The Problem of Writing About the Logoori Language

David Odden

This note reports on linguistic research which we have done on the Logoori language. It is a prelude to a descriptive grammar being prepared, and is intended for speakers of Logoori with no training in linguistics. The grammar has / will have many examples from the language, and a lot will be said about pronunciation of words, but without some background in how linguists write about pronunciation details, the facts being talked about won’t make much sense. We hope that this note will make speakers of Logoori aware of some features of the language that they may not have known of previously. In this note we will give example words, and if you are reading the file on a computer or phone, you should be able to hear the recordings of words that we provide. When you see a “play” button after a word or phrase, if you click / press it, the word should play (you may need to tell your computer to “allow” your sound player to play the file). Computer technology does not always work, so it is possible that you can’t hear the recordings.

           Our goal in the project is to accurately report the facts of the Logoori language. The first problem for accurate reporting is, how do we represent actual pronunciation in a written form? There is no single standard for writing Logoori, even though certain practices are often followed. (We should mention that the name of the language is an extreme example of the challenge of writing. There is the fact that the language is often called “Maragoli”. You will encounter various spellings such as Logori, Logooli, Llogori, Logoori.) We cannot rely on any existing writing scheme, because all schemes leave out scientifically-important facts of pronunciation about the language. There are three major features of pronunciation that are important in the language (those aspects are the main purpose of this note which we discuss first), and two other simpler problems discussed at the end. These facts of pronunciation are not consistently noted in writing. To understand the claims that we are making, we encourage speakers of the language to pronounce the words that we provide and think about their meanings, to see how pronunciation affects meaning. We also encourage you to compare your own pronunciation with the recordings contained here. 

           As you will discover from this note and the full grammar (and are no doubt aware, based on your experience with how people talk), pronunciations vary quite a lot in the language. The pronunciations in this version come from recordings of Editon Mulera, from Chanzeywe in South Maragoli. We are reporting these facts for scientific reasons. We do not assume anything about how words should be written in ordinary writing, and we make no claims about how words should be pronounced, we are just reporting how they are pronounced, by some people in some places. There are some advantages to having a unified spelling, where the word ‘day’ is spelled in just one way: whether that should be ilidiku, lidiku, iridiku, iddiku or something else is a very difficult decision. These spellings relate to actual differences in pronunciation.

Long Vowels

The first such problematic fact of pronunciation is about the length of vowels. The word kondora ‘to pick me up’ is pronounced differently from kondora ‘to see me’; ma vahere ‘they will inhale’ is pronounced differently from ma vahere ‘they will come to an end’; kokera ‘to age (of women)’ is different from kokera ‘to milk’. The difference in pronunciation is that in one case, the underlined vowel is short, and in the other case the vowel is long. This difference is recognised in many standard spellings of the language, though this practice is not always observed. These words would be written as koondora ‘to see me’, koondoora ‘to pick me up’, ma vahere ‘they will come to an end’, ma vaheere ‘they will inhale’, kokera ‘to milk’, kokeera ‘to age’. Similar examples are kuragira ‘to promise for’, kuraagira ‘to eat ugali’, kutura ‘to leave’, kutuura ‘to unload’, kuvura ‘to fail or lack’, kuvuura ‘to reveal’ and yaara ‘sue!’, yara ‘spread the bed!’.

