Archive for January 2009

Kaufman, Stephen A., “The Phoenician Inscription of the Incirli Trilingual: A Tentative Reconstruction and Translation,” MAARAV 14.2(2007): 7-26

January 26, 2009

The Incirli trilingual, not to be confused with Zincirl, was discovered in 1993. Unfortunately it had been exposed to the elements and is extremely weathered. Further complicating the reading, it was overwritten in Greek and is therefore sort of a stone palimpsest. A new website (much better than the old one) has been put together here  with pictures of the stone and more information. Pictures are also available from Inscriptifact. On seeing the stone, you will understand why this is presented as a tentative reading. 

The three languages of the original inscription are Phoenician, Hieroglyphic Luwian, and Neo-Assyrian. The two line Neo-Assyrian introduction has been read tentatively as well, but is not presented in this article. Unfortunately, Dr Kaufman reports that the Luwian and the rest of the cuneiform seem beyond reconstruction. 

The inscription is attributed Awarikku, King of the Danunites and commemorates a gift of land from Tiglath-Pileser III. Awarikku seems to have been rewarded for maintaining loyalty to Assyria during the well-known revolt led by Matiel of Arpad in the late 740′s BCE. It seems to follow the form of other lengthy stela texts by beginning with an introduction, the historical background for the erection of the stele, the narrative of the revolt, and finally ending with curses. I will leave the transcription and translation for you to read in the article, but summarize some of the more interesting findings and implications here.

First, Tiglath-pileser III is also referred to as Puˀ/wal with an intervocalic glide, spelled פאל quite clearly in at least one place (and probably the others) in contrast to biblical פול, vocalized Pûl. If it were pronounced according to the latter then one would expect פל in the Phoenician orthography. Also known as Pulu in later Babylonian records, this finally gives contemporary confirmation that Pul and Tiglath-Pileser III were indeed one and the same king. Also interesting is that Tiglath-pileser is spelled תכלתאפלסר (Tukulti-Apil-Ešarra) which corresponds to Assyrian orthography rather than the phonetic pronunciation as found in other Northwest Semitic texts (תגלתפלסר in 2 Kg 16:7 and תגלתפליסר in Samalian).        

More Assyrian influence is found in line 5 where Pul is referred to as פאל מל[ך] אשר רב “Puwal, the great king of Assyria”. Here, רב is the Assyrian word for great which seems to be reserved  for the Assyrian emperor as opposed to the “kings” of provincial areas and vassal states. See Hosea 5:13 and 10:6 where the Assyrian king is referred to as מלך ירב

Line 8 reads  אנך וריכס מלך ז בת מפש “I am (A)warikus, King of the dynasty of Mopsos”. Note the use of the demonstrative pronoun ז for the relative/genitive, similar to archaic Hebrew (יהוה זה סיני “YHWH of Sinai” Jd 5:5) and זי / די in Aramaic. This is not usually recognized by the grammars as a relative in Standard Phoenician which instead uses אש (though Old Byblian does use ז which was later replaced by אש).

Line 20-21 on the reverse reads וכרת ב.. ארצת אצר ותחת כל בת נפש which Dr. Kaufman translates (emphasis his), “Then I mined the treasure lands and beneath every tombstone.” Though a bit damaged, the reading כרת from כרה “to dig, mine” seems clear. In Is 3:20 וּבָתֵּ֥י הַנֶּ֖פֶשׁ refers to an “amulet worn around the throat”, but that seems to make no sense here. In later Aramaic and Post-Biblical Hebrew, נפש means “funerary monument”, thus something like “tombstone” is all that can make sense (I wonder, might this reading take some of the luster from Kuttamuwa’s so-called “soul”?).

Most importantly, perhaps, is the discussion within the narrative section of the practice of sacrificing a royal offspring in a crisis (see Mesha in 2 Kg 2:27) using the terms גזר, זבח and כפר. Unfortunately connections to the Punic molkomor and biblical Molek are not completely certain. In lines 11-12, Kaufman reads:

 וזבח מלך ארפד ליען הדד מלך וגזר מצפר כ ארפד פחד מלך אמ/שר חי/על

“And the king of Arpad (Matiel) sacrificed for the benefit of Hadad-Melek (or “for the purpose of a molk-offering for Hadad”), and redeemed [the human sacrifice] with butchered animal parts, because Arpad feared (a living molkomor/ the King of Assyria. He went up…)”

Thus in the first case הדד מלך can either be taken as one word, the god Hadad-Melek (See also Kaufman’s “The Enigmatic Adad-Milki”) or as two, “for Hadad, a molk“. In the second case we can either read that Arpad was afraid of a מלך אמר חי or that they were afraid of the מלך אשר, in which case על is a verb starting the next phrase “He arose…”

The Study of Classical Hebrew Poetry: Meter

January 23, 2009

The Standard Description of Hebrew Poetry distinguishes two main axes: parallelism and meter. In my last post I reviewed the study of parallelism, now I will take up the question of meter. I should note, however, that I find the distinction of parallelism from meter to be unnatural. As Jakobson noted, the basic feature of all poetry is the imposition of structured patterns on all levels of language, thus parallelism is merely the syntactic corollary to meter. The distinction of the two arose early in the study of Hebrew poetry based on the contrast with Western poetry, which seemed to rely more heavily on rigid metrical patterns (at least in its classical forms) while knowing nothing of the type of rigid semantic parallelism so obvious in Hebrew Poetry (again, at least in its classical form).

