Odd to us native English speakers, anyway.  To non-native speakers, our whole language is odd.  Anyway, I just felt like writing this, so here we go:

Magyar (Hungarian) has almost no relative, only distant ones:  Finnish and Estonian (which are not far apart from each other).

German and Arabic have four cases (meaning the noun changes to show its function in the sentence).  Latin has five, Russian six.  Finnish has, good lord, fifteen.

In Hebrew and Arabic, a ‘conjugation’ doesn’t mean what it means in French or Spanish.  It means actual changes to a root that make it into different verbs.

Irish has arcane rules that require changes to spelling and pronunciation at certain times.  Most of those spare Hs you see are really more like diacritical marks, that somehow affect the preceding letter.  FH, for example, is completely silent.  There are two different methods for this, and only some letters can be modified this way.  And of course, the two groups only have some overlap.  So C can become CH or GC.

French makes its spoken past tense with different helping verbs.  Thus, ‘he is died’ but ‘I have spoken.’

Bad news about Greek:  none of the accented syllables have accents where we intuit they should be.  Good news:  nearly every word incorporates an accent mark, so you’re going to know anyway.

Any word in Spanish over a certain number of syllables (four +, I think) has an accent mark to show the stress.  In French, however, there are no stressed syllables.  The marks you see affect only pronunciation.

Turkish has two Is, one without a dot.  The one without the dot is pronounced more like IH, like in ‘fish.’  All things considered, they adapted their language to the Latin characters pretty well.  It wasn’t even a century ago.

No matter how many Dutch people you ask, you will never properly pronounce ‘Schiphol.’  Do not bother.  Every Dutch word you try and read aloud, you are butchering, unless you actually studied Dutch.  Happily, speaking only three languages is considered borderline mentally challenged for a Dutch person, so they very likely speak another language that you also speak.

Swedish not only has a word for ‘so’ (try to define it sometime for a non-native speaker!), it sounds the same: så.  The ring makes an A a long O sound, so every time you said ‘awng-strom’ for Ångstrom in physics, you were wrong.  ‘oang-stroom, ‘ with vowels like ‘oat broom.’

Russian has no articles.  None.  (Neither does Swahili.)  This explains why Russians learning English find the concept so very challenging, much as we find verb aspect (a different verb for whether the action is completed or ongoing) a challenging aspect of Russian.  (In Spanish and French, this is handled by choice of tense.)

Latin is named for Latium, the part of ancient Italy where Rome was.

Swahili has noun classes based on what the noun represents:  people, objects, abstract concepts, etc.  One differentiates these with prefixes.  Thus, a group of Tutsi are ‘Watutsi’ (yes, that’s where the old dance name comes from).  The language is ‘Kiswahili.’

Afrikaans is essentially an evolved, simplified Dutch.  By the time it was named Afrikaans, it had gotten rather farther from Dutch than Australian from British, or arguably, Québécois French from Parisian French.

A lot of the swear words in Parisian French are scatological.  A surprising number in Québécois French are religious.

If you know swear words in Spanish, don’t use them.  They are taken by native speakers as sullying a beautiful language, and can mark you as a person of the lowest social class.  Also, say ‘mamá’ rather than ‘madre.’  The latter carries overtones that are a short throw from speaking of all Hispanic mothers anywhere, not a group you’d want to slight in even a minor way.

Swedish has two genders:  common and neuter.  They used to have three, but they combined masculine and feminine, as a national decision, about a century ago.  One of the advantages to having a native language spoken nowhere else in numbers:  whatever the Japanese say is Japanese, for example, no one’s in a position to debate with them.

Icelandic is so similar to Old Norse that any literate Icelander can read the ancient sagas without trouble.  Convenient, as there are just about zero illiterate Icelanders.

What makes English so hard for non-native speakers to learn well? It’s not our grammar, which isn’t too bad.  It is the inconsistent spelling, immense vocabulary, and multiple meanings a given word can have.  Proof of the difficulty comes in how atrociously even many native speakers write English–it’s hard even for us.

Arabic has 28 letters.  Most have four forms: initial, medial, final and isolated.  The ‘unfriendly’ letters, have only two forms:  final and isolated, as they cannot join to the next letter.  The dots are part of the letter; many letters can be told apart only by the number and location of dots.  Thus, a TH looks like a B except that TH has three dots above, and a B has one dot below.

Farsi uses the Arabic alphabet written in a slightly different style, and adds four letters not found in Arabic:  P, CH, ZH and a sort of G.

If Spaniards seem to lisp their Spanish (to ears used to LatAm Spanish), they don’t have speech impediments.  Z and soft C are pronounced TH in Castilian.  Una therbaytha, por favor.

Finns find the consonant blends of English terribly difficult to master.  That’s okay.  Everyone else in the world, except Estonians, finds everything about Finnish impossible to master.