Difference between revisions of "Tongan/Universal Dependencies"

From LING073
Jump to: navigation, search
(nsubj)
Line 143: Line 143:
 
Most Tongan complete sentences contain a noun subject, which requires the dependency relation, <code>nsubj</code>. Tongan is a VSO language, which means the subject of the sentence often occurs at the very end of the sentence, or in the middle.  
 
Most Tongan complete sentences contain a noun subject, which requires the dependency relation, <code>nsubj</code>. Tongan is a VSO language, which means the subject of the sentence often occurs at the very end of the sentence, or in the middle.  
  
na’e  kakau ha  fonu  kovi  ki    Tongatapu <br>
+
1. na’e  kakau ha  fonu  kovi  ki    Tongatapu <br>
 
PAST  swim  NRF    turtle  bad to Tongatapu <br>
 
PAST  swim  NRF    turtle  bad to Tongatapu <br>
 
‘A bad turtle swam to Tongatapu.’ <br>
 
‘A bad turtle swam to Tongatapu.’ <br>

Revision as of 14:18, 9 May 2017


Evaluation

Add an "Evaluation" section, and make a grid that includes the UAS and LAS scores from your four parsing attempts above, as well as the number of forms (and sentences) in each corpus.

Evaluation

UD Parser Evaluation
ton.withmorph.udpipe ton.nomorph.udpipe Number of forms Number of sentences
ton.annotated.ud.conllu UAS: 69.85%, LAS: 61.45% UAS: 61.83%, LAS: 53.44% 262 12
ton.annotated2.ud.conllu UAS: 39.82%, LAS: 21.24% UAS: 29.20%, LAS: 20.35% 113 8

Dependency Relations

Below is an overview of 5 dependency relations that each occur more than twice in my annotation.

A description of the relation, noting various ways it might be used in the language,
   Two examples of the relation from your corpus, preferably illustrating what you described.

Add a "Dependency relations" section, and make a subsection for each of five dependency relations that you used at least twice in your annotation. For each relation, provide:

aux

Tongan does not conjugate verb forms based on tense, aspect, mood, etc. Instead it uses tense markers that indicate whether the verb is imperative, or past, present, future tense. Since Tongan is a VSO language, and tense markers precede all verbs, it is usually the first word in a sentence. I have chosen to analyze these tense markers in my transducer as auxiliary verbs. Therefore, the verb depends on the auxiliary tense marker, and we use the relationship aux. Below are a few example where this aux dependency relation appears.


1. na’e ongo’i ‘e he kulii (‘a) e manupuna
PAST hear ERG REF dog ABS REF bird
‘The dog heard the bird.’
[1]

"<naʻe>"
	"naʻe" vaux past @aux #1->2
"<ongoʻi>"
	"ongoʻi" v trans @root #2->0
"<ʻe>"
	"ʻe" erg @case #3->5
"<he>" 
	"he" sdef @det #4->5
"<kuli>"
	"kuli" n sg @nsubj #5->2
"<ʻa>"
	"ʻa" abs @case #6->8
"<e>"
	"e" sdef @det #7->8
"<manupuna>"
	"manupuna" n sg @obj #8->2

2.

na’e kakau ha fonu kovi ki Tongatapu
PAST swim NRF turtle bad to Tongatapu
‘A bad turtle swam to Tongatapu.’
[2]


"<naʻe>"
	"naʻe" vaux past @aux #1->2
"<kakau>"
	"kakau" v intrans @root #2->1
"<ha>"
	"ha" ind @det #3->4
"<fonu>"
	"fonu" n sg @nsubj #4->2
"<kovi>"
	"kovi" adj postnom #5->4
"<ki>"
	"ki" pr @case #6->7
"<Tongatapu>"
	"Tongatapu" np @obj #7->2

case

The case dependency relation in my Tongan corpus is used for prepositions and absolutive/ergative case markers. The case relation is used for any case-marking element which is treated as a separate syntactic word. Here are two examples of the case dependency relation being used both in the preposition form and the ergative/absolutive form.

na’e sio (‘a) e kulii ki he sikulele
PAST see ABS REF dog to REF squirrel
‘The dog saw the squirrel.’

