:28): no-god, no-nation, no-death.
This is in fact the common way for analyzing compounds in modern Hebrew not only of the אי־צדק (injustice), אי־שימוש (disuse) type but also בין־לאומי (international), חד־צדדי (uni-lateral, one-sided), חד־שיח (monologue), דו־שיח (dialogue), על־אנושי (super-human), תת־עורי (hypodermic), אין־סופי (in-finite), קְדַם־מקצועי (prevocational), בתר־מקראי (post-biblical), חוץ־לשוני (extra-lingual), and others. It is true that these and similar compounds were developed under the influence of Indo-European equivalents and indeed a deliberate attempt was made to achieve Hebrew equivalents for idioms which are basically technical terms. Such compounds gradually became assimilated into the common language and generated compounds like חד־סטרי (one-way) and חד־פעמי (unique) which are not entirely parallel to the English forms (but compare the German einmalig = unique). The fact that such foreign words are analyzed as including prefixes, cannot dictate that the Hebrew analysis be done in the same way; that must be done in accordance with the manner in which these words are integrated into the Hebrew system. It will be demonstrated that there is really no difference between this form and the existing compounds in Hebrew. A compound of two or more words which become one indivisible word, so that at times the original elements are no longer recognizable, is almost unknown in classical Hebrew; of this form are the early words משהויין, (כל מה), כלום, בלימה, בליעל (plural of מה שהוא >משהו) and the modern words כדורגל, דחפור, רמזור, מִגְדָּלוֹר; in slang and affected speech compounds such as ארחי פרחי + טורה) ארחיפרחיטורה) "a student parade from the school of architecture," are formed.
There are, however, many combinations of more than oneword in Hebrew, which due to their wide use have become fixed formally with fixed semantic values; they can be called compounds. The compound is usually the necessary condition and the first step toward merging the separate elements into one word. Various types and levels of construction can be differentiated:
1) Where one is in construct state; two nouns are joined, the first ("nomen regens") is qualified by the second ("nomen rectum"). This is the reverse of the situation in Indo-European languages: for example מלאכת־יד = hand-work, the same order is common in a compound of a noun and its adjective מלאכה קשה = hard work. The opposite order is possible, as in English, if the first part is a noun of quantity or vague, and therefore of wider meaning than the second part, which limits scope of the first word, as in: שלש קלשון (I Sam. 13:21) – tri-dent, משנה כסף (Gen. 43:12; but Gen. 43:15 כסף משנה!) = double money, דרום אפריקה South Africa, צפון אמריקה North America (but קוטב הצפון = North Pole). There is no clear formal criterion to establish when a combination is a regular construct or when it is a fixed compound; the semantic content may helpbut it is not an absolute criterion. Still it may be pointed out that deviation from the normative grammatical rule which demands that the article be placed before the second word (nomen rectum) (בית הספר and not הבית ספר) does, to some degree, indicate that the combination is felt to be a fixed compound, and so we find התל אביבי, הבר מצוה (in the Bible המגן דוד הבעל קורא, (בית הלחמי. Indeed, this rule is obligatory with numerals: השבע עשרה (and never שבע העשרה) and is also the case for names of books such as השולחן ערוך, הבית יוסף. Another indication found in the older sources is the addition of the plural to the nomen rectum בית המבשלים (Ezek. 46:24), שלושה) בית פרסות) (Oholot 8:12), similarly the addition of the feminine ending to the nomen rectum as in Phoenician רבת הכהנים = רבכהנת. Still these possibilities have not become integrated into literary Hebrew, though they are found in slang as the plural of טוב מאד (the grade "Excellent") = טובמאודים. There are, however, some compounds, such as ברנש: plur. בר אוז>) ברוז, ברנשים): plur. ברוזים, fem. ברוזה, which reveal this formation principle. There is, therefore, no absolute formal differentiation between a fixed compound and the construct state of two nouns.
2) Where the form of the compound is two words joined by waw and at times even without it, as in: יום יומי) יום יום, דין וחשבון, משא ומתן). This last is the rule in the second decade (11–19) of the numerals as שבעה עשר (but שבעה ועשרים or עשרים ושבעה!). Regarding this type there is a syntactic test which indicates if it is a compound: if the adjective and the predicate are in the singular: המשא ו(ה)מתן נמשך, דין וחשבון מקיף. Already in the medieval Hebrew grammars we find עמם מסורה (segol =) שלש נקודות.
3) Another type is the compound of noun with an adjective as קרן־הקיימת, לבד(ה)גס, יום טוב, לשון (ה)רע. Here a possible formal test is the use of the article with the second word only.
In short, there is in Hebrew a basis for compound words becoming fixed lexical units, but the limits of the construction are not sharp. It is not difficult, therefore, to include in this category the formation which some see as second derivatives: such as בין לאומי. Since the two qualities which are commonly used to distinguish it from a compound, (a) the order in agreement with English and (b) the definite article being placed normally before the first term as הדו־תנועה (the "deviate" form due to hyper-correction דו־התנועה is also heard!), are not unique to this formation at all. It must also be noted that while in English most of the prefixes are not independent words the situation in Hebrew is the opposite. Only very few elements, foreign or loan words like אנטי, ארכי, do not function as independent words. The fact that most of the compounds have the relation suffix ־ִי, that is, they are used as adjectives, is a statistical fact, but is not grammatically meaningful, since forms which do not serve as adjectives such as תת־לשון (hypoglossus), תת־תזונה (undernutrition), are also found, though not as frequently.
In the discussion of compounds two types should not be excluded: (1) the derivation of words from commonly used abbreviations, עובד כוכבים ומזלות) עכו״ם) and thus לא היו דברים מעולם) להד״ם ;(סכין כף ומזלג =) סכו״ם ;עכומי) and thus להדמי (a word used in poetry); this form is especially common in military jargon, and mention need only be made that this washow the word סַמָּל (sergeant) was created (= סגן מחוץ למניין), and many names of weapons, as תותח ללא רתע =) תּוֹלָר); and (2) blending two elements taken from two different words and making one word out of them, as in מד + חניה=) מדחן (דחף + חפר) דחפור). Early examples are פלוני + אלמוני>) פלמני) and זוט + זעיר>) זוטר).
In nouns as in most pronouns and most of the verbal forms there are two genders, masculine and feminine, but only the feminine is normally marked. This mark in the singular establishes the gender whereas the feminine or masculine plural marks are not decisive. There are a number of nouns both for feminine forms of living things and inanimate objects which are feminine although they do not have the usual grammatical symbol and there are also nouns which are grammatically bisexual.
The feminine endings are
1) ־ָה as טובה (masc. מלכה, (טוב (masc. פקידה, (מלך (masc. שומרה, (פקיד (masc. יולדה, (שומר;
2) ־ַת, ־ֶת, as גבירְתְ>) גְּבֶרֶת but used as fem. of גֶּבֶר, not of שומרת, (גְּבִיר (masc. נוסעת, (שומר (masc. נוסע);
3) ־ִת as טַבָּחִית (masc. עתונאית, (טַבָּח (masc. עתונאי);
4) וּת–, the abstracting formative (see no. 9) which also implies the feminine mark;
5) It is common for the ending including ת to signify the feminine and at times ת as a root consonant is so taken by the speaker, thus פת, שבת (the root is פתת and the pattern is pi/aʿl) are feminine, and on the other hand the ending וּת-was not considered, especially in medieval Hebrew, a feminine mark.
All these suffixes are derived under different conditions from the primitive ending -at which still exist in the inflected forms. There are feminine nouns which do not have a feminine suffix such as שֵׁן, אָתוֹן, רָחֵל, עֵז, אֵם, אֶרֶץ, גֶּפֶן, בְּאֵר, אֶצְבַּע, עַיִן, יָד, and especially geographical names like לונדון, מצרים, בני־ברק, תל־אביב, ירושלים. There are even nouns which are used both as masculine and feminine like כּוֹס, לָשׁוֹן, שֶׁמֶשׁ, דֶּרֶךְ. In the course of time some nouns changed gender; an example is שדה which is masculine in biblical Hebrew and (commonly) today, but is feminine in tannaitic Hebrew. In some nouns which do not have a feminine ending there are even today ambivalences regarding the gender: the word שלד is feminine in literary language and masculine in the spoken, while the opposite is true regarding גֶּרֶב. Feminine nouns with no feminine symbol are a reflection of the division in the very early (prehistoric) period of Hebrew, when the criterion of sex was not the determining factor which was a different scale of values, probably one with many grades.
There are three numbers: singular, dual, and plural, but the dual is found only in nouns. Only the dual and plural are marked by special suffixes indicating their number. The suffixes are:
־ִין) ־ִים) for plural: generally used with nouns without the feminine suffix such as מלכים (sing. נופים, (מלך (sing. נוף) but it is sometimes also found in nouns with the feminine ending such as: שָׁנִים (sing. תְּאֵנִים, (שָׁנָה (sing. שְׂעוֹרִים, (תְּאֵנָה (sing. שְׂעוֹרָה). The suffix ־ִין, very common in talmudic and rabbinic literature, is also found elsewhere, but is uncommon in simple language. In that literature the suffix (ים (ין appearswith nouns that have a feminine ending such as שמיטין (sing. שמיטה) more so than in present-day literature.
־ָיוֹת) –אוֹת, ־ָוֹת) for plural: is used generally with names that have a feminine ending but is also found with masculine nouns (and so used is considered florid) and so there is not only מלָכות (sing. חידות, (מלכה (sing. חִידָה) but also קוֹלוֹת (sing. אָבוֹת, (קוֹל (sing. חובות, (אב, which only the context can indicate if it is the plural of חובה or of חוב. It should be noted that many nouns can be made plural in two ways, such as מדרשות, מדרשים. At times homonyms are differentiated by their plural forms as עֲצָמִים :עֶצֶם (substance), עצמוֹת (bones). The suffixes (־ָאוֹת (־ָיוֹת have been used since tannaitic times and are common in words dating from then. However, they are also used in new words (by analogy or to simplify the declension) and so we find not only the old words פרפראות (sing. מקואות, (פרפרת (sing. תיאטראות, (מקוה (sing, מרחצאות, (תיאטרון (sing. מרחץ) but also the new גִּמְלָאוֹת (sing. אוניברסיטאות, (גמלה (sing. פקולטאות, (אוניברסיטה (sing. פקולטה).
־ַיִם originally indicated duality, as in רַגְלַיִם, יָדַיִם, שְׂפָתַיִם, כְּנָפַיִם but those nouns which have this suffix maintain it even for the ordinary plural such as ארבע ידַיִם, שש כנָפַיִם. However, the function of this suffix to indicate the dual exclusively is retained in several nouns as we see from ארבע פעמִים/פעמַיִם, שלש שנים/שנתַיִם, and חמשה שבועות/שבועַיִם. Similarly there is a difference between גלגלַיִם (a tool) and שלשה גַּלְגַּלִּים, as there is between אוֹפַנַיִם (an apparatus) and אוֹפַנִּים, among others. This suffix is very productive in technical nomenclature. Basically the suffix ־ַיִם is added to the singular noun (in its inflected form) as שְׁנָתַיִם :שָׁנָה, רַגְלַיִם :רֶגֶל, but at times it is added to the plural form of the noun as in דּוֹרוֹתַיִם :דּוֹר.
In addition to the numbers mentioned above, there is in Hebrew a list of collective nouns and abstract nouns which appear with all of the suffixes mentioned but do not have a singular form (pluralia tantum), such as קִדּוּשִׁים, בְּתוּלִים, נְעוּרִים (the singular קדוש has a different meaning), גירושין (sing. גירוש has a different meaning), שָׁמַיִם, מַיִם, פיפיות, כְּלוּלוֹת, מְפֻרְסָמוֹת.
Every noun can appear in one of three states: absolute, construct, and with pronominal suffixes. Everything stated above regarding the gender and number refers to the absolute state. From the morphologic point of view the absolute includes nouns with a preposition or article (היום, למחר) since these only affect their syntax and not their form (their vocalism is not influenced). When the noun is in the construct or has pronominal suffixes its form usually changes; the feminine ending ־ָה changes to ־ַת and ־ַיִם, (ין)־ִים, to ־ֵי (the plural and dual are the same). Only a relatively small number of nouns do not change in declension (that is when in construct or with pronominal suffixes) while the majority change in accordance with the rules for vowel changes (See Phonology, 15. The Diphthongs and 16. The Interchange and Elision of Vowels). There is no fixed system in the grammar books to arrange the different ways of declining the nouns into a set number of classes as is the case for Greek or Latin. But at least 11 declensions can be identified and some scholars determine 14. A description of their qualities has no place in this review but belongs in a grammar.
Often the form of the noun in construct is the same as its form with pronominal suffixes as in דְּבַרְכֶם, דְּבַר־הנביא :דבר and שמחותינו, שמחות־הילדים :שמחות but at times it is the same as the absolute form as in כַּלְבִּי, כֶּלֶב־בית :כֶּלֶב and שׁוֹמַרְתֵּנוּ, שׁוֹמֶרֶת־לילה :שׁוֹמֶרֶת. The pronoun which is added to the noun expresses the concept of ownership – in its widest sense – relative to the subject of the pronoun. The form of the pronoun differs when in the singular, the plural, or the dual. This feature is not logical since the subject of the pronoun does not change if the noun is plural. It is the result of internal phonetic development and metanalysis (part of the noun in masculine plural is merged together with the pronoun), and is not paralleled in classical Arabic or in ancient Aramaic (in later Aramaic a similar situation developed); in biblical Hebrew there are still remnants of the early situation (אֲבוֹתָם = אֲבוֹתֵיהֶם). Following is the list of pronouns with examples of the declension of the noun קול.
|־ִי||my voice קוֹלִי||1st Person Sing.|
|־ְךָ||your voice קוֹלְךָ||2nd Person Sing. Masc.|
|־ֵךְ||your voice קוֹלֵךְ||2nd Person Sing. Fem.|
|־וֹ||his voice קוֹלוֹ||3rd Person Sing. Masc.|
|־ָהּ||her voice קוֹלָהּ||3rd Person Sing. Fem.|
|־ֵנוּ||our voice קוֹלֵנוּ||1st Person Plural|
|־כֶם||your voice קוֹלְכֶם||2nd Person Plural Masc.|
|־כֶן||your voice קוֹלְכֶן||2nd Person Plural Fem.|
|־ָם||their voice קוֹלָם||3rd Person Plural Masc.|
|־ָן||their voice קוֹלָן||3rd Person Plural Fem.|
|־ַי||my voices קוֹלוֹתַי||1st Person Sing.|
|־ֶיךָ||your voices קוֹלוֹתַיִךְ||2nd Person Sing. Masc.|
|־ַיִךְ||your voices קוֹלוֹתַיִךְ||2nd Person Sing. Fem.|
|־ָיו||his voices קוֹלוֹתָיו||3rd Person Sing. Masc.|
|־ֶהָ||her voices קוֹלוֹתֶיהָ||3rd Person Sing. Fem.|
|־ֶינוּ||our voices קוֹלוֹתֵינוּ||1st Person Plural|
|־ֵיכֶם||your voices קוֹלוֹתֵיכֶם||2nd Person Plural Masc.|
|־ֵיכֶן||your voices קוֹלוֹתֵיכֶן||2nd Person Plural Fem.|
|־ֵיהֶם||their voices קוֹלוֹתֵיהֶם||3rd Person Plural Masc.|
|־ֵיהֶן||their voices קוֹלוֹתֵיהֶן||3rd Person Plural Fem.|
These are the standard forms of the pronouns, but in poetry and in flowery style in general there are several variations found, especially in biblical Hebrew. The above pronouns are used also with the prepositions such as על, מן, את, עם; some with the pronominal suffixes usual for the singular nouns (as אִתּוֹ, עִמִּי) and some with those used for the plural noun (as תחתיו, עָלַי). It should be noted that in prepositions ־ָךְ as opposed to ־ֵךְ in nouns is found (עִמָּךְ, לָךְ) and ־ָנוּ in place of ־ֵנוּ; this latter being also used in the noun כל to form כֻּלָּנוּ.
Besides this, the synthetic, method for indicating possession, there exists an analytic method, by the use of the word של. Expressions such as קוֹלָם, קוֹלִי and הקול שלהם, הקול שלי, are of equal value. For this reason there are those who call של a "separated pronoun." Actually, however, this is a syntactic method, since של originates (already in the earliest stages of the language) from the relative clause in which the relative particle is attached to the preposition. Thus the expression הקול שלי is to be analyzed as "the voice which I have" exactly the same as הקול שצועק "the voice which is crying," and in certain instances the preposition ב fills the same function as the preposition ל, for example = הפחד שֶׁלִי = פַּחְדִיהַפַּחַד, שֶׁבִּי but the compound… של, since it is so common, became one word parallel to a pronoun and competed with the pronoun successfully in common speech and in those cases where there is a morphological difficulty in declining the noun (as in foreign words or words with the definite article, which cannot be joined to a pronoun) and personal names or where there is a semantic difficulty. Fundamentally, the preference for one way over the other is a matter of style and not of grammar.
Personal Pronouns. When the person is not expressed through pronouns which are attached to the noun (see section 11. Declension of Nouns) or the verbs (see section 20. Inflection of Objective Pronouns), they are independent words whose forms are as follows:
|אני, אנוכי||1st Person||אנחנו, אנו|
|אתה||2nd Person Masc.||אתם|
|את||2nd Person Fem.||אתן|
|הוא||3rd Person Masc.||הם, המּה|
|היא||3rd Person Fem.||הן, הנּה|
There are also several variations in form and usage in the ancient language which are no longer used. The first person singular אנוכי whose use declined and completely disappeared by the end of the biblical period, is now found in modern Hebrew not only in poetry but also in general use, for emphasis. The pronouns הֵמָּה and הֵנָּה are considered archaic forms, and are only used in poetic writing. These personal pronouns are used only as subjects in the sentence; the pronominal object is expressed by the pronominal element suffixed to the verb or the preposition, or to the word אות (absolute אֶת) as אוֹתָם, אותך. It should be noted that the double forms for first person in singular and plural, is a distinctive feature of Hebrew among the other Semitic languages, and only in Ugaritic is there a duplication in the first person singular pronoun.
These pronouns are used to indicate something before the speaker, whether close-by or far off (deictic use), or something which has already been mentioned in the discussion (anaphorically). Today the distinction between the near and far demonstrative is more precise then it was in the ancient language. For that which is near, זה (masc.), זו, זאת (fem.) = this (sing.) and הללו, אלו, אלה (poetry also ה) אל)) = these (pl.); for that which is far: ההוא (masc.), ההיא (fem.), הַלָזֶּה, הַלָּה (masc.), הַלָּז (masc. and fem.), הַלֵּזוּ (fem.) = that (sing.), and ההם (masc.) ההן (fem.) = those (pl.). אותן, אותם, אותה, אותו can be substituted for the above but also express intense identification, "that same." The fact that in the past there was no differentiation between near and far, is seen from expressions like זה לזה and מזה ומזה (not: מזה ומההוא, זה לההוא). In certain instances the demonstrative force of the definite article has been maintained, as in הלילה, היום and in literary style it is used more frequently not only in these fixed forms. The use of the definite article in a direct address as: !הדוד is basically demonstrative. The third person singular pronoun is used as a demonstrative in expressions such as הוא שאמרתי: "that is what I said" and היא שעמדה לאבותינו: "it is that which stood for our fathers." When compounded with the interrogative pronoun like מה זה, מי זה, or such words as עתה זה the demonstrative force of the pronoun is weakened and it becomes an element of the emphasis.
מי = "who," מה = "what," are not declinable, nor are אי) אלו, איזו, איזה) = "which." There is a variation ofמֶה :מַה the use of which is subject to the same rules as the definite article, but which is found unconditionally with the interrogative pronoun איזה. בַּמֶּה is actually the demonstrative, with the addition of the interrogative אי, found also in the words היכן =) איכן, איפה). In plural the forms אי אלה, אי אלו were shortened for phonetic reasons to (אילו(ה and became identical to the demonstrative. Due to ambiguity the singular forms איזה and איזו are also commonly used in the plural (איזה אנשים היו שם) and it should be noted that the demonstrative element זה which can come to strengthen the interrogative מי as stated above, appears in tannaitic literature also with איזה as: ?אי זו? איזו זה and this is positive evidence for the crystallization of the compound איזה (which appears in the Bible as two words, and in which זה is used to strengthen the interrogative אי), into the interrogative pronoun.
Except for the words אלמונים, אלמונית) אלמוני (פלונית) פלוני) ="somebody" and מאומה, כלום "nothing, something," there are no special words in Hebrew to express indefiniteness, and the interrogative words are used for that purpose as in מי לחיים, מי למות = "some for life – some for death," דבר־מה = "something" and especially in relative expressions such as עשה מי שעשה = "somebody did it." The words בן) אדם, איש) are also used to indicate the indefinite as is the pronoun אתה. (The indefiniteness of the subject is often indicated by use of the plural verb without a pronoun, as in: בעבר עשו את הדבר כך וכך היום עושים בדרך אחרת = "in the past they used to do it so; nowadays they do it differently.")
The relative in as far as it is expressed syndetically (cf. Syntax) is made up of אשר or ש plus the pronoun (generally the personal pronoun) but in certain syntactic situations the pronoun is not explicitly stated: האיש שיושב (= האיש שהוא יושב), "the man who is sitting," המקום שהייתי (= המקום שהייתי בו), "the place in which I was." In verbal sentences the relative is אשר, ש, while in nominal sentences the definite article has that function, as in האיש היושב= האיש אשר יושב (in biblical Hebrew the demonstrative is also used in this way (cf. Job 19:19) as is the definite article before the verb as in ההרימו (Ezra 8:25). A relative pronoun can be a subject or an object and come directly after a preposition as: את שראה "whom he saw," לשכמותו = "to one such as him," לשעבר = "in the past," על ידי שנתן = "by his having given," but generally the use of correlatives is preferred, as: על ידי העובדה שנתן, על ידי כך שנתן, לאיש כמותו, את מי שראה. Sophisticated style prefers to forego the use of the modern correlatives such as עובדה and כך.
Reflexive and Reciprocal Pronouns
Reflexivity can be expressed synthetically, by verbal conjugations or analytically; since the tannaitic period the analytic method is preferred over the synthetic. In this construction the possessive pronoun suffixed to the prepositions is used, עשיתי דבר זה בי ולא בו = "I did it to myself and not to him"; הוא ניסה את הנסיון עליו ולא על חברו = "he experimented on himself and not on his friend," or suffixed to certain nouns such as the limbs of the body, שאר(ושארב ומד = "his blood is on his head"; "it's his own fault"), דמו בראשו) ראש = "he looks after himself"), הוא גופו אמר לי) גוף= "he himself said to me") and others especially with the noun איבד עצמו לדעת :עצם ("he deliberately destroyed himself" – suicide), בא בעצמו ("he himself came"), etc. The noun איל, as in מאליהם, מאליו, is also used in this way. From this last noun Aramaic produced the word ממילא, which was borrowed in Hebrew, and is a word whose sole purpose is to express reflexivity.
In early Hebrew reciprocity was also expressed synthetically, as נדברו = "they spoke with each other," התראו= "they saw each other," שניהם מצטרפים = "they join each other," but in the post-biblical period this method was abandoned and reciprocity was expressed analytically by repeating the demonstrative, as זה עם זה, זה לזה, or by expressions like, איש… חברו, איש… רעהו, אישה… אחותה; e.g., דברו איש אל רעהו, דברו זה עם זה, etc.
This review of the various pronouns indicates that this part of speech in Hebrew has unique aspects not only from the formal point of view (one and two consonant roots; no differentiation of gender and number as in מי and איזה) but also in function, i.e., the lack of clear demarcation between demonstratives, relatives, interrogatives, and indefinitives. In other words even in modern Hebrew the early situation is clearly reflected; one pronoun can be used freely for all the above functions. The differentiation of function (which is not new) is the result of a long process and can be compared to the exchange, in Hebrew, of paratactic structures, common in ancient Hebrew, for hypotactic structures using well-defined conjunctions for different purposes.
In this category are to be classified all those words which are not nouns or verbs and whose common function is to indicate grammatical relationships. There are the following types:
Prepositions, which appear only with nouns, such as ב (in), ל (to), על (on), מן (from), על־פי (by).
Adverbs such as במאוד, מאוד (very), חנם (gratis), ישירות (directly), בעקיפין (indirectly), לא (not), אמנם (certainly). (The term adverb, however, does not accurately describe the situation in a language like Hebrew since this type of word can qualify a noun as well, גדול מאוד = "very big," גבּוֹר מאוד = "very strong," or זה אמנם שלחן = "this (is) surely a table.")
Conjunctions which join words and sentences such as גם, אף, ו and those which join only sentences like עד (ש), בשביל ש…, מפני ש…, כי, אם. The examples indicate that different layers are immediately recognizable. Primary words (i.e., whose origin is not obvious) such as: בוכ״ל which exist as proclitics, לא, אם; words whose origin is obvious, such as בעקיפין, (חן >) חנם, and words which are derived from sentences, such as כיצד, whose original form כאיזה צד is still found, יען (shortened imperfect form from יענה). Synchronically, two kinds can be recognized: (a) those which formally behave like nouns – having suffixed pronouns (על פיו, לו, בי) – and those which do not behave like nouns or verbs, and (b) those which are syntactically nouns and can serve as the predicate of a nominal sentence, such as דבר זה כיצד?, ראובן פה and those which are unable to be used in this way (conjunctions). The above examples indicate how easy it is to form prepositions in Hebrew, and in fact the development in this area is great (many examples being influenced by foreign constructions), as in: …הודות ל (older: בהתאם ל…, (בזכות, בגלל (older: למרות, (לפי (older: עם ש…, עם), which are considered less elegant than the older forms. Actually every noun can be used as an adverb, the criterion for the noun being not morphological but its syntactic use. In the sentence בשוגג ראובן עשה דבר זה ("Reuben did this thing unwittingly"), the word בשוגג can mean "as an unwitting person," but in שרה עשתה דבר זה בשוגג ("Sarah did this thing unwittingly"), it must be an adverb because there is no accord in gender between שוגג and שרה. So, too ראובן עשה ביודעים as opposed to עשה ביודע. The limited ability of the language to express adverbs as a special formal category is compensated by the syntactic devices mentioned as well as others.
The numeral in Hebrew is a unique phenomenon and extremely complicated, both morphologically and syntactically. In part there are parallels in the other Semitic languages, and it is presumably a common residue from proto-Semitic. The numerals are expressed in Hebrew (a) by words which indicate units עשר, (9) תשע, (8) שמונה, (7) שבע, (6) שש, (5) חמש, (4) אַרְבַּע, (3) שָׁלוֹשׁ, (2) שְׁנַיִם, (1) אֶחָד (10); (b) by the suffix ים– which is added to the numerals תשע־שלוש (30 = שלשים, 90 = תשעים) and עשר (20 = עשרים) and; (c) by the words 100 = מאה, 1000 = אלף, 10,000 = רבוא (archaic form and rarely used today). These basic words are compounded in different ways and from them are derived the various forms for particular use. There are different forms for the cardinal numerals, ordinal numerals, and fractions, but the ordinals and fractions exist only for the first ten, and must be expressed syntactically for the rest of the numerals.
There are two forms, the masculine: ארבע, שלש, שנים, אחד…, and the feminine: שתים, (אחד+ת) אחת
(< ארבעה, שלשה, (שִׁנְתַּיִם…, but it is exceedingly strange that in Hebrew, as in Semitic languages in general (for a discussion see Robert Hetzron, Journal of Semitic Studies, 12 (1967), 180ff.), there is no syntactic agreement between the cardinals (and any derived numerals) from 3–10 and their referents; the masculine numeral is used as an adjective or predicate for a feminine noun and the feminine numeral for a masculine noun; as in שלשה בנים (or בנות שלש) שלש בנות, (בנים שלושה). When the object counted is not referred to, the masculine or feminine numerals can be used. The numerals 1–10, 100, 1000 have both absolute and construct forms, but are used with the counted object in either form, not according to any grammatical rule, as in: אלפי אנשים, עשרת אלפים אנשים, אלפיםעשרה, שלושת אנשים, שלושה אנשים. The feminine construct form is also commonly used in the first decade for feminine nouns as in שלושת נשים, which is always the case with pronouns, and so not only שְׁלָשְׁתָּם but also שְׁלָשְׁתָּן (not, שלשן!). Pronominal suffixes are not used with the numerals 11–99; only in the early literature do we find חֲמִישֵׁיהֶם (II Kings 1:14). On the other hand, pronominal suffixes are used for 100 and 1000 as in במאתנו, באלפיהם ("in our century") as is the dual and plural. For the second decade of numerals two constructions are used; construct in the feminine numeral as עשרהשלוש, and the connection of the two terms without the waw in the masculine numeral as שלושה עשר (cf. no. 10). Note, too, that the second part of the numeral in the second decade has a different form than it has in the first decade, thus: עָשָׂר (not עֶשֶׂר) masc., עֶשְׂרֵה (not עֲשָׂרָה) fem. For 20–90 the connection with waw is common today as: עשרים ושלשה with the ten first followed by the unit but also the reverse order שלשה ועשרים which was common in different periods, and is not very uncommon in literature. However, in the second decade such a construction is considered exceptional (cf. Ez. 45:12 עשרה וחמשה שקל, ועשרים שקליםחמשה; it is the rule in the Aramaic of Elephantine), just as the compound by construct in the masculine חֲמֵשֶׁת עָשָׂר אֶלֶף (Judg. 8:10) or שְׁמׂנַת עָשָׂר אֶלֶף (Judg. 20:25) would be considered exceptional today.
