UNL2010

Revision as of 17:59, 4 November 2009

The guidelines here stated were first derived from the fully-manual enconverting experience of translating the integral text of Cratylus, by Plato, from English into UNL (the Cratylus Project), and have been continuously extended and amended in order to be as comprehensive as possible. They were built in order to normalize and standardize enconverting strategies (which otherwise would be subject to a somewhat undesirable variance; and to guide the development of the natural language generation grammars (which would benefit from these standards as a sort of English-to-UNL transfer grammar).

As a general enconverting policy, we have tried to follow the UNL 2005 Specifications (version of June 7, 2005, published by the UNL Centre) as close as possible. This is to say that texts should be treated as semantic networks, where paragraphs and sentences are represented as hypernodes, which in turn are represented as sets of binary relations between annotated nodes (representing both words, either simple, compound or complex — the so-called UWs; and clauses, either subordinate, embedded or coordinative — the so-called SCOPEs).

Nevertheless, these guidelines should not be taken as the UNL Specifications themselves, as long as 1) they are rather experimental and tentative; 2) they differ, in several points, from the current version of the Specifications; 3) they do not follow some of the existing enconverting policies; and 4) they are not provided and have not been approved yet by the UNL Centre.

GENERAL PRINCIPLES

These guidelines are derived from two main premises:

• The UNL representation is an interpretation rather than a translation of a given text. The main goal of the UNLization process is to represent the knowledge structure of the source text, which should be detached from its verbal structure. This means that the UNL representation should not be committed to replicate the lexical and the syntactic choices of the original, but should focus in representing, in a language-independent and non-ambiguous format, one of its possible readings, preferably the most conventional one.
• The UNL representation should be as semantically complete as possible. This means that, whenever possible, all the semantic valencies of the original should be saturated, including anaphora, ellipses, presuppositions and implicatures. The UNL representation will be therefore much more verbose and redundant than the original.

UNL EXPRESSION

For the time being, the network macrostructure has not been addressed, but it seems clear that the relation “nxt”, proposed by the UNLCenter to link sentences and paragraphs, is syntactic rather than semantic, and it is not appropriate for a network claimed to be mostly "semantic". Some alternatives have been considered (especially Discourse Representation Theory - DRT, proposed by Hans Kamp; and Rhetorical Structure Theory - RST, proposed by William Mann and Sandra Thompson), but they are still under investigation.

The document structure is also subject to change, and it is likely to move to a XML schema, which is still under development. For the moment, the syntax defined by the UNLCenter has been kept.

THREE-LAYERED REPRESENTATION

The basic assumption of the UNL approach is that the meaning conveyed by natural language sentences can be formally represented through three different types of semantic units: UWs, attributes and relations. This three-layered representation model is the cornerstone of UNL and its most distinctive feature over other semantic networks, which normally proposes only two levels: edges and vertices. Nevertheless, it poses several problems to the UNLization as the distinction between what is supposed to be represented by each unit is not always clear. In order to avoid superposition and to facilitate the enconversion process, we have tried to clearly identify the scope of each unit using the following procedures:

• RELATIONS should represent only syntactic relations (subject, object, complement, adjunct) with their corresponding semantic value;
• UWs should represent lexemes from open classes:
  • nouns, including proper nouns, abbreviations and acronyms;
  • adjectives;
  • full verbs;
  • adverbs and adverbials; and
  • numbers (to be always represented as Arabic numerals
• ATTRIBUTES should represent bound morphemes, closed classes and context-dependent information:
  • grammatical categories (gender, number, tense, aspect, mood, voice, etc)
  • determiners (articles and demonstratives);
  • adpositions (prepositions, postpositions and circumpositions);
  • auxiliary and quasi-auxiliary verbs (auxiliaries, copula, modals, coverbs, preverbs);
  • interjections;
  • conjunctions;
  • text structure (.@entry, .@topic, .@qfocus, .@emphasis, .@relative, etc);
  • speech acts (.@request, .@suggestion, .@offer, etc);
  • other context-dependent information (such as politeness, metaphor, irony, etc);

Pronouns and pro-forms are expected to be replaced by their antecedents and not to be represented in UNL, except in case of exophoric reference (indefinite pronouns, interrogative pronouns and personal pronouns that are not coindexed to any existing antecedent).

