Rhotacism, and how it can help your Latin

If you have studied a little Latin, you may have come across an important, yet rather annoying group of nouns. They belong to the third declension, are neuter in gender and end in -us in the nominative singular. They include words like tempus ‘time’, corpus ‘body’ and pectus ‘chest’.

They look like nice second-declension nouns, like masculine dominus, but the trouble begins when you look at their various declined forms. Let’s take corpus as the group’s representative.

The twelve forms of corpus ‘body’

With the exception of the nominative, vocative and accusative singular, the -us of corpus ending disappears and -or- appears, seemingly out of nowhere, bringing with it an awful headache for Latin students. I tell my students to remember whether or not a noun belongs to this group by thinking of an English word that descends from it. For example, tempus, corpus and pectus gave English the adjectives temporal, corporal and pectoral, which reveals the -us/-or- change.

However, we want to do better than tricks like this; rather, we want to understand this phenomenon. This article aims to show you how one linguistic theory, rhotacism, can explain these words.


The first step is to try to identify what the actual case-number endings of these twelve forms could be – that is, to separate the constant from the inconstant parts of the word, which have nothing to do with its ‘body’ meaning, but rather its function in the sentence. By comparing corpus with other nouns of the same gender and declension, these are the case-number endings that we can identify:

Nominativecorpuscorpor – a
Genitivecorpor – iscorpor – um
Dativecorpor – īcorpor – ibus
Accusativecorpuscorpor – a
Vocativecorpuscorpor – a
Ablativecorpor – ecorpor – ibus
The deconstructed forms of corpus ‘body’

The suffixes we have separated from the rest of the word are exactly the same case-number endings used by other third-declension neuter nouns, such as caput ‘head’ or nōmen ‘name’. Let’s follow that similarity and take a closer look at nōmen.

Nominativenōmen nōmina (nōmin – a)
Genitivenōminis (nōmin – is)nōminum (nōmin – um)
Dativenōminī (nōmin – ī)nōminibus (nōmin – ibus)
Accusativenōmennōmina (nōmin – a)
Vocativenōmennōmina (nōmin – a)
Ablativenōmine (nōmin – e)nōminibus (nōmin – ibus)
The twelve forms of nōmen ‘name’ and their division into stem and suffix.

Nōmen reveals something important. We can see that the case-number suffixes we have identified are simply added to the nominative/accusative/vocative singular form, with only a small change to its second vowel. The nom./acc./voc. singular form itself has no suffix; instead, it is both a stem and a complete, stand-alone word.

Being also a third-declension neuter noun, we should apply this analysis to the corpus group of words. When we do, we see that the -us on the end of corpus is not a case-number suffix. It is instead an inseparable part of the word’s stem, to which the real endings can be added. This brings corpus and company into line with other neuter nouns of the third and fourth declensions (such as caput or cornū), in that nothing is added to form their nominative/accusative/vocative singular forms.

This idea is corroborated by a sound change that took place in Old Latin: -us on the end of words like both corpus and dominus was originally -os. The older form of corpus therefore was *corpos, which has the same -o- vowel as the one we see in the corpor- forms. We can therefore deconstruct corpus into the following system of twelve forms:

Nominativecorposcorpos – a
Genitivecorpos – iscorpos – um
Dativecorpos – īcorpos – ibus
Accusativecorposcorpos – a
Vocativecorposcorpos – a
Ablativecorpos – ecorpos – ibus
The underlying structure of corpus

However, one crucial step remains. How do we get from words like *corposa and *corposibus, which we have reconstructed, to their attested forms, corpora and corporibus? This step is the sound change known as rhotacism.


Its name taken from the Greek letter rho, rhotacism is the change of a particular sound into [r], the sound known in English as a rolled or trilled r. Many sounds can become [r] over time, but a common candidate is [z], since both are pronounced with the tip of the tongue, and both are voiced sounds (i.e. the vocal folds are vibrating when we produce them).

When the case-number suffixes were first added to *corpos, the process of rhotacism began. This is the key to rhotacism; it was through being now sandwiched between two vowels that the *s in *corposa and *corposibus became voiced *z. This brought the consonant even closer to [r] and, in time, potentially though other intermediate stages, rhotacism occurred and produced the -or- forms of corpus, tempus and pectus as we know them. In brief, intervocalic *s became Latin [r].

