Munutuni: The Ugly Language

[Chengjiang]

Last night, madness struck me. I was thinking about how there are features from natlangs I find aesthetically pleasing that tend to be overrepresented in my conlangs, as discussed in this thread. What if I challenged myself by setting out to build a naturalistic conlang out of features I dislike? I give you Munutuni, the language of monotony!

As is often done, I'll start with the

Phonology

Design Principles

Munutuni is designed to have an overall sound that I find displeasing, and in particular monotonous. To that end:

• It has a small set of sounds, while also having a small number of environments for them and distinguishing just the right POA and MOA contrasts so that certain likely paths for allophony/sound change are made less likely
• It has a very small number of legal syllables, so that the average number of syllables per morpheme is high
• It has the least interesting prosody possible
• It is missing basically any type of sound, phonemically and phonetically, that I have to hold myself back from adding if I'm leaving it out of a language (defying my own instincts is going to be a big part of designing this language)

Segment Inventory

EDIT: Linguolabial counterparts to the bilabials were added later, along with a mid central vowel. This early version is preserved for reference.

Munutuni distinguishes seven consonants and three vowels. They are listed below with their orthographic representations.

/m n p t k f s/ m n p t k f s
/i u a/ i u a

I was considering including voiced stops, but I decided against it because most of the languages I'm aware of with especially small inventories have significant allophony in voiced stops, such as having them vary with fricatives, liquids, glides, or nasals. I also avoided other sounds that tend to shift especially easily, so all postvelars and all velars besides /k/ were out. I did decide to include some fricatives, in particular /s/ to reduce the motivation for the very common shift in small consonant inventories of [t] > [s] / _i. I left out liquids and glides because both are types of sound I find difficult to leave out. (I have a hard enough time making languages with only a single liquid!) I also left them out because of their tendency to color vowels, which I normally like but obviously don't want here. The vowels are the smallest vowel system that doesn't normally have massive allophonic variation.

Phonetics

The coronals are apico-alveolar. The stops are as classically tenuis as can be, with a VOT near zero in all circumstances. The vowels mostly stick near the extremes of the vowel space. There's very little allophony worth noting. As in most languages that don't phonemicize palatalization or labialization, all consonants are mildly palatalized before /i/ (or, strictly speaking, fronted in the case of /k/) and mildly rounded before /u/. Vowels may optionally nasalize in the vicinity of nasal consonants.

Phonotactics

The only allowed syllables are CV. All combinations of C and V are allowed, yielding 21 possible syllables.

Prosody

Munutuni has neither contrastive stress nor contrastive tone. Any phonetic stress it may have is volume-based, very weak, and tending to appear on alternating syllables beginning with the first syllable after a pause. All syllables have a uniform length, and thus the language is syllable-timed by default. This and the low-sonorant inventory give the language a characteristic staccato sound.

-----

Let me know what you think about the phonology so far. The main theme I'm going for is "boring, but not by way of being average", so if you think I've missed a major way to do that, tell me.

I have some ideas for grammar coming up. Get ready for lots of defective fusional declensions and conjugations that are still complex enough to warrant tables, and an extremely pervasive masculine/feminine distinction.

[Soap]

Ha. This actually looks like a language I might like. Even the name caught my eye, until I realized where it came from. I am interested to see the grammar, since your least favorite features seem to line up a lot with my favorites.

The only allowed syllables are CV. All combinations of C and V are allowed, yielding 21 possible syllables.

So hiatus and word-initial vowels are illegal?

... the very common shift in small consonant inventories of [t] > [s] / _i...

Ive been meaning to look into that, to see why it seems to be so common. In languages with allophonic palatalization, I can see a path like tʲ > č > š, where [š] is perceived as just an /s/, but it seems to happen even in languages without palatalization. All I can think of is that the stop is just vaguely more affricated before an /i/, and that the change is usually [t] > [ts] > [s] but with the middle step usually moving on to the last step fairly quickly.



Also, I just now went and looked up Chavakani, figuring that if a conlang full of features you hate sounds nice to me, maybe a conlang you love would be something I hate. It isnt, but I do have a couple of minor questions, and if youre interested, i can post on the Chavakani thread?

[Chengjiang]

Ha. This actually looks like a language I might like. Even the name caught my eye, until I realized where it came from. I am interested to see the grammar, since your least favorite features seem to line up a lot with my favorites.

Then this should be very interesting indeed. I was kind of hoping somebody would post who liked a lot of what I disliked.

So hiatus and word-initial vowels are illegal?

Yes.

Ive been meaning to look into that, to see why it seems to be so common. In languages with allophonic palatalization, I can see a path like tʲ > č > š, where [š] is perceived as just an /s/, but it seems to happen even in languages without palatalization. All I can think of is that the stop is just vaguely more affricated before an /i/, and that the change is usually [t] > [ts] > [s] but with the middle step usually moving on to the last step fairly quickly.

I think something like that tends to happen, yes. Lots of languages have [tʲ] be at least a little bit affricated anyway.

Also, I just now went and looked up Chavakani, figuring that if a conlang full of features you hate sounds nice to me, maybe a conlang you love would be something I hate. It isnt, but I do have a couple of minor questions, and if youre interested, i can post on the Chavakani thread?

Chavakani isn't actually especially stuffed with features I like (that'd be my oldest serious conlang, Janaharian), but it's got a decent number of them, I suppose. Feel free to revive its thread if you're interested. Maybe it'll help motivate me to work on Chavakani more.

-----

Also, for the first time I'm considering using a generator to create the vocabulary. I normally spend a lot of time agonizing over giving my conlangs words that sound "right" to me for what they refer to (that's part of why they tend to progress so slowly), and so I suppose the logical thing to do with Munutuni would be to try and find the wrongest-sounding words, but I think I would never get anything done on the language if I tried to do that, so I think I'll trust random series of legal syllables to create words that, at the very least, aren't what I would have come up with. Any suggestions for good ones? In particular, ones that can generate words with a variable number of syllables?

[Cedh]

Ha. This actually looks like a language I might like. Even the name caught my eye, until I realized where it came from. I am interested to see the grammar, since your least favorite features seem to line up a lot with my favorites.

Then this should be very interesting indeed. I was kind of hoping somebody would post who liked a lot of what I disliked.

I like this phonology too. Looking forward to the grammar...

Also, for the first time I'm considering using a generator to create the vocabulary. [...] Any suggestions for good ones? In particular, ones that can generate words with a variable number of syllables?

http://akana.conlang.org/tools/awkwords/
(Make sure to have a look at the Help page to see how you can configure it.)

[Chengjiang]

Thanks, all. To be honest, the main thing that's actually ugly to me about this phonology is the rhythmic effect of rigidly equal-timed CV syllables, and it's something I don't get from syllable-timed languages in general. Something about specifically both being syllable-timed and never having codas or null onsets grates on me. (Which, come to think of it, is actually kind of a rare combination, isn't it? I'm actually having trouble thinking of languages with obligatory CV syllables and no variation between long and short rimes. Even Old Japanese allowed V syllables word-initially.) The segmental phonology is largely a matter of leaving out things I like rather than putting in things I dislike, since there aren't many sounds, and there are no common ones, that I dislike per se. I expected at least a few people would like the phonology.

http://akana.conlang.org/tools/awkwords/
(Make sure to have a look at the Help page to see how you can configure it.)

Excellent. Yes, I think I will use this.

...starting with the

Personal Pronouns

I haven't devised all forms of these yet, but here are the citation forms of all 11 basic personal pronouns:

fanufa - 1st person singular masculine
fisuki - 1st person singular feminine
nipamu - 1st person plural masculine
nifikina - 1st person plural feminine
tuta - 2nd person singular masculine
sisisa - 2nd person plural masculine
putu - 2nd person feminine
tumu - 3rd person singular masculine
pipi - 3rd person singular feminine
pituti - 3rd person plural masculine
sipu - 3rd person plural feminine

As you can see, all of these are synchronically separate morphemes with no apparent derivation for gender or number, although a couple of them happened to end up with suspicious similarities, like the 1st person plural pronouns. Mixed-gender groups take the masculine form. Note also that in the 2nd person, feminine gender, the singular/plural distinction is lost.

[Chengjiang]

I realized that since much of this project will involve using features I've rarely used, I may have some typological concerns. My first one concerns verb conjugation. I think what I want to do with verbs is have them conjugate for gender and pluractionality, but not for person. Does this seem typologically plausible? I'm pretty sure there are at least a couple languages where verbs conjugate for gender but not person (throughout the whole paradigm) due to the verbs being historically participles, but I'm not really sure what the typological correlates of pluractionality are and if they'd create any problems with this setup.

[Chengjiang]

I'm also a bit concerned I may have overdone the "small number of distinct syllables" thing. 21 possible syllables is substantially fewer than any natlang I'm aware of has, and I'm concerned it would necessitate an unreasonably fast rate of speech.

I'm considering expanding the segment inventory a little (not too much!) if it seems non-naturalistic.

[Soap]

Well, how realistic do you want to be? My conlangs are a mix of elements both realistic and impossible. No real human culture would maintain a language with a phonology such that Babioppompompibopom is an everyday word, yet the grammar of Poswa stridently stays within the boundaries of well known human languages.

Im still interested to see the rest of the grammar if you decide to keep this project going.

[Chengjiang]

I'd prefer to keep this one fully naturalistic.

One small expansion I was considering for the phoneme inventory is linguolabials. They've been found in small natlang inventories, don't have any obvious phonological prerequisites, and I kind of dislike them because I find them difficult to distinguish from bilabials and get tongue-twisted between them and both labials and coronals easily.

[Hydroeccentricity]

Maybe you can extend the number of syllables with some hideous clusters? tn- and pm- sound appropriately ugly.

[mèþru]

tnuctipun

[DesEsseintes]

Maybe you can extend the number of syllables with some hideous clusters? tn- and pm- sound appropriately ugly.

tn is lovely! I have tn in many of my conlangs; in some it's even its own phoneme as a prestopped nasal.

I like this language, Chengjiang, but I would also find it hard to work with a language with only 21 permissible syllables. How about throwing in labialised velars or labiovelars, and perhaps /h hʷ/ to go with that? (I don't know whether that goes against your principle of only permitting sounds with little to no allophony...)

[Ryan of Tinellb]

Is it just me, or are the voiced versions actually quite nice, bm- and dn-? And the one that's a little harder to transliterate, gng-?

[Chengjiang]

OK, thanks for the suggestions, everyone. I've decided what I want to do. I'm adding the following three consonants and one vowel:

/n̼ t̼ θ̼/ v q x or nm tp sf
/ə/ e

These bring the number of legal syllables to 40, nearly doubling it and bringing it somewhat above Central Rotokas's 35 legal morae* and thus well within the range of plausibility. And I think adding this set of sounds is a good fit for a language fitting my own dislikes while keeping the number of syllables per morpheme high but not implausibly so; I've seen how many syllables morphemes in Rotokas tend to have.

I am undecided on what representation to use for the linguolabial consonants. The digraphs are a little more intuitive, but given what they are nothing automatically suggests their sound to an unfamiliar learner anyway, and the single letters have the advantage of being concise. Also I'm not sure what I want as far as making the romanization "ugly".

In a little bit I think I'll revise the personal pronouns somewhat, since the total absence of all of these sounds from that large a set of morphemes is a bit strange. I will keep certain things I like about this set, like nipamu vs. nifikina, and the (to me) awkward-sounding repetition of sisisa.

*Central Rotokas has long vowels and diphthongs, but as far as I know they don't form minimal pairs with sequences of short vowels. At any rate, it has about the number of morae per morpheme on average that I'd be going for with Munutuni, which has no syllables longer than one mora.

[Chengjiang]

Here are the revised personal pronouns, with segments somewhat re-randomized to fit the new inventory. I've decided, at least for the moment, to use v q x for the linguolabials, as above.

fanufa - 1st person singular masculine
fiseke - 1st person singular feminine
niqemu - 1st person plural masculine
nifekina - 1st person plural feminine
teqa - 2nd person singular masculine
sisisa - 2nd person plural masculine
qutu - 2nd person feminine
qumu - 3rd person singular masculine
pipi - 3rd person singular feminine
qiquti - 3rd person plural masculine
sepu - 3rd person plural feminine

Tell me what you think of these. My tongue is already stumbling quite a bit at having to alternate between tongue and lip connecting and tongue and lip doing separate things. I think this was a good idea.

