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

• The sequence oa at the beginning of the word is always a diphthong (SV or SW): oarecare'some' SV/CV/CW/CV, oaspete 'guest' SWC/CV/CV, oaste 'army' SWC/CV, oaze 'oases' SW/CV.
  
  
• The sequence ou at the end of the word always represents a diphthong (WS): birou 'office' CV/CWS, bou 'ox' CVS, erou 'hero' V/CWS, nou 'new' CVS.
  
  
• The sequence eu at the end of the word always represents a diphthong (WS): antreu 'hall' VC/CCWS, aoleu 'ouch' V/V/CWS, apogeu 'apogee' V/CV/CWS. The only exception is represented by the words which preserve the etymological orthography (bleu 'azure' CCV).

2.2.1.2.3. The phonological context

• After certain consonantic phonemes such as [f, g, h, s], the group of letters oa is always a diphthong (SV or SW): foame 'hunger' CSW/CV, goange 'insects' CSWC/CV, hoarde 'hordes' CSWC/CV, soare 'sun' CSW/CV. We have isolated over sixty phonematic contexts that define the value of the vowel and semivowel from the group oa.
  
  
• The sequence ia is always a diphthong after consonantic phonemes such as [g, l, r, s, t, z]: aliaj 'alloy' V/CV/WC, material CV/CV/CV/WC, siamez 'Siamese' SV/V/CWC, ahtiat 'craving' VC/CV/WC, parizian 'Parisian' CV/CV/CV/WC, plagiator 'plagiarist' CCV/CV/V/CWC; after [d, f, v], ie is a diphthong, with only a few exceptions (diacritic CV/V/CCW/CVC, but diavol 'devil' CSW/CVC).
  
  
• The sequence ie at the end of the word is always a hiatus (V/V) if it is preceded by a consonant: falie 'fault' CW/CV/V, mafie 'mob' CW/CV/V, bucurie 'joy' CV/CV/CW/V.

2.2.1.2.4. The accent is the simplest and safest way to distinguish between a hiatus and a diphthong: the sequence ai is a hiatus when /i/ is a stressed vowel (aici 'here' V/WC$, cais 'apricot tree' CV/WC, mozaic 'mosaic' CV/CV/WC); when /i/ is not stressed, the sequence ai can be either a hiatus (mozaic 'Mosaic' CV/CW/VC, naivitate 'naivety' CV/V/CV/CW/CV) or a diphthong (faimos 'famous' CVS/CWC, trainic 'durable' CCWS/CVC). In Romanian, the accent is not fixed, so it can be placed in any position. During flexion, it can change its place, for example from the final vowel of an infinitive on an anterior vowel at present indicative mood, first person singular: a dezdoi 'to straighten' CVC/CV/W > CVC/CWS, a ostoi 'to quiet' VC/CV/W > VC/CWS, a atribui 'to assign' V/CCV/CV/W > V/CCW/CVS, a bubui 'to thunder' CV/CV/W > CW/CVS, etc.

The sequence ea is always a diphthong after [v] (a avea 'to have' V/CSW, veac 'century' CSVC), with only one exception in the stock of our database, when [e] is stressed (hevea CV/CW/V). In order to establish the phonematic value of the group ea, we have isolated over sixty contexts.

As far as possible, we have avoided to use the stress criterion, because it becomes difficult to automatically syllabicate texts where the stress was not marked from the very beginning.

2.2.1.2.5. A recourse to the morphological structure of the word can be very useful. We shall give some examples to illustrate this.

After the consonants [d, m, n, p, v], the digraph oa represents a diphthong (SW or SV), with the exception of the compounds which are formed with a prefix, such as amino (aminoacid), auto (autoaprindere 'self-ignition'), hipo (hipoacid 'hypoacid'), mono (monoachene 'monoachenes'), pseudo (pseudoartist), termo (termoacustica 'the thermoacoustics'), added to words which begin with an [a].

2.2.1.2.6. Here are a few examples of letter groups representing V, S, $: aa, ae, aea, ai, aia, aie, aio, aiu, ao, aoa, au, aua, aue, aui, aya, ea, ee, eea, eeu, ei, aia, aie, eio, eioa, aiu, ao, eoa, eoi, eu, eua, aui, ie, ii, io, ioa, ioai, ioaie, ioi, iu, iua, iui, oa, oai, oaie, oau, oe, oi, oia, oie, oio, oiu, oo, ou, oua, oy, ua, ue, ui, uia, uie, uo, uu, ya, ye, etc.

2.2.1.2.7. These are a few simpler examples of analysis of vowel groups:

• The group /iu/ is always a diphthong at the beginning of the word, with a fully voiced [u] (iubire 'love' SV/CW/CV); at the end of the word, there can be three situations: (1) a diphthong formed of a fully voiced vowel [i] and a semivowel [w] (auriu 'golden' V/V/CWS, viu 'alive' CVS); (2) a diphthong in which the fully voiced vowel is [u] (beriliu 'beryllium' CV/CW/CSV); (3) [u] is a vowel which can form a syllable and the letter i has no phonematic value (colegiu 'college' CV/CW/C$V, calciu 'calcium' CWC/C$V); in medial position it may also create three situations: (1) both letters represent syllabic vowels (stagiune 'theatrical season' CCV/CV/W/CV, uniune 'union' V/CV/W/CV); (2) the letters represent a fully voiced vowel [u] and a semivowel (aiurit 'foolish' V/SV/CWC); (3) the letter i has no phonological value (giulgiu 'shroud' C$WC/C$V).
  
  
• The group of letters eo: at the beginning of the word it is always a hiatus (eolian 'aeolian' V/V/CV/WC); in medial position it is generally a hiatus (adeseori 'often' V/CV/CV/WCP); there are a few exceptions, when it is preceded by a consonantic group placed at the beginning of the word, such as [#bl] or [#fl] (a bleojdi 'to fade' CCSVC/CV), or followed by [r-p] and in compounds; it appears very rarely at the end of the word as a diphthong (aoleo 'ouch' V/V/CSW, huideo 'go to hell' CVS/CSW) or as a hiatus (video CW/CV/V, stereo CCW/CV/V).

191