Who Are the Waswahili?

Author(s): Carol M. Eastman

Source: Africa: Journal of the International African Institute , Jul., 1971, Vol. 41, No. 3
(Jul., 1971), pp. 228-236
Published by: Cambridge University Press on behalf of the International African
Institute

Stable URL: https://www.jstor.org/stable/1158841

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

and Cambridge University Press are collaborating with JSTOR to digitize, preserve and extend
access to Africa: Journal of the International African Institute

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
 [z228]

WHO ARE THE WASWAHILI?

CAROL M. EASTMAN

THE question as to who are the Swahili has long been controv
C. H. Stigand is responsible for the most widely accepted definiti
A Swahili . . . in the more confined sense of the word, is a descendant o
original Arab or Persian-Arab settlers on the East African Coast. In the b
the word it includes all who speak a common language, Swahili.'

As a field-worker doing linguistic research in East Africa on the Sw

this problem of identification became acute for me. Certain people
selves Swahili and others not, while some of those so considered may
self concept at all. With the standardization of Swahili and its growt
language in Tanzania and its use as a lingua franca all over East Afric
of definition is about to resolve itself. Swahili-speaking people are be
and Tanzanians, rather than Swahili or non-Swahili. A pattern, how
evident allowing me to categorize componentially the Swahili peopl
The Swahili language, while indeed a lingua franca is also the first and
for some East Africans. These people are Swahili in a sense that
For others, Swahili is a language used in order to communicate with t
neighbours, and friends from diverse linguistic backgrounds.
Linguistic criteria may be used to differentiate people who are calle
selves Swahili. Some useful indices are:

I. The amount and kind of loan words a speaker employs, e.g. Arabic, English,
Gujerati, etc.
2. The degree to which a speaker adheres to the grammatical rules of the lan-
guage, e.g. to what extent grammatical concord is used syntactically; whether
or not the intonation patterns are influenced by other languages.
3. Is Swahili the speaker's first and only language? If not, what are his other
languages ?

Complicating factors such as the teaching of Swahili to all children in many East
African schools and the implication of such a procedure that the ' best' speakers are
more truly Swahili than the' worst' also need to be taken into account.2 An approach
such as this can only show how many types of Swahili speakers there are.
Without resting our definition of the WaSwahili on criteria like these, i.e. how well
or how grammatically a person speaks, I thought it might be possible to arrive at an
analysis based on contrasts which correspond to East African conceptions of who
a Swahili is, the idea being that such a contrastive analysis would reflect what is
considered to be the inclusive connotation of the term Swahili as applied to a people
and what a person considers his own position to be within that connotation.

X Stigand, 1913, ' The Swahili', chapter 6, p. I 6. Polome, I967, chapter i ' The Language Situation'
S Regarding the people who speak Swahili see Section A, pp. x-8.

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
 WHO ARE THE WASWAHILI? 229

It seems that while a Swahili must have knowledg

all who know or use Swahili are considered or consid
other hand, in the broadest sense, to be a Swahili d
also being something else. For example, a member of
and speak Swahili. He may call himself a Swahili an
a Kikuyu to be termed a Swahili would be unusu
a statement, given later in this paper, as to whom
in which Swahili speakers of the 'suburbs' coul
up-country tribes such as the Kikuyu. Thus the ran
is very great indeed and is variable according to th
My field notes constantly indicate '. . . is non-
considers self Swahili ' or ' ... is considered to be an
with Muslim names,'. . . does not consider himself
use only 'Swahili' informants for my work. The v
as I looked for my informants will be indicated bel
For the purpose of my study of Swahili dialect, m
from one end of the resulting definitional continuu
Muslims of coastal parentage who had Muslim n
extent, geographically and culturally homogeneous
The basic criteria which emerged in my search f
Swahili speakers were:
i. Religion: Muslim, Christian, other
2. Geographical origin: self or parents born on th
3. Name: Muslim, Christian, other
4. Educational background: Koranic school, Miss
A. H. J. Prins' gives the following criteria which a
a type of civilization, and a Coastal habitat. But he also

I. '... any man from the coast who speaks Swahili a

2. ... any Muslim African from anywhere.'
3. one's inferior
4. anyone who is a mixed descendant of Africans, Ara
the exact ethnic makeup or proportion of these comp
East African coast or the islands adjacent to it'.
5. any coastal African who speaks Swahili either as hi
speech of his daily life. (Prins, pp. I 1-2)

Thus, Prins himself is not very clear in his definit

further that' The name Swahili is not applied by the
selves' (p. 12). He mentions that earlier writers did
was that many people did apply the term to thems
agreement with Prins's further observation that' In
however, it is always true that a man is never Swa
Prins also indicates that, at the time of his stu
to imply negative characteristics such as 'slave
I Prins, x96i, pp. II-i2.

