Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Loam]

FWIW = for what it's worth; conlang = constructed language, the usual term for these endeavors.

I probably got a bit carried away with my critique; linguistics is one of my few areas of semi-expertise, so I'm always excited when I have an opportunity to blather on about it. And sorry for jumping down your throat about phonology; I find Dwarven phonology fascinating (what's up with those vowels?) and I had a lot of info on-hand about it.

All third-person (or better yet, no marked person) is interesting, but if you want to keep it, I'd suggests dropping the pseudo-pronouns altogether. It just strikes me as unbelievable that a language would think it important to distinguish when the subject was the speaker/hearer, yet not have designated words for these subjects. Unless "speaker" and "hearer" were in the process of being "pronominalized" - perhaps a reduced form kut or kum would be in order? Frankly, stranger things have been known to happen with pronouns: the Old Norse indefinite pronoun nǫkkurr comes from a phrase meaning "I know not what kind": smiðr nǫkkurr "some smith/a certain smith (= I known not what smith)".

My issue on the orthography was, now that I think about it, really just with the appositives. German does, as you say, capitalize its nouns - so did English for a while, German just continued the practice; but even so that's not really a linguistic feature but an orthographic convention. However, as you point out, context serves to distinguish the meanings, and I think it's a good idea for words to be, in a certain sense, "malleable" - eshon could mean "good" or "goodness" depending on its place in the sentence, or somesuch. English does the same with many words (eg. "love"), though that's usually between nouns and verbs, not nouns and adjectives. We don't have as many substantives (adjectives used as nouns) in Modern English, but we used to - Beowulf, for instance, is often called se goda "the good (man)". We can still do this for abstract concepts ("the true, the good, and the beautiful") or collectives ("land of the free, home of the brave"), but not so much for single, concrete things.

On the issue of sentence structure generally: you might run into problems if the structure is too rigid. Even English is more fluid than people think; it sort of needs to be to be generative, especially in regard to phrases, subordinate and relative clauses, etc.
    Consider these arrangements of "The hero killed the dragon" (S=subject, V=verb, O=object):

        SVO: Hero kill dragon
        SOV: Hero dragon kill
        VSO: Kill hero dragon
        VOS: Kill dragon hero
        OSV: Dragon hero kill
        OVS: Dragon kill hero

SVO is, of course, the norm. VOS (killed the dragon, the hero did - Yoda-speak!) is understandable, if odd. SOV, I'd say, is understandable enough for an English speaker (the hero the dragon killed), but may cause confusion; at least the subject still comes first. VSO and OSV are maybe intelligible, but not as much: "killed, the hero, the dragon" sounds jumbled, while "the dragon the hero killed" is confusing. OVS is right out - "the dragon killed the hero" means the exact opposite of what we want it to.
    From this, I can see Dwarven having several different possible word-orders: SVO/SOV for the most used (these are the most common word orders IRL, across all languages, by some huge margin, like 90%), which probably means indicative, active, declarative sentences. VOS probably next most common (for emphasizing the action?); then maybe OSV in some rare constructions. VSO and OVS are probably too confusing to use without auxiliaries. That allows you 3-4 intelligible word orders, each of which could have a distinct grammatical meaning - this was the basis of my last attempt at Dwarven, so maybe I'm biased to the idea...
    This is, of course, assuming Dwarven has no case-markings whatsoever, even on pronouns (if there are pronouns). English can use potentially any one of these forms with pronouns, which mark subject or object: SOV "the hero him killed"/"he the dragon killed"; OSV "him the hero killed"/"the dragon he killed"; OVS "him killed the hero"/"the dragon killed he". These sound weird, but they are intelligible, thanks to case-marked pronouns. That's not saying you need cases or pronouns; just be aware (as you probably are) that simplicity in grammar comes at a cost of structural malleability - Dwarven poetry's gonna be hard.

I added some words of my own to Dwarven: "to be" was in, short and simple as it ought to be; "to go" was rud. Mostly some rules for word-creation would be 1) remember the phonemes/phonological rules, 2) avoid homophones, 3) try to fit the length/complexity of each word to the concept - a 12-letter word for "be" doesn't seem reasonable, since it's such an essential concept. Honestly the current lexicon is rather silly - why is there a unique word for "god-forsaken" in all four languages? - but that's what we've got.

