Chichewa tones
Chichewa (a Bantu language of Central Africa, also known as Chewa, Nyanja, or Chinyanja) is the main language spoken in south and central Malawi, and to a lesser extent in Zambia, Mozambique and Zimbabwe. Like most other Bantu languages, it is tonal; that is to say, pitch patterns are an important part of the pronunciation of words. Thus, for example, the word chímanga (high-low-low) "maize" can easily be distinguished from chinangwá (low-low-high) "cassava" not only by its consonants but also by its pitch pattern.
Tones (the rises and falls in pitch) also play an important grammatical role in Chichewa verbs, helping to distinguish one tense from another. They also play a part in intonation and phrasing and in distinguishing statements from questions.
Most nouns have either one high tone (usually on one of the last three syllables) or none, although a few nouns (mostly compounds) have two tones. Verbs also usually have no more than one tone per morpheme. Chichewa thus in some respects can be considered to be a pitch-accent language with a "mixture of accentual and tonal properties".[1]
Some specialists, notably Larry Hyman, however, have argued that the term "pitch-accent language" is an over-simplification and should be avoided.[2] He considers it best to consider such languages simply as one variety of tonal languages.
Overview
Lexical tones
Nouns
Certain syllables in Chichewa words are associated with high pitch. Usually there is one high pitch per word or morpheme, but some words have no high tone. In nouns the high pitch is usually in one of the last three syllables:[3]
- chímanga 'maize'
- chikóndi 'love'
- chinangwá 'cassava'
In a few nouns (often compound words) there are two high tones. If these tones are separated by only one unaccented syllable, they usually join in a plateau of three high-toned syllables; that is, HLH becomes HHH:
- chizólowezí 'habit'
- bírímánkhwe 'chameleon'
- gálímoto 'car'
- tsábólá 'pepper'
In addition there are a large number of nouns which have no high tone, but which, even when focussed or emphasised, are pronounced with all the syllables low:
- chipatala 'hospital'
- mkaka 'milk'
Because there is usually only one high tone per morpheme, it can be considered as a kind of tonal accent (see Pitch-accent language).[4] It differs from a stress-accent such as in English in that it always retains the same pitch contour (e.g. high-low, never low-high). It is also possible for an accent to contrast with a lack of accent:
- mténgo 'tree'
- mtengo 'price'
Verbs
Verbal roots too can have a lexical tone, but only on the final vowel -a of the verb. Most verbs, however, are toneless:
- thandiza 'help'
- thamangá 'run'
The tones are not inherited from proto-Bantu, but appear to be an independent development in Chichewa.[5]
Monosyllabic verbs are always toneless:
- -dya 'eat'
Often a verb has a tone not because the root has one but because a stative or intensive suffix is added to it:
- dziwa 'know'
- dziwiká 'be known'
Grammatical tones
One of the main features of the Chichewa verb system is that each tense is associated with one of about eight different tonal patterns; there are different patterns again when the verb is in a relative clause or is negative.[6] Thus the present simple (near future) is accented on the initial syllable, the remote perfect on the tense-marker -na-, the present continuous on the syllable following the tense-marker -ku-, the present habitual on the initial and penultimate, the remote imperfect on the tense-marker -ká- and the penultimate, the subjunctive on the final, the perfect is toneless, and so on:
- ndí-fotokoza 'I will explain'
- ndi-ná-fotokoza 'I explained (yesterday)'
- ndi-ku-fótokoza 'I am explaining'
- ndí-ma-fotokóza 'I (usually) explain'
- ndi-nká-fotokóza 'I used to explain'
- ndi-fotokozé 'I should explain'
- nd-a-fotokoza 'I have explained'
These tonal patterns apply (with minor modifications) whether the verb is long or short. In some dialects it is common in longer verbs to hear tonal spreading, so that for example the first verb above is heard as ndífótokoza and the second as ndináfótokoza. However, in the present habitual, the first tone does not spread. Tones will also spread, even in shorter verbs, when a verb is followed by an object, except for the subjunctive -é.
Negative verbs have different tonal patterns. For example, the negative infinitive, negative subjunctive, negative "not yet" tense, and negative future all have a single tone on the penultimate syllable. This tone causes all the other tones, such as the tone of the negative-marker sí, to be omitted:[7]
- si-ndi-dza-fotokóza 'I am not going to explain (at some future time)'
When used in a relative clause, several tenses have a different intonation. Usually there is an extra accent on the first syllable, and often another on the penultimate:
- á-nká-fotokóza 'who used to explain'
When the verb-root itself has a tone, this can be heard on the final vowel. However, if the tense has a penultimate or final accent, the tone cannot be heard:
- nd-a-thamangá 'I have run'
- ndi-ku-thámángá 'I am running' (with a plateau)
- ndí-ma-thamánga 'I usually run' (final tone deleted)
Intonational tones
Lexical and grammatical tones are not the only tones heard in a Chichewa sentence, but there are intonational tones as well. One common tone is a boundary tone rising from low to high which is heard whenever there is a pause in the sentence, for example after a topic or subordinate clause.
Tones are also added to questions. For example, the toneless word kuti 'where?' becomes kúti in the following question:
- kwánu ndi kúti? 'where is your home?'
Further details of intonational tones are given below.
How tones are pronounced
![](../I/m/Pitchtrack_of_sentence_'anadyetsa_nyani_nsomba'.jpg)
The accompanying illustration[8] shows the pitch-track of the following sentence:
- anádyétsa nyaní nsómba 'they fed the baboon (some) fish'
The first word, anádyétsa 'they fed', has an accent theoretically on the antepenultimate syllable. However, in this example, the peak (high point) of the accent does not coincide with the syllable but is delayed, giving the impression that it has spread to two syllables. This process is known as "tone doubling" or "peak delay", and is usual whenever a verb is followed by an object.
The second word, nyaní 'baboon', has an accent on the final syllable, but as usually happens with final accents, it has spread backwards to the penultimate syllable, showing a nearly level or gently rising contour, with only the initial n being low-pitched. Another feature of a final accent is that it tends not to be very high.[9]
If it comes at the end of a sentence a final-tone word such as nyaní can optionally be pronounced as nyăni with a rising tone on the penultimate and the final syllable low. But if a suffix is added, the stress moves to the new penultimate, and the word is pronounced with a full-height tone: nyaní-yo 'that baboon'.[10]
The third word, nsómba 'fish', has penultimate accent. Since the word ends the sentence, the tone falls from high to low.
Another point shown in the voice track is that the first accent is higher than the second. This is normal in declarative sentences and is known as downdrift. But when two tones come together, as in nyaní nsómba, there is no downdrift.
The intensity reading at the top shows that the intensity (loudness) is greatest on the penultimate syllable of each word.
Number of tones
Two pitch levels, high and low, conventionally written H and L, are usually considered to be sufficient to describe the tones of Chichewa.[11] In Chichewa itself the high tone is called mngóli wókwéza ('tone of raising'), and the low tone mngóli wótsítsa ('tone of lowering').[12] Some authors[13] add a mid-height tone but most do not, but consider a mid-height tone to be merely an allophone of nearby high tones.
From a theoretical point of view, however, it has been argued that Chichewa tones are best thought of not in terms of H and L, but in terms of H and Ø, that is to say, high-toned vs toneless syllables.[14] The reason is that H tones are much more dynamic than L tones and play a large role in tonal phenomena, whereas L-toned syllables are relatively inert.[15]
Tones are not marked in the standard orthography used in Chichewa books and newspapers, but linguists usually indicate a high tone by writing it with an acute accent, as in the first syllable of nsómba. The low tones are generally left unmarked.
Works describing Chichewa tones
The earliest work to mark the tones of Chichewa words was the Afro-American scholar Mark Hanna Watkins' A Grammar of Chichewa (1937). This was a pioneering work, since not only was it the first work on Chichewa to include tones, but it was also the first grammar of any African language to be written by an American.[16] The informant used by Watkins was the young Kamuzu Banda, who in 1966 was to become the first President of the Republic of Malawi.
Another grammar including Chichewa tones was a handbook written for Peace Corps Volunteers, Stevick et al., Chinyanja Basic Course (1965), which gives very detailed information on the tones of sentences, and also indicates intonations.[17] Its successor, Scotton and Orr (1980) Learning Chichewa,[18] is much less detailed. All three of these works are available on the Internet. J.K. Louw's Chichewa: A Practical Course (1987) [1980], which contained tone markings, is currently out of print.
From 1976 onwards a number of academic articles by Malawian and Western scholars have been published on different aspects of Chichewa tones. The most recent work discussing the tones of Chichewa is The Phonology of Chichewa (2017) by Laura Downing and Al Mtenje.
Four dictionaries also mark the tones on Chichewa words. The earliest of these was volume 3 of J.K. Louw's Chichewa: A Practical Course (1987) [1980]; A Learner's Chichewa-English, English-Chichewa Dictionary by Botne and Kulemeka (1991), the monolingual Mtanthauziramawu wa Chinyanja/Chichewa (c.2000) produced by the Centre for Language Studies of the University of Malawi (available online),[19] and the Common Bantu On-Line Chichewa Dictionary (2001) formerly published online by the University of California in Berkeley.[20]
So far all the studies which have been published on Chichewa tones have dealt with the Malawian variety of the language. There is no published information available on the tones of Chinyanja spoken in Zambia and Mozambique.
Some tonal phenomena
In order to understand Chichewa tones, it is necessary first to understand various tonal phenomena that can occur, which are briefly outlined below.
Downdrift
Normally in a Chichewa sentence, whenever tones come in the sequence HLH or HLLH, it is usual for the second high tone to be a little lower than the first one. So for example in the word ndímapíta 'I usually go', the tone of the syllable ndí is pronounced a little higher than the tone of pí. Thus generally speaking the highest tone in a sentence is the first one. This phenomenon, which is common in many Bantu languages, is known as 'downdrift'[21] or 'automatic downstep'.[22]
However, there are several exceptions to this rule. Downdrift does not occur, for example, when a speaker is asking a question,[23] or reciting a list of items with a pause after each one, or sometimes if a word is pronounced on a high pitch for emphasis. There is also no downdrift in words like wápolísi 'policeman', where two high tones in the sequence HLH are bridged to make a plateau HHH (see below).
High tone spreading ('HTS')
In some dialects a high tone may sometimes spread to the following syllable; this is known as 'High Tone Spreading' or 'Tone Doubling'.[24] So where some speakers say ndináthandiza 'I helped', others will say ndináthándiza.[25] Some phoneticians argue that what happens here, in some cases at least, is that the highest part or 'peak' of the tone moves forward, giving the impression that the tone covers two syllables, a process called 'peak delay'.[26] An illustration of peak delay can be seen clearly in the pitch-track of the word anádyetsa 'they fed', here pronounced anádyétsa, in Downing et al. (2004).
In order for HTS to occur, there must be at least 3 syllables following the tone, although not necessarily in the same word.
One very frequent use of spreading, at least in some dialects, is to link together two words (such as verb + object, or verb of motion + destination, or noun + possessive) into a single phrase. In transcriptions one frequently finds phrases such as kuphíká nyama 'to cook meat'[27] or kusámála mkázi 'to care for a wife',[28] chímánga chánga 'my maize'[29] mikángó iyo 'those lions'[30] etc. in which the second tone in each phrase is not original but due to spreading. The contexts where spreading occurs vary from one dialect to another.