Numerals from 2–10 demand plural nouns only (אישחמישה for example, is considered to be incorrect), from ten upward the noun can be either singular or plural (שלשים אנשים, שלשים איש) and it should be noted that the Academy of the Hebrew Language has suggested that the plural be used to prevent mistakes in the first decade as mentioned above. Combinations of numbers above 100 can be made in different forms (in addition to the possibility of a different order for the tens and units as pointed out above). 3755 can be שלשת אלפים ושבע מאות וחמשים וחמישה or מאות וחמשים וחמישהשלושת אלפים שבע or שלשת אלפים שבע מאות חמשים וחמישה. The last seems to be the most common. It should be noted that a number like 3715 is rendered generally מאות וחמשה עשרשלשת אלפים שבע so that the ו is in the last possible place in the compound.
There are two types: masculine and feminine. From 3–10 its form appears to be derived from a basic element whose pattern is paʿil, as חמישית, רביעי, שלישי, etc. For 1 both the numeral ראשון (from ראש) and אחד serve. The fact that the number "one" has an ordinal form which is not connected etymologically to the cardinals is common to other Semitic languages. From 11 on the ordinal does not have an independent form and the cardinal fulfills that function also: הבית העשרים וחמישה, השנה השלוש עשרה. The common use of the singular of the ordinal (from the third ten and up) הבית העשרים וחמישי (as in Arabic) is considered incorrect. Ordinality can also be syntactically expressed, by using the cardinal as a nomen rectum in the construct state: השנה העשרים וחמש = שנת עשרים וחמש, שנה חמישית = שנת חמש. Just as in the construct the numeral is always the nomen rectum so when it is an adjective it also comes after the noun, as, האיש השלישי and not the opposite.
For ½ the noun חצי is usually used. But, מחצית, מחצה, are also used; for ⅓, ¼, ⅕ on, the feminine ordinal form חמישית, רביעית, שלישית is used, but רביע, שליש or רבע (also חֹמש, (רֹבע are also possible. From ⅙ to 1/10 only the form with ־ִית is used. Note that while the suffixes ־יה, ־ית are elsewhere similar in origin and function (cf. section 9. Suffixes), in the numerals there is a differentiation: עשירית = 1/10, עשיריה = group of ten, חמישית = ⅕, חמישיה = a group of 5.
Multiples are often expressed by שבעתיים ;־ַיִם = ×7 (not 7×2).
Expressed by repeating the number as in שניםשנים.
In summation it may be said that there is nothing in Hebrew morphology to compare with the numeral for different forms and types of usages and syntax, and that the numeral best reflects the special nature of Hebrew morphology which includes, side by side, the very old with the new. "The uniqueness of Hebrew in our day and the source of its problems is that nothing in it has died and so there exist – and are in use – different chronological layers side by side, not on top of one another as in languages with a historic continuity" (Z. Ben-Hayyim, "An Ancient Language in a New Reality," Lešonenu La'am, Jerusalem, 1953, 43–44).
The verb, a part of speech easily identifiable according to all theories and in all languages (including, of course, Hebrew), can be characterized from three points of view: semantically, the verb denotes an action or a change of state or the existence of a state; morphologically, it is usually accompanied by an indication of its grammatical subject, whether as a separate word or as an affix or as both; and syntactically, it functions as the main part of the predicate (see Syntax).
Forms of the Hebrew verb that never (or only sometimes) include an explicit reference to the subject, such as the infinitives and the verbal noun, are considered borderline cases between verb and noun. There are two extreme views on what is to be included in the paradigm of a Hebrew verb. According to one view, the paradigm comprises only such forms as distinguish persons and tenses, and hence each בִּנְיָן ("conjugation") is taken to be a separate verb. For others, however, the paradigm of the verb includes also the conjugations, which are regarded as belonging to one verb if they have a common root. There are also some who treat two or more of the conjugations (but not all) as part of the same paradigm, for example פָּעַל – נִפְעַל or פִּעֵל – פֻּעַל – הִתְפַּעֵל. The best English translation for בִּנְיָן would be "verb pattern," not "conjugation." The latter term is however kept in the following, since it is commonly used in English works dealing with Hebrew morphology.
The verb paradigms are treated almost identically in all accounts, whether traditional or modern, scholarly or pedagogic. There is unanimity on the existence of seven principal conjugations, to which most verbs are related, and also of certain other patterns that have only rare and partial exemplification.
The most acceptable names for the conjugations are based on the form of the third person singular masculine past of a regular root, i.e., of a root having all its consonants in all forms of the paradigm. From the very beginning of Hebrew grammar in the Middle Ages the root פ׳ע׳ל׳ has been used for this purpose, under the influence of Arabic grammar. Accordingly, the names of the conjugations are הֻפְעַל, הִפְעִיל, הִתְפַּעֵל, פֻּעַל, פִּעֵל, נִפְעַל, פָּעַל (or הָפְעַל). However, because of the peculiarity of the consonant ע in Hebrew (see Phonology 6), this root has disadvantages which do not apply to the corresponding root in Arabic:
(1) ע cannot be doubled and therefore the names of the conjugations הִתְפַּעֵל, פֻּעַל, פִּעֵל lack the principal formal characteristics of these conjugations, namely, the doubling of the middle radical, e.g., ב in דִּבֵּר (dibber), ל in שֻׁלַּח (šullah), ג in הִתְרַגֵּל (hitraggel).
(2) When the ע is the first in a consonantal cluster, it is separated from the following consonant by a semi-vowel חטף (see above, §13), i.e., ă, ĕ, or ŏ. This type of vowel is not found, however, in corresponding verbs that do not have ע in that position. Contrast דָּרְשׁוּ (darĕšu) with פָּעֲלוּ (paʿălu), גְּדִי (gĕdi) with עֲדִי (ʿădi).
(3) The ע generally does not close a non-final syllable. When it does according to its pattern, it must be followed by a חטף. Contrast שִׁמְרִי (šimri) with גַּעֲרִי (gaʿări).
(4) The ע sometimes causes changes in preceding vowels, as in the last example (see below §22).
In the 19th century, scholars began to look for another root that would serve to denote the conjugations and decided on ק׳ט׳ל׳. This root has two advantages:
(1) none of its consonants has any peculiarities; and (2) it is found (sometimes in certain variations) in almost all Semitic languages. On the other hand, Hebrew grammarians have pointed out that this root had a serious disadvantage from the educational point of view: its meaning ("to kill") is unpleasant.
Recently, it has been recognized that it is in fact possibleto use the root פ׳ע׳ל׳, provided that it is transcribed phonemically without reference to the phonetic properties of the ע. For this purpose, we have repeated (i.e., doubled) the consonant in the Hebrew representation as well as in the transcription. The names of the seven conjugations, according to this approach are: paʿal (פָּעַל), nipʿal (נִפְעַל), piʿʿel (פִּעֵעל), puʿʿal (פֻּעַעל), hitpaʿʿel (הִתְפַּעֵעל), hipʿil (הִפְעִיל), hupʿal (הֻפְעַל). Since the transcription in Latin symbols is phonemic and not phonetic, the פ is always symbolized by /p/, even when phonetically it has the value [f]; likewise there is no need to distinguish between פַּתָּח and קָמָץ, both being symbolized by /a/.
Other names have been proposed for the conjugations, which identify them semantically (see below), in particular קַל ("light") and כָּבֵד ("heavy"). Originally, these names covered all seven conjugations: קַל for פָּעַל and כָּבֵד ;נִפְעַל for פֻּעַל, פִּעֵל, and הִתְפַּעֵל and כָּבֵד נוֹסָף ("supplemented heavy") or גּוֹרֵם ("causative") for הִפְעִיל and הֻפְעַל. However, the original application was later forgotten. In popular usage and in most textbooks, קַל alone has been retained, but it is restricted to the פָּעַל conjugation. Occasionally, הַבִּנְיָן הַגּוֹרֵם ("the causative conjugation") is used for הִפְעִיל and הַבִּנְיָן הַכָּבֵד ("the heavy conjugation") for פִּעֵל. The decision whether or not conjugations with a common root are to be treated as part of the same verb has great practical significance, not only grammatically but also lexicographically. Dictionaries of modern Hebrew list nouns according to their initial consonants, disregarding whether the consonants are radicals or not, so that, e.g., מִקְרָא appears under מ and תִּלְבּשֶׁת under ת. However, many follow the practice of dictionaries of biblical Hebrew in listing verbs, regardless of conjugation, according to the first radical, e.g., הִלְבִּישׁ under ל next to לָבַשׁ, and נִכְנַס under כ, next to כָּנַס, and sometimes, in the same way, even verbs without a corresponding form in פָּעַל, e.g., הִתְכּוֹנֵן under כ, because of its root כ׳ו׳נ׳. Some modern popular dictionaries, on the other hand, enter the form of the third person masculine singular past for each conjugation, e.g., הִתְרוֹמֵם under ה and נִכְנַס under נ, thus following (though probably unconsciously) the view that verbs from the same root in different conjugations are to be treated as independent verbs. (For the advantages of this view for grammar and semantics, see below.)
Those adopting the view (traditional for all Semitic languages) that conjugations of verbs from the same root constitute one paradigm must consider the semantic relationship between the various conjugations. In the extreme formulation of this view, every conjugation is said to have a particular meaning in relation to the "basic form" of the verb, the form פָּעַל. An attempt alongthese lines has been made, in particular, for biblical Hebrew (see Biblical Hebrew 9). It has resulted in the fabrication of imaginary forms that do not appear in the Bible, and also in the neglect of some forms that do, for example, the נִפְעַל conjugation.
A similar, though less extreme, position is generally taken in textbooks. These also treat the פָּעַל conjugation as the basic form, said to be semantically "simple," but they attempt to establish a semantic relationship for each root between the פָּעָל conjugation and other conjugations, it being assumed that each of these makes some addition to the basic meaning of the פָּעַל. The principal additional meanings, expressed synthetically by a change in the form of the verb, are said to be passive, reflexive, reciprocal, strengthening, durative, iterative, causative, change in state, declarative, and deprivative. Thus, for example, the נִפְעַל is said to express the passive when the active agent is found with the פָּעַל, e.g., שָׁבַר ("he broke") – נִשְׁבַּר ("it was broken") or הִפְעִיל, e.g., הִרְגִּיעַ ("he soothed") – נִרְגַּע ("he was soothed"); reflexive, e.g., נִזְהַר, נִשְׁמַר ("he took care of himself"); or reciprocal, e.g., נִדְבְּרוּ ("they spoke to one another"). The פִּעֵל is said to express the strengthening, e.g., שָׁבַר ("he broke") – שִׁבֵּר ("he smashed"); the durative, e.g., רָקַד ("he danced") – רִקֵּד ("he danced for a long time"); repetitive, e.g., קָבַר ("he buried") – קִבֵּר ("he buried many"); causative, e.g., לָמַד ("he learned") – לִמֵּד ("he taught"); or deprivative, e.g., שֵׁרֵשׁ ("he uprooted"). The פֻּעַל is said to be the passive equivalent of verbs with the same root in פִּעֵל. The הִתְפַּעֵל is explained as denoting the reflexive, e.g., הִתְרַחֵץ ("he washed himself"); reciprocal, e.g., הִתְלַחֲשׁוּ ("they whispered to each other"); passive chiefly when the active is in the פִּעֵל, e.g., בִּשֵּׁל ("he cooked") – הִתְבַּשֵּׁל ("it was cooked"); or strengthening, e.g., הִתְנַשֵּׁם ("he breathed strongly"). The הִפְעִיל is said to denote the causative, chiefly when the active is in the פָּעַל, e.g., מָלַךְ ("he reigned") – הִמְלִיךְ ("he made [him] a king"), and consequently changes the verb from intransitive to transitive, e.g., יָשַׁב ("he sat") – הוֹשִׁיב ("he caused to sit," "he set"), or from unitransitive to ditransitive, e.g., אָכַל ("he ate") – הֶאֱכִיל ("he fed"); a change of state, e.g., הֶעֱשִׁיר ("he became rich"), especially a change of color, e.g., הִלְבִּין ("it became white"); or declarative, e.g., צַדִּיק ("righteous") – הִצְדִּיק ("he declared as righteous," "he justified"). The הָפְעַל is considered the passive equivalent of the הִפְעִיל. In addition, some verbs with a "simple" meaning like that of the פָּעַל appear in other conjugations, e.g., נִכְנַס ("he entered"), רִחֵף ("he hovered"), הִתְנַגֵּד ("he opposed"), הִמְתִּין ("he waited").
However, this view of the semantic relationships of the conjugations, with the פָּעַל taken as the basic conjugation, does not sufficiently fit the facts. Even a partial examination of Hebrew verbs shows that, except for פֻּעַל and הָפְעַל (which almost always have a predictable relationship with פִּעֵל and הִפְעִיל respectively), we cannot automatically predict the meaning of a root in one conjugation from that of the same root in another conjugation. Though there are many instances of predictable semantic relationships between the conjugations, like those given above, in many instances verbs of the same root have no relationship at all or have an unpredictable relationship, e.g., דִּבֵּר ("he spoke") – הִדְבִּיר ("he subdued"); מָהַר ("he bought a wife") – מִהֵר ("he hastened"); בָּצַר ("he gathered grapes") – נִבְצַר ("it was withheld") – בִּצֵּר ("he fortified"); סָפַר ("he counted") – סִפֵּר ("he told") – הִסְתַּפֵּר ("he had his hair cut"); סָפַק ("he clapped hands") – סִפֵּק ("he supplied") – הִסְפִּיק ("he made enough") – הִסְתַּפֵּק ("he had sufficient"); הֶאֱמִין ("he believed") – הִתְאַמֵּן ("he trained himself"). It may well be that at an early period of the language, the conjugations constituted a paradigm of predictable semantic relationships similar to the paradigm of changes of person and tense within a conjugation, but as a consequence of the development of meanings of verbs throughout the history of the language, the conjugations cannot now be recognized as belonging to one paradigm.
Similarly, there was once a fixed semantic relationship between nouns of different patterns belonging to the same root. Indeed, even in contemporary Hebrew there are such relationships, e.g., between nouns denoting people with particular occupation such as סַפָּר ("barber") and the corresponding noun for the place of work, מִסְפָּרָה ("barber shop"). In general, new nouns have been formed in recent times on the appropriate patterns, e.g., קָטִיף ("season for picking fruit growing on trees") and תָּלִישׁ ("season for picking fruit growing on low bushes") for the seasons of agricultural work; מִרְפָּאָה ("clinic") and מִכְבָּסָה ("laundry") for places of work. Nevertheless, each noun is treated as an entirely independent noun; the semantic relationship between nouns of the same root has not resulted in their being considered one noun with various patterns.
Two forms have predictable relations when they fulfill two conditions: (1) if one of them exists, it follows that the other exists too; (2) when the meaning of one of them is known, the meaning of the other one is self-explanatory. Only forms having predictable relationships with other forms can be considered as belonging to the same paradigm, since only these can be freely used by a speaker though he has never heard them before and are unambiguous to the hearer though he has never encountered them before. For example, within one conjugation there are predictable relationships between forms varying only in person or tense, such as שָׁמַרְתִּי–שָׁמַרְתָּ or שָׁמַרְתִּי–אֶשְׁמֹר. Similarly, the relationships between verbs in פִּעֵל and הִפְעִיל and verbs of the same root in פֻּעַל and הָפְעַל, respectively, are virtually predictable. There is not complete predictability because some intransitive verbs in פִּעֵל, e.g., טִיֵּל ("he went for a walk") and רִחֵף ("he hovered"), and in הִפְעִיל, e.g., הִסְמִיק ("he became red") and הֶחְלִיד ("he [it] became rusty") either do not have corresponding forms in פֻּעַל or הָפְעַל or, if they do, these do not express a passive meaning.
Predictable relationships between פָּעַל and נִפְעַל (with respect either to the existence of one form if the other exists or to the stipulated semantic relationship) apply only to some verbs. Thus, there is no corresponding form in the other conjugation for יָרַד ("he descended"), נִבְהַל ("he was alarmed"), or נִשְׁבַּע ("he swore"), and a predictable active-passive relationship is lacking between יָשַׁב ("he sat") – נוֹשַׁב ("it was inhabited"); כָּנַס ("he assembled") – נִכְנַס ("he entered"); or יָשֵׁן ("he slept") – נוֹשַׁן("he was old"). Even less predictable are the relationships between פָּעַל and פִּעֵל. It is impossible to know whether the change from פָּעַל to פִּעֵל will entail strengthening, lengthening, repetition, causation, or some other meaning, which might be completely different from that of פָּעַל, e.g., שִׂחֵק ("he played") – שָׂחַק ("he laughed"), חִנֵּךְ ("he educated") – חָנַךְ ("he inaugurated"), בִּשֵּׁל ("he cooked") – בָּשַׁל ("it ripened"). Moreover, there are many verbs in פִּעֵל that have no corresponding forms in פָּעַל, e.g., טִיֵּל ("he went for a walk"), זִנֵּב ("he routed the rear"), חִדֵּשׁ ("he renewed"), חִיֵּךְ ("he smiled"), צִוָּה ("he commanded").
Similarly, it is impossible to be confident that the הִפְעִיל will express the causative of the פָּעַל, since many verbs in הִפְעִיל do not have any semantic connection with the same root in the פָּעַל, or have an unpredictable relationship, e.g., יָרַק ("he spat") – הוֹרִיק ("it became green"); סָרַט ("he scratched") – הִסְרִיט ("he filmed"); רָצָה ("he wanted") – הִרְצָה ("he lectured"). Furthermore, there are some verbs in הִפְעִיל whose passive is in נִפְעַל as well as in הָפְעַל (occasionally with some difference in nuance), e.g., הִרְתִּיעַ ("he deterred") – הִדְפִּיס, נִרְתַּע ("he published") – נִדְפַּס, and these somewhat disturb the predictability in relationship between הִפְעִיל and הָפְעַל.
There is certainly no predictable relationship between פָּעַל and הִתְפַּעֵל. Not only are there two large categories of semantic relationships, exemplified, on the one hand, by רָחַץ ("he washed") – הִתְרַחֵץ ("he washed himself"), and on the other, by כָּתַב ("he wrote") – הִתְכַּתֵּב ("he had a correspondence with [someone]"), but there are also many verbs appearing in only one of these conjugations, e.g., גָּזַל ("he robbed") and הִתְקָרֵר ("he caught a cold"), or which have independent meanings in the two conjugations, e.g., סָפַר ("he counted") – הִסְתַּפֵּר ("he had his hair cut"). The semantic relationships between other conjugations, such as פִּעֵל – הִתְפַּעֵל or פִּעֵל – הִפְעִיל are also unpredictable in similar respects.
An awareness of this situation requires a consideration of the conjugations not as an inflection of one verb but as a set of different verb patterns related by derivation. The relationships between פִּעֵל and פֻּעַל and between הִפְעִיל and הָפְעַל may perhaps be an exception to this generalization.
The inflection of a verb includes all forms of the verb that vary in pronominal subject or object, e.g., יִשְׁמֹר ("he will guard") – יִשְׁמְרֵהוּ ("he will guard him"), gender, number, tense, and modality. In traditional literary language, modal differences are chiefly expressed synthetically in the verb itself by certain additions to the normal forms, e.g., אֶשְׁמֹר – אֶשְׁמְרָה, but sometimes auxiliary verbs have this function (see Syntax). Differences in tense include not only the distinctions between past, present, and future, but also forms having a modal character, e.g., the imperative, and those that lack a time distinction, e.g., the construct infinitive, the absolute infinitive, and the action noun. These last three can be used in nonverbal functions as well as the present form which can also function as a noun in all respects. The affixed pronominal forms, which by their characteristics and by their place in the verb determine not only the subject and object but also the tense, are the inflectional morphemes, and are attached to the basic form, the inflectional base. The relationship between the inflectional base and the inflectional morphemes is similar to that between basic element and formative (see above 3. The Basic Element; 4. The Formative and the Morpheme), except that the combination of the latter results in a new word with a separate entry in the dictionary, since it is a derivational process and lacks the completely predictable semantic relationship of an inflectional process. In the פֻּעַל, הָפְעַל, and הִתְפַּעֵל conjugations, the same base serves for all the tenses, but in the other conjugations different bases are used, as can be seen in the table below. The existence of more than one inflectional base in a paradigm is not unique to the conjugations: there are some noun patterns with the same feature.
Inflectional Bases in the Conjugations
(In phonemic transcription, in which šewa is not marked; see Phonology: 13. The Šewa and Ḥaṭefs.) (See Table: Hebrew Grammar 1.)
The same inflectional morphemes are used for the same tense in the various bases, though between base and affixed morpheme there may develop sometimes transitional phones and other phonetic phenomena that can be described precisely in a few rules. The inflectional morphemes of past and imperative are suffixes, while those of future are prefixes, except that in the second and third person plural, and in the second person feminine singular there are also suffixes identical with those of the imperative. The inflectional morphemes of the participle are those for gender and number in nouns (see 11. The Declension of the Nouns) and do not vary with change of person. In addition, the forms of the participle in all the conjugations except פָּעַל and נִפְעַל are prefixed by the morpheme m (sometimes a vowel is inserted after the m, see below). The inflectional morphemes of the construct infinitive resemble the inflectional morphemes of nouns (see above 11. The Declension of the Nouns). Every verb which can be followed by an object (usually only a direct object) can take, after the inflectional morphemes already mentioned, an additional inflectional morpheme, the objective pronoun (for details, see below 21. Inflections of Weak Verbs).
|Note: The following changes affect the inflectional base of the הִתְפַּעֵל:|
|(1) When the first radical is ז, ס, צ, שׁ, or ׁש it precedes the t (see Phonology 6, 17). This change is optional if the first radical is ד or ט.|
|(2) The t is changed to d after ד and ז, and to ṭ after ט and צ.|
The following are the affixed pronominal inflectional morphemes denoting the subject:
|To the inflectional base (= – ) of the Past|
|Singular:||1st Person||Common||– ti|
|2nd Person||Masculine||– ta|
|3rd Person||Masculine||– ø|
|Plural:||1st Person||Common||– nu|
|2nd Person||Masculine||– tem|
|3rd Person||Common||– u|
|To the inflectional base (= – ) of the Future|
|Singular:||1st Person||Common||ʾ –|
|2nd Person||Masculine||t –|
|Feminine||t – i|
|3rd Person||Masculine||y –|
|Plural:||1st Person||Common||n –|
|2nd Person||Masculine||t – u|
|Feminine||*(t – na)|
|3rd Person||Masculine||y – u|
|Feminine||*(t – na)|
|To the inflectional base (= – ) of the Imperative|
|Singular:||2nd Person||Masculine||– ø|
|Plural:||2nd Person||Masculine||– u|
*The masculine forms of the same person are usually used instead of these. The form given in parentheses is the prevailing one in biblical Hebrew, but nowadays it is considered a possible variant only. To some extent, especially in the colloquial language, the masculine form of the second person plural in the past is also used for feminine. See also note 8 of the following section.
The following rules describe the principal changes affecting the form of the verb when the inflectional morpheme is affixed to the base:
(1) h at the beginning of the base is omitted after future and participle prefixes (but not when ב, כ, or ל come before the construct infinitive).
t + hiššamér → thiššamér → tiššamer
y + hatḥíl → yatḥíl
but l + hitgabbér → lhitgabbér
(2) When by adding a consonant before a base, the result is a form with a cluster of three consonants at the beginning, a vowel (generally i) is inserted between the first two consonants, namely between the prefix and the first consonant of the base. (This is a general rule which applies to the future and infinitive of paʿal, as well as to many other forms, e.g., below 7).
n + šmór → nšmór → nišmór
(3) After the prefix ʾ, the vowel e instead of i is used in rule 2 and when a preceding h is omitted (rule 1). See also the beginning of 22. Inflections of Weak Verbs.
ʾ + šmór → ʾšmór → ʾešmórʾ
+ hitgabbér → ʾhitgabbér → ʾitgabbér → ʾetgabbér
There are a few exceptions, chiefly in biblical Hebrew, where i follows the prefix ʾ.
(4) The vowel before the last consonant of the past base, whether é or í, changes to á before any suffix beginning with a consonant.
dibbér + ti → dibbárti; hitḥíl + ta → hitḥálta
(5) When the suffix is a vowel, the vowels é, ó, and á be-fore the last consonant of the base generally remain only in forms used in classical Hebrew, especially at the end of a sentence. ("Pausal forms.")
huggáš + a → huggáša
hadál + i → ḥadáli
šamár + u → šamáru
However, they disappear (become šewa) in the regular form of the verb, when the accent is on the syllable of the suffix.
šamár + á → šamrá (or šamәrá)
dibbér + u → dibbrú (= dibbәru)
See below for some uses of pausal forms in contemporary Hebrew. However, the vowel i does not disappear:
hitḥíl → hitḥílu
(6) The vowel í in the imperative and future base of הִפְעִיל changes to é before the suffix –na.
t + hatḥil + na → tatḥélna
And in the imperative before the suffix ø, í changes to é:
hathíl + ø → hatḥél
(7) If through the disappearance of é, ó, or á (according to 5) a cluster of three consonants is created at the beginning of the word, a vowel is inserted between the first two (as in 2). When the third consonant is b, k, p – it is realized as the corresponding fricative variant, i.e., v, x, f.
šmór + i → šmóri → šmrí → šimrí (imperative of פָּעַל), and similarly with construct infinitive, e.g.,
l + šmór + ó → l + šmró → l + šomró → lšomró.
(8) The vowels a and e in an open syllable before the accent disappear when the accent moves to the end of the bases:
labéš + im → labešim → lbešim
šamár + tém → šamartém → šmartém
In colloquial language, the suffix – tem is unaccented. It is used for both masculine and feminine. Since the accent does not shift, a and e do not disappear:
šamár + tem → šamártem (col.)
(9) The vowel é in the last syllable of the participle base (except for the base labéš) disappears when a suffix beginning with a vowel is attached to the base. (Such a suffix attached to a participle, which is a nominal suffix, is always accented. But see the next rule.)
mdabbér + ím → mdabbrím
šomér + á → šomrá
(10) If the suffix –t (but not –át) is attached to the participle base as the feminine inflectional morpheme, the result is a form ending in two consonants, which is treated like the segholates (see above 6).
nišmár + t → nišmárt → nišmáret → nišméret
mdabbér + t → mdabbért → mdabbéret
(1) The attachment of the suffix ø to the base usually does not affect the form of the base, except for phonetic changes, e.g., the change of קָמָץ to פַּתָּח in the last syllable of the base (but see rule 6, above).
šamár (= שָׁמָר) + ø → šamár (= שָׁמַר).
(2) The theoretical form of the base is also the form that is realized at the end of a sentence. In general, the "pausal forms" of all the persons are the forms from which it is possible to produce the regular forms by the rules of inflection detailed above. It should be further noted that some "pausal forms" are sometimes used in ordinary speech, and not necessarily at the ends of sentences, e.g., הַהַצָּעָה לֹא הוּבָנָה כָּרָאוּי, הָבוּ לָנוּ.
(1) For the objective inflection the forms of the verb containing the subjective pronoun serve as the inflectional base (see 19, 20). The inflectional morphemes denoting objective pronouns are as follows:
|3rd Person||Masculine||hu, w, o|
These inflectional morphemes are, in the main, attached to all the bases of the verb, but there is not always free variation in the inflectional morphemes of the third person singular, the choice of which sometimes depends on the nature of the base. Thus, the morphemes h and o are not affixed to the base šamárti, and others like it. The inflectional morphemes affixed to participle bases are generally the possessive inflectional morphemes of the noun (see above, 11).
(2) Changes in the Base. These inflectional morphemes generally cause changes in the vowels of the base, because the accent of the base usually moves forward when the inflectional morpheme is attached. As a result of the movement of the accent, vowels disappear, mainly according to regular phonetic principles (see Phonology 16).
šamárti + kém → šmartikém
(see above 20, rule 8.)
šamáru + ni → šmarúní
tišmóri + m → tišmrím
(the base vowels á, ó disappear, see above 20. Changes in the Base, paragraph no. 5.)
A striking change in the past form ending in tém or ténis the change of this suffix to tú when this form serves as the base for the objective inflection.
šmartém + ni → šmartún
(3) Transitional Phones. In several bases transitional phones are created between the base and the inflectional morpheme:
(1) The vowel í is added to the base of the second person feminine singular past:
šamárt + hu → šmartíhu
šamárt + nu → šmartínu
(2) t is added to the base of the third person feminine singular past with shift of accent:
šamára + ni → šmarátni
šamára + kém → šmaratkém
(3) The vowel á is added to the base of the third person masculine singular past:
šamár + nu → šmaránu
šamár + m → šmáram
But before the morpheme k (second person feminine singular) the inserted vowel is é:
šamár + k → šmarék
(4) The vowel é is added to the future bases ending in a consonant:
tišmór + ni → tišmréni
yilbaš + m → yilbašém
The vowel á of the future and imperative base of lbáš pattern does not disappear, unlike é and ó in other future and imperative bases. (See Phonology 16:2, 4.) Contrast future bases ending in a vowel, e.g.,
tišmóri + ni → tišmríni
yišmóru + m → yišmrúm
yilbášu + n → yilbašún
The transitional vowel á (and sometimes é) is affixed to the infinitive base:
lišmór + ni → lšmoréni (cf. 20.2)
laqáḥat + m → lqaḥtám
This concept is first discussed above in 2 (end), where it is stated that because of a phonetic characteristic of one of the radicals it may sometimes happen that two realizations of the same pattern may be different. The simplest change is that resulting from the peculiar phonetic characteristics of the gutturals א, ה, ח, and ע (see Phonology 6, Morphology 16). Their presence sometimes necessitates a vowel not found in the corresponding form without a guttural, e.g.,
In these cases, the high vowel i is replaced by a lower vowel, e or a, next to a guttural. (See also rule 3 in 20.) In addition, between a non-final guttural and the following consonant a vowel חטף is sometimes inserted, usually corresponding to the preceding vowel, e.g., yaḥǎmol, yeḥәzaq, yoʿǒmad, but also muʿǎmad. In the case of ה, ח, and ע if as the third radical it ends the word, and there is no vowel a preceding it, then a preceding a is inserted, e.g., niẓẓéaḥ, yaškíaḥ. With respect to א in this situation, see below 24.4. The impossibility of doubling the gutturals and ר (Phonology 8, Morphology 16) causes changes in preceding vowels when doubling is required in corresponding forms. The changes are from short vowels to long vowels, and are known in grammar as "compensation for the dageš." Sometimes, the change is expressed merely orthographically, e.g., in the pattern ydabber:
יְפָרֵשׁ, יְפָאֵר instead of יְפַרֵשׁ ‡, יְפַאֵר ‡
but sometimes it is also audible in modern pronunciation, e.g., in the pattern mdubbar:
מְיֹעָר, מְגֹהָץ instead of מְגֻהָץ ‡, מְיֻעָר ‡תֵּאֵר, פֵּרֵשׁ instead of פִּרֵשׁ ‡, תִּאֵר ‡
Apart from the gutturals and ר, there are two characteristics that cause changes in the forms:
(1) Assimilation, when one consonant completely assimilates another consonant, usually regressive assimilation.
‡yinpol → yippol; ‡yilqaḥ → yiqqaḥ
(2) Elision, when one of the radicals is א, ו, or י:qarʾú, but qratém
yašanta, but tišán (‡ tiyšán)
šaléw, but šalíti
The conditions for these changes are that, for assimilation, the assimilated consonant be at the end of a non-final syllable, and for elision, the elided consonant be at the end of any syllable. (See below, in this connection, forms of verbs with identical second and third radicals, e.g., נסב, סב.)
Typical of paradigms of assimilated forms is the presence of duplication of consonants in the middle of the verb. These paradigms are more often known as paradigms of defective verbs. This term derives from a study of the written language, since in the Hebrew script duplication is indicated by only one letter (with a point, "dageš forte," inserted in it). As a result, when a consonant is assimilated to its neighbor, one letter is missing from the script, that of the assimilated consonant. The term assimilated derives from a study of the phonetic characteristics of the language and an observation of the phonetic processes of assimilation as the principal characteristic typifying the membership of the root in this paradigm. However, there are a few instances in these paradigms of omission of a consonant and not its assimilation.
1. Defective פ״נ and Defective פ״י (Better: assimilated פ״נ and assimilated פ״י). The paradigm of this type with most roots is known as defective נ״פ (see Phonology 7), which comprises roots whose first radical is נ. This נ is assimilated to the second consonant of the root in certain circumstances, e.g.,
‡ yinpol (//yišmor) → yippol
‡ ninẓal (//nišmar) → niẓẓal
As stated at the end of 22, this assimilation occurs only when the assimilated consonant comes at the end of a non-final syllable. Most roots with נ as the first radical belong to this paradigm, but not all: sometimes the נ appears even when the condition exists for its assimilation. This is so when the second radical is a guttural, e.g:, yinham, yanʿim. (There are only two or three roots where the נ is assimilated to a guttural in some forms.) And it is so in a sizeable number of other roots, e.g., yanẓiḥu, yinbor, especially in verbs or forms that have been coined in recent times. It is worth pointing out that the facility for assimilating the נ has become a means of distinguishing between different meanings, e.g., yanbiṭ ("will cause to bud") versus yabbiṭ ("will look"), yangid ("will put contradictory items") versus yaggid ("will tell"), including meanings of verbal nouns, e.g., hangada versus haggada, hankara ("alienation") versus hakkara ("consciousness"). The root ל׳ק׳ח׳ is generally included in this paradigm because of the assimilation of its first radical, even though it is ל and not נ.
A parallel paradigm is that of defective פ״י, which comprises roots whose first radical –י– assimilates to the second. There are only six or seven roots in this paradigm, e.g., yẓb (yaẓẓib, hiẓẓib), yẓq (niẓẓaq, niẓẓoq). The bases of verbs that are defective פ״נ and פ״י hardly differ from those of regular paradigms, which are detailed in 20, and the same changes of base occur, which are included in those rules. The exception is that the imperative bases in פעל (lbaš, šmor) tend in the פ״נ and פ״י paradigms to lose the first radical without any substitution, e.g., pol (for npol), gaʿ (for ngaʿ). In these imperative forms there is therefore a true loss and not an assimilation (see Biblical Hebrew 10). With the root נתנ the last נ is assimilated to the following consonants in cases like natánti → nattáti and in לָתֵת. This last form is to be understood as a development of a feminine infinitive, in which t is added to the radicals. Other infinitives like this occur mainly in biblical Hebrew, e.g., לְאַהֲבָה, יְכֹלֶת. This construction as an infinitive is not productive nowadays, and most instances that are still used belong to the פ״נ or פ״י paradigms, e.g., לָדַעַת, לָרֶדֶת, לָטַעַת, לָגֶשֶׁת. A similar process has occurred with the biblical form לָלַת for the root י׳ל׳ד׳, which must be understood as laladt → lalatt → lalat.
2. Verbs with duplicated second radical or geminates are usually included among the defective paradigms. This can be justified not because in certain forms one letter is written symbolizing both the second and third radicals, e.g., סַבּוּ, but because in other cases the last radical is entirely omitted, e.g., נָסַב, סַב. But two points must be made clear: (a) In the inflections of roots in this paradigm, assimilation does not take place, since the two neighboring consonants written with gemination are identical; there is nothing unique in their being symbolized by one letter, since all gemination is symbolized in Hebrew script by one letter. Gemination is similarly symbolized by one letter in words like natánnu (נתנו), karátti (כרתי). (b) The omission of the third radical in forms like נָסַב, סַב is only a realization rule, a phonetic rule, and does not convey anything about the theoretical structure of the word. The omission occurs, therefore, on the final level of the language, since there is a general rule in Hebrew that gemination of consonants does not occur at the end of a word, one consonant alone remaining instead. A theoretical form like sabb changes, therefore, to sab without any assimilation, just as with words like חֹק, דֹּב, לֵב, in all of which gemination occurs with the last radical when it appears in the middle of the word, e.g.,חֻקּוֹ, דֻּבִּים, לִבִּי. (See Table: Hebrew Grammar 2.)
|Past Participle||Future Imp. Inf.|
The regular inflectional morphemes (19) are attached also to the bases of this paradigm, and rules 1, 3–5, 8–10 set out in 20 apply also to these bases, e.g.,
t + hitsobéb → thistobeb → tistobeb (20.1)
ʾ + hitsobéb → ʾestobeb (20.3)
sobéb + ti → sobábti (20.4)
sobáb + u → sobábu (pausal form) → sobәbu (20.5)
nasább + á → nasabbá → nsabbá (20.8)
m + sobéb + ím = msobәbím (20.9)
m + sobéb + t → msobébt → msobébet (20.10)
The principal phenomena that are peculiar to the rules for affixation of inflectional morphemes to bases of geminates are as follows:
Prefixed Morpheme: (1) a transitional vowel appears between the prefixed morpheme and the future bases of לַעָּפ if the accent is on the base. Before the base sobb the transitional vowel is a, while before the qall it is e:
y + sóbb → y + a + sóbb → yasóbb → yasób
y + sóbb + u → y + a + sóbb + u → yasóbbu
y + qáll + u → y + e + qáll + u → yeqállu
The same applies when ל is affixed to the infinitive base:
l + sóbb → lasóbb → lasób
Suffixed Morpheme: (2) when a suffix beginning with a consonant is attached to a base ending with gemination, the transitional vowel o is generally inserted between them, with e instead of o preceding – na:
sább + ti + sább + ó + ti → sabbótii
sább + tém → sább + o + tém → sabbotém
hissább + na → hissább + é + na → hissabbéna
(3) However, there are forms lacking this transitional vowel which lose the gemination of the base before a suffix beginning with a consonant, as if it was at the end of the word:
hissább + na → hissábna = [hissávna]
hussább + ta → husábta = [husávta]
hesébb + nu → hesábnnu = [hesávnu]
(4) The vowel e before gemination of the base is changed to i, and o to u when the accent is after the gemination:
sóbb + na → sóbb + é + na → subbéna
m + hesébb + ím → mesébbím → msibbím
(5) As usual with a consonant cluster whose first consonant is h, a half-vowel appears in this situation:
hasébb + nu → hasébb + ó + nu → hsibbónu → [hăsibbónu]
The term "mute" refers to the phonetic process occurring to various inflectional forms in which the consonants ו, א and י of the root cease to be pronounced, becoming mute. (With respect to ה, see below.) Sub-classification of these paradigms is based on the identity of the mute consonant and on its place in the root: mute פ״א; mute פ״י; mute ע״ו; mute ע״י; mute ל״א; mute ל״י. There are very few instances of mute פ״ו and mute ל״ו; these are therefore usually not treated separately, but are included under mute פ״י and mute ל״י respectively.
1. Mute פ״א. There are only five roots in this paradigm. א is mute chiefly in future forms of פָּעַל, e.g., יׁאפֶה, יׁאבֵד, יׁאכַל. There are a few instances of other forms with mute א, e.g., נׁאחֲזוּ. The vowel o appears in a syllable in which א is mute.
2. Mute פ״י. This paradigm comprises two subclasses, distinguished by the vowel of the syllable in which י is muted: חִירִיק, e.g., יִינַק, יִישַׁן; or צֵירֶה, e.g., יֵדַע, יֵשֵׁב. This latter subclass is assigned by many grammarians to defective פ״י paradigm, chiefly because the letter י is not written in many of the forms. It is true that in the inflectional base of future, imperative, and infinitive of פָּעַל the first radical is omitted (see below in the table of bases in this paradigm: réd). However, the defective paradigms are characterized not by omissions, but by the assimilation of a consonant to its neighbor, and in this paradigm there is no such assimilation (compare above, 23:1). Consequently, this subclass belongs to the mute paradigms. Historically, only forms with חִירִיק as vowel of prefix give convincing proof of an original י as first radical, while those with צֵירֶה suggest an original ו. The difference is not apparent in the future forms of פָּעַל, but it is very clear in the הֵישִׁיר, הֵינִיק :הִפְעִיל as opposed to הוֹשִׁיב, הוֹרִיד. The vowel o following the ה is the result of monophthongization of aw, the original forms probably being ‡hawrid, ‡hawšib (see Phonology 15), while הֵישִׁיר, הֵינִיק result from the monophthongization of ‡hayniq, ‡hayšir. (Evidence of this is also to be found in forms like מַיְשִׁירִים, מַיְמִינִים, where monophthongization has not taken place.)
Inflectional Bases of Mute פ״י Paradigm
The bases of פֻּעַל, פִּעֵל, and הִתְפַּעֵל are identical to those of regular verbs, see 19. (See Table: Hebrew Grammar 3.) The rules for inflection given in 20 apply to all inflections in this paradigm. Special attention should be paid to the following:
(1) The first radical י is mute when by rule 20.2 it is preceded by the vowel i:
t + yšán → tiyšán → tišán (rule 20.3 does not apply);
(2) Before the base for future, imperative, and infinitive of פָּעַל réd, the transitional vowel e is inserted:
t + réd → t + e + réd → teréd
(3) The infinitive base of פָּעַל given in parentheses follows a development characteristic of the segholates, e.g., rédt →rédet, and the transitional vowel a is inserted before it and after the prefix l, e.g., l + rédet → l + a + rédet → larédet (when the infinitive is in the construct state, no transitional vowel is inserted).
3. Mute ע״י–ע״ו. The distinction between ע״ו and ע״י is evident only in the inflectional base for the future, imperative, and infinitive of פָּעַל, exemplified by qum in ע״ו and širin ע״י.
Since this description is res tricted to contemporary Hebrew, there is no discussion of the difficult problem, still disputed, as to whether the roots of this paradigm were originally bilateral and at a later stage ו or י developed between the two consonants, or whether they were originally trilateral and subsequently the middle consonant, ו or י, was muted (see Biblical Hebrew 10, Morphology 2). For our purpose it is sufficient to point out that the name of this paradigm is based on the second possibility. On the other hand, there is evidence of the formation of regular trilateral roots (19) through the development into a consonant of a medial ו or י e.g., תִּוֵּךְ from בִּיֵּן, תּוֹךְ from ב׳י׳נ׳. However, it should be mentioned that most of the creations are in פִּעֵל (as in the above examples) or in פֻּעַל and הִתְפַּעֵל (e.g., הִתְגַּיֵּר, מְגֻוָּן), that is to say in patterns where the middle radical should be doubled and hence greater attention is paid to it. When the development of a medial ו or י does not take place, the inflection for פֻּעַל, פִּעֵל, and הִתְפַּעֵל in this paradigm is identical to that of geminates described in the preceding section.
Inflectional Bases of Mute ע״י–ע״ו Paradigm
The inflectional morphemes detailed in 20 are affixed to these bases according to rules 1, 3, 4, 5 (but rule 5 does not apply to the bases qám, nakón), 6, 8, 10 listed in 20. (See Table: Hebrew Grammar 4.) In addition, transitional vowels are formed according to rules 1 and 2 stated for the geminates (see section on verbs with duplicated radicals in Paradigms of Assimilated Forms above), with only a slight difference: the transitional vowel of the prefix is always a (except for the base of the verb בּושׁ – which serves both past and future – where it is e). The following rules are peculiar to this paradigm:
|Past and Participle||Future Imp. Inf.|
(1) The vowel ú of qúm changes to o before the suffix – na:
qúm + na → qómna
(2) The vowel ó generally changes to u before the transitional vowel ó:
nakón + t → nakón + ó + t → nkunót
Here are some additional examples of the production of forms:
(1) (rule 1) t + haqím → taqím
(2) (rule 3) ʾ + hikkón → ʾikkon → ʾekkón
(3) (rule 4) heqím + ta → heqámta
(4) (rule 5) huqám + u → huqámu (pausal form) →huqmú,
qám + u → qámu
nakón + a → nakóna
(5) (rule 6) haqím + na → haqémna
(6) heqámtém (rule 4) → hqamtém (rule 8) → [hăqamtém] (by rule 5 in the geminate paradigm)
(7) m + huqám + t→ muqámt (rule 1) → muqámet →muqémet (rule 10).
Examples of the rules for transitional vowels:
(8) t + qúm → t + a + qúm → taqúm (geminates – 1) but n + boš → nebóš
(9) heqim + ta → heqím + ó + ta (geminates – 2) →hqimóta (rule 8) → [hăqimóta] (geminates – 5)
For an example where a transitional vowel is not formed, see (3) geminates.
4. Mute Third Radical. With respect to the roots in this paradigm too, it can be argued that they were originally bilateral with a third consonant developing in the final position. Indeed, there is evidence in several cases of such a development (see above, Morphology 2: ז׳ה׳י׳, א׳ח׳י׳). However, the more common view is that these roots were originally trilateral and the final consonant was muted in certain circumstances. This view finds support in the history of Hebrew and other Semitic languages. It is worth pointing out that Hebrew has only a few instances of the existence of ו as a third radical (as in שָׁלַוְתִּי), apparently because already at anearly stage the ו in such roots was changed to י. It should also be mentioned that the popular name ל״ה is based on the written language and not the spoken, since while the third person masculine singular past forms of these roots are indeed written with final ה, a consideration of the whole of the inflection shows that the final consonant is really י, e.g., in forms like the passive participle רָצוּי or verbal noun in רְצִיָּה, פָּעַל rẓiya (like šmira); verbal noun in פִּעֵל niqquy (like dibbur); and in various noun forms, e.g., niqqayon. Finally, it may be noted that in many instances there is an overlap between roots in the paradigms ל״א and ל״י, e.g., הַשְׁקָאָה (as well as הַשְׁקָיָה) from הַרְצָאָה, ש׳ק׳י׳ from ר׳צ׳י׳. The reverse phenomenon is especially common: forms with a root from the paradigm ל״א but with their inflection following that of ל״י, e.g., ב׳ט׳א׳) בִּטּוּי). In mishnaic Hebrew this was the general practice, which was apparently reinforced through the influence of contemporary Aramaic, in which the ל״א paradigm was completely lost and its forms became identical to those of ל״י. However, it would not be correct to argue that this is a phenomenon restricted to mishnaic Hebrew: the transition is reflected in the pointing of the biblical text, and most of the instances can be explained as deriving from an internal phonetic development, namely, the elision of א when it is second in a consonantal cluster, e.g., חוֹטְאִים ← חוֹטִאים. Forms common in mishnaic Hebrew, e.g., יָצְתָה, מָצִינוּ serve nowadays merely as stylistic variants.
Inflectional Bases of Mute ל״א Paradigm
The inflectional morphemes listed in 19 are affixed to the bases of this paradigm and the rules detailed in 20 apply (with the reservations stated immediately below for rules 4 and 10). (See Table: Hebrew Grammar 5.) The following rules are specific to this paradigm:
(1) ʾ is mute at the end of a word, and in the middle of a word before a suffix beginning with a consonant:
qoré' → qoré
qará' + ti → qaráti
milléʾ + ta → milléta
m + hitmallé' + t → mitmallét
i.e., rule 10 of 20 does not apply to this paradigm, because the elision of ʾ prevents the creation of a segholate form.
(2) Instead of rule 4 of 20 the following rule applies to this paradigm: a and i in past bases (except of פָּעַל) change to e before suffixes beginning with a consonant:
niqrá' + ti → niqráti (rule 1) → niqréti
hiqrí' + nu → hiqrínu (rule 1) → hiqrénu
Note: In a small number of roots in the past base of פָּעַל the second vowel is always é, e.g., zamé', maléʾ.
5. Mute ל״י.
Inflectional Bases of Mute ל״י Paradigm. (See Table: Hebrew Grammar 6.) It is easy to see that the distinction between the past bases and the other bases in this paradigm lies primarily in the final vowel, which is á in the past and é in the other bases. The following rules specify what is characteristic of this paradigm:
(1) The suffix a of the third person feminine singular past changes into ta (and is then classed with suffixes beginning with a consonant).
(2) The infinitive is formed by the affixation of a special suffix ót to the base (and this is then classed with suffixes beginning with a vowel).
(3) The final vowel of all the bases is omitted before a suffix beginning with a vowel, and in such a case the accent is on the vowel of the suffix (this rule replaces rule 5 in 20):
baná + u → banú
haqné + i → haqní
l + naqqé + ót → lnaqqót
Note: To take account of "pausal forms," ancient inflectional bases must be considered, e.g., bakáy + u → bakáyu (rule 5 in 20).
(4) The vowel á in past bases changes to i when a suffix beginning with a consonant is added in bases of פִּעֵל, פָּעַל, and הִפְעִיל, and to e in other bases and occasionally in bases of פִּעֵל and הִפְעִיל:
baná + nu → banínu
niqqá + ti → niqqíti (or niqqéti)
hiqná + ta → hiqnéta (or hiqníta)
nibná + tém → nibnetém
In addition rules 1, 2, 3, 8 detailed in 20 apply to the inflection of this paradigm:
t + hibbané → tibbané (20, rule 1)
t + bné → tibné (20, rule 2)
ʾ + bné →ʾebné (20, rule 3)
ʾ + hibbané → ʾebbane (20, rule 3)
baná + tém → banitém (rule 4 of this paradigm) → bnitém (20, rule 8)
but l + hibbané + ót + m (objective pronoun) → lhibbanotám (without change of the vowel a of the base).
|Past||Participle||Future Imp. Inf.|
1: the influence of foreign works on hebrew syntax
1:1 the traditional grammars
1.2: neglect of syntax
1.3: conventional syntax in other languages
1.4: conventional hebrew syntax
2: THE PARTS OF SPEECH
2.1: the parts of speech and external reality
2.2: the classification into parts of speech
2.3: a criticism of the conventional classification
3: SENTENCE TYPES
3.1: structural analysis
3.11: Classification of sentences
3.12: The simple sentence
3.13: The multiple-unit sentence
3.14: The compound sentence
3.15: The complex sentence
3.2: pragmatic classification
3.21: The speaker's commitment
3.22: The speaker's attitude
4: THE PARTS OF THE SENTENCE
4.0: a syntactic framework
4.1: the division into the primary five parts
4.2: subject and predicator
4.21: Predicator and predicate
4.22: The copula
4.23: Nominal sentence and verbal sentence
4.24: Identifying sentence
4.25: Attributive sentence
4.26: Focusing sentence (extra-position)
4.27: Indefinite sentence
4.31: Morphological classification of adjuncts
4.32: Restrictive and nonrestrictive adjunct
4.33: Possessive pronoun as adjunct
4.34: The construct structure
4.341: classification of types of construct structure
4.35: Prepositional phrase adjunct
4.351: prepositional phrase adjunct as sentence remnant
4.352: prepositional phrase with verb transformed to noun
4.36: Adjunct before the "head" (center, nucleus)
4.38: Relative clause
4.41: Obligatory complement and obligatory preposition
4.42: Direct and indirect object
4.43: Transitive and intransitive verb
4.44: First and second object
4.45: Infinitive as object
4.46: Internal object
4.51: Circumstance adverbial
4.52: Types of conditional adverbial
4.53: Sentence adverbial
4.54: Prepositions as introducers of adverbials
5: DEPENDENCY WITHIN THE SENTENCE
5.11: Concord between subject and predicator
5.12: Predicator transformed into adjunct
5.13: Predicator transformed into circumstance adverbial
5.14: Lack of concord between subject and predicator
5.21: Pronoun concord in a focusing sentence (extra-position)
5.22: Pronoun concord in a relative clause
5.23: Copula concord
5.24: Other cases of concord
5.25: Lack of concord in adjunct sentences
5.3: determiner concord
5.4: restrictions on determination
5.6: the order of the parts of the sentence
6: LINKS BEYOND THE SENTENCE
6.1: anaphoric reference in sequentially related sentences
6.11: Personal pronouns
6.12: Demonstrative pronouns
6.2: fragmentary sentences
6.3: sentence connection
6:322: word order as indication of subordination
The first Hebrew grammarians devoted their attention chiefly to phonology and morphology, generally omitting special, ordered chapters on syntax. Some study was, however, made into the syntactical connections between adjacent or related words. This usually appears in traditional grammars when they deal with the system of accents. Many centuries passed before syntactic questions such as agreement of gender and number between different words in a sentence were first discussed comprehensively in a Hebrew grammar (Miqne Avram by Abraham de *Balmes, 1523).
This neglect seems to be due to the influence of the treatment of syntax in languages such as Latin, Greek, or Arabic, where the function of a word is generally shown by its form, and especially by the suffixes attached to it. The grammarian might therefore suppose that syntax essentially consists of such case suffixes, so that Hebrew, which lacks these suffixes, "has no syntax," or at least its syntax is not central to the language. The grammarian might therefore persuade himself that he should rather devote his energies to phonology and morphology. This conception of syntax continued to influence the treatment of Hebrew syntax even in the period of "scientific grammar," when philologists included a separate chapter on syntax in their Hebrew grammars. Until very recently, syntax was considered a study which attempted to reveal the logic behind language and thus external reality. For several centuries Latin was thought to be the most complete expression of logic and reality. Hence, while grammarians such as Gesenius, Ewald, or König and their modern counterparts such as Pereẓ and Segal treat Hebrew syntactic phenomena in great detail, their approach is not based on linguistic formal criteria derived from a study of Hebrew, but on categories of "reality" as reflected in Latin and as "laid bare" in Latin syntax.
It is well known that this defect has affected the treatment of syntax in other modern languages, including English. For example, grammarians have continued even recently to discuss the distinction between dative and accusative in English, as if the formal differences between them – noticeable in Latin but hardly at all in English – reflect relationships in external reality, and as if these relationships need to be considered in the syntax of every language, even where no distinction between them is made in the language.
Hebrew grammarians likewise saw Hebrew syntax as reflecting reality and the relationships existing in it, rather than as a formal study of the way words are linked and sentences are liked. For example, the Latin distinction between "direct object" and "indirect object" (see section Object below) is based upon the difference between a word that was an obligatory complement to a verb and was linked to it directly, i.e., without the word being preceded by a preposition, and a word which, while being an obligatory complement to a verb, needed a preposition before it. But in Hebrew what was called a "direct object" is under certain clear (and very frequent) conditions preceded by the preposition אֶת. "Direct object" in Hebrew, then, was applied not to a word with a certain status (or function) in the sentence, but to a word that designated a substance. That substance had a certain status in "reality" and had a certain relationship with another substance existing in the world. This relationship is realized by an action passed from this second substance to the one designated by the word which is "direct object." In other words, this syntax deals not with the grammatical relationships between words but with the relationships in the real world between what the words signify. A good illustration of this treatment of syntax appears in the comment usually quoted in the section dealing with the "direct object": "Sometimes the preposition לְ appears before the direct object instead of the preposition אֶת (for example in the biblical verse הָרְגוּ לְאַבְנֵר, II Samuel 4:30)." In this comment "direct object" is stated by the fact that a person is directly affected by the action, that is to say it is a person existing in the world that makes it "direct object" and not a linguistic relation existing in the sentence.
This conception of the word as reflecting reality is evident in what is traditionally the opening chapter of books on syntax: the chapter dealing with the "parts of speech." Since the early grammarians believed that reality was reflected best in Latin, it is precisely here that there is the greatest influence of Latin (and Greek) syntax. In these languages the function of a word can be recognized through its form, largely because of the many case suffixes that these languages have; hence, it was natural for the forms to serve as a basis for the treatment of functions. But in Hebrew it is exceptional for form and syntactic function to correspond, as in הַבַּיְתָה contrasting with הַבַּיִת. Only in the verb is there a regular correspondence, since it has merely one function, namely, to be the predicator of the sentence. It is reasonable to suppose that without the influence of foreign works on syntax the chapter on the "parts of speech" would have been the introduction to a treatment of morphology rather than of syntax. Indeed, the earliest medieval grammarians did include a discussion of the "parts of speech," which they then divided into three only: noun, verb, and particle (מִלָּה – literally "word"). They defined "noun" and "verb" semantically (for example, "a word denoting a substance or concept," "a word denoting an action or state"), while "particle" (comprising whatever was not regarded as "noun" or "verb") was defined by the function it had of linking other words. But, particles were also termed "sense words," since they supplied sense to the sentence. (See above 5. Parts of Speech in section Morphology above.)
