A Plea for a Broad Definition of Ancient Hebrew

Ancient Hebrew (עברית עתיקה Ivrít atiqá) is a conventional designation for the language in which the Hebrew Bible is composed. To be clear, not all of התנ״ך ha-Tanakh – an acronym composed of the first letter of the titles of its successive parts: תורה Torá, נביאים Neviím, andכתובים  Ketuvím – is written in Hebrew. Ezra 4:8-6:18; 7:12-26, Dan 2:4b-7:28, Jer 10:11, and a phrase in Gen 31:47 are written in ancient Aramaic (ארמית עתיקה Aramít atiqá), the lingua franca of a cross-section of the ancient Near East in the neo-Assyrian, neo-Babylonian, and Persian periods.

Ancient Hebrew’s primary period of attestation covers a period of more than a thousand years, from about 1000 bce – 135 ce. Extant texts of this time frame, inclusive of, but by no means limited to, texts preserved in the Hebrew Bible, are written on stone, pottery, metal, papyrus, and parchment, and make use of a total of 22 graphemes to denote up to 25 consonantal phonemes. A few graphemes, ו w, י y, ה h, andא  (the glottal stop), serve first in word-final position and later, in certain medial contexts, to mark vowels. But even in fuller orthographies, the texts do not mark most vowels, and vowel markers are multivalent; for example, ו covers both /o/ and /u/ vowels. For these reasons, the phonology, morphology, syntax, and prosody of ancient Hebrew is, in varying degrees, a matter of reconstruction. The texts also contain delimitation markers, spaces and/or dots, separating one word and one text block from another.

Varieties of ancient Hebrew (AH) are documented within its primary period of attestation. The corpus is alive with linguistic variation, conditioned by factors as diverse as register, genre, provenance, and diachrony. This is true across the texts that make up the Tanakh, and across epigraphic epistolary texts covering the entire chronological range, from Tel Qeiyafa, Lachish, and Arad to Bar Kochba. Another example: the linguistic profile of the sentence literature of Proverbs differs from that of Ben Sira as preserved in the Masada fragments; and that of Ben Sira from that of Pirqei Avot, an example of middle Hebrew (MH) on the other side of the artificial divide of the linguistic continuum of pre-modern Hebrew. The diverse forms of Hebrew in which non-sectarian and sectarian compositions found at Qumran were written are also examples of ancient Hebrew.

The Hebrew of post-biblical Jewish tradition preserved thereafter in Jewish tradition, the Hebrew of the Mishna, Talmud, Midrash, Piyyut, and so on, may be classified as middle Hebrew (עברית בינונית Ivrít benonít).¹ Modern Hebrew (NH, “new Hebrew’) developed as a secularization of middle Hebrew from the middle of the nineteenth century forward, and received a push from pioneers like Eliezer ben-Yehudah. It is now the spoken and written tongue of the state of Israel, with 8 million speakers.

The phonological detail of the Tanakh preserved in ancient reading traditions - Palestinian, Tibero-Palestinian, and Standard Tiberian; the Babylonian and Yemenite; in the case of the Pentateuch, the Samaritan reading tradition; but also, what can be gathered from biblical manuscripts with plene orthography found at Qumran; phonetic transcriptions of names and untranslated words in the LXX and larger Hellenistic Jewish tradition, and of entire passages in fragments of the second column of Origen’s Hexapla – said detail has roots in the pre-Hellenistic period, but nonetheless reflects developments which occurred in the Hellenistic, Roman, and Byzantine periods. This is the secondary period of attestation of ancient Hebrew, in which specific reading traditions were preserved and recorded to an ever greater extent and with ever greater precision.

The most important Hebrew manuscripts of the kind just mentioned, of the Tanakh or portions thereof, range in date from about 850 to 1100 ce; for the Samaritan Pentateuch, from about 1100 to 1400 ce. The pinnacle of achievement within this period of attestation was attained by the masoretic codices of the standard Tiberian tradition. Pandects that encompassed the entire Tanakh, Codex Aleppo (ca. 915) and Codex Leningradensis (1009) in particular, for the precision and wealth of phonological and prosodic detail, are among the greatest linguistic artifacts of all time. The accuracy and completeness of the reproduction of linguistic data, the inclusion of variant readings, the panoply of marginal and end notes (in Aramaic, not Hebrew!), is unrivaled in the history of text transmission.  

