on the role of south arabian and ethio-semitic within a comparative semitic lexicographical project

Genetic classification in general depends to a large extent on the criteria selected. Inspired by Kogan 2015 and other sources, this paper looks at specific lexical peculiarities and semantic traits in the South Semitic (mainly modern South Arabian and Ethio-Semitic) lexicon, in order to determine the value of the South Semitic lexicon for genetic classification within Semitic at large. [1] introduction This paper focuses on the role and importance of the lexicon for genetic classification, more specifically on the role of (mainly modern) South Arabian and Ethio-Semitic in this context, a role which recently has been highlighted by Kogan (2015). It is a truism that linguistic classification to a large degree depends on the criteria selected, i.e. phonological, morphological, and other data. Therefore, the main criteria, or rather shared innovations that are usually cited in this connection, are summarized in the following (cf. Faber 1997: 7–12): – East Semitic is characterized by the development of an adjectival ending -ūt (pl. m.) and by the dative suffixes -kum and -šum; – West Semitic is characterized by the suffix conjugation denoting past tense (as opposed to the Akkadian stative) and a prohibitive negator ʾal; – Central Semitic is characterized by a series of pharyngealized consonants, a prefix conjugation for nonpast without gemination of the second root consonant and the leveling of prefix vowels in this conjugation, the generalization of a -tsuffix (as opposed to -k-) in the suffix conjugation, and the development of a compound negative marker *bal; the Northwest Semitic part of Central Semitic (as being distinct from Arabic) is characterized by the change of word-initial w to y (except for the con-

-East Semitic is characterized by the development of an adjectival ending -ūt (pl.m.) and by the dative suffixes -kum and -šum; -West Semitic is characterized by the suffix conjugation denoting past tense (as opposed to the Akkadian stative) and a prohibitive negator ʾal; -Central Semitic is characterized by a series of pharyngealized consonants, a prefix conjugation for nonpast without gemination of the second root consonant and the leveling of prefix vowels in this conjugation, the generalization of a -t-suffix (as opposed to -k-) in the suffix conjugation, and the development of a compound negative marker *bal; the Northwest Semitic part of Central Semitic (as being distinct from Arabic) is characterized by the change of word-initial w to y (except for the con- junction w-) and a doubly marked plural (internal and external); further internal isoglosses apply; -South Semitic is characterized by the generalization of a -k-suffix (as opposed to -t-above) in the suffix conjugation and by the generalization of (ʾ)al as a verbal negator; Eastern South Semitic (Modern South Arabian), as opposed to Western South Semitic (Old South Arabian and Ethio-Semitic) features a pre-fixed definite article C(a), with C being one of the gutturals ʾ, ḥ, or h.Goldenberg (2013: 45f.) lists the following further classification criteria: -the distribution of the intraflexion ("broken plural") in Arabic, South-Arabian, and Ethio-Semitic; -internal vowel lengthening in the binyanim (forms III and VI in Arabic); -consistent use of -a-in all active forms of the suffix conjugation; -the existence of two prefix conjugations (Akkadian, Modern South Arabian, Ethio-Semitic); -the emergence of the -na ending (3pl.f, 2pl.f) in Central Semitic (cf.Hetzron 1976); -the isogloss p (Akkadian, Aramaic, Canaanite) vs. f (Arabic, South-Arabian, Ethio-Semitic. Representing the "traditional" criteria, Faber (1997) Hetzron (1976), to whom we owe the concept of "archaic heterogeneity" and "shared morpho-lexical innovations" in Semitic, had arrived at the following genealogical representation (apud Huehnergard and Rubin 2011: 262): OSLa volume 8( 1), 2016 [2] issues of lexical classif ication One aspect clearly affecting the lexical distinctiveness of Ethio-Semitic is the Cushitic substratum in Ethio-Semitic (cf.Leslau 1945Leslau , 1952;;Appleyard 1977).Appleyard (1977) identifies Cushitic loans especially in the following semantic fields: "man", comprising general terms, kinship terms, and parts of the body; "the domestic environment", comprising agricultural activities and implements, crops, domestic animals, food and its preparation, and the [realm of the] house; "the natural environment", comprising natural phenomena, flora, and fauna; "social organization", comprising law and government, economy, warfare, and religion; and "grammatical items", comprising pronouns, numerals, and particles.In addition, the core Semitic stock of the Modern South Arabian and Ethio-Semitic vocabulary also displays specific semantic traits in its lexicon, traits that may be due to linguistic contact or may even reflect early retentions.
Before turning to look more closely at a selected sample of Ethio-Semitic data, here is a brief overview of some issues that have the potential to complicate lexical classification, keeping the Ethio-Semitic scenario in mind.

