Saturday, 2 October 2010


Ernest Hemingway, Serengeti Plain, January 1934 (JFK Library)
For many years I've wanted to write something about 'bad Swahili', the mangled snippets of the language that appear with depresssing frequency in the works of otherwise literate European and American writers striving to add local spice to their fictional and factional accounts of life in East Africa. My principal inspiration was Ernest Hemingway's Green Hills of Africa, his lightly fictionalised account of a hunting safari in northern Tanganyika in early 1934. When I first read this book in Mombasa I was deeply unimpressed by its subject matter as well as the smattering of misspelled and ungrammatical Swahili that Hemingway used to colour his tale. Rather like the self-indulgent travel diaries of contemporary overlanders, it dwells too long on personal experiences and relationships, and caricatures Africans and their own relationship to the environment in which it is set. Instead my sympathies lie with the Austrian plantation manager 'Kandisky', who describes Hemingway's hunting as "this silliness of kudu" (2004a: 17) and invites him to "take a safari to study the natives" (2004a: 21). The real-life 'Kandisky', Hans Koritschoner, followed his own advice, and became the Tanganyika government sociologist Hans Cory, though commentators on Hemingway's book have generally missed this connection (e.g. Ondaatje 2003: 124-125).*

Researching a recent paper (see my post on Heritage, tourism, and slavery at Shimoni) brought me back to Hemingway, and a month ago I picked up a copy of Christopher Ondaatje's Hemingway in Africa (2003). Ondaatje's own mistranscribed Swahili ("Jumbo, habari, poleya safari") reminded me of Hemingway's, and I then bought new copies of the books that I'd left behind in Mombasa so that I could have another look. There certainly is some bad Swahili in Green Hills of Africa. Here are some of the misspelled words and phrases (all converted to lower case and italicised), with correct forms and glosses shown in parenthesis: b'wana (bwana, master), b'wana m'kumba (bwana mkubwa, big boss), doumi (dume, male), faro (faru, rhinoceros), manamouki (mwanamke, woman), m'uzuri (mzuri, good), n'dio (ndiyo, yes, it is), tarahalla (palahala, sable antelope), tendalla (tandala, kudu). Hemingway refers a number of times to his use of a dictionary when trying to communicate with Africans (2004a: 96, 115, 122, 129, 155, 156, 163, 165). This was probably A. C. Madan's English-Swahili Dictionary (first published in 1884), which is known to have been in Hemingway's library in 1941. Indeed the library list suggests that he may have owned two copies of it, or perhaps both the English-Swahili and Swahili-English volumes, the latter published in 1903 (Brasch and Sigman 2000: 238, No. 4147). Madan was a reliable source, and his work provided the foundation for Frederick Johnson's A Standard Swahili-English Dictionary (1939), which Hemingway obtained later, together with other works on Swahili. If Hemingway had used Madan then he shouldn't have made so many mistakes. But on his own account he couldn't always find the words that he was looking for, and he may well have misheard some of the terms that his hunting companions used, including the place names and other proper nouns that he also gets wrong. The insertion of an apostrophe to denote syllabic nasals in word-initial position (m'- and n'-) was once a relatively common practice, and Hemingway may well have picked this up from other sources. In his later story 'The Snows of Kilimanjaro' (2004b) it's notable that he (or his editor) gets the transcription of Bwana right.

Hemingway's hunting party, February 1934 (JFK Library)
So much for Hemingway's mistakes. More interesting, perhaps, is the evidence that Hemingway was reflecting contemporary Swahili usage, and in particular the pidginised variety of Swahili sometimes referred to as Kiset(t)la, the simplified idiom used by European settlers to communicate with their African employees, and vice versa. In his study of Kisetla, linguist Anthony Vitale declared that "The African novels of Hemingway (e.g., The Green Hills of Africa) and Robert Ruark (e.g., Uhuru, Something of Value) abound with utterances in KiSetla" (1980: 65, fn. 9). In Hemingway's case this is something of an exaggeration - most single Swahili words that he uses can't be construed as belonging to any specific idiom - but there are some words and phrases in his text that can be identified as typical Kisetla forms. These include lack of proper noun class prefixes and agreements (Wakamba instead of Kikamba, the Kamba language; manamouki kubwa sana instead of mwanamke mkubwa sana, or preferably jike kubwa sana, very large female), the use of hapana as a generalised negative (hapana m'uzuri for si nzuri, it's no good), and use of the simple imperative as a generalised verb form (piga for amepigwa, s/he's hit; piga m'uzuri for mpigie vizuri, hit him/her properly). These words and phrases are put in the mouths of both the European narrator and his African companions, suggesting a shared usage. Indeed Hemingway was well aware that he was using a pidginised Swahili, and makes this explicit when he writes: "I got the dictionary out of my pocket and made a sentence in pigeon Swahili" (2004a: 96), 'pigeon' being a variant of the now more common linguistic term 'pidgin'.

