HISTORICAL DICTIONARY PROJECT.

Frequent commenter Paul Ogden has created a Wikipedia article for what sounds like a wonderful thing, the Historical Dictionary Project for Hebrew. That second link is from the official site of the Academy of the Hebrew Language, which sponsors the project; they say:

The overarching goal of the HDP is to present the history and development of the Hebrew lexicon, from the earliest occurrences of words down through their most recent documentation. Whereas similar historical dictionary projects in Europe merely brought citations from texts of recent centuries, the Academy’s HDP is based upon Hebrew texts up until 1100 CE, and large selections of literature from the periods thereafter until the founding of the State of Israel. It was decided to begin with texts from the post-biblical period, and thus the database reflects more than 2000 years of Hebrew writing.

The Wikipedia article adds that “The complete lexical archive will contain at least 25 million quotations.” In short, a Hebrew OED. Every language needs one!

Comments

  1. Garrigus Carraig says:

    This is, of course, splendid news. But I can’t figure out whether this gets us closer to an Arabic OED, or further away from one.
    As I typed that, I started leaning sharply toward the latter.

  2. When was the earliest Hebrew written that can be securely dated? Inscriptions, perhaps?

  3. Google and WKPD still have the power to thrill me.
    http://phys.org/news182101034.html

  4. Google and WKPD still have the power to thrill me.
    The Gezer calendar also dates from about 10th century BC. It seems there is debate on whether the language is Hebrew or a Semitic neighbor.

  5. I can’t figure out whether this gets us closer to an Arabic OED, or further away from one.
    Closer, at least technically. The core software for the Hebrew dictionary was designed to accommodate the Semitic triliteral root system and its native orientation is right-to-left reading. This strongly suggests that the software would serve an Arabic OED as well.
    Political-cultural considerations almost certainly preclude the adoption of the software for an Arabic OED. Furthermore, Wiki says there are several Arabic language academies. For such a project, it would be best if these bodies joined forces. I suggest that an Arabic OED will be a long time coming.
    ______________
    It’s doubtful the Hebrew dictionary will ever be produced in traditional book form. Part of the reason is cost and the wholesale move to electronic-only editions of reference works.
    The other reason is the nature of the triliteral root system. Following is a brief example based on the root ‘a-v-d עבד. The inherent notion is that of work.
    Words formed from this root include oved עובד, worker, employee; evved עבד, slave; ma’abida מעבדה, laboratory; ibud עיבוד, processing (as in ‘working’ a piece of metal); worked מעובד (as in a ‘worked’ piece of metal); shi’abud שעבוד, lien (plus serfdom and the like); and the common verb to work עבד. There may be more words derived from this root but I can’t recall any others right now.
    An electronic dictionary can easily show all these related formations on the same page. By contrast, looking up one of these words in a conventional dictionary would never lead you to the others.
    Noteworthy: The name of the biblical prophet Obadiah and the Arabic name Abdullah share the same root עבד, which in this case means worship.

  6. Garrigus Carraig says:

    Political-cultural considerations almost certainly preclude the adoption of the software for an Arabic OED.

    I have the sense — not sure if it’s accurate — that reverence for the language of Quran has led to a kind of taboo in the Muslim world on applying historical linguistics to Arabic.

    An electronic dictionary can easily show all these related formations on the same page. By contrast, looking up one of these words in a conventional dictionary would never lead you to the others.

    If you were to make the paper dictionary, you could arrange it by roots instead of words. In that way it would be akin to the Chinese dictionaries that are arranged by radicals (do they still make those?), except better, because the process of extracting the root from an Arabic word is clearer (I think) than that of extracting the radical from a Chinese character.

  7. Aren’t all Arabic dictionaries arranged by roots? Certainly mine are, except for dialect dictionaries that use romanization anyway.

  8. If you were to make the paper dictionary, you could arrange it by roots instead of words.
    ____________________
    Aren’t all Arabic dictionaries arranged by roots?
    I don’t know about Arabic dictionaries. Basic Hebrew dictionaries are not so arranged, though comprehensive dictionaries often include several of the possible forms under the root. The other forms will also be listed in their alphabetical order throughout the work.
    Arranging by root alone is fine if the dictionary is for language specialists. Such a system is not practical for everyday use and is utterly unwieldy in a bilingual dictionary because many, and perhaps most people, are unable to tease out the root from the form they’re looking up.
    Example: The root נגע naga carries the basic meaning of touching. הוא נגע (hu naga) means he touched. Another construction of the same root yields הוא היגיע (hu higia [hard g]). It means he arrived.

