|
Annotation
The
annotation of the corpora is carried
out in several steps.
First step: manual and semi-automatic annotation
All texts are annotated in xml. The
syntactic annotation first of all serves the objectives of the project
and
reflects the actual working hypotheses. It does not contain all
morphological
information, but only that which is relevant for our research goal. Our
decisions of what to count as a token or 'word' or an intonational unit
is based on features in the spoken language and may perhaps be disputed
in some cases. The annotation comprises on the lowest level tokens with
their orthographic realisation, a lemma tag, a translation tag, and the
part of speech
classification. The second level consists of ntNodes for the
inflectional group (i.e. noun phrase in case of
arguments and adverbial phrase in case of non-arguments) with a
description of
the case markers used, clauses with clause type (e.g. simple, chained,
embedded,
etc.). As we have tried to keep our annotation as flat as possible we
have not (with only one exception) annotated ntNode-internal
ntNodes. Likewise we have generally treated complex verb syntagms as
units (tokens) and not as verbal phrases containing one or several
tokens. The higher units are clause, sentence, discourse units such as
direct or indirect speech, poems
or songs, and text segments (divisions).
Each verb token automatically receives an
identity number as an attribute. It also contains a special description
for features (such as auxiliary verbs, negation and locutionary
markers, etc., if necessary for the understanding), the frame, i.e.,
the
'canonical' argument structure as listed in the lexicon, and the
realFrame, i.e., the
realisation of the arguments in the clause. Each realComplement (i.e.,
argument
in the realFrame) automatically receives an identity number, based on
the verb
number. Empty arguments receive a corresponding attribute and a
reference tag
that specifies the antecedens in the text. The tag receives an
attribute "target"
with the identity number of the particular referee, which, in most
cases, will
be a realComplement. Reference tags will also specify the referees of a
deictic
element. Reference tags may also be specified via attributes according
to the type of reference (including 'invalid' references where the
options of a text cannot be solved) or the type of antecedens (e.g. in
the case of empty antecedens).
All tokens (or
"words") or
special readings are listed in two text-specific lexica (one for verbs
and one for all other items), each of which will be accessible via the
annotated corpora. Verbs will be listed with their stem
forms and their frames. Verb tokens
are linked to their English (or German) counterparts in the translation
file and vice versa, which
makes navigation between annotated text and translation easy (see also
our introduction to the visible corpora).
Second step: automatic evaluation of the data
The tags of non-realised realComplements will
contain an antecedent list with information about the distance between
reference and referee and
the possible shifts in roles and case marking between the referee.
Likewise the tags of realComplements that served as antecedent will
contain a similar anaphor list, specifying all anaphoric elements
referring to this particular item. Both lists are generated
automatically and allow fine-grained statistical evaluations.
Third step: visualisation
of the corpora
For a first introduction, see
Definitions:
I. Elements:
|
<tok>
|
the
minimal intonation unit (word), contains
|
|
<orth>
|
|
the
(reconstructed) orthographic form
|
|
attributes:
|
|
|
|
|
broken
|
|
for
discontinuative elements
|
|
graph
|
|
for
orthographic conventions
|
|
orig
|
|
original,
if a clearly misspelt form has been emendated
|
|
standard
|
|
standard
orthography as found in dictionaries
|
|
<lemma>
|
|
the
basic form as given in the dictionaries
|
|
<trans>
|
|
the
translation as given in the dictionaries
|
|
<pos>
|
|
part
of speech (see below)
|
|
(<desc>)
|
|
description
of grammatical features, containing:
|
|
|
|
<case>
|
|
|
|
<feature>
|
|
|
|
<ref>, the reference tags
|
|
|
|
the argument
structure, in the case of verbs (see below)
|
Several
non-verbal <tok>-s may form an <ntNode>; verbal
<tok>-s are not further analysed, but in the case of complex verb
syntagms one may understand the <tok> as an <ntNode> "VP"
|
|
<ntNode>
|
necessarily
specified by a <ntNodeCat>
the highest nominal phrasal unit bearing the inflection
(internal phrasal structures are typically not further annotated), may
contain <tok>-s and <clause>-s)
|
|
<clause>
|
necessarily
specified by a <clauseCat>
a unit of <tok> with maximally 1 verb <tok> on the main
level
(may contain further embedded <clause>, <q>; may not
directly contain <s>)
|
|
<s>
|
every
sentence <s> has exactly one finite or equivalent <clause>,
except for the alternative question, where quest.alt1 und quest.alt2
form two finite clauses (otherwise quest.alt2, for lack of the question
marker, could not be treated as interrogative).
|
|
<q>
|
direct
and indirect speech or thought
|
|
attributes:
|
|
|
|
|
direct
|
values |
|
|
|
|
y |
direct
speech (ends with finite verb & possibly {ces})
|
|
|
n |
indirect
speech (typically ending in a verbal noun plus LocPurp)
|
|
|
|
in
the case of work titles the default value is "unspecified".
|
<poem>
|
versified
text
|
<div>
|
division:
major discursive segment, paragraph, or chapter
|
<text>
|
the
highest node
|
|
additional
elements:
|
<comment>
|
recurring
information
|
<note>
|
explicatory
footnote
|
<lb/>
|
line
break (OTC)
|
<l_end/>
|
end
of verse line
|
<pb/>
|
page
break
|
<unclear_start/>
|
for
illegible or unintelligible passages
|
<unclear_end/>
|
for
illegible or unintelligible passages
|
<add_start/>
|
for
additions
|
<add_end/>
|
for
additions
|
II. Definitions of categories
(If
diacritics are not properly displayed, please select the UTF-8 charset
on your
browser.)
|
1.
