User:Janadume/sandbox/Swahili Grammar

Swahili Grammar is typical for a Bantu language, including a rich array of noun classes and agglutinativity, especially on its verbs. Swahili is, nevertheless, somewhat simpler than many other Bantu languages, possibly as a result of its former and continued use as a lingua franca.

Note on glossing
Glossing within this article is as per the Leipzig Glossing Rules with some additions. Note that the hyphens in the glossed examples are for the purpose of glossing only and are not used when writing Swahili. For example, "freedom of speech" is written uhuru wa kusema, and not *uhuru w-a ku-sema as given below.
 * "extention" refers to a verbal stem extention (ku- or kw-) which occurs in some situations with short verb stems.
 * "final vowel" refers to the final vowel -a present in all native Bantu verbs, which on its own means nothing specific but can be swapped with other vowels to indicate changes in tense, voice, polarity etc.
 * "positive"
 * "negative"

Nouns
As with other Bantu languages, Swahili nouns are grouped into noun classes (which may also be referred to as grammatical genders). Noun classes generally share many semantic features in common, although this is not a reliable means of determining to which class a noun belongs. A noun's class can frequently be identified by its prefix although nouns in a few classes frequently lack an identifying prefix. In other cases, a word may look like it belongs to one class when actually it belongs to another, such as mbwa 'dog', which may be mistaken for a class 1 noun (which would mean its plural would be in class 2 as *wabwa) but is actually a class 9 noun with the identical class 10 plural mbwa, or chembe 'particle', which looks like it may be a class 7 noun with the class 8 plural *vyembe although it too, is a class 9 noun with an identical plural.

In the following table, the number given to each noun class comes from the standard numbering of reconstructed proto-Bantu noun classes. This means that the noun classes given here can be directly compared with the noun classes of other Bantu languages. It also means that there are some missing numbers where Swahili lacks classes found in other Bantu languages.

Classes
Swahili noun classes are frequently described by their prefixes they take in singular and plural. For example, mtu 'person' has the plural form watu 'people', thus belonging with other nouns of this pattern to the "M-WA" class. Kitu 'thing' and its plural form vitu, likewise belong to the "KI-VI" class. While these class names are very easy for learners to remember, a more linguistically useful strategy is to use the numbering system for noun classes in Bantu languages which was devised by linguist Carl Meinhof. This system takes into account nouns with plural and singular forms in unexpected classes and can also be used to compare cross-linguistically with other Bantu languages. Proto-Bantu, the language from which all Bantu languages are descended, is believed to have had 22 noun classes, although no modern Bantu language preserves all of these. In standard Swahili, classes 12 and 13 are no longer present. Neither are those above 18. Classes 11 and 14 are now all but indistinguishable from one another.

Noun classes in Swahili up to 10 can be regarded as pairs, with the even numbers representing the plural class of the preceding odd number. Thus, class 2 forms the plural of class 1, class 8 forms the plural of class 7 etc. Classes with numbers higher than 10 are a little more irregular and may have no plural form, a plural form from a lower class, or may indicate gerunds or locations.

1 and 2 (M-WA)
Classes 1 and 2 are made up almost exclusively by nouns for people.

There is a very small number of nouns for animals in this class:

Many nouns in this class are derived from another noun or a verb. When the stem of a noun begins with a vowel, the class one prefix is mw-. In the class 2 plural form, the m- is frequently simply dropped. This applies to all nouns beginning with mwana, which on its own means 'offspring' but when compounded essentially means person.

Note that the plural form of mwanamke is wanawake, containing the plural form of mke 'wife'.

In other cases, the mw- is dropped and wa- added. Many of these cases derive from verbs.

Nouns whose stems begin with u may often be written with either mw- or mu-.

Genitive
The genitive, sometimes called "possessive" or "associative" is a construction involving two nouns in which the first noun, the "possessum" is indicated as being either owned by or in some way associated with the second noun, "the possessor". The possessor is introduced by the preposition -a, roughly equivalent to 'of', which takes a consonant prefix which agrees (i.e. changes according to) the noun class and animacy of the possessum. With the sole exception of class 1, these prefixes are identical to the verbal prefixes which occur before the aorist -a- prefix.

Genitive with animate possessum
When the possessum in a genitive construction is an animate noun (referring to a person or an animal), the genitive prefix w- is generally used. This follows the trend in the rest of the language whereby animate nouns, regardless of their noun class, are handled syntactically as though they belonged to the class 1/2.

A notable exception is with the animate nouns of the class 9/10, which have a plural form which is identical to the singular. In genitive constructions, these nouns usually retain the y- and z- forms associated with their own noun class which results in fewer cases where there is ambiguity as to whether the noun is singular or plural. When these nouns refer to animals, they may use w- in the singular and z- in the plural.

Genitive of personal pronouns
The personal pronouns have special genitive stems which are used with the same set of prefixes as the genitive preposition -a. For example:

Personal pronouns
Personal pronouns mark for person and number but not gender. Aside from wao 'they', all the full forms of personal pronouns involve a syllable reduplication. In the case of the second person plural, nyinyi, the variant ninyi is common. In certain circumstances, generally in informal speech when pronouns are unstressed, this reduplication is dropped. There are also notable variations present, such as mie for standard mimi.

With certain particles such as na 'and', 'with', when unstressed, a short form of the pronoun may be tacked onto the end causing the na to take the word stress.

The genitive forms require a prefix that agrees with the preceding noun, as outlined above in the section on the genitive construction.

The following table outlines the forms. Swahili is a pro-drop language, with verbal slots for subject and object, meaning that, in situations where the pronoun does not need to be emphasised, it is simply omitted and the information it conveys left up to the verb. There are, however, some verbal constructions (such as the present tense copula ni, the present tense negative copula si and the habitual verb form with hu-) which do not allow subject marking and in these instances, pronouns appear.

In the first of the following examples, the verb does not indicate the grammatical person or number, necessitating the pronoun. In the second, the verb indicates the grammatical person and number, meaning that the pronoun is only used in order to emphasise "we".

Demonstrative pronouns
Swahili's demonstrative pronouns come in three types, termed here "proximate", "medial" and "distal" although there is some controversy about what conditions their use, particularly that of the "medial" demonstratives. Some authors describe it as a three way distance distinction, with the proximate demonstratives being close to the speaker, the medial demonstratives being close to the listener and the distal demonstratives being far from both. Other authors describe the "medial" set as "referential", not specifying distance but referring back to something previously mentioned.

Verb Types
Swahili verbs come in essentially three types which inflect slightly differently:


 * 1) Long Bantu verbs
 * 2) Short Bantu verbs
 * 3) Foreign verbs

The long Bantu verbs are the largest group. They all end in -a which, in certain forms, changes to an -e (subjunctive) or an -i (present negative). Some examples of this group are -penda 'love/like', '-tazama 'watch' and -fundisha 'teach'.

Short Bantu verbs are a small group of verbs with a monosyllabic stem, such as -la 'eat', -nywa 'drink', -ja 'come' and -wa 'be'. Because they all end in the Bantu verbal suffix -a, the stem of each of these verbs is generally a single consonant, such as -l- 'eat'. The stem of drink is -nyw-, consisting of two consonants ('ny' + 'w' is /ɲw/). These verbs are extended with a prefixed -ku- in certain situations.

The group called 'foreign verbs' are recognisable because they do not end in -a. They are largely, but not exclusively of Arabic origin. Examples include -sahau 'forget', -ishi 'live' and -ripoti 'report', the latter of which comes from English, not Arabic. These verbs, lacking the -a suffix, do not have anything that can undergo the change to -i or -e that the other verbs have.

Additionally, there are the verbs -enda 'go' and -isha 'stop/finish' which are a somewhat borderline case. Although they are longer than the other short Bantu verbs and contain a vowel, they may behave as the short verbs, optionally taking the extention prefix -kw- in the same situations as the other short verbs take -ku-.

Infinitive
The infinitive of all Swahili verbs begins with ku- or, with a handful of verbs, kw-. For example: kuwa 'to be', kula 'to eat', kwenda 'to go' (or kuenda), kwanza 'to begin' (or kuanza), kwisha 'to end' (or kuisha), 'to stop', kuandika 'to write', kupika 'to cook', kupenda 'to like', 'to love'.

The negative form of the infinitive is formed with kuto-, for example kutopika 'not to cook'. This -to- cannot be stAoristThe aorist and thus triggers the extention morphemes appearance in the short verbs, such as kutokula 'not to eat', kutokuwa 'not to be' and kuto(kw)enda 'not to go'.

Dictionaries of Swahili generally list verbs using only the stem, with or without a preceding hyphen, so that all the verbs of the language are not arranged together according to alphabetical order. For example: -wa 'be', -la 'eat', -enda 'go', -anza 'begin', -isha 'end, stop', -andika 'write', -pika 'cook', -penda 'like, love'. Confusingly, however, some dictionaries and learning materials list the short verbs with their extention morpheme and a hyphen, such as -kula.

Verbal concord
Finite verb forms (with the exception of the infinitive and habitual forms) include a prefix indicating the subject of the verb by referencing, if animate, its number and person, or if inanimate, its noun class. There are also prefixes for objects which, in most situations, are not obligatory. Additional affixes can be used to turn the verb into a relative clause and these also depend on noun class and animacy. The forms are indicated in the following tables.

Tense, Aspect, Mood
Tense, aspect and mood (TAM) markers appear primarily in a slot after the subject and before any relative marker or object. A commonly used mnemonic is S-T-R-O-V-E, standing for Subject - Tense - Relative - Object - Verb - Extention. There is a large amount of interplay between TAM and polarity (positive or negative) and the marking of TAM may also extend to a changing of the final -a. Verbs which are loanwords from other languages, however, which are mostly from Arabic, may not end in -a and when this is the case, the ending does not change. The term tense will henceforth be used for simplicity to describe TAM combinations for the sake of simplicity, although it should be remembered that these are varying mixes of tense, aspect and mood.

Habitual
The habitual form of verbs  indicates a repeated or habitual action. It is formed with the prefix hu- which does not allow a preceding subject concord. For this reason, pronouns are needed to indicate the subject when this is not done by a noun. The syllable hu- is able to receive the word stress and thus the stem extention -ku- does not appear with short verbs.

There is no negative form of this TAM combination, however a fairly similar meaning can be achieved if necessary by preceding a present negative verb with huwa, which is the habitual form of -wa 'be'.

Aorist
The following table shows the main verbal structures in regard to TAM-marking and other verbal forms. In the table, an initial hyphen shows where the subject concord appears. The negative marker ha- shown here interacts with the singular, animate subject concords (ha- + ni- = si-, ha- + u- = hu-, ha- + a- = ha-) the form of the final vowel is indicated. For verbs not ending in -a in the infinitive, this is absent. The underlined vowels in the table cannot be stressed and where they appear before a short verb stem, they trigger the stem extention -ku- to appear, which takes the stress.