There are some verbal forms in which spreading does not occur; for example, the first tone in the present habitual does not spread: ndí-ma-thandíza 'I usually help'. The tone on -ngá- and -má- in negative tenses such as sí-ndí-má-thandiza 'I never help' and si-ndi-ngá-thandíze 'I can't help' also does not spread, although it may form a plateau with a 'bumped' tone: si-ndi-ngá-wá-phé 'I can't kill them'.[31]
Tonal plateau
It sometimes happens that the sequence HLH in Chichewa becomes HHH, making a tonal 'plateau'. A tonal plateau is common after the words á 'of' and ndí 'and':
- wápolísi 'policeman' (from wá 'of' and {{lang|ny|polísi 'police')
- chákudyá 'food' (from chá 'of' and kudyá 'to eat')
- ndí Maláwi 'and Malawi'
Before a pause the final tone may drop: chákúdya. Sometimes a succession of tones is bridged in this way, e.g. gúlewámkúlu 'masked dancer', with one long continuous high tone from gú to kú.[32]
Another place where a plateau is found is after the initial high tone of subordinate clause verbs such as the following:
- ákuthándiza 'when he is helping'; 'who is helping'
- ndítathándiza 'after I helped'
At the end of words, if the tones are HLH, a plateau is common:
- mpóngozí 'mother-in-law'
- kusíyaná 'difference'
However, there are exceptions; for example, in the word nyényezí 'star' the two tones are kept separate, so that the word is pronounced nyényēzī (where ī represents a slightly lower tone than í).[33]
Sequences with two or more toneless syllables between the high tones (e.g. HLLH) do not normally make a plateau in Chichewa. This also applies to words in which one of the syllables has been lost by contraction, such as:
- sánafé (contracted from sí-á-na-fé) 'he didn't die'
Tone shifting ('bumping')
When a word or closely connected phrase ends in HHL or HLHL, there is a tendency in Chichewa for the second H to move to the final syllable of the word. This process is known as 'tone shifting'[34] or 'bumping'.[35]
Local bumping
In one type of bumping (called 'local bumping') LHHL at the end of a word or phrase becomes LHHH (where the two tones are joined into a plateau):
- *nyumbá yánga > nyumbá yángá 'my house'
- *kupítá-nso > kupítá-nsó 'to go again'
- *ndinkápíta > ndinkápítá 'I used to go'
- *anámúpha > anámúphá 'they killed him'
At the end of a sentence the final high tone may drop again, reverting the word to anámúpha.
However, in certain verb tenses such as the Present Habitual when the tones are HHL, the penultimate tone is shifted to the final, but there is no plateau. Instead, the two tones are kept separate:
- ndímadyá 'I eat'
There is no bumping in HHL words where the first syllable is derived from á 'of':
- kwámbíri 'very much'
Non-local bumping
In another kind (called 'non-local bumping'), HLHL at the end of a word or phrase changes to HLLH or, with spreading of the first high tone, HHLH:
- *mbúzi yánga > mbúzi yangá (or mbúzí yangá) 'my goat'
- *bánja lónse > bánja lonsé 'the whole family'
- *chimódzimódzi > chimódzimodzí 'in the same way'
Again, the Present Habitual is an exception, and when the tones are HLHL, bumping does not occur:
- ndímapíta 'I go'
Reverse bumping
A related phenomenon, but in reverse, is found when the addition of the suffix -tú 'really' causes a normally word-final tone to move back one syllable, so that LHH at the end of a word becomes HLH:
- *ndipité 'I should go' > ndipíte-tú 'really I should go'
- *chifukwá 'because' > chifúkwa-tú 'because in fact'
Enclitic suffixes
Certain suffixes, known as enclitics, add a high tone to the last syllable of the word to which they are joined. When added to a toneless word or a word ending in LL, this high tone can easily be heard. Bumping does not occur when the tones are HLHL:
- Lilongwe > Lilongwé-nso 'Lilongwe also'
- akuthándiza > akuthándizá-be 'he's still helping'
But when an enclitic is combined with word which has penultimate high tone, there is local bumping, and the result is a triple tone:
- nsómba > nsómbá-nsó 'the fish also'
When added to a word with final high tone, it raises the tone higher (in Central Region dialects, the rising tone on the first syllable of a word like nyŭmbá also disappears):[36]
- nyumbá > nyumbá-nso 'the house also'
Not all suffixes are tonally enclitic in this way. For example, when added to nouns, locative suffixes (-ko, -po, -mo) do not add a tone, e.g.:
- ku Lilongwe-ko 'there in Lilongwe'
However, when added to verbs, the same suffixes add an enclitic tone:
- wachoká-po 'he's not at home' (lit. 'he has gone away from there')
Proclitic prefixes
Proclitic tenses
Conversely, certain prefixes transfer their high tone to the syllable which follows them. Prefixes of this kind are called 'proclitic'[37] or 'post-accenting'. The following tenses have a proclitic prefix:
The infinitive:
- ku-thándiza 'to help'
The Present Continuous, Recent Past, Recent Past Imperfective, and Immediate Imperative:
- ndi-ku-thándiza 'I am helping'
- ndi-na-thándiza 'I helped just now'
- ndi-ma-thándiza 'I was helping'
- ta-thándiza! 'help (now)!'[38]
The Present and Perfect participial tenses:
- ndí-kú-thándiza 'while (I was/am) helping'
- ndí-tá-thándiza 'after I helped'
Verbal Adjective with -ó- (derived from á 'of' + the Infinitive):
- wó-thándiza 'who helps'
Any tense with the infix -ngo- 'just'. But other prefixes used with -ngo- are no longer proclitic; instead, the accent goes on the prefix itself:
- nda-ngo-thándiza 'I have just helped'
- ndi-kú-ngo-thándiza 'I'm just helping'
- ndi-má-ngo-thándiza 'I was just helping'
Proclitic object-markers
Object-markers such as ndí- 'me', mú- 'him/her' etc. also become proclitic when added to an imperative or subjunctive:
- ndi-thándízé-ni! 'help me!'
- mu-ndi-thándízé 'could you help me?'
However, the aspect-marker -ka- 'go and' is not proclitic. With an imperative it is toneless, and with a subjunctive it puts tones on itself and the penultimate:
- ka-thandizé 'go and help!'
- ndi-ká-thandíze 'let me go and help'
Aspect or object-marker with a proclitic tense
When, as often happens, an aspect-marker or object-marker is added to a proclitic tense, extra tones appear on the verb stem, either on the penultimate or on the final:
- ku-má-thandíza 'to be always helping'
- ku-ká-thandizá 'to go and help'
- ku-mú-thandizá 'to help him'
Further details are given below.
Meeussen's Rule
Meeussen's Rule is a process in several Bantu languages whereby a sequence HH becomes HL. This is frequent in verbs. An example in Chichewa is the infinitive kú- + goná 'to sleep', where the addition of the proclitic kú- would normally be expected to produce ku-góná with two high tones; but by Meeussen's Rule the second tone is dropped, leaving ku-góna with a high tone on the penultimate only. In the Southern Region, a-ná-mú-thandiza 'he helped him' is pronounced a-ná-mu-thandiza, presumably also by Meeussen's Rule.[39]
Meeussen's Rule does not apply in every circumstance. For example, a tone derived from spreading is unaffected by it, e.g. ku-góná bwino 'to sleep well', where the tone of the verb-stem goná, having been deleted by Meeussen's Rule to make ku-góna 'to sleep', is replaced by spreading when the word is used in a phrase.[40]
Another instance where Meeussen's Rule does not apply in Chichewa is when the aspect-marker -ká- 'go and' is added to a verb, for example: a-ná-ká-thandiza 'he went and helped'. So far from being deleted, this tone in some circumstances can itself spread to the next syllable, e.g. a-ná-ká-thándiza.[41] The tone of an object-marker such as -mu- 'him' in the same position, however, is deleted by Meeussen's Rule and then replaced by spreading; it does not itself spread: a-ná-mú-thandiza 'he helped him'.
Tone of consonants
Just as in English, where in a word like zoo or wood or now the initial voiced consonant has a low pitch compared with the following vowel, the same is true of Chichewa. Thus Trithart marks the tones of initial consonants such as [m], [n], [z], and [dz] in some words as Low.[42]
However, an initial nasal consonant is not always pronounced with a low pitch. After a high tone it can acquire a high tone itself, e.g. wá ḿsodzi 'of the fisherman'[43] The consonants n and m can also have a high tone when contracted from ndí 'and' or high-toned -mú-, e.g. ḿmakhálá kuti? (short for múmakhálá kuti?) 'where do you live?'.[44]
In some Southern African Bantu languages such as Zulu a voiced consonant at the beginning of a syllable not only has a low pitch itself, but can also lower the pitch of all or part of the following vowel. Such consonants are known as 'depressor consonants'. The question of whether Chichewa has depressor consonants was first considered by Trithart (1976) and further by Cibelli (2012). According to data collected by Cibelli, a voiced or nasalised consonant does indeed have a small effect on the tone of a following vowel, making it a semitone or more lower; so that for example the second vowel of ku-gúla 'to buy' would have a slightly lower pitch than that of ku-kúla 'to grow' or ku-khála 'to sit'. When the vowel is toneless, the effect is less, but it seems that there is still a slight difference. The effect of depressor consonants in Chichewa, however, is much less noticeable than in Zulu.
Lexical tones
Lexical tones are the tones of individual words - or the lack of tones, since quite a large number of words in Chichewa (including over a third of nouns and most verbs in their basic form) are toneless and pronounced with all their syllables on a low pitch.
Nouns
In the CBOLD Chichewa dictionary,[20] about 36% of Chichewa nouns are toneless, 57% have one tone, and only 7% have more than one tone. When there is one tone, it is generally on one of the last three syllables. Nouns with a tone more than three syllables from the end are virtually all foreign borrowings, such as sékondale 'secondary school'.
Comparison with other Bantu languages shows that for the most part the tones of nouns in Chichewa correspond to the tones of their cognates in other Bantu languages, and are therefore likely to be inherited from an earlier stage of Bantu.[45] An exception is that nouns which at an earlier period had HH (such as nsómba 'fish', from proto-Bantu *cómbá) have changed in Chichewa to HL by Meeussen's rule. Two-syllable nouns in Chichewa can therefore have the tones HL, LH, or LL, these three being about equally common, but (discounting the fact that LH words are usually in practice pronounced HH) there are no nouns with the underlying tones HH.
The class-prefix of nouns, such as (class 7) chi- in chikóndi 'love', or (class 3) m- in mténgo 'tree', is usually toneless. However, there are some exceptions such as chímanga 'maize'. The three nouns díso 'eye', dzíno 'tooth', and líwu 'sound or word' are irregular in that the high tone moves from the prefix to the stem in the plural, making masó, manó, and mawú respectively.[46]
Toneless nouns
- chinthu 'thing'
- chipatala 'hospital'
- dzanja 'hand'
- Lilongwe 'Lilongwe'
- magazi 'blood'
- magetsi 'electricity'
- mayeso 'exam'
- mkaka 'milk'
- mlimi 'farmer'
- mowa 'beer'
- moyo 'life'
- mpando 'chair'
- mpira 'ball'
- msewu 'road'
- msonkhano 'meeting'
- mudzi 'village'
- munthu 'person'
- mzinda 'city'
- ng'ombe 'cow, ox'
- njala 'hunger'
- njira 'path'
- nyama 'animal, meat'
Nouns with final tone
As noted above, in isolation these words are actually pronounced bwālō, Chichēwā, etc., ending with two tones of mid height. Alternatively, in some dialects, they may be pronounced bwălò, Chichĕwà, with a rising tone on the penultimate and a low tone on the final.
