D4.4b: A new definition of Word Sketches for Norwegian

This report describes the new sketch grammar for Norwegian created within the scope of the project.

State of the art

The original sketch grammar for Norwegian contained only universal relations connected to the part-of-speech tagging, e.g. "noun on the left", "verb on the right", etc. The syntactic and semantic relationships among the words were not clear from the word sketch, and although there were some important collocations recorded, there was a lot of noise as well.

New sketch grammar

The new Norwegian sketch grammar was developed along the template used in the English and Spanish grammars within the Sketch Engine (so far the two most developed grammars), in terms of names of the grammatical relations (and their mapping to the template names which will enable matching the Norwegian word sketches to word sketches in other languages, in the Bilingual word sketch application) as well as in terms of the coding within the scope of the sketch grammar formalism which should simplify reading the grammar and make future modifications easy. The definitions of the relations are based on the Oslo-Bergen tagset for Norwegian.

In total, there is 19 relations in the current version of the grammar, covering the most important grammatical phenomena, such as modifiers of all parts of speech, subjects, objects, predicates and coordinations.

Figure 1: Word sketch for noun katt (cat)

As can be seen from Figures 1-3, modifiers are the words that modify or specify meaning of given word. This relation is called modifier of “word”. There are also few complement relations called nouns/verbs/adverbs modified by “word”, where the headword is the one who modifies the meaning (of noun, adjective, adverb or verb). These can be seen in Figure 3.

Figure 2: Word sketch for verb leke (to play)

Next grammatical phenomena to be described are subjects and objects, which can be either nominal or pronominal. These relations (objects of “word”, subjects of “word”, pronominal objects of “word”, pronominal subjects of “word”) can be seen in Figure 2. A somewhat special case is relation subject of “be word”, which is shown in Figure 3. Partially conjoining relation to subject of be “word” is “word” is… which can be seen in Figure 1 as well as relations “word” is a ... and … is a “word”. Another important grammatical phenomenon is predicate. Word sketches show two relations, verbs with “word” as object and verbs with “word” as subject. These relations can be seen in Figure 1.

Figure 3: Word sketch for adjective svart (black)

The last relation is “word” and/or, which shows frequent coordinations among words. This relation is shown in Figures 1-3.

The rules for the particular grammatical relations were developed in collaboration with native speakers of Norwegian who recorded their linguistic knowledge, needed to build a sketch grammar, as answers to specific language questions in a computer-aided environment. These answers were then manually reviewed and converted into a formal sketch grammar. The new version of the grammar was reviewed by native speakers of Norwegian and it was evaluated as clearly better and more usable than the previous universal grammar.

Last modified 5 years ago Last modified on May 30, 2017, 11:01:28 PM

Attachments (3)

Download all attachments as: .zip