Morphological Annotation

This page will give a detailed description about how the data is annotated at morphological layer (m-layer).

Positional Tagging


For m-layer annotation (aka POS tagging in general), we chose to annotate separate word tokens with morphological features in addition to single main POS. Having morphological features would be ideal and necessary for morphologically rich languages such as Tamil. Just by knowing those features it may be possible to identify certain syntactic phenomena (such as case markers can identify syntactic relations). So to include morphological information (such as case, person, number and etc) in addition to main POS, we decided to adopt the positional tagging scheme which has been successfully applied for the Czech language.

Figure 3.1: An Example Tagging of a Tamil Word
Positional Tag Example

In positional tagging, each token (word) is tagged with a fixed length string. Each position or character in the tag will  represent a particular morphological feature of the token.  The first position or character of the tag will indicate the broad word category (such as noun, verb etc) to which the token belongs and the 2nd position will indicate the detailed POS (for ex, what kind of verb? whether finite or non finite). The original Czech positional tagset includes 15 positions, and we designed a 9 positional tagset for Tamil. The Figure 3.1 illustrates the tagging of a Tamil word using positional tagging scheme.

Positional Tagset for Tamil


This section describes the positional tagset we designed for m-layer annotation of Tamil dataset. As mentioned in the previous section, positional tag for Tamil contains 9 positions. We describe each position in detail with all possible values a position can take. The Table 2.1 shows the Tamil positional tagset in general.

Table 2.1: Positional Tagset for Tamil
Position
 Tag
 #Num of Positions
1
2
3
4
5
6
7
8
9		 POS
Sub POS
Case
Tense
Person
Number
Gender
Voice
Negation
 14
48
07
04
03
02
05
02
02

Position 1: POS


Table 2.2: POS
Value Description
A
C
D
I
J
N
P
Q
R
T
U
V
X
Z		 Adverbs
Conjunctions
Determiners
Interjections
Adjectives
Nouns
Postpositions
Quantifiers
Pronouns
Particles
Numerals
Verbs
Unknown
Punctuations

Position 2: Sub POS


Table 2.3: SubPOS
Value
 Description
 Examples
#
 Sentence boundary The dot at  the end of the sentences
:
 Comma ,
=
 Numbers using digits 10, 15 etc.
b
 Comparative particle காட்டிலும்/kAttilum/'than', விட/vita/'than'
c
 Connective particle உம்/um/'also, and'
d
 Adjectival particle ஓடுகிற/OtukiRa/'which is running' , ஓடிய/Otiya/'which ran', ஓடும்/'which will run'
e
 Interrogative particle ஆ/A
f
 Civility particle உம்/um/'may please' : வரவும்/varavum/'may you please come'
g
 Particles, general ஆக/Aka, ஆன/Ana
h
 Reflexive pronouns தான்/TAn
i
 Interrogative pronouns யார்/yAr/'who',  எவன்/evan/'who he', எது/eTu/'which', என்ன/enna/'what'
j
 Lexical, Imperative verb ஓடு/Otu/'run', உதவு/uTavu/'help'
k
 Intensifier particle ஏ/E/'very, indeed, itself'
l
 Particle 'ஆவது/AvaTu' ஆவது/AvaTu/'at least'
m
 Particle 'மட்டும்/mattum' மட்டும்/mattum/'only'
n
 Complementizing nouns படி, மாதிரி/pati, mATiri/'manner, way', போது/pOTu/'when'
o
 Particle of doubt or indefiniteness ஓ/O
p
 Personal pronouns நான்/wAn/'I', நீ/wI/'you', அவன்/avan/'he,remote'
q
 Emphatic particle ஏ/E, தான்/tAn - ex: என்னையே/ennaiyE/'just me', ராமன்தான்/rAmanTAn/'It was
ram ...'
r
 Lexical, finite verb ஓடுகிறான்/OtukiRAn/'he is running', உதவுகிறாள்/uTavukiRAL/'she helps'
s
 Concessive particle உம்/um/ - ex: ஓடியும்/Otiyum/'although ran', உதவினாலும்/uTavinAlum/'evenif helps...'
t
 Lexical, verbal particle வந்து/vawTu/'having come', ஓடி/Oti/'having run', உதவி/uTavi/'having helped'
u
 Lexical, infinitive வர/vara/'to come', ஓட/Ota/'to run', உதவ/uTava/'to help'
v
 Inclusive particle உம்/um/'also' , for ex: ராமனும்/rAmanum/'also rAman'
w
 Lexical, conditional வந்தால்/vawTAl/'if come', ஓடினால்/OtinAl/'if ran', உதவினால்/uTavinAl/'if helped'
x
 Cardinals ஒன்று/onRu/'one', இரண்டு/iraNtu/'two'
y
 Ordinals ஒன்றாம்/onRAm/'first', இரண்டாம்/iraNtAm/'second', மூன்றாம்/mUnRam/'third'
z
 Lexical, verbal nouns வந்தது/vawTaTu/'the thing which came', ஓடியது/OtiyaTu/'the thing which ran'
A
 Adverbs, general விரைவாக/viraivAka/'quickly', அடிக்கடி/atikkati/'frequently'
B
 General referential pronouns யாரும்/yArum/'anyone', எவனும்/evanum/'anyone, he', எதுவும்/eTuvum/'anything'
C
 Coordinating conjunctions, general மற்றும்/maRRum/'and', அல்லது/allaTu/'or'
D
 Determiners, general அந்த/awTa/'that', இந்த/iwTa/'this', எந்த/ewTa/'which'
E
 Proper names கௌதமா/kauTamA/'Gautama', பிராக்/pirAk/'Prague'
F
 Specific indefinite referential pronouns யாரோ/yArO/'someone', எவனோ/evanO/'someone, he', எதுவோ/eTuvO/'something'
G
 Non specific indefinite pronouns யாராவது/yArAvaTu/'someone, atleast', எவனாவது/evanAvaTu/'someone,he, atleast', எதாவது/eTAvaTu/'something, atleast'
I
 Interjections, general ஆஹா/AhA/'aha'
J
 Adjectives, general அழகான/azakAna/'beautiful', நல்ல/walla/'good'
N
 Nouns, general நாற்காலி/wARkAli/'chair', பேருந்து/pEruwTu/'bus'
O
 Oblique nouns மர/mara/'wooden', புத்தக/puTTaka/'of book'
P
 Participial nouns ஓடியவள்/OtiyavaL/'she who ran', உதவியவர்/uTaviyavar/'he-she who helped'
Q
 Quantifiers, general கொஞ்சம்/konjcam/'a bit'
R
 Auxiliary, finite verb கொண்டிருக்கிறார்/koNtirukkiRAr/'he-she has been doing'
S
 Immediacy particle வந்ததும்/vawTaTum/'as soon as came', ஓடியதும்/OtiyaTum/'as soon as ran'
T
 Auxiliary, verbal participle கொண்டு/koNtu
U
 Auxiliary, infinitive பட/pata: passive infinitive
W
 Auxiliary, conditional பட்டால்/pattAl: passive conditional
X
 Unknown
Z
 Auxiliary, verbal nouns பட்டது/pattaTu: passive verbal noun (neuter)

