|
|
| Line 1: |
Line 1: |
| '''D-rules''' (declaration rules) are used to make statements inside the UNL<sup>arium</sup> framework. They are used to assign priorities (disambiguation rules), to map attributes and relations (mapping rules) and to state identities (replacement rules). | | '''D-rules''' (disambiguation rules) are used to prevent wrong lexical choices, to provoke best matches and to check the consistency of graphs, trees and lists. Differently from [[Grammar|T-rules]], they do not provoke transformations, but induce or block them, by assigning priorities to natural language phenomena. |
|
| |
|
| == Syntax == | | == Syntax == |
| D-rules are always composed of two fields, which are separated by "=". They are always ended by a ";". | | D-rules follow the general syntax: |
| <DEFINIENDUM> = <DEFINIENS>;
| |
| Where:<br />
| |
| *<DEFINIENDUM>, the term to be defined, may be:
| |
| **an attribute or a value of an attribute, of UNL (@pl, @male, @future) or of NL (POS, NOU, GEN, MCL, etc);
| |
| **a node, to be represented as a "string" (between parentheses) or as a [lemma] (between square brackets);
| |
| **a relation, either syntactic or semantic (such as "agt", "obj", "VS", "XP"), with the corresponding constituents; or
| |
| **a sequence of attributes, nodes and relations.
| |
| *<DEFINIENS>, the definition, may be:
| |
| **an attribute or a value of an attribute, of UNL (@pl, @male, @future) or of NL (POS, NOU, GEN, MCL, etc);
| |
| **a node, to be represented as a "string" (between parentheses) or as a [lemma] (between square brackets);
| |
| **a relation, either syntactic or semantic (such as "agt", "obj", "VS", "XP"), with the corresponding constituents;
| |
| **a sequence of attributes, nodes and relations; or
| |
| **a numeric value indicating the degree of certainty, ranging from 0 (=impossible) to 255 (=necessary).
| |
|
| |
|
| == Types of D-rule == | | STATEMENT=P; |
| There are three main types of D-rule:
| |
| *'''Disambiguation rules''' are used to state priorities.
| |
| *'''Mapping rules''' are used to state correspondences between UNL and NL.
| |
| *'''Replacement rules''' are used to state correspondences inside UNL or NL.
| |
|
| |
|
| === Disambiguation rules ===
| | Where<br /> |
| Disambiguation rules are used to state priorities. They follow the general syntax:
| | STATEMENT is the left side (condition) of a [[L-rule]] or a [[S-rule]]; and<br /> |
| <DEFINIENDUM> = VALUE;
| | P, which can range from 0 (impossible) to 255 (necessary), is the probability of occurrence of the STATEMENT<br /> |
| Where <VALUE> is a numeric value that may range from 0 (impossible, or FALSE) to 255 (necessary).
| |
| Disambiguation rules normally apply over the left side of [[L-rule]]s and [[S-rule]]s.
| |
|
| |
|
| ==== Examples of disambiguation rules ====
| | == Examples == |
| *("the",ART)("a",ART)=0; (It's not possible that the article "a" follows the article "the") | | *List structures |
| *("afraid",ADJ)(BLK)("of",PRE)=255; (It's absolutely necessary that the preposition "of" follows the adjective "afraid" after a blank space). | | **(ART)(BLK)(VER)=0; (an article (ART) may not precede a verb (VER)) |
| *VH(NOU)=0; (It's not possible that the head of a verbal phrase be a noun) | | **(ART)(BLK)(NOU)=255; (articles (ART) always precede nouns (NOU)) |
| *VH(VER)=255; (It's absolutely necessary that the head of a verbal phrase be a verb) | | *Syntactic structures |
| | | **agt(VER;ADJ)=0; (an adjective (ADJ) may not be an agent (agt) of a verb (VER)) |
| === Mapping rules ===
| | **agt(VER;NOU)=255; (agents (agt) of verbs (VER) are always nouns (NOU)) |
| Mapping rules are used to state correspondences between UNL and NL. They follow the general syntax:
| | **VS(VER;ADJ)=0; (an adjective (ADJ) may not be an specifier (VS) of a verb (VER)) |
| <DEFINIENDUM> = <DEFINIENS>;
| | **NS(NOU;DET)=255; (determiners (DET) are always specifiers (NS) of nouns (NOU)) |
| Where <DEFINIENDUM> is an element of UNL and <DEFINIENS> is an element of NL, or vice-versa.
| |
| | |
| ==== Examples of mapping rules ====
| |
| *From UNL to NL | |
| **@pl = PLR; | |
| **@past = PAS;
| |
| **@passive = PSV;
| |
| **@male = MCL;
| |
| **@past,@progressive = PAS,PGS;
| |
| *From NL to UNL
| |
| **PLR = @pl;
| |
| **PAS = @past
| |
| **MCL = @male
| |
| **PAS,PGS = @past.@progressive
| |
| | |
| === Replacement rules ===
| |
| Replacement rules are used to state correspondences inside UNL or NL. They follow the general syntax:
| |
| <DEFINIENDUM> = <DEFINIENS>;
| |
| Where <DEFINIENDUM> and <DEFINIENS> are both elements of UNL, or are both elements of NL.<br />
| |
| Replacement rules are mainly used to avoid redundancy and unnecessary proliferation of rules.
| |
| | |
| ==== Examples of replacement rules ====
| |
| *Inside UNL
| |
| **@dual = @pl; (replace the attribute @dual by the attribute @pl)
| |
| *Inside NL
| |
| **DUA = PLR; (replace the attribute DUA by the attribute PLR)
| |
D-rules (disambiguation rules) are used to prevent wrong lexical choices, to provoke best matches and to check the consistency of graphs, trees and lists. Differently from T-rules, they do not provoke transformations, but induce or block them, by assigning priorities to natural language phenomena.
Syntax
D-rules follow the general syntax:
STATEMENT=P;
Where
STATEMENT is the left side (condition) of a L-rule or a S-rule; and
P, which can range from 0 (impossible) to 255 (necessary), is the probability of occurrence of the STATEMENT
Examples
- List structures
- (ART)(BLK)(VER)=0; (an article (ART) may not precede a verb (VER))
- (ART)(BLK)(NOU)=255; (articles (ART) always precede nouns (NOU))
- Syntactic structures
- agt(VER;ADJ)=0; (an adjective (ADJ) may not be an agent (agt) of a verb (VER))
- agt(VER;NOU)=255; (agents (agt) of verbs (VER) are always nouns (NOU))
- VS(VER;ADJ)=0; (an adjective (ADJ) may not be an specifier (VS) of a verb (VER))
- NS(NOU;DET)=255; (determiners (DET) are always specifiers (NS) of nouns (NOU))