It has been somewhat traditional to write words like kutanga ‘to start’ , kogenda ‘to walk’ with a single vowel (in the 1996 New Testament translation, these and similar vowels are often written single), but these vowels are actually pronounced long. This fact is recognised in the dictionary of Joseph Ndanyi. The reason for this inconsistency is that there is a rule of the language which makes vowels long before combinations of m, n plus other consonant b, d, j, g, z in some contexts. Writing systems tend not to include pronunciation facts that you can figure out if you know the rule. But that rule is very complicated and does not apply everywhere, for example, it does not apply in izingugi ‘baboons’, nor does it apply in nzumbakaa ‘I am building’. The prefix rwa meaning ‘when’ has a long vowel in some cases (rwaandeekaa ‘when I cook’, rwaandeechi ‘when I cooked (recent)’), but a short vowel in other cases (rwandaadeeka ‘when I cooked (long ago)’ rwangedeekaa ‘when I am still cooking’). And, the fact that there is a rule does not mean that people know the rule, so from a practical perspective, it may be best to write long vowels as such, and not depend on a rule. Since this is a scientific work, the scientific readership will absolutely not know that rule. So if a simple rule cannot be given about pronunciation, our system of writing absolutely needs to include that information. We write double vowels whenever a vowel is pronounced long, and we will write single vowels whever a vowel is pronounced long, even if there is a way that the difference could be predicted.


Another important fact about the language, which has never been indicated in any spelling, is that Logoori is a tone language, meaning that voice pitch is part of the makeup of a word, just like the consonants and vowels are. Some examples of pairs of words having different tones are as follows:

kwiizuriza ‘to remember’; ‘to fill something’
vaamiga ‘they strangled’ ; ‘they strangled me’
kuviimba ‘to put on a roof’; ‘to swell’

When you say kwiizuriza with the meaning ‘to fill something’, it is pronounced differently from kwiizuriza meaning ‘to remember’.

           Another example of a tonal difference is yaakagura, which with one tone means ‘he has bought’   (if you are making an announcement that he has finally bought some thing), and with another tone it means ‘he bought’  (typically as an answer to a question like ‘where did this thing come from?’). One last example is umuundu yaavariza, where the word umuundu is pronounced differently, depending on whether the sentence means ‘the person counted’ or ‘the person who counted’ .

When we present examples of Logoori, we add accents to vowels to indicate those places where the voice goes up, which marks the difference between these words (ignore the exclamation mark for the moment)


‘to remember’


‘to fill’


‘he has bought’


‘he bought’

umúúndú yááváriza  

‘the person counted’

umúúndu yááváriza  

‘the person who counted’

In kwiizʊriza ‘to remember’, the voice remains low through the word, but in kwíízuriza ‘to fill something’, it is high on the first syllable then drops down after that. We say that the first syllable of ‘to fill something’ has a high tone, and the other syllables have low tone. All syllables in ‘to remember’ have low tone. We have the same pattern in yaakagura ‘he bought’ (all syllables are low toned) and yáákagura ‘he has bought’ (only the first syllable is high toned).

           In both váámíga ‘they strangled’ and váámiga   ‘they strangled me’, the voice starts high, but in ‘they strangled me’, it remains high through the second syllable , but it drops down on mi in váámiga ‘they strangled me’. In kuviimba  ‘to put on a roof’ the voice stays low throughout the word, and in kuvíimba ‘to swell’ it goes up on the second syllable víi, then falls back down.

           The tone pronunciation of a word is affected by what kind of sentence the word appears in. The word umuundu ‘person’, when said alone, has only low tones, and no high tones. In the sentence umúúndú yááváriza ‘the person counted’, the voice goes up on the second or sometimes first syllable and stays up until the syllable va. In umúúndu yááváriza ‘the person who counted’, the voice goes up for múú, goes back down for ndu, and then back up for yáá.

           The following words have a raised tone on the second from last vowel of the word, as indicated with the accent mark.

































The last two words illustrate another fact about Logoori pronunciation, that some words have more than one pronunciation. There is a rule explaining why you can say both iridíku and iddíku, which we give in the grammar

           In the following words, the voice remains low throughout, and no vowel has an accent mark.



































In these words, the voice is raised on the second to last vowel of the word.





















Compare the following words, where the voice is raised on the third to last vowel of the word.