  9. The root נגע naga
    The letter N is weak in Hebrew. It behaves somewhat like the N in English five, Dutch vijf and German fünf — sometimes it’s there, other times it’s not. Hebrew for pig is khazir, while in Arabic it’s khanzir (kh as in Scots loch). Does anybody know what this phenomenon is called?

  10. Tom Recht says:

    Those are really two separate phenomena. The five/fünf difference is due to loss of n with compensatory lengthening of the preceding vowel; the khazir/khanzir difference is due to assimilation of n to the following consonant (khazzir) followed by simplification of the geminate.

  11. marie-lucie says:

    Paul O: The letter N is weak in Hebrew
    Surely not initially as in the word naga. Your examples of “weak N” have n inside a word, following a vowel and preceding a consonant. In many languages (not all) this is a weak position for a consonant. If this position is occupied by a nasal consonant, the nasality is very often anticipated while uttering the preceding vowel, which becomes nasalizedL this occurs for instance in the usual pronunciation of the a in can, which is not the same as the a in cat), often leading to the disappearance of the nasal consonant as in English can’t or want, where the original spoken sequence [an] has turned into a single, nasal vowel with no trace of a consonantal [n]). If the preconsonantal position is occupied by a non-nasal consonant, this consonant is affected by the following one and may lose its specificity, hence “geminates” or pairs of identical consonants, as for instance in Italian cattivo from Latin captivus.
    Incidentally, you give naga as a root. Where is the third consonant? and if this is one of the relatively few biconsonantal roots, why does it have a final vowel?

  12. There’s an ayin: ngʿ.

  13. Tom Recht says:

    In Hebrew, preconsonantal n-loss gave not nasalization but gemination: nC < CC (although modern Hebrew has lost gemination altogether). And strange though it sounds, this does happen in initial position: e.g. the imperative of naga, with the template CCaC, is ga < *ngaʕ.
    As for the missing third consonant, it is/was ʕ (voiced pharyngeal fricative), but this sound has been lost (or, prevocalically, merged with glottal stop) in standard modern Hebrew.
    Actually, though, even Biblical Hebrew had roots with what you might call a ‘zero’ third consonant, which are not the same as biliteral roots: e.g. shata ‘drank (3sg.m. past)’, template CaCaC; shote ‘drinking (ppl. m.sg.)’, template CoCeC. The missing final C shows up as t in these verbs only in the infinitive, e.g. lishtot, template liCCoC; presumably it comes from a *t that has been lost in all other forms, though I don’t know why this would be.

  14. Marie-Lucie: the root of “naga” is the triconsonantal ngʕ. The pharyngeal consonant is not pronounced word-finally in most modern Hebrew dialects.
    Tom Recht: I’m not so sure about this ‘simplification of the geminate’. If I understand you, you mean something like *ħanzir > *ħaz:ir > ħăzi:r. The last change, from *a to a-flavored schwa (“ħataf pataħ”), is hard to justify within Hebrew.
    However, Aramaic has ḥzi:r; Arabic, Akkadian and Ugaritic have the n in the root. So perhaps the word is an early borrowing from Aramaic into Biblical Hebrew. I don’t know anything about Aramaic historical phonology, so I have no idea how it lost the proto-Semitic *n.

  15. Tom Recht says:

    On second thought, that missing final consonant might have been a *y, which does show up in some forms; IIRC the Semitic cognates of these verbs actually point to *y not *t — maybe someone can confirm this? In that case, the t in the infinitive becomes stranger yet.

  16. Tom Recht says:

    YM: Oh, you’re right – the ‘pig’ word does have hataf-patah and no dagesh, so there’s something odd going on in that particular case. But the general process can be illustrated with Paul’s example of higia ‘he arrived’: hingiaʕ > higgiaʕ > higia(ʕ).

  17. Tom: The loss of the geminates is a later process, and is universal in most dialects (IIRC only Yemeni Hebrew still preserves gemination contrast).

  18. Tom Recht says:

    Yes, of course. It actually isn’t a proper sound change, presumably, but a contact phenomenon due to the revival of Hebrew by speakers of languages which lacked gemination.

  19. Semitic cognates of these verbs actually point to *y not *t — maybe someone can confirm this?
    s¹ t y.

  20. Tom Recht says:

    Thanks, MMcM!
    I now realize that final -t in the infinitive occurs in some other classes of verbs as well. I suppose these infinitives started out as feminine verbal nouns or something of that sort.

