Caibidil a Ceathair Déag:

Initial Mutations (na hAthruithe Tosaigh) 

Causes of Initial Mutations in Irish

The main cause is the Irish way of speaking: the words of a sentence are pulled together and spoken without a pause. Pauses are almost always avoided, and some sentences seem so as if spoken as a single word. (there seems to be a quasi "horror vacui" = a fear of nothingness.)
The individual words affect each other in pronunciation (what is generally known as "Sandhi"), similar as it occurs in French (called there "liaison" , comp.: les amis [lezami]), in order to enable this unhindered flow of speech.
On the other hand, in German there is a short pause made after almost every word so that the beginnings and ends of the words remain seperate and unchanged. It's important to note that those words in German beginning in a vowel actually don't begin with a vowel sound but a glottal stop, which is not the case in Irish.
Another thing: Similar to the French, certain unspoken phonemes "reappear" under certain circumstances (e.g. the most often mute s in les). But as opposed to French, in Irish, these letters are mostly no longer included in the spelling. They make themselves known via the "t-, h-, n-prefix", "eclipsis" or similar phenomena.
Also different than French are these changes to the beginnings of words to their own right, and being used to express grammatical relationships to other words.


These mutations eased the flow of speech because through it, the obstacles presented by the stop sounds disappeared. Stops cause a stop in the flow of air, and in turn hinder the flow of speech.
The cause of lenition was generally in Early Irish the position of the consonant between two vowels, as well as within the word as over the word "limits."
If the word ended in a vowel and the next began in a consonant + vowel (which was mostly the case), this consonant was now between 2 vowels and was lenited.
Today the once final vowels have been mostly discarded, but lenition became in time a tool of grammar, being then used where once a vowel suffix once stood.
For example:  the feminine article was originally longer and ended in the vowel -a (*sinda). In time, it was shortened to an, but the following lenition was retained. This explains lenition of feminine nouns after the article.
The masculine article (*sindos) was *sindi in the genitive singular and thus incurred lenition.
The masculine possessive pronoun a is an old genitive singular of the personal pronoun. Just like the genitive singular of the masc. articles, lenition is also required afterwards.
Feminine nouns once ended in the nominative preferably in the vowel -a (as is the case in Latin), this is why the following adjectives and nouns in the genitive are still lenited today (while the lenition today is generalised and also occurs following those feminine nouns, which didn't once end in -a).

eclipsis and n-prefix

eclipsis and the n-prefix occur where today or once the preceding word ended in a nasal (m, n, ng, mostly n).
So were the endings of e.g. numbers seacht = seven, naoi = nine and deich = ten earlier surely in a nasal (comp. Latin septem, novem, decem). This -m is long since erased from Irish; but it lives on in the modern eclipsis. (e.g. seacht mbád = seven boats ). After ocht = eight is also, in analogy to 7,9,10 eclipsed, although the number 8 never ended in a nasal.
Because the preposition "i" once, as in German, was written "in", thus ending in a nasal, explains its following eclipsis. On the other hand, the eclipsis also made the -n in "in" superfluous, so that the preposition is today whittled down to "i".
It is easy to understand that the conjunction go (also today a rare preposition in the sense of "with") causes eclipsis when one views the originally related Latin counterpart cum. Also go (Old Irish co) ended once in a nasal.
Eclipsis was produced by these final nasals as a way of simplifying speech. A voiceless consonant is difficult to speak after -m or -n, and so it was simply re-voiced ( e.g. p to b, t to d), because -mb- and -nd- are easier to speak than -mp- and -nt-. In time, many of the end nasals disappeared but the initial mutation remained and was then generalised as a grammatical tool (e.g. also stretched on other words, e.g. ocht = eight).

After preposition + article, either eclipsis or lenition can occur, depending on the dialect. Why?: The original accusative singular form of the articles were *sindon (masc.) and *sindan (fem.). It endes then with an -n as a nasal, and this is why today eclipsis follows: ar an mbord = on the table.
Just as in German, there was the possibility of the accusative or the dative after certain prepositions (depending on the desired meaning: motion or position, in German e.g.: auf dem Tisch and auf den Tisch). The dative form of the article was now *sindu (masc.) or *sinda (fem.), which of course caused lenition (this is why in Ulster today still following preposition and article lenition is always used).
Today this lexical differentiation between dative and accusative following a preposition no longer is used. After most prepositions now follows always the dative or what remains of it. If lenition or eclipsis is used, ist now more a question of dialect (Connacht/Munster: ar an mbord = on(to) the table , Ulster: ar an bhord = on(to) the table )

In the genitive plural, the article once was also *sindan, from which eclipsis resulted.
The plural possessive pronouns ár, bhur, a are old genitive plural forms of the personal pronoun, probably ended in a nasal as well and therefore still cause eclipsis today.


The t-prefix is the remains of the -d- of the article, which is otherwise unseen.
As was mentioned by lenition, the article in earlier times was quite a bit longer.
The masculine article used to be *sindos. [ 1 ] This was shortened over time to the modern  an.
The suffix -os was lost before consonants, also the -d- fell out as "superfluous" (better said: -nd melted into -n) Preceding a vowel, the suffix -os was lenited, spoken as [oh] and lenthened the lifespan of the (post the omission of the -o-) leading d, in which it hardened the -d- to -t (sindos > sindoh > sindh > sint)
The -d in *sindos thus remained longer preceding vowels (*sind) and become unvoiced ("t"), so approximately *sint, it lost additionally its s (int). Today this is written a bit differently (an t-, e.g.: an t-ull = the apple ).
The feminine article ended in a (*sinda). A following word with s was lenited to sh, spoken [h] (e.g. *sinda shulis). Later on, the a was also dropped. The d of the article was devoiced through the [h] to t, the [h] itself became mute (*sint' shulis > *int' shulis) Today, one says and writes this as an tsúil = the eye.


Today one can explain the h-prefix as a "pronunciation help between two vowels " and it is also used in this way.
Actually it is quite often the lenited final -s of the preceding word, which is no longer to be seen. (lenited s becomes a spoken [h]).
As in the other Indoeuropean languages, endings in -s were once very prevalent (comp. the Latin suffix -us, Greek -os, Lithuanian -as, Old Celtic -os)
e.g. the possessive pronoun of the 3rd person sg. fem. a = her: It was once *esjas, whose s between vowels became lenited to [h]. This is why there is today an h-prefix preceding a vowel after a = her.
Similarly, the feminine article in the genitive singular ended in -s (*sindas), so that also here an h-prefix follows.
The (feminine) nominative plural article na was once also *sindas.

The masculine nominative article was then *sindos, but this suffix -os was probably weaker in its vowel and had another effect: the leniting final -s made the -d , after the devoicing of -o- , voiceless to -t, so that the t-prefix preceding a vowel remained (sindos > sindoh > sind-h > sint > an t-)

(Note: with * the old or antiquated forms also consciously denoted orthographically as incorrect forms)

navigation bar
initial mutations
Gramadach na Gaeilge

© Lars Braesicke 2000

Valid HTML 4.01!

[ 1 ] The similarity of sindos to the modern demonstrative pronoun sin is not coincidental.
Also in German, article and demonstrative pronoun stem from one source.
(comp. by: das Haus the word "das" could be the article or a demonstrative pronoun ["= that house "]. In the sentence: Das ist ein Haus , the "das" is always the demonstrative pronoun)