<pos>
|
ADJ
|
nominal
adjective (only derived bi- or multisyllabic forms)
|
ADV
|
plain
adverb
|
CONJ
|
conjunction
|
DEMfar
|
distal
demonstrative pronoun
|
DEMnear
|
proximal
demonstrative pronoun
|
EDEM
|
emphatic
demonstrative pronoun "exactly this/that one"
|
DEMfar.ADJ
|
distal
demonstrative adjective (compound)
|
DEMnear.ADJ
|
proximate
demonstrative adjective (compound)
|
EPRON
|
emphatic
pronoun (PRON + raŋ oder ñid); an additional
number
indicates the person
|
INTJ
|
exclamation,
interjection within the text
|
IQ
|
indefinite
quantifier/qualifier: tsam, "how much, as much"
|
NAME
|
personal
or place name
|
NOM
|
noun
|
NUM
|
number
word
|
ONM
|
onomatopoetica
|
PRON
|
personal
pronomen; an additional number indicates the person
|
PRON2voc
|
addressing
pronoun (WT wa, la)
|
-proNOM
|
pronominally
used ADJ, QADJ, RADJ
|
QADJ
|
quantitative
adjective (maŋpo 'many', gñiska 'both', etc.)
|
QoADV
|
quotation
adverb {ces}
|
QoI
|
quote
introduction na·re
|
QoM
|
quote
marker lo WT
|
QPRON
|
question
pronoun
|
QPRONrel
|
question
pronoun used as relative pronoun
|
RADJ
|
relational
(deiktic) adjective: gžan "other"
|
RFLX
|
reflexive
pronoun raŋ
|
RT
|
root:
monosyllabic elements of adverbs or adjective-compounds that do not
occur independently; AvP-s (adverbial phrases) are often derived from a
root with the help of a locational case marker
|
SEQ
|
sequentielle
marker 'and then' (formally DEMfar plus Abl)
|
SUM
|
sumariser:
·la-sogs(·pa) /·las-stsogs(·pa) 'etc.'
|
TOP
|
topic
marker ni
|
V
|
verb (any
non-finite form)
|
Vε
|
non-finite
use of mere stem (in the case of cross clausal group inflection: the
first members appear in a non-modified, thus seemingly finite form, but
the morphemes of the final non-finite verb operates also on the
preceeding verbs; stylistic device for closely related event sequences,
particularly frequent for ritual(ised) actions)
|
VFIN
|
finite use
of mere stem or complex periphrastic form (± final marker)
|
VN
|
verbal
noun
|
Vⁿ
|
verbally/predicatively
used VN (reanalysis as VN possible)
|
VCp
|
predicatively
used verb compounds (or nominalisations) governing arguments
|
VNst
|
nominally
used stem, e.g. TVP maslebkyi-bardu
|
discontinuative
elements
the first element remains without id and frame; both elements are
linked with pointers
|
discMV
|
discontinuative
modal verb construction
|
discSCV
|
discontinuative
serial verb construction
|
modifying
elemets (":x" for subcategories, ".x" for lexicalised items, and "-x"
for morphemes are joined to the main elements for the sake of a flat
annotation; one could have tagged them as features)
|
:anim
|
animate
|
:inan
|
inanimate
(including abstract terms)
|
:pers
|
person
(also in the case of titles); typically cogitative or intentional
persons above 7 years (small children and animals are typically not
included; but deamons, deities, talking animals, and supernaturally
intelligent children are intentional persons)
|
-cm
|
collective
marker dag
|
-df
|
determiniser
po/de WT
|
-iq
|
indefinite
quantifier/qualifier: tsam, "how much, as much"
|
-lq
|
limiting
quantifier {cig}
|
-pl, .pl
|
plural
marker (-pl kun, rnams) or plural form (.pl khoŋ)
|
.excl
|
exclusive
plural (only PRON1)
|
.incl
|
inclusive
plural (only PRON1)
|
|
2.
<feature> |
|
|
type="morph" |
morpheme
|
|
|
DM |
directive
marker, for commands, prohibitions, and optatives |
|
FIN |
(sentence)
final marker |
|
FM |
focus
marker |
|
NEG |
negation |
|
pQ |
preposed
question marker |
|
QF |
question
final marker |
type="AUX" |
auxiliary
verbs |
type="compound"
|
|
type="lwOpMV"
|
leftward
operating modal verb
|
type="MV"
|
modal
verbs
|
type="reduplication"
|
|
|
|
|
3.
<case> |
|
|
Abl
|
ablative (nas,
las)
|
Abs
|
absolutive
|
Aes
|
aesthetive
(=DatLoc), only WT, CT/OT only for yod 'have'
|
Com
|
comitative
(or sociative: daŋ)
|
DatLoc
|
dative/locative
(la)
|
Erg
|
ergative
(=Instr in OT/CT)
|
Gen
|
genitive
(kyi/gyi/gi/yi/-ḥi)
|
Instr
|
instrumental
(kyis/gyis/gis/yis/-s)
|
~Loc
|
locational
case (including postpositions; for frame and realFrame)
|
Loc
|
locative
(na)
|
LocPurp
|
locative/purposive
(tu/du/ru/-r/su)
|
PPosAbl
|
ablative
postposition
|
PPosCom
|
comitative
postposition
|
PPosLoc
|
locational
postposition
|
PPosInstr
|
instrumental
postposition
|
rel
|
relator
morpheme bas, WT basaŋ (is often taken as a special
form of the ablative las in comparative-like expressions; most
probably we are dealing with bimorphemes: la + s and ba
+ s (saŋ), whose functional value depends on the context)
|
|
4.
