Italian phonology

Notes:

• Between two vowels, or between a vowel and an approximant (/j, w/) or a liquid (/l, r/), consonants can be both singleton or geminated. Geminated consonants shorten the preceding vowel (or block phonetic lengthening) and the first geminated element is unreleased. For example, compare /fato/ [ˈfaːto] ('fate') with /fatto/ [ˈfatto] ('fact'). However, /ɲɲ/, /ʃʃ/, /ʎʎ/, /ddz/, /tts/ are always geminated word-internally.[1] Similarly, nasals, liquids, and sibilants are pronounced slightly longer in medial consonant clusters.[2]
• /j/, /w/, and /z/ are the only consonants that cannot be geminated.
• /t, d/ are laminal denti-alveolar [t̪, d̪],[3][4][5] commonly called "dental" for simplicity.
• /k, ɡ/ are pre-velar before /i, e, ɛ, j/.[4]
• /t͡s, d͡z, s, z/ have two variants:
  • Dentalized laminal alveolar [t̪͡s̪, d̪͡z̪, s̪, z̪][3][6] (commonly called "dental" for simplicity), pronounced with the blade of the tongue very close to the upper front teeth, with the tip of the tongue resting behind lower front teeth.[6]
  • Non-retracted apical alveolar [t͡s̺, d͡z̺, s̺, z̺].[6] The stop component of the "apical" affricates is actually laminal denti-alveolar.[6]
• /n, l, r/ are apical alveolar [n̺, l̺, r̺] in most environments.[3][5][7] /n, l/ are laminal denti-alveolar [n̪, l̪] before /t, d, t͡s, d͡z, s, z/[5][8][9] and palatalized laminal postalveolar [n̠ʲ, l̠ʲ] before /t͡ʃ, d͡ʒ, ʃ/.[10][11] /n/ is velar [ŋ] before /k, ɡ/.[12][13]
• /m/ and /n/ do not contrast before /p, b/ and /f, v/, where they are pronounced [m] and [ɱ], respectively.[12][14]
• /ɲ/ and /ʎ/ are alveolo-palatal.[15] In a large number of accents, /ʎ/ is a fricative [ʎ̝].[16]
• Intervocalically, single /r/ is realised as a trill with one or two contacts.[17] Some literature treats the single-contact trill as a tap [ɾ].[18][19] Single-contact trills can also occur elsewhere, particularly in unstressed syllables.[20] Geminate /rr/ manifests as a trill with three to seven contacts.[17]
• The phonetic distinction between [s] and [z] is neutralized before consonants and at the beginning of words: the former is used before voiceless consonants and before vowels at the beginning of words; the latter is used before voiced consonants (meaning [z] is an allophone of /s/ before voiced consonants). The two can contrast only between vowels within a word, e.g. [ˈfuːzo] 'melted' vs. [ˈfuːso] 'spindle'. According to Canepari,[19] though, the traditional standard has been replaced by a modern neutral pronunciation which always prefers /z/ when intervocalic, except when the intervocalic s is the initial sound of a word, if the compound is still felt as such: for example, presento /preˈsɛnto/[21] ('I foresee', with pre meaning 'before' and sento meaning 'I perceive') vs presento /preˈzɛnto/[22] ('I present'). There are many words in which dictionaries now indicate that both pronunciations, with [z] and with [s], are acceptable. Word-internally between vowels, the two phonemes have merged in many regional varieties of Italian, either as /z/ (Northern-Central) or /s/ (Southern-Central).

Vowels of Italian. From Rogers & d'Arcangeli (2004:119).

In Italian there is no phonemic distinction between long and short vowels, but vowels in stressed open syllables, unless word-final, are long at the end of the intonational phrase (including isolated words) or when emphasized.[23][24] Adjacent identical vowels found at morpheme boundaries are not resyllabified, but pronounced separately ("quickly rearticulated"), and they might be reduced to a single short vowel in rapid speech.[25]

Although Italian contrasts close-mid (/e, o/) and open-mid (/ɛ, ɔ/) vowels in stressed syllables, this distinction is neutralised in unstressed position,[23] where only the close-mid vowels occur. The height of these vowels in unstressed position is context-sensitive; they are somewhat lowered ([e̞, o̞]) in the vicinity of more open vowels.[26] The distinction between close-mid and open-mid vowels is lost entirely in a few Southern varieties of Regional Italian, especially in Northern Sicily (e.g. Palermo), where they are realized as open-mid [ɛ, ɔ], as well as in some Northern varieties (in particular in Piedmont), where they are realized as mid [e̞, o̞].