The division into three parts of speech was preserved in Hebrew grammar even when the division into nine parts of speech, traditional in Greek and Latin (and also in modern languages), entered Hebrew grammar. The nine were grouped under the three earlier parts as follows:
a. noun: 1. noun (substantive) 2. adjective 3. numeral 4. pronoun
b. verb: 5. verb
c. particle: 6. conjunction 7. preposition 8. adverb 9. interjection
The division into nine parts of speech is also largely based on the meaning of words. That is to say, the assignment of a word to a particular part of speech is generally decided not by its formal features nor by its function, but by the concept it denotes. As a result, the classification suffers from several defects:
(1) Not every word denotes a concept. For example, conjunctions merely denote that words are linked to each other. In practice, therefore, grammarians define different parts according to different criteria: meaning, function, and sometimes even form.
(2) The meaning of a word depends on its context, and therefore the same word type is likely to be considered as belonging to several parts of speech, depending on the context of the particular word tokens. For example, in the sentence הַשּׁוֹמְרִים מְטַיְּלִים הַלַּיְלָה (The watchmen are walking around tonight) the first word is considered a noun and the second a verb, and the same applies to the sentence הַמְטַיְּלִים שׁוֹמְרִים הַלַּיְלָה (The hikers are on guard tonight). The third defect of this classification follows from the previous two:
(3) The division into parts of speech does not establish exclusive sets, since many words belong to more than one part of speech. It is this third defect in particular that has led some prominent linguists to deny any value to the classification into parts of speech. Yet this classification, virtually in its entirety, is generally accepted even in the most modern Hebrew textbooks, although it is clear that in many ways it does not fit the facts of the Hebrew language. Thus, many scholars claim that there is no basis for distinguishing in Hebrew between noun and adjective, since every adjective can be considered a noun (e.g., גִּבּוֹר, חָכָם), and clearly many nouns originally served as adjectives (e.g., חַמָּה, לְבָנָה). It is true that this claim is made particularly for biblical Hebrew, but it is true also for modern Hebrew. It applies to the forms of the participle, which can be taken as nouns, as adjectives, or as verbs. This last possibility is especially evident in modern Hebrew, where the forms of the participle are given in the verb paradigm, though formally they resemble nouns.
A sentence is a syntactical unit built from a word or words of which each one (or a combination of them) fulfills a specific syntactical function as a "sentence-part." This unit can stand by itself, can sometimes be connected to other similar units – whether preceding it or following it, and whether they are articulated by the same speaker or by others – and it is intonated in a manner which members of that language-group recognize as a complete articulated unit which does not lack a continuation. The sentence has an additional typical attribute: it is recursive, i.e., this unit can include in it a further internal sentence or sentences, each one of which fills a function as a sentence-part (see below). Unlike many other languages, the Hebrew sentence – apparently also in its deep structure (see below 3.1: Structural Analysis) – does not have to include a verb (see below 4.23: Nominal Sentence and Verbal Sentence). On the other hand the Hebrew sentence may be a single word which is a verb, since the Hebrew verb includes a pronoun. Transformational rules (see below Structural Analysis) are likely to influence the sentence and reduce it to a single word which is not a verb; however, this attribute can be found in many languages.
Modern linguistic theory considers grammar to be a set of generative rules for the language. A central place is occupied by what are called transformational rules. Transformational rules are also generative rules, but they apply to sentences or parts of sentences derived from simpler generative rules. Transformational rules change the order of words in a sentence, produce conjoinings and dependencies within sentences or between sentences, replace words by other words or formatives by other formatives, delete sentences or parts of sentences, etc. According to this theory, one must distinguish between the surface structure of a language and its deep structure. The latter includes the generative and transformational rules and also most of the semantic links that appear "on the surface." So far, only a few works have been written on Hebrew grammar according to this theory or under its influence, and even these follow its earlier formulations. Consequently it is neither possible to describe here the "deep structure" of Hebrew, nor to survey the transformational rules that operate in the language, except in those areas where a few details have been discovered. The following description is therefore essentially a survey of surface structure, taking into account works written according to the classical method.
This classification is usual in books on Hebrew syntax, except that some authors treat together the multiple-unit sentence and the compound sentence, while some do not treat the compound sentence at all, because "there is nothing to deal with in the compound sentence except what we find in its parts as separate sentences" (Segal).
A simple sentence is a sentencein which each part is realized by one word. This seems the best definition, even though there are some problems with it. For example, several adjectives may be attached to one noun to form a noun phrase. At first sight each one might be considered adjectival to the same noun, and yet in most cases the sentence will be regarded as simple, e.g., הַמְּעִיל הַשָּׁחוֹר הֶחָדָשׁ שֶׁלִּי נִקְרַע (My new black coat was torn). Another problem is that sometimes the function of the predicator is realized by a verb phrase, one word of which is the main verb while the rest are auxiliaries, e.g., הַשִּׁיעוּר מַתְחִיל לְהֵרָאוֹת מְעַנְיֵּן (The lesson begins to appear interesting). Here the three final verbs together realize the function of predicator. The problem of adjectives in a noun phrase can be solved by recognizing that sometimes an adjective is not attached directly to the noun, which forms the nucleus of the noun phrase, but to the whole of the preceding phrase – the phrase of noun + adjective. In the above example, the correct analysis for the constituency of the adjectives is הַמְּעִיל הַשָּׁחוֹר 1) הֶחָדָשׁ 2) שֶׁלִּי 3) נִקְרַע (3 (2 (1, compare the English equivalent (My (new (black coat))) was torn. First the words הַמְּעִיל הַשָּׁחוֹר (the black coat) form a phrase; then to this phrase as a unit the following adjective הֶחָדָשׁ is attached, forming the phrase הַמְּעִיל הַשָּׁחוֹר) הֶחָדָשׁ); finally שֶׁלִּי is added, relating to the whole of the preceding phrase הַמְּעִיל הַשָּׁחוֹר) הֶחָדָשׁ) שֶׁלִּי)).This explanation is based on one of the important principles of structural linguistics, "the theory of immediate constituents" (ic). A close examination reveals that each of the adjectives belongs to a different adjectival category, i.e., has a different function within the sentence. There are indeed many more parts of the sentence than is traditionally supposed (see below). The problem with auxiliaries is solved by considering them morphemes attached to the center of the predicator ("the main verb") to give it some modal or aspectual nuance (see below 4.45 Infinitive as Object). The auxiliaries also vary in their function. If they are regarded as realizing certain parts of the sentence, then they too are additional parts of the sentence.
The multiple-unit sentence is a sentence in which one of the parts is realized by several words linked to each other by parataxis (sometimes expressed by a conjunction), e.g., "אָבִיךָ וְאִמְּךָ דּוֹאֲגִים לְךָ" (Your father and mother are anxious about you), "הָבֵא כַּפּוֹת וּמַזְלְגוֹת" (Bring spoons and forks). Some exclude from this type such sentences as have verbs that are linked paratactically, e.g., "הוּא הִתְרַחֵץ, הִתְלַבֵּשׁ וְיָצָא לַעֲבוֹדָה" (He washed, dressed, and went out to work); "וַיׁאכַל וַיֵּשְׁתְּ וַיָּקָם וַיֵּלַךְ וַיִּבֶז" ("and he did eat and drink, and rose up, and went his way. So Esau despised his birthright," Gen. 25:34), maintaining that such a structure should be classed as a compound sentence (see 3.14: The Compound Sentence). The motivation for this view is that a construction containing a subject and a predicator is considered to be a sentence, and this definition applies to the verb, every form of which contains a subject-pronoun. Since a group of consecutive sentences linked paratactically is termed a compound sentence, the sentences in the above example should be considered compound sentences. In an analysis of the sentence (to be more precise, an analysis of the surface structure of the sentence) this approach is advantageous.
The compound sentence is traditionally subclassified according to the type of linking: addition, contrast, choice, or result. It is obvious that this classification is essentially semantic. Though there is a practical need for it, it is doubtful whether it has a place in a theoretical treatment in syntax, at least as long as syntax is concerned only with surface structure (but see below 6.31: Coordination in section Links beyond the Sentence).
The complex sentence is defined as a sentence one or more of whose parts is realized by a sentence (rather than by a word or a phrase). In every complex sentence there is therefore an embedded sentence. Since the embedded sentence performs, as a sentence, a function within the complex sentence, one can say that it is subordinate in the complex sentence. Some therefore define a complex sentence as follows: It is a sentence consisting of at least two sentences, which are linked by the subordination of one sentence to another. According to this view, the subordinated sentence is termed the dependent sentence, and the subordinating one the main sentence. But one should rather say that one sentence is a part of the other, and, in the same way as a sentence-element expressed by one word, it generally relates to one of the parts of the sentence as an independent unit. For example, if its function is that of adjunct, it is linked to the noun (irrespective of the noun's function), as in הַיֶּלֶד שֶׁרָאִינוּ אֶתְמוֹל חָזַר כְּבָר לְבֵיתוֹ (The boy, whom we saw yesterday, has already returned to his home) or רָאִיתִי אֶת הַיֶּלֶד שֶׁבָּא מִן הַכְּפָר (I saw the boy who came from the village). Similarly "וְשָׁמֲמוּ עָלֶיהָ אוֹיְבֵיכֶם הַיּוֹשְׁבִים בָּהּ" ("and your enemies that dwell therein shall be astonished at it," Lev. 26:32) as opposed to "אַתֶּם נִלְחָמִים אוֹתָם כַּכָה יַעֲשֶׂה ה׳ לְכָל אוֹיְבֵיכֶם אֲשֶׁר" ("for thus shall the Lord do to all your enemies against whom ye fight," Josh. 10:25). Adjunct sentences are divided into relative clauses, as in the above examples, and adjunct-content sentences to be discussed below, in 4.38: Relative Clause. If adverbial, such a sentence is linked to the verb-predicator, e.g., חָזַרְנוּ מִן הַטִּיּוּל לִפְנֵי שֶׁשָּׁקְעָה הַשֶּׁמֶשׁ (We returned from the hike before the sun set). Even when the subordinated part is subject, it can be said to be chiefly linked to the predicator, e.g., מִי שֶׁטָּרַח בְּעֶרֶב שַׁבָּת יׁאכַל בְּשַׁבָּת (Whoever exerts himself on the Sabbath eve will eat on the Sabbath). The last example demonstrates the inappropriateness of the term "dependent" for the subordinated sentence, since it can realize one of the main functions of the sentence, that of subject, and hence "main" is likewise inappropriate for the subordinating part of the complex sentence. For other types of sentence structure, see 4.23: Nominal and Verbal Sentences, 4.24: Identifying Sentence, 4.25: Attributive Sentence, 4.26: Focusing Sentence, and 4.27: Indefinite Sentence.
Other analyses of sentences may be made according to the speaker's commitment to what is being said or according to his attitude to what is being said.
The speaker's commitment is discernible from the form of the sentence: declarative, exclamatory (or optative), and interrogative ("yes-no" question or "specific" question, i.e., question specifying type of information required). In modern Hebrew "yes-no" questions generally differ from declarative sentences merely in intonation, though sometimes – particularly in the written language – a question may be prefaced by a word indicating that the sentence is a question, e.g., הַאִם or כְּלוּם. In any event, the structure of the sentence remains the same when it serves as a "yes-no" question. In a specific question the sentence is introduced by the appropriate interrogative word, e.g., מָתַי (when), אֵיךְ (how), מַדּוּעַ (why). See below for the order of words in such questions. The same applies to the exclamatory sentence. Any declarative or interrogative sentence can be considered an exclamatory sentence when rendered by an exclamatory intonation. Investigation into Hebrew syntax must include intonation to allow for such a classification of sentences. But so far no research in this field has been published. We must therefore be content with the general observation that for a Hebrew sentence to be interpreted as a question it must be said with a rising tone, particularly toward the end, and in any case the last syllable must be heard as being on a higher pitch than the penultimate. On the other hand, in a declarative sentence the last syllable is lower in pitch than the penultimate. Specific questions vary and it is difficult to state what their characteristic intonation patterns are. However, it is important to point out that specific questions can function exactly as they are, as embedded (subordinated) sentences in a complex sentence, and then obviously they do not have an interrogative intonation.
The attitude of the speaker toward what is said in the sentence or toward one of the details in it, and the extent of his belief in what is said, can be expressed in three ways: (1) parenthetically, e.g., הַסּוּס, יִמַּח שְׁמוֹ, מִתְגָּרֶה בִּי (The horse – damn it – is annoying me);
(2) by a verb, by a subordinating expression, or by a sentence, the sentence transmitting the main content being subordinated to them, e.g., יִתָּכֵן שֶׁמָּחָר יֵרֵד שֶׁלֶג (It is possible that tomorrow snow will fall), יְהִי רָצוֹן שֵׁתֵּלֵד אִשְׁתִּי זָכָר ("May it be [God's] will that my wife bear a male child"), or תָבוֹא שֶׁאֱלָתִי (וְ) מִי יִתֵּן ("Would (lit. who will give) that my desire be fulfilled"). See also Subordinators in 6.321: Links beyond the Sentence – אֲשֶׁר; (3) through certain auxiliaries (modal auxiliaries) that are attached to the nucleus of the predicator, e.g., מִסְפַּר הַמּוֹרְדִים עָלוּל לִגְדּׁל (The number of rebels may increase). See 4.45: Infinitive as Object.
A syntactic analysis of surface structure means the identification of a string of words as a sentence and the identification of the function in the sentence of each word or group of words. The process of identification and analysis will be better understood if the sentence is compared to an elastic frame that can be expanded as required. The frame contains a string of words and each word or group of words appears within an inner frame, a frame symbolizing a part of the sentence. This conceptual framework underlies the definitions given above of types of sentences. Identification of a word's function in a sentence means determining in which inner frame to put the word; identification of sentence type means recognizing the composition of the inner frames in the external, sentence frame. The structure of the sentence is illustrated as follows with each term designating a frame making up a part of the sentence:
Note: The order of the parts of the sentence given from right to left is not intended to represent their actual order. On this, see below. The frames shown here are filled with words. They are "elastic," that is to say they can "stretch" and contain more than one word. When one or more frames of the parts of the sentence is filled by a sentence, the sentence frame covering all the frames is termed a "complex sentence" and the sentence filling one of the parts of the sentence is termed an "embedded sentence." The features of an "embedded sentence" are generally the same as those of other sentences. An embedded sentence can itself contain within one of its internal frames another embedded sentence. This feature of the sentence – its recursiveness – allows for the possibility, at least theoretically, of expanding it to an infinite length.
Traditionally, Hebrew syntax distinguishes five parts of the sentence: (1) subject, (2) predicator, (3) adjunct, (4) object, (5) adverbial. It should be noted that this does not correspond to the division into subject and predicate, which is traditional in the grammars of many languages. On this, see 4.2: Subject and Predicator; 4.21: Predicator and Predicate. The first two parts are called the principal parts of the sentence, and the other three the subsidiary parts of the sentence or complements. The adjunct is complement to any noun whatever its function may be; the object and adverbial are complements to the predicator, but see 4.53: Sentence Adverbial. Some parts of the sentence are traditionally subclassified. A distinction must be made between two types of subclassification: (1) a part of the sentence is designated variously according to the nature of the words realizing it, e.g., the usual distinction between different kinds of adverbial: place, time, cause, result, etc.; (2) a part of the sentence is itself divided into two parts, each of which denotes a different syntactic functions, e.g., a predicator can be said to be composed of two parts: copula and predicator. (Some designate as extended predicator the part of the sentence comprising both of these.) The first type of classification is generally based on non-syntactic surface-structure features. For example, the distinction between place adverbial and time adverbial is determined merely by the meaning of the word filling the frame adverbial. A frame complementing a verb and filled by אֶתְמוֹל (yesterday) or אַחֲרֵי אַרְבָּעִים יוֹם (after 40 days) is called time adverbial, whereas if it is filled by כָּאן (here) or by בִּרְחוֹב פְּלוֹנִי (at X Street) it is called place adverbial. It is doubtful whether such a classification is relevant to the surface structure analysis, though obviously there are many occasions even here when it is necessary to make such distinctions (cf. 4.5: Adverbial). On the other hand, the classification of a part of the sentence into different internal parts is clearly relevant to all levels of syntax, since each part fulfills a different syntactic function and is distinct from the other parts of the sentence. For example, it is not enough to say that the phrase מַתְחִיל לְהֵרָאוֹת מְעַנְיֵן (begins to seem interesting) is the predicator in the sentence הַשִּׁעוּר מַתְחִיל לְהֵרָאוֹת מְעַנְיֵן (The lesson begins to seem interesting). To describe the internal composition of this part of speech: one part functions as predicator-nucleus (מְעַנְיֵן – interesting), while the others are attached to it, their function being to express the aspect (מַתְחִיל – begins) or the modality (לְהֵרָאוֹת – to seem) of the predicator-nucleus. These deserve attention from writers on syntax and an appropriate term, such as predicator-auxiliaries. Generative rules are needed for the ways in which the predicator-auxiliaries combine with the predicator-nucleus. Unfortunately, this area has not yet been sufficiently investigated in Hebrew. In the literature on Hebrew syntax there are only a few scattered remarks on such distinctions. In what follows each of the traditional parts of the sentence is surveyed in turn, with comments where possible on any subclassification.
In syntax it is usual to define these two parts of the sentence in relationship to each other. The justification for doing so is that what determines whether a word fulfills the function of subject is the existence of a relationship between that word and another word with the function of predicator in the sentence. This relationship called Nexus by Jespersen – whether it exists between words actually appearing in the sentence or whether it exists only in the deep structure of the sentence – is a necessary condition for sentence status. It is not, however, a sufficient condition, since some types of Nexus appear in a frame which is not a "sentence," though it is the consequence of a transformation applied to a sentence, e.g., הֲלִיכַת הָרוֹפֵא (the doctor's walk) derived from הָרוֹפֵא הָלַךְ (the doctor walked) or אֲנִי חוֹשֵׁב אוֹתוֹ לְחָכָם (I consider him wise) the last two words of which are derived from הוּא חָכָם (He is wise). It is usual to define subject and predicator semantically, e.g., "The subject is the word denoting the substance spoken about in the sentence, the predicator is what is said about this substance." However, the question that the speaker is posing is not always amenable to an unequivocal answer. Moreover, sometimes it is clear that what is being spoken about in the sentence is not denoted by the word that is subject, but by a word with a different syntactic function. For example, in the sentence חַם לָהּ (She is warm, literally, Warm is to her) the topic of the sentence is third person singular feminine, but the corresponding pronoun is not the subject of the sentence. As elsewhere in syntax, one ought to use formal rather than semantic criteria to define "subject," "predicator," and the other parts of the sentence. If a straight definition (such as "The subject is…") seems too difficult, we can define the parts of the sentence operationally. The following is an example of such an operational definition of subject and predicator (following Ornan, The Syntax of Modern Hebrew): If one has a word that by itself constitutes a sentence, and if (1) the word is a verb, and one can substitute for it a combination of that verb and a subjective pronoun – הוּא (he), הִיא (she), etc., agreeing with it in gender and number, and this combination is likewise a sentence, then in this new sentence-frame the function of the subjective pronoun is termed "subject" and that of the verb is termed "predicator"; or if (2) the word is a noun, and one can substitute for it a combination of that noun and the verb הָיָה (be) agreeing with the noun in gender and number, and this combination is likewise a sentence, then in this sentence-frame the function of the noun is termed "subject" and that of the verb is termed "predicator." (On concord, see 5.11: Concord Between Subject and Predicator.) Substitution is an important factor in this definition. Indeed, it is a central principle in structural linguistics. According to this principle, words, phrases, or parts of words that are substitutable within a given frame, form a grammatical class. The above definition can be extended by the method of substitution to include all the words or phrases filling the function of subject in the given frame and all those filling the function of predicate. For that purpose the above definition must be supplemented: "Any word, or group of words, that can replace a word filling the function of subject, likewise fills the function of subject, provided that the resultant sentence does not thereby become deviant." A corresponding addition can be made for the definition of the predicator.
As noted above the division of the sentence usual in Hebrew syntax differs from that usual in the grammars of many languages, though Hebrew grammarians have not sufficiently considered the difference. In particular, a distinction should be made between the Hebrew concept נָשׂוּא (predicator) and the general concept "predicate." The parts of the sentence in the predicate are the predicator, the object, the adverbial, and any adjunct to these parts. The predicator is the nucleus of the predicate, with all the other parts in the predicate the complements of the predicator.
The predicator itself can be expressed by more than one word. Modal or aspectual predicator-auxiliaries were mentioned above (in 4.1: The Division into the Primary Five Parts, cf. 4.5: Adverbial). To these should be added the past and future forms of the verb הָיָה (be), since when the predicator is expressed by a noun or participle these may be combined with it to denote time, e.g., וְקַיִן הָיָה עוֹבֵד אֲדָמָה (Cain was a tiller of the earth). In this use the verb הָיָה fills the function of copula, which is also considered a predicator-auxiliary. Similarly, the forms of the third person pronoun – הֵן, הֵם, הִיא, הוּא – are used as copulas. This type of copula is used for emphasis (but 5.23: Copula Concord). Some consider the negative word אין as a copula, since "like the third person pronouns" it can be combined only with a noun or a participle (see below). However, it is more correct to consider as copula only the pronominal attached to this negative word. Thus, it is true that there is a copula in the sentence אֶסְתֵּר אֵינָהּ מַגֶּדֶת (Esther does not tell), but it is the pronominal suffix in אֵינָהּ (literally she-not), while in אֵין אֶסְתֵּר מַגֶּדֶת there is no copula. It should also be noted that הֵן, הֵם, הִיא, הוּא can be combined with the predicator even when it is a verb in the past or future, e.g., עֲצַת יְהׁוָה הִיא תָּקוּם (see 4.25: Attributive Sentence; 4.26: Focusing Sentence; 5.23: Copula Concord).
It is usual in Hebrew syntax to distinguish between nominal sentences and verbal sentences according to whether the predicator is a noun or verb. This distinction was borrowed from Arabic syntax, but in Arabic it depends on the first word of the sentence: if it is a noun, the sentence is nominal; if it is a verb, the sentence is verbal. Opinions differ when the predicator in Hebrew is a participle. In earlier Hebrew the participle was regarded as a noun and hence a sentence whose predicator was a participle was considered a nominal sentence. However, in modern Hebrew the status of a participle having the function of a predicator is identical with that of a verb, and consequently it is doubtful whether it is correct to consider such a sentence in modern Hebrew as a nominal sentence. Opinions also differ when the predicator consists only of a prepositional phrase, as in הַיֶּלֶד בַּבַּיִת (The boy [is] in the house). Generally, books on Hebrew syntax assign such sentences to the class of nominal sentences. Some maintain that a sentence whose predicate is a prepositional phrase has no predicator and therefore it cannot be a nominal sentence, but instead should be termed a verbal sentence without a predicator. The presence or absence of predicator (expressed by a verb) is the sole difference, according to this view, between these sentences and sentences such as הַיֶּלֶד יָשַׁב בַּבַּיִת (The boy sat in the house), הַיֶּלֶד הָיָה בַּבַּיִת (The boy was in the house). The word בַּבַּיִת (in the house) serves in the sentences exactly the same function of complement to the predicator (in this instance, adverbial).
Nominal sentences (in the restricted meaning of the term) where the state of determination of the subject and predicator is the same – whether they are both determined or both undetermined – are called equative or identifying sentences. With such sentences, e.g., מִלְחָמָה הִיא מָוֶת (War is death), it is sometimes impossible to decide which is subject and which is predicator except by the context. At all events, each of the parts identifies the other, the predicator being called the identifying predicator. Of particular interest are cases where the second part of the sentence is realized by a subordinate sentence, e.g., יְהׁוָה אֱלֹהֵיכֶם הוּא הַנִּלְחָם לָכֶם (The Lord your God is the one who fights for you). There is no basis for the view that in such a structure the first part יְהׁוָה אֱלֹהֵיכֶם is always the subject and the second part הַנִּלְחָם לָכֶם is the predicator. On the contrary, the first part usually has the function of predicator.
When the subject is determined and the predicator is undetermined, the predicator's function is to attribute what is denoted in the subject to the class possessing the characteristic denoted by the predicator, e.g., in יוֹסֵף הוּא פָּקִיד (Joseph is an official) the attribution is to the class of officials. Such a predicator is termed an attributive predicator.
This last structure formally belongs to the focusing sentence structures, but this term is usually assigned to sentences such as הַוָּתִיקִים – אִישׁ אֵינוֹ שָׂם לֵב אֲלֵיהֶם (The veterans – nobody pays any attention to them) or סַבָּא – הוּא אֵינוֹ מִתְעַיֵּף לְעוֹלָם (Grandpa – he is never tired). The structure of such a sentence is explained in current syntax as the result of a transformation from another sentence in which the first word of the focusing sentence appears in the second part of the sentence in place of the pronoun that agrees with it in gender and number: אִישׁ אֵינוֹ שָׂם לֵב אֶל הַוָּתִיקִים (Nobody pays any attention to the veterans), סַבָּא אֵינוֹ מִתְעַיֵּף לְעוֹלָם (Grandpa is never tired). With a focusing sentence (in the restricted sense of the term) the first word, or first endocentric phrase, is always to be considered the subject of the sentence, while whatever comes after it is the predicate. The predicate itself is an embedded sentence, and hence the focusing sentence is always a complex sentence. Others maintain that the focusing sentence is merely a simple sentence with a change in the order of the words. In any case, all agree that the noun appearing initially in a focusing sentence is very much more emphasized than it would be in a simple sentence.
It is worth noting that there are sentences without a subject, in particular where the predicator-nucleus is realized by an infinitive linked to a modal auxiliary, e.g., אֶפְשָׁר לְהַבְחִין בְּכָךְ מִיָּד ([It is] possible to discern it immediately), cf. 4.45: Infinitive as Object. However, many will argue that אֶפְשָׁר (possible) alone is predicator, and the string of all the other words in the sentence is the subject. In any case, this sentence is an indefinite sentence, that is to say a sentence whose understood subject is any man or men in general.
The adjunct differs from the other parts of the sentence in that by definition it cannot serve as nucleus for another part of the sentence, nor can it be linked to any part except a noun, irrespective of what function the noun fills in the sentence. Any word to which an adjunct serves as a nucleus, is considered in Hebrew syntax also as an adjunct.
From a morphological point of view, seven types of adjunct can be distinguished:
(1) attributive adjunct; (2) possessive pronoun adjunct (whether affixed or independent); (3) adjunct in the construct case; (4) prepositional phrase adjunct; (5) adjunct before nucleus; (6) appositive; (7) embedded sentence. These types are exemplified as follows:
(1) הָאִישׁ הַזָּקֵן הֶאֱזִין בְּסַבְלָנוּת (The old man listened patiently);
(2) הַכּוֹבַע שֶׁלִּי נָפַל לַמַּיִם (My hat fell into the water);
(3) קִירוֹת הַבַּיִת מְכֻסִּים אֵזוֹב (The walls of the house are covered with moss);
(4) הַזָּקֵן מִנַּהֲרַיִם הֵקִים אֶת הַמִּפְעָל (The old man from Nahrayim set up the enterprise);
(5) שְׁלוֹשָׁה סוּסִים דּוֹהֲרִים (Three horses are galloping);
(6) רְאוּבֵן, הַבְּכוֹר, יָרַד לְמִצְרַיִם (Reuben, the firstborn, went down to Egypt);
(7) הַנַּעַר, שֶׁלֹּא יָדַע בֵּין יְמִינוֹ לִשְׂמׁאלוֹ, הִסְכִּים בְּרָצוֹן (The lad, who could not distinguish between his right and left, agreed willingly).