As for the grammatical suffixes, you could probably just take off the doubled letter; but if you want to make new ones, I'd suggest giving language_DWARF a good look and maybe seeing what you can create from what's already there. IRL prepositions have some fascinating etymologies, so you can let your imagination run. Maybe, for instance, we can use a reduction of lînem "face": lîn or lîm "in front of (in space)" i.e. "facing" --> "in front of (in time)" i.e. "before"; possibly also --> "opposite" (one wall facing another), and thus --> "against" (if two people are fighting, they are usually facing each other).

[Solitarian]

I am glad that you became excited about the language; your critique is helpful, and I like encountering other linguistically minded people.

Yes, I would forget the pseudo-pronouns too. I do not recommend using them, but I realize that English speakers will want there to be some analogue with English rules, so I suggested Kutam and Fmid as possibilities. In a few of my examples I refer to the second person as reader instead of listener, so there is malleability there. I imagine that the language would eventually develop pronouns anyway (like with your example from Norse), and a word like Kutam would likely be one of the ones to become a pronoun. My codification is the language as handed to the first dwarves by the gods. What mutations come after that are none of my divine concern! The gods do seem pretty aloof, after all; all they do is curse statue-disturbers occasionally.

I really like that Dwarven words are not declined as verbs, nouns, etc. Like with my examples with Onol, there can be many fun possibilities. Urist onol-geth Mosus. Urist mountained Mosus. What does that mean? I don't know, but it probably wasn't good! I don't know what to do about appositions, though. They are currently quite awkward, albeit comfortably analogous to English appositions. What do you think could be done about them?

I thought about syntax a lot. I chose SVO because it was vaguely implied by the in-game nomenclature and because (this was the primary reason) English speakers would likely find it more comfortable. I did not want Dwarven to be so foreign that people would become dissuaded by its dissimilarity to English. I realize that most people are not linguists. Still, I understand your concern about excessive rigidity. I think SOV is actually possible, given how obvious the verb is with its tense marker and adverbs. "Urist deb-geth Shokmug" and "Urist Shokmug deb-geth" are both clear, I think, provided that the subject remains at the beginning. Still, this would face the German problem of verbs at the ends of clauses being confusing when the clauses are longer. Perhaps this abnormality could be saved for poetry. I reference German frequently because I deal with it frequently, and those cases are too much. There need not be six ways to say "the"! I gladly make Dwarven free of cases and pronouns.

I like rud as go. I'll adopt that. I think I'll keep var as be, though. I think it is more recognizable (beginning and ending with consonants, assuming the Dwarven R isn't too funky) and appears to be etymologically related to ver (become), which makes sense sematically. Yes, the complexity of a word should reflect its concept. I don't even like that my words for yes and no have two syllables, as that seems somewhat too complex for words that would be spoken so frequently, though this is far from unknown in real languages. Yes, the lexicon is rather silly. I suspect that it was somewhat arbitrarily and inconsistently produced, given that it is sometimes inconsistent and overly specific in certain areas while leaving other (far more important) concepts unmentioned. Further evidence of my suspicion is the fact that language_DWARF is alphabetical for a while, but then has many, many more words after Z that are not so organized.

I thought about doing something like that with postpositions, but I also wanted them all to begin with vowels, as they are affixed to words that end in consonants. I also wanted them to all be one or two syllables, meaning that I would have to trim other words so much that the etymology would no longer be apparent, making the effort pointless. I did take inspiration from the lexicon, though. For example, 'ist means against and is a part of Urist. We can imagine some kind of etymology there. I like your idea of developing those meanings from the word for face, but I think that would make the postpositions a bit unclear, as is sometimes true in English ("before" can mean "in front of" or "at an earlier time", so the sentence "I ran before him" is unclear and relies on context). I wanted to avoid such confusion by not combining the meanings unless they were very similar. Swedish has this problem too, as "i" and "på" have virtually identical (and very broad) meanings but are arbitrarily assigned to different (and not-so-different) contexts.

