Adam Ledyard and Sanae Sato
Lab 2: Miyako grammar basics


1. 	The choices made in the customization system, illustrated with examples from your 
	test suite. 


	As our first grammar of Ryukyuan (ISO: mvi), we have implemented following 
	features.

	  - Word Order
	    - SOV
	    - Determinar: yes
	    - The order of determinar: Det-Noun
	    - Aux verb: no
	  
	  - Number
	    - sg
	    - pl
	  
	  - Person
	    - First/Second/Third
	    - Inclusive/Exclusive in: 1sg-pl
	  
	  - Gender
	    - no gender
	  
	  - Case
	    - Nominative-Accusative
	      S&A  --> nom
	      O    --> acc  
	    - Additional Cases
	      dat
	      gen

	  - Other Feature
	    - PRONOUN
	      - a syntactive feature 
	      - nominal category
	      - use an existing value type: bool

	Although Ryukyuan pronoun has distinction in person and numbers in the surface 
	form. However there is no valence restriction in terms of person and number (i.e. 
	there is no subject-verb agreement), therefore we did not include that in our 
	test suite.


	(1) Transitive/Intransitive verb

	# Here is the example(s) form our test suite
	#-----------------------------
	#Example 1 Intransitive example
	Source: author
	Vetted: f
	Judgment: g
	Phenomena: wo
	pztu nu ff
	pztu=nu ff-0
	man=NOM come-NPST
	A man comes.

	#Example 2 Intransitive example
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: wo
	ff pztu nu
	ff-0 pztu=nu
	come-NPST man=NOM
	A man comes.

	#Example 3 Transitive example
	Source: author
	Vetted: f
	Judgment: g
	Phenomena: wo
	pztu nu aagu u naratar
	pztu=nu aagu=u nara-tar
	man=NOM song=ACC learn-PAST
	A man learned a song.

	#Example 4 Transitive example
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: wo
	naratar pztu nu aagu u
	nara-tar pztu=nu aagu=u
	learn-PAST man=NOM song=ACC
	Learned a man a song.

	#Example 5 Transitive example
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: wo
	naratar aagu u pztu nu
	nara-tar aagu=u pztu=nu 
	learn-PAST song=ACC man=NOM 
	Learned a song a man.

	#Example 6 Transitive example
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: wo
	aagu u naratar pztu nu
	aagu=u nara-tar pztu=nu 
	song=ACC learn-PAST man=NOM 
	a song learned a man.

	#Example 7 Transitive example
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: wo
	pzutu nu naratar aagu u
	pztu=nu nara-tar aagu=u 
	man=NOM learn-PAST song=ACC 
	a man learned a song.

	#Example 8 Transitive example
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: wo
	aagu u pztu nu naratar
	aagu=u pztu=nu nara-tar
	song=ACC man=NOM learn-PAST
	A man learned a song.

	#-----------------------------


	The sample sentence of intransitive verb ("a man comes") is made from two major 
	constituent (S & V).Therefore, 2 possible word orders:

	  S V   man=NOM come-NPST -- (Example 1)
	  V S  *come-NPST man=NOM -- (Example 2)

	Similary, the sample sentence of transitive verb ("a man learned a song") is made 
	from three major constituent (S O V).Therefore, 6 possible word orders:

	  SOV   man=NOM song=ACC learn-PAST -- (Example 3)
	  VSO  *learn-PAST man=NOM song=ACC -- (Example 4)
	  VOS  *learn-PAST song=ACC man=NOM -- (Example 5)
	  OVS  *song=ACC learn-PAST man=NOM -- (Example 6)
	  SVO  *man=NOM learn-PAST song=ACC -- (Example 7)
	  OSV  *song=ACC man=NOM learn-PAST -- (Example 8)

	Note that we assumed those "ungrammatical" sentencses are ungrammatical with our 
	limited reference/knowleadge. They could be correct sentences with native 
	speaker's intuition.



	(2) Cases

	# Here is the example(s) form our test suite
	#-----------------------------
	#Example 9 1SG Pronoun
	Source: author
	Vetted: f
	Judgment: g
	Phenomena: c
	ba ga iditar
	ba=ga idi-tar
	1SG=NOM exit-PAST
	I exited.

	#Example 10 1SG Pronoun
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: c
	ba nu iditar
	ba=nu idi-tar
	1SG=NOM exit-PAST
	I exited.

	#Example 11 1SG Pronoun
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: c
	ba u iditar
	ba=u idi-tar
	1SG=ACC exit-PAST
	I exited.

	#Example 12 1SG Pronoun
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: c
	ba ga iditar
	ba=ga idi-tar
	1SG=GEN exit-PAST
	I exited.
	#-----------------------------


	In Irabu Ryukyuan, there are 4 cases: 

	  CASE            | FORM    | EXAMPLE
	  ----------------+---------+---------------
	  nominative(nom) | =ga/=nu | ba-ga
	  genitive(gen)   | =ga/=nu | ba-ga
	  accusative(acc) | =u      | ba-a --> banu-
	  dative(dat)     | =n      | ba-n --> ban-

	There is a hierarchy of case particle usage, and pronoun only takes =ga as nom/
	gen form, and =nu elsewhere (although there are some minor distinctions, but it 
	was out of this lab2's scope).