In poetry, meter refers to the number of syllables in a poetic line and their type (unaccented/short, usually marked by the breve ˘, or accented/long, usually marked by the macron ¯). The foot is the basic building block of meter and describes the smallest pattern of syllables. Common examples include the iamb (˘ ¯), the trochee (¯ ˘), and the anapest (˘ ˘ ¯). Thus a line of Shakespeare in iambic pentameter contains five iambic feet (stress marked with capital letters):

Shall I com-PARE thee TO a SUM-mer’s DAY?

English meter then is based on rigid patterns of feet, which are arranged into patterns of lines, which are arranged into patterns of stanzas, etc. A poet, of course, has license to vary the pattern slightly when needed or for effect (eleven lines of da-DUM da-DUM da-DUM would make you go crazy). Thus, for example, an iamb can be inverted to a trochee (“THAT is” in the example below) or a final unaccented/short syllable may be considered extra-metrical (“the QUES-tion”):

to BE or NOT to BE THAT is the QUES-tion

As noted above, in classical Hebrew poetry it appears that semantic parallelism  is the major structural element and not regularity of meter. Combined with the difficulty of reconstructing the ancient pronunciation of the Hebrew texts, many scholars have followed Lowth in resisting a rigid description of meter beyond noting the number of “parallel members” in  bicolon or tricolon. Still, the quest for meter has proven too enticing for others, leading to many complicated schemes.   

The most widely accepted method is ascribed to Julius Ley and Edward Sievers, along with Karl Budde, and is generally described as an accent-counting system. Ley first developed the idea (Grundzüge des Rhythmus, des Vers- und Strophenbaues in der hebraischen Poesie, 1875) that the characteristic of a line of Hebrew poetry is the number of tone or accented syllables without regard for the number of unaccented syllables. Thus a poetic line could be measured by the number of “stress-units” (i.e. a construct chain would count as one unit since only the final word maintains its accent), irrespective of the type of feet (i.e. just “pentameter” rather than “iambic pentameter”). In standard Ley-Sievers notation, a bicolon can be described as 3 + 3 (two cola of 3 accents each),  2 + 2 + 2 (three cola of 2 accents), etc.

Budde built on Ley’s work by extensively examining the so-called “elegaic pentameter”, an un-balanced 3 + 2 line (3 accent groups in the first colon, but only 2 in the second) which was common in the Hebrew elegy or dirge. Budde concentrated on Lamentations chapter 3, where the acrostic form made the demarcation of poetic lines obvious (“Das Hebräische Klagelied”, ZAW 2, 1882). He named this pattern Qinah meter after the Hebrew קִינָה, “dirge”. For example, Lam 3:25 is a bicolon following the 3 + 2 pattern:

תִּדְרְשֶֽׁנּוּ לְנֶ֖פֶשׁ לְקוָ֔וֹ יְהוָה֙ ט֤וֹב
seeks-him to-the-one-who to-those-who-wait-for-him is-the-Lord good

Sievers contribution was the application of more rigorous phonetic principles (Studien zur hebraischen Metrik, 1901). Sievers counted the number of accented syllables per line, similar to Ley and Budde. However, while Ley felt that the number of unaccented syllables was relatively unimportant, Sievers argued that the basic metrical foot consisted of four “beats” – unaccented syllables counting one beat, but long, accented syllables counting two. Thus he affirmed the anapest (da-da-DAH) as the basic foot of Hebrew poetry. Seivers further developed rules for counting secondary stresses, handling monosyllabic words and particles, dropping stresses (as at the end of the line in the example above), etc.

The other widespread theory of meter is the iambic approach, generally traced back to Gustav Bickell (“Die hebraische Metrik I-II”, ZDMG 34 and 35, 1880/81). Bickell based his study on Syriac poetry (a native Semitic language in contrast to the classical and western poetry used by other scholars), arguing that the basic foot is not the anapest, but the iamb/trochee. Thus there is an alternating rhythm of stressed and unstressed syllables, though unstressed syllables can sometimes be dropped due to syncope. Sigmund Mowinckel has developed this idea further (see The Psalm’s in Israel’s Worship with a chapter on meter) as has Stanislav Segert. 

