If random letter sequences were actual words

[Raphael]

During my various recent internet-and-technology related (mis)adventures, at one time I came across an automatically generated file whose file name, among other things, contains the letter sequence "cgespzjvvskiaari". And that made me wonder: What if that were a real word? The first part looks perhaps a bit Czech or Polish to me, while the last part looks more like Finnish. Or perhaps someone could take it as a conlanging challenge: "Create a conlang in which 'cgespzjvvskiaari' is a word."

[xxx]

in 3SDeductiveLanguage(1Sign=1Sound=1Sense) cgespzjvvskiaari can be converted into meaningful units,
but I'm afraid the whole thing doesn't make sense...

[zompist]

During my various recent internet-and-technology related (mis)adventures, at one time I came across an automatically generated file whose file name, among other things, contains the letter sequence "cgespzjvvskiaari". And that made me wonder: What if that were a real word? The first part looks perhaps a bit Czech or Polish to me, while the last part looks more like Finnish. Or perhaps someone could take it as a conlanging challenge: "Create a conlang in which 'cgespzjvvskiaari' is a word."

I may not be getting the question, but if it's "Could cgespzjvvskiaari be a word" or "Is there a way to distinguish random gibberish from language"... well, the answers are yes and no respectively. This is something that comes up in cryptography rather than linguistics.

As an illustration, do you think "vagrearg-naq-grpuabybtl" could be the representation of a language? It is in fact; it's the rot-13 form of one of the phrases in your comment. Since language is pretty much unbounded, and codes are entirely so, you can't look at a bit of gibberish and decide that it is not a code (i.e. a very weird writing system) for a bit of language.

(Languages have phonotactics and thus maybe some limitations... but writing systems really don't. E.g. here's another poser: could "gnllhc gngnlnc s t kt dlc nms sphrp r" be English? Of course it can: again, it's a bit of your text written without vowels, right to left.)

[Raphael]

I may not be getting the question,

Not so much a question, more a random train of thought that started with me seeing the sequence and thinking something like (Hey, that looks Czech!" and "Hey, that looks Finnish!"

[hwhatting]

Well, to anyone who knows Czech, it doesn't look Czech at all

[Xhin]

It definitely isn't Czech, but it's plausible with something like this:

• "cg" represents some kind of palatal, or maybe two -- "g" is used as a palatal approximant in some languages, and "c" is plausibly a palatal fricative, so maybe something like /kj/
• The next sequence doesn't really make sense unless "j" is a vowel, maybe it was a palatal approximant that became a central vowel like /3/ or /6/, though I can't imagine how the language worked before that; the etymology would be unpronounceable. Maybe instead the language is already using a lot of vowels so when it was romanized, "j" was chosen for this one, the same way Welsh used "w". In this case you'd have some kind of agglutination between the words "cgesp" and "zjvv", with "zjvv" being pronounced /Z6v/ or something. Maybe "vv" represents /b_d/ while "v" represents /v/.
• "iaa" is tricky. Maybe for whatever reason "aa" is two /a/ or /{/ with a glottal stop in the middle, in which case this makes a bit more sense, though it definitely leads to a weird looking language.

[elgis]

If anyone wants to take on this challenge and needs more fake Czech-Finnish words, I have a word generator that can generate words in fake Czech/Finnish/etc.

Fake Czech words (in IPA):

xroɲiː stro ratiːdatɛlniː ɟiːvazarospotɛːta antka mɪlovɪna zakran cɪslat ɦolt mɛkspɛktropcɛʒɪsɛːr spoʒɪt ɦɛlɲiːk ziːskuːtr̩ ʃɛstɛrodaːrɛprɛdɪktaːlɲiː vaɲɪtsɛ rɛptɪka kor̝ɪliː vɪdɛnaːvatɛr̝ɪvostudɲɛjʃiː bralovat paːtɛk matsɛn ploʃko sna bavoslavɪ zaːvo baːbavɪjɛ ɪnvoksɪviː prosɲɪt vɪɟiːʃ rajts

Fake Finnish words:

s̠e̞ri mistoɑ kɑ̝rt̪ːine̞n bɑ̝nːɑ̝ryːnærɑkemie̯lunæytːø s̠o̞ːlo̞ hæʋirhe̞ilukɑːlihyʋæs̠tiʔ ruhjelemistɑː me̞rk̟is̠tæː pulːɑ̝nɑ̝ʋit̪e̞ʔ burboi hɑ̝jo̞nɑ̝lyːs̠ilmisliːnɑntɑ rɑ̝udut̪uk̟yk̟yine̞n mo̞no̞ft̪o̞ŋːin̪t̪ɑ̝ impit̪ɑ̝ɑ̝ikɑ̝ ilmɑjoreɑkteræ ɑ̝rt̪ɑ̝ ɑstɑː pyræhdelːɑtɑ t̪e̞ht̪i luoʋilːɑːn joulutuntɑsɑʋunlɑinen turɑːntilɑliːsɑtɑ oulutunːesseulotːi muːri elæin s̠iʋis̠e̞s̠ti pe̞hmæ jɑ̝lɑ̝ilukoulut̪yt̪ːæː liːe̞mmus̠ kuine̞

I don't know much about either language, so I can't say how "authentic" these words look.

[Torco]

but the real question is: are natural languages funky enough that it is plausible for any random sequence to be at least a viable unironic orthography for a real(istic) lang? I think the answer is clearly yes, at least for stuff that is not, you know, twenty ampersands in a row. cgespzjvvskiaari is not even that weird: there are less plausible actual words in some languages, like that town that ends in gogogoch in... wales?

like, take the sequence /kʁɛʃ.psifː.skɪˈaː.ɾi/. it's funky, but not funkier than english. g is a decent choice of for R if you already have rhotics, j for i is normal, especially if it's to be distinguished from I, long v as distinguished from short v is weird, but it could come from two morphemes. so how high in consonants per vowel can you go before such an exercise is not plausible ?

[Richard W]

but the real question is: are natural languages funky enough that it is plausible for any random sequence to be at least a viable unironic orthography for a real(istic) lang? I think the answer is clearly yes, at least for stuff that is not, you know, twenty ampersands in a row. cgespzjvvskiaari is not even that weird: there are less plausible actual words in some languages, like that town that ends in gogogoch in... wales?

But I'm not sure that Llanfair­pwllgwyngyll­gogery­chwyrn­drobwll­llan­tysilio­gogo­goch really is a word, and it's on a par with the fictional Tacarembo la Tumbe del Fuego Santa Malipas Zatatecas la Junta del Sol y Cruz. And it's not so rich in consonants once you remember that 'w' and 'y' are primarily vowels, and of course 'll' is just the usual lateral in the language.

[Torco]

granted, but any similar-length subset of that string is also nearly as funky as cgespzjvvskiaari