<ntNodeCat> definitions
|
AvP
|
adverbial
noun phrase
|
EP
|
expeditive
procedure (addressing the addressee in a speech act)
|
NP
|
noun
phrase as argument
|
NP-bound
|
bound
argument of a collocation
|
NP-suppl
|
supplement
of a light verb
|
ppAvP
|
postponed
AvP
|
ppEP
|
postponed
EP
|
ppNP
|
postponed
(full) NP
|
|
5.
<clauseCat> definitions
|
|
|
main
types:
|
|
1.
simple sentences (single clause)
|
|
|
simple
|
simple
sentence with finite verb without embedded or chained
structures (maximal 1 clause; exceptions: 1. simple-spch.intro; 2.
quest.alt1 plus quest.alt2)
|
simple-spch.intro
|
simple
sentence with non-finite introductory verbum dicendi
|
phrs-spch.intro
|
speech
introduction with na·re
|
|
2.
complex sentences (more than 1 clause)
|
|
2.a
clause chaining
|
chained
|
clause
chaining, non-finite Verb (linear, iconic representation of main story
line:
events 1, 2, 3 correspond to verbs 1, 2, 3)
|
chain.mod
|
modifying
clause without embedding
|
chain.purp
|
non-embedded,
i.e. preposed purposive clause ("in order to verb 1 / verb 2)
|
endchain
|
last finite
verb of a chain of clauses
|
endchain-spch.intro
|
last non-finite
introductory verbum dicendi of a chain of clauses
|
|
2.b
embedded (subordinated) clauses
|
emb.attr
|
embedded
attributive or modifying nominal clause (with genitive, preceding head
noun)
|
emb.cont
|
phrasal
content of thought or perception (Abs-marking)
|
emb.mod
|
modal,
causal, or temporal subordination
|
emb.nom
|
embedded
nominalised clause (following head noun)
|
emb.purp
|
purposive
clause
|
emb.prop
|
embedded proposition, indirect speech, thought, or
perception (Loc/Purp-marking)
|
emb.spch
|
embedded
speech, not introduced by verbum dicendi but preceded by agent
of final verbum dicendi(e.g. khos
"..." (ces) gsuŋs)
|
superord
|
superodinated
clause (finite verb)
|
superord-chained
|
superodinated
clause, chained (non-finite verb)
|
superord-emb.x
|
superodinated
clause, embedded
|
|
3.
paired constructions (the second member
is typically finite, unless specified with one of the markers of
chaining or embedding)
|
|
3.a
alternative questions
|
quest.alt1
|
first
part of alternative question (finite verb)
|
quest.alt2
|
second
part of alternative question (finite verb)
|
|
3.b
conditional constructions
|
cond
|
condition
clause (non-finite verb), necessary followed by cres
|
cres
|
result
clause (finite verb)
|
cond.intro
|
introductory
condition clause, almost like a question (if you don't know/don't you
know ..., this is x
|
|
3.c
relative clause constructions (pronominal)
|
relat1
|
first
part of relative construction (who dares)
|
relat2
|
second
part of relative construction (wins/that wins)
|
|
3.d
antithetical constructions
|
anti
|
antithetical
(konzessiv, adversativ: obwohl, zwar)
|
anti2
|
the
corresponding counterpart
|
|
marginal
types:
|
anaph-chained
|
anaphoric
sentence connection (the finite verb is repeated as non-finite verb)
|
ε:
|
in
the case of ellided verb, mostly ε:simple-spch.intro
|
el.chained
|
elliptical
chained, i.e., morphological ellipsis Vε (cross clausal group inflection: the
first members
appear in a non-modified, thus seemingly finite form, but the
morphemes of the final non-finite verb
operates also on the preceeding verbs)
|
excl
|
phrasal
exclamation
|
nominal
|
nominal
clauses
|
nominal-link.up
|
nominal
clauses with lhagbcas morpheme {ste}
|
postp.purp
|
postponed
purposive construction; is treated as finite verb (i.e. forms the end
of a sentence)
|
|
|
|
simple,
simple-spch.intro, endchain, endchain-spch.intro, phrs-spch.intro,
superord, cres, and quest.alt2, if not further modified, conclude
a 'finite sentence' <s>. The same holds for postp.purp.
|
|
6.
semantic categories of verbs (verb dictionary)
|
|
6.a.
valency
|
defined
here as the minimal number of obligatory arguments; facultative
arguments that are licensed by the verb semantics (such as LCT in the
case of motion verbs) are indicated by the use of "+", valency is
indicated on crucial arguments such as A or P
|
|
6.b.
semantic subclassification (open class and subject to reformulations)
|
general
subcategories (more or less Vendler's categories, modified for
[-control]):
|
non-dynamic:
|
state
|
[-control]
|
position
|
[+control]
|
dynamic:
|
process
|
[-control],
non-bound
|
activity
|
[+control],
non-bound
|
development
|
[-control],
slow and increasing change
|
transition
|
[-control],
sudden change
|
accomplishment
|
[+control]
|
achievement
|
[+control]
|
specific
subcategories:
|
predicative
|
[-control] yin & corresponding dynamic
auxiliaries ḥgro pr cha 'go', V2
|
resulting
state
|
[-control]
(like transition, but more result oriented)
|
exist
|
[-control] yod and other verbs that follow pattern
03b (e.g. skye
'be born'), V1+
|
possession
|
[-control]
V2
|
get
|
[-control]
V2
|
state-of-mind
|
[-control] V2
|
emotion
|
[±control]
V2
|
motion
|
[-control],
motion verbs, V1+
|
suffice
|
[-control],
instrumental for medium-argument possible (instrumental in Shamskat)
|
fill
|
[-control],
instrumental for medium-argument possible (genitive in Shamskat)
|
contact
|
[-control], contact and
separation, typically with comitative
(or collective realisation)
|
transfer
|
[+control]
V3 (word order typically eA R P)
|
deposit
|
[+control]
V2+ (neutral word order possibly eA P LCT)
|
load/unload
|
[+control]
V2+ (threefold variation: V2: P = load, or P = carrier /container, V3:
P = load)
|
movement
or translocation
|
[+control] V3
|
communication
|
[+control]
V2+
|
transformation
|
[+control]
V2
|
result
oriented
|
[+control] V2+
|
production
|
[+control]
V2
|
directional
activity
|
[+control]
V2
|
fix
|
[+control]
V3
|
keep
|
[+control] V2
|
|
7.