Mowinckel’s basic argument is that the Psalms are for singing, and this implies that there must be some regular metrical component. He then suggests a diachronic approach may be the best way to understand the apparent irregularity. In early folk poetry, parallelism dominates and it is the stress-group which is important, the most common form being the dipod, a bicolon of 2 + 2. However, the rhythm follows the natural prosody of the language and is therefore “irregular”. As the classical poetry develops however, the meter becomes more regularized, and, since Semitic languages are naturally iambic, it is only natural that the iamb comes to dominate. In fact, he finds that the Psalms show very regular meter when seen in an iambic rather than anapestic pattern.    

A third approach is found in the studies of Cross and Freedman, who count the total number of syllables in a line regardless of accent. Strictly speaking, this is not meant to be a description of meter since there is no analysis of the types of syllables or feet, but merely a quantitative description of the structure of poetic lines. Cross distinguishes two types of line based on length – longum (l) and breve (b). In archaic poetry there seems to be a preference for the patterns l:l or l:l:l. when a short line is used however, it is b:b::b:b, corresponding to l:l:l. Freedman has also analyzed Budde’s Qinah meter, finding that such lines also show an unbalanced pattern of 7 + 6 or 8 + 5 syllables. This is not to suggest that Hebrew poets were actively counting syllables, but the regularity of lines suggests that rhythm and meter is a component of Hebrew poetry, even if we cannot completely reconstruct its internal contour.

The Study of Classical Hebrew Poetry: Parallelism

January 22, 2009

I believe Michael O’Connor coined The Standard Description of Hebrew Poetry as a label for the basic framework by which most contemporary scholars analyze Hebrew poetry. This method has two basic components – parallelism and meter – which are combined to analyze the structure of a poem, ie versification, stanzas, etc. In this post I will give a brief survey of parallelism, leaving meter for another post.

The modern history of parallelism conventionally begins with Bishop Robert Lowth (De Sacra Poesi Hebraeorem Praelectiones Academicae, 1753 translated into English as Lectures on the Sacred Poetry of the Hebrews, 1815, and more systematically presented in the introduction to his commentary on Isaiah). As O’Connor remarks, Lowth is not to be credited with the discovery of parallelism itself, rather his contribution is in developing the connection of parallelism to poetic meter and versification. Lowth is also credited with coining the term parallelismus membrorum (parallelism of members). Lastly, Lowth’s work led to the recognition that the prophets also wrote in poetry, thus the corpus of Biblical Poetry was not limited to the Psalms.

Lowth assumed that Hebrew poetry had meter, but that its exact nature had been lost with the ancient pronunciation. However, the building block of Hebrew poetry was what he called the “sentential form”, that is, a unit of poetry usually corresponds to a simple sentence. This is then followed by one or more parallel sentences of equal length which either repeat, vary, or amplify the same idea, each member in the first unit having a complement in the following ones naturally creating a metrical cadence. Lowth further defines three basic types of parallelism: synonymous, antithetic, and synthetic.

This forms the basis for the Standard Description. Hebrew verse is analyzed as groups of lines (or better, cola) which are further subdivided into member units based on the parallelism across cola. For instance, Is 40:3:

c’ a’ b’ c b a
מְסִלָּ֖ה לֵאלֹהֵֽינוּ בָּעֲרָבָ֔ה יַשְּׁרוּ֙ דֶּ֣רֶךְ יְהוָ֑ה פַּנּ֖וּ בַּמִּדְבָּ֕ר ק֣וֹל קוֹרֵ֔א
a-highway-for-our-God in-the-desert make-straight the-way-of-the-Lord prepare in-the-wilderness a-voice-cries:

This verse is a bicolon (“A voice cries” seems to be outside of the poetic structure, at least based on the parallelism). The two cola are identified by their constituent parallel elements, which do not continue into Is 40:4. The parallelism is synonymous since each member in the first colon has a synonym in the second. Note that these members are not necessarily in the same order, nor is a member necessarily comprised of a single word.

George Buchanan Gray (The Forms of Hebrew Poetry, 1915) agreed basically with Lowth that Hebrew poetry contained meter, but that no system to his time described such meter satisfactorily. Building from Lowth, he pointed out that while the first two categories of parallelism are straightforward, the last (synthetic) had become a catch-all for any parallel element other than a simple synonym or antonym. He also added the category of incomplete parallelism, in which an element in one colon has no exact parallel in the second. For instance, Is 60:2:

c’ a’ c b a
לְאֻמִּ֑ים וַעֲרָפֶ֖ל אֶ֔רֶץ יְכַסֶּה־ הַחֹ֨שֶׁךְ֙ כִּֽי־הִנֵּ֤ה
the-people and-thick-darkness the-earth shall-cover The-darkness For behold::

Again, based on the parallelism, it appears that “For behold:” is outside of the poetic structure. Notice how the b-element (“shall cover”) has no complement in the second colon (this is also called ellipsis or gapping in the current literature). It should also be noted that Gray is the one who introduced the system of using letters (a.b.c. ||  a’.b’.c’) to mark the parallel members.

As is often the case in scholarship, later scholars have multiplied or modified Lowth’s categories of parallelism. Thus you will find reference to climactic parallelism, stair-step parallelism, etc. James Kugel has argued against the idea of simple synonymous parallelism, suggesting instead that the second parallel element always builds or adds to the first, and is never merely symmetrical.

