The ☃☃☃☃☃☃☃☃☃☃☃☃☃☃☃ language

[Nortaneous]

Inventory:
There is one phoneme, /☃/.

Phonotactics:
The following are permitted syllables:
* ☃
* ☃☃
* ☃☃☃
* ☃☃☃☃
* ☃☃☃☃☃
* ☃☃☃☃☃☃

Allophony:
* /☃/ is realized as [ɤ]; /☃☃/ is realized as [a]. Due to a historical shift from quantity to quality contrasts, /☃☃☃/ is realized not as [a:] but as . Sequences of more than three ☃s arise from historical patterns of absorption of /☃/s into a preceding /☃☃☃/, giving [jV] - but after frication and devoicing, /☃☃☃☃/ is [sɤ] and /☃☃☃☃☃/ is [sa]. /☃☃☃☃☃☃/, of course, is [ɕi].
* [ɤ] may be optionally rounded to [o].
* An epenthetic consonant is inserted before each syllable, except those beginning with sibilants: [h] before /☃/, [k] before /☃☃/, and [t] before /☃☃☃/, /☃☃☃☃/, /☃☃☃☃☃/, and /☃☃☃☃☃☃/.
* Syncope: every other vowel drops. Epenthetic consonants are preserved, except for [s], which devoices adjacent consonants -- which are otherwise voiced (note that [s] voices to [r]) -- and drops -- and [h], which drops without a trace.
* Intervocalic consonants are voiced unless devoiced by [h]. [h] itself voices to .
* Word-initial /☃/ drops and nasalizes the whole word, due to historical rhinoglottophilia. [h t k] nasalize to [m n ŋ]. However, this process is blocked by elided /☃/.

So the name of the language, /☃☃☃☃☃.☃☃☃.☃.☃☃.☃☃☃☃/, is pronounced [satokso]

An example sentence:
/☃.☃☃☃-.☃☃☃☃.☃.☃☃.☃-.☃☃ ☃☃.☃☃☃.☃.☃☃☃☃☃☃ ☃☃☃.☃☃☃.☃.☃☃☃☃.☃☃☃.☃☃☃☃☃.☃.☃☃.☃☃☃.☃☃☃☃ ☃☃☃☃.☃☃.☃☃☃.☃☃☃☃☃.☃/
[nisogok kados tidotiroktis soktiro]
COP-SENS.PRES-PFV DEM.PROX stupid Q
This is very stupid, isn't it?

[Sakir]

Is that a little snowman? With a little fez? What's he doing in unicode?

[äreo]

I dunno, man. I think the orthography may not be minimalistic enough.

[Pole, the]

I am defeated now. :<<<

[gach]

Right.

So what about the phonemic status of the separator /./? Since it's clearly vital to know its placement with respect to the surfacing elements /V/ and this is already a more abstract than necessary analysis, shouldn't it actually be treated an equally phonemic segment than the surfacing one yielding you a two segment system? Can you provide us any evidence why this couldn't or shouldn't be done (history doesn't count)?

[Rhetorica]

* ☃ -> ⚊
* ☃☃ -> ⚌
* ☃☃☃ -> ☰
* ☃☃☃☃ -> ☱
* ☃☃☃☃☃ -> ☳
* ☃☃☃☃☃☃ -> ☷

Bam. No more /./:

⚊☰☱⚊⚌⚊⚌ ⚌☰⚊☷ ☰☰⚊☱☰☳⚊⚌☰☱ ☱⚌☰☳⚊

[Pole, the]

* ☃ -> o
* ☃☃ -> a
* ☃☃☃ -> ı
* ☃☃☃☃ -> ö
* ☃☃☃☃☃ -> ä
* ☃☃☃☃☃☃ -> ï

‹oıöoaoa aıoï ııoöıäoaıö öaıäo›
[nisogok kados tidotiroktis soktiro]

[gach]

But how is that then different from an analysis with six distinct phonemes?

[Particles the Greek]

Have you considered using U+1f4a9?

[Torco]

But how is that then different from an analysis with six distinct phonemes?

I suck at linguistics and even I understood that it is a joke that hinges on the fact that the answer to your question is that it doesn't.
plot twist: its not and i'm tripping

[gach]

I suck at linguistics and even I understood that it is a joke that hinges on the fact that the answer to your question is that it doesn't.
plot twist: its not and i'm tripping

But that's hardly an interesting joke in comparison to discussing the discrepancy between data and descriptions having an agenda.

[Nortaneous]

Diphthongs and complex vowels are sometimes analyzed as two separate vowels in the same syllable, and long consonants can be analyzed as sequences of two separate consonants. Estonian has sequences of three separate consonants, and there was some Sami language that distinguished four lengths. So six lengths isn't all that far from there.

There are affixes with gradation effects, reducing or increasing the length:

/☃☃☃.☃.☃☃.☃☃☃.☃☃.☃.☃☃☃☃☃/
[tigatkara]
snow.NOM

/☃☃☃.☃.☃☃.☃☃☃.☃☃.☃.☃☃☃☃☃-☃☃.☃☃☃☃☃☃/
[tigatkaroki]
snow-LOC

/☃☃☃.☃.☃☃.☃☃☃.☃☃.☃.☃☃☃☃☃☃-☃☃☃☃.☃☃☃/
[tigatkariti]
snow-BEN

and

/☃.☃☃.☃☃☃.☃.☃☃/
[ŋanok]
snowman.NOM

/☃.☃☃.☃☃☃.☃.☃☃-☃☃.☃☃☃☃☃☃/
[ŋanokaɕ]
snowman-LOC

/☃.☃☃.☃☃☃.☃.☃☃☃-☃☃☃☃.☃☃☃/
[ŋanotot]

[Rhetorica]

Have you considered using U+1f4a9?

