In Latin, it seems some sounds that are pronounced like an "F" in Greek, are pronounced like a "P", why is this?

For example, we have the Greek word Phoenicians, and this word always seems to be pronounced like an "F" in all languages, except Latin, where we have statements such as in Virgil:

Daphni, tuum Poenos etiam ingemuisse leones interitum...

[Daphnis, even the Punic lions lamented your passing...]

So, the Phoenician city, Carthage is always pronounced apparently with a "P" by the Latins. Does this suggest that that the early Romans actually pronounced the letter "P" as an "F"? Or maybe there is confusion with the digamma? It's hard to believe this because Latin has the regular "F" sound in its alphabet, for example filius (son). So, why would it not be Foeni, instead of Poeni?

To make it more confusing, the Latins apparently transliterated phi as "PH", not "F". So, for example, Cicero calls his speeches against Antony the Philippicae. So, by that logic, Poeni should be Phoeni. However, one would wonder why not just "Foeni". Did Cicero pronounce "PH" differently than "F"?

Did any of the Latin grammarians, like Priscian or Quintillian, ever discuss this issue?

  • 5
    Consider visiting Latin.SE (which also covers Ancient Greek). There have already been a few similar questions there.
    – tum_
    Commented Feb 3, 2021 at 14:38
  • 5
    @user6726 whilst there is evidence for fairly free variation between p & f in Punic, the fact that at this stage Greek phi was not in fact a fricative seems a more likely culprit
    – Tristan
    Commented Feb 3, 2021 at 16:22
  • 2
    While this is certainly a good question for Latin Language, this does not make it off-topic here. Commented Feb 3, 2021 at 17:35
  • 1
    As the answer from Tristam below demonstrates, the technical issues are complex and are really only answerable by an expert in historical linguistics, so the Linguistics forum is the right place for this question. Commented Feb 3, 2021 at 18:10
  • 4
    @TylerDurden Latin.SE also answers questions about linguistics that relate to Latin, Greek, Punic, Etruscan, etc; I would say this is on-topic on both sites.
    – Draconis
    Commented Feb 3, 2021 at 19:27

2 Answers 2


I'm going to take a slightly different approach than Jk's answer, which does a good job coming at this from a Greco-Roman perspective. Instead, I'm going to focus on the Punic situation because it's a bit of an interest of mine

In this early stage of Greek (Classical Attic), we had three alveolar stops (I will come back to the bilabials in a bit, don't worry): Theta, Tau, & Delta, which represented /th/, /t/, & /d/ respectively

Meanwhile, Latin has two alveolar stops, /t/ & /d/

And Punic has three: Taw, Ṭet, & Dalt which represented /th/, /t/, & /d/ respectively. The exact nature of Punic Ṭet is uncertain, it may have simply been /t/ (this was likely the case in later Punic), an ejective /tʼ/ (this was likely the case in earlier Phoenician), or (probably the least likely, but depending on how you reconstruct early Amazigh emphatics might have existed in Amazigh-influenced Punic) a pharyngealised or velarised /tˤ ~ tˠ/

The three series of aspirated, tenuis, & voiced stops in Greek line up very well with the three series of aspirated (albeit usually just called plain or voiceless), emphatic (voiceless, but phonetically unaspirated), and voiced stops in Punic, and so Greek pretty consistently transcribes Punic Taw as Theta, Ṭet as Tau, and Dalt as Delta

Note that Tau & Theta appear to be the wrong way round (Theta gets its name and shape from Phoenician Ṭet, but seems to get its sound from Taw, whilst Tau gets its name from Taw but it's sound from Ṭet). This is somewhat of a mystery. Pretty much all other Semitic borrowings in Classical Attic seem to reflect the mapping given above with Greek aspirates being mapped to Semitic voiceless stops, and Greek tenuis stops being mapped to Semitic emphatics (likewise borrowings in the other direction from Greek into Semitic languages)

Meanwhile, most Latin speakers cannot reliably hear a difference between aspirated and tenuis stops, unless they are also Greek speakers, and so they tend to write Phoenician Ṭet & Taw both as ⟨t⟩ in borrowings. Educated speakers (and later Punic inscriptions in Latin script) do represent the distinction, writing Taw as ⟨th⟩ (just as they transcribe Greek Theta), but this still represents a stop at this stage

Ok, that's all well and good, but what about labials?

Well, Greek and Latin are essentially unchanged here having /ph/, /p/, & /b/, and /p/ & /b/ respectively

Punic however has no emphatic labial, only Pe & Bet. In the earliest stages of Punic, and in Phoenician, these were likely /ph/ & /b/ respectively

As such, the Greeks generally transcribed them using Phi & Beta, and the Romans as ⟨p⟩ & ⟨b⟩ in early borrowings (although some early transcription by Romans who likely also spoke Greek do prefer ⟨ph⟩ for Pe). So Greek Φοῖνῐξ (Phoînix) & Latin Punicus both represent the same Phoenician-Punic endonym written Pe-Waw-Nun-Yod i.e. pwny likely pronounced something like /pho:ni:/ (although vowels are generally difficult to reconstruct with much accuracy)

Later on (likely before the Second Punic War), Punic Pe started shifting toward /f/, likely being in free variation for a time, and we see no change in Greek borrowings and transcriptions at this time (Greek having no /f/ at this stage, and so using their nearest sound, the aspirate Phi), but Roman transcriptions & borrowings start using ⟨f⟩ to write Punic Pe (hence Sufes for the chief magistrate of Carthage, spelt špṭ in Punic), as well as some instances of Bet (likely indicating that it had shifted to /v/ before another consonant)

Later still, around the turn from BCE to CE, Greek goes through a big shift of its consonants. /ph/, /p/, & /b/ have now shifted to /f/, /p/, /v/ in most positions. The shift of Phi even catches on in Latin speakers' reading of Greek borrowings and so they start reading Latin ⟨ph⟩ as /f/, even though up until this point that was generally pronounced identically to p as /p/

So now the Greek Φοῖνῐξ is pronounced with an /f/ whilst the Latin Punicus retained its original /p/, and both fail to represent the original early Punic /ph/

With Cicero saying Philippicae, as an educated Roman, it's extremely likely he was fluent in Greek and pronounced it with a /ph/ as in the Greek contemporary to him

  • 1
    thanks for the edit jk, I still haven't quite got the hang of all the formatting here
    – Tristan
    Commented Feb 4, 2021 at 9:45

At some time in the history of the Greek languages, the letters Phi, Theta, and Chi represented aspirated consonants /ph/, /th/ and /kh/. The Romans felt that they were different enough from their native sounds /p/, /t/, /k/ (spelled ⟨c⟩) to deserve a special spelling ⟨ph⟩, ⟨th⟩, and ⟨ch⟩. This was the state of sounds at Cicero's time.

Later, the Greek consonants shifted and the aspirated stops became fricatives /f/, /θ/, and /ç/ or /x/. Modern Romance languages reflect this later shift for the letter Phi, but not for the other two that just lost their aspiration.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.