[2.1] Issues of choice of lemmata: lexical lists based on genetic cognates vs. lexical lists based on target language
Grosso modo, lexical lists can be based either on genetic cognates within a language family, irrespective of the precise meaning of the lexical items in the individual languages (e.g., Bergsträsser 1983: 210-223) or on lists of words with (approximately) the same meaning in a target language, which, however, need not be genetically related (e.g., Bennett 1998: 232-249).Kogan (2015) takes both strategies into consideration, but pays special attention to the second strategy, building his arguments on a kind of modified Swadesh list.The lemma "sun" in Semitic may serve to illustrate the situation.While East and Central Semitic (and also Epigraphic South Arabian, "ESA") use the root √ š-m-s in variations, the modern South Arabian languages resort to the lemmata for "day" (√ y-w-m) and "pre-noon" (√ ḍ-ḥ-y) respectively, roots which are also attested in other branches of Semitic but constitute lexical innovations in the cited South Semitic branches.Here is an overview of the lemma "sun" in Semitic (cf.also Leslau 1987: 149)

3] Issues of inner-family loan
An especially intricate issue is the distinction between genetically related cognates on the one hand and and inner-family loans on the other.The latter term refers to loan between languages of the same genetic subgroup, which may engender semantic specification of an indigenous term due to its interaction with a genetic cognate.In the case of the latter, usual lautgesetzlich correspondences need no longer obtain.Also the semantics need not be exactly maintained in the borrowing process.An example, featuring loans from Gəʿəz to Arabic, is the following (cf.

4] Issues of autochthonous vocabulary vs. loaned vocabulary in target languages
Of special interest is the case where lexical doublets emerge as a result of inner-family borrowing.In this case, the respective lemma is both attested autochthonously and in a loaned version (cf.Edzard 2015a for a contextualization of this issue).Typically, the borrowed lemma then has a specialized meaning.While Arabic, for instance, features the autochthonous word θāba 'he returned', one also finds the Aramaic loanword tāba 'he repented' (i.e."returned" in a moral sense).Brockelmann had already examined this issue in his Grundriss (Brockelmann 1908(Brockelmann -1913, vol. 1: 119), vol. 1: 119) The etymological origin a of a given word cannot be determined with certainty COMPARATIVE SEMITIC LEXICOGRAPHICAL PROJECT [227] OSLa volume 8( 1), 2016 in every case.Ethio-Semitic examples include the following (cf.Kogan 2015: 446-448): Cognate roots do not always appear in exactly the same order.The concept of "metathesis" is not meaningful in every case.Rather, at least in some cases, the semantics of a given term appear to be linked to the (non-ordered) set of the root consonants.[3] three sample e nt ries In the following, four sample entries will considered: √ n-f-s, √ d-b-r, √ r-k-b, and √ h-g-r, which shed light on the previously mentioned importance of South Arabian and Ethio-Semitic for classification.
This root is attested across Semitic and is relatively straightforward in the distribution of its semantic range.As this root served as a model example within the Doha project, context is also provided in this case.For the semantic connection between "life" and "tomb", see also Modern South Arabian (Mehri, Jibbāli, Ḥarsūsi, Soqoṭri) Mehri √ n-f(-s) -noun ḥə-nōf 'self' -verb: anōfes 'to make space'; əntəfūs 'to be safely delivered of a child'; šənfūs 'to welcome s.o.' -noun: nafs/nəfáws 'individual, soul, person' Regarding the importance of South Semitic, Kogan (2015: 578) This root has a wider semantic range.Indeed, in the case of this root we are probably looking at a variety of different homophonic lexical entries.Ethio-Semitic presents the additional semantic aspect of "monastery", i.e. something lying on back of a mountain.In Mehri and Amharic, one can observe interesting cases of metathesis.
East Semitic Akkadian √ d-b-r -verb: dupurrum 'to depart, recede ' -noun: madbaru, mud(a) [4] conclusion Even this short glimpse at the South Semitic (South Arabian and Ethio-Semitic) lexicon is revealing and rewarding.Clearly, South Arabian often features different or at least additional semantic traits in the lexicon.A systematic evaluation of sources such as Cohen 1970 -as well as the Semitic dictionaries with comparative evidence (notably Leslau 1987), will certainly bring to light more interesting material.The usefulness of lexicostatistics for genetic classification continues to be a point of discussion.
acknowledgme nts I would like to thank the NORHED project -Linguistic Capacity Building -tools for inclusive development of Ethiopia and the editors of this book.I also appreciate the comments and suggestions made by my two anonymous reviewers.

figure 2 :figure 3 :
figure 2: Stemma According to Huehnergard and Rubin 2011.Porkhomovsky (1997) and others have since then refined the previous model as follows.It is noteworthy that both Modern South Arabian ("MSA") and Ethio-Semitic ("Ethiopian") branch off at a quite high level in this model (apud Huehnergard and Rubin 2011: 263): : Aramaicšemšā ‫ܫ‬ ܶ ‫ܫܡ‬ ܳ ‫ܐ‬ edzard OSLa volume 8(1), 2016Diachronic loans within the same genetic subgroup (e.g., from Gəʿəz to Amharic) can also create confusion in the realm of phonology.A well-known (probably universal) phenomenon is the fact that proper nouns (place names and personal names) tend to be phonologically (and orthographically) more conservative than correlating common nouns and verbs.Within Ethio-Semitic, one can observe that the phonological structure of nouns in Gəʿəz containing gutturals is still preserved in loaned terms in Amharic (at least orthographically), whereas the gutturals are lost in the respective verbal forms belonging to the same root.Here is an example comprising four terms (cf.also Edzard 2015b: 189): ተማረ tämari 'student መጽሐፍ mäṣḥaf [mäṣhaf] 'book' ጻፈ ṣafä 'he wrote' ደኅና däḫna [dä(h)na] 'well' ዳነ danä 'he recovered' ሥዕል səʿəl 'picture' ሣለ salä 'he painted' [2.