Annotated Swadesh wordlists for the Surmic group (East Sudanic family).

Languages included: Narim [srm-nrm], Didinga [srm-did], Tennet [srm-ten], Murle [srm-mrl], Baale [srm-bal], Suri (Chai) [srm-cai], Mursi [srm-mrs], Me'en [srm-mee], Kwegu [srm-kwg], Majang [srm-maj].



Yigezu 2001 = Yigezu, Moges. A comparative study of the phonetics and phonology of Surmic languages. Ph.D. thesis, Université Libre de Bruxelles, 2001-2002. // Detailed study containing both descriptive data on most known Surmic languages and a historical-comparative study of the two main branches of Surmic (Southeast and Southwest Surmic). Includes a comparative etymological corpus and a 312-item comparative wordlist on ten languages, mostly collected by the author himself (Narim, Tennet, Didinga, Murle, Baale, Chai, Mursi, Me'en, Koegu, Majang). For many of these languages, the work still remains the best source of original data, despite occasional inaccuracies and misprints.

Bender 1971 = Bender, Lionel M. The Languages of Ethiopia: A New Le\-xicostatistic Classification and Some Problems of Diffusion. In: Anthro\-po\-lo\-gical Linguistics, 13, 5, pp. 165-288. // A lexicostatistical study of most of the languages of Ethiopia. Includes slightly modified Swadesh wordlists for a large number of Cushitic, Omotic, Ethiosemitic, and Nilo-Saharan (Nilotic, Surmic, Koman, etc.) languages. Unfortunately, the survey suffers from numerous inaccuracies of phonetic transcription and semantic glossing, making it practically unusable as a primary source for any of the languages concerned.


Tucker 1951 = Tucker, Archibald. Notes on Murle ('Beir'). In: Afrika und Übersee, 36, pp. 99-114. // Brief grammatical description of Murle with a short appended vocabulary. The latter also includes some comparative data on the closely related Longarim (Narim) and Didinga languages.

Stirtz 2011 = Stirtz, Timothy M. Laarim (loh) Tone. SIL International. // A paper on the prosodic properties of the Narim (Longarim, Laarim) language. Contains original field data collected by the author.


Driberg 1931 = Driberg, Jack H. The Didinga language. In: Mitteilungen des Seminars für orientalische Sprachen, 34-III, pp. 139-182. // A brief, somewhat antiquated and frequently inaccurate, description of Didinga grammar, accompanied with text examples and a large vocabulary.

Odden 1983 = Odden, David. Aspects of Didinga Phonology and Mor\-pho\-logy. In: Nilo-Saharan Language Studies. Ed. by Lionel M. Bender. Michigan: East Lansing, pp. 148-176. // A phonetic and grammatical description of Didinga, pretending to greater accuracy than Driberg 1931 but containing significantly less data.


Randal 1998 = Randal, Scott. A grammatical sketch of Tennet. In: Surmic Languages and Cultures. Ed. by Gerrit J. Dimmendaal and Marco Last. Köln: Rüdiger Köppe Verlag, pp. 219-272. // An important supplementary material to M. Yigezu's lists for Tennet, although, unfortunately, illustrative data are relatively scarce, since the sketch in question focuses much more on syntactic than phonological or morphological issues. Partially based on the author's earlier dissertation thesis "Nominal morphology in Tennet", which, however, contains almost no extra material.


Lyth 1971 = Lyth, R. E. A Murle grammar / A Murle vocabulary. University of Khartoum. // This detailed grammatical description of Murle, dating back to the 1940s, is also accompanied by a representative Murle-English / English-Murle vocabulary. Prosodic information is not included, and there are reasons to doubt the accuracy of phonetic transcription in some cases, but the source is very important at least in terms of its sheer volume.

Tucker 1951 = see above (Narim).


Yigezu & Dimmendaal 1998 = Yigezu, Moges; Dimmendaal, Gerrit J. Notes on Baale. In: Surmic Languages and Cultures. Ed. by Gerrit J. Dimmendaal and Marco Last. Köln: Rüdiger Köppe Verlag, pp. 273-318. // A brief sketch of Baale phonology and grammar, illustrated by some texts and a brief vocabulary. Contains very little lexical data that is not included in [Yigezu 2001], but transcriptional notation and semantics are generally more accurate, and some additional paradigmatic information (e. g. plurals of nouns) can be extracted as well.