Advances in the science of linguistics have also influenced the study of parallelism. The pioneer was Roman Jakobson (see Selected Writings: Poetry of Grammar, Grammar of Poetry, 1981), who states succinctly that “the poetic function projects the principle of equivalence from the axis of selection to the axis of combination”. Jakobson sees the hallmark of poetry to be the organization of language on all levels – phonetic, morphological, syntactical, lexical, phraseological – into characteristic types of patterns. Thus he saw parallelism in Hebrew poetry not as an oddity, but as the syntactic counterpart to regular metrical patterns.

Jakobson’s influence can be seen in the work of Adele Berlin (The Dynamics of Biblical Parallelism, 1985) who seeks to move beyond Lowth’s static categorization to describe the ways the different levels of language interact. Edward Greenstein, Terrence Collins (Line-Forms in Hebrew Poetry, 1978) and Michael O’Connor (Hebrew Verse Structure, 1980) have also analyzed parallelism at the level of syntax.

Lastly, Kugel’s discussion of parallelism came in the context of the argument that parallelism is not the defining feature of Biblical poetry, that poetry cannot be so easily distinguished from prose, and therefore that poetry itself is not a useful category for describing Biblical literature. He argues that poetry has no perceptible meter distinct from prose, that many so-called poems have weak or no parallelism, and that prose writers make use of parallelism as well. However, Jakobson considered poetics to be a function of all language, thus it should not be surprising to find poetic structures in prose. Thus, it seems to be an overstatement to argue against a category of Biblical poetry. However, what has resulted from Kugel’s criticisms is greater recognition of a continuum from prose to poetry.

Kaufman, Stephen A., “The classification of the North West Semitic dialects of the Biblical period and some implications thereof,” Pages 41-57 in Proceedings of the 9th World Congress of Jewish Studies. Panel Sessions: Hebrew and Aramaic Languages. Jerusalem: World Union of Jewish Studies, 1988.

January 20, 2009

Language classification is complicated in areas of language contact since interference from surrounding dialects and languages (ie borrowing, analogical change, etc.) tends to obscure the genetic connections (assuming that there is validity to the genetic model). Thus the Stammbaum  approach (ie family tree) tends to be modified by the wave theory of change, the idea that linguistic features radiate out from some cultural center into the surrounding dialects. As Garr has argued, this especially seems to be the case for Syria-Palestine in the first half of the first millennium which represents a linguistic continuum of dialects in contact with each other. Therefore, scholars meet with great difficulty when attempting to shoehorn some of these peripheral dialects into neat categories such as “Aramaic” or “Canaanite”.

The most notorious among the NW Semitic dialects have been Ugaritic, Samalian, and Deir Alla. Ugaritic is something of a special case since it is an earlier dead-end branch from the Late Bronze Age, a peripheral member the proto-Canaanite-Aramaic dialect continuum (Kaufman’s term), and no universal consensus seems to have been reached on its classification. There does seem to be a consensus that Samalian is properly Aramaic, again, a peripheral dead-end variety. In this paper, Dr Kaufman will take up the issue of Deir Alla. Both Randall Garr and Jo Ann Hackett have addressed Deir Alla in their dissertations, Garr grouping it with Aramaic but Hackett South Canaanite.   

The most common approach to the classification of dialects among students of Semitic has been to assemble lists of isoglosses and simply count the presence or absence of features in the dialect, grouping it where it shares the most in common. Quite often this approach suffers from a lack of methodological rigor, most notably the assumption that all isoglosses are equal. If one follows the genetic model of language development, then it makes sense that only shared innovations are significant for genetic subgrouping. This is because shared innovations demonstrate that a subgroup has continued to develop independently after a split in the family tree. 

Further, Kaufman argues that features of greater frequency in normal speech ought to be given more weight. While the absolute chronologies of the glottochronological method may be rightly criticized, the assumption of lexicostatistics that the basic vocabulary of a language has more resistance to change  seems fundamentally sound. The same can be said of common morphological and grammatical features. This approach seems consistent with the criterion of mutual intelligibility. While it is probably impossible to measure the mutual intelligibility of ancient languages, it stands to reason that two such languages must coincide in their basic vocabularies and fundamental grammatical structures.

Thus Kaufman begins by summarizing the features of Deir Alla in comparison to “Canaanite” and “Aramaic”.  He finds seven features that are either common to NW Semitic or inconclusive including the imperfect consecutive, the use of the infinitive absolute with cognate finite verb, and the apparent lack of a definite article. There are seven features in common with Canaanite like the Nifal and the 3fp form tqtln. Lastly there are seven features in common with Aramaic such as the masculine plural ending -n and the distinction in the third weak verb between jussive -y and indicative -h. The scales appear to be balanced, but after removing the features which come from doubtful readings or really cannot be clearly classified as Canaanite or Aramaic, only one feature remains in the Canaanite column while four remain in the Aramaic column. Here Kaufman introduces a lexical analysis, finding that the vocabulary of Deir Alla contains 75-80 words from common NW Semitic, 5-8 Canaanite, and 21-24 Aramaic. Further, in contrast to Canaanite, the Aramaic list is full of basic vocabulary items like ‘son’, ‘wine’, give’, ‘enter’, etc. This seems to tip the scale drastically toward Aramaic.

