Printer Friendly

South Ethiopian pronouns and verbs in an Arab grammatical text revisited after seventy years.

To the memory of Baruch Podolsky


Seventy years ago, Sidney Glazer edited in the pages of this journal what he called "a noteworthy passage from an Arab grammatical text." The fragment, going back to the famous fourteenth-century Arab grammarian Ab[] Hayy[]n, is indeed remarkable in several respects, not least because of the author's attempt to provide a kind of cross-linguistic survey of the origin of verbal inflection in five Near Eastern languages: Arabic, Persian, Turkish, Coptic, and Ethiopian Semitic.

Glazer's analysis of the text is focused exclusively on its implications for Arab grammatical theory, with no attention to the material evidence of any of the foreign languages treated by Ab[] Hayy[]n (hereafter AH). Our short note is intended to fill this gap in what concerns the Ethiopic material, which includes six elements: (1)







To the best of our knowledge, no systematic analysis of the Ethiopic forms in AH's text has been undertaken since the publication of Glazer's article. In his extensive review of Leslau's Etymological Dictionary of Gurage, Gideon Goldenberg (1987: 93) gave some attention to Au' s pronominal forms, attributing their language to "an earlier form of what is now East Gurage," (2) while in Baruch Podolsky's pioneering study of the historical phonology of Amharic (1991), AH is twice referred to in connection with the pertinent phonological phenomena. It is to Podolsky's book that the present authors owe their acquaintance with the passage under scrutiny.


The pronominal forms adduced by AU immediately betray their South Ethiopian origin. (3)

The first person plural form [TEXT NOT REPRODUCIBLE IN ASCII] (= * enya or * en(n)a) is compatible with cognate forms in the majority of modern South Ethiopian languages: Amharic [delta]nna, Harari ina, Soddo enna, Selti ina (Kane 1990: 1254; Leslau 1963: 29; 1979: 79).

In a similar way, the first person singular form [TEXT NOT REPRODUCIBLE IN ASCII] (= * eyya) has reliable parallels in the majority of gunnan-Gurage and East Gurage: Zway?[]ya, Masqan eyya, Ennemor eya, Chaha eya (Leslau 1979: 116). (4) Close to these forms is Argobba ay (Leslau 1997: 194). There is no explanation for the word-final hamza in the Arabic rendering, nor for the kasra that accompanies it.

The second person singular form [TEXT NOT REPRODUCIBLE IN ASCII] (= * ata) drastically reduces the choice, (5) as East Gurage is the only Southern Ethiopian sub-branch where such forms are attested: Selti, Wolane, and Zway ata (Leslau 1979: 102; Gutt 1997: 911; Meyer 2006: 165; 2005: 77).


1. The root

The sample root chosen by AH to illustrate the Ethiopian verbal paradigm is m-h-t ('to strike, to beat'), functionally corresponding to Arabic d-r-b. This root, going back to PS * mhs, is well attested in Geez and in the majority of Modern Ethiopian (Leslau 1963: 105; 1979: 437; 1987: 337). In today's East Gurage it is preserved only in Zway, but this is hardly a persuasive reason to exclude other languages of this sub-branch (or indeed other South Ethiopian languages) from consideration: the dialectal distribution of this root in AH's times may well have been much broader than today.

The presence of h in the Arabic rendering is more than expected for such an early date: as is well known, some of the South Ethiopian languages (Argobba, Harari) preserve h (< * h, * h, * h) up to this day, and this was certainly the case of (at least the earliest phases of) Old Amharic (Richter 1997: 549). Somewhat more intriguing is the shift * s > t, for which our document seems to provide the earliest piece of evidence known so far. It stands to reason that in the source language of Au the shift * s > t preceded the loss of gutturals and was accomplished at a fairly early date. (6)

2. The perfect

There is hardly anything special in the perfect mahat(a), except for the fact that there is no fatha under the third consonant in Glazer's Arabic text. Unless a misprint, (7) it might provide an interesting piece of evidence for the East Gurage affiliation of the text, since Zway is the only Ethiopian Semitic language where the final *-a of the perfect is regularly dropped: dalas 'he waited' (Meyer 2005: 112). Still, a more explicit notation with final suk[]n would probably be expected in such a case.

3. The prefix conjugation

Conversely, the prefix conjugation forms offer several riddles to be solved. First and foremost, the very nature of the base has to be established--is it the long form (the imperfect) or the short form (the jussive)?