  21. marie-lucie says:

    Tom Recht: In Hebrew, preconsonantal n-loss gave not nasalization but gemination: nC > CC
    In this case, the nasal consonant follows a more general rule valid for all consonants (or a larger subset of consonants).
    And strange though it sounds, this does happen in initial position: e.g. the imperative of naga, with the template CCaC, is ga < *ngaʕ.
    I had not thought of a pre-consonantal, as opposed to a pre-vocalic, initial position, but the dropping of the first vowel places initial n in pre-consonantal position, which is weak according to the other examples you give. Does this nasal loss occur only with the sequence ng, or also with the nasal m, and/or with other consonants in second place?

  22. Tom Recht says:

    Marie-Lucie: m doesn’t undergo assimilation in Hebrew. n is generally assimilated to the following consonant unless this is a pharyngeal or glottal, but there are exceptions (some morphologically conditioned, others sporadic). There are also occasional cases of assimilation of the other coronals l t d, but not regularly.

  23. From MMcM’s link, I found a short entry on the assimilating Hebrew n at that site’s forum.
    Note the word / בת bat daughter // בנות banot daughters / and recall British slang bint.

  24. Tom Recht says:

    Also ish ‘man’, pl. anashim; although in that word the singular, if from *ʔinš, should have a short vowel in Biblical Hebrew, but actually has long i:, which is hard to explain.
    The strangest fact about Hebrew etymology which I’ve learned thanks to this thread is that ish ‘man’ and isha ‘woman’ appear to be completely unrelated. The former has an Arabic cognate with s, the latter one with th. This is like finding out that, say, French chat and chatte were etymologically unrelated.

  25. From what I read in Klein, it’s not clear at all that ʔiš and ʔanašim (from ʔenoš) are related, so you may have here not just two but three separate etyma.
    The dagesh in גַּע /gaʕ/, the sing. masc. imperative of נגע /ngʕ/ ‘touch’, has nothing to do with n-assimilation or gemination. There’s a class of verb roots beginning with n, l or y where the imperative drops the first root letter. The dagesh here just shows that the /g/ is not changed to a fricative, which it never is at the beginning of a word.

  26. Tom Recht says:

    From what I read in Klein, it’s not clear at all that ʔiš and ʔanašim (from ʔenoš) are related
    Gesenius thinks the same; this is what I was getting at with my mention of the long-i problem.
    The dagesh in גַּע /gaʕ/, the sing. masc. imperative of נגע /ngʕ/ ‘touch’, has nothing to do with n-assimilation or gemination. There’s a class of verb roots beginning with n, l or y where the imperative drops the first root letter.
    But surely that’s only explicable by assuming the n or y (or l, though I think that only drops in one verb) were lost through assimilation? “Delete the first root consonant if it’s n or y” sounds strange as a morphological mechanism.
    Alternatively, you might speculate that n/y weren’t originally part of the root in such verbs at all, but are template prefixes that got incorporated into the root (which isn’t implausible for n- since that actually occurs as a template prefix); and that the imperative forms conserve the older, biliteral root. But this would entail that Hebrew originally had no triliteral roots beginning with n or y, which seems implausible (or else that it had such roots but they lost their n/y in the imperative by analogy, which is somewhat less so).
    I’m well out of my Semitic depth here, in any case; probably these questions are clearly answered by the comparative data.

  27. Tom Recht says:

    Another argument that initial n-loss in the imperative is due to assimilation is that, just as n doesn’t assimilate to a following glottal or pharyngeal, it doesn’t drop in the imperative before such consonants: e.g. /nhg/ ‘behave; drive’, impv. nhag; /nʕl/ ‘put on a shoe’, impv. nʕal, etc.

  28. Tom Recht: Well, after all English male < masculus and female < femina are unrelated, despite their resemblance, if not for which we’d pronounce the latter femmel.

  29. Tom Recht says:

    True, but fe- isn’t a productive feminine affix in English; the apparent relationship of ish/isha is more like that of prince/princess.

  30. Fem- has become so, though, in femfan and fembot; in the future, someone could easily assume that female < fem- + male.

  31. Clarification re word order in Hebrew dictionaries:
    The triliteral root and the past tense of a verb (3rd person singular male) are usually identical, and dictionaries do arrange their entries using this form. But many and perhaps most words derived from the root begin with another letter, and sometimes the first two letters are not part of the root. Utility thus demands that these other forms be shown in alphabetical order.
    In addition, Modern Hebrew has admitted a huge number of loan words (beyond the Greek and Latin admitted in Talmudic times and the Arabic admitted during the Middle Ages). These words, with a few serendipitous exceptions and a measure of the Arabic loans, do not follow the typical root pattern and therefore can only be shown in alphabetical order.

Speak Your Mind

*