The Most Batshit Natlang Competition!

[Drydic]

...that's pretty much what I said.

[Hallow XIII]

Yes, but you appeared to have confused my statement for "pie didnt exist lolololol!!!!!111!!11!!!!111!1!!1". So I went to clarify.

[Drydic]

No, I was objecting to it being called a conlang. It really doesn't fit, imo.

[Hallow XIII]

Okai. Reconlang.

[jmcd]

The person responsible for the dictionary entry has defended himself.

The upshot: it's a dialectal thing with the usual weirdness candidates of Lewis and Skye being among them as well as some other western dialects. Elsewhere it's rarer.

EDIT: On another note, is there an equivalent of this for conlangs?

Ceart gu lèor, leugh mi ann leabhar sin nuair a ionnsaich mi ann an Obar Dheathain ach cha tug mi aire. Thug mi aire do rudan eile mar [D] an àite [r].

Oh cool I was reading in that book when I was studying at Aberdeen but I didn't notice that. I did notice other things like [D] for [r].

[Hallow XIII]

The person responsible for the dictionary entry has defended himself.

The upshot: it's a dialectal thing with the usual weirdness candidates of Lewis and Skye being among them as well as some other western dialects. Elsewhere it's rarer.

EDIT: On another note, is there an equivalent of this for conlangs?

Ceart gu lèor, leugh mi ann an leabhar sin nuair a dh'ionnsaich mi ann an Obair Dheathain ach cha tug mi aire dha. Thug mi aire do rudan eile mar [D] an àite [r].

Oh cool I was reading in that book when I was studying at Aberdeen but I didn't notice that. I did notice other things like [D] for [r].

[Yiuel Raumbesrairc]

PIE-as-shown is definitively something near an a-posteriori conlang (if not one); the very fact that mixes between dialects are entirely possible makes it likely that the PIE-as-shown has never been spoken. A PIE happens to have existed, and it is probably close to PIE-as-shown (though there could be striking differences as well), but PIE and PIE-as-shown are probably not the same thing.

That would be like mixing Edo sound change /ai/ -> /ee/ with Minako sound change /auta/ -> /oota/. And if anyone did that with the current various dialects of Japan, well it would be called a conlang.

[Karinta]

Archi has both batshit phonology, being a Caucasian language, and a batshit grammar. The verbal morphology and the extremely fusional aspects of its grammar make it batshit insane in that department.

Also, Navajo. Verbs are insane, noun classifiers, all of that. However, the phonology's okay.

[baradsonoron]

The wikipedia article refers to a certain Nader from 1984. It can either mean "he had had a bunchberry plant", or "Abandon all hope, ye who try to learn this language".

Or "help me i'm choking".

[Morrígan]

Archi has both batshit phonology, being a Caucasian language, and a batshit grammar. The verbal morphology and the extremely fusional aspects of its grammar make it batshit insane in that department.

It's hard to say - Archi isn't really anymore extreme than Khwarshi. Honestly, all the Dagestanian languages that still have noun-class markers are pretty wild.

[Whimemsz]

I am a little disappointed no one ever brought up Algonquian languages! They have a pervasive animate/inanimate distinction, proximate/obviative distinction, and direct/inverse alignment. In my experience other languages and families described as having direct/inverse alignment and/or a proximate/obviative system usually pale in comparison (the only really comparable example that I know of is Kutenai--whose system is no doubt similar to the Algonquian one due to contact--but even in Kutenai the inverse system is only present for third persons). Also Arapaho and Gros Ventre lack phonemic low vowels (vowel system: /i ɨ~u ɛ ɔ/ plus some diphthongs and triphthongs plus length), so I guess my vote has to go for them specifically. (Also relevant are the batshit sound changes that characterize the Arapahoan languages)

(NB: The only real fight I can envision would be put up by Kiowa-Tanoan languages, which do have a form of inverse alignment, have incredibly elaborate person inflection on verbs, have nifty phonologies, and also sport inverse number systems. Although if I'm being truly unbiased I have to admit the verb systems of Athabaskan languages are ... absolutely insane...)

[Morrígan]

Kutenai

Oh... yeah. Kutenai. I think I blocked that out - Dryer loves talking about Kutenai.

I remember Ichiskiin being fairly weird, but it's Penutian and probably not that odd for a language of North-Western America.

[baradsonoron]

I vote Navajo because it's so hard to count in it, let alone do verbs (Is a towel large and flexible or skinny and flexible? What about a length of fabric?)

[marconatrix]

I never said it was bad per se. It's one of the reasons why I prefer Scottish Gaelic over Irish actually (the others being back unrounded vowels and the continued existence of word-final [ɣ]). It's just that for the purpose of pronouncing a written text, this sort of thing can catch you off guard. I still don't properly know how to do it with t, especially since that lovely consonant can take either [x] or [ʃ] as its preaspiration phoneme.

The preaspiration is only [ʃ~s] in "rt" (and for some speakers "rd") otherwise it's just [h], but mostly in stressed syllables after a short vowel.

You want real batshit, how about Manx spelling ...

[Vardelm]

You want real batshit, how about Manx spelling ...

No kidding. My grandfather was from the Isle of Man, so I've looked at the Manx language a bit. Just..... wow....!!!

[marconatrix]

You want real batshit, how about Manx spelling ...

No kidding. My grandfather was from the Isle of Man, so I've looked at the Manx language a bit. Just..... wow....!!!

It's actually not too far from Scots Gaelic, you have to kind of read it with your eyes shut, if that makes sense ...