[Salmoneus]

I'm a little confused. Why is it called 'the ugly language' when you're trying hard to make it as beautiful as possible?

[mèþru]

Because it is ugly from Chengjiang's point of view.

[Salmoneus]

Because it is ugly from Chengjiang's point of view.

Really? That's interesting to learn. While you're at it, could you explain the concepts of "irony" and "humour" to me, perhaps?

[mèþru]

I strongly suspected it was sarcasm, but I wasn't sure. I had a tendency to find patterns that my peers don't because of different knowledge areas and get frustrated. To counter that, I started trying to explain what is obvious to me alone. I now often point out what is obvious to everyone because I didn't know it was obvious.

[Chengjiang]

Grammar: Design Principles

Munutuni's grammar is designed to maximize morphology necessary for agreement while minimizing morphology that elucidates meaning and could lead to words being left out due to context providing the necessary information. To that end, most inflectional paradigms, while large, will be heavily defective, necessitating redundant marking. Verbs will also specifically not conjugate for person, discouraging pro-drop. (I realize that doesn't make pro-drop impossible; see Japanese.)

Here are some other decisions I've made:

• The MSA is split ergative based on TMA, with more of it being ergative than accusative.
• All cases have non-intuitive uses, due in part to older cases becoming homophonous.
• There are no genderless words in the language whatsoever, regardless of part of speech, though all other forms of marking have gaps.
• "Conjunctions" do not exist. Their role is filled by adpositions and verbs.
• Morphology is a mixture of prefixing, suffixing, and infixing, leaning toward prefixing.
• Word order is inflexible (SVO) despite the extensive morphology, since so much of the case marking is defective and verbs do not conjugate for person.

I realize this system seems ripe for regularizing collapse, but I also know that there are some natlangs with incredibly Byzantine morphology. If the amount of memorization required starts to look unreasonable, I think I'll have what I'm describing here be a somewhat calcified court/official standard and not necessarily reflective of the more regularized speech of commoners.

Here's some other sample vocabulary, with meanings not assigned yet. All of it is nouns in the citation form, the absolutive singular. Some of it may get prefixes later, depending on how much I want to have a zero morpheme for this case and number.

xavasa pese tiqeva punavafa mutufupa viki fikupi vasipe fenikuqe pekexu vimequ kufuki xumuqavu sekape xuxi xemetixu sivitika fuxa pipeva tuxesena mikife kevuxa kusu miki qunupase qeqimine faniki piqeqe vute sexapi mipakipa fiqeqi vatu vufu sapuxa mepumave mefepa qevexamu sisete kife pumufe pimani xevuxe samefuku vive kakeqepa qesami pukuva kisi fekefuva sikimu mimumi setupu fuki kakime nuqavi vaqu kisuvuni tisuqe qixika tinutefi katimani kuvuxa xunu xunipi vevu memape qufivu veva kumukupa tifafefa vivipu qufesa sesa tika vafe xupisi tifiti nasetetu tesa supuxe mexi kanutasi munuqipu sapaka kufuxu mafita sapiqa tetiteka pexixa ketafe musune maqumi kumu nenufima piquna sitena temu viqeme kuxupa

[Soap]

Munutuni's grammar is designed to maximize morphology necessary for agreement while minimizing morphology that elucidates meaning and could lead to words being left out due to context providing the necessary information. To that end, most inflectional paradigms, while large, will be heavily defective, necessitating redundant marking. Verbs will also specifically not conjugate for person, discouraging pro-drop.

I like this idea. If I were to make a degenerate "i dont like it" language from one of my own conlangs, I would do something very similar. I'd keep all of Poswa's verbal endings, which are often two or three syllables long, but have some of them merge with others due to sound changes in such a way that they no longer function and must be coupled with otherwise redundant standalone morphemes. It would be as if Latin speakers were forced to translate "of the house" as de illīus casae instead of just casae.

I believe Albanian requires a definite article even for adjectives, as well as one for nouns. If your definite article is three syllables long, as above, that would make it even more absurd.