‘bean leaves’


‘animal blood’


‘potash sieve


‘flat land’


‘roof peak’


‘leaf bundle’


‘culvert pipe’






‘bird breastbone’



The tonal pronunciations of words is not absolutely rigid, and there are some general rules about other pronunciations that are possible. For example, the words indurúme, iganísa may also tend to be pronounced as indúrúme, igánísa, with raising of the voice on both the second and third from last syllables. Similarly, ekeségese and amagáraba may also be pronounced as ekéségese and amágáraba with raising of the voice on the third and fourth syllables from the end. The pronunciations indúrúme, igánísa, ekéségese and amágáraba may be more common than indurúme, iganísa, ekeségese and amagáraba, at least for some people. The grammar will explain the rules which make all of these pronunciations possible.

           One rule which pervades the language is that when a word has a high tone, the vowels before it may also have the same high tone. The word umugera ‘river’ has no high tone when said alone, and it also has no high tone if combined with certain other words.

umuger(a) umwaangu

‘quick river’

umugera mulla

‘1 river’

umugera vuza

‘only a river’

Actually, in the pronunciation of the first example, the phrase is pronounced umuger-umwaangu, eliding the first vowel at the end of the word umugera. This is another aspect of Logoori pronunciation that is explained in the grammar. In this note, the final vowel will be written so that you will recognize the word, but the elided vowel is put in parentheses in order to indicate that it is not actually said in the recording.

           But when other words (ones which have a high tone) follow this same noun, the pronunciation of the noun changes.

umúgér(á) úmútáámbi

‘long river’

umúgér(á) úmúzíllu

‘cold river’

Compare the pronunciation of ‘long river’ and ‘quick river’: you should be able to hear that in ‘long river’ the voice goes up higher and stays up until the last syllable, but in ‘quick river’ the voice is not as high, and it slowly drops down throughout the utterance.

           One final feature of tonal pronunciation needs to be pointed out. In umúgér(á) úmútáámbi and umúgér(á) úmúzíllu, the tone is constantly raised within the noun umugera, to the final syllable of the next word. This is different from the pronunciation of ekedet(e) ikitaambi ‘long finger’: the voice goes up on de and possibly ke, then drops a little, continues at a less-high level on the syllables te ki taa, then drops down to a low level.



When high tones have such a shift downwards, we mark that with a little raised exclamation point: ekédé!t(é) íkítáámbi. A similar example is the word uvuchíma ‘ugali’, when combined with the adjective vuráhi ‘good’: we may hear uvuchí!m(á) úvúráhi .

           Some nouns have more than one high tone in them. For example, in these nouns, the first syllable has a high tone, then the tone drops part-way down.


‘tobacco plant’



In these nouns, the first high tone is on the second and possibly first syllable, and a second lowered high is on the last two syllables.




‘ant type’









Here are more examples, where the word begins with some low toned syllables then has a high toned syllable, next a drop in pitch, and finally high toned syllables all the way to the end of the word.




‘tent canvas ’


‘a mushroom’


‘bird (whydah)’

Notice the difference in pronounciation between these pairs of nouns.








‘roof peak’


‘clay paste’


‘bean leaves’

In both sets of words, the tone becomes high towards the beginning of the word and then drops down. In the words on the right, the tone drops further down than it does in the words on the right. The same pattern exists in these pairs of words.






‘tent canvas’




‘a mushroom’




‘bird (whydah)’



You may find yourself confused about our claims if you don’t agree with these pronunciations. The fact is that there is a lot of variation in how words are pronounced, so we have observed pronunciations for ‘rabbit’ including kɪsʊ́ʊ́!ngʊ́ra, kɪ́!sʊ́ʊ́ngʊ́ra, kɪ́sʊ́ʊ́!ngʊ́rá, and kɪsʊ́ʊngʊ́ra. As far as we can tell, the word mrógoori always has this tone, and not *mrógóori; ‘people’ may be vaandʊ or avaandʊ, but never *váandʊ or *aváandʊ. By including a number of examles, we hope you will at least agree with the majority of pronunciations that we provide. Our point is that pronunciations of words are not completely fixed across the language.