- bwaló 'open area'
- chaká 'year'
- Chichewá 'Chichewa'
- chinangwá 'cassava'
- galú 'dog'
- madzí 'water'
- maló 'place'
- masó 'eyes'
- mawú 'word'
- mnyamatá 'boy'
- mundá 'garden'
- mutú 'head'
- mwalá 'stone'
- mwaná 'child'
- njingá 'bicycle'
- nyanjá 'lake'
- nyumbá 'house'
- nzerú 'wisdom'
- ufulú 'freedom'
- ulendó 'journey'
- Zombá 'Zomba'
Nouns with penultimate tone
- Bánda 'Banda'
- bánja 'family'
- bóma 'government'
- búngwe 'organisation'
- chikóndi 'love'
- chitsánzo 'example'
- dzíko 'country'
- gúle 'dance'
- Maláwi 'Malawi'
- mankhwála 'medicine'
- máyi 'mother, woman'
- mbáli 'side'
- mbéwu 'seed, crop'
- mfúmu 'chief'
- mkángo 'lion'
- mpíngo 'church, congregation'
- mténgo 'tree'
- mtíma 'heart'
- mvúla 'rain'
- mwamúna 'man'
- ndaláma 'money'
- nkháni 'story'
- nsómba 'fish'
- ntchíto 'work'
- ntháwi 'time'
- nyímbo 'song'
- tsíku 'day'
- vúto 'problem'
Nouns with antepenultimate tone
- búluzi 'lizard'
- chímanga 'maize'
- khwángwala 'crow'
- maséwero 'sport'
- mbálame 'bird'
- mphépete 'side, edge'
- mpóngozi 'mother-in-law'
- msúngwana 'teenage girl'
- mtsíkana 'girl'
- námwali 'initiate'
- njénjete 'house-cricket'
- síng'anga 'witch-doctor'
This group is less common than the first three. Many of the words with this tone are loanwords from Portuguese or English such as:
- bótolo 'bottle'
- kálata 'letter'
- mákina 'machine'
- mbátata 'sweet potatoes'
- nsápato 'shoe'
- pépala 'paper'
Nouns with two tones
A small number of words (mostly compounds) have more than one tone. Some are compounded with á 'of', which has a high tone, so that wá + ntchíto 'man of work' becomes wántchíto 'worker' with two tones joined in a plateau. In words with the sequence HLH like wá-pólísi 'policeman' the high tones usually bridge to make HHH. Verbal forms such as Lólémba 'Monday' (short for Lá kúlémba 'the (day) of writing') are also originally plateaux. Plateaux are marked with underlining:
- chákúdyá 'food'
- Láchíwíri 'Tuesday'
- Lólémba 'Monday'
- Lówéruka 'Saturday'
- wákúbá 'thief'
- wántchíto 'worker'
- wápólísi 'policeman'
- wódwála 'sick person'
- wóphúnzira 'pupil'
- zófúnda 'bedclothes'
- zóóna 'truth'
The prefix chi- in some words adds two high tones, one following chi- and one on the final. The first may spread forwards and the second backwards, but the two tones are kept separate with the second lower than the first:
- chizólowezí (or chizólówezí) 'habit, custom'
- chilákolakó 'desire'
- chipwírikití 'riot'
- chithúnzithunzí 'picture'
- chitsékereró 'stopper'
- chivúndikiró 'lid'
- chiwóngoleró 'steering-wheel'
If there are only three syllables following the prefix chi-, the two tones link into an HHH plateau:
- chikwángwání 'banner, sign'
- chipólówé 'violence, riot'
The (L)HHH pattern is also found in a few other words (all of them compounds):
- kusíyáná 'difference'
- masómphényá 'vision'
- mkámwíní 'son-in-law'
- tsábólá 'pepper'
A triple tone is also found in:
- bírímánkhwe 'chameleon'
In the following compound nouns, however, the high tone of the second element (móto 'fire' and bédi 'bed') is lost:
- gálímoto 'car'
- shítíbedi 'bed sheet'
The following nouns have two separate tones and no plateau. The second tone is lower than the first:
- nyényezí (also nyényezi) 'star'[47]
- kángachépe 'small bribe, tip' (lit. 'though it may be little')
Adjectives
Adjectives in Chichewa are usually formed with word á (wá, yá, chá, zá etc. according to noun class) 'of', which has a high tone. The high tone tends to spread to the following word. When there is a sequence of HLH, the tones will bridge to make HHH:
- wá-bwino 'good'[48]
- wá-ḿkázi 'female'
- wá-mwámúna 'male'
Combined with an infinitive, á and ku- usually merge (except usually in monosyllabic verbs) into a high-toned ó-:
- wó-ípa 'bad'
- wó-pénga 'mad'
- wó-dwála 'sick'[49]
- wá-kú-yá 'deep'
Some people make a slight dip between the two tones:
- wô-dúla 'expensive'[50]
Possessive adjectives are also made with á-. As explained in the section on bumping, their tone may change when they follow a noun ending in HL, LH or HH. The concords shown below are for noun classes 1 and 2:
- wánga 'my'
- wáko 'your'
- wáke 'his, her, its' (also 'their' of non-personal possessors)
- wáthu 'our'
- wánu 'your' (of you plural, or polite)
- wáo 'their' (or 'his, her' in polite speech)
The adjective wína 'another, a certain' has similar tones to wánga 'my':
- wína (plural éna) 'another, a certain'
The adjective wamba 'ordinary', however, is not made with á and has a low tone on both syllables. The first syllable wa in this word does not change with the class of noun.
Pronominal adjectives
The following three adjectives have their own concords and are not formed using á. Here they are shown with the concords of classes 1 and 2:
- yénse (plural ónse) 'all of'
- yémwe (plural ómwe) 'himself'
- yékha (plural ókha) 'only'
As with possessives, the high tone of these may shift by bumping after a noun ending in HL or LH or HH.[51]
With these three the high tone also shifts before a demonstrative suffix: yemwé-yo 'that same one', zonsé-zi 'all these'.[52] (In this they differ from wánga and wína, which do not shift the tone with a demonstrative suffix, e.g. anthu éna-wa 'these other people'.) The tone also shifts in the word álí-yensé 'each, each and every', in which áli has the tones of a relative-clause verb.[53]
The following demonstrative adjectives (shown here with the concords for noun classes 1 and 2) usually have a low tone:[54]
- uyo (plural awo) 'that one'
- uyu (plural awa) 'this one'
- uno (plural ano) 'this one we're in'
- uja (plural aja) 'that one you mentioned'
- uti? (plural ati?) 'which one?'
The first of these (uyo), however, can be pronounced úyo! with a high tone if referring to someone a long way away.[55] The word uti/ati? can also acquire an intonational tone in certain types of questions (see below).
Numbers
Chichewa has the numbers 1 to 5 and 10. These all have penultimate high tone except for -sanu 'five', which is toneless. The adjectives meaning 'how many?' and 'several' also take the number concords and can be considered part of this group. They are here illustrated with the concords for noun classes 1 and 2 (note that khúmi has no concord):
- munthu mmódzi 'one person'
- anthu awíri 'two people'
- anthu atátu 'three people'
- anthu anáyi 'four people'
- anthu asanu 'five people' (toneless)
- anthu khúmi 'ten people'
- anthu angáti? 'how many people?'
- anthu angápo 'several people'
The numbers zaná '100' and chikwí '1000' exist but are rarely used. It is possible to make other numbers using circumlocutions (e.g. 'five tens and units five and two' = 57) but these are not often heard, the usual practice being to use English numbers instead.
Personal pronouns
The first and second person pronouns are toneless, but the third person pronouns have a high tone:[56]
- ine 'I'
- iwe 'you sg.'
- iyé 'he, she'
- ife 'we'
- inu 'you pl., you (polite)'
- iwó 'they, he/she (polite)'
These combine with ndi as follows:
- ndine 'I am'
- ndiyé 'he is' (etc.)
- síndine 'I am not'
- síndíyé 'he is not' (also 'surely')
Monosyllables
The following monosyllabic words are commonly used. The following are toneless:
- ndi 'it is, they are'
- ku 'in, to, from'
- pa 'on, at'
- mu (m') 'in'
The following have a high tone:
- á (also wá, yá, zá, kwá etc. according to noun class) 'of'
- ndí 'with, and'
- sí 'it isn't'
These words are joined rhythmically to the following word. The high tone can spread to the first syllable of the following word, provided it has at least three syllables:[57] They can also make a plateau with the following word, if the tones are HLH:
- Lilongwe 'Lilongwe' > á Lílongwe 'of Lilongwe'
- pemphero 'prayer' > ndí pémphero 'with a prayer'
- Maláwi > á Máláwi 'of Malawi'
The word (in Punjabi) has a tone when it means 'of' following a noun of class 16:
- pa bédi 'on the bed' but pansí pá bédi 'underneath (of) the bed'
It also has a tone in certain idiomatic expressions such as pá-yekha or pá-yékha 'on his own'.
Ideophones
The tones of ideophones (expressive words) have also been investigated by linguists.[58] Examples are: bálálábálálá 'scattering in all directions' (all syllables very high), lólolo 'lots and lots' (with gradually descending tones), bii! 'very dark or dirty' (low pitch). It can be seen that the tonal patterns of ideophones do not necessarily conform to the patterns of other words in the language.
Lexical tones of verbs
Chichewa verbs are mostly toneless in their basic form, although a few have a high tone (usually on the final vowel). However, unlike the situation with the lexical tones of nouns, there is no correlation at all between the high-toned verbs in Chichewa and the high-toned verbs in other Bantu languages. The obvious conclusion is that the high tones of verbs are not inherited from an earlier stage of Bantu but have developed independently in Chichewa.[59]
When a verbal extension is added to a high-toned root, the resulting verb is also usually high-toned, e.g.
- goná 'sleep' > gonaná 'sleep together'
Certain extensions, especially those which change a verb from transitive to intransitive, or which make it intensive, also add a tone. According to Kanerva (1990) and Mchombo (2004), the passive ending -idwa/-edwa also adds a high tone, but this appears to be true only of the Nkhotakota dialect which they describe.[60]
High-toned verb roots are comparatively rare (only about 13% of roots),[61] though the proportion rises when verbs with stative and intensive extensions are added. In addition there are a number of verbs, such as peza/pezá 'find' which can be pronounced either way. In the monolingual dictionary Mtanthauziramawu wa Chinyanja 2683 verbs are given, with 10% marked as high-toned, and 4% as having either tone. In the Southern Region of Malawi, some speakers do not pronounce the tones of high-toned verbs at all or only sporadically.
The difference between high and low-toned verbs is neutralised when they are used in a verb tense which has a high tone on the penultimate or on the final syllable.
Three irregular verbs, -téro 'do so', -tére 'do like this', and -táni? 'do what?', have a tone on the penultimate syllable.