I like the expanded phonology, though for me, orthography really means a lot. I have a very hard time reading q as anything but /q/ or x as anything but /x/ or a similar sound. I went through and pasted the sample text into MS Word and replaced all {x q} with {w b}. because I associate linguolabials with bilabials since neither of them has a "sharp" element in the sound. I expected the result to look like something I would make, but it didnt. I think the relentlessly predictable ... and monotonous .... CV rhythm turns me off, so I'm actually going to say that in the end I actually share your dislike of the sound of this language.

[mèþru]

One of my conlangs also requires articles/demonstratives for adjectives. Gender agreement is actually done via taking the same article/demonstrative (which have three genders: masculine, feminine and neuter, but the last one only exists in nominalised words).

[mèþru]

One of my conlangs also requires articles/demonstratives for adjectives. Gender agreement is actually done via taking the same article/demonstrative (which have three genders: masculine, feminine and neuter, but the last one only exists in nominalised words).

[Chengjiang]

I like this idea. If I were to make a degenerate "i dont like it" language from one of my own conlangs, I would do something very similar. I'd keep all of Poswa's verbal endings, which are often two or three syllables long, but have some of them merge with others due to sound changes in such a way that they no longer function and must be coupled with otherwise redundant standalone morphemes. It would be as if Latin speakers were forced to translate "of the house" as de illīus casae instead of just casae.

Exactly! Although I prefer doing (mostly) naturalistic languages, I guess I do have a little bit of engelanger mentality because I regularly dislike "inefficiency" in a language's structure.

I believe Albanian requires a definite article even for adjectives, as well as one for nouns. If your definite article is three syllables long, as above, that would make it even more absurd.

I was going to have definiteness be part of the nominal declension for more memorization, but I will admit leaving it a separate word that has to agree gives more opportunities for pure volume of cruft.

I like the expanded phonology, though for me, orthography really means a lot. I have a very hard time reading q as anything but /q/ or x as anything but /x/ or a similar sound. I went through and pasted the sample text into MS Word and replaced all {x q} with {w b}. because I associate linguolabials with bilabials since neither of them has a "sharp" element in the sound.

I'm definitely open to suggestions about the Roman orthography. I decided on these symbols because (aside from <v>) they often seem to get used as "whatever" characters in natlangs, and for some reason I wanted to keep the relentless CVCVCV rhythm looking as regular in text as in speech and so I didn't want to use digraphs, but I do realize these characters call up other associations.

I might do something similar to your suggestion, although I think I'd switch <w> for <v> so that <b> represents a stop and <v> a fricative, leaving <w> for the nasal. At least in that case the most obvious spelling pronunciation wouldn't be as far off as what I have been using.


That said, I don't expect to ever make this a "general consumption" conlang. It's mostly just a challenge for myself and, hopefully, some amusement for other board members.

And thus far, it really has been a challenge, in no small part because in order to make it make sense I have to weigh exclusion of features I like versus inclusion of features I dislike. For example, I like dual number, but (for some reason) dislike trial number, which implies the presence of a dual; in the end I decided leaving out the dual was more important, especially since natlangs generally limit trial to pronouns anyway.

I expected the result to look like something I would make, but it didnt. I think the relentlessly predictable ... and monotonous .... CV rhythm turns me off, so I'm actually going to say that in the end I actually share your dislike of the sound of this language.

It kind of crept up on you, eh? Interesting.

[Chengjiang]

Here's a weird idea I had that I'm not sure is attested in any natlangs: verbs and adjectives that are suppletive by gender. Just as some languages have suppletive verb stems by tense in some cases (e.g. forms of the copula in many Indo-European languages), Munutuni would have different stems for some common verbs by gender rather than using the normal masculine and feminine inflections of one stem. For instance (warning: very non-finalized morphology ahead):

tipe-nuvesa
MASC.SG.AORIST-speak.MASC

versus

xumu-kakive
FEM.SG.AORIST-speak.FEM

with the prefixes being the normal conjugation morphology of verbs in this gender/actionality/TAM.

Does this look like something that could happen? I was thinking, in order to make it especially displeasing to me, of making some of these pairs of suppletive stems have a slight difference in connotation between stems.

This phenomenon happening with both verbs and adjectives was inspired in part by pairs of adjectives in some languages (especially English) that are more or less synonyms but are almost exclusively used with different genders and have gendered connotations, e.g. "pretty" versus "handsome".