All these types of adjunct appear to be transformed from other structures. With uncertainty as to the source of adjuncts of type (5), all have their source in the predicator (cf. 5.12: Predicator Transformed into Adjunct) or, in some cases, in another part of the predicate. Thus, the following set of sentences can be seen as the source of the adjuncts in the above examples:
(1) הָאִישׁ זָקֵן; הוּא הֶאֱזִין בְּסַבְלָנוּת (The man is old; he listened patiently);
(2) יֵשׁ לִי כּוֹבַע; הוּא נָפַל לַמַּיִם (I have a hat; it fell into the water);
(3) לַבַּיִת יֵשׁ קִירוֹת; הֵם מְכֻסִּים אֵזוֹב (The house has walls; they are covered with moss);
(4) הַזָּקֵן גָּר בְּנַהֲריִם (קָשׁוּר בְּנַהֲרַיִם); הוּא הֵקִים אֶת הַמִּפְעָל (The old man lived in Nahrayim (he is connected with Nahrayim); he set up the enterprise);
(6) ראוּבֵן הוּׁא הַבְּכוֹר; הוּא יָרַד לְמִצְרַיִם (Reuben is the first-born; he went down to Egypt);
(7) הַנַּעַר לֹא יָדַע בֵּין יְמִינוֹ לִשְׂמׁאלוֹ; הוּא הִסְכִּים בְּרָצוֹן (The lad could not distinguish between his right and left; he agreed willingly). At present it is not clear what the source is for an adjunct denoting quantity. On concord with the adjunct, see 5.12: Predicator Transformed into Adjunct, and 5.3: Determiner Concord.
Only a few works dealing with Hebrew syntax mention the distinction between restrictive adjunct and nonrestrictive adjunct, sometimes merely to indicate that a nonrestrictive adjunct "is not an adjunct." An example of a restrictive adjunct would be if a man having three sons and wanting to say something about the eldest says בְּנִי הַגָּדוֹל לוֹמֵד כְּבָר בָּאוּנִיבֶרְסִיטָה (My grownup son is already studying at the university). The function of the word הַגָּדוֹל (grown-up) is that of restrictive adjunct, distinguishing this son from the others. An example of a non-restrictive adjunct would be if a man with one son wants to say something about him and wants incidentally to mention that he is grown-up; he says בְּנִי הַגָּדוֹל לוֹמֵד כְּבָר בָּאוּנִיבֶרְסִיטָה (My grown-up son is already studying at the university). The function of the word הַגָּדוֹל (grown-up) is then that of nonrestrictive adjunct. This distinction is important, and in practice has also a formal expression, particularly in intonation, but sometimes also in punctuation. There is no pause between the nucleus and a restrictive adjunct: the pitch of the latter rises slightly, and it has greater stress. On the other hand, there is a slight pause between the nucleus and a nonrestrictive adjunct: the pitch of the latter falls slightly, and it has a lighter stress. If the nonrestrictive adjunct is long, it is usual to put a comma before it. Usually there is no comma before a restrictive adjunct, even when it is a subordinate embedded sentence (despite the official rules for punctuation, which require a comma before every adjunct that is a sentence).
The structural ambiguity of the adjunct can be explained in transformational grammar. The adjunct is transformed from the predicate of another sentence, this sentence being the continuation of a preceding sentence. The same subject serves these two sentences, both of which are deleted by a deletion transformation and hence do not appear in the text. But the subject of the second sentence does not always refer to the same quantity of substances or material that the subject of the first sentence refers to. When it refers to a lesser quantity, the adjunct in the transformed sentence is a restrictive adjunct; when it refers to the same quantity, the adjunct is nonrestrictive. The sources of the above examples are therefore in the following two sets of sentences:
(a) He has three sons. יֵש לוֹ שְׁלוֹשָׁה בָּנִים
One of them is grown-up. אֶחָד מֵהֶם גָּדוֹל
He is studying at the university. הוּא לוֹמֵד בָּאוּנִיבֶרְסִיטָה
His grown-up son is studying at the university (restrictive adjunct). בְּנוֹ הַגָּדוֹל לוֹמֵד בָּאוּנִיבֶרְסִיטָה
(b) He has one son. יֵשׁ לוֹ בֵּן אֶחָד
He is grown-up. הוּא גָּדוֹל
He is studying at the university. הוּא לוֹמֵד בָּאוּנִיבֶרְסִיטָה
His grown-up son is studying at the university (nonrestrictive adjunct). בְּנוֹ הַגָּדוֹל לוֹמֵד בָּאוּנִיבֶרְסִיטָה
Some explanatory comments on several of the types of adjuncts enumerated above are called for. The possessive pronoun as adjunct: in "deep grammar" its source is in a sentence denoting possession, e.g., יֵשׁ לוֹ אָח (He has a brother) → אָחִיו or הָאָח שֶׁלּוֹ (his brother). The possessive pronoun affix, e.g., in אָחִיו, and the independent possessive pronoun, e.g., שֶׁלּוֹ, are not entirely free variants, but sometimes the appearance of one or the other is conditioned (see below). It is worth noting that in written Hebrew the use of the affix is between ten and fifteen times more frequent than the use of the independent form. As far as can be ascertained from the few studies in this area, the use of the affix is greater in spoken Hebrew, but more substantial studies are required before one can establish the relative frequency with any certainty.
The most obvious conditions favoring the appearance of the independent form of the possessive pronoun are the following:
(1) when a second possessive pronoun is used to emphasize an affixed possessive pronoun, e.g., כַּרְמִי שֶׁלִּי לֹא נָטַרְתִּי (I did not tend my own vineyard);
(2) when the nucleus is a proper noun, e.g., יִטֹּשׁ אֶת מֶנְדֶלִי שֶׁלּוֹ לֹא יַעֲזׁב וְלֹא (He will not desert his Mendele);
(3) generally with a foreign or borrowed word, e.g., הַטֶּלֶפוֹן שֶׁלּוֹ מְצַלְצֵל (His telephone is ringing);
(4) with a noun-numeral, e.g., בִּשְׁנוֹת הַשְּׁלוֹשִׁים שֶׁלּוֹ (In his thirties);
(5) with a noun in the construct state, e.g., דִירַת הַשְּׂרָד שֶׁלּוֹ (His official residence);
(6) with a word that was not originally a noun, e.g., אַתָּה מוּכָן לְהָפֵר אֶת הַ״בְּרׁגֶז״ שֶׁלְּךָ? (Are you ready to cancel your anger?);
(7) with a phrase that is used metaphorically, e.g., בְּאַרְבַּע אַמּוֹת שֶׁלָּנוּ גַּם (Even within our "four cubits");
(8) when the nucleus has two meanings and the rarer meaning is intended, e.g., בָּאַדְמוֹרִים =) הָעִיר הִתְפַּרְסְמָה בַּצַּדִּיקִים שֶׁלָּהּ) (The city was famous for its "pious men" = ḥasidic rabbis);
(9) when the nucleus is used euphemistically, e.g., !זׁאת מַצִּיעִים לִי הַ״יְּדִידִים״ שֶּׁלִּי (My "friends" suggest it to me!). Haim Rosén (see bibliography) has argued that the difference in usage between the two forms corresponds to the difference between inalienable possession (e.g., the family relationship or the parts of the body) and alienable possession. This proposal seems dubious.
An endocentric phrase consisting of nouns, or words that have nominal function, the order of which cannot be changed without changing the meaning of the phrase, is said to be in the construct state. This phrase may be in three structures:
(1) Close construct state, when two nouns are linked without interruption (except for the definite article), e.g., בֵּית הָאִישׁ (The man's house). On the changes in form of the first noun, see above, Morphology. The second noun does not change.
(2) Loose construct state when the word שֶׁל interrupts between the two nouns, e.g., הַבַּיִת שֶׁל הָאִישׁ (The house of the man).
(3) Reduplicated construct state, when a possessive pronominal affix agreeing in gender and number with the second noun is attached to the first noun, and the word שֶׁל is put between the nouns, e.g., בֵּיתוֹ שֶׁל הָאִישׁ (The (his) house of the man). The two last structures are termed dismembered construct states. There are other ways as well of making the construct state discontinuous, for example by the preposition לְ or מִן, e.g., צִנְצֶנֶת מִזְּכוּכִית (A glass container). It is difficult to say under what conditions the three types of construct states are in free variation and when one of them must be used. But it is clear that there are certain phrases that can only be used in one type of construct state, e.g., חֲנֻכַּת הַבַּיִת (the inauguration of the home), זוּטוֹ שֶׁל יָם (the floor of the sea). With other phrases, the meaning changes if a different type is used, e.g., זֶה דְּבַר הַמְּפַקֵּד (This is the message of the commander), זֶה דָּבָר שֶׁל הַמְּפַקֵּד (This item belongs to the commander); בֶּן עֶשְׂרִים (twenty years old), בְּנָם שֶׁל עֶשְׂרִים (the son of twenty). In a construct state consisting of two words the nucleus is usually the first word, while the second word is the adjunct, e.g., עֲבוֹדַת אֱלִילִים (the worship of idols). For other possibilities, see 4.36: Adjunct before the "Head." If it consists of three or more words, usually the second and later words are each adjunct to the immediately preceding word, and the combination is in turn adjunct to the immediately preceding word. For example in the phrase עֲבוֹדַת אֱלִילֵי זָהָב (the worship of idols of gold), זָהָב (gold) is adjunct to אֱלִילִים (idols) and אֱלִילֵי זָהָב (idols of gold) is adjunct to עֲבוֹדָה (worship):
In an analysis of surface structure, the words in construct state are classified semantically; in deep grammar they can be classified according to the function they perform in the underlying structure from which the construct state has been transformed.
The following are the chief meanings attributed to the nomen rectum (the last noun of the construct phrase; following Pereẓ): (1) the owner of what is denoted by the nomen regens (the last but one of the phrase), e.g., גַּן הָאִכָּר (the farmer's garden); (2) the material from which is made what is denoted by the nomen regens, e.g., כְּלֵי כֶּסֶף (vessels of silver); (3) the genus of what is denoted by the nomen regens, e.g., עֲצֵי שִׁטִּים (trees of acacia wood); (4) the characteristic of the nomen regens, e.g., לְשׁוֹן שֶׁקֶר (an expression of falsehood); (5) the limit of application for the characteristic expressed in the nomen regens, e.g., נְקִי כַּפַּיִם (clean of hands); (6) the content of the nomen regens, e.g., סִפְרֵי מוּסָר (books of ethics); (7) the agent of the action expressed as a verbal noun in the nomen regens, e.g., נְשִׁיכַת שׁוּעָל (the bite of a fox); (8) the object of the action expressed as a verbal noun in the nomen regens, e.g., הַדְלָקַת נֵר (the lighting of a candle); (9) the instrument used for the result expressed in the nomen regens, e.g., שְׂרוּפוֹת אֵשׁ (burnt by fire); (10) the place of the nomen regens, e.g., אַרְזֵי לְבָנוֹן (cedars of Lebanon); (11) the time of the nomen regens, e.g., חֲזוֹן לַיְלָה (the vision at night); (12) the cause for the fact in the nomen regens, e.g., חוֹלַת אַהֲבָה (sick through love); (13) the result of the nomen regens, e.g., גִּשְׁמֵי בְּרָכָה (rains of blessing); (14) the purpose of the nomen regens, e.g., מִזְבַּח קְטׁרֶת (altar of incense). In addition, sometimes the nomen rectum denotes the name of the nomen regens, e.g., נְהַר פְּרָת (the river of Euphrates), and sometimes it emphasizes the nomen regens or its quantity by repetition of the same word in the plural, e.g., עֶבֶד עֲבָדִים (slave of slaves). When the nomen regens is בַּעַל (master) or אָדוֹן (lord) the rectum is its property (the converse of 1 above). According to transformational theory every phrase in the construct state is a transformation from another structure, for example (cf. Ornan):
(The farmer has cows) פָּרוֹת־הָאִכָּר → לָאִכָּר יֵשׁ פָּרוֹת
(The fish has a sword – its characteristic) דַּג־הַחֶרֶב → לַדָּג יֵשׁ חֶרֶב
(The pen is made of iron) עֵט־בַּרְזֶל → הָעֵט עָשׂוּי מִבַּרְזֶל
(The moon is in the form of a sickle) חֶרְמֵשׁ־הַיָּרֵחַ → הַיָּרֵחַ הוּא בְּצוּרַת חֶרְמֵשׁ
(The land is called "Canaan") "אֶרֶץ כְּנַעַן → הָאָרֶץ נִקְרֵאת "כְּנַעַן
(The Jordan flows through the valley) בִּקְעַת־הַיַּרְדֵּן → הַיַּרְדֵּן זוֹרֵם בְּתוֹך בִּקְעָה
(The child is healthy) בְּרִיאוּת הַיֶּלֶד → הַיֶּלֶד בָּרִיא
(Jackals howl) יִלְלַת תַּנִּים → תַּנִּים מְיַלְּלִים
(The dog (habitually) guards the flock) כֶּלֶב צׁאן→ הַכֶּלֶב שׁוֹמֵר אֶת הַצּׁאן
(The old man walks in the morning) טִיּוּל הַבּׁקֶר (שֶׁל הַזָּקֵן) → הַזָּקֵן מְטַיֵּל בַּבּׁקֶר
Since the structure of the construct state can be transformed from a large number of different sources, it is clear that it has a large number of possible meanings. However, in practice many meanings are ruled out, since the speaker knows the meanings of the words and the context in which the structure appears. There is in fact no ambiguity in a construction such as פָּרוֹת הָאִכָּר (the farmer's cows) since the meaning of the words allows only one possible interpretation, namely that the farmer is owner of the cows (and not the reverse, for example). Nevertheless, there are instances where the construct state is ambiguous. This might arise, e.g., when the nomen regens is a verbal noun derived from a transitive verb, since the rectum can then be agent of the action (subject in a background sentence) or recipient of the action (object in a background sentence). For example, בִּחִירַת הַנָּשִׂיא (the choice of the president) could be interpreted as a transform of הַנָּשִׂיא בָּחַר (The president chose) or מִישֶׁהוּ) בָּחַר אֶת הַנָּשִׂיא) (X chose the president). It has been claimed that in such cases the close construct state is selected for one meaning and the loose one for the other, בְּחִירַת הַנָּשִׂיא being interpreted solely as "X chose the president" while the transform of "the president chose" would be הַבְּחִירָה שֶׁל הַנָּשִׂיא (cf. Haim Rosén in the bibliography). An examination of considerable material drawn from newspapers and modern literature does not support the claim.
An adjunct realized as a prepositional phrase is the result of one or the other of two transformations:
(1) It may be a remnant of a sentence in which it functioned as an adverbial. When the sentence is embedded in another sentence, the subject and predicator are deleted and the adverbial becomes an adjunct to the subject of the other sentence. This is illustrated in the following example:
It is worth demonstrating how this explanation appears in the usual formulation of the generative-transformationalists. The two sentences are first placed one after the other:
Representation of הָאוֹטוֹבּוּס יְאַחֵר לָצֵאת
(The bus will leave late): npi +vpj
Representation of הָאוֹטוֹבּוּס נוֹסֵעַ לִירוּשָׁלַיִם:
(The bus travels to Jerusalem): npi + vpk + pp
npi + vpj + npi + vpk + pp
(1) (2) (3) (4) (5)
However, under certain conditions (e.g., when the subjects of the two sentences are two instances of the same nominal structure, if the referent of the subject is identical in both sentences), a transformation applies which changes the order of the words. In place of the order given above (left to right) the words are ordered (1) (3) (4) (5) (2), i.e:
npi + npi + vpk + pp + vpj
The second sentence is parenthetically included, as it was, in the first: הָאוֹטוֹבּוּס (הָאוֹטוֹבּוּס נוֹסֵעַ לִיִרוּשָׁלַיִם) אֵחֵר לָצֵאת. Now an obligatory transformation applies, which deletes the second instance of np, and adds instead a relative (שֶׁ, אֲשֶׁר, or הַ). The result is npi + še + vpk + pp + vpj, i.e., הָאוֹטוֹבּוּס שֶׁנּוֹסֵעַ לִירוּשָׁלַיִם אֵחֵר לָצֵאת. However, this sentence can again be transformed as follows:
|Structural Description:||npi +||še +||vp +||pp +||vpj|
The relative ש and the internal predicator vp are deleted, leaving only npi + pp + vpj: הָאוֹטוֹבּוּס לִירוּשָׁלַיִם אֵחֵר לָצֵאת
(2) The prepositional phrase is a remnant as before. However, the verb which is complemented is not deleted, but transformed into a noun, the phrase changing from complement of a verb to complement of a noun, that is to say an adjunct. For example:
הֲלִיכָתוֹ לְבֵית הַסֵּפֶר → הוּא הָלַךְ לְבֵית הַסֵּפֶד
הַטִּיּוּל בַּבּׁקֶר → הַיְּלָדִים טִיְּלוּ בַּבּׁקֶר
We should treat as a special case prepositional phrase adjuncts introduced by מִ =) מן) (from) when they designate the place of origin or of action of what is denoted by the noun serving as nucleus, e.g., הַזָּקֵן מִנַּהֲרַיִם (the old man from Naharayim), הַמְכַשֵּׁפָה מִפָּרִיס (the witch from Paris). It is not clear what is the source sentence from which these adjuncts are transformed.
A pre-nucleus adjunct generally denotes quantity, and it comprises cardinals, dividers, measures, and words such as הַרְבֵּה (much), רׁב (the majority of), שְׁאָר (the rest of), קְצָת (a little), מִבְחַר (the best of).
Cardinals agree with the nucleus in gender, and in the case of units of measurement, also in number, e.g., חֲמִשָּׁה אֲנָשִׁים (five men), חָמּשׁ אַמּוֹת בַּד (five cubits linen), שְׁנֵים עָשָׂר דּוּנָם אֲדָמָה (twelve dunams of land). The numbers 3–10 are likely to be in the construct state before the nucleus, especially when the latter is determined, number 2 generally so, while number 1 appears after the nucleus when the latter is singular (אִישׁ אֶחָד), and in the construct state before the nucleus when it is determined and plural (אַחַד הָאֲנָשִׁים).
Nouns for containers, such as בַּקְבּוּק (bottle), פַּח (can), may serve as measuring units for liquids or for bulk solids such as קֶמַח (flour), or פֵּרוֹת (fruit), provided the reference is to massproduced vessels of fixed size, e.g., הֶחָבִית מְכִילָה שְׁלוֹשָׁה פַּחִים (The barrel contains three cans), אַרְבַּע תֵּבוֹת תַּפּוּזִים (four boxes of oranges). It has not yet been established whether there is in modern Hebrew a systematic semantic difference (as Haim Rosén has claimed) between measures appearing in a close construct state, e.g., שְׁנֵי שַׂקֵּי קֶמַח (two sacks of flour), in a loose construct state, e.g., שְׁנֵי שַׂקִּים שֶׁל קֶמַח, or in apposition, e.g., שְׁנֵי שַׂקִּים קֶמַח. Similarly nouns for shapes, provided reference is to shapes with a more or less fixed size, can serve as measures for solids, e.g., שְׁנֵי כִּכְּרוֹת לֶחֶם (two loaves of bread).
Two nouns one of which has the function of adjunct to the other, but without their being in the construct state relationship, are said to be in apposition. An appositive is transformed from a noun predicator. If the predicator from which it is transformed functioned as identifier in an identifying sentence (cf. 4.24: Identifying Sentence), the appositive also functions as an identifying appositive, e.g., מִרְיָם הִיא הָאָחוֹת הָרָאשִׁית (Miriam is the matron). בְּחֻפְשָׁה הִיא נִמְצֵאת (She is on leave) → מִרְיָם, הָאָחוֹת הָרָאשִׁית, נִמְצֵאת בְּחֻפְשָׁה (Miriam, the matron, is on leave). When the predicator is attributive (cf. 4.25: Attributive Sentence), the appositive is an attributive appositive, e.g., לֵוִי הוּא ד״ר לְמִשְׁפָּטִים (Levi is a doctor of law). הוּא הִתְמַנָּה לְמַרְצֶה (He has been appointed lecturer) → לֵוִי, ד״ר לְמִשְׁפָּטִים, הִתְמַנָּה לְמַרְצֶה (Levi, a doctor of law, has been appointed lecturer). Books on syntax generally note that the appositive follows the nucleus. Hence, in שְׁלֹמׁה הַמֶּלֶךְ (Solomon the king), הַמֶּלֶךְ is said to be appositive, while in הַמֶּלֶךְ שְׁלֹמֹה (King Solomon), the proper noun שְׁלֹמֹה is said to be appositive. But on the basis of the semantic identity of the two phrases, it has been proposed that an attributive noun denoting status, occupation, or title that is attached to a proper noun should be considered an appositive even when it precedes the proper noun, e.g., הד״ר לְמִשְׁפָּטִים לֵוִי הִתְמַנָּה לְמַרְצֶה (Doctor of Law Levi has been appointed lecturer). In such cases the appositive has a determiner.
As with the adjectival adjunct, all appositives can be divided into restrictive and nonrestrictive. Other types of appositives are appositional compounds, e.g., כֻּרְסָה = מִטָּה (divan bed) and quantifying apposition, e.g., שְׁלוֹשָׁה אֲנָשִׁים (three men). Certain introductory expressions appear before identifying appositives, e.g., כְּלוֹמַר (that is to say), דְּהַיְנוּ (that is), בְּיִחוּד (especially). Another characteristic of the identifying appositive is that the preposition before the nucleus is sometimes repeated before the following appositive. It seems that this only applies when the appositive is nonrestrictive. Such a repetition is obligatory when the nucleus is a pronoun and the following appositive is a noun, e.g., אָמְרוּ צָלָיו צַל רַבִּי צֲקִיבָא (They said about him, about Rabbi Akiva).
An adjunct sentence is a transformation of a complete predicate and not just of a predicator, e.g., הַמַּחֲזֶה הָצֳלָה אֶמֶשׁ לָרִאשׁוֹנָה (The play was put on last night for the first time). הוּא וַדַּאי יִזְכֶּה לִהַצְלָחָה (It will certainly meet with success) → הַמַּחֲזֶה שֶׁהָצֲלָה אֶמֶשׁ לָרִאשׁוֹנָה וַדַּאי יִזְכֶּה לְהַצְלָחָה (The play which was put on last night for the first time will certainly succeed). The condition for this transformation is that the noun appearing in one sentence will also appear in the other sentence. In the transformed adjunct sentence this noun is deleted and is replaced by a pronoun agreeing with it in gender and in number. However, if the noun functioned as subject of the sentence before its transformation, then generally gender and number concord with the predicator is sufficient and a pronoun is not inserted. See further on this, 5.22: Pronoun Concord in a Relative Clause. The indicators of the subordination of the adjunct are שֶׁ and אֲשֶׁר (virtually in free variation), and ה under certain conditions (cf. 6.32: Subordinators in Links beyond the Sentence). Adjunct sentences may also be asyndetic, cf. 6.322: Word Order in Links Beyond the Sentence. Another category of adjunct sentence must also be distinguished, namely the adjunct content-sentence, e.g., הַהַשְׁצָרָה שֶׁיֵּשׁ חַיִּים צַל הַמַּאְדִּים נִתְבַּדְּתָה (The supposition that there is life on Mars has been proved false). The source for the content-adjunct is not in the predicate of a preceding sentence but in the object sentence of a preceding sentence. It is formed when the verb complemented by the object is converted into a noun. The source of the content-adjunct in the above example is in the object of the following sentence. שִׁצֲרוּ שֶׁיֵּשׁ חַיִּים צַל הַמַּאְדִים (It was supposed that there is life on Mars) → הַהַשְׁצָרָה שֶׁיֵּשׁ חַיִּים צַל הַמַּאְדִּים (The supposition that there is life on Mars…), cf. the similar phenomenon in 4.35: Prepositional Phrase Adjunct and 4.352: Prepositional Phrase with Verb Transformed to Noun, a prepositional phrase as adjunct. There is no element in a content-adjunct which agrees in gender and number with the noun to which the adjunct is attached. Introducing words of the subordination of the content-adjunct are שֶׁ or כִּי. If the content-adjunct is transformed from a question, the interrogative word serves as an introducing word of subordination, e.g., הוּא דָּאַג׃ מַה יַּצֲשֶׂה בָּעִיר הַגְּדוֹלָה? זֶה לֹא נָתָן לוֹ מָנוֹחַ (He worried: What would he do in the big city? This gave him no rest) → הַדְּאָגָה מַה יַּצֲשֶׂה בָּעִיר הַגְּדוֹלָה לֹא נָתְנָה לוֹ מָנוֹחַ (The anxiety about what he would do in the big city gave him no rest). When the content-adjunct begins with an infinitive there is no other introductory word of subordination, e.g., נִכָּר אֶצְלָם הָרָצוֹן לְהִתְבַּלֵּט (In them could be seen the desire to excel).
Grammars of European languages and of Arabic, also accepted in Hebrew grammars, have long defined the object semantically (e.g., "The word denoting the substance to which the action expressed in the predicator passes is called the direct object. If the action is merely connected with it, the word is called the indirect object").
The syntactic definition of an object is based on its being obligatory, or "close," complements of the verb-predicator. Optional complements are adverbials (cf. 4.5: Adverbial). In many instances it is possible to distinguish sharply between an obligatory complement, e.g., the prepositional phrase consisting of ב and a following noun as complement to the verb הִשְׁתַּמֵּשׁ (use) and an optional complement, e.g., the same phrase as complement to the verb הָלַךְ (walk). הִשְׁתַּמֵּשׁ בַּחֶדֶר (He used the room), as opposed to הָלַךְ בַּחֶדֶר (He walked in the room). Usually the preposition introducing an obligatory complement cannot be changed, for example, we cannot replace the preposition ב linked to the verb הִשְׁתַּמֵּשׁ by another preposition. Sometimes there is a restricted range of permissible substitutions though generally only one additional preposition is allowed, e.g, … נִתְמַנָּה כ… = נִתְמַנָּה ל (He was appointed as…). Sometimes a change of preposition effects the meaning of the verb, e.g., …קִנֵּא ב (He envied) … ≠ …קִנֵּא ל (He suspected). A preposition introducing an obligatory complement is called an obligatory preposition. It can be considered a part of the lexical entry for the verb. Although at first sight the obligatory preposition must always accompany its verb, there are certain conditions, apparently varying with particular items, under which it can be omitted. In all probability one should speak of varying degrees of obligatoriness in Hebrew (cf. 5.5: Obligatoriness). Moreover, the same verb may appear also without requiring a particular preposition. Since the obligatory preposition is part of the verb's lexical entry, it must be concluded that such a verb should be given two separate lexical entries, one when the obligatory preposition is a part of it, and the other when the verb appears without an obligatory preposition. Generally the two entries will have different meanings, e.g., צָבַד (He worked) – צָבַד צַל (He worked upon); הִתְגַּלְגֵּל (He wandered around) – …הִתְגַּלְגֵּל ל (He was transformed into); הִשְׁתַּגֵּצַ (He became mad) – הִשְׁתַּגֵּצַ אַחֲרֵי (He longed desperately for). The difference can cause ambiguities since a particular preposition not required by the verb in a certain occurrence can nevertheless be attached to it as an optional complement. Hence, the combination of the same verb and preposition can be followed by either an object or by an adverbial. In such instances, of course, the distinction is not so easy to make. At all events, dictionaries do not adequately distinguish between prepositions that are obligatory to a certain degree, and optional prepositions.
The terms "direct object" and "indirect object" derive from European or Arabic grammars. They were originally intended to distinguish between objects preceded by a preposition and those linked directly to a verb without an intervening preposition. Hebrew, however, has a preposition – אֶת – which appears before a direct object. Thus, the use of this term in Hebrew does not correspond to its original use. On the other hand, אֶת generally appears only before an object which is a determined noun, and many writers point to this as justification for the use of the term in Hebrew. It has also been argued that אֶת should not be regarded as a preposition at all, but merely as an indicator of determination. In practice there is no essential syntactic distinction between direct object and indirect object, since all objects are obligatory complements. A subclassification of obligatory complements based on the nature of the obligatory preposition should include אֶת, even though אֶת can be replaced under certain conditions by Ø, which is merely a variant of אֶת. It seems that here too the influence of foreign grammars has been excessive.
These terms are entailed by the preceding terms. Generally, those defining object semantically will define "intransitive verb" and "transitive verb" semantically, e.g., "a verb whose action passes to another body is a transitive verb," while an intransitive verb is a verb "whose action does not pass to others, but affects only the actor." It is obvious that in such definitions "verb" means the lexical entry of the verb, comprising all its forms in all its occurrences. A moment's thought will show, as Jespersen has shown, that the action of many transitive verbs does not pass to another body. The syntactic approach should be applied here too and each verb classified in the sentence in which it appears. That is to say, one should not refer to a verb in this respect as a concept comprising all the possible forms distinguished in the grammar, but as a given form appearing in a given sentence. The tokens of the verbs and not their entries or their types should be classified as transitive and intransitive. In the sentence הַמַּלָּח מְצַשֵּׁן סִיגָר (The sailor is smoking a cigar), מְצַשֵּׁן is considered a transitive verb since it has an obligatory (close) complement, while in the sentence הַמַּלָּח מְצַשֵּׁן (The sailor is smoking) or הַמַּלָּח מְצַשֵּׁן בִּלְהִיטוּת (The sailor is smoking eagerly) it is an intransitive verb since it does not have a complement or it has an optional complement.