The view held in Mtenje (1986) that Chichewa also has 'rising-tone' verbs has been dropped in his more recent work.[62]
Low-toned verbs
- bwera 'come'
- chita 'do'
- choka 'go away'
- dziwa 'know'
- fika 'arrive'
- fotokoza 'explain'
- funa 'want'
- funsa 'ask'
- ganiza 'think'
- gula 'buy'
- gulitsa 'sell'
- gwira 'take hold of'
- imba 'sing'
- khala 'sit, live'
- kumana 'meet'
- lankhula/yankhula 'speak'
- lemba 'write'
- lowa 'enter'
- mwalira 'die'
- nena 'say'
- ona 'see'
- panga 'do, make'
- patsa 'give (someone)'
- pereka 'hand over'
- pita 'go'
- seka 'laugh'
- sintha 'change'
- tenga 'take'
- thandiza 'help'
- uza 'tell'
- vala 'put on (clothes)'
- vuta 'be difficult'
- yamba 'begin'
- yankha 'answer'
- yenda 'go, walk'
Monosyllabic verbs such as the following are always low-toned, although nouns derived from them, such as imfá 'death', can have a tone:
- ba 'steal'
- dya 'eat'
- fa 'die'
- mva 'hear'
- mwa 'drink'
- pha 'kill'
- tha 'finish, be able'
High-toned verbs
- bisá 'hide (something)'
- bisalá 'be hidden'
- dabwá 'be surprised'
- dandaulá 'complain'
- goná 'sleep'
- iwalá 'forget'
- kaná 'refuse'
- kondá 'love'
- lakwá 'be in error'
- lepherá 'fail'
- phunzirá 'learn'
- siyá 'leave'
- tayá 'throw away'
- thamangá 'run'
- topá 'be tired'
- tsalá 'remain'
- vulazá 'wound'
- yenerá 'ought'
- zimá/thimá 'go out (of fire or lights)'
Verbs with either tone
- ipá 'be bad'
- khotá 'be bent'
- kwiyá 'be angry'
- namá 'tell a lie'
- pewá 'avoid'
- pezá 'find'
- sowá 'miss, be missing'
- thokozá 'thank'
- yabwá 'irritate, make itch'
Stative verbs
Most intransitive verbs with the endings -iká, -eká, -uká, -oká derived from simpler verb-stems are high-toned. This is especially true when a transitive verb has been turned by a suffix into an intransitive one:
- chitiká 'happen' (cf. chita 'do')
- duká 'be cut' (cf. dula 'cut')
- dziwiká 'be known' (cf. dziwa 'know')
- funiká 'be necessary' (cf. funa 'want')
- masuká 'be at ease' (cf. masula 'free')
- mveká 'be understood' (cf. mva 'hear')
- onongeká 'be damaged' (cf. ononga 'damage')
- theká 'be possible' (cf. tha 'finish, manage to')
- thyoká 'be broken' (cf. thyola 'break')
- vutiká 'be in difficulty' (cf. vutitsa 'cause a problem for')
However, there are some common exceptions such as the following which are low-toned:
- oneka 'seem' (cf. ona 'see')
- tuluka 'come out, emerge' (cf. tula 'remove a burden')
Intensive verbs
Intensive verbs with the endings -its(its)á and -ets(ets)á always have a high tone on the final syllable, even when derived from low-toned verbs. A few intensive verbs with the endings -irirá or -ererá are also high-toned:[63]
- yang'angitsitsá 'examine carefully' (cf. yang'ana 'look at')
- onetsetsá 'inspect' (cf. ona 'see')
- mvetsetsá 'understand well' (cf. mva 'hear')
- menyetsá 'beat severely' (cf. menya 'hit')
- pitirirá 'go further' (cf. pita 'go')
- psererá 'be overcooked' (cf. psa 'burn, be ripe')
Grammatical tones of verbs
Tonal melodies of positive tenses
In addition to the lexical tones which go with individual words, Chichewa also has grammatical tones which apply to verbs. Each different tense has its own tonal pattern or melody, which is superimposed on top of whatever lexical tone the verb-stem itself may have.[64]
There are at least eight different tonal patterns in ordinary positive tenses, while others are used in negative tenses or relative clause tenses.[65]
Sometimes two tenses have the same tense-marker and are distinguished by tone alone, as with the Present Habitual and the Past Imperfective. The Present Habitual has two tones, one on the subject-marker and the other on the penultimate, while the Past Imperfective has a tone following the tense-marker:
- ndí-ma-thandíza 'I help'
- ndi-ma-thándiza 'I was helping'
In the examples, hyphens have been added for clarity; they are not used in the standard orthography.
Toneless
Some tenses, such as the Perfect, are toneless. In toneless tenses all the syllables are pronounced low, unless the verb-stem itself has a high tone (e.g. nd-a-goná 'I have lain down'):
- nd-a-fotokoza 'I have explained'
- nd-a-werenga 'I have read'
- nd-a-ona 'I have seen'
- nd-a-dya 'I have eaten'
Other toneless tenses are the simple Imperative (thandiza! 'help!'), the Potential (also called 'Permissive') (ndi-nga-thandize 'I can help'),[66] and the -ka- tense meaning 'when' or 'if' (ndi-ka-fotokoza 'if/when I explain').[67]
Final tone
One tense, the Subjunctive, has a high tone on the final vowel:
- ndi-fotokozé 'I should explain'
- ndi-werengé 'I should read'
- ndi-oné 'I should see'
- ndi-dyé 'I should eat'
Proclitic
The Infinitive, Present Progressive, Recent Past, and Past Imperfective all have a proclitic tone, that is, a high tone is heard on the syllable immediately following the tense-marker:[68]
- ku-thándiza 'to help'
- ndi-ku-thándiza 'I am helping'
- ndi-na-thándiza 'I helped' (just now)
- ndi-ma-thándiza 'I was helping'
If the verb is monosyllabic, the tone often spreads backwards or regresses to the penultimate:
- ndi-ku-dyá (pronounced ndi-kú-dyá or ndi-kú-dya)[69] 'I am eating'
Tense-marker tone
In the Past Simple/Remote Perfect, whether made with -na- or -da-, the tone comes on the tense-marker itself. This tone can also spread to the next syllable in longer verbs in some dialects:[70]
- ndi-ná-fotokoza (or ndi-ná-fótokoza) 'I explained'[71]
- ndi-ná-werenga (or ndi-ná-wérenga) 'I read'
- ndi-ná-ona 'I saw'
- ndi-ná-dya 'I ate'
Another tense with tense-marker tone is the perfect conditional:
- ndi-kadá-fotokoza (or ndi-kadá-fótokoza) 'I would have explained'
Subject-marker tone
The Present Simple, which often refers to events in the near future, has a high tone on the initial syllable, that is, on the subject-marker ndi- 'I'.[72] In some dialects this high tone will spread (or shift forward) to the second syllable in longer verbs (that is, verbs where the high tone is followed by at least three toneless syllables):[73]
- ndí-fotokoza (or Template:Nd'''í'''-f'''ó'''tokoza) 'I will explain'
- ndí-werenga (or Template:Nd'''í'''-w'''é'''renga) 'I will read'
- ndí-ona 'I will see'
- ndí-dya 'I will eat'
Subject-marker and tense-marker
In some dialects the two future tenses with -dzá- and -ká- have a tone on both the subject prefix and the tense-marker:[74]
- ndí-dzá-fotokoza 'I will explain (at a future time)'
However, in other dialects the first tone is dropped, and only the tense-marker has a tone:
- ndi-dzá-fotokoza
Subject-marker and penultimate
The Present Habitual has a high tone on the initial syllable and another on the penultimate. But if the verb-stem is monosyllabic, the second tone is heard on the final syllable instead of the penultimate.[70] Because of downdrift, the second high tone is slightly lower in pitch than the first.[75]
- ndí-ma-fotokóza 'I usually explain'
- ndí-ma-werénga 'I usually read'
- ndí-ma-óna 'I usually see'
When the verb is monosyllabic, the second tone is on the final, unless there is an object-marker:
- ndí-ma-dyá 'I usually eat'[76]
- ndí-ma-mú-dya 'I usually eat it (e.g. sugar)'
In this tense the first tone never spreads to -ma-.[77]
The Future Continuous also has this tonal pattern:
- ndí-zi-(dza)-werénga 'I shall be reading'
Initial, peninitial, and penultimate
The Remote Past (with -naa- or {{lang|ny|-daa-) is similar, but the first tone usually spreads to the second syllable:
- ndí-ná-a-fotokóza 'I explained (but...)'
- ndí-náa-dyá 'I had eaten'[78]
Another tense with this tone-pattern is the not commonly used 'continuative' subjunctive:[79]
- tí-báa-fotokóza 'let us continue explaining'
Tense-marker and penultimate
The Remote Past Imperfective also has two tones, but on the tense-marker and the penultimate. The second tone disappears if the verb is monosyllabic. The tone on -ká- never spreads:[80]
- ndi-nká-fotokóza 'I was explaining'
- ndi-nká-werénga 'I was reading'
- ndi-nká-óná or ndi-nká-oná 'I was seeing'
- ndi-nká-dya 'I was eating'
The same pattern is found in the Necessitative (= Imperfective Subjunctive). In some dialects this tense is formed with -dzi- instead of -zi-:[81]
- mu-zí-werénga 'you must (always) read'
Lexical tone combined with tense
As noted above, some verb stems have an inherent lexical tone, for example goná 'lie down', siyá 'leave (something)', yang'anitsitsá 'examine carefully'. When such a verb is used to make a tense, the tone can usually be heard on the final syllable. The tone can be heard most easily when the tense-melody is toneless, as in the Perfect tense:[82]
- nd-a-siyá 'I have left (it)'
However, it can also be heard when there is another high tone in the verb, provided this tone is not in one of the last two syllables:
- ku-khúlulukirá 'to forgive'
If the tense-pattern puts a tone on the antepenultimate syllable (so that the verb ends HLH), the two tones will link up, making a plateau:
- ku-gónáná 'to sleep with one another'
However, if the tonal pattern of the tense puts a tone on the penultimate or the final syllable, the tone of the verb-stem cannot be heard; in the first case because it is deleted by Meeussen's Rule,[83] and in the second, because the two tones coincide:
- ku-góna (not *ku-góná) 'to lie down'
- ndi-goné 'I should lie down'
Modifications with shorter verbs
Certain modifications of the tones of a tense take place when the verb-stem is a short one, i.e. of only one syllable, such as -dya 'eat', or two syllables, such as ona 'see'.
In most tenses, a penultimate tone moves to the final if the verb is monosyllabic. But if there is an aspect-marker or object-marker, the penultimate tone goes on that:
- ndí-ma-píta 'I usually go'
- ndí-ma-dyá 'I usually eat'[84]
- ndí-ma-ká-dya 'I usually go and eat'
In the 'not yet' tense with -na-, however, the penultimate tone remains penultimate:
- si-ndi-ná-dya 'I have not yet eaten'
It also remains penultimate in the negative future with -dza- or -ka-:
- si-ndi-dzá-dye 'I will not eat'
When the shortening of a verb would mean that the verb ends LHH, the second tone disappears:
- ndi-nká-dya 'I used to eat'
When a verb has two syllables, in most cases, if the ending would normally be HHL, the second tone shifts to the final, usually making a tonal plateau:
- ku-ká-ményá 'to go and hit'[85]
- ti-ná-mú-phá 'we killed him'[86]
- ndi-nká-pítá or ndi-nká-pitá "I used to go'
However, when a monosyllabic verb is preceded by -ná-ká- 'went and', this bumping does not occur:[87]
- ti-ná-ká-pha 'we went and killed'
Tonal patterns of negative tenses
Negative tenses in Chichewa tend to have a different tonal pattern from the corresponding positive ones.[88]
In general, negative intonations can be divided into two groups:
- (a) Those with a tone on the negative-marker sí- (which spreads to the subject-marker). This intonation is used for negative past tenses and negative present tenses.