Ancient Hebrew as pronounced today according to religious and/or academic convention is emphatically not the way it was pronounced during its primary period of attestation. The differences occasionally rise to the level of divergent grammatical analysis. This is undeniable once coeval sources of evidence for the pronunciation of the language are taken into account: in particular, place names and personal names from Judah attested in neo-Assyrian, neo-Babylonian, and late Babylonian documents. For example: banay barqa for בני ברק benei beraq; ḫazaqia(u) for חזקיהו ḥizqiyahu; iaua for יהוא jehu.² On the basis of the same evidence and Greek transcriptions, it is clear that the inventory of consonantal phonemes in ancient Hebrew in the primary period of attestation was larger at first. ע and ח, not just ש, did double-duty and stood for one of two possible phonemes. Nonetheless, the merger of these phonemes, and of שׂ /ś/ to ס /s/, was well underway in the last phases of AH in the primary period of its attestation. It is also clear that in the course of transmission as documentable in the secondary period of its attestation, some features of AH grammar came to be poorly understood and were less than adequately transmitted, for example, qal passives; infinitive absolutes; paragogic nuns; rarely, the vocalization of waw before a yiqtol. The phonology of Tiberian Masoretic Hebrew contains within it a great number of clues to earlier stages of ancient Hebrew phonology; nonetheless, it represents a development away from the phonology of ancient Hebrew as known to us from the primary period of its attestation. For example, sporadic defective spellings of vowels occurring beyond the nucleus of the stressed syllable in epigraphic Hebrew suggest that particular classes of verbs and nouns were stressed in all contexts in the First Temple period as they are in Tiberian Masoretic Hebrew in pausal position only. Thus we find וע֫ת instead of ועתה; שלח֫ת in place of שלחתי; הי֫ת instead of היתה. Other examples: insertion of an anaptyctic vowel in final position consonant clusters and before final position pharyngeals and laryngeals (but not the glottal stop), over against a progressive lack thereof in earlier data, working backwards; raising of /a/ to /i/ in closed unaccented syllables, e.g. Tiberian MT מגרש migraš over against magras (Jerome) and מזמור mizmor over against μαζμωρ mazmor (Secunda); reduction of /a/ to shwa in unaccented syllables, e.g. מגדו megiddo, over against cuneiform magiddu; mutation of /i/ to /a/ in originally closed accented syllables, e.g. גת gath over against gint (or gimt, with partial dissimilation) in neo-Assyrian sources. Finally, evidence from Persian period Aramaic, a language in an adstrate relationship with coeval Hebrew then and thereafter such that features of Aramaic made their way into Hebrew and features of Hebrew into Aramaic in communities in which both were employed, suggests that fricativization of non-emphatic plosives (בגדכפת) in post-vocalic position is a phenomenon which gathered steam from the Persian period on. The weakening or elision of laryngeals/ pharyngeals and the non-initial glottal stop, on the other hand, characterized spoken Hebrew in some times and places far more than others, and must be considered a possibility even in the earliest of times.

The sense of the texts in the Tanakh, the particular nuances of words and phrases in context, the eloquence of its rhetoric, the content of its instruction, and the poetics of its narrative, prayer, and prophecy, all this and more is a deserving object of study and a source of fascination for people of diverse persuasions. The language and literature in question can now be studied against a richer linguistic background than ever before. Whoever chooses not to study the Tanakh in the light of what can be learned from linguistic analysis of the entire corpus of ancient Hebrew texts, and of the corpora of extant texts in cognate languages on top of that, is committed to a self-defeating enterprise. Whoever chooses not to study biblical Hebrew in full awareness of the distinction between its primary and secondary periods of attestation is doomed to retroject elements of the latter into his understanding of the former.

¹ The division of Hebrew into three parts, old, middle, and new, is in accord with a tradition among Semitists that goes back to Gotthelf Bergsträsser (see bibliography below).

² Note the uncontracted diphthong in banay (so also LXX Codex B), the norm, presumably, in First Temple Judahite; barqa points to a qatl formation, whereas MT beraq comes across as an Aramaized version of baraq; ḫazaqia(u) points to a sentence name Yáhu prevailed, possibly reanalyzed as an apocopated yiqtol + yáhu in MT Hebrew (y^e)ḥizqiyahu Yáhu will prevail, or my strength is Yáhu; iaua for יהוא points to a sentence name He is Yah, with the long form of the 3sg pronoun, as in Qumran Hebrew: the forms hú’a and hí’a, spelled הא in both cases originally, may have been the norm in First Temple Judahite, analogous to héma and héna in the 3pl. Caution is in order because of the known tendency of proper names to fossilize earlier linguistic features.