When in the same sentence there are two objects with the relationship between them of subject-predicator, i.e., nexus (cf. 4.2: Subject and Predicator), it is usual to call them first object (the object performing the function of subject in that relationship) and second object (performing the function of predicator), e.g., הָרוֹפֵא חָשַׁב אֶת הַחוֹלֶה לְבַדַּאי (The doctor considered the patient an impostor), underlying which is the sentence הַחוֹלֶה בַּדַּאי (The patient is an impostor). Only certain verbs can appear in such a sentence, verbs denoting the attitude or opinion of the person designated in the subject to what is designated in the first object. This attitude, or an action resulting from this attitude, is expressed in the second object. Thus, in the above the attitude of the doctor to the patient is expressed in בַּדַּאי (an impostor). Similarly, מוֹצֵא אֲנִי מַד מִמָּוֶת אֶת הָאִשָּׁה (I find woman more bitter than death) – הָאִשָּׁה (woman) is first object, מַד מִמָּוֶת (more bitter than death) second object. Here another structure should be mentioned, namely sentences in which the object is a subordinate sentence beginning with a subordinator, e.g., הָרוֹפֵא חָשַׁב שֶׁהַחוֹלֶה בַּדַּאי (The doctor thought that the patient was an impostor). In biblical Hebrew the word הִנֵּה (behold) often opens the subordinate sentence, e.g., וּפַרְעֹה חֹלֵם…וְהִנֵּה מִן הַיְאֹר עֹלֹת שֶׁבַע פָּרוֹת ("and Pharoah dreamed… and, behold, there came up out of the river seven kine," Genesis 41:1, 2) and וַיַּרְא וְהִנֵּה חָרְבוּ פְּנֵי הָאֲדָמָה ("and he looked, and behold, the face of the ground was dried," Genesis 8:13). See also 4.51: Circumstance Adverbial.
A complex sentence whose predicator is a verb denoting saying or thinking and whose object is an inner sentence, e.g., הִבְטַחְתִּי שֶׁאָ בוֹא (I promised that I would come) with שֶׁאָ בוֹא (that I would come) as object, can be expressed also as הִבְטַחְתִּי לָבוֹא (I promised to come) with לָבוֹא (to come) as object. The infinitive לָבוֹא is transformed in this case from אֲנִי אָבוֹא (I will come), containing subject and predicator. Sometimes the infinitive has its own complements, e.g., in the sentence הַמפַקֵּד דָּרַשׁ מִן הַחיָּלִים לְהִשָּׁמַע לוֹ (The officer required the soldiers to obey him), לוֹ is object, obligatory complement to לְהִשָּׁמַע (to obey) while לְהִשָּׁמַע לוֹ (to obey him) is object of דָּרַשׁ (required). However, sometimes an identical surface structure should not be treated in this way because the first verb in such a combination is an auxiliary verb while the infinitive is the nucleus of the phrase with the function of predicator. "Auxiliary verb" has a wider range in this sense than is accepted for some languages, including English. These verbs complementing the nucleus of the predicator comprise modals and aspectual verbs, viz. verbs denoting the speaker's attitude toward the content of the sentence, the attitude of the person designated as subject toward the content of the rest of the sentence, or the point of time in the action, its duration, its recurrence, etc. For example: הַגֶּשֶׁר עָלוּל לְהִשָּׁבֵר (The bridge is likely to break), הַמְּפַקֵּד נֶאְֶלַץ לְהַמְתִּין (The officer was forced to wait), הַגֶּשֶׁר מַתְחִיל לְהִשָּׁבֵר (The bridge is starting to break), הַגֶּשֶׁם מוֹסִיף לָרֶדֶת (The rain continues to fall), הָאכָּר נוֹהֵג לִשִׁתּוֹת (The farmer is accustomed to drink). It ought to be added that besides the infinitive (the most usual form), the nucleus in such combinations may also take on such forms as participle, e.g., הִתְחִיל מְפַקְפֵּק (He began doubting); verbal noun preceded by the preposition ב, e.g., הִרְבָּה בַּאֲכִילָה (He ate a lot, literally: He increased in eating); or finite verb identical in person and tense to the auxiliary, the two verbs being coordinated by the conjunction ו, e.g., חָזַר וְקָרָא (He again read, literally: He returned and read); another aspect is expressed by repeating the same verb itself: הֵם הָלְכוּ וְהָלְכוּ (They walked for a long time). There have been hardly any studies in this area of Hebrew, and there is still no complete list or categorization of these auxiliaries.
An internal object is the term applied to a verbal noun functioning as object to a verb of the same root. The internal object is usually not an obligatory complement. It has one of two functions: (1) to emphasize the verb serving as predicator, e.g. גְּנֵבָה וּלְהִסְתֵלֵּק לְגְזֹל גְּזֵלָה, לִמְעֹל מְעִילָה, לִגְנֹב (To embezzle (misuse), to rob, to steal – and to disappear). This use is a modern counterpart of the use of the infinitive absolute in biblical Hebrew, e.g., הָלוֹךְ הָלְכוּ הָעֵצִים (The trees have surely gone); (2) to serve as nucleus to an adjunct when the combination of nucleus and adjunct functions as adverbial to a predicator, e.g., the phrase יְשיבָה כְּבֵדָה וּמְאֻשֶּׁשֶׁת (a heavy and firm sitting) in the sentence כְּבֵדָה וּמְאֻשֶּׁשֶׁת יָשַׁב הַכַּפְרִי יְשִׁיבָה (The villager sat heavily and firmly). A direct object is usually not determined, and therefore the preposition אֶת rarely precedes it.
The adverbial is an optional complement of the predicator (or of the sentence as a whole, cf. 4.53: Sentence Adverbial). Adverbials generally begin with a preposition, if we exclude a few words considered adverbs, e.g., פֹּה (here), אֶתְמוֹל (yesterday), יַחְדָּו (together), הֵיטֵב (well), or temporal words, e.g., יוֹם (day), which in this function generally appear without anything added before them (though two adjacent instances of such words may appear, e.g., יוֹם יוֹם (every day), שָׁנָה בְּשָׁנָה (year by year), similarly טִפִּין טִפִּין (drop by drop)) and if we exclude the locative expressed by a noun to which is added an unstressed a, e.g., צָפוֹנָה (northwards), הַבַּיְתָה (home(wards)). The prepositions used for this purpose are the same prepositions introducing obligatory complements, except for אֶת. (See 4.51: Circumstance Adverbial.) In Hebrew syntax, as in the syntax of other languages without cases, it is usual to classify adverbials not formally – a method used in languages with cases – but according to content. Thus, often the following adverbials are distinguished, or at least some of them: place, time, cause, purpose, manner, measure, circumstance, condition, concession, and result. Not all of these appear in every book, nor do the authors agree on the ascription of a phrase to the same adverbial. Thus, there are differences with respect to phrases denoting duration of time, e.g., עָבַד שָׁלוֹשׁ שָׁעוֹת (He worked three hours). Different authors designate such a phrase as time adverbial, measure adverbial, or manner adverbial. Studies on the deep structure of adverbials have scarcely been written, apart from some work on the circumstance adverbial, which is recognized as a transformation of a predicator under certain conditions, e.g., הַיַּלְדָּה חָזְרָה עֲיֵפָה (The girl returned tired) ← ←הַיַּלְדָּה חָזְרָה (The girl returned); הִיא הָיְתָה עֲיֵפָה בְּאוֹתוֹ זְמַן (She was tired at that time). It may be supposed that research in this area will show that the traditional categories of adverbials, now based on semantic distinctions in surface structure, derive from deep structure.
The circumstance adverbial is also called "circumstance adjunct," since like adjuncts it agrees in gender and number with subject or object, e.g., הַיְּלָדִים יָצְאוּ שְׂמֵחִים (The children went out happy), פָּגַשְׁתִּי אוֹתָם מְאֻשָּׁרִים (I met them happy). In both instances the agreement derives from the same source. The adjunct is transformed from a predicator – an adjective or participle – and so is the circumstance adverbial, except that with the latter the predicator denotes not a permanent phenomenon, but one that is contemporaneous with the action expressed by the predicator in our sentence. Thus, in the sentence הַיַּלְדָּה חָזְרָה עֲיֵפָה (The girl returned tired) the girl is said to be tired at that time. If tiredness was a permanent characteristic, the adverbial עֲיֵפָה would have been changed into an adjunct: הַיַּלְדָּה הָעֲיֵפָה חָזְרָה (The tired girl returned). On the other hand, the circumstance adverbial does not have to be attached to a subject or object, and these do not function as nucleuses to it. Moreover, the circumstance adverbial is not determined, even when the noun it is related to is determined. This adverbial can be expressed by a participle form preceded by the preposition ב, e.g., הוּא נִכְנַס לָעִנְיָן בְּמִתְכַּוֵּן (He entered into the affair intentionally). The adverbial differs in these features from the adjunct. (For the difference between circumstance adverbial related to the object and second object, cf. 4.44: First and Second Object.) Since the predicator is the source for both circumstance adverbial and adjunct, we cannot accept the suggestion that the circumstance adverbial be termed "circumstance predicator." The transformation of the predicator does not necessarily produce a circumstance adverbial. Furthermore, terms for the parts of the sentence in surface grammar are not generally based on their transformations from deep structure. A circumstance adverbial can be realized by a complete sentence. This sentence, considered a subordinate sentence, is linked to the independent part by the conjunctions בְּלֹא שֶׁ, בְּלִי שֶׁ, כְּשֶׁ, and ו, or it is juxtaposed to the independent sentence without a conjunction, e.g., הַיַּלְדָּה הִצְבִּיעָה עַל הַשּׁוֹדֵד כְּשֶׁהִיא רוֹעֶדֶת מִפַּחַד (The girl pointed to the robber while she was trembling with fright), הוֹלֵךְ לוֹ יְדִידֵנוֹ בַּדֶּרֶךְ, הַשָּׁמַיִם הַכְּחֻלִּים מֵעַל רֹאשׁוֹ וְהַנַּחַל לִימִינוֹ (Our friend walks along, the blue skies above his head and the brook on his right). In circumstance sentences, the predicator is realized by a participle form (or the sentence lacks a predicator, cf. 4.23: Nominal Sentence and Verbal Sentence). See also 5.13: Predicator Transformed into Circumstance Adverbial.
The conditional adverbial is unique among the adverbials. In the rare instances when it is realized as a nominal phrase in a simple sentence it will normally begin with… בְּמִקְרֶה שֶׁל (in case of), but it is chiefly realized as a subordinate sentence in a complex sentence. The conditional adverbial is called רֵישָׁה (protasis) whether it appears at the beginning or end of the sentence, while the rest of the sentence is called סֵיפָה (apodosis). The conditional adverbial can be distinguished grammatically, and not just semantically. Moreover, the two chief categories – real condition and hypothetical condition – are also formally distinguishable. Conditional sentences also have their own intonation patterns.
A "real condition" denotes something that has happened, is happening, or will happen and whose existence entails a result expressed in the apodosis part of the sentence. The chief signs of an adverbial of real condition are (1) special subordinating conjunctions – כַּאֲשֶׁר, כְּשֶׁ־, אִם; (2) the word order in the protasis; (3) the place of the protasis in the sentence; (4) the dependence of the tense of the verb in the superordinate part on that of the verb in the conditional part. Sometimes several of these signs come together, cf. 6.322: Word Order as Indication of Subordination in section Links beyond the Sentence.
A "hypothetical condition" is one which at the time it is said is known not to be fulfilled. The speaker speculates as to the possible results if the condition had been fulfilled. The traditional conjunctions for a hypothetical condition are אִלּוּ, לוּ, and in the negative אִלּוּלֵא, לוּלֵא. The introductory אִלְמָלֵי) אִלְמָלֵא), used mainly in literature, is sometimes interpreted as positive and sometimes as negative. In modern Hebrew (also in some places in biblical Hebrew) אִם is used also for a hypothetical condition. This use is accompanied by a verb form consisting of הָיָה (verb "be") plus participle, e.g., אִם הָיִיתָ אוֹמֵר לוֹ (If you had said to him), which is usually interpreted as a hypothetical condition indicating that the speaker knew the other had not said it. On the other hand, אִם אָמַרְתָּ לוֹ(If you said to him) is interpreted as a real condition, denoting that the speaker does not know whether the other had said it or not, but it is certainly possible that he said it. The sign of an unreal condition here is not a special conjunction (אִלּוּלֵא, לוּלֵא, לוּ as opposed to אִם) but the form of the verb (הָיָה + participle as opposed to past tense). The verb in the apodosis of a hypothetical condition also has the form הָיָה plus participle, irrespective of the verb in the protasis.
A double condition is one in which the speaker sets out both the result of the fulfillment of the condition and the result of its lack of fulfillment. This structure is also known as תְּנַאי בְּנֵי גָד וּבְנֵי רְאוּבֵן (a condition of the children of Gad and the children of Reuven), cf. Numbers 32:29–30.
An emphatic condition with negative followed by positive (see below), is apparently related to the double condition and is derived from it by a deletion transformation. For example, לֹא יִכָּנֵס אָדָם לְמַחֲנֶה צְבָאִי אֶלָּא אִם (כֵּן) הֻרְשָׁה לְכָךְ (A person may not enter a military camp unless he is expressly permitted), which is presumably before the transformation לֹא יִכָּנֵס אָדָם לְמַחֲנֶה צְבָאִי אִם לֹא הֻרְשָׁה לְכָךְ; יִכָּנֵס אָדָם לְמַחֲנֶה צְבָאִי אִם (כֵּן) הֻרְשָׁה לְכָךְ (A person may not enter a military camp if he is not expressly permitted; a person may enter a military camp if he is expressly permitted).
A concessive sentence is a conditional sentence the content of whose apodosis is reversed as far as can be determinedfrom its presumed protasis and the subject matter of the whole sentence. It also appears to be derived from a double condition, where neither the fulfillment nor the lack of fulfillment of the condition can change the result. For example, the sentence אִם לֹא תַּעֲמֹד עַל שֶׁלְּךָ – לֹא תְּקַבֵּל, וְגַם אִם תַּעֲמֹד עַל שֶׁלְּךָ – לֹא תְּקַבֵּל (If you don't defend your own, you will not receive anything, and even if you defend you own you will not receive anything) can be contracted to אִם לֹא תַּעֲמֹד עַל שֶׁלְּךָ וְגַם אִם תַּעֲמֹד עַל שֶׁלְּךָ – לֹא תְּקַבֵּל (If you don't defend your own and even if you defend your own, you will not receive anything). (The introductory formula for such a structure can be ״אִם…אוֹ אִם…״, ״בֵּין שֶׁ…וּבֵין שֶׁ…״, ״גַּם אִם…(וְ)גַם אִם״). If from such a structure the condition which is more probable in the context is deleted, the result is a concessive sentence: גַּם אִם תַּעֲמֹד עַל שֶׁלְּךָ – לֹא תְּקַבֵּל (Even if you defend your own you will not receive anything). A concessive sentence can also come from a series of conditional sentences in which one element is changed every time until the series comprises a wide range of topics the last of which is the converse of the first. When only the last is expressed, the rest of the possibilities are understood, deduced a fortiori. For example, in אֲפִילוּ אִם יִקְרְאוּ לִי אֶבּוֹלִיצְיוֹנִיסְטְ, לֹא אַסְגִּיר אֶת הַכּוּשִׁי (Even if they call me an abolitionist, I shall not hand over the Negro), what is also clear is that עַל אַחַת כַּמָּה וְכַמָּה לֹא אַסְגִּיר אֶת הַכּוּשִׁי אִם יִקְרְאוּ לִי בִּשְׁמוֹת גְּנַאי פָּחוֹת חָרִיפִים, אִם לֹא יִקְרְאוּ לִי כְּלָל בִּשְׁמוֹת גְּנַאי, אוֹ אִם יְשַׁבְּחוּנִי עַל כָּךְ (All the more so, I will not hand over the Negro if they call me names that are less derogatory, if they do not use any derogatory names against me, or they praise me for it).
Some adverbials do not complement the predicator, but are comments adding details to what is said in the sentence as a whole, e.g., בְּמַקְלִי עָבַרְתִּי אֶת הַיַּרְדֵן (With my staff I crossed the Jordan), דַּרְכָּם נִמְשְׁכָה בִּשְׁתִיקָה(Their journey continued in silence). These are not predicator adverbials but sentence adverbials or situation adverbials. A subordinate clause can also realize this function, e.g., הוּא נִכְנַס לְעֶצֶם הָעִנְיָן בְּלֹא שֶׁנִּתְכַּוֵּן לְכָךְ (He went into the heart of the matter though he did not intend to do so). Following N. Chomsky's works on transformational grammar, it has been claimed by some authors, that in many cases place and time adverbials, as well as some other adverbials, should be considered as sentence adverbials, e.g., פֹּה הַנְּעָרוֹת נָאוֹת (Here the girls are nice) (see Rubinstein, Lešonenu, 35).
Grammars do not give a complete list of prepositions introducing classes of adverbials, but a large number can be extracted from the examples they give. A fuller list is provided of conjunctions introducing subordinate sentences functioning as adverbials. Below is a list of the main prepositions and conjunctions serving as introducers to adverbials:
Place Adverbial: prepositions – עַל פְּנֵי, עַל יַד, עַל גַּבֵּי, עַל, מִתַּחַת לְ, מִמַּעַל לְ, אֵצֶל, אֶל, מִ, לְ, בְּ; conjunctions – בְּמָקוֹם שֶׁ, בְּמָקוֹם אֲשֶׁר (and other prepositions preceding מָקוֹם שֶׁ, or אֲשֶׁר מָּקוֹם, eg. אֶל מָקוֹם שֶׁ, מִמָּקוֹם אֲשֶׁר).
Time Adverbial: prepositions – עַד, לִפְנֵי, אַחֲרֵי, אַחַר, מִ, לְ, כְּ, בְּ; conjunctions – כָּל זְמַן שֶׁ, בְּשָׁעָה שֶׁ, לִכְשֶׁ, מֵעֵת שֶׁ, מָשָּׁעָה שֶׁ, מִדֵּי, כָּל עוֹד, בְּעוֹד, כֵּיוָן שֶׁ, כָּל אֵימַת שֶׁ, בְּעֵת שֶׁ, קֹדֶם שֶׁ, טֶרֶם שֶׁ, טֶרֶם, מֵאָז, אַךְ, לְאַחַר שֶׁ, and also the above prepositions (except לְ, בְ) in combination with שֶׁ, e.g., לִפְנֵי שֶׁ, אַחֲרֵי שֶׁ, מִשֶּׁ, כְּשֶׁ, or (except מִ, לְ, בְ) in combination with אֲשֶׁר, e.g., אֲשֶׁר, אַחֲרֵי אֲשֶׁר, כַּאֲשֶׁר עַד אֲשֶׁר, לִפְנֵי.
Manner Adverbials: prepositions – יוֹתֵר מִ, כְּמוֹ, מִתּוֹךְ, מִ, לְ, כְּ, בְּ.
Measure Adverbials: prepositions – עַד, כְּ.
Cause Adverbials: prepositions – בְּשֶׁל, לְרֶגֶל, מִפְּנֵי, בִּגְלַל, מִ, בְּ, הוֹדוֹת לְ, מֵחֲמַת, בִּשְׁבִיל; conjunctions – שֶׁ, מֵאַחַר שֶׁ, לְפִי שֶׁ, מִכֵּיוָן שֶׁ, שׁכֵּן, כִּי בַּאֲשֶׁר, עֵקֶב, יַעַן כִּי, יַעַן, הוֹאִיל וְ …, מִשּׁוּם שֶׁ, עַל, and several of the above prepositions followed by שֶׁ, e.g., מִפְּנֵי שֶׁ, בִּשְׁבִיל שׁ.
Purpose Adverbials: prepositions – כְּדֵי, בִּשְׁבִיל, לְשֵׁם, לְמַעַן, לְ; conjunctions – כְּדֵי שֶׁ, בִּשְׁבִיל שֶׁ, לְמַעַן אֲשֶׁר, לְמַעַן.
Conditional Adverbials: prepositions – בְּמִקְרֶה שֶׁל; conjunctions – לִכְשֶׁ, בִּזְמַן שֶׁ, כְּשֶׁ, כִּי, אִם (and other introducers of time adverbials) כָּל מִי שֶׁ, מִי שֶׁ.
Concessive Adverbials: prepositions – לַמְרוֹת, עַל אַף; conjunctions – אַף כִּי, גַּם כִּי, גַּם אִם, עִם כָּל, בְּכָל, אִם כִּי, אַף אִם, אַף שֶׁ, אַף עַלפִּי שֶׁ, אֲפִילוּ.
Result Adverbials: prepositions – וְ, עַד כְּדֵי, עַד לְ, עַד (in a negative sentence or in a rhetorical question); conjunctions – עַד כִּי, שֶׁ, עַד שֶׁ, עַד אֲשֶׁר.
Equative Adverbials: prepositions – כְּאִלּוּ, כְּמ וֹ, כְּ, בְּ; conjunctions – כְּשֵׁם שֶׁ, כְּמוֹ שֶׁ, כְּשֶׁ, כַּאֲשֶׁר, בְּמִדָּה שֶׁ, כָּל כַּמָּה שֶׁ, כְּכֹל אֲשֶׁר, כְּמוֹ שֶׁ, כִּלְעֻמַּת שֶׁ, כְּדֶרֶךְ שֶׁ.
The independent part of the sentence is introduced by correlative conjunctions, e.g., אַף, כָּךְ, כָּכָה, כֵּן.
Comparative Adverbials: preposition – מִ; conjunctions – יוֹתֵר מִמַּה שֶּׁ, יוֹתֵר מִשֶּׁ, מֵאֲשֶׁר, מִשֶּׁ. Sometimes the whole sentence is a rhetorical question: the comparative adverbial opens with a conjunction such as הֲלֹא, הִנֵּה, אִם, וּמַה, הֵן while the independent part is introduced by a correlative conjunction, e.g., כָּל שֶׁכֵּן, עַל אַחַת כַּמָּה וְכַמָּה, קַל וָחֹמֶר, וְאֵיךְ.
Below are examples of other uses:
Instrumental Adverbial: בְּ (colloquially also עִם) – בְּפַטִּיֹשׁ הִכָּה (He hit with a hammer).
Price Adverbial: בְּ (also בְּעַד) – קָנוּ בְּכֶסֶף רָב (They bought with a great deal of money).
Concomitant Adverbial: בְּ (also עִם) – יָצְאוּ בִּרְכוּשׁ גָּדוֹל(They went out with a lot of property).
Coordinate Adverbial: הָלַךְ עִם חֲבֵרוֹ – עִם (He went with his friend).
Material Adverbial: הַקִּיר בָּנוּי מִלְּבֵנִים – מִ (The wall is built of bricks).
Oath or Promise Adverbial: e.g., בְּחַיַּי שֶׁ, שֶׁכֹּה אָמוּת אִם, חֵי נַפְשִׁי שֶׁ.
In Hebrew, as in other languages, sometimes a word or a form involves the appearance of another word or form in the same sentence or in a neighboring sentence. The reciprocal relationship between the words is called "dependency." Dependencies within the sentence are classified as (1) concord in gender and number (5.11–5.23), (2) concord of determination (5.3–5.4), (3) obligatory appearance (5.5), and (4) order of parts of sentence (5.6).
The basic concord rule in Hebrew is the rule requiring that as far as possible the predicator should agree in gender and number with the subject. When the predicator is a verb, participle, or adjective, it always appears in a form agreeing in gender and number with the subject (see below 5.14: Lack of Concord between Subject and Predicator; 5.25: Lack of Concord in Adjunct Sentences on instances of lack of concord), e.g. הַזְּמַן קָצָר, הַמְּלָאכָה מְרֻבָּה, בַּעֲלֵי הַבַּיִת דוֹחֲקִים וְהַפּוֹעֲלוֹת עֲצֵלוֹת (The time is short, the work is great, the masters of the house are pressing, the female workers are lazy). When the predicator is an adjective or participle, agreement of gender between subject and predicator in the plural is always determined by the singular form of the subject. That is to say, it is irrelevant whether the plural suffix of the noun is ־ִים or וֹת–. If the noun is masculine, the plural suffix of the adjective predicator is always ־ִים, e.g., הָאָבוֹת זְקֵנִים (The fathers are old), since in the singular we have הָאָב זָקֵן, and of course הַבָּנִים חָרוּצִים (The sons are diligent) with a singular הַבֵּן חָרוּץ. Similarly, with a feminine noun the plural suffix of the adjective is always וֹת–, e.g., הַתְּאֵנִים יְבֵשׁוֹת (The figs are dry), since the singular is הַתְּאֵנָה יְבֵשָׁה, and obviously הַבָּנוֹת חָרוּצוֹת (The daughters are diligent) with a singular הַבַּת חָרוּצָה.
It is different when the predicator is a noun. Nouns do not always have gender inflection, nor is the meaning of the plural form always the same as that of the singular. Therefore, when such a noun realizes the function of predicator, it is sometimes impossible for the predicator and subject to agree in gender or number, e.g., הַשָּׁלוֹם הוּא תִּקְוָתֵנוּ (Peace is our hope), מִלְחָמָה הִיא מָוֶת (War is death), הָאַחֲרוֹן הַמַּחְסָנִים הֵם הַמִּכְשׁוֹל (The stores are the last obstacle). Nevertheless, sometimes a noun that normally is not inflected for gender did receive gender inflection when used in new ways. For example, the noun כּוֹכָב (star) was used as predicator for both male and female when it was first applied to an actor, e.g., גְּרֶטָה גַּרְבּוֹ הִיא כּוֹכָב גָּדוֹל (Greta Garbo is a great star). But after a time the form כּוֹכֶבֶת was created, e.g., יִשְׂרְאֵלִית דָּלִיָּה לָבִיא הִיא כּוֹכֶבֶת (Dalia Lavi is an Israel star).
When the predicator is transformed to the function of adjunct (see 4.31: Morphological Classification of Adjuncts), the agreement with subject is preserved, i.e., when the predicator in the source construction agrees with the subject, the adjunct transformed from it agrees in gender and number with its nucleus, e.g., הַזְּמָן הַקָּצָר לֹא הִסְפִּיק (The short time was not sufficient), הַפּוֹעֲלוֹת הָעֲצֵלוֹת פִּטְפְּטוּ (The lazy female workers chattered), אֶסְתֵּר הַמַּלְכָּה הָיְתָה יָפָה (Esther the queen was beautiful). But if the predicator in the source construction is a noun that cannot be inflected, then the appositive, which is the result of the transformation, does not necessarily agree in gender and number with its nucleus, e.g., הַשָּׁלוֹם, תִּקְוָתֵנוּ, עוֹדֶנוּ רָחוֹק (Peace, our hope, is still far off), הַקַּרְקַע, סְלָעִים, לֹא תֵּעָבֵד (The ground, rocks, will not be tilled). When a predicate containing a verb as predicator is transformed into a subordinate sentence with the function of adjunct (see 4.38: Relative Clause), the predicator continues to agree with the nucleus in gender and number, e.g., הַיְּלָדִים [הַיְּלָדִים שִׂחֲקוּ בֶּחָצֵר] לֹא הִרְגִּישׁוּ בַּמִּתְרַחֵשׁ (The children [The children played in the courtyard] did not notice what was happening), הַיְּלָדִים שֶׁשִּׂחֲקוּ בֶּחָצֵר, לֹא הִרְגִּישׁוּ בַּמִּתְרַחֵשׁ(The children, who played in the courtyard, did not notice what was happening).
Similarly, the circumstance adverbial also agrees in gender and number with the subject or object, since it is also a transform of the predicator (cf. 4.51: Circumstance Adverbial). There is also a requirement of concord between a circumstance sentence and one of the parts of the independent sentence in which it appears, e.g., הִיא דִּבְּרָה וְעֵינֶיהָ מַבְרִיקוֹת (She spoke, as her eyes were flashing). In many circumstance clauses, however, no concord exists, at least in the surface structure, e.g., וַיַּעֲמֹד בְּמִשְׁעוֹל הַכְּרָמִים, גָּדֵר מִזֶּה וְגָדֵר מִזֶּה (And he stood in the path of the vineyard, a fence here and a fence here = between two fences).
