F, S, and H in Classical Pronunciation

Classical Latin f, s, and h do not form one natural phonological class, but they share a practical problem: familiar European spellings invite later or English values. Reconstructed Classical f is a voiceless fricative broadly like English f; s remains voiceless /s/ even between vowels; and h represents /h/ in the educated Classical norm, although it was socially and historically unstable.

These broad conclusions are stronger than every fine phonetic detail. A labiodental [f], made with lower lip and upper teeth, is the responsible classroom value for f, but no recording establishes identical force in every Roman community. Evidence for h proves both its prestige pronunciation and its weakening. A careful reconstruction states both facts instead of declaring either “always fully pronounced” or “always silent.”

The practical map

Both f*s are voiceless. Hold the lower lip lightly against the upper teeth and let air pass continuously; unlike *p, f has no complete oral closure. The long ī in fīlius belongs to the vowel. It neither voices the following consonant nor marks the stressed syllable directly.

F: a voiceless fricative

Latin f is conventionally transcribed /f/ and taught with the English sound in fine. Ancient descriptions identify it as labiodental. Fine differences of force are less important for a beginner than preserving the lower-lip contact, voicelessness, and continuous airflow.

The f in fāma and forum never takes the /v/ heard in English of. Consonantal v in volat is /w/ in this guide’s reconstructed Classical convention, so f and v are not the English pair /f v/. Later developments toward fricative v belong to other periods.

In officium, ff is a long fricative divided of-fi-ci-um. The first o is short, while of- is heavy because the first half of /fː/ closes it. By contrast, faciō begins with a single /f/. English spelling often doubles a consonant merely to signal a preceding vowel; Latin requires the consonantal duration itself.

In educated Greek loans, ph represents an aspirated stop /pʰ/ in a careful Classical reconstruction, not native /f/ and not two syllabified consonants /p.h/. Many speakers and later periods merged or adapted this foreign sound, and ecclesiastical pronunciation normally reads ph as /f/. That later rule should not be projected back without a label; see Pronouncing Greek Loanwords for the larger system.

S: voiceless in every position

Classical s is /s/ at the beginning, middle, and end of a word. It does not automatically become /z/ between vowels, as English s often does in rose, music, or easy. Thus rosa contains /s/, not /z/.

This is a useful English–Latin minimal comparison: the English translation rose ends in /z/, but Latin rosam has voiceless /s/. Translation gives the meaning, not the sound. The Latin o is short as well; familiar English pronunciation predicts neither feature.

The intervocalic s of miser and the s of servus and causam remain /s/. Historically, inherited intervocalic s often changed to r in earlier Latin—a process visible in alternations and called rhotacism—but the s that survives in Classical words is not thereby licensed to voice. A historical change explains part of the lexicon; it is not a synchronic pronunciation rule.

Final s is also pronounced. It can close a syllable and contribute to weight, and it participates in the sound texture of prose and verse. Do not erase it merely because some modern languages weaken final consonants.

The final -s of servus and bonus remains consonantal in careful Classical speech. Exact assimilation across word boundaries may occur in fluent language, but the lexical form does not end in an open vowel.

Double ss is longer than single s. In possum, divide pos-sum and sustain the frication. This makes the first syllable heavy without making its o long.

H: pronounced, weak, and socially marked

In the educated late Republican and early imperial norm, written h represented /h/, a voiceless glottal fricative like the beginning of English house. Roman comments on omitted and wrongly added h show that the sound was already unstable. Some speakers weakened or lost it; others inserted it hypercorrectly because correct aspiration carried social prestige.

Those “errors” are evidence that h was not simply absent. One cannot be mocked for putting a sound in the wrong place unless the community recognizes a norm governing that sound.

In a careful reconstructed Classical reading, pronounce /h/ in hodiē and Horātius. Do not make it the harsh velar or pharyngeal fricative found in some modern languages. A light breath through an open glottis is enough.

The h of habēsne and hunc is audible in the educated target. The construction is ordinary (informal) direct questioning; careful aspiration need not make it formal or affected. Overemphasizing every h can be as misleading as deleting it.

Between vowels, h begins the following syllable and does not close the preceding one: ve-hō, ni-hil. It is prosodically weak. In poetic elision, initial h normally does not prevent a final vowel from eliding before the following word, a fact that supports its limited consonantal weight.

Read nihil as ni-hil and habeō as ha-be-ō. The h is present, but it does not turn ni- into a closed syllable. Some later pronunciations and lexical traditions treat nihil differently; the current analysis is the Classical target.

In standard ecclesiastical pronunciation, written h is normally silent, with traditional exceptions in mihi and nihil where h may be pronounced /k/. This coherent later convention must not be mixed silently with Classical hard c, Classical diphthongs, and Classical vowel quantity.

What ancient spelling does not settle

Ordinary Roman spelling marks f, s, and h, but it does not transcribe every allophone. The presence of h in a normalized literary text can reflect educated spelling even when a particular speaker weakened it. Conversely, an inscriptional omission may reveal real weakening but cannot prove that every contemporary speaker had no /h/.

Terence gives an authentic (literary, dramatic) line, here reproduced without modern macrons:

The repeated h is pronounced in the careful Classical teaching norm. A macronized aid may separately show homō sum: hūmānī nihil ā mē aliēnum putō, but an attributed quotation should follow its edition rather than silently acquiring pedagogical marks.

The exact phonetic force of f and h, or small contextual changes in s, is less certain than the contrastive facts. A broad /f s h/ transcription is more responsible than an over-detailed performance claimed as the only possible voice of Rome.

Common Mistakes

1. Voicing f because an English cognate or function word does.

2. Voicing intervocalic s.

3. Silencing every h in a Classical reading.

4. Pronouncing ph as a sequence p + h.

5. Treating a macron as stress.

Key Takeaways

• Use a voiceless fricative broadly like English f for native Latin f; hold ff longer.
• Pronounce Classical s as voiceless /s/ in every position, including between vowels and word-finally.
• Pronounce a light /h/ in careful educated Classical Latin while recognizing ancient weakening and hypercorrection.
• Do not let h close the preceding syllable, and do not split Greek ph into separate p and h.
• Label ecclesiastical silence or later loanword values instead of mixing them into a Classical reading.