Interestingly, this conclusion fits nicely with Dion’s analysis of Samalian. Every feature which Samalian shares with Old Aramaic against Canaanite is also shared by Deir Alla with only one exception, but none of the so-called Canaanite features of Samalian are found at Deir Alla. This suggests that Old Aramaic, Samalian, and Deir Alla shared a period of joint development after the split into the Canaanite and Aramaic branches. Both Deir Alla and Samalian seem to have then split from the main Aramaic branch at the same time. 

Now for the implications. Not only was Syria-Palestine a region of linguistic continuum, but also a literary continuum. The Deir Alla text has obvious parallels to the Bible, not only the character of Balaam but language, style, and phraseology. For instance, the “Last words of David” in 2 Sam 23:1-7 is an oracle introduced by the same phrase as the Deir Alla text, n’m PN n’m hgbr… 

More interesting though is the way Aramaic and Canaanite features are “mixed” in the Deir Alla text. Much of the Aramaic-like vocabulary was dismissed by those arguing for a Canaanite affiliation since it also occurs in the Hebrew Bible. But where does it occur? Usually in passages like Job or the wisdom of Lemuel (Prov 30) that are regarded as “Aramaizing” (and declared therefore to be post-exilic). However, Deir Alla suggests a different solution – the text may simply be written in a Trans-Jordanian pre-Exilic dialect which has a mix of Aramaic and Canaanite features. In fact, most such passages are not only connected to Trans-Jordanian characters, but are representing their direct speech. Thus what we may have are Hebrew authors in Hebrew texts attempting to represent Trans-Jordanian speech.

Happy MLK Day

January 19, 2009

Last year I found this paper, “Light on the Old Testament from the Ancient Near East“, that Martin Luther King wrote in Seminary under James Bennett Pritchard. I was a couple weeks late then, but I thought it would be fitting to note it again.

Online Amharic Lexicon

January 13, 2009

I have been reading the Ethiopian languages Ge’ez and Amharic in preparation for my Semitics exam, and I have found this site – – very helpful. What is great is that you can search either English or Amharic and the results are returned with Latin transliterations. Plus, in Amharic mode you can enter root consonants transliterated into Latin script (no need for an Amharic keyboard) and it will return anything close to a match. I wish someone had an Arabic dictionary like this, it takes me forever to try to enter things in Arabic.

Greenberg, J. H., “A Quantitative Approach to the Morphological Typology of Language,” International Journal of American Linguistics 26 (1960): 178-94.

January 10, 2009

Greenberg suggests that for a discipline to properly be considered “science”, it must move beyond mere description to comparison and classification. In linguistics this has been achieved through the respected discipline of “comparative linguistics” using the historical-genetic method to group languages into families with common ancestors. However, there is another method called “typological”, which (up to Greenberg’s day) was not so well received. Greenberg argues that the typological method has been damaged by confusion between the two, especially where typological criteria have been used to establish historical-genetic relationships.

The historical-genetic method classifies languages into families based on shared features, usually common forms related by sound and meaning. If two languages agree in a considerable number of these forms (not through borrowing) then they can be tied to a common ancestor. However, one can also compare languages that cannot be said to have a common ancestor. For instance, all languages must express comparison, but the number options for this are actually limited: a special comparative inflection of the adjective (English great-er), use of the preposition “from” (Semitic), use of a verb meaning “surpass” (African languages), etc. Some of these are more common than others, and the geographic distribution crosses genetic boundaries. Thus typological comparison also groups languages into “families”, but these are based only on the absence or presence of a given feature. However, the grouping is arbitrary and will change depending on which features are selected.

Some of these classification schemes are useless, but some can be very useful such as the classic 19th century division of languages into isolating, agglutinative, and inflecting. While the scheme itself has many weaknesses, it has hit upon something of fundamental importance to language – the morphological structure of a word. Sapir took up this classification as the foundation of his book Language and attempted to more carefully define the terms. In the end he saw a very basic form to language – every language has a stock of roots with concrete meanings (ie table, eat) as well as forms to express abstract relational ideas between the concrete terms (such as case endings). This opposition of concrete (I) and abstract (IV) marked the ends of the scale. Between the two are derivational elements (II) which impact the meaning of the root, but do not relate syntactically to the whole sentence (such as the –er in farmer). Lastly, concrete-relational elements (III) primarily affect sentence syntax, but have a concrete meaning. All languages use I and IV, but II and III are dispensable.