reference tags may contain
additional (non-standardised) information mainly for annotator (open
class), e.g.
|
ø
|
no
reference implied or possible
|
impersonal
|
generic
statements
|
speaker
|
|
adressee
|
|
situation
|
for
references on clauses or sequences of clauses
|
VNref
|
missing
argument is constitutted by the VN itself (speak-er)
|
VNhead
|
missing
argument is head of
the VN (X, who speaks)
|
III. Argument structure: valency and semantic roles
(If
diacritics are not properly displayed, please select the UTF-8 charset
on your
browser.)
Annotation
in the text:
|
<frame>
|
contains:
<complement>, (<alt>)
|
attributes:
|
inherited
|
for
alternative questions
|
type, type2
|
for
special types, such as light verbs or collocations
|
<alt>
|
for
complement alternations, contains one or more complements
|
attributes
|
position
|
|
<complement>
|
contains:
<role>, <case>
|
attributes:
|
position
|
(in case
of alternations)
|
status
|
values:
|
|
|
bound
|
for
collocation partners
|
|
inherited
|
for
alternative questions
|
|
obligatory
|
default
status
|
|
omissible
|
facultative
arguments
|
|
supplement
|
for light
verbs
|
<realFrame>
|
contains:
<realComplement>
|
attributes:
|
derived,
derived 2
|
derived
frames (following a derivational scheme)
|
|
values:
|
|
|
cause
|
additional
causation argument for [-control] verbs
|
|
collective
|
argument
reduction, two assymetric arguments reduced to one
|
|
collocation
|
change in
status of complements
|
|
comparative
|
additional
relation-argument for predication verbs
|
|
exceptive
|
special
negation patterns
|
|
experiencer
|
additional
experiencer argument
|
|
explication
|
extended
frames with arguments that are usually left unspecified (e.g. an SRC
argument for motion verbs) or usually represented collectively
|
|
honorific
|
impersonal
ablative construction, used for persons of highest status
|
|
idiomatic
|
|
|
medium
|
additional
medium arguments for special verbs
|
|
MV
|
[-control]
modal verbs, change of roles; in WT also change of case
|
|
MVcs
|
causative
modal verbs
|
|
narrIntro
|
narrative
introduction corresponding to German es
war einmal
|
|
possessor
|
argument
reduction (especially common for body-related events or states)
|
|
potentialis
|
original
function of stem IV ('imperative' stem), change of roles; in WT also
change of case
|
|
reciprocal
|
argument
reduction, two assymetric arguments reduced
to one
|
|
reflexive
|
possible
change in subject case marking
|
|
refPara
|
additional
reference parameter for predication verbs
|
|
SCV
|
serialised
conjunctive verbs (aka vector verbs), change in case or argument reduction
|
order
|
for
changes in word order, relative to the canonical frame
|
type, type2
|
for
special types, such as light verbs or collocations
|
<realComplement>
|
contains:
<role>, (<case>), (<ref>)
|
attributes:
|
id
|
based on
verb token number, possible targets of <ref> tags
|
position
|
(in case
of alternations in canonical <frame>)
|
status
|
values:
|
|
|
additional
|
|
|
bound
|
for
collocation partners
|
|
bound-empty
|
non-realisation
of collocation partners
|
|
empty
|
non-realisation
of
|
|
inherited
|
non-realisation
of inherited arguments
|
|
inherited-empty
|
non-realisation
of
|
|
omitted
|
non-realisation
of facultative arguments
|
|
postponed
|
postponed
realisation outside the clause
|
|
supplement
|
for light
verbs
|
|
supplement-empty
|
non-realisation
of light verb complements
|
<role>
|
|
atributes
|
feature
|
values:
|
|
|
+anim
|
animate
|
|
-anim
|
inanimate
|
|
+hum
|
human
|
|
-hum
|
non-human
|
|
+spec/given
|
specific
or given argument
|
|
adj
|
argument
is an adjective
|
|
compound
|
(argument
reduction)
|
|
multipleRealisation
|
logical
argument represented by more than one case-marked NPs
|
|
possConstr
|
possessor
construction (argument reduction)
|
microRole
|
specific
micro roles (open class)
|
possMicroRole
|
possible
micro roles
|
<case>
|
|
|
Annotation
in the verb dictionaries:
|
<entry>
|
may
contain two ore more <subEntry>, otherwise see structure
<subEntry>
|
<subEntry>
|
contains:
<written>, <val>, <primScheme>, (<derScheme>),
<semCat>, <frames>, <meaning> or <trans>,
<lemma>
|
attributes:
|
idiomatic
|
values:
|
|
|
fig
|
figurative
employment
|
|
figEtym
|
figura
etymologica
|
|
onom
|
onomatopoetic
|
type
|
collocation
|
|
|
idiomatic
|
|
|
impersonal
|
|
|
lightV
|
light verb
construction
|
|
stripped
|
special
verb reading, conventional omission of one ore more arguments
|
<written>
|
contains:
<wrI>, <wrII>, <wrIII> (OT and CT), <wrIV>,
<neutral>, (<misspelt>, <rdg>)
|
|
<wrI> to <wrIV>
|
the four
possible slots for Tibetan verb stems; slot III and IV, implying an
intentional agent, are by definition not filled in the case of
[-control] verbs, slot II
remains empty in the case of copula verbs;
empty slots are indicated by "-"; slots for which the annotated text
has no attestation are filled with "?"; only differing verb forms are
indicated; in case of attested syncretic forms, the number of the lower
slot is given (eg. <wrI>dgos</wrI>,
<wrII>I</wrII>, <wrIII>-</wrIII>,
<wrIV>-</wrIV>;
|
|
<neutral>
|
neutral
stem as used in certain non-finite forms
|
|
<misspelt>
|
clearly
misspelt variants
|
|
<rdg>
|
variant
spellings in the text
|
<val>
|
valency;
we define the valency value as the minimal number of obligatory core
arguments. Core arguments are those arguments that are licensed by the
verb meaning, thus LCT is a (non-obligatory) argument of motion verbs,
but not of most other verbs.