Hemingway's first East African safari took place in the decade that the existence of Kisetla and related Swahili pidgins in Kenya was first brought to general attention. A translation of the Gospel of Mark into 'Kitchen Swahili' was published 1931. A humorous article on 'Kisettla' ("by J. W.") was printed in the East African Standard in 1932 and later circulated as an illustrated pamphlet. In 1933 a language teaching manual was written for the King's African Rifles, using the soldiers' own idiom which is referred to variously as KiKAR, Kikeya, and Kivita (Newell 1933; Mutonya and Parsons 2004). The first of many editions of F. H. Le Breton's Up-Country Swahili Exercises for the Soldier, Settler, Miner, Merchant, and their Wives was published in 1936. Although it didn't pretend to describe Kisetla, it did reproduce many of its best-known features, including those mentioned above and others, such as the the generalised use of mingi to mean 'many'. This was no doubt the kind of Swahili spoken between the Kenya-based Europeans and Africans in the Hemingways' party, as well as the up-country Tanganyikans that they came across who were also able to converse in the pidginised lingua franca. Snatches of it embellish films made in East Africa from this time onwards, including the Hollywood movies inspired by Hemingway's own work (Carrier 2010). And despite the inroads made by modern education and the media, something like it can still be heard in up-country Kenya and in particular settings elsewhere in East Africa. It is used, for example, by some Asian shop-keepers when addressing their African staff and customers, and is frequently caricatured. Even Christopher Ondaatje slips into pidgin practice when he reports saying "Pole kusumbua, wewe", translated as "Sorry for causing you trouble" (2003: 108). The meagre literature on Swahili pidgins refers to different varieties, including Kihindi, Kishamba, and others already mentioned above. But these remain largely unresearched, and it may be that they are best thought of as a continuum of forms with a common core. And I haven't touched on Hemingway's second trip and later writings about Africa. Hii ni kazi mingi sana, mimi hapana taka fanya sasa.

* In his family history, written in 1956, Hans Cory referred to his encounter with the Hemingways and his appearance in Green Hills of Africa: "I am Kandinsky [sic], and though the conversation did not take place exactly as quoted, the events happened as described, and the breakdown of my lorry, etc. is true. Hemingway and his wife were very kind to me. I was their guest for three days, and we had many amusing and interesting conversations."

My thanks to Helle Goldman and Ray Abrahams for sharing their own thoughts and experiences of bad and pidgin Swahili with me, and to Neil Carrier for the additional inspiration provided by his recent workshop paper and presentation on the use and abuse of Swahili in Hollywood movies. The usual disclaimer applies.


Brasch, James D. and Joseph Sigman 2000 [1981]. Hemingway's Library: A Composite Record (electronic edition). Boston: John F. Kennedy Library.

Carrier, Neil 2010. Kiswahili Hollywood style: linguistic use and abuse in the movies. Paper presented to 
the VIII European Swahili Workshop, Contemporary Issues in Swahili Ethnography, University of Oxford, 19-21 September 2010.

Cory [Koritschoner], Hans 1956. Our Family Chronicles. Online in the archive of The Institute on World War II and the Human Experience, Florida State University.

Hemingway, Ernest 2004a [1935]. Green Hills of Africa. London: Arrow Books.

Hemingway, Ernest 2004b [1936]. The Snows of Kilimanjaro. In Hemingway, E. The Snows of Kilimanjaro. London: Arrow Books.

Le Breton, F. H. 1951 [1936]. Up-Country Swahili Exercises for the Soldier, Settler, Miner, Merchant, and their Wives and for all who deal with up-country natives without interpreters (11th edition). Richmond, Surrey: R. W. Simpson and Co.

Madan, Arthur Cornwallis 1884. English-Swahili Dictionary. Oxford.

Madan, Arthur Cornwallis 1903. Swahili-English Dictionary. Oxford.

Mutonya, Mungai and Timothy H. Parsons 2004. KiKAR: a Swahili variety in Kenya's colonial army. Journal of African Languages and Linguistics 25: 111-125.

Newell, H. W. 1933. Notes on Ki-Swahili as Spoken by the K.A.R.. Manuscript in the Kenya National Archives, Nairobi. [cited in Mutonya and Parsons 2004: 125]

Ondaatje, Christopher 2003. Hemingway in Africa: The Last Safari. Toronto: HarperCollins.

Vitale, Anthony J. 1980. Kisetla: linguistic and sociolinguistic aspects of a pidgin Swahili of Kenya. Anthropological Linguistics 22 (2): 47-65.

Wilkes, Hamilton Paget (translator) 1931. Habari Njema: kama aliandikwa kwa mkona [sic] ya Mariko. London: British and Foreign Bible Society.

W., J. Undated [1932.] Kisettla. (Pamphlet.) [Nairobi:] The East African Standard. [some sources give the reference as The Kenya Weekly News, 23 December 1955, 24-25.]


  1. Really interesting article, once again, Martin. Kisettla lives on in language I hear every day at the Karen dukas (I mean, maduka ya karen)to me - and I admit I was taught 'classical Swahili' by a teacher: I didn't acquire it myself - it sounds funny and deprecating, but some non-Swahili mothertongue Africans do speak 'kitchen Swahili' too.

  2. I agree, this was very interesting! Now, a question: after having reread Green Hills of Africa, did your opinion of it change, in terms of its literary value? Is it a book you would recommend or not?

  3. Living in Kenya 1998 - 2005 I found this "Kisettla" quite frequently - used by Kenyans among themselves and also communicating with me. Surely not because I am Mzungu, I speak relatively clean Standard Swahili, Tanzanian style. My impression was that it is a widespread colloquial Swahili used among Kenyans with less formal education. Amongst people who passed Secondary school, the level of Swahili has become much better. But even here I got recently a lot of "mingi" from University graduates to whom Iinsisted to talk Swahili.