Suri (Chai)

Last & Lucassen 1998 = Last, Marco; Lucassen, Deborah. A grammatical sketch of Chai, a Southeastern Surmic language. In: Surmic Languages and Cultures. Ed. by Gerrit J. Dim\-men\-daal & Marco Last. Köln: Rüdiger Köppe Verlag, pp. 375-436. // Rather brief grammatical sketch of the Chai language of Suri, accompanied with a small, but seemingly accurately transcribed, vocabulary.

Abbink 1993 = Abbink, J. Suri-English Basic Vocabulary. In: Frankfurter Afrikanistische Blätter, 5, pp. 39-64. // A small vocabulary of Chai (contrary to the title, not at all restricted to "basic" vocabulary in the Swadesh understanding), somewhat inferior to the data in Last & Lucassen 1998, since the author pays little attention to the morphological constituency of Chai items (particularly verbs).


Turton et al. 2008 = Turton, David; Yigezu, Moges; Olibui, Olisarali. Mursi-English-Amharic Dictionary. Addis-Ababa: Ermias Advertising. // Relatively large dictionary of Mursi, oriented mainly at native speakers and therefore not particularly accurate with phonology (i. e. no indication of prosody); various semantic inaccuracies and typos also render the source less useful than it could have been.

Turton & Bender 1976 = Turton, David; Bender, M. L. Mursi. The Non-Semitic Languages of Ethiopia. Ed. by M. Lionel Bender. East Lansing: Michigan State University, pp. 533-562. // A brief grammatical sketch of Mursi that contains some relevant paradigmatic information, e. g. on pronouns and suppletive verbal stems.

Siebert & Caudwell 2002 = Siebert, Ralph; Caudwell, Simon. Sociolinguistic Survey Report of the Melo (Malo) and Mursi Languages of Ethiopia. SIL International. // An SIL list of Mursi basic lexicon. Useful as a control source, but seems to contain rather numerous mistakes and typos.


Ricci 1972 = Ricci, Lanfranco. Materiali per la lingua Mekan. In: Rassegna di Studi Etiopici, 25, pp. 90-455. // A detailed description of Me'en grammar, based on the Tishena dialect and accompanied with a large dictionary that also includes potential etymological parallels in other Surmic languages and beyond. Most of the data actually come from a much earlier collection, accumulated by Fulvio Sudano in 1939-41; their phonetic and semantic quality are highly dubious in many instances.

Will 1989 = Will, Hans-Georg. Sketch of Me'en grammar. In: Topics in Nilo-Saharan Linguistics. Ed. by M. Lio\-nel Be\-nder. Hamburg: Helmut Bus\-ke Verlag, pp. 129-150. // A brief description of Me'en grammar (based mainly on the Tishena dialect).

Will 1991 = Will, Hans-Georg. Me'en: a Bodi-Tishena dialect compari\-son. In: Journal of Ethiopian Studies, 24, pp. 97-113. // A brief phonetic, grammatical, and lexical comparison of the two main dialects of Me'en.

Will 1993 = Will, Hans-Georg. Me'en phonology. In: Journal of Ethiopian lan\-guages and literature, 3, pp. 61-80. // A brief description of Me'en phonology (based mainly on the Tishena dialect).


Hieda 1990 = Hieda, Osamu. Koegu, a preliminary report. Journal of Swahili and African Studies, 1, pp. 97-108. // A brief sketch of Kwegu (Koegu) and its speakers. Contains some notes on the divergence between Koegu (Muguji) and Kwegu (Yidinic) dialects.

Hieda 1991 = Hieda, Osamu. Koegu Vocabulary, with a reference to Ka\-\-ra. In: African Study Monographs, Suppl. 14, pp. 1-70. // A classified vocabulary of the Koegu (Kwegu) language (Muguji dialect), collected by the author and accompanied with the corresponding equivalents in Kara (a neighbouring Omotic language that exerts a heavy influence on modern Kwegu).

Hieda 1998 = A sketch of Koegu grammar: Towards re\-constructing Proto-Southeastern Surmic. In: Surmic Languages and Cultures. Ed. by Gerrit J. Dim\-men\-daal & Marco Last. Köln: Rüdiger Köppe Verlag, pp. 345-376. // A relatively brief sketch of Kwegu (Koegu) grammar (Muguji dialect).