3.1. The jussive. At first sight, the jussive hypothesis has much in its favor. The stem *-[C.sub.1][C.sub.2]a[C.sub.3]-is, of course, the normal base of the jussive throughout Ethiopian Semitic. In addition, this interpretation can provide a promising solution for another riddle of AH's verbal forms, viz., the difference in the vocalic shape of the prefixes (fatha for y-, but kasra (8) elsewhere); exactly this distribution is attested in the jussive paradigm of gunnan-Gurage (Cana nagraz 'let me become old' vs. yagraz 'let him become old'). (9) But in spite of its positive aspects, the jussive hypothesis cannot be accepted in view of two weighty arguments against it.

Firstly, AH's Arabic equivalents ([TEXT NOT REPRODUCIBLE IN ASCII], etc.) are not expected to correspond to jussive forms in Ethiopian: had the jussive meaning been intended, [TEXT NOT REPRODUCIBLE IN ASCII] (or similar) would have been used. This argument might not be compelling: the Ethiopian long imperfect was undoubtedly very unusual to an Arab ear and could have easily been substituted by the corresponding jussive forms (in their turn, structurally identical to the Arabic imperfect) with no attention to their morphosyntactic properties.

Far more serious is the second argument: the first person singular prefix of the jussive is * le - throughout Southern Ethiopian, whereas AU' s form clearly reflects? e-. Moreover, there are good reasons to trace the * le-prefix back to the proto-language of this branch (Bulakh and Kogan 2010: 279), which makes it virtually inconceivable for a Southern Ethiopian language of the fourteenth century to bypass this fundamental innovation.

The jussive interpretation must therefore be rejected. How does the alternative solution of the imperfect fare?

3.2. The imperfect. The imperfect hypothesis does not suffer from any of the aforementioned deficiencies: from the functional point of view it corresponds exactly to [TEXT NOT REPRODUCIBLE IN ASCII] in the Arabic original, and the first person singular prefix of the imperfect is *? e-in proto(-South-) Ethiopian, preserved as e- in most languages of this branch.

The main difficulty connected with the imperfect hypothesis is the structure of the base. Indeed, mhat- is not compatible with any of the structures attested for the imperfect of verbs with the (historical) second guttural in Southern Ethiopian--neither the "e-type" represented by Argobba yagaher-'he works' and Harari yalahsi 'he licks' (Leslau 1997: 66, 70; 1958: 51), nor the "e-type" attested in Amharic yesem-'he kisses', Argobba yeg[]r 'he works', Wolane yelix 'he sends' (Leslau 1995: 533; 1997: 70; Meyer 2006: 58, 68). Theoretically, -mhat-could be analyzed as a special variety of the former type (* m[]ht- > * mahat- > -mhat-), but the degree of probability of such a development is quite low.

The imperfect interpretation also fails to provide an explanation for the vocalic difference in the prefixes: throughout Ethiopian Semitic, the only vowel attested in this morphological position is e. (10)

3.3. More conflicting features. In the two preceding sections we have discussed a few morphological features that preclude direct identification of Alys verbal paradigm with any attested South Ethiopian language. There are, however, more such phenomena whose attestation is less ubiquitous, yet broad enough to exclude from consideration many of the key languages of the Southern branch of Ethiopian Semitic--including those that seem the most likely candidates on the evidence of the pronominal paradigm.

The first feature to be mentioned here is the suffix * -n(a), regularly attached to the first person plural forms of both the imperfect and the jussive in Argobba, Gogot, Muher, Masqan, and West Gurage: Chaha ne-zarfe-na 'we plunder', ne-graz-na 'let us grow old' (Leslau 1983: 16-17).

The second feature, characteristic of the East Gurage languages, is undoubtedly related to the first one (cf. Goldenberg 1977: 481-82) and consists in the replacement (again both in the imperfect and the jussive) of the first person plural prefix * ne-by the prefix of the first person singular: Sel. i-nakti-na 'we beat' (cf. i-nakt 'I beat', Gutt 1997: 921).

In several other languages, the prefix of the first person plural (either in the imperfect only or in both the imperfect and the jussive) is characterized by the secondary addition of the element? e-: Tigre? en- (imperfect), Amharic en(ne)- (imperfect/jussive), Gafat enne-(imperfect) (Raz 1983: 55-56; Leslau 1995: 301, 347-48; 1956: 101).