That's out of the Basic Multilingual Plane; you can't use it here.

[Kereb]

CAN YOU PLEASE TO SEND NUMERALS 1-10 IN ☃☃☃☃☃☃☃☃☃☃☃☃☃☃☃ LANGUAGE

[kusuri]

Yes, please.
I would love it if Janko actually added them to his collection.

[Risla]

1. ☃
2. ☃☃
3. ☃☃☃
4. ☃☃☃☃
5. ☃☃☃☃☃
6. ☃☃☃☃☃☃
7. ☃☃☃☃☃☃☃
8. ☃☃☃☃☃☃☃☃
9. ☃☃☃☃☃☃☃☃☃
10. ☃☃☃☃☃☃☃☃☃☃

[Torco]

how about
1 ☃
2 ☃☃
3 ☃☃☃
4 ☃☃☃☃
5 ☃ ☃☃☃☃
6 ☃☃☃ ☃☃
7 ☃ ☃☃☃ ☃☃
8 ☃☃ ☃☃☃☃
9 ☃ ☃☃ ☃☃☃☃
10 ☃☃ ☃ ☃☃☃☃

[KathTheDragon]

Alternative analysis:
☃
☃:
☃::
☃:::
☃::::
☃:::::

[Nortaneous]

man i *did* get jankoed

sending him torco's but what's the pattern? it looks familiar but i can't place it

[Kereb]

man i *did* get jankoed

sending him torco's but what's the pattern? it looks familiar but i can't place it

it makes sense if you stick some math symbols in it

5 ☃ + ☃☃☃☃
6 ☃☃☃ × ☃☃
7 ☃ + (☃☃☃ × ☃☃)
8 ☃☃ × ☃☃☃☃
9 ☃ + (☃☃ × ☃☃☃☃)
10 ☃☃ × (☃ + ☃☃☃☃)

[Nortaneous]

^ oh, how'd i miss that? that reminds me of how it's possible to generate any non-negative integer up to at least 100 with four fours if you put the symbols in the right place...

as for the properties of the language:

disregarding nasals, there are 30 possible two-underlying-syllable sequences, where <c> is the palatalized fricative and <g> is the velar nasal bc lazy

Code: Select all

  1   2   3   4/5 6 <- C2
1 hoh hok hot hos hoc
2 kah kak kat kas kac
3 tih tik tit tis tic
4 soh sok sot sos soc
5 sah sak sat sas sac
6 cih cik cit cis cic

and then you get nasals -- so the only phonetic syllables that can occur word-initially, ignoring codas, are [ho ka ti so sa ci mo~ ga~ ni~ so~ sa~ ci~] -- which is a clue that something very interesting is going on on the phonemic level

but after that, i *think* it's possible to get any CV(C) sequence, except nasals can't occur after voiceless consonants in a word

[gach]

Diphthongs and complex vowels are sometimes analyzed as two separate vowels in the same syllable, and long consonants can be analyzed as sequences of two separate consonants. Estonian has sequences of three separate consonants, and there was some Sami language that distinguished four lengths. So six lengths isn't all that far from there.

Analysing length and diphthongs as arising from combining separate single phonemes certainly works and is very often the preferable way to go. In fact I think that people give a phonemic status to long vowels and diphthongs far too often when there seems to be no reason for doing it and analysing them as strings of simple vowels without an intervening syllable break would give a much simpler description of what's going on.

The four consonant lengths of word internal consonantisms which can mostly be found in historical varieties is Saami are a bit extreme and the fact that they've mostly collapsed into two or three lengths tells very much about their stability. The complex system of consonant length is furthermore connected to the consonant gradation so it's never exactly possible to just freely sprinkle around the length grades as people gaining new vocabulary would want them. Anyway, in the case of this project naturalism is hardly what you are aiming at and to me there isn't too much value trying to motivate things on how natural languages do them.

So to return to the analysis of the phoneme inventory, it's equally possible to interpret it as multiple units differing in phonemic length or as a single surfacing element which gets combined in various ways. You can just never get down to an analysis where you can grant phonemic status to only a single segment. What determines the correct parsing of a sequence if you don't know where a break /./ must land? And if their positions can't be predicted, then why shouldn't their assignment be granted phonemic status as well?

A truly monophonemic system would generate words with no internal phonemic structure at all. These would differ from each other only by their length, i.e. the number of phonemes contained in the string, and knowing the length of a word would uniquely predict its shape.

[Nortaneous]

Should the syllable break count as a phoneme?

Are there any natlangs where /a.a/ contrasts with /a:/, or /a:.a/ with /a:.a/ or /a.a.a/? Or where /s.s/ contrasts with /s:/?

Are there any natlangs where syllabification is unpredictable -- so /as.ta/ could contrast with /a.sta/ or /ast.a/, or /as.a/ with /a.sa/?

[tezcatlip0ca]

Are there any natlangs where syllabification is unpredictable -- so /as.ta/ could contrast with /a.sta/ or /ast.a/, or /as.a/ with /a.sa/?

English could count for this, at least marginally on a morphemic level: basket vs. glass kit, distress vs. this tree.

[Yng]

Isn't over word boundaries kind of cheating?