Bibliography

Shmuel Aḥituv, Echoes from the Past: Hebrew and Cognate Inscriptions from the Biblical Period (tr. and ed. Anson F. Rainey; Jerusalem: Carta, 2008) [the vocalization of the inscriptions according to the principles of Tiberian Masoretic Hebrew, however helpful to the average reader, is unscientific and without historical foundation]; Gotthelf Bergsträsser, Introduction to the Semitic Languages. Text Specimens and Grammatical Sketches (trans. with notes and bibliography and an appendix on the scripts by Peter T. Daniels; Winona Lake: Eisenbrauns, 1983 [1928]) [a classic; still worth reading]; Margaretha L. Follmer, The Aramaic Language in the Achaemenid Period. A Study in Linguistic Variation (OLA 68; Leuven: Peeters, 1995 [must background reading for those interested in the study of linguistic variation across the corpus of ancient Hebrew]: W. Randall Garr, Dialect Geography of Syria-Palestine 1000-586 b.c.e. (Philadelphia: University of Pennsylvania Press, 1985) [Garr distinguishes between “Northern Hebrew” (Israelite) and “Southern Hebrew” (Judahite) on pp. 38-39 and 232-235]; Takamitsu Muraoka and Bezalel Porten, “Phonology,” in A Grammar of Egyptian Aramaic. Second Revised Edition (HdO 32; Leiden: Brill, 2003) 1-42 [the problems that arise in reconstructing the phonology of ancient Aramaic in its primary period of attestation are analogous to those encountered in the case of ancient Hebrew]; Gary Rendsburg, “Ancient Hebrew Phonology,” in Phonologies of Asia and Africa (Including the Caucasus) (ed, Alan S. Kaye; technical advisor, Peter T. Daniels; Winona Lake: Eisenbrauns, 1997) 65-83 [Rendsburg’s claim that “the [standard Tiberian] Masoretic Text more or less accurately reflects the pronunciation (or at least one pronunciation) of ancient Hebrew in the first millennium b.c.e., i.e., the time of the composition of the biblical books” (68) is belied by his own exposition; it does not accurately reflect such, except through the prism of phonological developments within the biblical period (1200-100 bce) and subsequent to it, developments standard Tiberian Hebrew preserves, which is not to say that the radical reconstructions of the phonology of ancient Hebrew proposed by Zellig Harris and Klaus Beyer, respectively, are on the right track; those reconstructions err in the opposite direction by projecting elements of “proto-Semitic” onto Hebrew of the First Temple period; for the rest, Rendsburg’s assumption that varieties of ancient Hebrew beyond “standard literary Judahite Hebrew” are isolable in the extant corpus of ancient Hebrew is undoubtedly correct, but, from a methodological point of view, he has things backwards: in order to have reliable comparative data to work with, the individual linguistic profiles of discrete corpora such as Iron Age epigraphic Hebrew, Job, Song of Songs, Qohelet, Ben Sira as found in the Masada fragments, Hodayot, 4QMMT, and so on, need to be thoroughly explored, compared with each other, and compared with “standard literary Judahite Hebrew,” before idiosyncratic features within the core corpus of literary Jerusalemite Hebrew are associated with alternative dialects, “Israelian” or some other; for Rendsburg’s latest attempt to isolate "Israelian" features, see idem, “Israelian Hebrew Features in Deuteronomy 33,” in Mishneh Todah: Studies in Deuteronomy and Its Cultural Environment in Honor of Jeffrey H. Tigay (ed. Nili Sacher Fox, David A. Glatt-Gilad; and Michael J. Williams, Winona Lake: Eisenbrauns, 2009) 167-184]  

In a subsequent post, I will take up the issue of linguistic variation across the corpus of ancient Hebrew in more detail.

Comments

[Stacy]

Amazing post! Thanks for sharing

[JohnFH]

Thank you, Stacy, for your blogging from a Muslim woman's perspective.

[lu]

This doesn't help me at all I was looking for stuff about what their land looked like and what their government was like. But, no I read all of this to find it's not what I wanted

[Hebrew Scholar]

Thanks for this definition of Ancient Hebrew. It might be useful to your readers to offer a similar definition of "Archaic Hebrew", which is often used to designate some elements of Biblical (or Ancient) Hebrew which are not easily explained, and therefore often lumped together and claimed to be archaic.

[JohnFH]

I'm not sure it's possible to speak of texts written in "Archaic Hebrew," though it stands to reason, on non-linguistic grounds (for example, considerations relative to the history of the religion of Israel), that texts like Deut 32 and Judg 5 contain "archaic Hebrew" features.