[Taffer]

In regards to the lack of vocabulary: have you seen GoblinCookie's Expanded Dictionary?

[Loam]

Okay, I've looked through your propositions again, and I have a more developed critique.

This started by trying to find a good way to do appositives. Here's my suggestion: do them like English, and just put them directly after the noun (subject or object) that they describe:
   Urist Obur vagúshgeth Måmgoz Arkimlur
   Urist hero kill-PAST dragon  dwarf-foe
   "Urist, the hero, killed the dragon, enemy of dwarves"
I'd say that's easily understandable. For clarity in writing we can put commas around them -- Urist, Obur, vagúshgeth Måmgoz, Arkimlur. -- but remember orthography =/= grammar. If something can't be understood when spoken, it shouldn't really be in the language.

This structure, however, necessitates some changes. Chiefly, unmarked dative (i.e. indirect object) won't work -- we had it following direct objects, but with the above appositive structure that would cause confusion (the hero killed the dragon for the enemy?). Similarly, unmarked genitives (possessives and "of ___" constructions) would no longer work as they do in DF names, by simply following the noun, for the same reason (Urist of the hero? The dragon of the enemy?).

My suggestion for both of these: again, do like English, and use some pre- or postpositions. You pretty much already do this with the possessive -long, so it's clearly not out of the question.

With that said, I'll submit the rest of my critique, following the order of your OP:

Nouns:
   Subject and (Direct) Object will be easy to identify in SVO order if the verb is always clearly marked, as currently it is (with tense, mood etc.). See the above example. SOV might also work, haven't tried it much.
   I already mentioned my genitive/dative suggestions.
   No number (singular/plural) is fine, can be marked by adjectives (many-, few-, etc.) or ignored.
   Re: nouns from verbs, I feel that agents and verbals (teach-er vs. teach-ing) should be distinguished (although language_words disagrees with me here), just for clarity's sake.


Adjectives:
   Names show adjectives are agglutinated to nouns; the head noun should always be the right-most element:
      [ADJ]+[ADJ]+[ADJ]+[HEAD NOUN]
   names also show that nouns can be appended like adjectives:
      libash-emar "axe-animal"
      this is probably a good way to make "genitives"
      such nouns go after all the adjectives; so, a Dwarven compound looks like:
         [ADJ]...+[NOUN]+[HEAD NOUN]
      datan-dastot "iron-sword" > "an iron sword"
      emen-datan-dastot "strong iron sword"
      emen-datan-råsh-dastot "strong iron death-sword"
   Problems:
      compounding could make it hard to use comparative forms marked with prefixes:
         tel-emen-nashon-dastot > "very strong" or "very bloody"?
      unless:
         emen-tel-nashon-dastot > "strong, very bloody sword"
   Adjectives probably follow an implicit order, like English: we don't say *"the green big tree" but "the big green tree". We aren't taught this order but we know it instinctively. I'll have to look up what it is.


Verbs:
   Marking tense will clearly mark verbs, which helps intelligibility with our strict word-order;
      if S and O are always separated by V+tense we can tell them apart
   Verbal babin ought to be "be friends with" rather than "befriend" - stative rather than transitive
   In your suggestion, the verb-phrase is like:
      [ADV]...+[HEAD VERB]+[TENSE]+[MOOD]
      this could work, but it might be better to keep the head verb in a designated position, like nouns are: left-most or right-most
         I say left-most: verb always begins the verb-phrase
         in which case, more important info should be closer to the verb; as you go outward it gets less "essential"
            so perhaps VERB-tense-mood-etc... (adv?)
   Adverbs are a notoriously difficult to classify: they're not really the "adjectives of verbs"; some adjust entire phrases rather than just the verb... It may be better to keep adverbs separate, since verbs have so much else going on.
      Adverbs also tend to be rather simple compared to adjectives
      I think using a suffix, like -ak, is simple and effective (if a bit boring)
   Passive Voice:
      could just be like English: to be + past participle
         varudiz deb(geth?)
         how are past participles formed? (predicate) adjectives...?
      but there's other options: maybe a verb "slot"? like:
         Shokmug debvergeth
         cheese  eat-PASS-PAST
         "Cheese was eaten"
         
