Dtjytyk

I was talking today with an English co-worker about whether he says "an H-1B visa" or "a H-1B visa", which hinges on whether one says "aitch" or "haitch" for the letter H.

And I noticed that unlike other languages such as German or French or Romanian that change the indefinite article based on things like gender or case, in English we change the indefinite article based on the sound of the following word.

Are there other languages that change the indefinite or definite article based on what the following word sounds like?

In Italian, both the indefinite and the definite article change in spelling and pronunciation depending on the following sound, in the masculine gender.

Before vowels, the masculine indefinite article is un and the definite article is l' (elision of lo):

• 
  un albero (a tree), l'albero (the tree)

Before single consonants or consonants followed by semivowels or liquids, the masculine indefinite article is un but the definite article is il, and this includes the affricate consonants /t͡ʃ/ (spelled 'c' before 'i' or 'e') and /d͡ʒ/ (spelled 'g' before 'i' or 'e'):

• 
  un ramo (a branch), il ramo (the branch)


• 
  un fiore (a flower), il fiore (the flower)


• 
  un treno (a train), il treno (the train)


• 
  un cielo (a sky), il cielo (the sky)


• 
  un gioco (a game), il gioco (the game)

Before double consonants (generally 's' followed by a consonant, but other combinations arise in loanwords, especially of Greek origin, including with the double consonant 'x'), and before 'z' which is pronounced as the affricate /t͡s/ or /d͡z/, the masculine indefinite article is uno and the definite article is lo:

• 
  uno spazio (a space), lo spazio (the space)


• 
  uno psicologo (a psychologist), lo psicologo (the psychologist)


• 
  uno xilofono (a xylophone), lo xilofono (the xylophone)


• 
  uno zaino (a backpack), lo zaino (the backpack)

Feminine articles are una (indefinite) and la (definite), but they get elided into un' and l' before vowels:

• 
  una foglia (a leaf), la foglia (the leaf)


• 
  un'ombra (a shadow), l'ombra (the shadow)

These changes are not dependent on the noun the article connects to, but the sound immediately following it, so for example we have:

• 
  l'albero (the tree), but il grande albero (the big tree)


• 
  un gioco (a game), but uno speciale gioco (a special game)


• 
  uno zaino (a backpack), but un piccolo zaino (a small backpack)


• 
  una foglia (a leaf), but un'altra foglia (another leaf)


• 
  un'ombra (a shadow), but una cupa ombra (a dark shadow)

These behaviors can be at least partly explained by the phonotactics of Italian: whenever uno is used, the phonotactic constraints wouldn't allow for un unless the 'n' were dropped entirely; the same issue gives rise to lo instead of il, which look very different on the surface, but both come from different parts of the Latin word illum or illud (a demonstrative).

The elided forms with an apostrophe can be explained by a tendency to eschew hiatus.

A famous example is the Arabic language where the the definite article al assimilates to one half of the potential following consonants called Sun letters in Arabic grammar. So it is an-Nil "the Nile" or ash-shams "the sun", but al-qamar "the moon".

Quite similarly to Italian (see @LjL very complete answer) and a few other Romance languages, French does this for indefinite and definite articles, but not really the same way English does, in the sense that what we call "Liaison" is very common, though it rarely changes the writing of the articles.

Indefinite articles:

• 
  Un verre /œ̃ vɛʁ/ (a glass)


• 
  Un arbre /œ̃.n‿aʁbʁ/ (a tree)

Both mean "a" but the second one will have its N pronounced as a "transition" letter between the /œ̃/ and the /a/ to avoid hiatus. The writing is unchanged, though.

A similar process happens with the definite article:

• 
  Le verre /lə vɛʁ/ (the glass)


• 
  L'arbre /l‿aʁbʁ/ (the tree)

Here, "le" becomes "l'", again to avoid hiatus, but this time the writing changes to reflect the absence of the E.

I've only written a few examples, but this process also happens for a plethora of other articles, in a similar fashion.

Basically, it happens, but the writing only gets modified if a sound disappears completely because of the change.

Catalan masculine singular definite articles.

• /l/ before a vowel sound.


• /el/ before anything else.

What triggers this allomorphy is clearly the sound, not the letter, as we see l'interval /linteɾval/ but el iode /eljod/, because of the glide /j/. Also l'hivern /liveɾn/ but el hiat /eljat/.

I think it is pretty common to "change the indefinite or definite article (that's called allomorphy) based on what the following word sounds like" in many languages. Italian has three masculine singular definite articles, /el/, /lo/, and /l/, depending on the following sound. In English, the definite article the is pronounced /ðə/ or /ðiː/, sometimes even /ðɪ/.

Maybe the point of interest for you comes from the arbitrary insertion/dropping of /h/, not from the allomorphy.

Most of the famous examples in Europe and the Mediterranean have been mentioned, but we should add the languages where the definite article is simply a suffix, for example the core languages of the Balkan Sprachbund, like Albanian and Romanian, and Armenian, which works very much like they do in this regard.

Turkish does not really have an article but objects are marked for definiteness with a suffix. In that case the connector varies but also the vowel, because of vowel harmony. So -u, -yu, -i... are all realisations of the same morpheme, and it varies according to a vowel that need not even be directly adjacent.

In Spanish, before a feminine noun with stressed initial /a/, the usual feminine singular definite article la has an alomorph el (which is identical to the masculine singular definite article).

• la vaca (f)


• el agua (f)


• el libro (m)

Why is “agua” masculine in singular form and feminine in plural? “El agua” / “Las aguas” (Spanish SE)