|
<primScheme>
|
primary
scheme or clause type
|
attributes:
|
type
|
values
|
|
|
collective
|
preferred
collective expression
|
|
collocation
|
|
|
idiomatic
|
|
|
impersonal
|
conventionalised
usages
|
|
possessor
|
preferred possessive expression |
|
stripped
|
special
verb reading, conventional omission of one ore more arguments
|
<derScheme>
|
derived
scheme or clause type
|
attributes:
|
type,
typeComb
|
cause
|
additional
causation argument for [-control] verbs
|
|
collective
|
argument
reduction, two arguments assymetric reduced to one
|
|
collocation
|
change in
status of complements
|
|
comparative
|
additional
relation-argument for predication verbs
|
|
exceptive
|
special
negation patterns
|
|
experiencer
|
additional
experiencer argument
|
|
explication
|
extended
frames with arguments that are usually left unspecified (e.g. an SRC
argument for motion verbs) or usually represented collectively
|
|
honorific
|
impersonal
construction, used for persons of highest status
|
|
idiomatic
|
|
|
medium
|
additional
medium arguments for special verbs
|
|
narrIntro
|
narrative
introduction, corresponding to German es
war einmal
|
|
possessor
|
argument
reduction (especially common for body-related
events or states)
|
|
potentialis
|
original
function of stem IV ('imperative' stem), change of roles; in WT also
change of case
|
|
reciprocal
|
argument
reduction, two arguments assymetric reduced
to one
|
|
reflexive
|
possible
change in subject case marking
|
|
refPara
|
additional
reference parameter for predication verbs
|
|
SCV
|
serialised
conjunctive verbs (aka vector verbs), change
in case or argument reduction
|
<semCat>
|
semantic
subcategory
|
<frames>
|
contains
one or more <frame> elements
|
<frame>
|
see above
|
<trans>
|
translation
|
<meaning>
|
contains
one or more <trans> and/or additional <colloc>,
<scv>, <suppl>, <usg> elements
|
|
<colloc>
|
collocation
partner, contains <w>, <def> or <tr>
|
|
<scv>
|
serialised
conjunctive verb construction; contains <w> (type = verb),
<tr>
|
|
<suppl>
|
light verb
complement; contains <w>, <tr>
|
|
|
<w>
|
word
|
|
|
attributes:
|
|
|
type
|
pos
|
|
|
syntax
|
if word is
case-marked
|
|
|
<def>
|
definition,
in case of onomatopoetic expressions
|
|
|
<tr>
|
translation
of <w>
|
<lemma>
|
|
Arguments
(overview)
It should be noted that
in the following we present semantically based macro roles that are
defined by
1. semantic properties, such as function and agency
2. their position in the frame in relation to other arguments
3. their syntactic behaviour
Future analysis may show that some of these
roles can be further reduced and subsumed under syntactic roles.
Nicholas Tournadre (2009) proposes the following syntactic roles for
Tibetan:
S
'sole argument' (semantically undergoer or agent, also used for verbs
with higher valency, but more peripheral second arguments)
A
'actor' (for ergative-marked first arguments, whether they are
semantically agents or not)
R
'recipient' / C
'ceptor' (for dative-marked first arguments, typically only possessors
and re-ceptors in modern Tibetan, but also per-ceptors in Ladakhi and
in some other Indian and Caucasian languages; for this reason, he
agrees that C = 'ceptor' might be a better designation)
B
'beneficiary' (for dative-marked second non-peripheral arguments,
basically the semantic recipient)
P
'patient' (for all non-marked second or third arguments)
Tournadre, Nicholas (2009). " Core
Grammatical Roles in Tibetan with special reference to their
syntactic behaviour in subordinate clauses."
Presentation given at the University Tübingen, January 2009.
V=1 (S)
[-control]
|
|
Existence (EXST1)
|
existence
predication: absolutive
|
Undergoer (U1)
[+control]
|
[-control]
state, motion, or result verbs: absolutive
|
Agent (A1)
|
not acting
upon another participant [+control], motion, position verbs: absolutive
|
Agent (eA1)
|
effecting
(?) agent without Patient or Target: ergative
|
|
|
V=2
|
|
1.
'subject'-like (S, C, A) |
|
[-control]
|
|
HEAD
|
'subject'
of predication "be" or light verb "go X": absolutive
|
Possessor (POSS)
|
'subject'
of "have", "get" and similar verbs (subcase of existence predication,
where the non-obligatory LCT turns into an obligatory subject): aesthetive
|
Experiencer1
(Ep2)
|
'subject'
of receptive or perceptive verbs OT/CT: ergative; Ladakhi: aesthetive
|
Experiencer2
(Ee2)
|
'subject'
of emotion verbs: absolutive
|
Experiencer3
(Eg2)
|
'subject'
of sleb: ergative;
second argument: LCT (transitional frame; the verb is attested in OT
with pattern 08: eA2 & P2 or Ep2 & AFC2; in modern Tibetan it
follows pattern 03a: U2 & LCT)
|
[+control]
|
|
Agent (A2)
|
not acting
upon another participant, [+control]: absolutive
|
Agent (eA2)
|
not acting
upon, but towards another participant, directional activity: ergative
|
|
|
2.