In some instances there is no concord between subject and predicator. Usually, this results from a difference between the grammatical gender or number of the noun and the natural gender or number of the person or entity denoted by the noun (Hebrew has no neutral), as in the following cases:
(1) Collective noun. When the noun denotes a group of individuals, its form is singular, but its predicator can be in the plural, agreeing with the content rather than the form, e.g., הַחֶבְרֶה רָצוּ לָלֶכֶת (The group wanted to go. Colloquial). The number of collective nouns with this usage seems to be fewer in modern Hebrew than in earlier periods of the language.
(2) Proper nouns that have the plural or dual form are always combined with a predicator in the singular and take the grammatical gender corresponding to the natural gender of the person, e.g., רַחֲמִים הָלַךְ (Raḥamim went off), תָּמָר חָזְרָה(Tamar returned), יוֹנָה הָלַךְ (Jonah went off (when the reference is to a male)), יוֹנָה הָלְכָה (Jonah went off (when the reference is to a female)).
(3) Names of countries are always in the feminine, whatever the form of the noun, e.g., וַתִּכָּנַע מוֹאָב בַּיּוֹם הַהוּא (Moab surrendered on that day), מִצְרַיִם שַׁלְחָה צָבָא לְתֵימָן (Egypt sent an army to Yemen), יַרְדֵּן הִכְרִיזָה עַל מַצָּב חֵרוּם (Jordan proclaimed a state of emergency), אַרְצוֹת־הַבְּרִית יָזְמָה וְעִידַת שָׁלוֹם (The United States initiated a peace conference).
(4) Pluralis majestatis. The noun is in the plural not to indicate plurality but out of respect to the person designated. Modern Hebrew includes in this category אֱלֹהִים (God) and also בְּעָלִים (owner), e.g., לֹא נִמְצָא בְּעָלָיו שֶׁל הַכֶּלֶב (The owner of the dog was not found).
Gender and number concord in a sentence is sometimes not the basic concord between subject and predicator (5.11–5.13), but the result of a transformation deleting a noun and replacing it with a pronoun. The substituted pronoun always agrees in gender and number with the deleted noun. Since there is usually another instance of the same noun elsewhere in the sentence, the pronoun agrees in gender and number to this other instance, as in the example given in 4.26: the sentence אִישׁ לֹא שָׂם לֵב אֶל הַוָּתִיקִים (Nobody paid any attention to the veterans) is transformed into a focusing sentence when the noun הַוָּתִיקִים (the veterans) is taken from its place and put initially while in its place is introduced a third person masculine plural pronoun, agreeing with it. Since the deleted noun followed a preposition, the substituted pronoun also follows the preposition and hence is attached to it. The resulting sentence is אִישׁ לֹא שָׂם לֵב אֲלֵיהֶם – הַוָּתִיקִים (The veterans, nobody paid any attention to them). The same applies when the focused part functioned as subject in the source sentence, e.g., סַבָּא אֵינוֹ מִתְעַיֵּף אַף פַּעַם (Grandfather never gets tired). When the subject is extracted from this sentence and placed initially, a third person masculine singular pronoun is introduced in its place, agreeing with סַבָּא (Grandfather). Since the noun is not preceded by a preposition, neither is the pronoun. The resulting sentence is סַבא – הוּא אֵינוֹ מִתְעַיֵּף אַף פַּעַם (Grandfather, he never gets tired). The pronoun agreeing with the focused part is called referring pronoun or "binder". As explained in 4.26: Focusing Sentence, the first noun in a focusing sentence is the subject and the rest of the sentence is the predicate. Consequently, the basic concord rule between subject and predicate applies here too; the element in the predicate agreeing with the subject is the referring pronoun.
When the predicate of a focusing sentence is transformed into an adjunct sentence (4.38: Relative Clause), the concord between subject and predicate is transformed into concord between the noun nucleus and the adjunct sentence: הַוָּתִיקִים [הַוָּתִיקִים – אִישׁ לֹא שָׂם לְב אֲלֵיהֶם] הֵחֵלוּ לְהִתְאַרְגֵּן בְּאִרְגּוּנִים נִפְרָדִים (The veterans [The veterans, nobody paid any attention to them] began to organize themselves in separate organizations → הַוָּתִיקִים, שֶׁאִישׁ לֹא שָׂם לֵב אֲלֵיהֶם, הֵחֵלוּ לְהִתְאַרְגֵּן בְּאִרְגּוּנִים נִפְרָדִים (The veterans, to whom nobody paid any attention, began to organize themselves in separate organizations). Even when the focused part was subject in the source sentence, e.g., סַבָּא – הוּא אֵינוֹ מִתְעַיֵּף אַף פַּעַם (Grandfather, he never gets tired), the whole of the predicate can be transformed into an adjunct sentence. However, if there is already subject-predicator concord in this sentence, for example when the predicator is a verb or there is a copula in the predicate, then this concord is usually sufficient. In this case, the adjunct sentence does not contain the subject pronoun, e.g., סַבָּא, שֶׁאֵינוֹ מִתְעַיֵּף אַף פַּעַם, חִיֵּךְ בְּסַלְחָנוּת(Grandfather, who never gets tired, smiled forgivingly) and not… סַבָּא, שֶׁהוּא אֵינוֹ מִתְעַיֵּף אַף פַּעַם, חִיֵּךְ בְּסַלְחָנוּת‡, or הָאִישׁ שֶׁהָלַךְ בַּדֶּרֶךְ, לֹא חָזַר (The man, who went off on a journey, did not return), and not הָאִישׁ, שֶׁהוּא הָלַךְ בַּדֶּרֶךְ, לֹא חָזַר ‡. This does not apply to a nominal sentence that does not have subject-predicator concord, e.g., הַמַחְסָנִים הֵם הַמִּכְשׁוֹל הָאַחֲרוֹן (The stores[they] are the last obstacle), the source of which is in the identifying sentence הַמַּחְסָנִים – הַמִּכְשׁוֹל הָאַחֲרוֹן (The stores – the last obstacle), cf. 4.24: Identifying Sentence. The relative clause formed from the predicate includes the pronoun הֵם, which agrees with the subject: הַמַּחְסָנִים שֶׁהֵם הַמִּכְשׁוֹל הָאַחֲרוֹן עוֹלִים בָּאֵשׁ (The stores, which are the last obstacle, are going up in fire). On whether the subject pronoun (הוּא, הִיא, הֵם, הֵן) is copula or referring pronoun of a focused part, see below 5.23: Copula Concord.
The source of the copula (see 4.22: The Copula) is a referring pronoun in a subject-focusing sentence, cf. the examples in 5.21: Pronoun Concord in a Focusing Sentence and 5.22: Pronoun Concord in a Relative Clause. However, the subject pronoun is also used in sentences where the focused part is not felt to be emphasized in any way (4.24: Identifying Sentence). This appears to have happened for two reasons:
(1) Subject-focusing sentences became common in nominal sentences in the present tense by analogy with nominal sentences in the past and future tenses, that is speakers tended to insert a word agreeing with subject in gender and number between subject and predicator, or more precisely, to link such a word to the predicator (not necessarily putting it before the predicator) as they do with sentences in the past or future. Since this word (an inflected form of הָיָה (be)) does not introduce any emphasis to the sentence in the past or future, the emphasis is also lost in sentences with subject pronouns in the present tense.
(2) In certain constructions that lack subject-predicate concord, and particularly in sentences without a predicator the desire for "leveling" activates speakers, i.e., the need is felt to add something that will produce subject-predicate concord, in order that such constructions can enter the regular framework of Hebrew sentences, in which there is subject-predicator concord. We can explain in this way the obligatory appearance of the pronoun as a copula in sentences such as הַפִּגּוּר הוּא בְּיִצּוּר דְּשָׁנִים (The delay is in the production of fertilizers), הָעֲלִיָּה הִיא בַּשָּׂכָר (The rise is in salary), which are transformed from sentences פְּלוֹנִי מְפַגֵּר בְּיִצּוּר דְּשָׁנִים (X is lagging in the production of fertilizers), הַשָּׂכָר עָלָה (The salary rose) respectively. (See E. Rubinstein, for another explanation.) Usually the copula agrees with the subject. However, there are cases where the copula agrees with the predicator, when several words separate the subject from the copula and the copula is next to the predicator. In the colloquial language, and sometimes in writing, some use זֹהִי, זֶהוּ, זֶה, זֹאת (or זֹאתִי) as copula. There is also quite frequent use of הִנּוֹ or הִנֵּהוּ and other inflected forms of הִנֵּה as copula.
(1) A possessive pronoun agreeing with a noun mentioned after it is to be found in the double construct state: בֵּיתוֹ שֶׁל הָאִישׁ (the man's house), חֶרְדָתָם שֶׁל הַהוֹרִים (the parents' dread), cf. 4.34: The Construct Structure.
(2) A pronoun attached to a preposition and referring to a noun mentioned after it is to be found in the apposition structure אָמְרוּ עָלָיו עַל רַבִּי עֲקִיבָא (They said about him, about Rabbi Akiva), cf. 4.37: Apposition.
(3) A demonstrative pronoun introducing an identifying sentence agrees in gender and number with the noun appearing as part of the complement in the identifying sentence, and not with the noun in the preceding sentence to which the pronoun refers, e.g., הַשְּׁאֵלָה יְדוּעָה, אֲבָל זֶה עִנְיָן אַחֵר (The question is well-known, but this is a different matter); זוֹ פִּסְקָה חֲדָשָׁה (… this is a new paragraph); אֵלֶּה דְּבָרִים נְדוֹשִׁים (… these are matters that are well-known), cf. also 6.12: Demonstrative Pronouns. The same applies if the reference is to something mentioned after the identifying sentence, e.g., זֹאת הַתּוֹרָה (This is the law) or אֵלֶּה הַחֻקִּים (These are the statutes), when the details come after such a sentence.
The demonstrative pronoun זֶה can be attached to an undetermined noun or to an undetermined construct state. The noun or construct state then becomes determined. That is to say, בַּיִת זֶה is equivalent to הַבַּיִת and בֵּית אֲבָנִים זֶה to בֵּית הָאֲבָנִים. Agreement in gender and number between pronoun and preceding noun is obligatory.
(1) Adjunct content-sentences (see 4.38: Relative Clause), which are not transformed from a predicate but from an object sentence (on which there is no obligatory concord in gender and number with anything outside the sentence), do not require concord with the noun nucleus transformed from the verb-predicator in the source sentence.
(2) Introductory expressions, especially for adverbial sentences (see 4.54: Prepositions as Introducers of Adverbials), e.g., בְּאֹפֶן שֶׁ, שָׁעָה שֶׁ, מָקוֹם שֶׁ. The sentence subordinated by such an expression does not include a referring pronoun agreeing with אֹפֶן, שָׁעָה, מָקוֹם, etc., e.g., כָּל פַּעַם שֶׁנַּפְשׁוֹ הָיְתָה מָרָה עָלָיו (every time that he felt depressed). It means that the noun in the introductory expression loses its semantic force, in whole or in part, and becomes entirely or virtually a grammatical word. Its "adjunct sentence" is not really an adjunct sentence. The introductory expression can also subordinate a sentence with a different function, for example as subject, e.g., ?הַאִם לֹא הִגִּיעָה הַשָּׁעָה שֶׁתִּתְאַחֵד חָכְמַת יִשְׂרָאֵל עִם שְׂפַת יִשְׂרָאֵל (Has the time when the wisdom of Israel will be united with the language of Israel not come?).
(3) Pronoun substitute. In some adjunct sentences the place adverbial שָׁם (there) replaces a pronoun attached to a preceding preposition, e.g., הַחַלּוֹן שֶׁהִצְטוֹפְפוּ שָׁם יוֹנִים…(the window where the doves crowded, literally the window which doves crowded there). The substitution of שָׁם for an inflected preposition is found in the Bible, e.g., אֲנִי מֵבִיא אֶתְכֶם שָׁמָּה אֶרֶץ כְּנַעַן, אֲשֶׁר (the land of Canaan, where I shall bring you).
(4) If in the source the nucleus was the object of אֶת in the adjunct sentence, the appearance of אֶת with an inflected pronoun agreeing in gender and number with the nucleus is not obligatory. For stylistic reasons it is normally omitted, unless the omission will lead to ambiguity. For example, חֹמֶר שֶׁשָּׁמַע מִפִּי מוֹרָיו אוֹ קָרָא בֶּעָבָר (material that he heard from his teachers or read in the past) and not חֹמֶר שֶׁשָּׁמַע אוֹתוֹ מִפִּי מוֹרָיו אוֹ קָרָא אוֹתוֹ בֶּעָבָר (literally, material that he heard it from his teachers or read it in the past). It is normally possible to add in any such adjunct sentence the preposition אֶת inflected to agree with the nucleus.
(5) When the nucleus is a verbal noun or abstract noun with the same root as the predicator in the adjunct sentence. In such a case there is no referring pronoun in the adjunct sentence, e.g., הָאֵבֶל שֶׁמִּתְאַבֶּלֶת עַל בְּנָהּ (the mourning which she mourns for her son). This phenomenon is presumably connected with the characteristics of the internal object.
(6) "Space words." When the nucleus is only required for grammatical purposes, namely for the attachment of an adjunct sentence, words devoid of semantic content, such as indefinite pronouns or demonstrative pronouns, e.g., מַה, כָּךְ, זֶה, or words with a very general semantic content, e.g., דָּבָר (thing), עֻבְדָּה (fact) are used. Such a nucleus does not require the presence of a referring pronoun in the adjunct sentence that is attached to it, whether the nucleus sums up what is said before or whether it comes after an obligatory preposition as introduction to what is to follow. For example: לְאַבָּא אָמְרָה כִּי אֶת אֲרוּחַת הַצָּהֳרַיִם הוּא אוֹכֵל הַיּוֹם בְּמִסְעָדָה, מַה שֶּׁהוּא שָׁמַע מִתּוֹךְ שְׁתִיקַת הַכְנָעָה (She told Father that he is eating lunch today at a restaurant, which he heard in submissive silence). However, sometimes the "space words" are not entirely devoid of semantic content and denote something more than a word like דָּבָר or עֻבְדָּה, e.g., בִּמְלֹאת שִׁבְעִים שָׁנָה לְהַדָּחַת בֵּית הַמְּלוּכָה הַפּוֹרְטוּגָלִי, תַּאֲרִיךְ שֶׁחָל אֶתְמוֹל (At the end of seventy years following the dethronement of the Portuguese monarchy, a date which fell yesterday…) or 1941 ל־ 1939 הוּא נֶחְבַּשׁ בְּמַחֲנֶה רִכּוּז, מָקוֹם שֶׁמֵּת אוֹ נֶהֱרַג בֵּין (He was imprisoned in a concentration camp, a place where he died or was killed between 1939 and 1941).
Determination is not recursive. It follows that the four methods of determination exclude each other: (1) the definite article; (2) a proper name; (3) a noun in combination with a possessive pronoun; (4) a nomen regens in the construct state. Determiner concord exists between a nucleus-noun and an adjective serving as its adjunct, i.e., either both words are determined or neither are determined: נְיָר חָלָק (smooth paper), הַנְּיָר הֶחָלָק (the smooth paper). In mishnaic Hebrew and to some extent in biblical Hebrew it is possible for the adjunct to be determined while its nucleus is not determined, e.g., מַיִם הָרָעִים (the evil waters), יָם הַתִּיכוֹן (the Mediterranean Sea), יוֹם הַשִּׁשִּׁי (the sixth day). This phenomenon is seldom found in modern Hebrew. Determiner concord does not have its source in the deep grammar. It is a surface phenomenon: the adjunct adjective receives determination even though the predicator, which served as its source, was not determined. The cause of determination is attraction: determination of the noun attracts determination of the adjunct adjective attached to it. The attraction of the definite article from the noun to the adjunct applies even to a demonstrative pronoun following a noun with the definite article (see section 5.24: Other Cases of Concord), since corresponding to בַּיִת זֶה (this house) we can have הַבַּיִת הַזֶּה without a difference in meaning being felt between these phrases. The same applies to the construct state relationship. Here too if the noun is determined by the definite article and the pronoun זֶה follows, the article is also attached to the pronoun בֵּית אֲבָנִים זֶה – בֵּית הָאֲבָנִים הַזֶּה (this house of stone). In the phrase הַבַּיִת הַזֶּה there is therefore a double determination. This is a clear example of redundancy, unique in determination, since determination is not recursive, i.e., a noun that is already determined cannot accept additional determination.
The non-recursiveness means that the definite article cannot be combined with a proper name, nor with a noun with an inflected possessive, nor with a nomen regens; a proper name cannot be combined with the definite article (unless the name has been changed into a common name, or if the definite article is part of the name, e.g., הַלִּבָנוֹן, הַיַּרְדֵן (the Jordan, the Lebanon)) nor with a possessive pronoun and it cannot be a nomen regens; a noun inflected for the possessive cannot be a proper name (unless it became an independent item, not connected with the common noun), cannot be combined with the definite article, and cannot be a nomen regens; a nomen regens cannot be a proper name unless it is a shortened name, e.g., רִאשׁוֹן for נַחְלַת, רִאשׁוֹן לְצִיּוֹן for נַחְלַת יְהוּדָה; it cannot be combined with the definite article or with a possessive pronoun, but if the construct state is taken to be a compound the definite article can precede it, e.g., הַכַּדּוּרֶגֶל (the football). It should be added, however, that two (or more) people with the same name may be referred to with a definite article preceding the name in plural "שְׁתֵּי הַשִּׁפְרוֹת." Determination by the demonstrative pronoun זֶה, which may precede or follow the noun, does not come under this rule: זֶה can come in addition to the above four methods. But if זֶה follows a noun with the definite article or the construct state with the definite article, the definite article must precede: הַזֶּה (see 5.3: Determiner Concord). When זֶה follows a noun determined by another method, the definite article is not obligatory but optional: כּוֹבָעִי זֶה (or, כּוֹבָעִי הַזֶּה) (my hat), אַבְרָהָם הַזֶּה, אַבְרָהָם זֶה (this Abraham), but בַּיִת זֶה (this house) – הַבַּיִת הַזֶּה, בֵּית אֲבָנִים זֶה (this house of stone) – בֵּית הָאֲבָנִים הַזֶּה.
Obligatoriness in the widest sense refers to the obligatory appearance of a word or form as a result of the existence of another word or form in the same sentence. It signifies roughly what is signified by "dependency" in the sentence but from the standpoint of one element in the dependency, either the active element or the passive one. For example, when the word אֶתְמוֹל (yesterday) appears in a verbal sentence the verb normally is in the past. Here, what has been made obligatory is the element signifying that the verb is in the past form, while what made it obligatory is the word אֶתְמוֹל. A further instance is the infinitive following certain auxiliaries, e.g., הִתְאַמֵּץ, מֻכְרָח, עָשׂוּי, עָלוּל (cf. 4.45: Infinitive as Object). Other auxiliaries require either an infinitive or a certain form of the verb, i.e., a participle, or a verbal noun, e.g., הִתְחִילוּ לְהָבִין (they began to understand), הִתְחִילוּ מְבִינִים (they began understanding), הִמְעִיט לְדַבֵּר (he spoke less, literally: he lessened to speak), הִמְעיטבְּדבּור (literally: he lessened in speech). In a narrow and more usual sense, obligatoriness is used to denote the obligatory appearance of a particular preposition when there is in the same sentence a particular verb requiring the preposition. For example, the appearance of the preposition בְּ is required by the verb הִשְׁתַּמֵּשׁ (cf. 4.41: Obligatory Complement and Obligatory Preposition). Compared with English, the rules are less stringent in this respect, since there are various cases when certain "obligatory" prepositions can be omitted without affecting the meaning of the verb. In most cases when an obligatory preposition is retained, the meaning of the verb without the preposition differs discernibly from its meaning with the preposition, e.g., דָּלַק (burn) – דָּלַק אַחֲרֵי (pursue after); הִתְגַּלְגֵּל (roll) – הִתְגַּלְגֵּל לְ (change to). When the verb does not have another meaning, then under certain conditions the obligatory preposition can be omitted, e.g., ְל after הִתְכַּוֵּן (intend) can be omitted before a subordinate sentence beginning with שֵׁ, e.g., הִתְכַּוַּנְתִּי שֶׁתָּבוֹאוּ (I intended you to come). It is also possible to retain the preposition לְ here, but it would then be necessary to add a space-word between לְ and שֶׁ, e.g., הִתְכַּוַּנְתִּי לְכָךְ שֶׁתָּבוֹאוּ, cf. 5.25, (6): Lack of Concord in Adjunct Sentences – "Space Words." This area of degrees of obligatoriness, i.e., when the obligatory preposition can be omitted, has not been sufficiently investigated in Hebrew. The preposition אֶת is required (after certain verbs) only before a determined noun (it may be omitted, however, for stylistic reasons) but if an undetermined noun follows, its omission is obligatory. Some phrase the rule differently: "Before an undetermined noun the obligatory preposition אֶת is changed to the obligatory preposition Ø." Obligatory prepositions are retained in many cases when the verb is converted into a verbal noun, e.g., הָרְדִיפָה אַחֲרֵי הֲנָאוֹת הַחַיִּים (the pursuit after the pleasures of life), הַזִּלְזוּל בַּסַּכָּנָה (the contempt for danger). This area too requires further investigation.
Concord between different parts of the sentence in gender, number, and person; marking of the subject pronoun in the verb form (Morphology – 19. The Inflection of the Verb); presence of prepositions, particularly presence of אֶת to mark the "direct object" – all these allow a reduction in Hebrew of restrictions on the order of the parts of the sentence. While comprehensive investigations have not been undertaken into the order of parts of the sentence in modern Hebrew, it is possible to say in general that word order is fairly free, and usually what the speaker wishes to emphasize he says at the beginning of the sentence. Below is given a list of restrictions (more or less accepted by all) on the order of the parts of the sentence. Nevertheless, in some instances the strict observance of them is more a matter of style than of syntax. The influence of word order in other languages can sometimes be discerned.
(1) When the sentence begins with an object or adverbial and the predicator is a verb, the predicator follows the object or adverbial and the subject comes after the predicator. The rule applies particularly when the verb is in the past or future, e.g., אֶתְמוֹל הִפְלִיגָה הָאֳנִיָּה – (Yesterday sailed the boat) הָאֳנִיָּה הִפְלִיגָה אֶתְמוֹל – (The boat sailed yesterday). In colloquial Hebrew, sometimes in writing also, there are cases where this rule is not kept.
When two complements of the predicator occur in the sentence, for example two different objects or an object and an adverbial, the speaker is free to give them in any order. But if one of them is an inflected preposition, i.e., the complement includes a personal pronoun, the inflected preposition precedes the second complement. If both of them are inflected prepositions, the order is free, unless one of the words is אֶת with a pronominal inflection, in which cases it comes first. For example: ל ֹומתֶא דֶלֶיַה תֶא יִתי ִאר (I saw the child yesterday); רָאִיתִי אֶתְמוֹל אֶת הַיֶּלֶד (I saw yesterday the child) – but: רָאִיתִי אוֹתוֹ אֶתְמוֹל (I saw him yesterday) and not רָאִיתִי אֶתְמוֹל אוֹתוֹ‡ (literally: I saw yesterday him). לָקַחְתִּי אֶת הַסֵּפֶר מֵאֲחוֹתִי (I took the book from my sister); לָקַחְתִּי מֵאֲחוֹתִי אֶת הַסֵּפֶר (I took from my sister the book) – לָקַחְתִּי אוֹתוֹ מִמֶּנָּה (I took it from her) and not לָקַחְתִּי מִמֶּנָּה אוֹתוֹל‡ (literally I took from her it).
(3) Interrogative words, coordinating conjunctions, and the various subordinating conjunctions appear at the beginning of the sentence, whether the sentence is independent or subordinate. But a few subordinating words such as אִם כֵּן, עַל כֵּן, לָכֵן, sometimes occur not at the beginning of the sentence they are connecting but within it, e.g., לָכֵן שָׁאַלְתִּי אוֹתוֹ מַדּוּעַ לֹא בָּא (Therefore I asked him why he did not come), שָׁאַלְתִּי אוֹתוֹ, עַל כֵּן, מַדּוּעַ לֹא בָּא (I asked him, therefore, why he did not come). The coordinating word גַּם (also) links the word after it to one of the words preceding it in that sentence or in a preceding sentence. Hence, גַּם can appear in various places in the sentence. In speech גַּם sometimes occurs after the word it is linking (undoubtedly under the influence of foreign languages). For example: גַּם שָׁתִיתִי [וְ]אָכַלְתִּי (I ate [and] also I drank); הָיָה שָׁם גַּם הַנָּשִׂיא (Also the president was there). The latter sentence implies a previous statement that others were there.
(4) There is still no adequate investigation in modern Hebrew of the order of the parts of the noun phrase, i.e., the order of the various adjuncts relative to the nucleus. But if there is no special reason for changes, the order seems to be as given below (the degree of confidence in this order is sufficiently high for the first five parts, though in the rest it is less; the parentheses denote that it is possible to omit that part and pass on to the next, and still preserve a noun phrase): (quantity adjunct +) noun phrase nucleus (+ nomen rectum adjunct) (+adjective adjunct) (+separate possessive pronoun adjunct) (+demonstrative pronoun) (+prepositional phrase adjunct) (+appositional adjunct) (+subordinate sentence adjunct). For example: שְׁנֵי מְעִילֵי הַצֶּמֶר הַחֲדָשִׁים שֶׁלִּי הָהֵם מֵאַנְגְלִיָּה, מַתְּנַת דוֹדִי, שֶׁהִגִּיעוּ בְּדִיּוּק לִפְנֵי שָׁבוּעַ, נֶעֶלְמוּ. (Those two new woolen coats of mine from England, my uncle's gift, which arrived exactly a week ago, have disappeared; literally: Two coats of wool – new – mine – those – from England – the gift of my uncle – which arrived exactly a week ago – have disappeared). "A special reason for change" (above) includes the wish for emphasis, an afterthought, and the length of the adjunct, particularly when it is a subordinate sentence. An attached possessive pronoun may accompany any noun.
(5) Any noun can be placed initially as a focused part (cf. 4.26: Focusing Sentence). Though as a result the general word order is changed, within the source sentence now serving as predicate the word order remains as it was, with the referring pronoun taking the place of the focused noun, cf. also 6.322: Word Order as Indication of Subordination.
Although speech consists not only of the combination of words, but also of the combination of sentences, the links between sentences in the same discourse have not been described in books on syntax, and in fact have not been given a linguistic description at all. Those dealing with the combination of sentences do so within theories of rhetoric or composition and pay attention not to grammatical questions but to literary and logical structure. In this respect some changes occurred in the early 1950s, when Z.S. Harris began linguistic analyses of a whole discourse, which can include much more than one sentence and can sometimes consist of a dialogue between two or more speakers. In doing so, even though some previous scholars had already dealt with this topic, he laid the foundations for the development of modern linguistic views on deep structure. As stated in the section 3.1: Structural Analysis above, there are few descriptions in Hebrew as yet which are based on those assumptions, still there are works in Hebrew that describe some of the material related to the problem of links beyond the sentence, for example the uses of the various subordinating conjunctions. The following survey covers these topics: (1) anaphoric references in sequentially related sentences; (2) elliptical sentences related to previous sentences; (3) ways of combining sentences.
Those rules of pronoun concord applying within the sentence, e.g., in a focusing sentence or in a relative clause, apply also when the pronoun is in a different sentence, which follows the one with the noun, even though the two sentences do not have any other grammatical links. Moreover, the pronoun in the new sentence (if indeed the second instance of the noun has been deleted and a pronoun has replaced the noun) normally cannot be omitted as happens under certain conditions in adjunct sentences. Thus, the deletion of the second instance of a noun and its replacement by a personal pronoun happens not only in a focusing sentence or in a relative clause but in general, whether the first noun is in the same sentence (in surface structure) or in a preceding sentence. The personal pronoun therefore agrees in gender and number with the noun, even though only the first instance of the noun remains. For example: רָאִיתִי אֶת הַיֶּלֶד. הַיֶּלֶד חָזַר לְבֵית הַסֵּפֶר (I saw the child. The child returned to school), רָאִיתִי אֶת הַיֶּלֶד. הוּא חָזַר לְבֵית הַסֵּפֶר (I saw the child. He returned to school).