- (b) Those with no tone on the initial syllable of the verb, but a tone on the penultimate and sometimes also on the tense-marker. This intonation is used for all negative future tenses, tenses meaning "not yet", and also for the negative infinitive and subjunctive.
Consequently in some tenses, there are two negative intonations with different meanings. For example, the Remote Perfect (Simple Past) has a tone on sí- and the penultimate when it means "didn't", but a single tone on the penultimate when it means "have not yet":[89]
- sí-ndí-na-thandíze 'I didn't help'
- si-ndi-na-thandíze 'I haven't helped yet'
The negative Present Simple / Near Future also has two intonations, depending on whether the meaning is present or future:
The negative tonal melodies are as follows:
sí- only
The Present Simple, in its habitual, non-future sense, has a tone on the initial syllables only. This second tone does not spread even in longer verbs and is presumably itself due to spreading:[92][93]
- sí-ndí-thandiza 'I don't help'[94]
However, if an object-marker is added, there is an extra tone on the penultimate syllable:
- sí-ndí-mu-thandíza 'I don't help him'[95]
In monosyllabic verbs, in the Central Region, the tone on the subject-marker disappears in the Present Simple and the tones are:
- sí-ndi-dya 'I don't eat'
sí- and penultimate
The negative Remote Perfect tense with -na- or -da- has tones on the negative prefix sí- and the penultimate. The first tone may spread. The ending changes to -e:
- sí-ndi-na-thandíze 'I did not help'
- sí-ndi-na-fotokóze 'I did not explain'
In a monosyllabic verb, the second tone goes on the final, unless there is an object-marker:
- sí-ndi-na-dyé 'I didn't eat'
- sí-ndi-na-mú-dye 'I didn't eat it (e.g. the sugar)'
The negative perfect participial tense has tones very similar to the above, with the first tone on the initial syllable and the second on the penultimate. The first tone may spread:[96]
- ndí-sa-na-thandíze 'without my having helped', 'before I help(ed)'
- ndí-sa-na-dyé 'before I eat/ate'[96]
- ndí-sa-na-mú-dye 'before I eat/ate it (e.g. the sugar)'
sí-, tense-marker, and penultimate
The Remote Past Imperfective has three tones when negative. The first may spread. The n of nká is syllabic and low-toned, so there is no plateau between the first and second tone:[97]
- sí-ndi-nká-thandíza 'I didn't use to help'
In a monosyllabic verb, the third tone is lost:
- sí-ndi-nká-dya 'I didn't use to eat'
sí- and proclitic tone
Tenses with proclitic tense-markers, such as the Present Continuous, have a tone on sí and a second tone immediately following the tense-marker:[98]
- sí-ndí-ku-thándiza 'I'm not helping'
- sí-ndí-ma-thándiza 'I wasn't helping'
- sí-ndí-na-thándiza 'I didn't help just now' (rarely used)
sí- and aspect-marker
If the aspect-marker -ma- is added to any negative tense, it carries a high tone; otherwise the tones are similar to the same tenses without -ma-:
- sí-ndí-má-thandiza 'I never help'[99]
- sí-ndí-má-dya 'I never eat'
When an object-marker is added there is an additional tone on the penultimate:
- sí-ndí-má-mu-thandíza 'I never help him'
The negative Perfect Conditional can also optionally have this intonation:
- sí-ndí-kadá-thandiza 'I would not have helped' (but also {{lang|ny|sí-ndí-kada-thandíza)
Penultimate tone only
Any negative tense with future meaning, or with the meaning 'not yet', has only one tone, on the penultimate syllable. Other tones, such as those of the negative prefix sí- and the object-marker, are suppressed:
- si-ndi-thandíza 'I won't help'[100]
- si-ndi-dza-thandíza 'I won't help' (in future)[101]
- si-ndi-dza-mu-thandíza 'I won't help him'
- si-ndi-na-thandíze 'I haven't helped yet'
The same intonation, with a tone on the penultimate, is found in non-finite tenses such as the negative Infinitive and negative Subjunctive, which have the negative-marker -sa- following the subject-marker:
When these patterns are used with a monosyllabic verb such as -dya 'eat', the tone usually goes on the final syllable:
However, in the forms with -dza- and -na- the tone goes on the penultimate:
- si-ndi-dzá-dya 'I will not eat' (at a future time)
- si-ndi-ná-dye 'I have not yet eaten'
It also goes on the penultimate if a monosyllabic verb is preceded by an object-marker:
- si-ndi-mú-dya 'I won't eat it' (e.g. the sugar)
- ku-sa-mú-dya 'not to eat it'[102]
Tense-marker and penultimate
The negative Future Continuous and negative Potential have tones on tense-marker and penultimate. The tone on the tense-marker does not spread:[105]
- si-ndi-zí-thandíza 'I won't be helping'
- si-ndi-ngá-thandíze 'I won't be able to help'
In a monosyllabic verb, the two tones coincide:
- si-ndi-ngá-phe 'I cannot kill'
The negative infinitive with the aspect-marker -ma- has the same tones:
- ku-sa-má-thandíza 'to be never helping'
In a monosyllabic verb, the second tone is lost by Meeussen's Rule:
- ku-sa-má-dya 'to be never eating'
- si-ndi-ngá-dye 'I can't eat'
Relative clause intonation
Introduction
Certain tenses have a different tonal pattern when used in certain kinds of dependent clauses. Stevick calls this intonation the 'relative mood' of the verb,[106] since it is frequently used in relative clauses; however, it is also used in a range of other dependent clauses, such as conditional clauses ('if...'), cleft sentences ('it is he who...'), and adverbial clauses of time, place, manner, and concession. Often the use of relative clause intonation alone can show that a verb is being used in the meaning of a relative or conditional clause.[107]
How it is made
The dependent-clause intonation generally has two high tones, one on the initial syllable and another on the penultimate. High tones between these two are suppressed. The first high tone may spread. When the tones are HLHL, some dialects have bumping; for example, ndí-na-gúla '(which) I bought' can become ndí-ná-gulá.[108]
- a-ná-thandiza 'he helped' becomes á-na-thandíza (or á-ná-thandíza)
- wa-thandiza 'he has helped' becomes wá-thandíza (or when used as a noun or adjective {{lang|ny|wá-thándíza)
- ndi-nga-thandize 'I can help' becomes ndí-nga-thandíze[109]
But when the tense has a proclitic prefix the second high tone comes on the syllable after the tense-marker. There is usually[110] bridging of the two tones, as well as a bridging with the tone of the relative pronoun améne:[111]
- améné á-kú-thándiza 'who is helping'
- améné á-má-thándiza 'who was helping'
If the verb has only one syllable, the second high tone is dropped at the end of a phrase:
- watha 'it has finished' becomes wátha 'which has finished'[112]
The participial tenses also use the relative clause intonation.
- ndíkúthándiza 'while I am helping'
- ndítáthándiza 'after I helped', 'having helped'[113]
- ndísanathandíze 'with me not yet having helped', 'before I help(ed)'
If the verb tense already has a high tone on the initial syllable, such as the Present Simple, Present Habitual, and Remote Future, there is no change when the verb is used in a relative clause.[114] Negative tenses, such as s-a-na-thandíze 'he hasn't helped yet', are also unaltered when used in dependent clauses.
Where it is used
The relative clause intonation is used most often in relative clauses, for example after améne or yemwé 'who'. The tone of améne spreads to make a plateau with the high tone of the relative clause verb:
- malálánjé améné múnágulá alí kuti?
'where are the oranges which you bought?'[115] - ápézá galú améné ánámusowétsa
'he will find the dog which he lost'[116]
The relative clause intonation is also used in temporal clauses after paméne 'when' and in clauses of manner after momwé 'in the way in which', which are derived from the same roots:
- amavála paméné ndímálówa
'he was getting dressed when I entered' - momwé ánkachitíra
'as he used to do'
It is also used if the word améne or paméne is omitted, the relative clause intonation alone showing that the verb is being used in a relative way:
- amadwála kwámbíri álí mwaná
'he was very sick when he was a child'[117] - mvúlá ílí kugwá
'when rain is falling'[118] - mwezí wá-tha
'the month which has finished, i.e. last month'[119]
Questions with ndaní? 'who?' and nchiyáni? 'what?' are expressed as cleft sentences, using relative clause intonation:
- wákhálá pa-mpando ndaní?
'who is sitting on the chair?' (lit. 'the one who has sat down on the chair is who?') - cháchítiká n'chiyáni?
'what's happened?' (lit. 'the thing that has happened is what?')
The dependent clause intonation is also used in conditional clauses, except those with the toneless -ka- tense. An example can be observed in the following proverb, where the dependent verb has a different intonation from the main verb:
- ndíkanadzíwa, ndikanáphika thereré
'If I'd known (that you were going to come back from the hunt empty handed), I'd have cooked some vegetables!'[120]
It is similarly used after ngati 'if' and in clauses after ngakhále 'although':
- sízíkudzíwíká ngati wámwalíra (or wámwálirá)
'it isn't known if she has died' - ngakhálé mvúla íkubvúmba
'even when rain is falling'[121]
However, when ngati means 'as if', the ordinary intonation is used.
Tones of -li ('am', 'are', 'is')
As well as the word ndi 'is/are' used for identity (e.g. 'he is a teacher') Chichewa has another verb -li 'am, are, is' used for position or temporary state (e.g. 'he is well', 'he is in Lilongwe'). The tones of this are irregular in that in the Present Simple, there is no tone on the subject-marker.[122] For the Remote Past, both á-naa-lí and a-ná-li[123] can be heard, apparently without difference of meaning. In the dependent Applied Present (-lili), used in clauses of manner, the two tones make a plateau.
In main clauses:
- Present: a-li 'he is, they are'
- Recent Past: a-na-lí 'he was (today)'
- Remote Past: á-náa-lí (ánaalí) or a-ná-li 'he was' (in the past)
- Persistive: a-kada-lí (or a-ka-lí) 'he is still'
In dependent clauses:
- Present: á-li 'when he is/was'
- Persistive: á-kada-lí 'when he is/was still'[124]
- Applied Present: momwé á-lílí 'the way that he is'
The dependent-clause form of the Persistive tense is frequently heard in the phrase pá-kada-lí pano 'at the present time' (literally, 'it still being now').
Aspect-markers
After the tense-marker, there can be one or more aspect-markers, which add precision to the meaning of the tense.[125] Altogether there are four aspect-markers, -má- 'ever, usually, always', -ká- 'go and', -dzá- 'in future', and -ngo- 'just', which are always added in that order, though not usually all at once. These infixes add extra high tones to the verb.