"<naʻe>"
	"naʻe" vaux past @aux #1->2
"<sio>"
	"sio" v trans @root #2->0
"<ʻa>"
	"ʻa" abs @case #3->5
"<e>"
	"e" sdef @case #4->5
"<kuli>"
	"kuli" n sg @nsubj #4->2
"<ki>"
	"ki" pr @case #5->7
"<he>"
	"he" sdef @det #6->7
"<sikulele>"
	"sikulele" n sg @obj #7->2

na’e siofi ‘e ha sikulele ‘a Pulotu
PAST watch ERG NRF squirrel ABS Pulotu

‘A squirrel watched Pulotu.’

"<naʻe>"
	"naʻe" vaux past @aux #1->2
"<siofi>"
	"siofi" v trans @root #2->0
"<ʻe>"
	"ʻe" erg @case #3->5
"<ha>"
	"ha" ind @det #4->5
"<sikulele>"
	"sikulele" n sg @nsubj #5->2
"<ʻa>"
	"ʻa" abs @case #6->7
"<Pulotu>"
	"Pulotu" np @obj #7->2

amod

amod in my corpus has been used for adjectives that are modifying nouns. Adjectives can precede or follow a noun in Tongan. A special type of adjective, which I used amod for, are plural markers. The morphology of lemmas does not change according to plurality, instead Tongan uses plural markers to indicate whether a noun is in its plural form.

nsubj

Most Tongan complete sentences contain a noun subject, which requires the dependency relation, nsubj. Tongan is a VSO language, which means the subject of the sentence often occurs at the very end of the sentence, or in the middle.

1. na’e kakau ha fonu kovi ki Tongatapu
PAST swim NRF turtle bad to Tongatapu
‘A bad turtle swam to Tongatapu.’
[3]


1. "<naʻe>"
	"naʻe" vaux past @aux #1->2
"<kakau>"
	"kakau" v intrans @root #2->1
"<ha>"
	"ha" ind @det #3->4
"<fonu>"
	"fonu" n sg @nsubj #4->2
"<kovi>"
	"kovi" adj postnom #5->4
"<ki>"
	"ki" pr @case #6->7
"<Tongatapu>"
	"Tongatapu" np @obj #7->2


2. na’e siofi ‘e ha sikulele ‘a Pulotu
PAST watch ERG NRF squirrel ABS Pulotu

‘A squirrel watched Pulotu.’

"<naʻe>"
	"naʻe" vaux past @aux #1->2
"<siofi>"
	"siofi" v trans @root #2->0
"<ʻe>"
	"ʻe" erg @case #3->5
"<ha>"
	"ha" ind @det #4->5
"<sikulele>"
	"sikulele" n sg @nsubj #5->2
"<ʻa>"
	"ʻa" abs @case #6->7
"<Pulotu>"
	"Pulotu" np @obj #7->2

det

Tongan has three different determiners. A definite article, he/e, indefinite article, ha, and a semi-definite article, he/e. Most articles (at least as seen in my corpus), are semi-definite article he/e. Articles tend to immediately precede the noun they are describing.


1. na’e tuli au ‘e he kulii
PAST chase me ERG REF dog
‘The dog chased me.’

"<naʻe>"
	"naʻe" vaux past @aux #1->2
"<tuli>"
	"tuli" v trans @root #2->0
"<au>"
	"au" prn postd pl excl sg @obj #3->2
"<ʻe>"
	"ʻe" erg @case #4->5
"<he>"
	"he" sdef @det #5->6
"<kuli>"
	"kuli" n sg @nsubj #6->2

2.

na’e mohe e kulii ‘i he hala
PAST sleep REF dog in REF street
‘The dog slept in the street.’

"<naʻe>"
	"naʻe" vaux past @aux #1->2
"<mohe>"
	"mohe" v intrans @root #2->0
"<e>"
	"e" sdef @det #3->4
"<kuli>" 
	"kuli" n sg @nsubj #4->2
"<ʻi>"
	"ʻi" pr @case #5->7
"<he>"
	"he" sdef @det #6->7
"<hala>"
	"hala" n sg @obj #4->2
  1. The Syntax and Semantics of Tongan Noun Phrases http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.535.1919&rep=rep1&type=pdf ]]
  2. The Syntax and Semantics of Tongan Noun Phrases http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.535.1919&rep=rep1&type=pdf ]]
  3. The Syntax and Semantics of Tongan Noun Phrases http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.535.1919&rep=rep1&type=pdf ]]