Finally, the old prefix of the first person singular *? e- was replaced by y[delta]- in the imperfect in Wolane (yesabr- 'I break') and Zway (yenakl- 'I take') (Meyer 2006: 110; 2005: 96).

None of these four features is in any way reflected in AH's list.


The pronominal evidence analyzed in the first section strongly suggests that the source-language of AH should be identified with East Gurage, more specifically with Zway or Selti. (11) In view of its internal consistency, it seems justified to take Au's pronominal system at face value and to treat it as superior to the contradictory verbal evidence analyzed under "The Verb" above.

Indeed, the main deficiency of AH's verbal paradigm is not that it does not fit the proposed East Gurage affiliation, but that it is not compatible with any Ethiopian Semitic language. One may reasonably wonder whether some of the conflicting features could be discarded as late innovations chronologically posterior to AU. Such a reconstructive approach may help to eliminate some of the obstacles (notably those dealt with under 3.3 above, as they are not common South Ethiopian), but most of the truly fundamental difficulties (such as the lack of le- within the jussive hypothesis and the unusual structure of the base within the imperfect one) would invariably persist.

Under such conditions, it seems legitimate to question the very reliability of AH's verbal paradigm (or at least the form in which it has reached us). After all, AH's primary interest was the morphological shape of the prefixes and the way they are attached to the base, not the base itself (nor its endings, nor the concomitant phonological processes). This would provide a more meaningful explanation of the sometimes surprisingly Arabic-like guise of AH's verbal forms. (12) We do not feel competent to judge whether this Arabizing trend was initiated by the great grammarian himself or introduced by later copyists.

Authors' note: Critical remarks and suggestions by Wolfhart Heinrichs, John Huehnergard, and an anonymous reviewer are gratefully acknowledged.


Bulakh, M., and L. Kogan. 2010. The Genealogical Position of Tigre and the Problem of North Ethio-Semitic Unity. Zeitschrift der Deutschen Morgenlandischen Gesellschaft 160: 274-302.

Glazer, S. 1942. A Noteworthy Passage from an Arab Grammatical Text. JAOS 62: 106-8.

__. 1947. Manhaj as-s[]lik: Ab[] Hayy[]n' s Commentary on the Alfiyya of Ibn M[]lik. New Haven: AOS.

Goldenberg, G. 1977. The Semitic Languages of Ethiopia and Their Classification. Bulletin of the School of Oriental and African Studies 40: 461-507.

__. 1987. Linguistic Interest in Gurage and the Gurage Etymological Dictionary. Annali Istituto Orientale di Napoli 47: 75-98.

Gutt, E.-A. 1997. A Concise Grammar of Silt'e. In Trilingual Dictionary Silt'e-Amharic-English, ed. E. H. M. Gutt and H. Mohammed. Pp. 895-957. Addis Ababa: Addis Ababa Univ. Press.

Hetzron, R. 1968. Main Verb-Markers in Northern Gurage. Africa 38: 156-72.

__. 1972. Ethiopian Semitic: Studies in Classification. Manchester: Manchester Univ. Press.

Kane, T. L. 1990. Amharic-English Dictionary. Wiesbaden: Harrassowitz.

Leslau, W. 1956. Etude descriptive et comparative du Gafat (Ethiopien meridional). Paris: Librairie C. Klincksieck.

__. 1958. The Verb in Harari. Berkeley and Los Angeles: Univ. of California Press.

__. 1963. Etymological Dictionary of Harari. Berkeley and Los Angeles: Univ. of California Press.

__. 1979. Etymological Dictionary of Gurage, vol. 3. Wiesbaden: Harrassowitz.

__. 1983. Ethiopians Speak: Studies in Cultural Background, pt. V: Chaha-Ennemor. Wiesbaden: Franz Steiner.

__. 1987. Comparative Dictionary of Ge fez (Classical Ethiopic). Wiesbaden: Harrassowitz.

__. 1995. Reference Grammar of Amharic. Wiesbaden: Harrassowitz.

__. 1997. Ethiopic Documents: Argobba. Grammar and Dictionary. Wiesbaden: Harrassowitz.

Littmann E. 1944. Altamharisches Glossar. Rivista degli Studi Orientali 20: 473-505.

Meyer, R. 2005. Da.s. My: Deskriptive Grammatik einer Ostguragesprache. Cologne: Rudiger Koppe.