[Hallow XIII]

The point is, Manx is basically Scots Gaelic spelt with english rules and and added treat from Welsh in that it writes schwa as <y>. This produces such wonderful cognate pairs as ScG ithidh and Manx eeee, "eating" (where ee is the verb root "eat" and ee is the VN marker).

I never said it was bad per se. It's one of the reasons why I prefer Scottish Gaelic over Irish actually (the others being back unrounded vowels and the continued existence of word-final [ɣ]). It's just that for the purpose of pronouncing a written text, this sort of thing can catch you off guard. I still don't properly know how to do it with t, especially since that lovely consonant can take either [x] or [ʃ] as its preaspiration phoneme.

The preaspiration is only [ʃ~s] in "rt" (and for some speakers "rd") otherwise it's just [h], but mostly in stressed syllables after a short vowel.

Note that I made that post months ago. Your explanation is appreciated, but no longer needed at this point.

[Richard W]

I mean I think Thai is regular, but it's far more complex than English, which I guess is a different question.

It works well enough for monosyllables, but for longer words you have to chop the word up into syllables, and that is *computationally* difficult, and strictly speaking, impossible. There are also irregularities in how the tone of one syllable affects the tone of the next - this is an orthographic issue, not a phonological one. Oh, and of course, you also have to chop phrases up into words - space is the main punctuation mark in native mainland SE Asian scripts.

http://en.wikipedia.org/wiki/Lao_alphabet - here you have multiple letters for quite a lot of consonants and at least half the vowels, and even more of them merge together for the coda (which I guess must be morphophonemic – Korean does this too incidentally). I think Lao is tonal, which would explain it... except WP hasn't explained that.

1) WP did and does explain that the different initial consonants are for the three different sets of rules for determining tone - there is (and was when Finlay's post was made) a column headed 'tone class'.

2) The different letters for the coda were only morphophonemic in so far as Indic loans (and coinings?) are concerned. The spelling reforms abolished these differences - final /t/ is always written the same way in reformed Lao. There may be some differences for recent loan words. Of course, the article may also be intended to partially address older spellings - the article doesn't distinguish Lao script and Lao writing system.

3) Apart from /ai/, the multiple vowel sequences for the same sound are for the difference between open and closed syllables, or at least, they are if you treat the glottal stop as non-phonemic.

4) For /ai/, I'm not sure that the distinction between /ai/ and /aɯ/ is completely extinct, which would explain the preservation of two distinct symbols.

[Trebor]

Pfft, tone doesn't make a language crazy.

Well, if tone terracing is an added feature, it seems to me that pronunciation and comprehension will be seriously impeded.

Try learning Akan/Twi and let me know how far you get.

[Trebor]

I have to wonder how much of the perceived batshittery of Somali is just due to bad analysis on the Wikipedia page. It could easily be cleaned up into something more sensible, although I'm not sure how accurate this would be:

Code: Select all

  m   n      
  b t d ʧ ɖ k g ʡ   ʔ
  f s   ʃ   x   ħ ʕ h
      r     
      l   j   w

So the only things that are weird then are the retroflex plosive and the vowels.

What do you find strange about the Somali vocalic inventory? According to what I've read, they include short and long versions of the five Spanish/Italian vowels and not-quite-phonemic pharyngealized versions, or something like that.

Also, isn't the epiglottal stop in your table actually uvular?--As in, for instance, qaamuus "dictionary", a clear Arabic borrowing.

[Trebor]

http://www.armazi.com/georgian/

It's been the standby georgian grammar online for probably a decade, though through several different sites. I don't know about its accuracy as I haven't made myself look through it.

The link isn't working anymore... I'd be interested to know where to find that great description of Georgian grammar now.

[Trebor]

I'll nominate Tariana, not because of its phonology but because of the insane amount of stuck it packs into its syntax and morphology. This is mostly thanks to it being an Arawak language heavily influenced by neighbouring Tucano languages, so it ended up with the grammatical categories of both families more or less.

Where do you recommend we read about this language online?

[Trebor]

+1 for English. Or at least the western Germanic languages.

1. Northwestern European languages are weird: perfect tense with "have", inverting word order to form polar questions, relative pronouns.

What about phrasal verbs in English? I really don't envy ESL students for having to learn them.

It is funny to see Zulu mentioned as one of the centers of areas of rareness in that article. Weird things in grammar I always call "wtf-features", that is features of a language that make you go wtf. I think Zulu has a very normal grammar, and it, and other South-African Bantu languages, don't have a few of the wtf-features other Bantu languages have such as subject-object reversal and locative noun classes.

I'd like to learn more about such odd Bantu features. If you have any links, please share.

[marconatrix]

The point is, Manx is basically Scots Gaelic spelt with english rules and and added treat from Welsh in that it writes schwa as <y>. This produces such wonderful cognate pairs as ScG ithidh and Manx eeee, "eating" (where ee is the verb root "eat" and ee is the VN marker).

Not quite, "eeee" is the 2pl imperative (SG ithibh), if the thing you're being instructed to eat is feminine, then "eat (ye) it!" becomes "eeee ee!", honest In SG, some dialects at least, replace the 'th' which would be realised as a weak [h] or simply hiatus by [ç] which makes things a bit clearer. Likewise in a few other words, _ruith_ 'run' can be something like [rœç].

[marconatrix]

I have to wonder how much of the perceived batshittery of Somali is just due to bad analysis on the Wikipedia page...

The weirdist thing I remember about Somali is that there are no prepositions in noun phrases, they are prefixed to the sentence-final verb, often with stuff in between. Hasn't it been suggested that PIE prepositions were originally verb-modifying prefixes like this?