Word-final stressed /ɔ/ is found in a small number of words: però, ciò, paltò.[27] However, as a productive morpheme, it marks the first person singular of all future tense verbs (e.g. dormirò 'I will sleep') and the third person singular preterite of first conjugation verbs (parlò 's/he spoke', but credé 's/he believed', dormì 's/he slept'). Word-final unstressed /u/ is rare, [28] found in onomatopoeic terms (babau),[29] loanwords (guru),[30] and place or family names derived from the Sardinian language (Gennargentu,[31] Porcu).[32]

When the last phoneme of a word is an unstressed vowel and the first phoneme of the following word is any vowel, the former vowel tends to become non-syllabic. This phenomenon is called synalepha and should be taken into account when counting syllables, e.g. in poetry.

In addition to monophthongs, Italian has diphthongs, but these are both phonemically and phonetically simply combinations of the other vowels, with some being very common (e.g. /ai, au/), others being rarer (e.g. /ɛi/) and some never occurring within native Italian words (e.g. /ou/). None of these diphthongs are however considered to have distinct phonemic status because their constituents do not behave differently than they would in isolation (and all occur in isolation), unlike the diphthongs in some languages like English and German. Grammatical tradition makes a distinction between 'falling' and 'rising' diphthongs; however, since rising diphthongs are composed of one semiconsonantal sound [j] or [w] and one vowel sound, they are not actually diphthongs. The practice of referring to them as 'diphthongs' has been criticised by phoneticians like Luciano Canepari.[19]

Word-initial consonants are geminated after certain vowel-final words in the same prosodic unit. There are two types of triggers of initial gemination: some unstressed particles, prepositions, and other monosyllabic words, and any oxytonic polysyllabic word.[35] As an example of the first type, casa ('house') is pronounced [ˈkaːsa] but a casa ('homeward') is pronounced [akˈkaːsa]. This is not a purely phonological process, as no gemination is cued by the la in la casa 'the house' [laˈkaːsa], and there is nothing detectable in the structure of the preposition a to account for the gemination. This type normally originates in language history: modern a, for example, derives from Latin AD, and today's geminate in [akˈkaːsa] is a continuation of what was once a simple assimilation. Gemination cued by final stressed vowels, however, is transparently phonological. Final stressed vowels are short by nature, if a consonant follows a short stressed vowel the syllable must be closed, thus the consonant following the final stressed vowel is drawn to lengthen: parlò portoghese [parˈlɔpportoˈɡeːze] 's/he spoke Portuguese' vs. parla portoghese [ˈparlaportoˈɡeːze] 's/he speaks Portuguese'.

In standard Italian, syntactic gemination occurs mainly in the following two cases:[44]

• After end-stressed words (such as sanità, perché, poté, morì and so on).
• After the words a, che, chi, come, da, do, dove, e, fa, fra, fu, gru, ha, ho, ma, me, mo' (in the phrase a mo' di), no, o, qua, qualche, qui, so, sopra, sta, sto, su, te, tra, tre, tu, va, vo.

Syntactic gemination is the normal native pronunciation in Central Italy (both "stress-induced" and "lexical") and Southern Italy (only "lexical"), including Sicily and Corsica (France).

In Northern Italy and Sardinia, speakers use it inconsistently because the feature is not present in the dialectal substratum and is not usually shown in the written language unless a new word is produced by the fusion of the two: "chi sa"-> chissà ("who knows" in the sense of goodness knows).

The above IPA symbols and description refer to standard Italian, based on a somewhat idealized version of the Tuscan-derived national language. As is common in many cultures, this single version of the language was pushed as neutral, proper, and eventually superior, leading to some stigmatization of varying accents. Television news anchors and other high-profile figures had to put aside their regional Italian when in the public sphere. However, in more recent years the enforcement of this standard has fallen out of favor in Italy, and news reporters, actors, and the like are now more free to deliver their words in their native regional variety of Italian, which appeals to the Italian population's range of linguistic diversity. The variety is still not represented in its wholeness and accents from the South are maybe to be considered less popular, except in shows set in the South and in comedy, a field in which Naples, Sicily and the South in general have always been present. Though it still represents the basics for the standard variety, the loosened restrictions have led to Tuscan being seen for what it is, just one dialect among many with its own regional peculiarities and qualities, many of which are shared with Umbria, Southern Marche and Northern Lazio.