	The sample sentences (Example 9 - Example 12) illustrate case usage of 1st person 
	singular pronoun (ba=ga idi-tar "I exited") :

	  ba=ga idi-tar   1SG=NOM exit-PAST -- (Example 9)
	  ba=nu idi-tar  *1SG=NOM exit-PAST -- (Example 10) 
	  ba=u  idi-tar  *1SG=ACC exit-PAST -- (Example 11)
	  ba=ga idi-tar  *1SG=GEN exit-PAST -- (Example 12)


	(3) Pronoun (Inclusive/Exclusive)

	# Here is the example(s) form our test suite
	#-----------------------------
	#Example 13 1PL exclusive Pronoun
	Source: author
	Vetted: f
	Judgment: g
	Phenomena: c
	banti ga iditar
	banti=ga idi-tar
	1PL.excl=NOM exit-PAST
	We exited.

	#Example 14 1PL inclusive Pronoun
	Source: author
	Vetted: f
	Judgment: g
	Phenomena: c
	duu nu iditar
	duu=nu idi-tar
	1PL.incl=NOM exit-PAST
	We exited.

	#Example 15 1PL inclusive Pronoun
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: c
	duu ga iditar
	duu=ga idi-tar
	1PL.incl=NOM exit-PAST
	We exited.


	#Example 16 no ad position
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: c
	duu iditar
	duu idi-tar
	1PL.incl exit-PAST
	We exited.

	#Example 17 ad position in wrong order
	Source: author
	Vetted: f
	Judgment: u
	Phenomena: wo
	ga duu iditar
	ga=duu idi-tar
	NOM=1PL.incl exit-PAST
	We exited.
	#-----------------------------


	In Irabu Ryukyuan's 1st person pronoun, there is a distinction between inclusive
 	and exclusive (only in plural form). This distinction needs special attention in 
	implementing into grammar matrix because inclusive form "duu-" is semantically 
	pronoun but syntactically noun, therefore it can not take case marker "=ga", 
	which is only used with pronoun.                                                  

	  1ST Plural | Case | Example ("Our Child")
	  -----------+------+----------------------------------
	  banti      | =ga  | banti-ga ffa (1PL.EXVL=GEN child)
	  duu-       | =nu  | duu-nu ffa   (1PL.INC=GEN child)

	Therefore, we added extra feature PRONOUN +/- to illustrate this phenomena.

2. Properties of Miyako in our test suite not covered in the grammar yet:

	The feature of tense is not covered yet so test sentences with simple past
	and non-past are present but not accounted for.  
	
	Non-past is indicated by nothing being added to the stem and the simple
	past is indicated by adding a suffix to the verb -tar
	
	STEM  |  Non-Past  |  Past  
	------+------------+--------
	ff    |    ff      | ff-tar
	idi   |    idi     | idi-tar
	
	The morphemes are not present in the lexicon and only the surface forms,
	`ff' and `iditar' are present.
	
	Plural suffixes on the pronouns are not defined in our grammar as well.
	Other than that, the test sentences are very simple at this point since
	agreement between the nominal and the verb is not needed so only case
	and word order have been adressed.
	
	
3.  Summary of Results for Lab2

	We have 100 percent coverage over the test suite so far.  It parses all the 
	test sentence it should but there is a bit of over generation where 3 
	ungrammatical examples are parsed which are:
	
		ba nu iditar
		ba=nu idi-tar
		1SG=NOM exit-PAST
		I exited.
	
	Here the particle used on the pronoun is reserved only to non-pronouns
	and has something to do with the PRONOUN feature not being required to agree.
	
		duu ga iditar
		duu=ga idi-tar
		1PL.incl=NOM exit-PAST
		We exited.
		
	Here again we see the same problem with PRONOUN feature disagreement. This
	is the noun that is semantically a pronoun but syntacticly a noun.
	
		ba ga iditar
		ba=ga idi-tar
		1SG=GEN exit-PAST
		My exited.
		
	Here the particle `ga' can mean either mean `nom' or `gen'.  There is no
	destinction that blocks this being parsed ONLY as CASE nom.
	
4.  	The customization seems to be having trouble with the particle and pronoun
	interaction.  The ambiguity between the `nom' particle and `gen' particle. 
	Also, the agreement constraint for the PRONOUN feature needs to be remedied.

	Another thing to note is that we realized there was more that can be added
	to this test suite to help with the grammar related to case.  The genetive
	and dative case aren't looked at closely enough and there are a few more
	cases besides these main four that we have outlined so far.  This will be 
	dealt with later since we have run out of time for this update.