Position 3: Case


Table 2.4: Case
Value
 Description
A
 Accusative
D
 Dative
I
 Instrumental
G
 Genitive
L
 Locative
N
 Nominative
S
 Sociative

Position 4: Tense


Table 2.5: Tense
Value
 Description
 Example
 Tag
D
 Past
 தெரிவித்தார்/TeriviTTAr/'informed, he'
 Vr-D3SHAA
F
 Future
 உதவும்/uTavum/'it will help'
 Vr-F3SNAA
P
 Present
 செல்கிறார்/celkiRAr/'he is going'
 Vr-P3SHAA
T
 Tenseless
 இல்லை/illai/'neg auxiliary'
 Vr-T3PNAA

Position 5: Person


Table 2.6: Person
Value
 Description
 Example
 Tag
1
 1st person
 மேற்கொண்டேன்/mERkoNtEn/'undertook,I'
 Vr-D1SAAA
2
 2nd person
 நீங்கள்/wIngkaL/'you,polite'
 RpN-2SH--
3
 3rd person
 கட்டினார்/kattinAr/'he built'
 Vr-D3SHAA
X
 Person not marked
 not used so far
 not used so far
Rules for Gender Person Attribute
  1. 1st person is always tagged with "Animate gender".
  2. If a person is not marked then use "X" which means person not marked.

Position 6: Number


Table 2.7: Number
Value
 Description
Example
 Tag
S
 Singular             
இந்தியா/iwTiyA/'India'
 NEN-3SN--     
P
 Plural                  
நாடுகள்/wAtukaL/'countries'
 NNN-3PN--      
X
 Number not         marked
not used so far
 not used so far

Position 7: Gender


Table 2.8: Gender
Value
 Description
Example
 Tag
F
 Feminine
வருவாள்/varuvAL/'she will come'
 Vr-F3SFAA
M
 Masculine
ஆடவனின்/Atavanin/"adult's(he)"
 NNN-3SM--
N
 Neuter
எடுத்தது/etuTTaTu/'took it'
 Vr-D3SNAA
H
 Honorific (both F & M)
அவர்/avar/'he or she, polite'
 RpN-3SH--
A
 Anitmate (humans)
யார்/yAr/'who'
 RiN-3SA--
I
 Inanimate (non humans)
not used so far
 not used so far
X
 Gender not marked
not used so far
 not used so far

Position 8: Voice

Tamil passive sentences are always constructed with auxiliary verb படு/patu/'experience'.  Only  the auxiliary verb படு/patu/'experience' receives this feature value "P". All other verbs receive "A"

Table 2.9: Voice
Value
 Description
Example
 Tag
A
 Active
எடுத்தது/etuTTaTu/'took    it' Vr-D3SNAA
P
 Passive
படுகிறது/patukiRaTu/'being....[past participle]'
 VR-P3SNPA

Position 9: Negation



Table 2.10: Negation
Value
 Description
Example
 Tag
A
 Affirmative
பின்பற்ற/pinpaRRa/'to follow'
 Vu-T---AA
N
 Negatedமுடியாது/mutiyATu/'cannot'
 VR-T3SN-N

Rules for Negation
  1. This feature is marked on wordforms which are morphologically marked for negation. The word form should be negated with "AT"
  2. If a person is not marked then use "X" which means person not marked.