• In Tuscany (though not in standard Italian, which is derived from, but not equivalent to, Tuscan dialect), voiceless stops are typically pronounced as fricatives between vowels.[45] That is, /p t k/ → [ɸ θ h/x]: e.g. i capitani 'the captains' [iˌhaɸiˈθaːni], a phenomenon known as the gorgia toscana 'Tuscan throat'. In a much more widespread area of Central Italy, postalveolar affricates are deaffricated when intervocalic so that in Cina ('in China') is pronounced [in t͡ʃiːna] but la Cina ('the China') is [laʃiːna], and /ˈbat͡ʃo/ bacio 'kiss' is [ˈbaːʃo] rather than Standard Italian [ˈbaːt͡ʃo].[46] This deaffrication can result in minimal pairs distinguished only by length of the fricatives, [ʃ] issuing from /t͡ʃ/ and [ʃː] from geminate /ʃʃ/: [laʃeˈrɔ] lacerò 's/he ripped' vs. [laʃːeˈrɔ] lascerò 'I will leave'.
• In nonstandard varieties of Central and Southern Italian, some stops at the end of a syllable completely assimilate to the following consonant. For example, a Venetian might say tecnica as [ˈtɛknika] or [ˈtɛɡnika] in violation of normal Italian consonant contact restrictions, while a Florentine would likely pronounce tecnica as [ˈtɛnniha], a Roman on a range from [ˈtɛnnika] to [ˈtɛnniɡa] (in Southern Italian, complex clusters usually are separated by a vowel: a Neapolitan would say [ˈtɛkkənikə], a Sicilian [ˈtɛkkɪnɪka]). Similarly, although the cluster /kt/ has developed historically as /tt/ through assimilation, a learned word such as ictus will be pronounced [ittus] by some, [iktus] by others.
• In popular (non-Tuscan) Central and Southern Italian speech, /b/ and /d͡ʒ/ tend to always be geminated ([bb] and [dd͡ʒ]) when between two vowels, or a vowel and a sonorant (/j/, /w/, /l/, or /r/). Sometimes this is also used in written language, e.g. writing robba instead of roba ('property'), to suggest a regional accent, though this spelling is considered incorrect. In Tuscany and beyond in Central and Southern Italy, intervocalic non-geminate /d͡ʒ/ is realized as [ʒ] (parallel to /t͡ʃ/ realized as [ʃ] described above).
• The two phonemes /s/ and /z/ have merged in many varieties of Italian: when between two vowels within the same word, it tends to always be pronounced [z] in Northern Italy, and [s] in Central and Southern Italy (except in the Arbëreshë community). A notable example is the word casa ('house'): in Northern Italy it is pronounced [ˈkaːza]; in Southern-Central Italy it's pronounced [ˈkaːsa].
• In several Southern varieties, voiceless stops tend to be voiced if following a sonorant, as an influence of the still largely spoken regional languages: campo /ˈkampo/ is often pronounced [ˈkambo], and Antonio /anˈtɔnjo/ is frequently [anˈdɔnjo].

The various Tuscan, Corsican and Central Italian dialects are, to some extent, the closest ones to Standard Italian in terms of linguistic features, since the latter is based on a somewhat polished form of Florentine.

Very little research has been done on the earliest stages of phonological development in Italian.[47] This article primarily describes phonological development after the first year of life. See the main article on phonological development for a description of first year stages. Many of the earliest stages are thought to be universal to all infants.

Provided here is a rendition of the Bible, Luke 2, 1–7, as read by a native Italian speaker from Milan. As a northerner, his pronunciation lacks syntactic doubling ([ˈfu ˈfatto] instead of [ˈfu fˈfatto]) and intervocalic [s] ([ˈkaːza] instead of [ˈkaːsa]). The speaker realises /r/ as [ʋ] in some positions.

