Нашел у себя статью Баруха об ударениях в иврите. Правда, она на английском. Но считаю нужным поместить ее здесь, потому что Барух часто на нее ссылался.
STRESS SYSTEM IN MODERN HEBREW
Dr. B. Podolsky
paper presented at the
Conference on Afroasiatic typology
University of Colorado, Boulder,
There is an age-old tradition of viewing Semitic languages as lacking phonemic stress. Some linguists attempted to explain various forms in related languages as reflecting different ancient stress patterns in Proto-Semitic (e.g. Dolgopolsky 1978).
As a rule Hebrew, both Classical and Modern, is viewed as having a simple stress system which tends to be word-final, with a few exceptions.
Since Modern Hebrew sprang into being in early 20th century on the basis of Classical Hebrew, let us begin with a presentation of the stress system in the classical language.
In the old system, that of the Biblical Hebrew as it appears in its vocalized form, the stress is primarily word-final; in certain categories of words it falls on the penultimate syllable. These categories fall into two groups:
A. due to phonetic reasons
B. due to structural reasons.
To group A belong:
A1. The so-called segolate words, i.e. words which had once ended in a consonant cluster. At a certain period of time this final cluster became unpronounceable and was split by an auxiliary vowel, in most cases e (denoted in vocalization by the segol sign, hence the name segolate); whenever the last or the last but one consonant is a guttural, the inserted vowel is a; when the last or the last but one consonant is y, the auxiliary vowel is i. The auxiliary vowel is unstressed and stress is on the penultimate syllable. Exx. *yald > yéled ‘child’, *sipr > séfer ‘book’, *gubh > góvah ‘height’, *na‘r > ná‘ar ‘youth’, *bayt > báyit ‘house’, *biky > béxi ‘weeping’.
A2. Whenever the final consonant is a guttural (‘, ת, h) preceded by a non-low vowel (that is any but a), an auxiliary unstressed a vowel is inserted before the guttural:
*šabu‘ > šavúa‘ ‘week’, *gaboh > gavóah ‘high’, *śameת > śaméaת ‘glad’, *qali‘ > qalía‘ ‘bullet’ (the length of the vowels is unmarked since it is irrelevant to the problem).
Whenever the condition is missing, i.e. when a suffix is added and the guttural is not in word-final position, there is no auxiliary vowel and the stress falls on the last syllable: yaldí ‘my child’, *baytuhu > beto ‘his house’, ś#meתa ‘glad f.sg.’. This means that the auxiliary vowel is not part of the phonemic representation of the word and has no influence on the stress system.
The group B is a different story. Here we find the penultimate stress dictated not by phonetic reasons but by the morphological structure of the word.
B1. The directional suffux of the nouns actually turning them into adverbs is unstressed: ha‘íra ‘to the city’ from ha-‘ir ‘the city’, ×afóna ‘northward, to the north’, daróma ‘southward, to the south’.
B2. In certain bisyllabic pronominal suffixes the stress is on the last but one syllable: yaldénu ‘our child’, horéxa ‘your parents’, śadéhu ‘his field’.
B3. There are four nouns which have a stressed vowel -í- before the pronominal suffix: ’avíxa ‘thy father’ (cf. betIxá ‘thy house’), ’aתíxa ‘thy brother’, תamíxa ‘thy father-in-law’, píha ‘her mouth’.
B4. In suffixal conjugation of verbs (past tense) the suffixes of first and second person (except second person plural) are unstressed when followed by a word boundary, so that the last syllable of the stem carries the stress: katávti ‘I wrote’, haláxta ‘you (m.sg.) went’, unlike k#tavtém ‘you (m.pl.) wrote’.
B4.1. Whenever a pronominal suffix is added, the stress moves one syllable to the right: ra’íti ‘I saw’ - r#’ití+xa ‘I saw you (m.sg.)’, ra’ínu ‘we saw’ - r#’inú+m ‘we saw them (m.pl.)’.
B5. In various forms of two root types (gzarot): mediae infirmae (with medial yod or waw) and mediae geminatae (roots of the type 122) the stress in suffixal and prefixal conjugations is on the stem: qámu ‘they stood up’ (root qwm), yaqúmu ‘they will stand up’.
B6. The same is true with respect to the causative (hiph‘il) stem of any verb: hisbíru ‘they explained’, yaxnísu ‘they will put in’.
