A link to the last of these type threads, for anyone interested. I haven't worked on that project for a while, though, but FWIW my suggestions are there.
So, first: congratulations on your foray into "conlanging"; this your first one? It's reasonably complete (more complete than mine usually get), and simple enough to be picked up easily - though perhaps a bit too simple for its own good! That aside, it's an impressive amount of effort to put in, and you ought to be proud of your accomplishment.
I
do, however, have a few problems with your construction of Dwarven. These have little, I think, to do with opinion (I personally prefer more synthetic languages myself, but I'll keep such suggestions to myself); these are problems that I think weaken your language and may ultimately undo it, on a structural level.
First - and this possibly strays into the realm of opinion - your appeal to the in-game use of language to find clues as to its construction. As you found out, there aren't (m)any. This is because the
only thing the language is used for in-game is names, which (as I've said before) often follow unique conventions unrelated to the grammar/semantics of the language; if you expect to come up with a complete grammar for, say, Latin based on the names of Roman towns and rulers
alone... well, you can see that wouldn't work. This also explains the paucity of the lexicon (language_DWARF): one doesn't often find prepositions and pronouns in names, so it's no surprise that they don't show up here. We might expect an analytic language to use a
lot of pre-(or post-)positions/other function words; and as for pronouns... I don't know of any language that lacks them, certainly not 1st and 2nd person. I would suggest, if you're willing to invent a word for "dwarf," you can invent a word for "I" and "you" as well, whether or not it appears in the raws.
Second - you rely over-much on orthography to make meanings clear: capitalization of nouns, use of apostrophes to combine verb tenses, hyphens for combined numerals, and especially marking appositives with apostrophes. None of these would be viable in a real language for the simple reason that languages are
spoken: if something doesn't
sound different, it can't be distinguished from other things.
Now you might think that, since no one's going to try to
speak this language, that doesn't matter. I suppose that's true enough - but do you want to make a secret code for forumites, or an actual language that makes sense in the context of the world? Dwarves speak to each other after all - I doubt Dwarven is purely a written language.
Third and most important - you have not paid the least attention to Dwarven phonology! which is actually the easiest study to perform on the present lexicon, and one in which we can make some relevant observations, unlike in Dwarven onomastics.
You say (to my chagrin) that pronunciation doesn't matter because the language only appears as text. Leaving aside my above comments on language vs. secret code, you needn't worry: phonology is not primarily concerned with the actual sounds produced (that's phonetics), but with the
system of sounds in a language - what phonemes are understood? how do they work together? what are the constraints? These investigations ought to come before any suggestions of morphology or syntax: after all, you need to know what sounds are possible and meaningful before you can string them together. Again, we're not worried about pronunciation, which we can't discover anyway since we have no spoken samples, and no existing related languages to compare to. But we can still answer some phonological questions by turning to the lexicon...
Sparse as it is, language_DWARF actually does a good job of revealing Dwarven phonology. There are some 2,000+ lexemes listed, quite enough to make reasonably certain claims about all of Dwarven - at least in terms of what sounds appear where, and how often. As for what those sounds are... well, that's another (very interesting) question.
Let's start with some basic assumptions. We ought to assume
a priori that 1) each letter (in some cases two letters, see below) corresponds to a distinct phoneme - there are no "silent letters," and no orthographic difference is made between merely allophonic sounds; 2) one letter does not represent two distinct phonemes, like Modern English
c; 3) Modern English orthographic conventions have been applied where needed. Thus I submit that sh and th be taken as their English equivalents, as in
sheep and
thought respectively. We'll talk about ng later.
Without these assumptions we couldn't really begin, and besides they seem reasonable enough. Let's move on.
Out of ~2,000 lexemes not a single one ends in a vowel. It would seem that Dwarven consists
solely of closed syllables: we can see in the lexicon VC and CVC types, but no *V or *CV. We can take that as a phonological rule:
all syllables must end in a consonant.
Dwarven phonology consists of 17 consonants:
b t d c k g f v th s z sh m n ng* l r (*ng is of suspect value; see below)
Most common to least:
s r l t n m k g d th b sh z ng v c fPoints of interest:
1) no p to go with b. This is strange if b, d, g represent voiced stops, and t, k voiceless, as in English.
2) actually, there's no voiced pair for c, th, or sh (sh isn't so surprising: that was the case in English for a long time).