Meanwhile, the -os on the end of the nom./acc./voc. singular forms, unaffected by rhotacism, would become -us in an unrelated development. To sum up, we have the following stages:

StructureStage 1Stage 2Final Stage
*corpos + a*corposa*corpozacorpora
*corpos + ibus*corposibus*corpozibuscorporibus
The development of corpus, corpora and corporibus out of their original structure.

This also explains the English adjectives already mentioned, corporal, pectoral and temporal. These were formed in Latin through the addition of the suffix -ālis. Since -ālis begins with a vowel, it triggered rhotacism on the final consonant of the original noun; *corposālis and *temposālis became corporālis and temporālis.


It was not only this group of nouns that rhotacism affected; it had a language-wide effect, and we can find traces of it everywhere.

Firstly, let’s look at the nouns ōs ‘mouth’, flōs ‘flower’ and mūs ‘mouse’, which have a very different appearance to corpus, while mūs and flōs are also masculine in gender, not neuter. Once again, when the relevant case-number endings were added to these words, in nearly all cases the final [s] found itself hemmed in between two vowels and its change into [r] began. For example, the addition of the genitive singular suffix -is resulted in the forms ōris, flōris and mūris.

Esse, the Latin verb for ‘to be’, is a tricky verb to learn. It is very irregular. However, it is also important in any language, and doubly so in Latin, because its various forms also provide the endings for several tenses; for example eram not only means ‘I was’, but is also used as a verb ending, as in vīderam ‘I had seen’. Hence, it needs learning, and rhotacism can help with this task.

We should first recognise that the present stem of the word is es-. It is from es- that the verb gets its meaning of ‘to be’, and we can see the stem in the infinitive esse and the present tense forms est ‘it is’ and es ‘you are’. Like other verbs in the present tense, es and est are formed with the addition of the simple endings -s and -t (specifically es-s and es-t). In other tenses however, the endings that are added to es- start with vowels, such as the endings for the imperfect active -am, -ās, -at and –ō, -is, -it for the future. This, having sandwiched the [s] between two vowels, causes rhotacism; thus, *esam, *esat and *esis become eram, erat, eris, etc.

Rhotacism is also responsible for the -re that defines present infinitives. The ‘original’ ending for the infinitive was -se, but it changed when attached to the present stem and its final vowel sound. The rhotacised ending can be seen in the vast majority of infinitives like amāre ‘to love’, monēre ‘to warn’ and venīre ‘to come’, while the original -se has survived in a handful of verbs whose present stems end in consonants, such as esse ‘to be’ and nōsse ‘to know’.

Furthermore, we can observe rhotacism below the level of whole words. We are taught that there are six genitive plural endings, –ārum, -ōrum, –um, –ium, –uum and -ērum, but we can do better than this. The ‘real’ genitive plural suffix is what unites them, -um, while what remains originates elsewhere. Since the vowel in -um was able to cause the rhotacism of the preceding consonant, we can deconstruct –ārum, -ōrum and -ērum as the results of –ās-um, -ōs-um and -ēs-um. The three affixes that emerge, –ās-, -ōs– and -ēs-, are identical to the accusative plural endings of their respective declensions, and fairly similar to the genitive singular endings –ās and -is. On this matter at least, I do not want to advance any theories for the reasons behind this system (I am aware it likely emerged in pronouns rather than in nouns) or for the precise nature of –ās-, –ōs– and –ēs-; I mention it only to serve as an example of the explanatory power of rhotacism.


Rhotacism is a sound change that transforms a given consonant in a language into [r], something that has been important for the development of Latin as we know it. Most significant is the change from intervocalic [s] to [r], traces of which are still visible in many words, and which is a source of apparent irregularity – yet rhotacism is a very regular change, and recognising its effects is one way to dispel a lot of confusion.

5 thoughts on “Rhotacism, and how it can help your Latin

    1. That is a superb point – yes indeed, it has been suggested that the ř sound was a short-lived intermediate stage between *z and [r] in Latin, which is a very funny thought for me, prehistoric Romans sounding so Czech! As for the sound itself in Czech, however, that ř emerged from a different origin – it emerged out of the consonant [r] when positioned before the vowels [e] and [i], which are both what we call front vowels, produced with the tongue pushed forward.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: