Preliminary comments

Zulu has 14 noun classes.  Every noun class has a prefix that attaches
to nouns of that class. One noun stem can take different
prefixes. Some of the classes are singular/plural pairs, e.g. class 4
nouns are plurals of corresponding class 3 nouns. However some prefix
changes are more derivational. So, 'umu-ntu' means 'person, 'aba-ntu'
means 'people', and 'ubu-ntu' means 'humaneness'.  Verbs are required
to agree with the noun class of their subject, by means of prefixes
called subject concords.  There are also optional object concords.
These noun classes appear across Bantu languages, and a standardized
numbering system (Class 1, Class 2, etc.) has been created to identify
corresponding classes. I have labeled the noun class prefixes as C#
where # is the class number, e.g. C1, C2.  Subject concords on verbs
are SC#, Object concords are OC#.  First and second person subject
concords are specified with the person and number information first
(e.g. 1SG.SC). It is not to late to change this notation if there is a
better strategy.  NB: Because the numbering system is common across
all Bantu languages, the Zulu class numbers are not completely linear.
Zulu does not have classes 12 and 13, and 16 only appears in some
fossilized locative terms.

My examples are all in a future tense that specifically refers to the
immediate future.  Most of the other tenses either had variation
attached to whether the verb was followed by an object or adverb, or
had some phonological variation that I decided not to deal with if I
didn't have to.  One challenge that I had was in deciding how to
analyze the verb roots.  One of my grammars and the dictionary I have
both analyze the roots as including the final 'a' (e.g. -bona,
'see'). Under this analysis, the future tense of 'see' would be:

SC-zo-bona

where SC is the Subject Concord. The problem is verb forms where the
final 'a' changes.  The grammar mentioned above says that in those
cases the final 'a' is dropped with .  The dictionary lists the other
forms under separate entries. The other grammar analyzes the root as
being without the final 'a', analyzing the future tense of see as:

SC-zo-bon-a

This takes care of verb forms where the ending is different. However,
this grammar does not say what that final -a might represent; it calls
it simply "the suffix" (where -zo- is the future tense marker). It
also seems like a complex analysis for a simple grammar like the one
I'm building, especially considering I'm using an a-ending verb form.
I think the second analysis is both more extensible and more
linguistically satisfying, so I decided to use it and label the final
-a as though it were part of a circumfix with the tense marker.  I am
not convinced this was the right decision, but it's the one that
appears in these examples.

The standard orthography for Zulu does not include notation for tone,
and so I have not included it.

My grammars have proven to be unhelpful in finding examples.  The
grammar's examples are not in future tense, or only use first or
second person arguments, or are full of adverbials or other extraneous
information that I am not ready to include in my grammar.  Most of my
sample sentences are modeled after samples provided in the text, but
none are taken word-for-word.  The site isiZulu.net has a message
board that appears to be frequented by native speakers of Zulu; I
think I will try asking there for grammaticality judgments.


Basic Word Order

Zulu is an SVO language.  While there are many circumstances where it
is possible to omit arguments, as far as I can tell they cannot appear
in any other order. One of my ungrammatical examples is OVS, which
could in theory be mistaken for SVO, because there is no case
marking. The example is actually ungrammatical, but only because the
subject concord on the verb does not match the noun class of the first
argument.  I have therefore marked it as ungrammatical and listed the
phenomena as both word order and agreement.

I have not found any mention of ditransitive verbs, so I don't know if
there aren't any or if they just weren't covered in my grammars.
Until I find out, I am not including any ditransitive test data,
because I can't make grammaticality judgments.

Pronouns

Zulu has pronouns, but my grammars give the impression that they are
not used much. One of my grammars did not mention them at all; the
other never used them outside of the section titled "Pronouns".
Instead, the subject and object concords on the verb are used to
convey information about the arguments.  Still, because there are
pronouns and they may be useful for agreement information later on, I
have included some pronoun examples.  I have made test sentences for 4
non-third person pronouns: 1sg, 1pl, 2sg, 2pl.  There is no case
distinction. I only used (what I think are) grammatical sentences.  I
did not include any third-person pronouns, as every noun class has its
own pronoun, and that seemed excessive and unnecessary for this stage
of the project.

The Rest of the NP

Zulu does not have mandatory determiners, but it does have some
optional demonstratives that I thought would fit here.  Demonstratives
have a 3-way deixis split: close to the speaker, close to the
addressee, and far from both. One of the grammars describes these as
Position 1, Position 2, and Position 3; I have used this notation in
my morpheme-by-morpheme glosses.  The demonstrative can appear either
before or after the noun (or alone as a pronoun, but I have not
included examples reflecting this). Demonstratives must agree with the
noun class of the word they modify, which for 14 noun classes means
there are 42 demonstratives.  To test all of them, both before and
after the noun, would be 84 sentences just to get the grammatical
examples.  Instead, I give the 6 grammatical constructions for a class
1 noun, followed by after-the-noun examples demonstrating the
agreement requirement using a class 9 noun and the class 1 noun from
before.  I think my strategy will be to make an agreement rule, and
then add more demonstratives as I add nouns from more classes.

Argument Optionality

Of the four patterns discussed in the assignment instructions, Zulu
best matches the Pronomial Affixes category. I mentioned before that
Zulu verbs have mandatory subject concords and optional object
concords that each agree with the noun class of the argument they
correspond to.  As long as the concord is present, it is fine to leave
the argument unexpressed.  Indeed, this appears to be the primary way
of marking first- and second-person arguments (as opposed to using
pronouns). The subject concord is mandatory whether the argument
appears or not.  The object concord is optional if the argument is
expressed (i.e. it can co-occur with an overt object, or on its own)
but is mandatory if the argument is not expressed.  It is possible for
a verb to appear with no overt arguments at all.

Agreement

I have already mentioned the subject and object concords on the verbs,
which are what I covered in this section. I previously touched on
demonstratives, and while those should be expanded, I just focused on
verbs here.

I did not include examples where the subject concord is the same for
two classes.  For example, the subject concord for both class 1 and
class 3 is u-.  Since I don't have a means of differentiating them
yet, it seemed silly to list the u- prefix twice.

Verbs only need to agree with noun class.  The subject concord and
object concord(if present) each need to agree with their respective
arguments.  Mixing e.g. every subject concord with every noun class
seems excessive, so I have chosen a less-complicated approach.  The
assignment instructions said that I needed a noun of every type and a
verb of every type, so I decided to hold either the argument class or
the concord class constant while varying the other. I used an
intransitive verb to make test sentences that show one noun class
subject with every subject concord, and every noun class with one
subject concord.  I then did the same for the objects, using a
transitive verb.  I threw in some first- and second- person examples
for good measure.  It was a lot of work, but it will help encourage me
to actually get all the noun classes working.

The Remaining Features

Zulu nouns don't show any case marking.  There are differences between
subject and object concords, however, so I was thinking of making some
examples to specifically highlight those differences instead.  It
could be that i've covered it sufficiently in the examples I have,
though.

All the other features I should be able to cover, although neither of
the grammar books I have talk about imbedded clauses or matrix yes-no
questions.  I will need to do some more legwork to get data for those.