EDIT by Minion: This once started as a side track in "Worst Summaries"

Oh god, someone get this bitch the Language Construction Kit! Being a hobbyist linguist who creates his own langauges has ruined me from enjoying "languages" created by authors/artists/etc... who don't know what their doing because inevitably, they all recreate English+kh spelled worse and harder to pronounce.

"Yhō-Shétah" How is this to be pronounced? "Ee-hoh-SHAY-tah"? Are all the Hs silent? If so, if H occurs syllable-initially, is it also silent? Why in god's name do you have a stress system and a vowel-lenght system (or appear to have)? Granted, Latin did as well, but none of its daughters do, because it almost never lasts very long (lingually speaking). I would wager that Vulgar Latin had only stress, as does its daughters.

"Xehy-a-Gi" How much you wanna bet that the G in that word is, in this case, pronounced like a J, and not hard? How much you wanna bet that that X is pronounced like a Z, and not the CH in Scots "Loch"? Ooh, look, a lone-A, without an accompanying H, I wonder if it's pronounced as schwa (uh).

"Shj-Eh" Remember kids, English+Kh, spelled worse and harder to pronounce. In this case, probably impossible without unwritten interveining schwas.

Oh, look, it's your standard 1-10 (has no one ever heard of other bases for counting, even though they occur all the fucking time in languages throughout the world (you don't even have to look outside of Western Europe: check out that semi-base-20 shit that French does!). I bet speakers of this "language" have a bitch of a time distinguishing 8 and 0 in a mildly loud room. How about 1? Is that pronounced like "though" or "dough" or "duh-hoh"? Is 9 "ilk-sah" or "luk-sah" or how the fuck do you pronounce that?

She really likes her "Shé"s which appear throughout, but seem to have no consistant meaning. Given that her "langauge" definitely seems to have a set of roots, this is not a good thing.

Making a conlang (constructed langauge) is no small task, and you can find several works where it's done very well: Quenya from LotR, Klingon from Star Treck, the Atlantean language from Disney's Atlantis (it and Klingon are both by Marc Oakrand), and Navi from Avatar; and recently, Dothroki for the Game of Thrones HBO series. You can literally learn these and use them in a community of other speakers. They aren't English+Kh. They have genuinely naturalistic grammars that are both interesting to someone interested in learning them as well as interesting to any linguist who might wish to analize them.

When I see this sort of thing, this "dictionary", I can't help but think that even Zaminhoff's Esperanto is better than this.

Kawahara Reiki essenstially did this, writing fanfiction of his own SAO under a pen name, and also wrote a crossover with Accel world, where Kirito fights the pig.

 

For real, bro.

 

I have always been intrigued by linguistics and found that link to extremely interesting. Have any suggestions for more reading, whether they be books or websites?

 

http://languagelog.ldc.upenn.edu/nll/ is a good source of entertaining linguistic tidbits.

 

http://www.incatena.org/

But yeah, if you, in general, follow the advice given in the LCK, you pretty much can make a genuine language that humans could learn to speak and use in everyday life.

 

I'm a bit late but I thought this was worth its own thread and moved the posts.

Perhaps some more people are interested in further resources. I have seen authors trying to create some kind of Parseltongue (with more than just countless 's') or some ancient language, and more often than not it's weird.

But then - every time I see some Turkish newspaper it looks weird as well.

 

will return to this thread for help in constructing languages later xD

 

Oddly enough, an as-yet-unmade screenplay I had a hand in developing was originally more than half in an original language. A fully comprehensive, completely impenetrable original language.

The author stood by it throughout, and to be honest it was actually pretty impressive... until we came to changing the wording of certain lines. It was like watching a nuclear meltdown on someone's face. The language would probably have worked, but certain lines sounded better as they were, and because it was direct translation and supposedly would stand up to scrutiny, we couldn't bullshit the subtitles.

The end result? The story was transposed from a fantasy setting to a loosely-historical Balkans-area medieval one, and the dialogue that had been in the original language was rewritten in Turkish.

The moral of this story? You're going to butcher the language enough to not have to take full credit for it. I personally butchered Anglo-Norman and Old English in HPAC, I've seen more than enough authentic languages fucked around with creatively or accidentally... ultimately, when a lot of the authors in question struggle to make sense with their fucking mother tongue, they're better off steering clear of one they've 'created'.

 

I'm creating three languages at the moment for use in a fantasy novel, which I've also begun writing.

Two of the cultures in the story resemble medieval Germanic societies (i.e. the Franks and the Norse), so I've created two new languages that are basically Germanic languages, but new branches on that family tree. I know that technically in a fantasy setting there would be no Germanic language family at all, but for the sake of the reader's comfort, I'm not going to make these cultures and their languages too exotic. The story will contain many of the staples of the real world's medieval society, so I figure the corresponding languages also should not deviate too much from what could be expected of a Germanic tongue of the middle ages. Anyway, even Tolkien's Elvish languages weren't sprung completely from his imagination. He used Finnish and Latin as major sources of influence and inspiration. In the same way, I'm using Old Frankish, Old High German, Old Norse, Old Low Franconian, etc.

The third language is inspired by Avestan, Sanskrit, and Ottoman Turkish. It will belong to a powerful empire, which itself is inspired by the Achaemenid Empire, the Great Seljuq Empire, and the Ottoman Empire.