Vowel Qualities

In ordinary writing, Logoori uses the 5 vowel letters a e i o u which have letters in the Latin alphabet. The problem is that there are 7 vowels in the language: the letters i and u stand for a total of 4 different vowels. Linguistics has special symbols to make that difference clear: we can write [i] or [ɪ], [u] or [ʊ]. To show that there are 4 vowels for 2 ordinary letters, we will look at some forms of the verb, used to give orders. There is a difference between how you would say ‘ask us!’ if you are speaking to one person, versus to a group of people. If you say it to one person, you say koteevé, and if you say it to a group of people, you say koteeví. In the examples, we translate this difference as ‘(plural)’.


‘ask us!’


‘(plural) ask us!


‘speak ill of us!’


‘(plural) speak ill of us!’


‘scrape it!’


‘(plural) scrape it!’


‘arrange them!’


‘(plural) arrange them!’

You can see that in these particular examples, the form for speaking to one person ends in e, and the form for speaking to more than one person ends in i. There are some other changes in pronunciation that go with that difference -e versus -i. When a verb stem ends with y, it changes pronunciation before i so that it becomes very “light”, in fact may not even be there at all. In addition, the consonants k and g change to ch and j (some people may not do this – this is another area of variable pronunciation in the language).

In the form for addressing one person, the last vowel actually changes: it may also be i, depending on what vowel comes before it. The rule is that when the preceding vowel is u or i, the final vowel of the singular is i, and when the preceding vowel is one of a, e, o, the final vowel is e. (There are some complications, where you can get i after a, e, o which has to do with the consonant between the two vowels – we talk about this in the grammar). You might expect that ‘bite him!’ is pronounced the same as ‘(plural) bite him!’, but actually the singular and plural forms of the orders are pronounced differently: there are two kinds of i in the language, as well as two kinds of u. The singular order has the vowel which we write as [ɪ], and the plural order has the vowel which we write as [i].


‘bite him!’


‘(plural) bite him!’


‘buy it!’


‘(plural) buy it!’


‘cross it!’


‘(plural) cross it!’


‘hold it!’


‘(plural) hold it!’

There are very few situations where the only difference between words is the choice of i versus ɪ, or u versus ʊ – there is usually some other difference in tone or consonants that can be used to figure out which word is being said. For example, kʊrúga    ‘to cook ugali’ and kʊrʊ́ka    to vomit’ have the vowels u and ʊ, but the consonants k and g are also different. Tones are different as well as vowels, but you can compare the two kinds of vowels in the words ɪkɪ́rɪ́tʊ   ‘chest’ and ɪkɪritu    ‘heavy’. This difference in vowels figures into how a number of particular words are pronounced. The vowel that would be ordinarily spelled “i” in to fold’ is a different vowel from the one in ‘to protect’.



to fold


to bathe


to hide (something)



to protect


to watch


to watch



to be blunt


to overcome


to chase away


to snatch


to cover



to strain


to reach


to cover


to extract


to push


to pay

Unsolveable matters

There are some problems with our plan to write Logoori in a way that reflects pronunciation, which do not have a solution as far as we can see. One of those problems is the matter of “l” versus “r”. There is a generally rule that the consonant is pronounced “l” when it is doubled (mʊlla ‘1’ referring to ‘river; person’), but as “r” when it is single (ndara ‘1’ referring to ‘cow’). Even so, the “r” quality is different from “r” of English (in fact, in English, there is a lot of dialect difference in how “r” is pronounced). We have found that the r-like pronunciation of Logoori sounds like it is “mixed with l” depending on the context and the individual. For example, the l/r at the beginning of the word ‘brick’ may sound more like lidá!fáari or ridá!fáari, for some people. It usually does not sound like l in words like vara ‘those ones’. However, “l” does appear in some cases where we would not expect it, for example, the adjective kamíili will, for some speakers, be pronounsed with a clear single l. The reason for this is clear, that the word derives from Swahili kamili, and some speakers therefore pronounce it with clear l. Likewise, the word ‘Luo”, ʊmjálwo, is as far as we know pronounced with “l” and not “r”, even though the usual rule predicts it should sound like r.