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
 230 WHO ARE THE WASWAHILI?
occupational position, and a general "boorish ", unciviliz
look on life' although some people answering such a de
Arab descent.
Today, however, the situation seems to be changing. An example of this may be
seen among the members of the Arab League in a place such as Mombasa, Kenya.
The members today purposely emphasize their self concept of being Swahili rather
than Arab. If a person is considered a Swahili today, he is assumed to be of' good '
lineage. It may be argued that this is a matter of political expediency following the
attainment of independence in the countries of East Africa and the subsequent
development of nationalism. As mentioned above, all distinctions such as these
discussed here are being discouraged with the intent that people should think of
themselves as Africans first, and then as Kenyans or Tanzanians.
My findings indicate that all Swahili are to some extent Swahili users, no Swahili
may also be an Asian or European, and that no one is primarily a Swahili. The two
extremes are Muslim-Arab Coastal Swahili and non-Muslim Up-Country African
Swahili. All people who are considered Swahili today are native-born Kenyans or
Tanzanians. The references in this paper to coastal and non-coastal areas refer to the
coasts, off-shore islands, and interiors of those two countries alone. The non-
Muslim-Swahili tend to refer to Muslim Swahili speakers as 'Arabs' while the
Muslim-Swahili call the non-Muslims 'Africans '. The non-Muslims confuse Arabs
with other Muslims since many Muslims descend from Arabs or call themselves
Arabs. The Muslims call the non-Muslims Africans, implying that they have no
Oriental background.
Since independence there is a growing trend in both countries for all to consider
themselves Africans. At the same time, with the current development of Swahili
literature and pride in Swahili as a national language, it seems that the extremes are
now merging. Indeed, the term is today rarely used in any derogatory sense. The
decline in numbers of people who use the Arabic script' is hastening this merger,
although, among literary men, a distinction is made between those who write tradi-
tional poetry (using either Roman or Arabic script) who regard themselves as 'true'
Swahili, and those who use modern literary forms and Roman script who are regarded
more as Africans than Swahili.
People on the island of Pate and in parts of northern Kenya made it plain to me
that in their opinion those who speak the Bajun dialect of Swahili are Bajun and not
Swahili. The Bajun made this clear as did their neighbours on Lamu. It seemed to me,
however, subjectively, that the Bajun took pride in being Bajun while at the same
time the inhabitants of Lamu were proud to be Arab, Swahili, or African. The Bajun
are Muslim, Coastal, of mixed parentage (i.e. generally of Persian and/or Arab plus
African descent), with Muslim names and users of the Arabic script.z
Using the four contrasts mentioned above, I found that eight groups of people
who fall under the broad term Swahili could be distinguished. The following labels
may be attached to the feature complexes defining each type:

Harries, 1962, chapter 2. Arab-Swahili have ' tribal' roots. The Bajun report
2 Among the 'Arab' Swahili, just as among the that there are Chinese-Bajun among them. See
'African', there are many ethnic groups. As you Eastman and Topan, I966.
have Giriama Swahili or Digo Swahili, so, too, the

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
 WHO ARE THE WASWAHILI? 231
I. Mwarabu safi 'Pure Arab'
2. Mwarabu 'Arab'
3. MSwahili safi ' Pure Swahili'
4. MSwahili' Swahili'
5. Mmasihii wa Pwani ' Christian of the Coast '
6. Mwafrika wa Pwani 'African of the Coast'
7. Mkristo wa Bara ' Up-Country Christian'
8. Mwafrika wa Bara ' Up-Country African'

' A g 11 3 B C

Muslim + + + + -
Coastal origin + + - -
Muslim name + + - -3
Christian name + - + 4
Koranic school + - + - 5

Distinctive Feature Matrix: Swahili

(blanks represent redundancy features)

This feature matrix may be illustrated in the following diagram:

Swahili

These labels are only labels. Any East African hearing them in the context of
Swahili would understand them as designating within broad limits a particular feature
matrix. The terms are all used and understood to mean what their make-up indicates.
Thus, as pointed out earlier, there are Christians, Arabs, and Africans who, in no
I Mmasihii is the Swahili word for ' Christian' word in Swahili for ' Christian' and is more com-
and is more common on the Coast. Mkrisfo is a loan monly used by Up-Country or non-native speakers.
Q

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
232 WHO ARE THE WASWAHILI?
way, could also be classed as Swahili. Often to disting
Arab, he would be called a Swahili while a Coastal
Swahili to distinguish him from a Coastal non-Christia
Up-Country African, he tends to think of all Coastal p
with no sharper divisions. If he is a Coastal Christian,
people as Swahili. But, an analysis such as this contras
validity for the researcher. Prins stated that, '... the
an epithet of reference and hence an important sociolo
ever used for self-identification '. This is perhaps becau
always going to be a different kind of Swahili from onese
one or a few features of contrast.
The following is an analysis of the term Swahili person. It assumes only that by the
general term Swahili is meant one who speaks the kiSwahili language and is a native
of Kenya or Tanzania. Categories 7 and 8 actually define people who are not commonly
thought of as Swahili. These categories are included because they present minimal
contrasts with the Coastal African and Coastal Christian who are thought of as
Swahili, for in many instances these people think of themselves as Swahili. These last
two categories refer to the Up-Country Christian and African Kenyan or Tanzanian
citizen who uses Swahili.
From the matrix and tree analysis, it would appear that a person designated
MSwahili is one who is a Muslim, but who has a non-Muslim name (i.e. a Bantu one
instead) and who did not receive Islamic schooling: he differs from an Mswahili
Safi in that the latter received training at a Koranic School, from a Mwarabu who has
a Muslim name, and from a Mwarabu Safi in both name and education. All four are
Swahili by religion. This distinguishes them from the other four who only have
language and nationality in common with this group but who are still Swahili in
that respect.
Two examples of the definition of a Swahili follow, which were supplied to me by
a native Zanzibar woman of Comoro Island ancestry and by a man from Up-Country
Tanzania. Both regarded themselves as Swahili. These indicate that the preceding
analysis is in some ways too specific and in others too general, and that a number of
the indices of contrast may need reworking. They show that a working definition
such as that derived from my analysis cannot apply uniformly but serves only as an
indicator of the range of applicability of the term.

SUMMARY OF STATEMENT OF ZANZIBAR WOMAN

The people referred to as WaSwahili are not so called on a linguistic basis alone.
The term MSwahili (pl. WaSwahili) varies with the time and place of reference apart
from the individual using the term. In Zanzibar, for instance, at one time an MSwahili
was anyone whose father and mother were Africans regardless of their birthplace.
The term Mwafrika was rarely used until the early nineteen-fifties.
An MSwahili would be anyone answering to one of these combinations:
i. Both parents Africans born in Zanzibar or Pemba or neighbouring coastal
areas (Mombasa, Lamu, Tanga, Mafia).

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
 WHO ARE THE WASWAHILI? 233
2. Father African, mother Indian, Arab, or other, regardless
language spoken at home was Arabic (a very rare case) or an I
and regardless of whether or not the person's physical feature

Any African born on Zanzibar (even though the parents may hav
mainland) speaking Swahili and another language in addition wou
as MSwahili by an Arab or Indian or Up-Country African.
An African whose father was an indigenous African and whose m
of the Comoro Islands is still referred to as MSwahili. On the oth
father or both parents came from the Comoro Islands, the child i
even if he and his siblings were born on Zanzibar. Even descenda
came from the Comoros five or six generations back are not regar
people.
In contrast, the inhabitants of the East African Coast known as the Shirazi-
descendants of the Persians who intermarried with Zanzibar natives-are regarded
as true WaSwahili. They speak only Swahili and live mainly in villages north of
Tumbatu island and on the island itself.
An Arab who marries an African on Zanzibar, Lamu, or anywhere on the Coast of
East Africa will have his offspring regarded as Arabs by his clan. However, a female
Arab marrying an African will have her offspring regarded as WaSwahili.
A Comorian woman who marries an Arab has Arab offspring. A Comorian man
who marries an Arab woman will have Comorian offspring.
The question of religion, skin colour, features, given name, and mode of attire
seems trivial in describing an MSwahili on the Coast-but, to an African from the
mainland, these are the criteria used.
Factors such as the ability to read the Koran, or to speak grammatical Arabic
(which few East African Arabs can do) or the degree of adherence to Arabic culture
are poor criteria for defining WaSwahili.