The technique of how II, III, and IV (all the elements which are somehow marked for syntax) are used is described as a) isolating (based on significant order of elements, John hit Bill marks John as subject, Bill as object), b) agglutanitive (like good + ness > goodness), c) fusional (like deep + th > depth), and d) symbolic (internal changes such as drink/drank/drunk). Note that the contrast between aglutanitive and fusional is itself somewhat superficial. The Semitic languages are labeled Complex Mixed-relational because all four concepts are present. In concept II they use techniques d) then c). In concept III they use c) then d). For IV they use a), though this is a weak development.

Greenberg has basically adapted Sapir’s model and developed quantitative measures. First, Greenberg introduces an index to measure the degree of synthesis based on the number of morphemes per word. The lower limit is obviously 1.00, and the upper limit while theoretically infinite seems to be practically 3.00. Analytic languages will give low ratios, synthetic higher, and polysynthetic the highest of all.

Secondly, Greenberg measures “agglutinavity” by the ratio of agglutinative constructions to morph juncture. A construction is considered agglutinative when both morphemes are automatic, which is to say that any allomorphic variation is predictable and regular. Thus leaves is made of two morphemes /leaf/ + /s/ both of which vary predictably to /leav/ and /z/ respectively. Thus the ratio is 1.00. A language with a high value for this index will be agglutinative, a low value fusional.

The third index analyzes the type of morphemes related to the number of words: root morphemes, derivational morphemes, and inflectional morphemes. This measures the basic contrast of Sapir’s categories I and IV. A language which can have more than one root per word is a compounding language, an idea that Sapir nowhere mentions. The fourth analyzes the ratio of prefixes and suffixes. The last index measures the use of word order, inflection, and concord (matching of gender, number, etc) to indicate syntactic relationships. Note that morphemes such as the Latin case ending –um, which indicates case, number, and gender, is counted separately for each function it performs.

Greenberg next spends some time discussing the problem of finding suitable definitions for concepts like morphs, morphemes, roots, etc. For instance, some things are obviously morphemes, some are obviously not, but some are in between. For example, is /deceive/ a single morpheme or is it made from /de/ + /ceive/? Greenberg suggests the model of a square, that is if there are four meaningful sequences in a language AC, BC, AD, BD then each may be considered a morpheme. An example is the English eating : sleeping : : eats : sleeps, where A is eat-, B is sleep-, C is -ing and D is –s.

Lastly, Greenberg calculates the indices for eight languages: Sanskrit, Anglo-Saxon, Persian, English, Yakut (related to Turkish, but with less borrowing from Arabic), Swahili, Annamite, and Eskimo. Based on an admittedly small sample, Greenberg’s results seem to corroborate the usual non-quantitative descriptions of these languages as synthetic or polysynthetic and agglutinative or non-agglutinative.

Go Gators!

January 9, 2009

In case I haven’t mentioned it before, I happen to have a BS in Mechanical Engineering from a certain university located in Gainesville, Florida. Go Gators!

Faber, Alice, “Genetic Subgrouping of the Semitic Languages,” Pages 3-15 in The Semitic Languages, Edited by R. Hetzron, London: Routledge, 1997.

January 9, 2009

One of the major pursuits of science has always been classification. Thanks to Linnaeus, the predominant taxonomy tends to be a fixed hierarchy (ie kingdom – class – order – genus – species). Thus a major pursuit of linguists has been the classification of languages and dialects into such groupings of parent-child (genetic) relationships. These groupings are usually based on the similarities between languages/dialects. However, it is difficult to determine whether a given similarity is a result of common ancestry or merely the influence of one language on another, or even just chance. Thus the preferred method is to group languages based on only shared innovations. Further, morphology is more useful for such analysis than phonology or lexicon.

In this article, Faber outlines the traditional approach to the Semitic languages as well as a new approach first introduced by Robert Hetzron based on shared innovations. The traditional grouping of the Semitic languages tended to be based more on cultural and geographical than linguistic features. The major groups are East Semitic (of which only Akkadian and Eblaite are attested) and West Semitic. West Semitic is further divided as follows:

West Semitic

Northwest Semitic



South Semitic


Southeast Semitic

Modern South Arabian


Old South Arabian


The major change introduced by Hetzron (and modified by Huehnergard) is to group Arabic and Northwest Semitic together under the label Central Semitic, which is distinguished from South Semitic (Southeast Semitic in the traditional model). Arabic shares features with both Northwest Semitic and South(east) Semitic, thus classification depends on which of these features are shared innovations and which may be the result of chance or structural similarities in the languages.

The features Arabic shares with Northwest Semitic include:

1. The realization of the emphatic consonants through pharyngealization (in Ethiopian and Modern South Arabian they are followed by a glottal stop).

2. The prefix conjugation yaqtulu which replaced the form yaqattal with doubling of the middle radical, attested in Akkadian (iparras), Ethiopic, and South Arabian.