-má-
The aspect-marker -má- 'always, ever, generally' usually adds two high tones, one on -má- itself and one on the penultimate syllable. For example, when added to the 'when' tense, which has the toneless tense-marker -ka-, it adds two tones:
- ndi-ka-má-thandíza 'whenever I help'
However, in the Present Habitual, the tone on -má- is lost:[126]
- ndí-ma-mu-thandíza 'I usually help him'
On the other hand, in the negative Present Habitual, the tone on -má- is retained while the penultimate tone is lost, unless there is an object-marker:
- sí-ndí-má-thandiza 'I never help'[127]
- sí-ndí-má-mu-thandíza 'I never help him'
-má- can also be used as a tense-marker itself to make the Imperfect tense, in which case its high tone is proclitic:
- ndi-ma-thándiza 'I was helping'
- sí-ndí-ma-thándiza 'I wasn't helping'[128]
-ká- and -dzá-
The aspect-markers -ká- 'go to' and -dzá- 'in future' or 'come to' have the same tones as each other. In the Remote Perfect, the tone is high, and in longer verbs sometimes spreads:[40]
- ndi-ná-ká-thandiza (or ndi-ná-ká-thándiza) 'I went to help'
In a monosyllabic verb in this tense, unlike with the object-marker, the tone of -ká- does not bump to the final:[129]
- ti-ná-ká-pha 'we went and killed'
Contrast:
- ti-ná-mú-phá 'we killed him'
In tenses which have a tone on the penultimate, -ká- and -dzá- usually lose their tone:
- ndí-ma-ka-thandíza 'I (always) go to help'
- sí-ndí-na-ka-thandíze 'I didn't go to help'
- si-ndi-dza-píta 'I won't go'[130]
- ku-sa-ka-thandíza 'not to go and help'[131]
But in monosyllabic verbs, where -ká- comes in penultimate position, it receives the tone:
- ndí-ma-ká-dya 'I usually go and eat'
The high tone is also lost in the Imperative, where there is a tone on the final syllable only:
- ka-thandizé 'go and help!'
- ka-dyé 'go and eat!'[132]
But in the positive Subjunctive -ká- and -dzá- retain their tone, and there is a second tone on the penultimate. In two-syllable verbs this second tone is bumped to the final. In monosyllabic verbs it is dropped:
- ndi-ká-thandíze 'I should go and help'
- ndi-ká-óné 'I should go and see'
- ndi-ká-dye 'I should go and eat'
-ngo-
The aspect-marker -ngo- 'just' is derived from the infinitive, and like the infinitive it is proclitic, that is to say, there is a high tone on the syllable after -ngo-. (The syllable -ngo- itself is always toneless.)
In toneless tenses, the syllables before it are toneless:
- i-ngo-bwéra! 'just come!'
- nda-ngo-bwéra 'I have just come'
In all other tenses, there is a high tone on the syllable preceding -ngo-:
- ndí-má-ngo-thándiza 'I usually just help'
If -ngo- is added after a proclitic tense-marker such as the Present Continuous -ku-, the tense-marker has a high tone and is not proclitic:[133]
- ndi-ku-thándiza 'I'm helping'
- ndi-kú-ngo-thándiza 'I'm just helping'
Object-markers
Between the aspect-markers and the verb stem, it is possible to add an object-marker such as -mú- 'him/her/it' or -zí- 'those things' etc. In some tenses, the object-marker retains its tone, but in some the tone is lost. In others, the object-marker adds a tone, but not on itself.
With toneless tenses
When an object-marker is added to a toneless tense, such as the Perfect, it has a high tone. (In verbs of three syllables or longer, the tone may spread in some dialects.)[134][135]
- nda-mú-thandiza 'I have helped him'
- ndi-ka-mú-thandiza 'if I help him'
- ndi-nga-mú-thandize 'I can help him'[136]
With penultimate-tone tenses
In tenses which have penultimate tone the object-marker is usually toneless:
But when the verb is monosyllabic, the penultimate tone goes on the object-marker itself:
With the Remote Perfect
In the Remote Perfect (Past Simple) tense, which has the tense-marker -ná- or -dá-, the tone of the object-marker is lost in southern dialects.[141] In dialects with high tone spreading, the object-marker has a high tone, but it is thought that this tone is due to spreading from the previous syllable, as it does not spread further to the following syllable, unlike the tone of -ká-:[40]
- ndi-ná-mu-thandiza (or ndi-ná-mú-thandiza) 'I helped him'
Contrast:
- ndi-ná-ká-thandiza (or ndi-ná-ká-thándiza) 'I went and helped'
With monosyllabic verb-stems, however, the high tone of the object-marker is retained and (unlike the tone of -ká-) undergoes bumping, making a plateau:[142]
- ti-ná-mú-phá 'we killed him'
Contrast:
- ti-ná-ká-pha 'we went and killed'
With proclitic tenses
Where the tense-marker or aspect-marker has proclitic tone, for example in the infinitive, Present Continuous, or Recent Past, there is an additional tone on the penultimate in longer verbs, or on the final in verbs of 2 or 3 syllables. This second tone disappears if the verb has only one syllable:[143]
- ku-mú-fotokozéra 'to explain to him'
- ku-mú-thamangítsa 'to chase him'
- ku-mú-thandizá 'to help him'
- ku-mú-ményá 'to hit him'
- ku-mú-pha 'to kill him'
This is similar to the effect of adding the aspect-marker -má-, except that in three-syllable verbs with -má- the second tone is heard on the penultimate:
- ku-má-thandíza 'to be always helping'
With Subjunctive and Imperative
In the Subjunctive and Imperative the high tone of the object-marker becomes proclitic and is heard on the syllable which follows it. The final vowel changes to -e and has a tone. There is a plateau if the verb stem has three syllables:[144]
- ndi-mu-thándízé 'I should help him'
- mu-chi-gúle 'you should buy it'[145]
- mu-fótokozeré 'explain to him!'
- mu-thándízé 'come and help him!'
If the verb has two syllables, the tone on the final is dropped by Meeussen's rule:
- (ndi-)pátse-ni 'give me!'
- mu-ndi-pátse 'you should give me'
If the verb is monosyllabic, the tone remains on -mú-, and the high tone of the final vowel is dropped:
- mú-dye 'eat it!'[146]
- mu-mú-dye 'you should eat it'
When the Imperative is preceded by the aspect-markers ta- ('do it now!') or i-ngo- ('just do it!'), which are proclitic, the rules for proclitic verbs given above apply:
- i-ngo-mú-thandizá! 'just help him!'
- ta-ndí-fotokozéra! 'explain to me now please'
When -ka- or -dza- are added to an Imperative, they are toneless, and the tone of the object-marker remains proclitic:
- dza-mu-thándízé 'come and help him!'
But when -ká- or -dzá- is added to the Subjunctive, there are tones on -ká- and the penultimate, and the object-marker loses its tone:
- mu-dzá-mu-thandíze 'you should come and help him'
Reflexive-marker
The reflexive-marker -dzí- in most dialects has exactly the same tones as object-markers such as -mú-, but for some speakers in parts of the Central Region there is also an extra tone on the penultimate or final syllable.[147]
- a-ná-dzí-thandiza 'he helped himself' (Lilongwe dialect)
- a-ná-dzí-thandíza 'he helped himself' (Nkhotakota dialect)
Intonational tones
In addition to the ordinary lexical tones which go with individual words, and the grammatical tones of verb tenses, other tones can be heard which show phrasing or indicate a question.
Boundary tones
![](../I/m/Myers_boundary_tone_illustration.png)
Quite often, if there is a pause in the middle of sentence, such as might be indicated by a comma in writing, the speaker's voice will rise on the syllable just before the pause. This rising tone is called a boundary tone.[148] A boundary tone is typically used after the topic of a sentence, at the end of a dependent clause, after items on a list, and so on. The illustration included here of the sentence Mwamúna, ámulamúlá amáyi ('A man, he rules women')[149] clearly shows the rise in the voice on the last syllable of the word mwamúna, which is here taken to be the topic of the sentence.
A typical sentence where the dependent clause precedes the main clause is the following:
- mu-ka-chí-kónda↗, mu-chi-gúle[150] 'if you like it, please buy it'
As Kanerva points out, the rising boundary tone is not used when the order is reversed and the dependent clause follows the main clause.[151]
Another kind of tone considered to be a boundary tone, but this time a low one, is the optional fall in the speaker's voice at the end of sentences which causes the final high tone on words like chákúdyá 'food' to drop to become chákúdya. The end-of-sentence boundary tone is marked L% in Myers' illustration.
Both Kanerva and Stevick also mention a 'level' boundary tone, which occurs mid-sentence, but without any rise in pitch.[152]
Tones of questions
Wh-Questions
Questions in Chichewa often add high tones where an ordinary statement has no tone. For example, with the word kuti? 'where?', liti 'when?', yani 'who?' or chiyáni 'what?' some people add a tone on the last syllable of the preceding word. This tone does not spread backwards, although it may form a plateau with an antepenultimate tone, as in the 3rd and 4th examples below:
- alí kuti? 'where is he?'[153]
- ndiwé yani? 'who are you?'
- mu-ná-fíká liti? 'when did you arrive?'[154]
- mu-ká-chítá chiyáni? 'what are you going to do there?'[155]
But as Stevick points out, not all speakers do this, and others may say mu-ná-fika liti?[156]
When kuti? 'which place?' or liti? 'which day?' are preceded by ndi 'is', they take a high tone on the first syllable:
- kwánú ndi kúti? 'where is your home?'[157]
It appears that with some speakers the high tone after ndi is heard on the final syllable in forms of this adjective which begin with a vowel; but with other speakers it is heard on the first syllable:[158]
- mipando yáikúlu ndi ití? 'which are the large chairs?'
- msewu wópíta ku Blántyre ndi úti? 'which is the road going to Blantyre?'
A high tone also goes on the final syllable in ndaní? '(it is) who?' (which is derived from ndi yani?)[159] Before this word and nchiyáni? '(is) what?', since such questions are phrased as a cleft sentence or relative clause, the verb has its relative-clause intonation:
- wákhálá pampando ndaní? '(the one who is) sitting on the chair is who?', i.e. 'who is it who's sitting on the chair?'
The relative-clause intonation of the verb is also used when a question begins with bwánji? 'how come?', but not when it ends with bwánji?, when it has the meaning 'how?'
Yes-no questions
With yes-no questions, intonations vary. The simplest tone is a rising boundary tone on the final syllable:
- mwa-landirá? 'did you receive it?'[160]
A more insistent question often has a HL falling boundary tone on the last syllable. Pitch transcriptions show that the voice rises up on the penultimate and falls on the final:[161]
- mwa-landirâ? 'did you receive it?'
But in other dialects, it seems that this fall may begin on the penultimate syllable:
- ku Zómba? 'in Zombá?'[162]
If there is already a penultimate high tone it may simply be raised higher:
- ku Baláka? 'in Baláka?'[163]
Alternatively, there can be two successive falling tones, one on the penultimate, and another on the final.[164]
Sometimes, however, there is no particular intonational tone and the question has the same intonation as a statement, especially if the question starts with the question-asking word kodí.[165]
When there is a choice between two things in a disjunctive question, the first half of the question ends in a high boundary tone, but the voice drops in the second half:[166]
- mukufúná khófí (H%), kapéná thíyi (L%)? 'do you want coffee or (would you prefer) tea?'