Bender 1983 = Bender, M. Lionel. Majang phonology and morphology. In: Nilo-Saharan Language Studies. Ed. by M. Lionel Ben\-der. Michigan: East Lansing, pp. 114-147. // Grammar sketch of Majang with a lot of illustrative lexical data.

Cerulli 1948 = Cerulli, Enrico. Il linguaggio dei Masongo nell'Etiopia Oc\-cidentale. In: Rassegna di Studi Etiopici, 7, 2, pp. 131-166. // An early grammatical sketch and a very brief vocabulary of Majang.


1. General.

For the sake of general uniformity, we consistently select [Yigezu 2001] as our main source of data, despite the fact that the wordlists collected by M. Yigezu are not always perfectly corroborated by alternate sources (including more detailed and/or more recent works), especially in terms of phonetic detail (inaccuracies in prosodic marking and in determining ATR quality of the vowels have been noticed, sometimes even in the form of simple typos, as the same word may be transcribed differently in different parts of Yigezu's thesis). In general, however, the work seems at least more reliable than the majority of older publications on Surmic.

Another comparative source of data on Surmic basic lexicon is [Bender 1971], a large paper that includes a collection of wordlists collected by the author on the Ethiopian Surmic languages: Murle, Zilmamu, Tirma, Mursi, Me'en, and Kwegu (also Mesengo = Majang, but this has largely been superseded by Bender's later publication [Bender 1983]). Unfortunately, none of these wordlists can be used as primary sources, and the utmost care has to be exercised even in using them as control sources: comparison with more detailed and reliable dictionaries shows that semantic inaccuracies are quite frequent, not to mention phonetic mistakes.

Morphological segmentation of Surmic words is approximate. For instance, nominal stems in Surmic tend to have the shape CVCV, where the second vowel is likely to be an old desemanticized nominal suffix ("fossilized classifier"). However, we prefer to include morphological segmentation only in those cases where suffixation is productive, i. e. when it is possible to identify a suffixal element by comparing singular and plural forms (thus, Narim tàmːù 'rain' without segmentation, but k=ʋ̀rʋ̀ð-ʌ̀ 'dog' because the final vowel changes to -e in the plural form, etc.).

I. Narim

There are no detailed descriptions or dictionaries for the Narim (Longarim) language. The main source is [Yigezu 2001]; for additional perspective, we also list equivalents from the older source [Tucker 1951] and the newer source [Stirtz 2011] where they are available. In a couple of cases where no equivalent for the Swadesh meaning has been found in Yigezu 2001, we allow substituting the corresponding equivalent from [Tucker 1951] to minimize gaps, although this is slightly risky, since every now and then older and newer sources differ as to the exact equivalent (either the result of semantic inaccuracies or researchers describing slightly different dialects).

II. Didinga

There are two phonetic and grammatical sketches that specifically focus on Didinga: [Driberg 1931] and [Odden 1983]. The latter, as could be expected, is on the whole more adequate in its description of the nuances of various aspects of Didinga phonetics and grammar, but has the disadvantage of containing a very limited amount of illustrative material. [Driberg 1931], on the other hand, includes a fairly large vocabulary of the language. Despite occasional inaccuracies in transcription (including systematic inaccuracies, such as mistaking Didinga tone for stress, etc.) and semantic notation, this vocabulary is very useful as a means of corroboration of Yigezu's wordlist material.

III. Murle

There is at least one detailed grammatical description of Murle available, together with a large vocabulary [Lyth 1971]; unfortunately, it dates back to the 1940s and was not prepared by a professional linguist, so the accuracy of phonetic notation is often questionable (prosodic information, for instance, is not noted at all). On the other hand, the sheer size of the vocabulary allows to fill in several Swadesh slots that are not taken care of in Yigezu's lists, as well as correct some of Yigezu's mistakes. Additionally, [Tucker 1951] is a brief grammatical sketch of Murle with a brief accompanying vocabulary; the source has been taken into account, but it adds very little information to what is already available from Yigezu's and Lyth's data. For comparative and control purposes, we also list Murle forms that are adduced in [Bender 1971].

IV. Baale

Baale, also known as Kacipo-Balesi, is a difficult language for lexicostatistics; rather clearly belonging to the Southwestern branch of Surmic, it has, nevertheless, been subjected to serious areal influence on the part of Suri (Southeastern branch), borrowing some basic lexicon from that language. Unfortunately, the only usable source for Baale, outside of the wordlist in [Yigezu 2001], is the brief grammar sketch and small lexicon in [Yigezu & Dimmendaal 1998], meaning that the same author is responsible for both sources and no control sources are available.