3) of voiced-voiceless pairs (again, based on English orthography) only k, g have similar distributions. s is almost 3x as prevalent as z, and t, v are ~1.5x as common as d, f.
4) what is c anyway? It's not paired with anything else, and we already have k. Maybe an affricate, like English
ch?
5) the low occurences of v, c, f seem to be partially due to the fact that these consonants never appear finally.
What about consonant clusters? It seems anything is permissible at the syllable boundary (mab-dug "ale", as-dos "bad", nok-zam "battle" etc.); but as for "true" consonant clusters, I can only find one: <st>, attested both as syllable-initial (stalkòb "boil") and final (gast "cleave") [other instances like zustash are at syllable boundary: zus-tash].
Geminates (doubled consonants) are attested, but only at the syllable boundary: they never begin or end a syllable. Gemination is possible for all consonants except c, f, v, th, and ng.
ng could be a cluster (n + g, as English "finger"), or a single sound (as English "sing"). Evidence for the latter would be that ng appears initially and finally, but there are no comparable clusters *mb, *nt, or *nk (or *nc) in those positions, as we might expect. Evidence for the former would be the relatively low frequency of ng compared to other nasals (102 instances vs. n = 524, m = 518) and the lack of an attested geminate *ngng.
Now for vowels (25 phonemes):
a â á à ä å e ê é è ë i î í ì ï o ô ó ò ö u û ú ù Most common to least (by series/base glyph):
a o i e uAssuming each of those glyphs represents a distinct (if related) vowel sound, as I've said we should (why would they mark something that wasn't significant?), that's an immense vowel inventory. The basic glyphs (without accents) seem pretty regular; what could the accents mean, though?
Unaccented glyphs vastly outnumber accented forms. We see 874 instances of unaccented
a, compared to 130 instances of
all accented a-forms
combined. For each vowel "series", the unaccented form is ~88% of the total vowels, while accented forms make up just 12% (u is an outlier, with 91% accented and 9% un-, probably due to the smaller number of variants).
As we can see, all vowels have the circumflex (â), acute (á) and grave (à) accents; a, e, i, and o have the diaresis (ä); and only a has the ring (å) (suggestively, perhaps, this averages to five forms per series - maybe å ought better to be ü, but that doesn't appear on the ASCII tileset?).
The distributions of these accent marks are fairly regular: circumflex, diaresis, and ring hover between 3-4% of total series instances (high outliers are î at 4.7%, û at 4.5%). Acute and grave hover around 2%, a little more than half as common as the other accents (á and à are low outliers at 1.8% and 1.4% resp. - possibly their proportions are driven down by the (suspect) presence of å).
What could these accents represent? I can't possibly answer that right now. That they could be tonal markers, as has been suggested, seems unlikely given how uncommon they are. More probably they represent regular alterations of the base vowels - prosodic features? length? pitch? - possibly conditioned by phonological environment or etymological origin. More study is needed, especially on what kinds of environments these glyphs appear in, but I'm not sure we'll ever know what they really represent.
What's the upshot of all this? Mostly, you should re-align your linguistic creations to match what we can glean about Dwarven phonology. We know syllables can't end in geminates, so we can't have *morr (more), *zott (hypothetical), *-akk and *-obb suffixes, etc. We only have one allowable cluster, st, so *shlis, *-arg, *sjun, *mfogz, etc., can't be Dwarven - and, unfortunately, neither can *dvergr.
NOW: none of that wall-o'-text was said (written?) to dishearten you - rather, to push you towards even more thorough investigations and compositions! If you pay more attention to the phonology, and largely ignore the very flat and un-linguistic way the language is used currently, I think you can come up with something not just fun to use, but a systematic element of a complete, coherent, and consistent world.
So about the whole word-for-dwarf situation... I stumbled on an old thread discussing the topic. Might have some ideas worth stealing.
I believe my suggestion was
arkim "race (of people)"; a lot of ethnic groups IRL just call themselves "people," so the Dwarves could do that too. And if there's no singular/plural distinction in Dwarven,
arkim can also mean just "person" - i.e. Dwarf.
The real question is: what do they call anyone else? "The tall ones" (humans)? "Green(-skins)" (goblins)? "Firewood" (elves)?
Author
Topic: The Dwarven Language Codified (Read 82140 times)