Voiced and Voiceless S

Catalan contrasts voiceless /s/ with voiced /z/. The difference alone can change meaning: Central Catalan caça “hunting” is /ˈkasə/, while casa “house” is /ˈkazə/. English speakers already use both consonants, but Catalan assigns them to spellings differently; many Spanish speakers must build /z/ as a new meaning-bearing category rather than treating it as an automatic variant.

This page teaches the general standard contrast with Central Catalan vowel quality as the pronunciation baseline. The /s/–/z/ contrast is widespread, though (regional: Central Valencian and other apitxat areas) some varieties merge voiced /z/ into voiceless [s].

Feel the voicing contrast

/s/ and /z/ have essentially the same narrow channel near the alveolar ridge. With /s/, the vocal folds do not vibrate; with /z/, they do. Put two fingers lightly on your throat and alternate ssss–zzzz. The second should produce a buzz without turning into an extra vowel.

Those are real lexical contrasts, not merely careful spelling distinctions. If casa is pronounced with /s/, it can be heard as caça. Lengthening the consonant is not the solution: voicing is what separates the pair.

The first two rows are minimal pairs in the general pronunciation: the surrounding sounds match and /s/ versus /z/ carries the meaning difference. Pressa–presa is a useful near-minimal pair, but many Central speakers also distinguish the quality of the stressed e. It should therefore train voicing without being presented as proof that only one sound differs.

The main spellings inside a word

The spellings become predictable once position is considered. Between vowels, a single s normally represents /z/, while ss represents /s/. The letter z commonly represents /z/, and c before e, i or ç before a, o, u represents /s/.

The spelling rule is positional, not “one letter, one sound.” Initial s in sala is /s/, but intervocalic s in casa is /z/. Conversely, doubling ss between vowels prevents voicing: passa has /s/. See S, SS, C, Ç, and Z for the full spelling system.

Final s changes according to the next sound

At the end of a word, s is [s] before a pause or a voiceless consonant. Before a voiced consonant, it normally becomes [z] by regressive assimilation: the vocal folds start vibrating early for the next sound. Compare dos pans with [s] before voiceless p and dos gats with [z] before voiced g.

In els veïns, final s stands before voiced v, so [z] is the natural connected realization. This is not careless speech and is not restricted to an informal register; it is a regular phonological process in fluent standard Catalan.

Before a following vowel, final s is likewise voiced and links into the next syllable: dos amics is naturally [do.zəˈmiks] in Central Catalan, and les ales has a linking [z]. A real pause blocks the process, so an isolated dos ends in [s]. (regional: apitxat Valencian) Prevocalic final s may remain voiceless, consistently with the regional lack of lexical /z/.

Lexical z and contextual voicing are not the same thing

The /z/ in casa belongs to the word itself: it remains voiced whether the word is spoken alone or inside a sentence. The [z] in dos amics is different. The word dos ends in underlying /s/, as its pronunciation before a pause shows, but that /s/ becomes [z] because the following vowel is voiced. Linguists call the first contrast phonemic and the second change assimilation.

This distinction explains why spelling does not change at a word boundary. Writing dos amics with s preserves the identity of dos; pronunciation adjusts automatically to the next sound. It also explains why a pause matters: En vull dos. Amics meus també en volen has [s] before the full stop, while the uninterrupted phrase dos amics has [z].

English also voices some plural endings, but Catalan applies its own phrase-level rule regardless of whether final s is a plural marker. Spanish does not provide a safe model here: standard Spanish lacks the same lexical /s/–/z/ opposition, and many speakers carry [s] across boundaries where general Catalan requires [z].

Stress and vowel reduction still matter

Minimal pairs are only useful if the rest of each word is accurate. In Central Catalan, both casa and caça are stressed on the first syllable, and their final written a reduces to versus /ˈkasə/. A Spanish-like final [a], or an English-style diphthong on the stressed vowel, obscures the Catalan rhythm even if /s/ and /z/ are correct.

Keep the stressed [a] steady—do not let it drift toward an English “ah-uh”—then make the consonant contrast, then reduce the final vowel. Catalan spelling becomes much more predictive when consonant context, stress, and vowel reduction are read together.

Regional systems and careful pronunciation

(regional: apitxat Valencian) In Central Valencian around València and in some other western localities, historical voiced fricatives are devoiced, so /z/ may merge with [s]. In that regional system, casa and caça can sound alike. This is a coherent dialect feature, not random “bad pronunciation.”

Outside those areas, replacing /z/ with /s/ is commonly associated with Spanish interference and can neutralize contrasts. (formal) Careful speech does not require exaggerating /z/, but it does preserve it where the speaker's standard variety has the opposition. A light, continuous buzz is enough.

Slow and connected models

In a slow but natural reading, maintain intervocalic /z/ in casa and rosa. Do not insert pauses between dos and amics: even careful connected speech voices the final consonant. In a deliberately word-by-word citation, a pause could restore [s], but that is not the target for a sentence.

Common Mistakes

1. Pronouncing every written s as /s/.

2. Using one s for intervocalic /s/ in spelling.

3. Keeping final s voiceless before a vowel.

4. Copying Spanish vowel quality into the minimal pair.

Key Takeaways

Catalan /s/ and /z/ contrast in words such as caça and casa. Between vowels, ss normally spells /s/ and single s spells /z/; c/ç also spell /s/, while z spells /z/. Final s assimilates to the next sound, becoming [z] before vowels and voiced consonants. Preserve Central stress and vowel reduction, and recognize apitxat devoicing as a labeled regional system rather than applying it to the general standard.