Annotated Swadesh wordlists for the Hadza group (Hadza family).

Languages included: Hadza [hdz-hdz].


Sands, Maddieson & Ladefoged 1996 = Sands, Bonny; Maddieson, Ian; Ladefoged, Peter. The Phonetic Structures of Hadza. In: Studies in African Linguistics, 25.2, pp. 171-204. // One of the most accurate and reliable descriptions of the phonetics of Hadza to date, accompanied by a 232-item wordlist recorded from seven Hadza informants in August 1991.

Sands 1998 = Sands, Bonny. Eastern and Southern African Khoisan. Eva\-lu\-a\-ting Claims in Distant Linguistic Relationships. Ed. by Rainer Vossen. Köln: Rüdiger Köppe Verlag. // Monograph dedicated to the issue of genetic relations between the various "Khoisan" languages. Contains basic lexicon lists for multiple languages, including Hadza; unfortunately, the list is somewhat idiosyncratic, specially tailored for "Khoisan needs", and only coincides to a small extent with the standard Swadesh list.

Sands Ms. = Sands, Bonny. Hadzabe vocabulary. // Typed manuscript, containing a large wordlist elicited by B. Sands in 1991/1992. Phonetic notation and semantic accuracy are generally of a very high quality, except that the data do not contain tonal notation.

Bleek 1956 = Bleek, Dorothea F. A Bushman Dictionary. American Oriental Society: New Haven, Connecticut. // (A huge (almost 700 pages) collection of comparative data on Khoisan that includes both Dorothea F. Bleek's own collection and data from numerous other researchers published up until the 1930s (W. Bleek, L. Lloyd, etc.). Transcription quality varies in between all the different sources, but is generally unreliable, quite typical of all Khoisan data published before the second half of the XXth century. Nevertheless, the edition still contains a wealth of priceless data, particularly on extinct North and South Khoisan languages.)

Dempwolff 1917 = Dempwolff, Otto. Beiträge zur Kenntnis der Sprachen in Deutsch-Ostafrika. In: Zeitschrift für Kolonialsprachen, 7, pp. 309-325. // A brief paper containing three wordlists (Iraqw, Datooga, and Hadza) collected by O. Dempwolff in 1910-11; the Hadza wordlist, along with E. Obst's data published in 1912, remains one of the oldest sources on the language and is thus of considerable historic interest.

Tucker, Bryan & Woodburn 1977 = Tucker, Archibald; Bryan, Margaret; Woodburn, James. The East African Click Languages: a phonetic comparison. In: Zur Sprach\-ge\-schichte und Ethnohistorie in Afrika. Ed. by Wilhelm J. G. Möhlig, Franz Rottland & Bernd Heine, pp. 300-323. // A brief description of the phonetic structures of Sandawe, Hadza, and Dahalo, richly illustrated by lexical material.


1. General.

Although the general bibliography on the language of Tanzania's Hadzabe is already quite large, we still lack a comprehensive, accurately compiled dictionary of Hadza; consequently, the wordlist has to be compiled from several different sources. As the fundamental basis for the wordlist, we choose the published and unpublished results of fieldwork conducted by Bonny Sands: namely, the 232-item lexical list published in [Sands, Maddieson & Ladefoged 1996] (containing tonal notation) and, in order to fill in the remaining gaps, the much larger - but tonally unmarked - wordlist contained in [Sands Ms.].

In between themselves, these sources allow for the compilation of a comprehensive wordlist. However, the notes section also lists alternate variants from several older sources: (a) Otto Dempwolff's 1917 paper, one of the oldest sources on Hadza (E. Obst's paper from 1912, technically the first collection of data on Hadza, is not listed because the phonetic notation is of extremely poor quality); (b) the results of D. Bleek's fieldwork, included in [Bleek 1956]; (c) the collective paper [Tucker, Bryan & Woodburn 1977], presenting a phonetic analysis of data collected by James Woodburn.

NB: Most of the morphological segmentation applied to the data is confined to separating the root from the productive synchronic suffixes (gender and number for nouns, personal markers for verbs, etc.), in which we generally follow Sands et al. However, one of the peculiarities of Hadza nominal stems is that a statistically significant part of them shows a limited set of word-initial vowels (ʔa=, ʔi=, ʔu=) or combinations of vowels with laryngeal ɦ- (ɦa=, ɦi=, ɦe=, ɦo=). From a structural point of view, they produce a strong impression of representing fossilized class prefixes, and even though there is almost no internal evidence for separating them from the "bare" root (other than a large group of supposedly "prefixless" forms in D. Bleek's materials, which may in reality represent incorrect segmentation), it seems reasonable to formally segment them as prefixes in order to provide a more reliable basis for external comparison. All such situations are explicitly marked in the notes.