An additional problem is that the so-called "Zilmamu" language, often said to be a dialect of Baale and currently represented in our materials only by a sketchy wordlist in [Bender 1971], while showing unquestionable Surmic affiliation and plenty of matches with Southwestern languages, has a large number of items that have nothing in common with the respective Surmic equivalents. With this data, it is all but impossible to show if "Zilmamu" is really a dialect of Baale, or if it represents a completely autonomous branch of South Surmic. For the sake of tradition, we list the "Zilmamu" forms in the notes section next to notes on Baale proper, but it should be remembered that the issue of "Zilmamu" affiliation remains an open one and requires additional data.

V. Suri

Of the two primary dialects of Suri, Chai and Tirma(ga), only Chai currently satisfies the requirements for a proper lexicostatistical treatment (very little data on the whole and not a single proper vocabulary is available as yet for Tirma). In addition to Yigezu, two more lexicographical descriptions for Chai have been published: [Abbink 1993] and [Last & Lucassen 1998], both comprising brief vocabularies of several hundred entries. Of these, [Last & Lucassen 1998] is particularly important, since the description is an appendix to the authors' grammatical sketch of the language and shows understanding of its grammatical peculiarities, such as the rather complex suppletivism between imperfective and perfective verb stems, very important for lexicostatistical purposes. For the sake of consistency, we still treat [Yigezu 2001] as our main source, but regularly supplement it or even correct it with data taken from [Last & Lucassen 1998], especially since Yigezu regularly only adduces the perfective stems of verbs, omitting their suppletive imperfective counterparts altogether.

As for Tirma, currently we only list the forms that were easily extracted from M. L. Bender's Tirma wordlist in [Bender 1971]. Although it shows a few lexical discrepancies between Chai and Tirma, these cannot be ascertained due to the overall unreliable nature of Bender's lists.

VI. Mursi

Although lexical data on Mursi are available from at least two additional sources - the extensive dictionary [Turton et al. 2008] and the wordlist in [Siebert & Caudwell 2002], which is itself larger than most standard SIL lists - none of the available sources are completely reliable, suffering from lack of transcriptional and semantic accuracy; this emerges rather clearly from their comparison with each other, as well as the somewhat more accurately collected data of the closely related Suri language. (The situation is even worse with M. L. Bender's brief wordlist of Mursi in [Bender 1971], taken here into consideration for comparative purposes). Therefore, although we still use [Yigezu 2001] as our primary source for the sake of consistency, the final judgement on the choice for primary slot rests on careful comparison of all three sources. In at least one case ('die'), we had to replace Yigezu's entry with a different item for lack of confirmation. We can only hope that sometime in the future, a definitive dictionary for Mursi will eventually appear.

VII. Me'en

The Me'en (Mekan) language in Ethiopia is represented by two primary dialects, sometimes even defined in literature as different languages: Bodi and Tishena. Of these, Bodi is used by a minority of speakers as compared to Tishena, but is sometimes referred to as more archaic in a number of features. Both dialects have been studied, and a certain amount of lexical material is available on each of them; in particular, M. Yigezu's data in [Yigezu 2001] seems to reflect the Bodi dialect, whereas earlier sources, such as F. Sudano's notes that were used in [Ricci 1972], and H.-G. Will's series of papers on Me'en usually focus on Tishena.

Unfortunately, all the sources of data are either too incomplete or too unreliable to allow us the separate construction of two different wordlists for Bodi and Tishena. Based on what was possible to ascertain from a brief comparative description of the two dialects in [Will 1991], as well as scraps of information from other sources, it may be concluded that lexicostatistical variation between Bodi and Tishena probably does not exceed 5-6% at most; such a discrepancy only warrants the presence of two different lists if the available sources are completely reliable, otherwise it would be too easy to introduce crucial errors that reflect semantic inaccuracy on the part of field workers rather than valid historical data.

Consequently, our strategy is as follows. For the sake of uniformity, [Yigezu 2001] (Bodi dialect) is used as the default source for the wordlist. To confirm his data, in the notes section we also quote the relevant Tishena equivalents from [Ricci 1972] and from H.-G. Will's papers, particularly from [Will 1991] where it is possible to see the phonetic and occasional lexical differences between Bodi and Tishena. Where there seems to be sufficient evidence for a mistake on Yigezu's list, we take the liberty of replacing Yigezu's word for Will's (this happens in one well-justified case, namely, the word for 'star' q.v.). We also list the Me'en forms in [Bender 1971] (they seem to rather agree with Yigezu's Bodi than Tishena, but not always).