B7. In the archaic pausal forms of the verb (in nouns there is only change in the stem vowel, not in the stress) the stress is on the stem and not on the suffix: šamáru (instead of šamrú) ‘they kept’, yišmóru (instead of yišm#rú) ‘they will keep’.
In descriptions of Modern Hebrew it is usually mentioned that stress is phonemic in certain words, like táam ‘taste (noun)’ – taám ‘he tasted’, or bíra ‘beer’ – birá ‘capital city’ (Glinert 1989, 10: ‘Word stress is on the last syllable, unless specifically transcribed with an acute accent. It is occasionally phonemic, e.g. taám – táam ‘tasted ~ taste’.). Otherwise the system is presented as identical to that of the classical language. Only rarely do we find mention of the stress on the third syllable from the end (Falk 1996, p. 6).
Actually the picture is different. There are instances not only of antepenultimate stress, like ámbulans, télefon, but also on the fourth syllable from the end: kibúcnikiyot ‘female kibbuts members’ or kámeriyot ‘camera adj. f. pl.’, but these are usually not mentioned.
Stress has not been described, to the best of my knowledge, as a system; its behavior in the paradigm was not considered at all.
In fact the stress pattern in Modern Hebrew is quite different from what seems to be the picture when one views only the basic lexical forms. When we look at the whole paradigm, a new and quite interesting system is unrolled in front of our eyes.
To begin with, the classical system is preserved in most cases, the most obvious exception being that the 2nd p. pl. suffixes in past tense (see remark in B4 above) are unstressed in the spoken language: katávtem / katávten ‘you m./f. pl. wrote’, although the literary norm is like in the classical language.
Important remark: even though Modern Hebrew has lost the pharyngeal consonants phonetically (ת > x, ‘ > zero), phonemically, or to be more precise morphophonemically, they are preserved and distinguished from the velar x and the zero which stems from aleph [/], so that A2 rule is retained: šavúa‘ > šavúa ‘week’, śaméaת > saméax ‘glad’, qalía‘ > kalía ‘bullet’. H in final position is usually dropped; in more careful speech there occurs metathesis -ah > -ha: gavóah > gavóa ~ gavóha ‘high’.
Forms of verbs with pronominal suffixes (B4.1) are found in the written language only and are read according to the classical norm.
Pausal forms are archaisms which are found in modern language as set expressions retaining the classical stress.
Still there are numerous cases in Modern Hebrew in which stress cannot be explained by these rules.
As has been mentioned above, stress can fall on one of the four syllables counting from the end. What is more important, in some forms the stress moves to the suffix whereas in other cases it remains on the stem. Consider the following examples: gan ‘garden’ – pl. ganím, but gen ‘gene’ – pl. génim; milyón ‘million’ – pl. milyónim, but milyonó ‘his million’, milyonéy ‘millions of (pl. constr.)’.
In order to explain this phenomenon we’d like to introduce the notion of two types of stress: heavy and light, which are found both in lexical stems and in affixes and which differ in their behavior. We shall notate the light stress with the acute sign and the heavy stress with circumflex: ganím vs. gênim.
In stems the light stress is usually found in verbs and in old words inherited from previous stages.
It can be either word-final or penultimate: rexóv – rexovót ‘street, streets’, yéled – yeladím ‘child, children’.
The heavy stress is found primarily in recently borrowed words, but also in abbreviations which function as words (like mankâl ‘manager’, abbreviated from menahél klalí), and in a few words of children’s lexicon, like bûba ‘doll’. In some cases words from children's lexicon penetrate the speech of the grown-ups, like glîda – glîdot ‘icecream’. Sometimes there is an opposition, e.g. the grown-ups use bulím ‘post stamps’ whereas children collect bûlim.
Both types of stress can be found not only in words but also in morphemes, e.g. the plural suffixes -ím, -ót have the light stress, whereas -âim, -aôt have the heavy stress, e.g. gêto – getaôt ‘ghetto’. The word univêrsita ‘university’ has two plural forms: univêrsita +-ót = univêrsitot and univêrsita +-aôt = universitaôt.
The pronominal suffix (first person sg. possessive) -î has a heavy stress: axî ‘my brother’. On the other hand, the derivational suffix -í has a light stress: rusí ‘Russian’ vs. sîni ‘Chinese’ (see below).
