Database of the Romanian Language Phonetics and Phonology

Felicia ªerban, Dana Bucerzan, Luciana Peev, Lidia Bibolar

1. Synopsis

The Database of the Romanian Language is a computerised dictionary that contains in a structured form 60,000 basic entries with phonetic and morphological information. The structures of the database have been shaped so as to allow the elaboration of theoretical studies as well as a wide range of applications in text processing.

The dictionary is supposed to be flexible. One can easily extract the necessary information for any specific application. New information can be added (syntactic or semantic attributes) for further developments.

From the entire database project, we intend to present the section of phonetics and phonology in detail. The other sections (morphology, lexicon and semantics as well as stylistics) will be only briefly mentioned.

The computerised lexical and phonological system 'FONO' is an informatised product (a soft) which has been conceived to allow the solving of a wide range of phonetic and phonological problems. It can also be useful in the area of processing Romanian texts. The analysis soft includes a set of rules that are applied in the very moment of writing the words.

The algorithm of 'FONO' uses the phonetic and phonological information from the database to allow

• solving of the syllabication of the Romanian words;
• statistical studies on the frequency of the phonological units in Romanian texts;
• lexical and morphological studies about the migration of accents in the paradigm of a word;
• studies on the phonological sequence of words;
• studies of phonetic problems.

The soft includes a 'filter' which, for a certain sequence of letters (coded or in clear), selects the list of all the words which include it, at the beginning, at the end, or in any other position.

A facility of adding observations about any word which can be grouped in a different 'dictionary' has been also created.

2. Elements of phonology and orthography of the Romanian language

2.1. Letters and phonemes

It is indeed very difficult to perform automatic syllabication (from a letter to a phoneme) in Romanian, even though some say that its system of writing is phonetic (more accurately, phonological), i.e. one writes exactly as he or she pronounces the word [1,2]. Firstly, it is not that easy to establish an equivalence between letters and phonemes, because certain graphical signs can be used to represent different phonemes.

The Romanian language phonemes can be: consonantic, vocalic, semivocalic (semiconsonantic), and 'pseudovocalic' (non-syllabic deaf vowel that is only a marker of the palatalization of the preceding consonant). We use the following symbols: V = unstressed vowel sound; W = stressed vowel sound; S = semivowel; P = pseudovowel (soundless i); $ = the letter e or i that has no phoneme value; C = consonant; GG = monophonematic consonantic sequence; # = the pause between words; / = syllabic cut; / / = phonological transcription; [ ] = phonetic transcription.

188