(1) A demonstrative appearing in a sequentially related sentence but not functioning as adjunct sometimes refers back to a noun in a previous sentence. However, if its function in the sequentially related sentence is that of subject or predicator, it must agree in gender and number with the predicator or subject respectively of the sequentially related sentence and not with the noun in the preceding sentence to which it refers (cf. 5.24, (3): Other Cases of Concord). For example: הֵבֵאתִי אוֹתוֹ אֶל הַדִּירָה; ״זֶה יִהְיֶה בֵּיתְךָ״ אָמַרְתִּי (I brought him to the apartment. "This will be your home," I said). The same applies to the demonstrative כָּזֶה and its inflectional variants, e.g., יֵשׁ לִי עֲשָׂרָה פְּקִידִים גְּרוּעִים; אֵינֶנִּי רוֹצֶה עוֹד פָּקִיד כָּזֶה (I have ten rotten clerks; I don't want another clerk like that). (See H. Rosén, ʿIvrit Tova, for another explanation.)
(2) Sometimes the demonstrative does not refer to a noun in the preceding sentence, but to the whole of the content of the sentence or to a part of it. In such a case there is no requirement for concord with a particular element in the preceding sentence. Generally the demonstrative is then זֶה, e.g., הַיְּלָדִים שָׁכְבוּ לִישׁוֹן. זֶה הָיָה בְּשָׁעָה עֶשֶׂר (The children went to bed. That was at ten). A sequentially related sentence with such a structure can introduce stories, e.g., ……זֶה הָיָה בַּחֲנֻכָּה (It happened at Ḥanukkah). By doing so, the narrator plunges the reader straight into the story, making him feel that he is not at the beginning of the story. Another possible explanation: זֶה refers in these cases to something unknown which is to be explained later, so that the reader becomes anxious to know what is coming.
(3) The definite article should also be mentioned here, since its appearance before a noun indicates a reference to the previous appearance of the noun and confirms that the two instances of the noun have the same referent, e.g., אִישׁ הָיָה בְּאֶרֶץ עוּץ, אִיּוֹב שְׁמוֹ. וְהָיָה הָאִישׁ הַהוּא תָּם וְיָשָׁר (There was a man in the land of Uz, whose name was Job. And that man was perfect and upright). The fact that sometimes the definite article appears before the first appearance (in surface structure) of a noun does not invalidate this claim, since in such a case it may be supposed that a previous sentence containing a non-determined instance of the noun has been deleted as a result of a transformation. Here belong also various adjuncts, restrictive and nonrestrictive alike (4.32: Restrictive and Non-restrictive Adjunct). There is no doubt that the sentences in which these appear are linked to preceding sentences, whether they are retained or deleted.
Usually every sentence has two parts, a subject and a predicate. However, in Hebrew, as in other languages, there are short sentences that cannot be divided in this way. Some have called them single-element sentences, analyzing them according to their surface structure. But from the point of view of deep grammar they should be considered as remnants of normal sentences, with subject and predicate, from which some parts (subject and/or predicate or parts of the latter) have been deleted. Here belong expressions of agreement with or opposition to a preceding sentence (usually said by another speaker), e.g., כֵּן (Yes!), לֹא (No!), אוּלַי (Perhaps!), these being remnants of the predicate. Sentences such as רוּחַ. שֶׁמֶשׁ. חַם (Wind. Sun. Warm.) are used in literature and not necessarily in sequentially related sentences. The part that would complete them and make them into proper sentences has been omitted, because it can be understood or because it is unimportant, and not because it has been mentioned previously.
Two juxtaposed and linked sentences may be related in one of two ways: (1) The two sentences have equal status syntactically and are linked by coordination (see 3.14: The Compound Sentence above); (2) The two sentences have a different syntactic status, one of them being subordinate to the other (see 3.15: The Complex Sentence above).
It is customary to classify the kind of relations between two coordinated sentences by the relation of the content of the sequentially related sentence to that of the preceding sentence. The relation may be (1) addition; (2) contrast; (3) alternative; (4) explanation or conclusion; (5) result.
Coordinated sentences may be juxtaposed without a coordinating marker between them (asyndetic coordination), but generally in modern Hebrew the coordination is marked by a coordinator (syndetic coordination). Some coordinators mark only one of the above types of coordination, others mark more than one kind. Coordinations of types (1) and (3) can comprise more than two consecutive sentences. In that case the coordinator can come merely between the last sentence and the one preceding it. When the coordination is expressed without a coordinator, it generally requires a special intonation, particularly for coordinations of explanation and result, and a significant pause before the sequentially related sentence suggests the content-type of the coordination. Sometimes the coordination is marked by a coordinator before the first sentence. Every such coordinator has available a correlative between that sentence and the sequentially related sentence, e.g., …לֹא רַק…, כִּי אִם גַּם (Not only… but also…). Sometimes, however, the correlative is omitted, the speaker being content with the appropriate intonation.
(1) coordination of addition: …לֹא דַי … אֶלָּא ; …וְ… ; … גַּם …לֹא רַק אֶלָּא גַּם ; …אַף… ; …וְלֹא עוֹד אֶלָּא שֶׁ… ; (וְ) גַּם ; גַּם… ; שֶׁ…; …לֹא זוֹ בִּלְבַד … אֶלָּא שֶׁ ; לֹא רַק… (כִּי אִם) גַּם … ;
(2) coordination of contrast: …אֶלָּא… ; … ; …וְ… ; …אֲבל… לֹא… כִּי אִם… ; אֵין… אֶלָּא… ; אוּלָם… ; …אַךְ… ; רַק… ; …אֶפֶס ;
(3) coordination of alternative: …אוֹ שֶׁ… אוֹ… ; אוֹ שֶׁ… אוֹ שֶׁ… …אוֹ;
(4) coordination of reason or conclusion: …מִכָּאן שֶׁ… כְּלוֹמַר… ; …שֶׁהֲרֵי…;
(5) coordination of result: …לָכֵן… ; …עַל כֵּן… ; …לְפִיכָךְ
A subordinated sentence realizes a function within another sentence (see 3.15: The Complex Sentence above). From the point of view of the other sentence "a link outside the sentence" does not apply, but from the point of view of the subordinated sentence that is the nature of the link.
Subordination is chiefly marked in Hebrew by words exclusively used to mark subordination, by words that mark either coordination or subordination, and by the order of the sentence elements, whether within the subordinated sentence or in the place of the subordinated sentence within the super-ordinate sentence. Intonation is also criterial for the nature of the sentence, but since descriptions of Hebrew intonation have not yet been published there will merely be occasional comments in this area. The following survey will cover only (1) words marking subordination; (2) the order of the sentence elements.
שֶׁ is the most general subordinator. It can introduce an adjunct sentence and an object sentence (after verbs of saying), and can combine with other words to introduce various adverbial sentences (cf. 4.54: Prepositions as Introducers of Adverbials). In historical grammar שֶׁ is considered to have evolved from an ancient demonstrative pronoun, the ancient subordination marker זוּ (found in biblical Hebrew) being pointed to as a transitional form between demonstrative pronoun and subordination marker. But even if this claim can be proved right historically, it should not be taken into account in a consideration of the function of שֶׁ in modern Hebrew. It now serves solely as a subordination marker and there is no trace in it of an ancient demonstrative pronoun. No concord of any kind applies between it and what precedes it, and therefore it does not serve as "substitute for the subject," as some authors have alleged.
אֲשֶׁר. Although this word is typical of biblical Hebrew (where שֶׁ appears only in late passages), modern Hebrew uses it too. Indeed, אֲשֶׁר is found in all levels of the contemporary language, particularly as a stylistic variant for שֶׁ, when the latter occurs too often for the speaker's taste. This variation is restricted, since שֶׁ cannot be replaced in all its uses by אֲשֶׁר, this being one of the reasons for the relative infrequency of אֲשֶׁר as compared with שֶׁw in literary modern Hebrew. אֲשֶׁר cannot introduce the following structures which are related in deep structure and can be transformed from one another: (1) adjunct content-sentences (see 4.38: Relative Clause); (2) object sentences which can be transformed into adjunct content-sentences; (3) subject sentences derived from such object sentences by the change of the predicator verb from active to passive, or by its replacement by modal predicators such as טוֹב (good), יָפֶה (fine), הַלְוַאי (would that), חֲבָל (a pity), which express the attitude of the speaker to what is said in the subordinated part. The following are examples:
|Object content-sentence||Adjunct content-sentence|
|יָדַעְתִּי שֶׁרוּת עוֹבֶדֶת||הַיְּדִיעָה שֶׁרוּת עוֹבֶדֶת|
|(I knew that Ruth works)||(the knowledge that Ruth works)|
|Subject content-sentence||Subject content-sentence|
|טוֹב שֶׁרוּת עוֹבֶדֶת||נוֹדַע לִי שֶׁרוּת עוֹבֶדֶת|
|(It's good that Ruth works)||(It became known to me that Ruth works)|
In such sentences אֲשֶׁר can only be used exceptionally, in highly rhetorical language. Likewise, אֲשֶׁר does not introduce indirect speech, and generally it does not precede the participle.
כִּי introduces indirect speech, adjunct content-sentences, object sentences from which adjunct content-sentences can be transformed, and subject sentences derived from these object sentences by a change of the verb-predicator from active to passive. כִּי therefore is used in all structures not open to אֲשֶׁר. However, כִּי cannot introduce a subject sentence when the predicator is an initial modal expression such as הַלְוַאי (would that), חֲבָל (a pity), הַלְוַאי שֶׁיֵּרֵד שֶׁלֶג (would that snow will fall), but not הַלְוַאי כִּי יֵרֵד שֶׁלֶג ‡. Similarly, כִּי cannot introduce relative clauses. Another function of כִּי is to introduce adverbial sentences of cause (and in biblical Hebrew, also time and condition adverbials).
הַ introduces relative clauses that begin with the participle form of the verb, which agrees in gender and number with the nucleus.
ו mostly marks coordination rather than subordination. However sometimes, and especially in literature, it introduces a sentence that is or seems to be subordinated to the preceding sentence, when the subordinate sentence is one of circumstance, comparison, result, or purpose, e.g., הַנַּעַר יָצָא לַדֶּרֶךְ וְיָדָיו רֵיקוֹת (The lad went out and his hands were empty) – circumstance; הַשֶּׁמֶשׁ זָרְחָה, הַשִּׁיטָה פָּרְחָה, וְהַשּׁוֹחֵט שַׁחַט (The sun shone, the acacia blossomed, and the slaughterer slaughtered) – comparison; בַּקֵּשׁ רַחֲמִים וְיִסָּלַח לְךָ (Seek mercy and you will be forgiven) – result; ?מַה נַּעֲשֶׂה וְנִנָּצֵל (What shall we do and we shall be saved?) – purpose. See also 4.54: Prepositions as Introducers of Adverbials.
Generally, the order of the parts of the sentence in a subordinated sentence does not differ from that in an independent sentence (5.6: The Order of the Parts of the Sentence). But word order in relative clauses and conditional sentences may influence the nature of the link between them and the sentences in which they appear. In conditional sentences the place of the subordinated sentence may determine the nature of the link, usually together with other factors, such as intonation.
(1) A relative clause beginning with an inflected preposition that agrees in gender and number with the nucleus (see 5.21: Pronoun Concord in Focusing Sentence; 5.22: Pronoun Concord in a Relative Clause) can occur without the initial subordinator שֶׁ (or אֲשֶׁר) e.g., רָאִיתִי אֶת הַבַּיִת בּוֹ גָּר אָחִיךָ (I saw the house in which your brother lives), as opposed to רָאִיתִי אֶת הַבַּיִת שֶׁאָחִיךָ גָּר בּוֹ. It is usual to term such a relative clause as an asyndetic relative clause, since it lacks an initial conjunction. It may be argued that a marker of subordination is present, though not the usual שֶׁ, the inflected preposition filling also that function besides its function within the relative clause. In biblical Hebrew an asyndetic relative clause may occur also without this condition e.g., יֹאבַד יוֹם אִוָּלֵד בּוֹ. A relative clause without an inflected preposition may appear without a marker of subordination if it has a verb in the future agreeing with the nucleus and preceded by the negative participle בַּל, e.g., אֱמוּנָה בַּל־תְּעֻרְצַר (a faith that cannot be uprooted). It is impossible to add the subordinator שֶׁ or some other subordinator in initial position here. In such a structure the negative participle לֹא can replace בַּל but it is then possible to add שֶׁ or אֲשֶׁר, e.g., כֹּבֶד שֶׁלֹּא יְתֹאַר (weight that cannot be described) or כֹּבֶד לֹא־יְתֹאַר.
(2) Conditional sentences (see 4.52: Types of Conditional Adverbial) can be expressed without an initial conditional word under certain restrictions: (a) the protasis must come first, (b) the predicator must come first in the protasis. The absence of the conditional word is usual in legal language, e.g., מָצָא אָדָם חֵפֶץ, יְבִיאֶנּוּ לְתַחֲנַת הַמִּשְׁטָרָה הַסְּמוּכָה (A man has found an object, he shall bring it to the nearest police station). The future form of the second verb expresses not the condition, but the intention of the whole sentence to serve as a permanent instruction in all cases to which the condition applies. Similar conditional sentences are found in literature. However, they are not instructions, but refer to recurring events. In these the second verb may also be in the past, e.g., מָצָא כַּפְתּוֹרִים – הֱבִיאָם לְאִמּוֹ; מָצָא מַסְמְרִים – הֱבִיאָם לְאָבִיו (He found buttons – he brought them to his mother; he found nails – he brought them to his father). Proverbs may likewise have this form, e.g., in biblical Hebrew, מָצָא אִשָּׁה – מָצָא טוֹב (He found a wife – he found a good thing). When such sentences are said orally, the conditional sentences have characteristic intonation patterns. In the colloquial language, too, there may occur conditional sentences without an introductory conditional word. Only sentence order, word order, and intonation show them to be conditional, e.g., יִרְצוּ – יֹאכְלוּ, לֹא יִרְצוּ – לֹא יֹאכְלוּ (They will want – they will eat; they won't want – they won't eat). A change of sentence order, of word order, or of intonation will necessitate a conditional word, e.g., יֹאכְלוּ, אִם יִרְצוּ; לֹא יֹאכְלוּ, אִם לֹא יִרְצוּ (They will eat if they want; they will not eat if they don't want).
The following is not intended to serve as an exhaustive bibliography, but rather as an aid to the interested reader seeking additional information. It includes entries of three types:
(1) General research works on Hebrew linguistics and contemporary language problems.
(2) Publications containing particularly extensive bibliographical material.
(3) Publications presenting a wide variety of approaches.
a. general works: F.E. Koenig, Historischkritisches Lehrgebaeude der hebraeischen Sprache und historisch-comparative Syntax der hebraeischen Sprache, 3 vols. (1881–97); E. Kautzsch (trans. A.E. Cowley), Gesenius' Hebrew Grammar (191028); H. Bauer and P. Leander, Historische Grammatik der hebraeischen Sprache des Alten Testaments, 3 vols. (1918–22); G. Bergstraesser, Wilhelm Gesenius' hebraeische Grammatik, 3 vols. (1918–29); M.B. Sznejder, Torat ha-Lašon bĕhitpattĕḥutah, 2 vols. (1923–40); Š.Š. Kantoroviẓ, Diqduq ha-Śafa ha-ʿivrit lĕ-kol Signoneha, 2 vols. (1928); M.H. Segal, A Grammar of Mishnaic Hebrew (1927); idem, Diqduq lĕšon ha-Mišna (1936); M. Lambert, Traité de grammaire hebraïque (1938); Y. Avinery, Hekal Rashi, 4 vols. (1940–49); D. Yellin, Diqduq ha-Lašon ha-ʿivrit (1942); P. Joüon, Grammaire de l'hébreu biblique (19472); P. Joüon-T. Muraoka, Ẓ. Har-Zahav, Diqduq ha-Lašon ha-ʿivrit, 6 vols. (1951–55); N.H. Ṭur-Sinai, ha-Lašon wĕ-ha-Sefer, 3 vols. (1954–59); B. Klar, Meḥqarim wĕ-ʿiyyunim ba-Lašon, ba-Šira u-va-Sifrut (1954); W. Chomsky, Hebrew the Eternal Language (1952; Heb., 1967); R. Meyer, Die hebraeische Grammatik (19552, 19663), adaption of G. Beer's Grammar; E.Y. Kutscher, Ha-Lašon wĕ-ha-Reqaʿ ha-Lĕšoni šel mĕgillat Yĕšaʿyahu (1959); A. Sperber, A Historical Grammar of Biblical Hebrew (1966); A. Kašer (ed.), Balšanut Ḥiššuvit (1969); R. Macuch, Grammatik des Samaritanischen Hebraeisch (1969, see Z. Ben-Ḥayyim, in: Biblica, 52 (1971)); C. Rabin, in: Current Trends in Linguistics, vol. 6 (1971), 304–46. add. bibliography: M. Mishor, "The Tense System in Tannaitic Hebrew" (dissertation, Heb. Univ., 1983); N. Braverman, "Ha-Milliyot ve-To'arei ha-Po'al Bilšon ha-Tanna'im" ("Particles and Adverbs in Tannaitic Hebrew (Mishnah and Tosefta)," dissertation, Heb. Univ. Jerusalem (1995); P. Joüon, A Grammar of Biblical Hebrew, trans. and rev. by T. Muraoka, 2 vols. (1993). b. problems of modern hebrew: E.M. Lipschuetz, Vom lebendigen Hebraeisch (1920); I. Garbell, Fremdsprachliche Einfluesse im modernen Hebraeisch (1930); Ẓ. Har-Zahav, Lĕšon Dorenu (1930); Y. Klausner, Diqduq Qaẓar šel ha-ʿivrit ha-Ḥaˇdaša (1935); Y. Avineri, Kibbuš ha-ʿivrit bĕ-Dorenu (1946); Y. Epstein, Hegyone Lašon (1947); Y. Klausner, Ha-Lašon ha-ʿivrit Lašon Ḥayya (19492); E.M. Lipschuetz, Kĕtavim, vol. 2 (1949); A. Avrunin, Meḥqarim bi-lĕšon Bialikwĕ-Yalag (1953); Z. Ben-Ḥayyim, "Lašon ʿattiqa bi-Mĕẓiʾut Ḥaˇdaša," in: Lešonenu Laʿam, nos. 35–37 (1953); Ḥ. Rosén, Ha-ʿivrit Šellanu (1956); Y. Klausner, Ha-ʿivrit ha-Ḥaˇdaša u-vĕʿayoteha (1957); Y. Livni, Lašon Kĕ-Hilḵatah (1957); A. Bar-Adon, Lĕšonam ha-Mĕdubberet šel ha-Yĕladim bĕ-Yiśraʾel (1959); Y. Pereẓ, ʿivrit ka-Hālaḵa (19612); R. Sivan, Ẓurot u-Mĕgammot bĕ-Ḥidduš ha-Lašon ha-ʿivrit bi-Tĕqufat Tĕḥiyyatah (1963); A. Bendavid, Lĕšon Miqraʾ u-Lĕšon Ḥaḵamin (1967–692); M. Ben-Asher, Hitgabbĕrut ha-Diqduq ha-Normaṭivi (1969). add. bibliography: M. Bar-Asher, L'Hébreu Mishnique, études linguistiques (1999). c. phonology and morphology: E.A. Speiser, in: jqr, 16 (1925/26), 343–82; 23 (1932/33), 233–65; 24 (1933/34), 9–46; M. Segal, Yĕsode ha-Fonetiqa ha-ʿivrit (1928); J. Cantineau, in: Bulletin d'études orientales, 1 (1931), 81–98; L. Gulkowitsch, Die Bildung von Abstraktbegriffen in der hebraeischen Sprachgeschichte (1931); E. Porat, Lĕšon Ḥaˇḵamim (1938); C. Sarauw, Ueber Akzent und Silbenbildung in den aelteren semitischen Sprachen (1939); H. Birkeland, Akzent und Vokalismus im Althebraeischen (1940); C. Brockelmann, in: zdmg, 94 (1940), 332–71; G. Lisowski, Die Transkription der hebraeischen Eigennamen des Pentateuch in der Septuaginta (1940); Z. Ben-Ḥayyim; in: Lešonenu, 11 (1941), 83–93; E. Brønno, Studien ueber hebraeische Morphologie und Vokalismus der Hexapla (1943); D. Yellin, Tolĕdot Hitpattĕḥut ha-Diqduq ha-ʿIvri (1945); Y. Cantineau, in: Bulletin de la Société de Linguistique, 46 (1950), 82–122; R.W. Weiman, Native and Foreign Elements in a Language (1950); G.J. Botterweck, Der Triliterismus im Semitischen (1952); Ẓ. Elner, ʿAl Ḥuqqeha ha-Naʿaˇlamim šel ha-Nĕgina ha-ʿivrit (1952); V. Christian, Untersuchungen zur Lautund Formenlehre des Hebraeischen (1953); Y.G.P. Gumpertz, Mivṭĕ'e Śĕfatenu (1953); Z. Ben-Ḥayyim, Studies in the Traditions of the Hebrew Language (1954); I. Garbell, in: Homenaje a Millās-Vallicrosa, 1 (1954), 647–96; S. Moscati, Preistoria e storia del consonantismo ebraico antico (1954); Z. Ben-Ḥayyim, ʿivrit wa-aʾramit Nosaḥ Šomĕron, 5 vols. (1957–77); Ḥ. Rabin, in: Sefer Ṭur-Sinai (1960), 169–206; idem, in: Tarbiẓ, 30 (1960/61), 99–111; Š. Morag, Ha-ʿivrit še-bĕ-Fi Yĕhude Teman (1963); Ḥ. Yalon, Mavoʾ lĕ-Niqqud ha-Mišna (1964); K. Koskinen, in: zdmg, 114 (1964), 16–58; K. Aartun, ibid., 117 (1967), 247–65; E. Jenni, Das hebraeische Piʿel (1968); R. Mirkin, in: Lešonenu, 32 (1968), 140–52; Y. Kutscher, in: Sefer Zikkaron l-Vinyamin de-Vries (1969), 218–51; W. Weinberg, in: huca, 40–41 (1969–70), 1–32; E.
Brønno, Die Aussprache der hebraeischen Laryngale nach Zeugnissen des Hieronymus (1970); J. Blau, On Pseudo-Corrections in Some Semitic Languages (1970), 23–42, 114–25. add. bibliography: Y. Yeivin, Masoret ha-Lašon ha-Ivrit ha-Mištakkefet ba-Nikkud ha-Bavli (1985); O. (Rodrig) Schwartzwald, Dikduk u-Metzi'ut ba-Po'al ha-Ivri (1981); ibid., Perakim be-Morfologi'ah ivrit (2002); U. Ornan, Ha-millah ha-Aḥaronah: Mangenon ha-Tezurah shel ha-Millah ha-Ivrit (2003). d. syntax and style: S.R. Driver, A Treatise on the Use of the Tenses in Hebrew (18923); E. Koenig, Stilistik, Rhetorik, Poetik in Bezug auf die biblische Literatur (1900); Y.Ḥ. Ṭaviov, More ha-Signon (19022); J. Nodel, Der zusammengesetzte Satz im Neuhebraeischen, auf Grund der Mischna, der Tosefta und Midraschim (1928); G.R. Driver, Problems of the Hebrew Verbal System (1936); Y. Pereẓ, Taḥbir ha-Lašon ha-ʿivrit (1942); F.R. Blake, A Resurvey of Hebrew Tenses (1951); M. Gottstein, Taḥbirah u-Millonah šel ha-Lašon ha-ʿivrit še-bi-Tĕḥum Hašpaʿatah šel haʿaravit (1951); E. Lemoine, Theorie de l'emphase hebraïque (1951); M.M. Bravmann, Studies in Arabic and General Syntax (1953); C. Brockelmann, Hebraeische Syntax (1956); ʿU. Ornan, in: Lešonenu, 25 (1961), 35–51; F. Rundgren, Das althebraeische Verbum, Abriss der Aspektlehre (1961): E.J. Revell, A Structural Analysis of the Grammar of the Manual of Discipline (1962); U. Ornan, Ha-Ẓerufim ha-Šemaniyim bi-Lĕšon ha-Sifrut ha-ʿivrit ha-Ḥaˇdaša (1964); Y. Pereẓ, Mišpaṭ ha-Ziqqa ba-ʿivrit lĕ-Ḵol Tĕqufoteha (1967); Ḥ.B. Rosén, ʿivrit Ṭova (19672); E. Rubinstein, Ha-Mišpat ha-Šemani (1969); Y. Hayon, Relativization in Hebrew (1969); U. Takamitsu, Emphasis in Biblical Hebrew (1969); I.F. Anderson, The Hebrew Verbless Clause in the Pentateuch (1970); E. Rubinstein, Ha-Ẓeruf ha-Poʿoli (1971). add. bibliography: A. Abbadi, Taḥbir ha-Siyaḥ shel ha-Ivrit ha-Ḥadashah (1988); S. Fassberg, Sugi'ot be-Taḥbir ha-Mikra (1994); M.Z. Kaddari, Taḥbir ve-Semantikah be-Ivrit she-le-aḥar ha-Mikra, 2 vols. (1991–95). e. lexicon and semantics: J. Levy, Woerterbuch ueber die Talmudim und Midraschim, 4 vols. (1876–89); Ḥ.Y. Kohut, Sefer he-ʿaruḵ ha-Šalem meʾet Natan Ben-Yeḥiʾel, 8 vols. (1878–92); S. Krauss, Griechische und lateinische Lehnwoerter im Talmud, Midrasch und Targum, 2 vols. (1898–99); M. Jastrow, A Dictionary of the Targumim etc., 2 vols. (1903); F. Brown, S.R. Driver, and C.A. Briggs, A Hebrew and English Lexicon of the OT, Based on… Gesenius (1907); E. Ben-Yehuda, Millon ha-Lašon ha-ʿivrit ha-Yešana ve-ha-Ḥadašah, 16 vols. (1908–50); W. Gesenius, Hebraeisches und aramaeisches Handwoerterbuch ueber das at (ed. by Buhl; 192117); Y. Klatzkin, Oẓar ha-Munnaḥim ha-Pilosofiyim, 4 vols. (1928–33); S. Krauss, Tosefothe-ʾarukh ha-Šalem (1937); Y. Kenaʿani, Oẓar ha-Lašon ha-ʿivrit, 1–11 vols. (1948–71); L. Koehler, W. Baumgartner, Lexicon in Veteris Testamenti libros & Supplementum (1953, 1958, 19673); F. Zorell, Lexicon Hebraicum et Aramaicum Veteris Testamenti (1954); D. Sedan, ʾAvne Śafa (1967); Š.E. Loewenstamm, Y. Blau, Oẓar Lashon ha-Miqraʾ(ט-א), 3 vols. (1957–60); J. Barr, Semantics of Biblical Language (1961); Y.Y. Kutscher, Millim ve-Toldotehen (1961); R. Sappan, Darḵe ha-Sleng (1963); idem, Millon ha-Sleng ha-Yiśrĕ'eli (1965); Ẓ. Scharfstein, Oẓar ha-Millim wĕ-ha-Nivim (19643); Y. Avineri, Yad ha-Lašon (1965); A. Even-Šošan, Millon Ḥadaša, 5 vols. (1949–52); idem, Ha-Millon he-Ḥadaš, 7 vols. (1966–1970); Ḥ. Yalon, Mĕgillot Midbar Yĕhuda, Divre Lašon (1967); Y. Avineri, Gĕnazim Mĕgullim (1968); N. Stutchkoff, Oẓar ha-Safá ha-ʿIvrit (1968); M.Ẓ. Qaddari, Ha-Ḥiyyuv bi-Lĕšon ha-Mĕgillot ha-Gĕnuzot (1968); idem, Mi-Yrušat Yĕme ha-Benayim (1970); Ḥ. Rabin and Ẓ. Radday, ʾOẓar ha-Millim (1970). add. bibliography: J. Schweka, Rav Millim, Ha-Millon ha-Shalem Ivri-Ivri, 6 vols. (also on cd and updated version at: www.cet.ac.il/ravmilim) (1997); A. Even-Shoshan, Millon Even-Shoshan: Meḥuddash u-Me'udkan Lishnot ha-Alpayim (ed. Moshe Azar), 6 vols. (2003). f. periodicals:Lešonenu (1929– ); Lešonenu La-ʿam (1949– );Qunteresim le-ʿinyene ha-Lašon ha-ʿivrit (1937–8, ed. Ḥ. Yalon); ʿInyene Lašon (1942–3, ed. Ḥ. Yalon); Balšanut Ivrit Ḥofšit (1969– ). add. bibliography: Meḥkarim be-Lashon (1985– ).