VIII. Kwegu

The Kwegu (Koegu) language in Ethiopia is represented by two main dialects, which Osamu Hieda [1990, 1991] defines as "Koegu" and "Kwegu" respectively. However, since this orthographic convention may be confusing (and relies on a subtle phonetic peculiarity that may or may not be relevant for all speakers of the respective dialects), we will refer to the language on the whole as "Kwegu" (its normative name in the Ethnologue and Glottolog systems), and to the dialects as "Muguji" (the Kara name for Koegu) and "Yidenic" (the Bodi name for Kwegu), respectively.

Extensive data are available only on the Muguji dialect: in addition to Yigezu's wordlist, there is a detailed classified vocabulary [Hieda 1991] and a grammar sketch [Hieda 1998]. These have been seriously criticized by Yigezu for various inaccuracies, including phonetic mistakes (e. g. postulating a five-vowel system instead of a seven-vowel system for the language) and lexical mistakes (e. g. underestimating the influence of Kara, a neighbouring Omotic language, on the lexicon). Nevertheless, Yigezu's wordlists, as is clearly seen on the material of other languages than Kwegu, are far from perfect themselves, and the use of Hieda's materials as a "control source" for Yigezu is required, at least until more reliable materials have been assembled.

In the notes section, we also list some discrepancies between the Muguji and Yidenic dialects the way they are presented in the brief paper [Hieda 1990]. For the most part, they are of a purely phonetic nature, but at least in one case ('dog') there is a lexical discrepancy (a recent borrowing from Kara in Muguji vs. a word of unclear origin - possibly an old pre-Surmic substrate remain - in Yidenic). However, as of now it is impossible to estimate the lexicostatistical difference between the two dialects due to lack of reliable data on Yidenic.

IX. Majang

Majang (Masongo) is the most highly divergent Surmic language, constituting one of its two primary branches (Northern Surmic). Several sources of Majang data are available, but not a single definitive grammar or vocabulary have been produced so far. The primary slot is filled in with the corresponding equivalent from [Yigezu 2001]; in the notes section, we also make heavy use of works by M. L. Bender, most notably the brief grammatical sketch [Bender 1983], and of the much earlier description in [Cerulli 1948], to which a very small vocabulary is attached. Due to the age of Cerulli's work, and to the general phonetic and semantic inaccuracies that may be suspected in Bender's and Yigezu's materials alike, no single entry can be qualified as completely reliable; however, most of the time the sources agree with each other at least as to the lexical (if not necessarily phonetic) details.

2. Transliteration.

As far as our main source on most Surmic languages is concerned, M. Yigezu's data generally follow IPA standards and therefore only require "cosmetic" alterations when transliterated to UTS. Specific notes on other sources are as follows:

(a) Driberg 1931 for Didinga: Driberg's dh = UTS ð, th = UTS θ, ny = UTS ɲ; long vowels ā, ī, ō, ū are transliterated as aː, iː, oː, uː. It is not quite clear what Driberg means by (said to be pronounced "as ai in air"): most likely, this is just a standard long correlate to short e, perhaps uttered with a slight automatic trace of diphthongization - respectively, we transliterate it as . Additionally, except for e and ɛ, Driberg does not distinguish between +ATR and -ATR vowels. He also very rarely indicates tone (only in those cases where words with different tones form minimal pairs), but regularly indicates stress. For accuracy's sake, we preserve his stress notations, even though D. Odden (1983) doubts the importance of stress for Didinga as such.

(b) Odden 1983 for Didinga: The only important retransliteration is that of Odden's lower-case vowels a, e, i, o, u as +ATR vowels (UTS ʌ, e, i, o, u) and his upper-case vowels A, E, I, O, U as -ATR vowels (a, ɛ, ı, ɔ, ʋ). Additionally, Odden explicitly marks the high tone, leaving the low tone unmarked; we systematically mark both (high and low ).