3. Levelling of the prefix vowels. In Akkadian the prefixes of the active, non-derived stems are ˀa and ta, but ni and yi. However, in Central Semitic, all of the prefixes for a particular verb stem have the same vowel, either a or i. The vowel a was generalized in Arabic, while Hebrew preserves the a-i alternation (yilmad but yåqum).

4. Generalization of -t- in the kataba suffix conjugation. In the Akkadian stative, the 1cs form is parsā-ku while the 2ms form is parsā-ta. In Hebrew, Aramaic, and Arabic, the 1cs suffix form becomes katab-tu on analogy to the 2ms form katab-ta. In Ethiopic the reverse occurs as the 2ms becomes -ka.

5. Development of the compound negative marker *bal. Farber suggests that these forms may have developed from an inherited Afroasiatic negative particle *b combined with either negative or asseverative *la.

The features that Arabic shares with South(east) Semitic include:

1. The unconditioned sound change *p > /f/.

2. The existence of verbal stems with a long first vowel – kātaba, takātaba.

3. The broken plural, ie plurals formed by adding a prefix and/or internal changes rather than adding a suffix.

Those who group Arabic with South(east) Semitic (Blau, Diem) argue that the first five features represent convergence or diffusion, while the last three are shared innovations. Those who group Arabic with Northwest Semitic (Hetzron, Huehnergard, Goldenberg) argue that the first five are shared innovations while the last three represent common retentions from an older phase of language which have been replaced in Akkadian and Northwest Semitic.

Ginsberg, H.L., “The Northwest Semitic Languages,” Pages 102-124, 293 in The World History of the Jewish People, Vol 2: The Patriarchs. Edited by B. Mazar. Tel-Aviv; New Brunswick, 1970.

January 8, 2009

In this classic article, Ginsberg discusses the subdivision and development of the Northwest Semitic family of languages. NWS can be first divided into Canaanite and Aramaic groups. Canaanite is further comprised of a Phoenic group (Phoenician, Punic, Ugaritic, and Philistine/Ashdodite?) and a Hebraic group (Hebrew, Moabite, Ammonite and Edomite). Aramaic is comprised of Common Aramaic and Samalian.

Canaanite can be distinguished from Aramaic based on the following features:

1. Lexicon – הלך “to go” has secondary root ילך in imperfect, words like גג “roof”, שלהן “table” are confined to Canaanite.

2. Morphology – 3fp preformative t-, use of pōlel and hitpōlel for D and tD hollow verbs, formation of the participle based on the stem of the imperfect in hollow verbs, stative verbs, and the nifal, change of preformative vowel ya- to yi- when following vowel is a, attenuation of a to i in first syllable of derived conjugations.

Within Canaanite, the Phoenic can be distinguished from the Hebraic group by the following features:

1. Lexicon – for example יתן “to give”, כון “to be”, פעם “foot”, הרוץ “gold”, קרת “city” (note that the last three occur in Biblical Hebrew, but only as rare or poetic terms).

2. Syntax – no imperfect consecutive tense, instead use of infinitive absolute as a narrative tense.

In the 2nd Millenium, the only attestation of NWS comes from Ugaritic and the Canaanite reflexes/glosses in the Amarna letters. Some Biblical texts may date to the 2nd Millenium, most notably the Song of Deborah in Jd 5 (c. 1100 BCE). The oldest significant Phoenician inscription is the Ahiram sarcophagus from c. 1000 BCE.

Ugaritic maintains almost the entire inventory of Proto-Semitic consonants. It is also characterized by the use of the š causative in contrast to h-/ˀ-/y- and the use of the Gt. Ugaritic retains case endings and lacks a definite article. There is no evidence of a shift from ā > ō in contrast to the Amarna letters as well as Hebrew and Phoenician. The Amarna letters also frequently use a t- prefix instead of y- for the third masculine plural.

By 1000 BCE, Canaanite had gone through several changes:

1. First is the loss of case endings in Hebrew and Phoenician, though Phoenician retains the -i of the oblique case at least before pronominal suffixes. The short vowels at the end of verbal forms have also disappeared. Stress therefore came to rest on the final syllable causing the vowel to lengthen except when the final syllable is doubly closed (it ends with two consonants).

2. The third feminine singular ending -at > a in the Hebrew and Phoenician verb, as well as the Hebrew noun. In the Phoenician noun this change may have been blocked by the retention of the oblique -i ending.

3. Along with the shift of ā > ō, the diphthongs reduce (aw > ō and ay > ē) in Phoenician. This occurred variously in the other Canaanite languages as well.

4. In Phoenician the consonants were reduced to 22. Hebrew, while using 22 graphemes, still distinguished š and ś, and likely from and ġ from ˁ in the early pronunciation.