The main changes concerning the present UNL Specifications are the following:

RELATIONS

The set of relations is exactly the same as defined in the UNL 2005 Specifications, with a single difference: the introduction of the relation “voc”, which stands for vocative, because of sentences like:

[S:20] Son of Hipponicus, there is an ancient saying, that "hard is the knowledge of the good.”

In cases like this, the attribute .@vocative cannot be applied, because the vocative (“Son of Hipponicus”), which seems to be semantically isolated from the main clause, cannot be introduced by any other relation of the UNL Specification.

ATTRIBUTES

The set of attributes has been substantially increased to represent information concerning grammatical categories, determiners, adpositions and conjunctions. The main additions are the following:

• gender: @male, @female
• degree of comparison: @more, @less, @equal, @most, @least
• demonstrative: @proximal, @medial, @distal
• preposition: @under, @below, @above, @after, @before, etc.
• conjunction: @before, @after, etc.
• relative (for the main entry of relative clauses): @relative

The decision to represent closed classes as attributes instead of UWs has led to a different way of representing several natural language phenomena:

this X

UNL Centre: mod(X, this)

These guidelines: X.@proximal

X is under Y

UNL Centre: plc(X, under), obj(under, Y)

These guidelines: plc(X, Y.@under)

bigger than Y

UNL centre: man(big, more), bas(big, Y)

These guidelines: bas(big.@more, Y)

etc.

Additionally, the following general principles were adopted:

• interjections, filled pauses, phatic expressions and short answers should be represented by the null UW (to be represented as "00") together with the attribute indicating the corresponding speech act (.@confirmation, .@surprise, etc).
• the attribute .@entry (mandatory in every scope, including the main one) should be placed at the left (source) side of at least one relation;
• the difference between mentioning and using a word (which is a quite frequent situation in a metalinguistic text such as Cratylus) should be represented by the attribute .@mention (which is not the same as "quotation");
• attributes should be used in alphabetical order (“.@entry.@past” instead of “.@past.@entry”).

UNIVERSAL WORDS

The syntax of UWs is exactly the same proposed in the UNL 2005 Specifications, but there have been many changes in the process of labeling and defining UWs. The current set of UWs — provided by the UNL Center and available at http://www.undl.org/unlsys/uw/UNLKB_tree.txt or through the UW Gate — has proved already not to be stable and not to be consistent (see MARTINS & NUNES 2005). Because of that, we have decided to use the set of UWs provided by the UNL Wordnet (UNLWN) Project, available at http://www.ronaldomartins.pro.br/unlwordnet. Specific procedures for labeling UWs not included in the UNLWN are referred to in the English-to-UNL Grammar (see ABBREVIATION; ACRONYM; ADJECTIVE; ADVERB; COMMON NOUN; CONJUNCTION; DETERMINER; INTERJECTION; PREPOSITION; PRONOUN; PROPER NOUN; QUANTIFIER; VERB)

Considering those changes, we have prepared these guidelines, which comprise the main grammar categories and detailed instructions on how to represent them in UNL. In order to address English grammar categories, examples and metalanguage were extracted from many different grammars and scientific papers on English. The main references adopted for the present version were the following:

• QUIRK, R. & GREENBAUM, S. (1996[1973]). A University Grammar of English. Harlow, England: Addison Wesley Longman Ltd.
• MURPHY, R. (1991[1985]). English Grammar in Use. Cambridge: Cambridge Unviersity Press.
• SWAN, M. (1980[2000]). Practical English Usage. Oxford: Oxford University Press.
• SERPA, O. (1984[1967]). Gramática da Língua Inglesa. Rio de Janeiro: MEC/FAE.
• WIKIPEDIA, at http://www.wikipedia.org.

Finally, we ought to stress that the enconverting standards here presented are tentative and provisional, and they are subject to improvements and changes as soon as they were proved not to be the most adequate ones. In order to provide such enhancements, we would invite UNL Society members and other people interested in UNL to criticize them, to propose alternatives and to help us build an English-to-UNL transfer grammar as comprehensive as possible.