Principles for Annotation
Prefer Core Roles over Non-core Roles
Thus:
Dependency graph for sentence: Kim bought a house. Token bought is labeled POSSESSION-INIT and has an edge labeled possessor to token Kim and an edge labeled possessum to token house.
Dependency graph for sentence: Kim wants to swim. Token wants is labeled META-NORM and has an edge labeled participant to token Kim and an edge labeled scene to token swim.
Rather than:
Dependency graph for sentence: Kim bought a house. Token bought is labeled POSSESSION-INIT and has an edge labeled CAUSER to token Kim and an edge labeled possessum to token house.
Dependency graph for sentence: Kim wants to swim. Token wants is labeled META-NORM and has an edge labeled EXPERIENCER to token Kim and an edge labeled scene to token swim.
Try Adding Arguments
When not sure how to frame a predicate, it can help to think of additional
arguments that the predicate could have and that are just not realized in this
instance. For example, what is the role of It in It is not important? If we
add an argument, we might get It is not important for the process, where
INFLUENCE seems to fit with It as influence and the
process as affected. Thus:
Dependency graph for sentence: It is not important ( for the process ). Token not is labeled NEG and has an edge labeled negated to token important. Token important is labeled INFLUENCE and has an edge labeled influence to token It and an edge labeled SCENE to token not and an edge labeled affected to token process.
Similarly, what is the role of her petals in She adjusted her petals? We can add an argument to adjust to make it clearer:
Dependency graph for sentence: She adjusted her petals ( to how she liked them ). Token adjusted is labeled COMPARISON-INIT and has an edge labeled CAUSER to token She and an edge labeled compared to token petals and an edge labeled reference to token liked.
Similarly, seems could be MESSAGE or REALITY. Because we can add an
EXPERIENCER argument, we prefer MESSAGE:
Dependency graph for sentence: Kim seems happy ( to Sandy ). Token seems is labeled META-MESSAGE and has an edge labeled participant to token Kim and an edge labeled scene to token happy and an edge labeled EXPERIENCER to token Sandy.
Alternative Framings
Often, both a literal and a more figurative reading of a predicate are salient
and compete for the framing. You can annotate both possibilities: separate the
literal and the figurative frame/role by the symbol >>.
Dependency graph for sentence: One loves the sun when one is so sad. Token loves is labeled EVALUATION and has an edge labeled EXPERIENCER to token One and an edge labeled evaluee to token sun and an edge labeled TIME>>CONDITION to token sad.
Dependency graph for sentence: some sort of miraculous apparition must emerge from the bud. Token emerge is labeled LOCATION-DEINIT>>INIT and has an edge labeled has-location>>initiated to token apparition and an edge labeled location>>CAUSER to token bud.
Dependency graph for sentence: Embarassed , she coughed two or three times. Token coughed is labeled CONTAINMENT-DEINIT>>EXPRESSION and has an edge labeled ACCOMPANIER to token Embarassed and an edge labeled container>>sender to token she and an edge labeled TIME to token times.
If two framings seems equally plausible to you and are not in a
literal-figurative relationship, use || instead of >>.
Use Comments if Not Sure
You are encouraged to use comments to note down any doubts and thoughts.
Comments can be added to the end of any line using the # symbol:
[TIME] generally (19) # ?
[has-time] fair (20)
[has-time] to - middling (24)
Obliqueness
Some predicates, such as friend or identical, denote an inherently symmetric relationship. Thus, their arguments are semantically virtually indistinguishable. In such cases, we default to giving the syntactically less oblique realized argument the first role, and the more oblique one, if any, the second. For our purposes, obliqueness is defined as follows:
| Obliqueness | Syntactic role | Dependency label |
|---|---|---|
| least oblique | Subject | nsubj |
| direct object | obj | |
| indirect object | iobj | |
| most oblique | oblique nominal, nominal modifier | obl, nmod |
Thus we have, e.g.:
Dependency graph for sentence: Kim 's friend. Token friend is labeled BOND and has an edge labeled tied to token Kim.
Dependency graph for sentence: I am Kim 's friend. Token friend is labeled BOND and has an edge labeled tied to token I and an edge labeled tied-to to token Kim.
Dependency graph for sentence: Kim is comparing apples with oranges. Token comparing is labeled COMPARISON-MESSAGE-INIT and has an edge labeled EXPERIENCER to token Kim and an edge labeled compared to token apples and an edge labeled reference to token oranges.
Terminology
This is a list of terms and how they are used in this manual (not necessarily in linguistics in general, though I try to be compatible with widespread usage).
| Term | Explanation |
|---|---|
| adjunct | D is an adjunct of H if D is a dependent of H and not an argument, thus is assigned a role not by H but only by the adjunct construction – this means it must have a non-core role |
| argument | D is an argument of H if D is a dependent of H and is assigned a role by it – this can be a core role or a non-core role |
| dependent | if D, H are predicates, D is said to be a dependent of H if there is an edge from D to H in the annotation (either syntactic or non-syntactic) |
| event | a synonym for eventuality; sometimes also used as a synonym for process; I try to avoid the term due to its ambiguity |
| eventuality | a state or process |
| head | H is a head of D iff D is a dependent of H |
| predicate | a content word (noun, verb, adjective, or adverb) that evokes a frame |
| process | a dynamic eventuality, i.e., a transition over time through two or more states |
| scene | a synonym for eventuality |
| situation | a synonym for eventuality |
| state | a static eventuality, thus a state of affairs that does not include change over time |