         Shokmug debvergeth esmulak
         cheese eat-PASS-PAST full-ADV
         "The cheese was completely eaten"


Possessives: another problem with genitive constructions
   obviously Uristshokmug (Urist-cheese) is confusing: how do we know its not "dagger-cheese" (sharp cheddar)?
   -long could work, but I might put it on the possessor, not possessed:
      Shokmug Uristlong debvergeth
      "Urist's cheese was eaten"
      
   but some languages (Hebrew for one, I think) have a "construct state" that the possessed noun is put in, while the possessor is unmarked:
      Queen [of Sheba] / [Sheba's] Queen vs. [Queen of] Sheba / Sheba['s Queen]
   Something like this could yield:
      Mosus debgeth Shokmuglong Urist
      "Mosus ate Urist's cheese"
   Which is close to your suggestion. Your way is probably fine.
   But I think we need to address the broader question of how to mark genitives generally. Pre/postpositions are probably the best bet.


Question Words: let's wait on these - questions are wierd, and these involve pronouns


Numbers:
   The numerals are fine
   For numbers over 10:
      10+# = -teen: zeznir 11, zeznob 12 etc.
      #+10 = nobzez 20, mezzez 30, vorzez 40 etc.
      this is just like you have it, but without - and i (again, I'm trying to reduce purely orthographic distinctions).
      by this token, zezzez 100 (but maybe we need a special word for this)
      and so on...
   this, however, could cause confusion with compounding:
      Nob-zez-nir-ustos-arkim
      two-ten-one-angry-dwarf
      "Twenty-one angry dwarves" - I guess it's clear enough
      #s should always be the first adjective(s) in a compound
      
   ordinals: just -t will create unacceptable clusters (*nirt, *zezt), but something like -it could work; perhaps the -i- could be "fleeting" and disappear if not needed, as in:
      Nirtonol "first mountain" - -onol takes the -t- to make correct syllable onset
      but Niritdastot "first sword" - -dastot refuses -t- as syllable onset, so -i- needed
   that's probably too confusing, though


Particles and Conjunctions:
   only thing I'd say here is: these don't need endings to "make them particles/conjunctions" - they should be their own semantic category
   (most should be monosyllabic as well)


Example Sentence: deliberately convoluted, trying to stress-test:

urist arkimobur telemendatandastotesh nogleshak vagúshudizzilir nobzeznirustosomerlur kodoravor

Urist dwarf-hero most-strong-iron-sword-with savage-ADV kill-PRES-"must" two-ten-one-angry-green-foe dawn-before

"Urist, the dwarven hero, with his very strong iron sword savagely must kill twenty-one angry green enemies before dawn."

(NB: I used -zilir imperative as a "must" mood, designating obligation)
Note the lack of orthographic markings to distinguish nouns, verb tense/mood, postpositions, etc. I think it's clear enough without these, which is good. But they can easily be put back in for extra (written) clarity:

Urist Arkimober Telemendatandastot'esh nogleshak vagúsh-udiz'zilir Nobzeznirustosomerlur kodoravor.

Doesn't look (or sound) too bad.

Now, what we've got here is almost an English-cipher; I think there's much more we could do to make Dwarven more unique without straying into the realm of "too weird for non-linguists to want to learn". Mostly stuff with verbs; then there's idiomatic constructions to account for, etc. But this is good so far. What'd'ye think?

[Sanctume]

I speak Tagalog which is a dialect in the Philippines. 
My best friend wanted to practice speaking, and asked about how verbs are transformed.

Then without really knowing how I knew, I deciphered something in the language at least when it came to verbs.

Example the word "eat" (English) | ka-in (Tagalog, 2 syllables without the dash) | debgeth (Dwarf)

Eat. | Kain. | Debgeth
English: "You eat."  or "Eat." - a command to act, implied to who the speaker is speaking to (you).
Tagalog: "Kain." or "Kain ka." - ka is word for you.
Dwarf: "Debgeth." or "You degbeth."

Past tense, I observe was adding "um" between first letter and second letter.
Ate | Kumain | Dumebgeth
English: "I ate."
Tagalog: "Kumain ako." - ago is work for me.
Dwarf: "I Dumebgeth."