5. Lastly, by 1000 BCE all the Canaanite languages have the definite article haCC-.

The Phoenic group is comprised of Phoenician, Punic, Ugaritic, and perhaps Philistine/Ashdodite. Phoenician has a broad diffusion chronologically (lasting almost 1500 years) and geographically. It is divided into two dialects: Byblian and Common Phoenician. Byblian is characterized mainly by the archaic pronominal suffixes: 3ms -h and -w, 3fs -h, and 3mp -hm. Common Phoenician, on the other hand, uses two sets. The first set is 3ms -o, 3fs -a, pl -m, and the second is 3ms -yu, fs -ya, pl -nm. The first set is used following consonants while the second set is more common (perhaps due to retention of the oblique -i ending). Other characteristics of Common Phoenician include  yi- as the preformative of the causative rather than hi-, bal as the negation rather than , and the particle of non-existence ˀy in contrast to ˀēn in Hebrew, Ugaritic, and Moabite.

Punic is the language of the Phoenician colonies in the Western Mediterranean, most notably Carthage in North Africa. It is attested from the early 8th century BCE through the 2nd century CE, but continued to be spoken until the 6th century CE. Early Punic only differs from Common Phoenician in writing the ms and fs suffixes (-o and -a) with א.

The Hebraic group is comprised of Hebrew, Moabite, Ammonite, and Edomite (though we have no extant texts of the latter two). The Hebraic group is distinguished most notably by the verb עשה “to make, do” and the imperfect consecutive construction. Moabite retains the Gt, the mp ending -n, and the feminine ending -t.

Hebrew can be divided into the Biblical Period (1225 BCE- 63 BCE) and the Roman Period (63 BCE – c. 200 CE), though admittedly this division is socio-political in nature and not linguistic. The Biblical period is further divided into the Golden Age (1225 BCE – 500 BCE) and the Silver Age (500 BCE – 63 BCE). The earliest text is probably the Song of Deborah (c.1100) while Deutero-Isaiah (c 550-500) is the last author of the Golden Age. Esther is probably the last work of the Silver Age though the language is archaizing and thus typologically older than Canticles for example. Qumran Hebrew deliberately imitates Biblical Hebrew, but is recognizably later.

Hebrew is more archaic than the Phoenic group in the use of 3mp suffix -hēm instead of -nēm, the use of the pual, hophal, and qal passive (which is not supplanted completely by the niphal until the early Silver Age). However, Hebrew is less archaic then the Phoenic group in the complete loss of the case system and the loss of the Gt.

The Hebrew modal system has many interesting characteristics. The jussive is distinguished from the imperfect by the so-called energic nun: yišmrehû “let him watch it”, but yišmrennû “he will watch it”. Prohibition is expressed by אל while ordinary negation uses לא (not bal as in Phoenician). A permanent or solemn command can be expressed by the infinitive absolute, and a permanent prohibition by לא + imperfect.

Several changes occurred in the Silver Age. Most notably the more frequent use of ש for אשר, the introduction of many Persian loanwords by way of Aramaic, and imitation of Aramaic. For instance אשר and ש usurp the function of the conjunction כי, similar to the use of ד in Aramaic.

Most of the changes in the Roman Age are due to the increasing influence of Aramaic. Several parts of the verbal system fell out of use including the infinitive absolute, the cohortative, the so-called consecutive tenses, the infinitive construct with a subjective recta, and the Qal passive. On the other hand, new tense forms developed from the combination of the participle with היה. Other changes include the use of את + a pronominal suffix as a demonstrative and many borrowings from Greek.

The Aramaic sub-family includes Common Aramaic and Sam’alian, a local vernacular of the kingdom of Sam’al that was superceded by Common Aramaic in the 8th Century, at least for royal inscriptions. Distinctive features of Aramaic include:

1. Dropping of ˀa from the number one – ḥad.

2. Shift of to a phoneme resembling q, which was again shifted later to ˁ.

3. The relative particle .

4. The vowel pattern a-i in all derived active conjugations.

5. The assimilation of initial y in roots ydˁ and yṯb in certain conjugations.

6. The generalization of the nominative construct for all other cases (except before 1cs suffix) for the words “father” ˀabū-, “brother” ˀaḥū-, and “father-in-law” ḥamū.

The earliest Aramaic inscriptions maintain almost the entire Proto-Semitic consonantal inventory as well as the old passive conjugations. The jussive is distinguished from the imperfect and negated by אל. Instead of the nifal or Gt, there is a tG. Common Aramaic lost the case endings as well as the feminine ending -t. In the Hellenistic period, the initial hit- reduced to ˀit in the reflexive conjugations, while the analogous reduction of ha- to ˀa- in the causative was completed by the Roman period. Also in the Roman period, the characteristic vowel reduction took place. The bgdkpt letters softened sometime before the 5th century CE, but it is not known exactly when.

In the Roman period there is also a noticeable division between Eastern and Western Aramaic dialects. Some characteristics of the Eastern dialects include:

1. The use of -ē for the mp emphatic ending (cf. -ayyā).

2. The emphatic state no longer indicates determination.

3. The disappearance of the pronominal suffixes with -in-.

4. The substitution of the jussive form with prefixed l- (later n-) for the third person imperfect.


Get every new post delivered to your Inbox.

Join 154 other followers