'object'-like
(peripheral, P, B)
|
Existence (EXST2)
|
'object'
of "have" and "get" verbs: absolutive
|
Affecting (AFCp2)
|
'object'
or stimulus of perception verbs: absolutive
|
Affecting (AFCp2, AFCe2)
|
'object'
or stimulus of emotion verbs: dative/locative
|
Patient (P2)
|
absolutive
|
Target (TAR)
|
not an
'object', but the target of directional activities (including directed
attention): dative/locative
|
|
|
V=3 (A-B-P,
peripheral)
|
|
Agent (eA3)
|
ergative
|
Recipient (R)
|
animate,
typically on second position: dative/locative
|
Patient (P3)
|
thing
transferred or positioned: absolutive
|
Location (LCT)
|
inanimate,
position variable: locational marker
|
Source (SRC)
|
argument
from which P is taken or moved away: ablative
|
|
|
Other
semantically motivated (core) arguments:
|
Attribute (ATR)
|
predication
with copula: absolutive
|
Resulting
Attribute (RSA)
|
predication
with light verb: absolutive
|
Resulting
State (RST)
|
resulting
state of [-ctr] transition locative/purposive
(dative/locative)
|
Produced
State (PST)
|
resulting
state of [+ctr] transformation (V=3): locative/purposive
(dative/locative)
|
Location (LCT)
|
static or
dynamic; verbs of existenc, position, motion, deposit: locational marker
|
Source (SRC)
|
spatial
origin, particular motion and transfer: ablative
|
Contacted (CTC)
|
also in
case of separation: comitative
|
Mental
State (MST)
|
cognitive
or emotional state with light verb "come": absolutive
|
Content (CONTdir)
|
content of
a direct proposition (including proposition nouns), communication verbs
(finite sentence OT/CT mostly followed
by the QoADV {ces}): absolutive
|
Content (CONTind)
|
content of
an indirect proposition, communication verbs (non-finite sentence): locative/purposive
|
|
|
Peripheral arguments:
|
|
Bene-
(Male-)ficiary (BEN)
|
dative/locative, postposition (only
additional argument/adjunct)
|
Instrument
(INSTR)
|
only
microroles: cause, medium, organ: instrumental
|
Standard
of comparison (CREL)
|
contrastive
relation ablative or bas
= rel
|
Parameter (PARA)
|
relational
argument for adjectivals absolutive
|
Path (PATH)
|
absolutive argument of
particular motion verbs (transversing)
|
|
|
derived
roles:
|
|
reflexive
Agent (A1rfl; eA1rfl)
|
absolutive or ergative
|
reciprocal
Agent (A1rcp; eA1rcp)
|
absolutive or ergative
|
causative
Instigator (CIN)
|
ergative
|
causee
Undergoer (CU1)
|
absolutive
|
causee
Agent (CA1 etc)
|
CA1: absolutive, WT (CT in certain constructions) CA2: dative/locative
|
Experiencer
expansion (Eo;
Eoo)
|
Ladakhi: aesthetive
|
demoted
Agent (demA)
|
honorific
impersonal construction: for persons of high status, mainly for verba
dicendi and verbs related to some sort of speech act (e.g.
establish a custom), in WT also acts of giving ablative
|
demoted
Experiencer (demE)
|
ablative for
body-part
with [-ctr] verbs
|
ME1, ME2, eME2,
aME2
|
[-ctr] modal verbs (ME: absolutive,
eME: ergative, aME: aesthetive)
|
|
|
additional possible microroles (open class), e.g.
|
direction, goal
|
for LCT
|
time
|
for LCT
|
manner
|
for LCT
|
purpose
|
for LCT
|
Argument
combination
template OT/CT
(If
diacritics are not properly displayed, please select the UTF-8 charset
on your
browser.)
valence
|
"subject" like role
|
"object" like
|
other core arguments
|
1
|
EXST1
(Existence)
|
|
1
|
U1
(Undergoer)
|
1
|
A1 (Agent thamidadpa)
|
2
|
HEAD
|
|
ATR
(Attribute) yin
|
RSA
resulting Attribute ḥgro
|
RST
resulting State ḥgyur
|
2
|
POSS(essor)/LCT
|
EXST2
(Existence)
|
|
2
|
Ep2 (Experiencer)
|
AFCp2
(Affecting)
|
2
|
Ee2 (Experiencer)
|
AFCe2
(Affecting)
|
U2
(Undergoer)
|
|
LCT
(Location)
|
SRC
(Source)
|
CTC
(Contact)
|
2
|
A2 (Agent
thamidadpa)
|
|
LCT
(Location)
|
SRC
(Source)
|
CTC
(Contact)
|
eA2 (effecting
Agent
thadadpa)
|
TAR(get)
|
P2 (Patient)
|
|
3
|
eA3 (effecting
Agent
thadadpa)
|
R(ecipient)
|
P3
(Patient)
|
|
|
SRC
(Source)
|
LCT (Location)
|
PST (produced State)
|
|
|
peripheral and
facultative arguments
|
|
|
BEN(e/Maleficiary)
|
INSTR(ument)
|
|
|
REL
|
PATH
|
Case
|
CONTdir
(content of proposition)
|
Abs
|
Erg/Instr
|
CONTind
(content of proposition)
|
Aes/~Loc
|
Abl
|
CREL
(standard of "comparision")
|
Com
|
|
Annotation example:
first sentence of OTC
(without evaluation of references)
(If
diacritics are not properly displayed, please select the UTF-8 charset
on your
browser.)