The plural construct suffix -êy has a heavy stress, as well as pronominal suffixes: milyônim, milyonô, milyonêy ‘millions, his million, millions of’.
What is most important is how a word behaves when a suffix is added to it. The rule is as follows:
light + light > unstressed - stressed: sapár + ím > saparím ‘hairdresser, -s’
light + heavy > unstressed - stressed: sapár + êy > saparêy ‘hairdressers of’;
heavy + light > stressed - unstressed: milyôn + ím > milyônim ‘million, -s’;
heavy + heavy > unstressed - stressed: milyôn + êy > milyonêy ‘millions of’.
In other words the stress tends to move to the right (in transcription, not in the original Hebrew right-to-left script) unless a heavy stress precedes a light one.
Some borrowed affixes carry a heavy stress, like -îst, -îzm, whereas others demand the preceding syllable to be stressed; cf:
kibúc, kibucím, kibucêy ‘kibbutz’ vs.
kibûcnik, kibûcnikim, kibûcnikit, kibûcnikiyot ‘kibbutz member (m.sg., m.pl., f.sg., f.pl.)’. Here the suffix -nik precludes movement of the stress. Such case can be notated as ˆ-nik.
There are a few borrowed words in which the stress moves in a slightly different way:
dôlar – dolârim (obviously due to the Yiddish dólar – dolárn), têlefon – telefônim, âmbulans – ambulânsim, ôtobus - otobûsim. Here although the added affix has a light stress, the heavy stress of the stem moves to the penultimate syllable.
In derived words the character of the suffix dictates the behavior of the word. Thus, the borrowed suffix -îst turns the noun into a heavily stressed one, like bicuîst – bicuîstim ‘man of good practical performance’ < Hebrew bicúa (Classical bi××ūa‘) ‘fulfilment, realization’. On the other hand, the loanword čêlo ‘cello’ with the Hebrew suffix -án produces a new word with a light stress: čelán ‘cellist’ – pl. čelaním.
There are also two specific functions of the stress: stress shift one syllable to the left can distinguish a personal name from a common noun, e.g. Xáyim vs. xayím ‘life’, Yóna vs. yoná ‘pidgeon’, Sára vs. sará ‘woman-minister’.
The other specific usage of a similar stress shift in personal names denotes familiarity or intimacy versus formality. Thus, one says Davíd Ben Gurión, but my friend Dávid.
A radio journalist announces: -I have invited the well-known singer Yehorám Gaón – Shalóm Yóram! Here in addition to stress shift we see the omission of the weak consonant h together with the preceding vowel.
There exists also a very interesting connection of the stress pattern with morphology.
In gentilic names with the light word-final stress we find two types of feminine: anglí – angliyá – anglít ‘English’, rusí – rusiyá – rusít ‘Russian’, where the form in -yá refers to a human being; it need not be a noun but can also function as an adjective: soféret angliyá ‘an English female writer’. Very few (three or four) exceptions have been found, like yisreelí ‘Israeli’, litaí ‘Lithuanian’ – fem. only yisreelít, litaít.
In names with a heavy stress on the stem only one feminine form exists: sîni – sînit ‘Chinese’, amerikâni – amerikânit ‘American’. A plausible explanation of the phenomenon is that the feminine suffix -yá must be stressed; but it cannot win over the heavy stress on the stem, and hence only the form in -t exists.
In a few adjectives of recent coinage the stress is moved from the last to the penultimate syllable creating a distinction: pertaining to a language group versus to a specific language, e.g. germâni ‘Germanic’ : germaní ‘German’; români ‘Romance’ : romaní ‘Roumanian’.
Dolgopolsky, A. 1978. On phonemic stress in Proto-Semitic, Israel Oriental Studies VIII: 1-12.
Falk, Y.N. 1996. Metrical structure in Modern Hebrew, Hebrew Linguistics 40: 5-19.
Glinert, L. 1989. The grammar of Modern Hebrew. Cambridge University Press.
Mel’ֶuk, I. and B. Podolsky. 1996. Stress in Modern Hebrew nominal inflection, Theoretical Linguistics 22(1/2): 155-194.
Podolsky, B. 1991. “The problem of word accent in Modern Hebrew”. In: H. Mukarovsky (ed.). Proceedings of the Fifth International Hamito-Semitic Congress, Band 2: 277-281.