Other idiomatic tones
Some speakers add intonational tones also with the toneless word kale 'already', making not only the final syllable of kale itself high but also the last syllable of the verb which precedes it:
Other speakers do not add these intonational tones, but pronounce ndavina kale with Low tones.
Occasionally a verb which is otherwise low-toned will acquire a high tone in certain idiomatic usages, e.g. ndapitá 'I'm off' (said on parting), from the normally toneless pita 'go'. This can perhaps also be considered a kind of intonational tone.
Focus and emphasis
In European languages it is common for a word which is picked out for contrast to be pronounced on a higher pitch than the other words in a sentence, e.g. in the sentence {{they fed the baboon fish, not the elephant}}, it is likely that the speaker will draw attention to the word baboon by pronouncing it on a high pitch, while the word fish, which has been mentioned already, will be on a low pitch. This kind of emphasis is known as 'focus'. In tonal languages it appears that this raising or lowering of the pitch to indicate focus is either absent or much less noticeable.[168]
A number of studies have examined how focus is expressed in Chichewa and whether it causes a rise in pitch.[169] One finding was that for most speakers, focus has no effect on pitch. For some speakers, however, it appears that there is a slight rise in pitch if a word with a tone is focussed.[170] A toneless word, when in focus, does not appear to rise in pitch.
A different kind of emphasis is emphasis of degree. To show that something is very small, or very large, or very distant, a Chichewa-speaker will often raise the pitch of his or her voice considerably, breaking the sequence of downdrift. For example, a word such as kwámbíri 'very much' or pang'óno 'a little' is sometimes pronounced with a high pitch. The toneless demonstrative uyo 'that man' can also acquire a tone and become úyo! with a high pitch to mean 'that man over there in the distance'.[55]
Tonal minimal pairs
Sometimes two nouns are distinguished by their tone patterns alone, e.g.
- mténgo 'tree' vs mtengo 'price'
- khúngu 'blindness' vs khungú 'skin'
Verbs can also sometimes be distinguished by tone alone:
- ku-lémera 'to be rich' vs ku-lémérá 'to be heavy'
- ku-pwéteka 'to hurt' vs ku-pwétéká 'to be hurt'
There is also a distinction between:
- ndi 'it is' vs ndí 'and, with'
However, minimal pairs of this kind which differ in lexical tone are not particularly common.
More significant are minimal pairs in verbs, where a change of tones indicates a change in the tense, or a difference between the same tense used in a main clause and in a subordinate clause, for example:
- ndí-ma-werénga 'I usually read' vs ndi-ma-wérenga 'I was reading'
- sá-na-bwére 'he did not come' vs sa-na-bwére 'he has not come yet'
- ndi-kaná-pita 'I would have gone' vs ndí-kana-píta 'if I had gone'
- chaká chatha 'the year has finished' vs chaká chátha 'last year'
- ndí-kana-dzíwa 'if I had known' vs ndi-kaná-dziwa 'I would have known'
Reduplicated words
Reduplicated words are those in which an element is repeated, such as chipolopolo 'bullet'. The tones of these have been extensively studied in the literature.[171]
Reduplication in nouns
In nouns, the two elements join as follows (note that hyphens have been added here for clarity, but are not used in the standard orthography of Chichewa).
LL + LL becomes LLLL (i.e. there is no additional tone):
- chi-ng'ani-ng'ani 'lightning'
- chi-were-were 'sex outside marriage'
LH + LH becomes LHLL (i.e. the second tone is dropped):
- chi-masó-maso 'adultery'
- m-lengá-lenga 'sky, atmosphere'
- chi-bolí-boli 'wooden carving'
HL + HL becomes HLLH (or HHLH), by 'bumping':
- chi-láko-lakó (or chi-lákó-lakó) 'desire'
- chi-thúnzi-thunzí (or chi-thúnzí-thunzí) 'picture'
- a-múna-muná (or a-múná-muná) 'real men'
Reduplication in adverbs
When adverbs are reduplicated, however, and there is an element of emphasis, the first two types have an additional tone. Thus:
LL + LL becomes LLHL (i.e. there is an additional high tone on the second element):
- bwino-bwíno 'very carefully'
- kale-kále 'long ago' (or 'far off in the future')
LH + LH is also different when emphatic, becoming LHHH (or in the Southern Region HHHH):
- kwení-kwéní (or kwéní-kwéní) 'really'[172]
- okhá-ókhá (or ókhá-ókhá) 'only, exclusively'
When a three-syllable element is repeated, there is no special change:
- pang’óno-pang’óno 'gradually'
- kawíri-kawíri 'often'
Reduplication in verbs
A high tone following a proclitic tense-marker does not repeat when the verb is reduplicated:[173]
- ku-thándiza-thandiza 'to help here and there'
However, a final or penultimate tone will usually repeat (unless the verb has only two syllables, in which case the middle tone may be suppressed):[174]
- ti-thandizé-thandizé 'let's help here and there'
- ndí-ma-thandíza-thandíza 'I usually help here and there'
- á-ma-yenda-yénda 'they move about here and there'
Reduplication in ideophones
Ideophones (expressive words) have slightly different types of reduplication. Moto (1999) mentions the following types:
All high:
- góbédé-góbédé (noise of dishes clattering)
- bálálá-bálálá (scattering of people or animals in all directions)
All low:
- dzandi-dzandi (walking unsteadily)
High on the first syllable only:
- chéte-chete (in dead silence)
See also
References
- Downing & Mtenje (2017), pp. 132–4.
- Hyman (2009).
- Kanerva (1990), pp. 12-14.
- Clark (1988), pp. 51ff.
- Hyman & Mtenje (1999b), p. 122f.
- Downing, L.J. and Al Mtenje (2017), The Phonology of Chichewa (OUP), Chapter 7.
- Downing & Mtenje (2017), p. 190.
- Downing et al. (2004), p. 174.
- cf. Louw (1987), vol. 3, pp. 22, 60
- Moto (1983), p. 206; Mtenje (1986), p. 206; Stevick et al. (1965), p. 20; Louw (1987), vol. 3, p. 3.
- Mchombo (2004), p. 11.
- Mtanthauziramawu wa Chinyanja (c.2000), p. vi.
- e.g. Kulemeka (2002), p. 15.
- Myers (1998).
- Hyman (2000).
- Watkins (1937).
- Stevick, Earl et al. (1965) Chinyanja Basic Course Archived 2016-03-07 at the Wayback Machine
- Scotton & Orr (1980) Learning Chichewa.
- Mtanthauziramawu wa Chinyanja (c.2000); Kishindo (2001), pp. 277-79; Kamwendo (1999).
- Mtenje (2001).
- Myers (1996); see also Myers (1999a).
- Yip (2002), p. 148.
- Myers (1996), pp. 36ff.
- For the latter term cf. Downing & Mtenje (2017).
- Cf. Mtenje (1986), p. 240 vs. Mtenje (1987), p. 173.
- Myers (1999b).
- Kanerva (1990), p. 59; Mchombo (2004), p. 22.
- Moto (1983), p. 207.
- Moto (1983), p. 204.
- Mchombo (2004), p. 24.
- Downing & Mtenje (2017), p. 193, 194.
- cf. Stevick et al. (1965), p. 194.
- Louw, vol. 3, s.v. nyenyezi.
- Kanerva (1990), pp. 151-215.
- Hyman & Mtenje (1999a), p. 100.
- Moto (1983).
- Mchombo (2004), p. 17.
- Stevick et al. (1965), p. 111.
- Mtenje (1986), p. 240.
- Hyman & Mtenje (1999a), p. 104.
- Hyman & Mtenje (1999a), p. 126.
- Trithart (1976), p. 267.
- Kanerva (1990), p. 71, 66, 70; cf. Stevick p. 111.
- Stevick et al. (1965), p. 83.
- Hyman & Mtenje (1999b), p. 121.
- Kanerva (1990), p. 39.
- cf. MWC, entry 'nyenyezi'.
- Stevick et al. (1965), p. 119.
- Stevick et al. (1976), p. 176.
- Stevick et al. (1965), p. 155, cf. p. 101-2, 214.
- Kanerva (1990), p. 175.
- Stevick et al. (1965), pp. 69, 101.
- Kanerva (1990), p. 177.
- Stevick et al. (1986), pp. 248-9.
- Kulemeka (2002), p. 91.
- Stevick et al. (1965), p. 163, 165.
- Moto (1983), pp. 204f.
- Moto (1999).
- Hyman & Mtenje (1999b), p. 122f.
- Kanerva (1990), pp. 16-17; see Hyman & Mtenje (1999b), p. 127.
- Hyman & Mtenje (1999b), p. 124.
- Mtenje (1986), pp. 169, 206f.
- Hyman & Mtenje (1999b), p. 135.
- Mtenje (1987).
- Downing & Mtenje (2017), pp. 132-204.
- Downing & Mtenje (2017), pp. 140, 171.
- Maxson (2011), p. 86.
- Downing & Mtenje (2017), pp. 147–9.
- Mtenje (1986), p. 272.
- Mtenje (1986), pp. 203–4.
- Mtenje (1986), p. 195; (1987), p. 173.
- Maxson (2011), p. 79; Kanerva (1990), p. 22.
- Mtenje (1987), p. 174
- Downing & Mtenje (2018), p. 165.
- See the voice-track in Myers (1986), p. 83.
- Hyman & Mtenje (1999a), p. 99.
- Mtenje (1995), p. 7n.
- Hyman & Mtenje (1999a), p. 108.
- But see Downing & Mtenje (2017), p. 154.
- Downing & Mtenje (2017), p. 154.
- Downing & Mtenje (2017), p. 155.
- Kanerva (1990), p. 20.
- Kanerva (1990), p. 25.
- Hyman & Mtenje (1999a), p. 99.
- Hyman & Mtenje (1999a), p. 101.
- Hyman & Mtenje (1999a), p. 103.
- Hyman & Mtenje (1999a), p. 126.
- For different negative tenses see Mtenje (1986), p. 244ff; Mtenje (1987), p. 183ff; Kanerva (1990), p. 23; Downing & Mtenje (2017), pp. 174–195.
- Mtenje (1987), p. 183.
- Mtenje (1987), p. 203.
- Mtenje (1986), p. 248.
- See discussion in Mtenje (1987), p. 203.
- Downing & Mtenje (2017), p. 177.
- Mtenje (1987), p. 203.
- cf. Stevick et al. (1965), p. 124, p. 243.
- Downing & Mtenje (2017), p. 187.
- Downing & Mtenje (2017), p. 181.
- Downing & Mtenje (2017), p. 174.
- Mtenje (1986), p. 245.
- Mtenje (1986), p. 248.
- Mtenje (1987), p. 183.
- Downing & Mtenje (2017), p. 186.
- Stevick et al. (1965), p. 222.
- Downing & Mtenje (2017), p. 189.
- Downing & Mtenje (2017), p. 194.
- Stevick et al. (1965), p. 147.
- Mchombo (2004), pp. 17–18.
- Downing & Mtenje (2011), p. 9; Stevick et al. (1965), p. 148.
- Stevick et al. (1965), p. 196.
- But cf. Stevick (1965), p. 159.
- Downing & Mtenje (2017), pitchtrack p. 259.
- Stevick et al. (1965), p. 168.
- Downing & Mtenje (2017), p. 166.
- Downing & Mtenje (2017), p. 199.
- Stevick et al. (1965), p. 148.
- Downing & Mtenje (2011).
- Maxson (2011), p. 86.
- Stevick et al. (1965), p. 291.
- Stevick et al. (1965), p. 168.
- Chakanza, J.C. (2000). Wisdom of the People: 2000 Chinyanja Proverbs. CLAIM Blantyre (Malawi) p. 241.
- Stevick et al. (1965), p. 287.
- Maxson (2011), p. 79; Stevick et al. (1965), p.1.
- Mpanje (1983), p. 141; Stevick et al. (1965), p. 156.
- Watkins (1937), p. 100.
- Hyman & Mtenje (1999a), pp. 94f.
- Mtenje (1995), p. 7n.
- Mtenje (1986), p. 245.
- Mtenje (1986), p. 246.
- Hyman & Mtenje (1999a), p. 126.
- Mtenje (1986), p. 274.
- Hyman & Mtenje (1999a), p. 101.
- Kanerva (1990) p. 44 (with pausal form).
- Downing & Mtenje (2017), pp. 166–7.
- Kanerva (1990), p. 24.
- Downing & Mtenje (2017), p. 143.
- Stevick et al. (1965), p. 197.
- Kanerva (1990), p. 33.
- Stevick at al. (1965), p. 264.
- Hyman & Mtenje (1999a), p. 101.
- Downing & Mtenje (2017), p. 189.
- Mtenje (1986), pp. 240-41.
- Hyman & Mtenje (1999a), p. 126.
- Hyman & Mtenje (1999a), p. 102.
- Mtenje (1995), p. 7; Stevick et al. (1965), p. 222.
- Kanerva (1990), p. 147.
- Mtenje (1995), p. 23; Kanerva (1990), p. 44.
- Hyman & Mtenje (1999a), pp. 103, 127.
- Myers (1996), pp. 29-60.
- Myers (1996), p. 34
- Kanerva (1990), p. 147.
- Kanerva (1990), pp. 138ff.
- Kanerva (1990), p.138; Stevick et al. (1965), p. 21.
- Stevick et al. (1965), p. 201.
- Stevick et al. (1965), p. 15.
- Stevick et al. (1965), p. 31.
- Stevick et al. (1965), p. 26.
- Stevick et al. (1965), p. 12.
- Stevick et al. (1965), pp. 254-6.
- Stevick et al. (1965), p. 268.
- cf. Myers (1996), p. 35; Hullquist, C.G. (1988) Simply Chichewa, p. 145.
- Downing (2017), p. 382; Downing (2008), p. 61; cf. Hullquist, C.G. (1988) Simply Chichewa, p. 145.
- Stevick et al. (1965), p. 34, 48, 53, 54.
- Stevick et al. (1965), pp. 225, 19.
- Downing (2017), p. 382
- Stevick et al. (1965), pp. 42, 47, 75, 119.
- See the voicetrack in Downing & Mtenje (2017), 264.
- Stevick et al. (1965), p. 176.
- Cruttenden (1986), pp. 10ff; Downing (2008).
- Kanerva (1990); Myers (1996); Downing (2004); Downing (2008).
- Downing & Pompino-Marschall (2011).
- Mtenje (1988); Kanerva (1990), pp. 37, 49-54; Hyman & Mtenje (1999a), pp. 107-124; Myers & Carleton (1996); Moto (1999).
- Stevick et al. (1965), p. 66.
- Hyman & Mtenje (1999a), pp. 107, 114.
- Stevick et al. (1965), p. 301.
Bibliography
- Botne, Robert D. & Andrew Kulemeka (1991). A learner's Chichewa-English, English-Chichewa dictionary. Köln: Rüdiger Köppe Verlag.
- Clark, Mary (1988). "An Accentual Analysis of the Zulu Noun", in Harry van der Hulst & Norval Smith, Autosegmental Studies in Pitch Accent. Dordrecht: Foris.
- Cibelli, Emily (2012). "The Phonetic Basis of a Phonological Pattern: Depressor Effects of Prenasalized Consonants". UC Berkeley Phonology Lab Annual Report, 55-71. (Now published in Joaquín Romero, María Riera (2015) The Phonetics–Phonology Interface: Representations and methodologies, 335, 171-192.)
- Cruttenden, Alan (1986). Intonation. Cambridge University Press.
- Downing, Laura J. (2008). "Focus and Prominence in Chichewa, Chitumbuka, and Durban Zulu", ZAS Papers in Linguistics 49, 47-65.
- Downing, Laura J., Al D. Mtenje, & Bernd Pompino-Marschall (2004). "Prosody and Information Structure in Chichewa", ZAS Working Papers in Linguistics 37, 167-186.
- Downing, Laura J. & Al D. Mtenje (2010). "The prosody of relative clauses in Chichewa". Workshop on Relative Clauses in Bantu Languages LAM, Paris, 8-9 January 2010.
- Downing, Laura J. & Al D. Mtenje (2011). "Un-WRAP-ping prosodic phrasing in Chichewa", in Nicole Dehé, Ingo Feldhausen & Shinichiro Ishihara (eds.). The Phonology-Syntax Interface, Lingua 121, 1965–1986.
- Downing, Laura J. & Bernd Pompino-Marschall (2013). "The focus prosody of Chichewa and the Stress-Focus constraint: a response to Samek-Lodovici (2005)". Natural Language and Linguistic Theory, Volume 31, Issue 3, pp 647–681.
- Downing, Laura J. (2017). "Tone and intonation in Chichewa and Tumbuka". In Laura J. Downing & Annie Rialland (eds) Intonation in African Tone Languages. de Gruyter, Berlin/Boston, pp. 365–392.
- Downing, Laura J.; Mtenje, Al (2017). The Phonology of Chichewa. Oxford University Press.
- Hyman, Larry M. & Al D. Mtenje (1999a). "Prosodic Morphology and tone: the case of Chichewa" in René Kager, Harry van der Hulst and Wim Zonneveld (eds.) The Prosody-Morphology Interface. Cambridge University Press, 90-133.
- Hyman, Larry M. & Al D. Mtenje (1999b). "Non-Etymological High Tones in the Chichewa Verb", Malilime: The Malawian Journal of Linguistics no.1.
- Hyman, Larry M. (2000). "Privative Tone in Bantu". Paper presented at the Symposium on Tone ILCAA, Tokyo, December 12–16, 2000.
- Hyman, Larry M. (2007). "Tone: Is it Different?". Draft prepared for The Handbook of Phonological Theory, 2nd Ed., Blackwell (John Goldsmith, Jason Riggle & Alan Yu, eds)
- Hyman, Larry M. (2009). "How (not) to do phonological typology: the case of pitch-accent". Language Sciences 31, 213–238.
- Kamwendo, Gregory H. (1999). "Work in Progress: The Monolingual Dictionary Project in Malawi", Malilime: The Malawian Journal of Linguistics no.1.
- Kanerva, Jonni M. (1990). Focus and Phrasing in Chichewa Phonology. New York, Garland.
- Kishindo, Pascal, 2001. "Authority in Language: The Role of the Chichewa Board (1972–1995) in Prescription and Standardization of Chichewa". Journal of Asian and African Studies, No. 62.
- Kulemeka, Andrew T. (2002). Tsinde: Maziko a Galamala ya Chichewa, Mother Tongue editions, West Newbury, Massachusetts.
- Louw, Johan K. (1987). Pang'onopang'ono ndi Mtolo: Chichewa: A Practical Course. UNISA Press.
- Maxson, Nate (2011). Chicheŵa for English Speakers: A New and Simplified Approach. Assemblies of God Literature Press, Malawi.
- Mchombo, Sam (2004). Syntax of Chichewa. Cambridge University Press.
- Mchombo, Sam; Moto, Francis (1981). "Tone and the Theory of Syntax". In W.R. Leben (ed.) Précis from the 12th Conference on African Linguistics, Stanford University, pp. 92–95.
- Moto, Francis (1983). "Aspects of Tone Assignment in Chichewa", Journal of Contemporary African Studies, 3:1.
- Moto, Francis (1999). "The tonal phonology of Bantu ideophones", in Malilime: The Malawian Journal of Linguistics no.1, 100-120.
- Mtanthauziramawu wa Chinyanja/Chichewa: The first Chinyanja/Chichewa monolingual dictionary (c2000). Blantyre (Malawi): Dzuka Pub. Co. (Also published online at the website of the Centre for Language Studies of the University of Malawi.)
- Mtenje, Al D. (1986). Issues in the Non-Linear Phonology of Chichewa part 1. Issues in the Non-Linear Phonology of Chichewa part 2. PhD Thesis, University College, London.
- Mtenje, Al D. (1987). "Tone Shift Principles in the Chichewa Verb: A Case for a Tone Lexicon", Lingua 72, 169-207.
- Mtenje, Al D. (1988). "On Tone and Transfer in Chichewa Reduplication". Linguistics 26, 125-55.
- Mtenje, Al D. (1995). "Tone Shift, Accent and the Domains in Bantu: the Case of Chichewa", in Katamba, Francis. Bantu Phonology and Morphology: LINCOM Studies in African Linguistics 06.
- Mtenje, Al D. (2001). Comparative Bantu On-Line Dictionary (CBOLD) Chichewa Dictionary. This is an abridged version of Clement Scott and Alexander Hetherwick's 1929 Dictionary of the Nyanja Language, with tones added by Al Mtenje.
- Myers, Scott (1988). "Surface Underspecification of Tone in Chichewa", Phonology, Vol. 15, No. 3, 367-91.
- Myers, Scott (1996). "Boundary tones and the phonetic implementation of tone in Chichewa", Studies in African Linguistics 25, 29-60.
- Myers, Scott (1999a). "Downdrift and pitch range in Chichewa intonation", Proceedings of the 14th International Congress of Phonetic Sciences, vol. 3, pp. 1981–4.
- Myers, Scott (1999b). "Tone association and f0 timing in Chichewa", Studies in African Linguistics 28, 215-239.
- Myers, Scott (2004). "The Effects of Boundary Tones on the f0 Scaling of Lexical Tones". In Proceedings of the 4th International Symposium on Tonal Aspects of Languages 2014, pp. 147–150.
- Myers, Scott & Troi Carleton (1996). "Tonal Transfer in Chichewa". Phonology, vol 13, (1) 39-72.
- Nurse, Derek & Gérard Philippson (2006). "Common tense-aspect markers in Bantu". Journal of African Languages and Linguistics. 27 (2): 155–196.
- Scotton, Carol Myers & Gregory John Orr, (1980). Learning Chichewa, Bk 1. Learning Chichewa, Bk 2. Peace Corps Language Handbook Series. Peace Corps, Washington, D.C.
- Stevick, Earl et al. (1965). Chinyanja Basic Course. Foreign Service Institute, Washington, D.C. Archived 2016-03-07 at the Wayback Machine
- Trithart, Lee (1976). "Desyllabified noun class prefixes and depressor consonants in Chichewa", in L.M. Hyman (ed.) Studies in Bantu Tonology, Southern California Occasional Papers in Linguistics 3, Los Angeles, CA: University of Southern California, 259-86.
- Watkins, Mark Hanna (1937). A Grammar of Chichewa: A Bantu Language of British Central Africa, Language, Vol. 13, No. 2, Language Dissertation No. 24 (Apr.-Jun., 1937), pp. 5–158.
- Yip, Moira (2002). Tone. Cambridge University Press.