Future tense, I observe was prefixing the first syllable to the word.
I will eat.
Kakain ako.
I deddebgeth.

Present tense, I observe was prefixing with (the first syllable twice with "um" between first letter and second letter.)
I am eating.
Kumakain ako.
I am dumeddebgeth.

[Solitarian]

I note that some of your suggestions are already in the codification.

Appositions can be like surnames! That occurred to me as I was reading your suggestion. John Smith = John the smith, John who is a smith. I like where you put it syntactically, but I think an orthographic distinction is necessary, otherwise syntax becomes confusing. Commas or apostrophes are fine. The way I originally codified it is ok, then. Besides, I think these could be marked in speech with a pause and tonal shift. That seems reasonable.
I think you example was a bit needlessly confusing, though. Arkimlur could mean dwarven foe, as "arkim" could be taken as an adjective of "lur". I would say either 1) Lur-longArkim (foe belonging to dwarves) or 2) Lur, Arkim'ist, (foe against dwarves). I agree that this could be marked with postpositions, and it can be so marked with the ones I supplied.

I think an unmarked dative object is fine, as it actually is marked with postpositions most of the time. That would be necessary anyway, and people would want to do that to clarify meaning.

The problem with agents seems unsolvable, given how inconsistent language_DWARF and language_words are. Some of the words defy codification. Perhaps we could treat those as exceptions to the rule. Perhaps we could mark agents with an -ark ending as a shortened form of arkim. So Amalark (teach person) is teacher. Does that sound good?

Yes, adjective order is important, as I noted. The first adjective is the most important, and the last adjective is the least important. "tel" is placed before whatever it modifies. I think the verb position needs to stay as I have it, as it becomes too confusing otherwise. Sometimes adverbs are just the adjectives of verbs, so they should be treated like the adjectives of nouns. I think the - before the tense marker would be pronounced as a pause or glottal stop, so orthographically representing it is sensible. Adverbs that aren't merely the adjectives of verbs can either go into that "extra information" area at the end of the clause or be agglutinated onto something if that is really necessary.

The passive construction needs to remain the same as the adverb construction, so verdeb-geth needs to be correct, not debver-geth. "Debver" would be eat-become, the become of eating. To me that implies that eating causes some kind of transformation. When I am hungry, I can use pizza to eat-become full.

Meh, babin can be either, depending on context. It is vague anyway. Like with Onol as a verb, context is necessary to understand what it is supposed to mean.

-long being on the possessor or possessed makes no difference, I think, given that they remain together in the word and the syntax is unaffected. Shokmug-longUrist and Shokmug-Uristlong are both intelligible and almost the same, so I don't think it matters. The postposition 'ar means of / from, so it and -long can be used to mark possession / genitive things.

The -i- in numbers is not purely orthographic; it is pronounced. Furthermore, the - is necessary to distinguish powers of ten. Each - is a step down. Without it, reading numbers becomes very confusing, as there is just a big jumble of nobzeznirzeznobzeznobfim. Also, there is a word for 100: zezbog! I made it for exactly that reason. Why are nirt and zezt unacceptable? Why can there be to t before dastot? Just enunciate; finish saying the t, then say the d. I think the variable -i- is definitely too confusing. Swedish does something like that with its numbers: trettio = 30, trettioett = 31. In trettio, the o is pronounced, while in 31 it is not pronounced... sometimes. It is inconsistent and confusing.

I think you're right: conjunctions should be monosyllabic. Mine are too long, and I realized that they can never be agglutinated, so their status as tiny words by themselves would make them obvious enough. I will change them. I like eshob and asdob, though. I think eshob is easy to say and is believable as a yes word. I imagine that either the s or d in asdob would disappear at some point.

Also, tel is a superlative marker, so telustos does not mean "very angry", but rather "angriest / most angry". I do not like your deliverately difficult stress-test because it breaks the syntax. All the extra information belongs at the end of the sentence, after the object. Also, imperative forms necessarily can only be second-person, so using it as "must" seems weird to me. I think must should be its own word. My "havt" is arbitrary, though. I think we could develop something better than that.