    Since our goal is to describe how people do pronounce words (and not how we think people should pronounce words – we have no basis for making that judgment), this means that we ought to write “l” in those cases where a person clearly say “l” rather than “r”. Therefore, in those cases – in the grammar – where we notice clear “l”, we will present the example spelled with l. We may not always notice one of these unexpected  clear-l sounds. The point is that even though the difference between “r” and “l” is subject to a rule (single “r”, double “l”), single “l” is also entering the language.

    Another problem regards what might be spelled “ny”. A simple illustration of the problem is a comparison of kweenya ‘to want’ and yeenyí ‘he wanted’ – “ny” is usually pronounced very differently in those words. The tip of the tongue is flatter and further forwards in kweenya, but the back of the tongue is more advanced and raised in yeenyí. However, these two types of ny are not always pronounced differently, and we think that for some people there is no distinction. We have to resort to a special letter, in order to indicate which pronunciation is used in speech. In our writing, we will use the special letter ɲ to indicate the kind of sound in kweenya, and we use “ny” for the sound in yeenyí. Here are some examples of the two sounds (from a speakers who pronounces them differently):

    ɲasáye ‘God’    vaanya ‘mothers’    mɲó!róró ‘chain’

    ɪɲʊ́ʊ́mba , ɪnyʊ́ʊ́mba ‘house’

    kʊnyɪɪra ‘to stretch’    kʊɲára ‘to be able’

    kʊnyʊ́ʊnya ‘to give the breast’ koɲoora  ‘to get’

    komooɲa ‘to gossip’, komoonya ‘to cause to gossip’

    kʊɲááɲa ‘to eat’, kʊɲáánya ‘to cause to eat’, kʊnyáánya ‘to cause to eat’

    kʊ́ʊ́ɲɪmbɪra    to sing for me      (not pronounced *kʊ́ʊ́nyɪmbɪra)

    ɲʊmbákáa      I am building (not pronounced *nyumbakaa)

Again, there are rules related to pronunciation, but the rules are so complex that on practical grounds we have to write the two sounds differently. One rule is that before the vowel [i] (but not always before [ɪ]) you only get ny – this is why ɲ in kweeɲa changes to ny in the past tense yeenyi. Another rule is that ɲ meaning I; me’ remains ɲ before [ɪ] or [ʊ] in a verb, but other examples of ɲ may be pronounced as ny, for example ɪɲʊ́ʊ́mba or ɪnyʊ́ʊ́mba. A third rule is that n and ɲ at the end of a verb changes to ɲ on order to make the ‘make someone’ form, for example komooɲa ‘to gossip’, komoonya ‘to cause someone to gossip’.

A big puzzle to us is the word kʊniara ‘to wither’ which has an n-sound that is different from both ny and ɲ. Our best solution is to write this as ni: this is the only words that we know of with this sound.

Our Writing System and Existing Orthographies

There have been a number of spelling systems employed for Logoori over the past century, all of which have been confined to the standard Latin alphabet. Consequently, the pronunciation features discussed above have mostly been ignored. The existence of long vowels has been encoded in the spelling, indeed this is one feature of pronunciation that has been incorporated into most spelling practices such as Bible translations and school primers. However, as noted above in the discussion of vowel length, conventional spelling notates long vowels only in some of the contexts where they arise. Even though rogeendo "journey" has a long vowel, since it comes before a nasal plus consonant combination, it is usually not indicated in spelling. Let us compare parallel text in two Bible translations. Taking examples from two Logoori Bible translations (Matthew 1:18), we find divergence between vowel-length marking of the two texts where some long vowels are written as short (marked in bold)

Liivulwa lia Yeso Kristo liali kuli ndi.
Liivulwa lia Yeso Kristo liali kuli ndi.

