As Julius Caesar said, “Wehnee, weedee, weekee!”

Or: Why do I pronounce Latin words like that?

---

A question came up recently in the middle of an enjoyably linguistic conversation, concerned with the way I personally pronounce Latin words. The question was essentially ‘why?’

My interlocutor, a friend with only a little Latin learning, was curious about how I pronounce certain letters. What caught his attention was that I pronounce Latin vīnum ‘wine’ and videō ‘I see’ with an initial /w/ sound (as in English weed) and cēdō ‘I go’ with an initial hard /k/. Many others would say ‘veenum‘, ‘videhoh‘ and ‘chehdoh‘ (or ‘sehdoh‘) instead.

You may recognise this as what’s widely known as reconstructed Latin pronunciation. This system is the result of scholarly efforts to reconstruct how the Romans themselves spoke Latin by using the Latin they have left us in writing. It stands in contrast to ecclesiastical pronunciation, which is characterised by various late-Roman and post-Roman changes. These make for a very crude dichotomy though; Latin in the time of the Roman state had as much variation (both across time, geography and society) as any living language, while ecclesiastical Latin has never by any means been uniform, nor its features limited to churches only.

My preference for the reconstructed pronunciation is also not thorough; it would be hard to stick to everything we have learned about the Latin of Ancient Rome. It’s more of a preference for a handful of features – and even these I don’t use in all circumstances! While I might say “vēnī, vīdī, vīcī” as ‘wehnee, weedee, weekee‘, I still pronounce its author’s name like the majority of Anglophones do when talking about him in English – i.e., Julius Caesar with a /d͡ʒ/ (as in juice) and a /s/ (as in the salad). Et cetera is still ‘et setera‘ for me too, don’t worry.

Despite all this, I reckon it could still prove interesting to write down in this article why my Latin pronunciation has picked up these reconstructed features. For the sake of space, it concentrates on two particularly salient features: the aforementioned sounds /w/ (written V) and /k/ (written C). The piece is apologetic, but certainly not polemic against people who say things differently. It’s my way of explaining myself and, in doing so, sharing some things that could be beneficial for your own Latin.

---

How do we know how the Romans spoke their Latin?

The Romans, despite the sanitation, the medicine, the education and the aqueducts, never got around to inventing tape recorders. This means that we don’t have the same type of evidence for spoken Latin that phoneticians have for modern languages – but we can still use what we do have, which is text. Lots and lots of text.

It comes from a range of times, locations and authors. It includes not only standardized Latin spelling, but also plenty of lovely non-standard variants; each one is a potential glimpse into how its author thought it best to denote the sounds of their speech. We can be confident that written Latin (standardized or not) mirrors the spoken language quite closely; unlike English or French, Latin was not so hung up on reflecting words’ origins and history in their spelling, but rather aimed to match one sound to one symbol, and was perfectly happy to adapt the letters of existing Italian alphabets to that purpose.

In Vox Latina, his magnum opus on the sounds of Latin, W. S. Allen sets out six sources from which the pronunciation of Latin in the classical era can be reconstructed:

• (1) specific statements of Latin grammarians and other authors regarding the pronunciation of the language
• (2) puns, plays on words, ancient etymologies, and imitations of natural sounds
• (3) the representation of Latin words in other languages
• (4) developments in the Romance languages
• (5) the spelling conventions of Latin, and particularly scribal or epigraphic variations
• (6) the internal structure of the Latin language itself, including its metrical patterns

Allen 1978: viii

These give us a wealth of useful information!

Although Allen admits for (1) that many were writing towards the end of the Roman era (in the west), it’s certainly helpful to have the words of actual Roman grammarians writing about their own language, often in precise terms. For instance, the fourth-century scholar Gaius Marius Victorinus writes that “we pronounce the letter F by pressing the bottom lip with the upper teeth...”, which really isn’t far from the modern phonetic term of ‘labiodental’ used for the letter’s pronunciation in English today.

Borrowed Latin words, as per Allen’s third point, are equally interesting and crop up in all sorts of places. For example, how do we know that the letter C in the words Caesar and centum (‘hundred’) used to be pronounced /k/? Why not with a /s/, as in English Caesarean and century? Look no further than the gospels, which were first (probably) written in the other big language of the Roman empire, Ancient Greek. In Jesus’ famous saying about giving to Caesar what is his, the Greek word is Kaîsar, while the officer at the foot of the cross, the commander of a hundred Roman soldiers, is a kenturíōn. Both are spelled with the Greek letter kappa, which uncontroversially stands for the consonant /k/. Consequently, this is why a German emperor was a Kaiser.

Now, this is all good stuff – but it’s not without its limitations. The relevant sources are mostly the language of the Roman male elite, many of whom were writing according to a linguistic standard and at a late date. Such evidence does not give us the full window into Roman Latin that we would like. It requires us to speak in general terms; some people at some earlier point pronounced C and V as /k/ and /w/. These pronunciations were not universal, even at a time as early as the first century AD, when we have the first evidence for the shift of /w/ to /v/ (as in vase).

Multiple pronunciations most likely existed side by side, used according to various sociolinguistic factors, such as age, location, present company or the particular register desired. As a great example of Allen’s second point, Cicero, writing de Divinatione in the first century BC, revealed quite unintentionally that /w/ was one of the sounds of his Latin and one of the ways he might pronounce the letter V.

In a story about a fig-seller, Cicero gives evidence for the sound of the Latin letter V.

The man shouts "cauneas!" ('Caunian figs!'), which Cicero also interprets as the warning 'cave ne eas!' ('beware, don't go!')

This only works if the V in "cave" stood for [w], as in 'way'.