Using real languages as a broad basis for my own is working well for me. I'm even creating root word lists for each language (that is, words of Proto- forms of the languages), so that I have a clear idea of how each language would have evolved and diverged from other, related languages in the world.

I've completed two maps that use these languages: here and here. They overlap somewhat. True to real life, not all of the place names on the maps correspond to actual words in the languages. Many are derivative, altered forms, short forms, etc.

 

He's got all the stuff you need to create a language right there on the site I linked, but he's got a lot more interesting stuff in the books. The advanced book is newly published and worth it. Unfortunately, I don't believe that there's an e-pub format. Sorry.

---------- Post automerged at 15:11 ---------- Previous post was at 15:09 ----------

Looks and sounds awesome, Sechrima!

---------- Post automerged at 15:57 ---------- Previous post was at 15:11 ----------

Ultimately, I view language creation as an essentially 5 step process.

1) Decide on your basic syntax. This is counterintuitively at the beginning, but has far-reaching consequences for maintaining naturalism. Linguistic Universals (which are more like statistical corelations that occur a lot, rather than hard and fast rules) show, for example, that it's no coincidence that Objects follow Verbs in English, Relative Clauses follow Nouns (that Adjectives preceed them is anomolous, actually), English has Prepostions rather than Postpostions, and English has closed syllables (syllables that end with a consonent rather than a vowel). All of these are corelated. On the other hand, it's also no coincidence that Japanese has all the opposite features: Objects preceed Verbs, Adjectives and Relative Clauses preceed Nouns, Japanese has Postpositions, and with only two exceptions, Japanese syllables are almost universally open.

These are only just FOUR such corelations.

2) Decide on the sounds. Newblets tend to do English+Kh, and probably add meaningless or multi-meaning apostrophes. Newbies (one step above Newblets), who have learned what IPA is, tend to just throw the entire IPA into the mix, so the go from about 30 sounds to 170. I suppose that's fine if you're trying to do a Khoisan Re-lex, but not very wise. Try to keep that consonant inventory in the mid-teens to the mid-twenties.

3) Decide thow they go together. Does your conlang allow mtblatxt? Or is it like the majority of langauges on the planet that barely allow much beyond kalemitopela? English is near the complex end of the spectrum, ostensibly allowing as much as CCCVVCCC (strengths = [stɹɛŋθs]) , though in practice, most of English is limited to CV(G)C (like, hate, dead, dog = [laɪk], [hɛɪt], [dɛd], [dɒg]). Georgean has as many as eight consonants on either side of a single vowel, but again, this is fairly rare, and tends, more often than not, when it does happen, to be a result of grammatical particles stacked upon one another.

4) Decide on your grammar. Pre/Postpostions (Infixes are also not unheard of!), inflections? Stacks of particles (aggludinations), complete reliance on word order?

5) Attach meaning to roots that are made according to the decisions made in (2) and (3), and then using the grammar you made in (4), derive words. For example, with the verb "write" in English, we can derive write, written, writing, wrote, writer, writes. That's not a whole lot, but take a look at what you can do with the Arabic root KTB:

So, as you can see, you really don't need a terribly large inventory of roots, just enough to let your grammar fill the word-space.

Here's one of my favorite resources for Language creation: http://rickm.net16.net/essays.html

*G = glide: W/Y

 

Not strictly implausible. Though the following is probably a little more likely, given what you mentioned about a bunch of words from various languages being cobbled together.

What you essentially described is a pidgin. They arise naturally through contact of various kinds; c. f., the Finno-Russian pidgin of yester-century used for fur-trading and the like.

Get you sum Chinese, English, Russian, etc . . . speakers together for an extended period, and that's what you're going to get. Now, have them bring their wives along to the base and you've put the pedal to the metal in terms of Pidgin formation.

Now have these military families have kids who grow up not only learning this Pidgin, but using it to communicate with their friends. This is basically the formation of a creole. They grow up speaking the creole with each other and their mother tongues at home--so generation two is bilingual X/creole. They get married with someone from a different native background, so they only speak the creole at home and they have kids who only speak the creole, and wonder what those weird noises grandma and grandpa make are unless they're speaking a really shitty version of creole to their grandkids.

NOW, go down about five more generations so that all or almost all of the original pidgin and original creole speakers are dead, and all that's left is creole speakers, and you've got five generations of sound changes that make it virtually unintelligible to the original speakers were they to be resurrected and listened in on a conversation in the language they originated.

Read any of John McWhorter's work on creole languages and you'll get a really good sense of what they are and what they aren't.

Here's an example of a real English-based creole langauge in Papua New Guinea.

Further Reading:
http://en.wikipedia.org/wiki/John_McWhorter

http://rickm.net16.net/lexical_semantics.html -- I highly recommend this. Rick Morneau here shows that not only can you get by with an astonishingly small set of roots (He's essentially got just 100), but his conlang (detailed in Lexical Semantics), Latejami, has over 6,500 words and a very rich and versitile grammar. I just wanted to point this out because if you say "too simple" you have to really get simple -- simpler in fact than a creole. Latejami manages with just 100 roots (+113 classifiers: each root has a basic, broad meaning or set of meanings, which can be narrowed considerably by the classifier).

Anyway, good luck.