2. Transliteration.

The UTS system, being essentially based on the IPA with minor changes, is quite close to the completely IPA-based system in [Sands, Maddieson, & Ladefoged 1996]; transcription systems for Hadza employed by other researchers generally require more transliterational efforts. It should also be noted that serious discrepancies are attested in various researchers' interpretation of the phonological structure of Hadza (see notes below).

UTS Sands, Maddieson, Ladefoged 1996 Sands Ms. Bleek 1956 Tucker, Bryan, Woodburn 1977
p p p p p
p ph ph
b b b b b
f f f f f
m m m m m
w w w w w
t t t t t
t th th
d d d d d
n n n n n
r r r r r
c ts ts ts ts
ʒ dz dz dz dz
tsʼ ts' tsʔ tsʼ
s s s s s
č ch c
ǯ j j
čʼ tʃʼ chʼ tʃʔ
š š sh ʃ ʃ
ƛ tʎ̥ tl tl ~ kl tl
ƛʼ tʎ̥ʼ tl' tlʔ ~ klʔ tlʼ
ɬ ɬ hl xl hl
l l l l l
y j y y y
k k k k k
k kh kh
g g g g g
k' k
ŋ ŋ ng ŋ ŋ
ʔ ʔ ' ʔ ʔ
ɦ ɦ h h h
ǀ / ǀk ~ ǀg /
ǀʼ ŋ̥ǀʼ /' ǀ
ǀʰ ǀh /h
ɳǀ ŋǀ n/ ǀn n/
! k! ! !k ~ !k ! ~ !!
ŋ̥!ʼ !' ! !ʼ ~ !!ʼ
!h !h ~ !!h
ɳ! ŋ! n! !n n! ~ n!!
ǀǀ kǀǀ // ǀǀk ~ gǀǀ //
ǀǀʼ ŋ̥ǀǀʼ //' ǀǀ //ʼ
ǀǀʰ ǀǀh //h
ɳǀǀ ŋǀǀ n// ǀǀn n//
a a a a a
e (= ɛ) e e e ~ ɛ e ~ ɛ
i i i i i
o (= ɔ) o o o ~ ɔ o ~ ɔ
u u u u u


1. Aspirated stops and affricates seem to be an integral part of Hadza inventory, but are not systematically marked in [Sands Ms.], and other sources allow for some variation as well; we mark everything as originally transcribed in the sources.

2. The laryngeal h, usually written in most sources as such, is recognized by B. Sands et al. as voiced aspiration and consistently transcribed as ɦ. We follow this convention, but it must be remembered that there is no phonological opposition between ɦ and h in Hadza.

3. There are three commonly recognized click influx types (dental ǀ, alveolar !, lateral ǀǀ) and three commonly recognized click accompaniment types (zero, glottal stop, nasalized) in Hadza. In [Tucker, Bryan, & Woodburn 1977], the authors distinguish between the "hard" palato-alveolar click influx (!) and the "flapped" palato-alveolar click influx (!!), but even according to them, these mostly appear to be in free variation.

4. As for the accompaniments, existence of a fourth accompaniment (aspirated) is postulated in [Tucker, Bryan & Woodburn 1977] and somewhat agrees with the data of D. Bleek, but is denied in [Sands, Maddieson, Ladefoged 1996]; we preserve the aspirated accompaniment in our transliteration wherever it is present, but it should be noted that in most cases, ǀʰ, , ǀǀʰ are probably just free variants of ǀ, !, ǀǀ. Furthermore, it must be noted that the glottal stop accompaniment in Hadza, according to [Sands, Maddieson, Ladefoged 1996], is regularly accompanied with weak prenasalization, which they consistently mark; since this prenasalization is automatic, we omit it in our transliteration.

5. It seems that Hadza does not have a +/-ATR phonological distinction, but some sources (Tucker, Bryan, & Woodburn 1977 and Bleek 1956 in particular) still try to maintain phonetic difference between e, o and ɛ, ɔ; we do not omit them in our transliteration, but it should be remembered that the difference is primarily positional.

6. In [Sands, Maddieson, Ladefoged 1996], combinations of velars with the labial glide are judged as separate labiovelar phonemes (, etc.), whereas in most other sources they are usually treated as clusters (kw, etc.). Since even the wordlist at the end of the mentioned paper transcribes these elements as kw, etc., we unify the cluster notation for all sources.

7. The tonal structure of Hadza has not been studied particularly well, and there are conflicting descriptions: e. g. Sands et al. only distinguish two general tones (high and low), whereas Tucker, Bryan & Woodburn postulate up to four register tones plus two contour tones. We transliterate the tonal notations in all sources in as strict accordance with the UTS system as possible; however, tonal notation in general for Hadza cannot be considered reliable.

Database compiled and annotated by: G. Starostin (last update: December 2012).