(c) Lyth 1971 for Murle: Lyth's old transcription is generally adapted to reflect the same values as suggested by Yigezu's description of the system, thus dh = ð, ny = ɲ, c = ɕ, j = ʓ. The voiced consonants b, d may actually be implosive (ɓ, ɗ), as would be suggested by comparison with other sources. However, we prefer not to modify them, since Lyth himself says nothing about implosive articulation, and dialectal fluctuations between voiced and implosive articulation are quite common for various Sudanic languages. Concerning vocalism, Lyth's notations are left intact, except for the vowel ä, which he explains as "as in French 'de'"; to avoid ambiguity, it has been tentatively recoded as ǝ.

(d) For Baale, transcriptional systems in [Yigezu 2001] and [Yigezu & Dimmendaal 1998] are slightly different: the latter omits certain phonetic details (e. g. consistently transcribes the intervocalic voiced velar as -g- as opposed to fricative -ɣ- in [Yigezu 2001]) and renders the low tone in [Yigezu 2001] as mid tone (i. e. ā instead of , etc.; the low tone is said by the authors to only have been encountered as part of a composite "mid-low" tone, i. e. the result of tonal samdhi).

(e) Abbink 1993 for Chai (Suri): The author regularly uses ch to denote the palatal affricate (UTS ɕ) and ny to denote the palatal nasal (UTS ɲ), but occasionally also uses for seemingly the same thing; this looks like an inconsistency, so we "neutralize" both graphical variants to ɲ.

(f) Turton et al. 2008 for Mursi (orthographic transcription):
Vowels: -ATR vowels ɛ, ɔ are transcribed as ê, ô in Mursi orthography.
Consonants: (1) Implosives ɓ, ɗ are transcribed as bh, dh; (2) Glottal stop ʔ is transcribed as apostrophe ʼ ; (3) Alveolar affricates and fricatives č, ǯ, š are transcribed as ch, j, sh respectively; (4) Nasals ŋ, ɲ are transcribed as ng, ny.
Prosody: Remains completely unmarked.

(g) Ricci 1972 for Me'en: The situation in this source is somewhat confused. The author prepares his description of Me'en based on notes that were collected much earlier (1939-1941) by F. Sudano. In those notes, Me'en forms were transcribed using Italian orthography (e. g. rendering the palatal nasal ɲ as gn, the post-alveolar affricate š as sc/i/, etc.). This transcriptional system is sometimes left intact (e. g. in the Italian-Me'en index on pp. 428-453), but is usually modified by Ricci to reflect a more "linguistic" approach, although that still does not prevent him from occasional "Italianisms", such as consistently rendering velar k as c.

For the sake of accuracy, we present Ricci-Sudano's entries in two forms: tentative transliteration to UTS, put down in cursive, and Ricci's modified transcription, included in curly brackets (according to our usual convention for "orthographic" variants). Sudano's original transcription is not included. The major changes from Ricci's transcription to UTS are as follows:

[Ricci 1972] UTS
c k
ts c
ng ŋg

We also omit all prosodic diacritics in our transliteration; although "grave" and "acute" accentuation are regularly marked in Sudano's notes, there is no precise explanation of what it is they are really supposed to mean, and in any case, it is dangerous to trust prosodic information recorded as early as in the mid-20th century (especially keeping in mind that even newer data, e. g. Yigezu's notation vs. H.-G. Will's, have numerous discrepancies in the sphere of prosody).

(h) Hieda 1991 for Kwegu: This source uses a simplified orthography that is largely diacritics-free for printing reasons. The following conventions are observed: 'b, 'd > UTS ɓ, ɗ (implosive consonants); ts' > UTS (ejective alveolar affricate); c, j, c', sh > UTS ɕ, ʓ, ɕʼ, ʆ (palatal affricates and fricatives); ny > UTS ɲ (palatal nasal); ng' > ŋ (velar nasal); ? > UTS ʔ (glottal stop). High tone is marked as in the original source; low tone is unmarked in the original source and converted to UTS .

(i) Cerulli 1948 for Majang: This source utilizes a rather specific orthography. Its special symbols are transliterated into the UTS as follows:

[Cerulli 1948] UTS
p` β
ʼ ʔ

Note: The consonant is described by Cerulli as a "cacuminal explosive", but in the absolute majority of encountered forms it regularly corresponds to the implosive ɗ in Yigezu's and Bender's materials. We take the liberty of re-transcribing it as ɗ for the purposes of compatibility (and also because it is highly likely that the consonant was indeed implosive, mistaken for an explosive with a different place of articulation).

Database compiled and annotated by: G. Starostin (last update: June 2015).