SUMMARY OF STATEMENT BY SUKUMA MAN

There are five main divisions of the people who fall into the category of
WaSwahili;

I. WaSwahili wa Pwani-the Swahili People of the Coast

2. WaSwahili wa Unguja-the Swahili People of Zanzibar
3. WaSwahili wa Bara-the Swahili People of Up-Country
4. WaSwahili wa MaShamba-the Swahili People of the ' Suburbs '
5. WaSwahili wa Tamaduni-Swahili Scholars

A Swahili person speaks Swahili and no other vernacular on most occasions. Today
there are Muslim and non-Muslim Swahili. Historically, the Swahilis are descendants
of Arabs, Persians, and Chinese who intermarried with Africans dwelling within
twenty miles of the shore all along the East African Coast. As a result of the ivory
trade, Swahili began to move to such places as Tabora, Ujiji, and Kigoma. Most
Up-Country Swahili today are still found near the commercial centres.

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
234 WHO ARE THE WASWAHILI?
The Swahili people of the Coast may be either Muslim
are in contact with outside influences (i.e. the mass me
cultures, etc.). These people may be further divided into
(a) The Swahili people from Tanga who are'... pretty c
there and there is a tradition for people from distan
go to Tanga to train to be Sheikhs and the like.'
(b) The Swahili people of Mvita who ' speak a strange ty
less cultured than those of Tanga '.
(c) The Swahili people of Mombasa. These Swahilis are
Swahilis to the Swahilis of Zanzibar and are, thus, t
Coastal groups.

The Swahili people of Zanzibar are mostly Muslim.

than the Coastal Swahili because of little outside influe
Swahili.
The Swahili people of Up-Country may be divided int

(a) The Swahili people of Tabora who speak Swahili wit

accent.

(b) The Swahili people of Ujiji and Kigoma who speak Swahili with th
accents.

The Swahili people of the suburbs (lit. the fields or plantations) are t
who live in areas around urban centres. They are usually new arrivals in
and do not know the prevalent language. They are people in transition w
ungrammatical form of Swahili and they are generally non-Muslim.
The Swahili scholars comprise those people who live anywhere in East
being well-trained Islamic theologians and Muslims.

DISCUSSION AND CONCLUSION

These two statements, as may easily be seen, contradict partly what is

in this paper. Therein lies the crux of the problem. The statement by
girl brings out the fact that whether a child is Swahili or not depends up
father is. Here, too, we see that there are groups of people, such as the
to whom none of the labels in my analysis directly apply. They are no
or African than Swahili-they are Comorians. Culturally, however, they
analysis more as Swahili than as non-Swahili. The argument that the pr
absence of a Koranic School education is not a valid criterion for definin
is a good one. It is valued in so far as education of any kind is inappropr
ferentiating peoples. It is valid, too, in so far as Koranic schooling exclu
young and the female-many of whom indeed are WaSwahili. However,
to diffoentiate the 'pure' Arab and the ' pure' Swahili from Arabs and
general, it seems useful. A wife or child of a Mwarabu Safi would also be Mwa
The point that criteria such as religion, skin colour, features, given n
type of dress are used by mainland Africans to distinguish coastal peopl
Swahilis, or Africans, is a salient one. The classification of the WaSwahili
Sukuma man is a case in point.

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
 WHO ARE THE WASWAHILI? 235
His classification, too, is very broad, including all those wh
often than any other vernacular. This classification includes
(Category 7, incidentally, would include this particular man
A comparison of his classification with my categories break
I. Mwarabu Safi
2. Mwarabu WaSwahili wa Pwani
3. MSwahili Safi WaSwahili wa Unguja
4. MSwahili WaSwahili wa Tamaduni
5. Mmasihii wa Pwani
6. Mwafrika wa Pwani
7. MKristo wa Bara WaSwahili wa Bara
8. Mwafrika wa Bara WaSwahili wa Mashamba

It must be pointed out that Coastal origin or contact somewhere in the p

to be an integral part in defining the WaSwahili. It is interesting that the
girl's statement assumes coastal or Zanzibar origin regardless of where th
father or mother came from while the Sukuma man takes into account m
albeit from coastal areas.
The question, 'Who are the WaSwahili? ' is by no means answered here. Perhaps
a more extensive study setting out to deal with this particular question which would
consider various linguistic criteria as well as these attitudinal ones of ' self' and
'others ' would be fruitful.