__. 2006. Wolane: Descriptive Grammar of an East Gurage Language (Ethiosemitic). Cologne: Rudiger Koppe.

Nosnitsin, D. 2003. Ase. In Encyclopaedia Aethiopica, ed. S. Uhlis, vol. 1. Pp. 364-65. Wiesbaden: Harrassowitz.

Podolsky, B. 1991. Historical Phonetics of Amharic. Tel Aviv: n.p.

Raz, Sh. 1983. Tigre Grammar and Texts. Malibu: Undena Publications.

Richter, R. 1997. Some Linguistic Peculiarities of Old Amharic Texts. In Ethiopia in Broader Perspective: Papers of the XIIth International Conference of Ethiopian Studies, Kyoto, 12-17 December 1997, ed. K. Fukui et al., vol 1. Pp. 543-51. Kyoto: Shokado.

Strelcyn, S. 1968. Le passage s > t en amharique comme objet d'etudes synchroniques et diachroniques. Rocznik Orientalistczny 31: 127-34.

Wagner, E. 1968. Drei Miszellen zum sudostsemitischen Verbum. Wissenschaftliche Zeitschrift Universitat Halle 17: 207-15.

(1.) Here and elsewhere below the relevant forms are adduced as in Glazer 1947: 230, which provides a better text than Glazer 1942.

(2.) Reference courtesy of Wolfhart Heinrichs.

(3.) As duly recognized already by Podolsky (1991: 23, 38).

(4.) For Selti, Leslau gives both iy[] (1979: 116) and iha. (ibid., 31). Only the in Gutt 1997: 911.

(5.) Contrast Podolsky 1991; 23: "It might have been Old Amharic or some other southern Ethio-Semitic language."

(6.) This circumstance practically excludes Amharic as a source-language of AH, as the loss of h preceded the shift * s> t in Amharic: cf. such Old Amharic forms as sabba (* sbh) 'it appeared, it shone' (Littmann 1944: 499; for the shift * s > t cf. also Strelcyn 1968; Podolsky 1991: 21-22). An alternative possibility is that the Arabic rendering does not reflect the shift * s > t at all; one may hypothesize that the Ethiopian glottalized affricate [TEXT NOT REPRODUCIBLE IN ASCII] [ts'] was so different from the Arabic pharyngalized sibilant [TEXT NOT REPRODUCIBLE IN ASCII] [s] that the emphatic dental stop could he perceived by an Arab as a more suitable rendering (cf. the Amharic title aye rendered as hatt[] in al-'Umar[]'s Mas[]lik al-abs[]r f[] mam[]lik al-ams[]r [Nosnitsin 2003]).

(7.) Note that a appears without parentheses in Glazer's transliteration embedded into the English translation (1942: 108), so also Goldenberg 1987: 93. It is nevertheless significant that the final fatha is missing also from the 1947 text, which seems to be considerably more accurate as far as the vocalic signs in the Ethiopian forms are concerned.

(8.) Certainly equivalent to Ethiopic [delta].

(9.) For the diachronic background of this remarkable phenomenon, see Wagner 1968.

(10.) One may wonder, however, whether the distinction is indeed genuine to AU, who mentions only i in his commentary to the Ethiopian verbal paradigm ([TEXT NOT REPRODUCIBLE IN ASCII], boldface added). On this question, see further below in the concluding section.

(11.) Not Wolane, where the first person singular pronoun is?ihe/yihe (Meyer 2006: 165).

(12.) Apart from the "short" structure of the base and the unexpected a-vocalism of some prefixes, this reasoning may apply to one important feature of AH's verbal forms not discussed before, viz., the final -u in the prefix-conjugation forms. In theory, this damma might reflect a genuinely Ethiopian morphological element: the declarative marker -u in Zway (always attached to the final lexeme with a consonantal Auslaut in a declarative sentence [Meyer 2005: 305-6]) or the Main Verb Marker -u in Soddo, Gogot, and Muher (Hetzron 1968). However, a secondary amalgamation with the Arabic imperfect marker -u is at least no less likely.

COPYRIGHT 2011 American Oriental Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2011 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bulakh, Maria; Kogan, Leonid
Publication:The Journal of the American Oriental Society
Geographic Code:6ETHI
Date:Oct 1, 2011
Previous Article:Istikshaf in Islamic jurisprudence and modern law.
Next Article:Why al-Jahiz needs Slonimsky's Earbox.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters