Katie Sickles
LING 567
lab5 write-up

	Icelandic has both case and agreement so I chose the first package.
The first thing I did was create a noun-lex type heirarchy. The common-
noun-lex is restricts the common nouns to third person and INFLECTED -
(for lexical rules). Then common nouns are broken down into the genders
fem, masc, and neut. Each of these restricts the GEND value to the
gender.
	Icelandic has a complex noun declension system. There are different
inflections for each case and number for each gender. There are four
classes with different inflection in each gender as well as a strong
and a weak declension. This could be represented as an ALTS feature.
However, for the sake of clarity, I have chosen to break down noun
types instead. I figured that this would be easier to debug and 
understand conceptually.
	For each gender I have the four classes as subtypes. I have chosen
to ignore the strong/weak declension since this does not appear in my
test suite. In the masc noun classes, I have chosen a middle step with
m-cl1to3-noun-lex since classes 1-3 have the same nom and acc declensions
for singular nouns.
	For the case lexical rules, I only focused on nom and acc since no
example of dative or genative is in my test suite. This is also the
reason why I have not written any plural lexical rules yet. The number
are both reflected in the suffix, so they are combined in the lexical
rules.
	All of the case lexical rules are working except for the class 4
masc noun which has a complex declension. The class 4 masc noun I have
in my test suite is "mašur" which is "man". When this word is in the 
acc form (as well as dat and gen), "mann" is the root. I have not yet
decided if I should I make "mann" the base root and have "mann-ur" be
the nom form as a morpheme transciption sytle.
	Icelandic has a complex verb declension system as well. However,
all the verbs in my test suite conjugate the same, so I only wrote
lexical rules to cover this one type of declension. The subject-verb
agreement is working for all cases.
	I have chosen to save dealing with determiners for a later lab
since they are complex in Icelandic. There are free determiners as
well as suffixed ones, and each appear in certain cases. I may decide
not to represent the free determiner since it is rarer and the more
specialized case.

note:	I have been having a lot of trouble with emacs/tsdb on linux.
I have tried several distributions and am working with someone familiar
to linux, but have not been able to get it running yet. I have included
a batch parse from the most recent grammar. It is the test.results
file in the matrix folder. I will make it to the treehouse sometime on
Monday to run the actual test suite and email the tsdb part in late.