Casual Conversation

Shameless plug to !linguistics@mander.xyz . This sort of question is welcome there.

Latin already did a bloody mess of those suffixes:

• if you were born in Roma, you'd be romanus ("Roman")
• if you were born in Eboracum (modern York), you'd be eboracensis ("Yorkese")
• if you were born in Gallia (roughly modern Belgium and France), you'd be gallicus ("Gallic")

In turn those suffixes used to mean different things:

• that -anus was originally just -nus. Inherited from Proto-Indo-European *-nós; you'd plop it after verbs to form adjectives.
• that -icus was originally just -cus; from PIE *-kos, but you'd plop it into nouns instead.
• nobody really knows where -ensis is from but some claim that Latin borrowed it from Etruscan.

Then French and Norman inherited this mess, and... left it alone? Then English borrowed all those suffixes. But it wasn't enough of a mess, so it kept its native -ish suffix, that means the exact same thing. That -ish is from PIE *-iskos, and likely related to Latin -cus.

And someone from Afghanistan is an Afghan? How did the word get shorter not longer? 🤔

There's some awareness among English speakers that "[$adjective]istan" means roughly "country where the [$adjective] people live", so the suffix is simply removed: Afghanistan → Afghan, Tajikistan → Tajik, etc.

That -istan backtracks to Classical Persian ـستان / -istān, and it forms adjectives from placenames.

In turn it comes from Proto-Indo-European too. It's from the root *steh₂- "to stand", and also a cognate of "to stand". So etymologically "[$adjective]istan" is roughly "where the [$adjective] people stand". (inb4 I'm simplifying it.)

Also, why is a person from India called an Indian, but the language is called Hindi? This breaks my brain…

Note that India doesn't simply have different "languages"; it has a half dozen different language families. Like, some languages of India are closer to English, Russian, Italian etc. than to other Indian languages.

That said:

• "India" ultimately backtracks to Greek Ἰνδός / Indós, the river Indus; and Greek borrowed it from Old Persian 𐏃𐎡𐎯𐎢𐏁 / Hindūš. That ending changed because it's what Greek does.
• "Hindi" comes from Hindi हिंदी / hindī, that comes from Classical Persian هِنْدِی / hindī. That hind- is the same as in the above, referring to the lands around the Indus (India), and the -ī is "related to".

Now, why did Greek erase the /h/? I have no idea. Greek usually don't do this. But Latin already borrowed the word as "India", showing no aspiration.

Philippines --> Filipino? They just saw the “Ph” and decided to use an “F”? 🤔

So, the islands were named after Felipe II of Spain. And there's that convention that royalty names are translated, so "Felipe II" ended as "Philip II" in English. And so the "Islas Filipinas" ended as "Philippine Islands".

...but then the demonym was borrowed straight from Spanish, including its spelling: filipino → Filipino.

Note that this mess is not exclusive to English. As I hinted above, Latin already had something similar; and in Portuguese for example you see the cognates of those English suffixes (-ese/-ês, -an/-ano, -ic/-ego... just no -ish).

Except that for Portuguese simply inheriting the Latin suffixes wasn't enough, you got to reborrow them too. So you end with etymological doublets like -ego (see: Galícia "Galicia" → galego "Galician") and -co (see: Áustria "Austria" → austríaco "Austrian").

Then there's cases where not even speakers agree on which suffix applies, and it's dialect-dependent; e.g. polonês/polaco (Polish), canadense/canadiano (Canadian).

Besides afegão vs. Afeganistão (Afghan vs. Afghanistan), another example of a word where the demonym is shorter than the geographical name is inglês vs. Inglaterra (English vs. England). But it's the same deal: -terra is simply -land, so people clip it off.

There's also the weird case of "brasileiro" (Brazilian), that -eiro is a profession suffix. Originally it referred to people extracting brazilwood, then the country name was backformed from that.