<text>
<s><punct>~||</punct>
<ntNode>
<clause>
<ntNode>
<tok><orth>Dri·gum-btsan·po</orth>
<lemma>Dri·gum-btsan·po</lemma>
<trans>-</trans>
<pos>NAME</pos></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>Abs</case></desc></ntNode>
<ntNode>
<tok><orth>sku</orth>
<lemma>sku</lemma>
<trans>body
(hon)</trans>
<pos>NOM:inan</pos></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>Abs</case></desc></ntNode>
<tok
check="ok"
id="v1"><orth>chuŋ·ba·ḥi-tshe</orth>
<lemma>chuŋ</lemma>
<trans>be
small</trans>
<pos>VN</pos>
<desc>
<frame>
<complement><role>U1</role><case>Abs</case></complement></frame>
<realFrame
derived="refPara">
<realComplement
id="v1c1"><role>U2</role></realComplement>
<realComplement
id="v1c2"><role>PARA</role></realComplement></realFrame></desc>
</tok><clauseCat>emb.attr</clauseCat>
<desc><case>PPos</case></desc></clause><ntNodeCat>AvP</ntNodeCat>
<desc><case>Abs</case></desc></ntNode><q
direct="y">
<s>
<clause>
<ntNode>
<tok><orth>mtshan</orth>
<lemma>mtshan</lemma>
<trans>name
(hon)</trans>
<pos>NOM:inan</pos></tok><ntNodeCat>NP-bound</ntNodeCat>
<desc><case>Abs</case></desc></ntNode>
<ntNode>
<tok><orth>jir</orth>
<lemma>ji</lemma>
<trans>what</trans>
<pos>QPRON</pos></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>LocPurp</case></desc></ntNode>
<tok
check="ok"
id="v2"><orth>gdags</orth>
<lemma
n=":2:">ḥdogs</lemma>
<trans>name,
give as name</trans>
<pos>VFIN</pos>
<desc>
<frame>
<complement><role>eA4</role><case>Erg</case></complement>
<complement><role>R</role><case>~Loc</case></complement>
<complement><role>CONTind</role><case>~Loc</case></complement>
<complement
status="bound"><role>bArg</role><case>Abs</case></complement></frame>
<realFrame
order="inv:bArg-CONTind;q">
<realComplement
id="v2c1" status="empty"><role>eA4</role><ref>impersonal</ref></realComplement>
<realComplement
id="v2c2" status="empty"><role>R</role><ref
target="v1c1"></ref></realComplement>
<realComplement
id="v2c3"><role>CONTind</role></realComplement>
<realComplement
id="v2c4" status="bound"><role>bArg</role></realComplement></realFrame></desc>
</tok><clauseCat>simple</clauseCat></clause></s></q>
<clause>
<tok><orth>šes</orth>
<lemma>ces</lemma>
<trans>so,
thus</trans>
<pos>QoADV</pos></tok><punct>||</punct>
<ntNode>
<tok><orth>ma·ma</orth>
<lemma>ma·ma</lemma>
<trans>nurse</trans>
<pos>NOM:pers</pos></tok>
<tok><orth>Gro·ža·ma-Skyi·brliŋ·ma·la</orth>
<lemma>Gro·ža·ma-Skyi·brliŋ·ma·la</lemma>
<trans>-</trans>
<pos>NAME</pos></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>DatLoc</case></desc></ntNode>
<tok
check="ok"
id="v3"><orth>dris·na</orth>
<lemma>ḥdri</lemma>
<trans>ask</trans>
<pos>V</pos>
<desc>
<frame>
<complement><role>eA2+</role><case>Erg</case></complement>
<complement
status="omissible"><role>R</role><case>~Loc</case></complement>
<complement><role>CONTdir</role><case>Abs</case></complement></frame>
<realFrame
order="inv:CONTdir-R">
<realComplement
id="v3c1" status="empty"><role>eA2+</role><ref></ref></realComplement>
<realComplement
id="v3c2"><role>R</role></realComplement>
<realComplement
id="v3c3"><role>CONTdir</role></realComplement></realFrame></desc>
</tok><clauseCat>chained</clauseCat></clause><punct>|</punct>
<clause>
<ntNode>
<tok><orth>ma·ma·ḥi</orth>
<lemma>ma·ma</lemma>
<trans>nurse</trans>
<pos>NOM:pers</pos>
<desc><case>Gen</case></desc></tok>
<tok><orth>mchid·nas</orth>
<lemma>mchid</lemma>
<trans>speech</trans>
<pos>NOM:inan</pos></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>Abl</case></desc></ntNode><punct>|</punct><lb
n="0002"/>
<q direct="y">
<s>
<clause>
<ntNode>
<tok><orth>Skyi·brag</orth>
<lemma>Skyi·brag</lemma>
<trans>-</trans>
<pos>NAME</pos></tok>
<tok><orth>mar·ba</orth>
<lemma>mar·ba</lemma>
<trans>golden,
red</trans>
<pos>ADJ</pos><note>...</note></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>Abs</case></desc></ntNode>
<tok><orth>ni</orth>
<lemma>ni</lemma>
<trans>-</trans>
<pos>TOP</pos></tok>
<tok
check="ok" id="v4"><orth>rñil·tam</orth>
<lemma>rñil</lemma>
<trans>crumble,
collapse</trans>
<pos>VFIN</pos>
<desc><feature
type="morph">QF</feature>
<frame>
<complement><role>U1</role><case>Abs</case></complement></frame>
<realFrame>
<realComplement
id="v4c1"><role>U1</role></realComplement></realFrame></desc>
</tok><clauseCat>quest.alt1</clauseCat></clause>
<clause>
<tok
check="ok" id="v5"><orth>ma·rñil</orth>
<lemma>rñil</lemma>
<trans>crumble,
collapse</trans>
<pos>VFIN</pos>
<desc><feature
type="morph">NEG</feature>
<frame>
<complement><role>U1</role><case>Abs</case></complement></frame>
<realFrame>
<realComplement
id="v5c1" status="empty"><role>U1</role>
<ref
target="v4c1"></ref></realComplement></realFrame></desc>
</tok><clauseCat>quest.alt2</clauseCat></clause><punct>|</punct></s>
<s>
<clause>
<ntNode>
<tok><orth>Daŋ·ma</orth>
<lemma>Daŋ·ma</lemma>
<trans>-</trans>
<pos>NAME</pos></tok>
<tok><orth>ḥbri·spaŋs</orth>
<lemma>ḥbri·spaŋs</lemma>
<trans>grazing land for the ḥbri, the female of a yak</trans>
<pos>NOM:inan</pos></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>Abs</case></desc></ntNode>
<tok><orth>ni</orth>
<lemma>ni</lemma>
<trans>-</trans>
<pos>TOP</pos></tok>
<ntNode>
<tok><orth>myes</orth>
<lemma>mye</lemma>
<trans>fire</trans>
<pos>NOM:inan</pos></tok><ntNodeCat>NP-bound</ntNodeCat>
<desc><case>Instr</case></desc></ntNode>
<tok check="ok"
id="v6"><orth>tshig·gam</orth>
<lemma>tshig</lemma>
<trans>burn [-ctr]</trans>
<pos>VFIN</pos>
<desc><feature type="morph">QF</feature>
<frame>
<complement><role>U2</role><case>Abs</case></complement>
<complement
status="bound"><role microRole="cause">INSTR</role><case>Instr</case></complement></frame>
<realFrame>
<realComplement id="v6c1"><role>U2</role></realComplement>
<realComplement id="v6c2" status="bound"><role microRole="cause">INSTR</role></realComplement>
</realFrame></desc></tok><clauseCat>quest.alt1</clauseCat></clause>
<clause>
<tok check="ok"
id="v7"><orth>ma·tshig</orth>
<lemma>tshig</lemma>
<trans>burn [-ctr]</trans>
<pos>VFIN</pos>
<desc><feature type="morph">NEG</feature>
<frame>
<complement><role>U2</role><case>Abs</case></complement>
<complement
status="bound"><role microRole="cause">INSTR</role>
<case>Instr</case></complement></frame>
<realFrame>
<realComplement id="v7c1" status="empty"><role>U2</role><ref
target="v6c1"></ref></realComplement>
<realComplement id="v7c2" status="bound-empty"><role microRole="cause">INSTR</role>
<ref target="v6c2"></ref></realComplement></realFrame></desc>
</tok><clauseCat>quest.alt2</clauseCat></clause><punct>|</punct></s>
<s>
<clause>
<ntNode>
<tok><orth>mtsho</orth>
<lemma>mtsho</lemma>
<trans>lake</trans>
<pos>NOM:inan</pos></tok>
<tok><orth>Dam·le·dbal·mtsho</orth>
<lemma>Dam·le·dbal·mtsho</lemma>
<trans>-</trans>
<pos>NAME</pos></tok><ntNodeCat>NP</ntNodeCat>
<desc><case>Abs</case></desc></ntNode>
<tok><orth>ni</orth>
<lemma>ni</lemma>
<trans>-</trans>
<pos>TOP</pos></tok>
<tok check="ok"
id="v8"><orth>skams·sam</orth>
<lemma>skam</lemma>
<trans>be, become dry</trans>
<pos>VFIN</pos>
<desc><feature type="morph">QF</feature>
<frame>
<complement><role>U1</role><case>Abs</case></complement></frame>
<realFrame>
<realComplement id="v8c1"><role>U1</role></realComplement></realFrame></desc><note>...</note>
</tok><clauseCat>quest.alt1</clauseCat></clause>
<clause>
<tok check="ok"
id="v9"><orth>ma·skams</orth>
<lemma>skam</lemma>
<trans>be, become dry</trans>
<pos>VFIN</pos>
<desc><feature type="morph">NEG</feature>
<frame>
<complement><role>U1</role><case>Abs</case></complement></frame>
<realFrame>
<realComplement id="v9c1" status="empty"><role>U1</role>
<ref target="v8c1"></ref></realComplement></realFrame></desc>
</tok><clauseCat>quest.alt2</clauseCat></clause></s></q>
<tok><orth>šes</orth>
<lemma>ces</lemma>
<trans>so, thus</trans>
<pos>QoADV</pos></tok>
<tok check="ok"
id="v10"><orth>mchi</orth>
<lemma>mchi</lemma>
<trans>say, speak</trans>
<pos>VFIN</pos>
<desc>
<frame>
<complement><role>eA2+</role><case>Erg</case></complement>
<complement
status="omissible"><role>R</role><case>~Loc</case></complement>
<complement><role>CONTdir</role><case>Abs</case></complement></frame>
<realFrame derived="honorific">
<realComplement id="v10c1"><role>demA</role><case>Abl</case></realComplement>
<realComplement id="v10c2" status="omitted"><role>R</role><case>~Loc</case>
<comment><ref targType="ø" target="v3c1"></ref></comment></realComplement>
<realComplement id="v10c3"><role>CONTdir</role><case>Abs</case></realComplement></realFrame></desc>
</tok><clauseCat>endchain</clauseCat></clause><punct>|</punct></s></text>
|
Layout:
Christoph Singer. Responsible for the content: B. Zeisler. Last
modified: 31.03.2009
|