— Danny Bate (@DannyBate4) August 31, 2021

But does this mean that the pronunciation of V as shifted /v/ or some intermediate sound in was completely unknown to Cicero? Not necessarily – who knows how he might have spoken in each particular circumstance? We have to be cautious with the evidence.

This shift is ultimately the reason why Latin words with V are pronounced /v/ in English and the Romance languages (e.g. French vin, Italian vino, Portuguese vinho, as well as English vine). Such sound changes that Latin underwent would lead to the new sounds of both Romance and Latin itself in its later medieval and ecclesiastical lives.

---

So, why should I care?

Frankly, you don’t have to. Latinist or not, I will admit that it’s possible to go through one’s life happily paying no heed to theories of phonological reconstruction. However, if you know Latin or are currently learning it, these theories may be very beneficial for your knowledge. For me, the advantages of being familiar with reconstructed Latin pronunciation boil down to two: for things inside Latin, and for things outside it.

1: Internal Benefits

By “internal”, I mean to say that are parts of Latin grammar that make more sense when you bear in mind that that the Romans originally used the letter V to stand for either the vowel /u/ or the consonant /w/. These sounds are therefore used in Latin’s reconstructed pronunciation for the letter V – but not the consonant /v/.

This was very reasonable of the Romans; the sounds /u/ and /w/ are very similar. In fact, /w/ is technically known as a semivowel and is really the same sound as the vowel /u/. The difference is just where in a syllable it’s used. When in the central position of the syllable (its nucleus), the sound is a vowel. When placed either side of the nucleus, that same sound is a semivowel. For a good introduction to syllable structure and what I mean by all of this, I recommend this video:

You can demonstrate the similarity of /w/ and /u/ for yourself: take any English word that begins with a /w/ (like wasp or water) and swap the consonant for the vowel /u/ (as in food). Strung together, the resulting ‘ooater‘ and ‘ooasp‘ still sound almost identical to the originals.

All of this is why the Romans grouped the vowel and the semivowel under a single letter: V. Since then though, V and U have emerged as separate letters to distinguish consonants from vowels. This is now common practice in writing Latin, as well as modern languages.

Now to apply this to Latin grammar. Let’s take the very important perfect tense. It’s the one with the tense/person endings -ī, -istī, -it, -imus, -istis, -ērunt, and it’s a keystone of Latin. If you want to express past-tense actions like ‘I loved‘ or ‘I taught‘, you need the perfect tense:

But how do you add the abstract endings above to specific verbs? You need to add them to a specific form of the verb, known as its perfect stem. Stems are basically a mid-point between the root of the word (the meaningful bit) and its endings, and the perfect stem is confusingly formed in many ways, which you need to be familiar with to get the hang of the perfect tense. For love the perfect stem is amāv-, but for teach the stem is docu-, giving us amāvī ‘I loved’ and docuī ‘I taught’.

Different verbs, different stems? Multiple things to learn? Not really – they’re formed with the same affix! It’s only current practice to write amāv– and docu– with -v- and -u- to distinguish between vowel and semivowel; the Romans would have written them simply as AMAV– and DOCV–, both formed with the affix -V-.

There are further examples. This is why English has the verb solve but also the noun solution. To the Romans, the root of this verb was simply SOLV-. The V in SOLV- could be pronounced as either /w/ or /u/, depending on what followed: either a vowel (as in solvō ‘I loosen’) or a consonant (as in solūtiō ‘loosening’).

This is because the following T consonant in solūtiō requires the V sound to be a syllable nucleus and therefore a vowel; meanwhile, the Ō vowel in solvō puts V to the edge of the second syllable, making it a semivowel. It was later developments in speech and spelling that have led to the bigger difference between English solve and solution.

Likewise, the present-tense verbs sequor ‘I follow’ and loquor ‘I speak’ have the past participles secūtus ‘followed’ and locūtus ‘spoken’, with the same alternation between semivowel and true vowel. This is still reflected in English sequence/persecute and eloquence/elocution.

2: External Benefits

In a word, these other advantages are to do with etymology. Time and time, I find that remembering that /w/ was once a sound of Latin reveals interesting and helpful links to languages beyond Latin.

It’s no great surprise that Latin vīnum is related to English wine. Specifically though, the relationship of one of ancestry; vīnum is the origin of wine, which still preserves its initial /w/ consonant.

This is by no means the only example though! By undoing Latin’s shift from /w/ to /v/, we can appreciate better how the five Latin words vidēre (the origin of English vision), vādere (the origin of invade, evade and Spanish vamos!), vermis (the origin of vermin), ventus (the origin of ventilation) and verbum (the origin of verbose) are distant relatives of the five inherited English words wise, wade, worm, wind and word.

It’s not only at the beginning of words that Latin /w/ can be connected to other sounds in other languages. Until I stopped to remember that English nerve comes from Latin nervus, and that nervus would have been pronounced ‘nerwus‘, their connection to Greek neûron (the origin of neuron and neurology) was not at all obvious. Here Latin /w/ corresponds comfortably to the Greek /u/ vowel – they are effectively the same sound after all, though their positions within the words have turned them into a semivowel and a vowel, respectively.

I don’t think I’m alone in finding etymological connections useful for more than just wowing one’s friends; they’re how I best remember new vocabulary. The more links I identify, the more words I can remember.

---

I hope that this has demonstrated why a little knowledge of Latin’s older sounds can go a long way, as well as where that knowledge comes from. I firmly believe that any language at any point in time is learnable – but also that complementing your learning with some of its history can make your job much easier. Certain features of a language make most sense when one considers the linguistic context in which they first emerged, such as how some perfect stems are formed, or the exact relationship between sequor and secūtus.

All of this is therefore why I like to use reconstructed Latin pronunciation; it’s helped me before, and no doubt it will help me again. I will admit, I think it can also sound quite cool.

END.