Felicia ªerban & al. * Database of the Romanian Language Phonetics and Phonology

Some prefixes and elements of the compound words are easily recognisable: acro, agro, algie, anglo, arier, atot, auto, balneo, biblio, cvadri, cvazi, echi, edru, juris, latrie, lipo, manie, mecano, medico, mente, metrie, mono, moto, neuro, orto, penta, poli, pseudo, radio, rare, scop, semi, tele, trans, tri, uni, vice, zaur, etc. [12]. Others are built from a sequence which is too short to enable automatic recognition: bi (bioxid 'dioxide') compared to bio (biochimist 'biochemist'), pre (preaviz 'previous notice' CCV/V/CWC) and prea (preaplin 'superabundance' CCSV/CCWC). Such cases need more information than a simple marker of the presence of a suffix or another element of composition.

It must be taken into account the fact that Romanian orthography requires morphological syllabication even when the elements which compose the word do not exist as independent units in Romanian: ab/laut (not a/blaut, as phonetic syllabication requires), auto/psie 'autopsy' (not autop/sie), steto/scop 'stethoscope' (not stetos/cop), trans/atlantic (not tran/satlantic).

2.2.3. Conclusion

Romanian syllabication rules presuppose a high degree of linguistic culture of the person who writes. They are also very complicated for a possible formalisation. The need for their simplification was stated on many occasions [12]. Thus, the elaboration of automatic syllabication is a very laborious operation.

3. Morphology

The computerised morphological system 'LEX' is a specialised software for solving the theoretical and practical problems of Romanian morphology.

The main fields of usage of the system are:

• the morphological automatic analysis of a Romanian text;
• the generation of the whole paradigm of a word starting from the lemma and using all the associated attributes;
• a spell-checker for Romanian texts;
• a didactic aid for computer-assisted teaching of the Romanian language in high school and university;
• a system of statistical studies on the Romanian vocabulary;
• a system for lexical studies on literary works.

The paradigm is automatically generated from the coded root and the associated flexional class [15].

The Database of the Romanian Language as a source/target language, as an interdisciplinary research derived from the projects mentioned above, is conceived for teaching and learning Romanian as a foreign language. It offers vocabularies on various levels and specialities (restricted lexicon, enlarged lexicon, complete lexicon for special fields, such as medicine, economy, etc.) with the indispensable grammatical information in a number of languages for those who are interested in this topic.

4. Lexicography and stylistics

4.1. Author dictionaries

At the beginning, the stylistics section has to combine lexicography with the study of individual style. Because textual analysis means a full knowledge of the lexical inventory, author dictionaries are needed. A first variant of such a dictionary has the form of lexical concordances in the works of great Romanian poets and prose writers. A similar work has been done at the Centre for Text Analysis (Faculty of Letters, Cluj-Napoca). Besides the interest that such lexicographic studies may present for those interested in the respective writers' work, it may also be a valuable source for completing the information on the development of Romanian language from different regions.

4.2. The thesaurus-dictionary of the Romanian language

The project of an explanatory, historical, normative, and etymological dictionary was initiated by the Romanian Academy, under the supervision of the great linguist Sextil Puºcariu. After World War, this activity was shared between Bucharest, Cluj-Napoca and Iaºi. The last volumes are presently in course of elaboration. Twelve volumes (12 tomes) have been printed up to the present moment.

An electronic variant could help its circulation in a more practical form for its users. Because it has been written in a period of almost a century, its volumes do not mirror the same stage in the evolution of the Romanian language and do not use identical sources. In its electronic variant, the dictionary must be actualised as an inventory of words, meanings, attestations, and could become a continuously enriched lexicographic work.

192