2:1 In quei giorni, un decreto di Cesare Augusto ordinava che si facesse un censimento di tutta la terra.
2 Questo primo censimento fu fatto quando Quirino era governatore della Siria.
3 Tutti andavano a farsi registrare, ciascuno nella propria città.
4 Anche Giuseppe, che era della casa e della famiglia di Davide, dalla città di Nazaret e dalla Galilea si recò in Giudea nella città di Davide, chiamata Betlemme,
5 per farsi registrare insieme a Maria, sua sposa, che era incinta.
6 Proprio mentre si trovavano lì, venne il tempo per lei di partorire.
7 Mise al mondo il suo primogenito, lo avvolse in fasce e lo depose in una mangiatoia, poiché non c'era posto per loro nella locanda.

The differences in pronunciation are underlined in the following transcriptions; the velar [ŋ] is an allophone of /n/ and the long vowels are allophones of the short vowels, but are shown for clarity.

A rough transcription of the audio sample is:

2:1 [iŋ ˈkwɛi ˈdʒorni un deˈkreːto di ˈtʃeːzare auˈɡusto ordiˈnaːva ke si faˈtʃɛsse un tʃensiˈmento di ˈtutta la ˈtɛrra
2 ˈkwɛsto ˈpriːmo tʃensiˈmento fu ˈfatto ˈkwando kwiˈriːno ˈeːra ɡovernaˈtoːre dɛlla ˈsiːrja
3 ˈtutti anˈdaːvano a ˈfarsi redʒiˈstraːre tʃaˈskuːno nɛlla ˈprɔːprja tʃitˈta
4 ˈaŋke dʒuˈzɛppe ke ˈeːra dɛlla ˈkaːza e dɛlla faˈmiʎʎa di ˈdaːvide dalla tʃitˈta di ˈnaddzaret e dalla ɡaliˈleːa si reˈkɔ in dʒuˈdeːa nɛlla tʃitˈta di ˈdaːvide kjaˈmaːta beˈtlɛmme
5 per ˈfarsi redʒiˈstraːre inˈsjeːme a maˈriːa swa ˈspoːza ke ˈeːra inˈtʃinta
6 ˈprɔːprjo ˈmentre si troˈvaːvano ˈli ˈvɛnne il ˈtempo per ˈlɛi di partoˈriːre
7 ˈmiːze al ˈmondo il swo primoˈdʒeːnito, lo avˈvɔlse iɱ ˈfaːʃe e lo deˈpoːze in ˈuːna mandʒaˈtɔːja poiˈke non ˈtʃeːra ˈpɔsto per ˈloːro nɛlla loˈkanda]

The Standard Italian pronunciation of the text is:

2:1 [iŋ ˈkwei ˈdʒorni un deˈkreːto di ˈtʃeːzare auˈɡusto ordiˈnaːva ke ssi faˈtʃesse un tʃensiˈmento di ˈtutta la ˈtɛrra
2 ˈkwesto ˈpriːmo tʃensiˈmento fu fˈfatto ˈkwando kwiˈriːno ˈɛːra ɡovernaˈtoːre della ˈsiːrja
3 ˈtutti anˈdaːvano a fˈfarsi redʒiˈstraːre tʃaˈskuːno nella ˈprɔːprja tʃitˈta
4 ˈaŋke dʒuˈzɛppe ke ˈɛːra della ˈkaːsa e ddella faˈmiʎʎa di ˈdaːvide dalla tʃitˈta ddi ˈnaddzaret e ddalla ɡaliˈlɛːa si reˈkɔ in dʒuˈdɛːa nella tʃitˈta ddi ˈdaːvide kjaˈmaːta beˈtlɛmme
5 per ˈfarsi redʒiˈstraːre inˈsjɛːme a mmaˈriːa ˈsuːa ˈspɔːza ke ˈɛːra inˈtʃinta
6 ˈprɔːprjo ˈmentre si troˈvaːvano ˈli ˈvenne il ˈtɛmpo per ˈlɛi di partoˈriːre
7 ˈmiːse al ˈmondo il ˈsuːo primoˈdʒɛːnito, lo avˈvɔlse iɱ ˈfaʃʃe e llo deˈpoːse in ˈuːna mandʒaˈtoːja poiˈke nnon ˈtʃɛːra ˈposto per ˈloːro nella loˈkanda]

• Italian orthography
• Italian language
• Italian grammar
• Syntactic gemination
• Wikipedia help page for IPA for Italian – includes English approximations
• Italian pronunciation guide at Wiktionary

• The online edition (2007) of the Dizionario d'ortografia e di pronunzia (DOP), a pronouncing dictionary of standard Italian