REFERENCES

EASTMAN, CAROL M., and FAROUK, M. T. TOPAN. I966. ' The Siu: Notes on the People and Their
Swahili, vol. 36/2, pp. 22-48.
HARRIES, LYNDON. 1962. Swahili Poetry. Oxford: Clarendon press.
POLOME, EDGAR C. I967. Swahili Language Handbook, Center for Applied Linguistics.
PRINS, A. H. J. I96I. The Swahili-Speaking Peoples of Zanzibar and The East African Coast. Dar
London: International African Institute (Ethnographic Survey of Africa Part XII).
STIGAND, C. H. I913. The Land of Zinj. London.

Resume

QUI SONT LES WASWAHILI?

A CETTE question difficile, il est couramment repondu par une definition de Sti
Swahili descendent d'un des premiers immigrants arabes ou persans etablis sur l
l'Afrique de l'Est, et comprennent, au sens le plus large, tous ceux qui parlent Sw
Cet article traite de nombreux autres points particuliers a considerer avant de
une definition. Bien que leur origine c6tiere et la connaissance de la langue Swah
des facteurs determinants, cela n'implique pas que ceux qui connaissent ou utilise
langue soient consideres eux-memes comme des Swahili. L'auteur a utilise le
suivants permettant de reconnaitre un veritable Swahili: (i) religion; (2) ori
graphique; (3) appellation; (4) systeme d'education. Selon ces criteres on peut
groupes en 8 categories: (i) Arabe authentique; (2) Arabe; (3) Swahili authe
Swahili; (5) Chretiens de la c6te; (6) Africain de la cote; (7) Chretiens apatrides; (
cains apatrides.

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms
236 WHO ARE THE WASWAHILI?

II y a cependant quelques groupes, tels les habitants des iles Comores, auxque
de ces descriptions ne peut s'appliquer; ils ne sont pas plus arabes ou africains
ils sont Comoriens, bien que culturellement, ils soient proches des Swahili.
On peut dire que tous les Swahili sont a quelque degre pros des locuteurs Sw
Swahili ne peut etre en meme temps asiatique ou europeen; aucun n'est primit
Swahili. Les deux categories extremes sont constituees par les Swahili c6tiers
arabes, et les Swahili africains non-musulmans apatrides. Sont consideres
comme Swahili tous les groupes originaires du Kenya ou de Tanzanie. Les
musulmans ont tendance a considerer les Swahili musulmans comme des Arab
que les Swahili musulmans les appellent 'Africains '. Depuis l'independance
tendance croissante pour tous les habitants du Kenya et de Tanzanie a se consi
Africains. En meme temps, grace a la diffusion progressive de la litteratu
manifeste une certaine fierte de voir la langue Swahili reconnue comme langu
il semble que les positions extremes soient en train de se confondre. Aujourd'
'Africain ' est rarement utilise dans un sens pejoratif.

NOTES FOR CONTRIBUTORS TO AFRICA

CONTRIBUTIONS should be addressed to the Editor. Articles should not excee

including footnotes and references. Longer papers can be accepted in excep
stances if a subsidy can be provided to meet the cost of printing additiona
Articles should be typed in double spacing, on one side of the paper only
margins to allow for editorial marking. Clean copy ready for printing sho
Authors are advised to keep a copy of their texts as the return cannot be gu
in African languages should be underlined and special characters kept
Footnotes may be placed at the bottom of the page on which they occur o
the end. Bibliographical references should be cited only briefly (i.e. author a
text or in footnotes, and an alphabetical list of references given in full at
article. Where appropriate the text should be divided by suitable headings an
Tables outside the text should be typed on separate pages, numbered co
given headings. Maps and other line-drawings should be submitted in final
for size) on good-quality paper.
Contributors are asked to provide a summary of their articles of 250
translation into French or English.
Galley proofs are submitted to authors but only essential corrections can
additional matter cannot be inserted at proof stage. The Editor reserves th
any corrections or alterations he may deem necessary.
Contributors receive 25 free offprints of their articles; extra copies can be
price on request, if ordered before publication.

This content downloaded from

197.221.210.239 on Tue, 01 Jun 2021 06:45:15 UTC
All use subject to https://about.jstor.org/terms