Lwa Josefu yaali ni yakakwa Maria, mama wa Yeso, navuzwa nivakikiri                              
Lwa Josefu yaali ni yaakakwa Maria, mama weeve,                  ni vakikiri kokola keselero

kuva harara, kuli musaza nu mukali,
nukuva halala kuli musaza nu mukali,

yanyooleka naveeye ninda                       ya Roho Mutakatifu
yaanyoleka naaveeye ni inda ku zingulu zya Roho Mutakatifu

Our linguistically-oriented system would render these two versions as

ríívʊrwá rya yééso kríísto ryááli kʊ́rɪ́ !ndɪ́. rwá jósefʊ yáárɪ nɪ yáákakwá maryá, mááma wa yééso, navʊ́zwa  nivákɪkɪrɪ kʊvá harara, kʊ́rɪ́ ʊ́msááza nʊ́m̀kári.                yááɲóóreka náave nɪ́ɪnda ya róóho ʊmtakatɪ́ɪ́fʊ
ríívʊrwá rya yééso kríísto ryáári kʊ́rɪ́ !ndɪ́. rwá jósefʊ yáárɪ nɪ yáákakwá maryá, maama wé!évwé, nɪ vákɪkɪrɪ kókóra ekeseerero kʊvá harara kʊ́rɪ́ ʊ́msááza nʊ́m̀kári, yááɲóóreka náave nɪ́ɪnda kʊ zingúlu zya róóho ʊmtakatɪ́ɪ́fʊ

While these two published translations agree in writing musaza  “husband” with a short vowel, it is generally found in other written sources that the word has a long vowel: this is also our consistent finding working with speakers. The names Yeso Kristo, Josefu, Roho Mutakatifu are written with short vowels in the second-from-last syllable of the word, but they are pronounced with long vowels. Perhaps in this instance, an overriding factor is a goal of maintaining consistent spelling across Bible translations, indeed it seems that Logoori versions of words originally from other languages tend to be spelled according to Swahili conventions 

The inconsistency in writing pure Logoori yaakakwa ~ yakakwa or yanyooleka ~ yaanyoleka probably relates to the fact that there are two tonally-different verb forms that these texts could represent. There is the form [yááɲooreka]    which means "she is in the state of having received", and a different form [yááɲóóreka]    "she received (at least 3 days earlier)". Which form is most appropriate in this context? Without an explicit system for encoding tone (as we have offered), how are speakers most likely to read yaanyoleka versus yanyooleka? Likewise, [yaakákwá]    "he just paid dowry" and [yáá!kákwá]    "he has paid dowry" differ in meaning and tone, and it is likely that the spelling differences result from an attempt to say that one form would be more appropriate as the translation in this case, but it is hard to say just based on the spelling which pronunciation and therefore meaning is intended.

We want to point out a problem with excessive attention to details of pronunciation in writing systems. When the words kʊ́rɪ́ and ʊmsááza are combined, the vowel ɪ in kʊrɪ is deleted: one generally says kʊ́r-ʊ́msááza. Tones as pronounced in words said my themselves change when put in sentence context (see the previous discussion of  umúúndu yááváriza ‘the person who counted’ and umúúndú yááváriza ‘the person counted’). The practical problem is that if, as a speaker of Logoori, one is to produce linguistically-detailed examples such as we offer, you must carefully reflect on what the actual pronunciation of a word is in a particular context – this is frankly too complicated for a useful writing system to be adopted by non-specialist language speakers. This system of writing is thus a technical adjunct to ordinary spelling, comparable to dictionary notations that tell you how Americans or British people pronounce words like "Caesar", "economy", "leopard" or "flour".