To Sanctume:

I don't like that suggestion because it is declination, which Dwarven doesn't do. Also, Dwarven words never end in vowels, so that system would produce many awkward consonant clusters. With the current agglutinative system, your Tagalog suggestion would make verbs very difficult to read.

[Loam]

I note that some of your suggestions are already in the codification.

Yeah, they were really just notes I was taking as I read, so I was just hashing some things out.

I think you example was a bit needlessly confusing, though. Arkimlur could mean dwarven foe, as "arkim" could be taken as an adjective of "lur". I would say either 1) Lur-longArkim (foe belonging to dwarves) or 2) Lur, Arkim'ist, (foe against dwarves). I agree that this could be marked with postpositions, and it can be so marked with the ones I supplied.

You're spot-on about Arkimlur, thanks for catching that. It ought to mean "a foe who is a dwarf," in fact; Arkimetar "dwarf-king" = "king who is a dwarf (or at least rules dwarves...?)". That just goes to show why a well-marked genitive is important. Your suggestions are both good.

I think an unmarked dative object is fine, as it actually is marked with postpositions most of the time. That would be necessary anyway, and people would want to do that to clarify meaning.

Well, yeah, it would be marked (with prepositions), and therefore not "unmarked." It doesn't need a case-ending, if that's what you mean; it just needs to be differentiated from DO or S by something other than position (if we do appositives in the above way). So this is fine.

The problem with agents seems unsolvable, given how inconsistent language_DWARF and language_words are. Some of the words defy codification. Perhaps we could treat those as exceptions to the rule. Perhaps we could mark agents with an -ark ending as a shortened form of arkim. So Amalark (teach person) is teacher. Does that sound good?

This is the reason I choose to ignore the way words are actually used in-game, because it just makes no sense. But yeah, exceptions to a rule sounds good. Marking agents is fine, but (and I'm gonna get phonological again) we can't use -ark because -rk is not an allowed cluster in final position (remember, attested Dwarven only allows -st). -ar could work (and sounds like English -er), but we might take some inspiration from the lexicon here. Note that "hammer" is nil, while "hammer-er" is ùnil - does this suggest Dwarven uses a prefixed ù to mark agents (maybe ùn if the noun begins in a vowel)? On the other hand, since "hammerer" has strong cultural connotations to Dwarves, perhaps this is only a survival of an old, no longer productive agentive prefix?

I think the verb position needs to stay as I have it, as it becomes too confusing otherwise. Sometimes adverbs are just the adjectives of verbs, so they should be treated like the adjectives of nouns. I think the - before the tense marker would be pronounced as a pause or glottal stop, so orthographically representing it is sensible. Adverbs that aren't merely the adjectives of verbs can either go into that "extra information" area at the end of the clause or be agglutinated onto something if that is really necessary.

Whether or not adverbs are "adjectives of verbs," verbs are not nouns: there is no morphological reason that verbs should be modified in the same way as nouns, and in fact (in my opinion) there is morphological reason to modify them in a different way - to help distinguish verbs from nouns. But, it's not really a big deal. We can keep them on the front of the verb.
I have thoughts about the "other information" area that I'll give below.

The passive construction needs to remain the same as the adverb construction, so verdeb-geth needs to be correct, not debver-geth.

I disagree. Passive voice is not adverbial: "becomely eat" does not mean the same as "become eaten." Why not use another verb "slot" like I suggested: we already have tense and mood, why not voice?
    [ADV]... + [VERB] + [TENSE] + [VOICE] + [MOOD]  or something like that.
You could mark these with orthography if you want, but I think the meaning is clear even without that.

"Debver" would be eat-become, the become of eating. To me that implies that eating causes some kind of transformation. When I am hungry, I can use pizza to eat-become full.

I think you misunderstand what I'm doing. I'm proposing different systems than you did; in this case I was 1) not appending adverbs to the front of verbs, 2) using -ver- solely as a passive-voice marker. There would be no confusion; saying there would be confusion in your system is beside the point, because I'm not talking about that.

-long being on the possessor or possessed makes no difference, I think, given that they remain together in the word and the syntax is unaffected. Shokmug-longUrist and Shokmug-Uristlong are both intelligible and almost the same, so I don't think it matters. The postposition 'ar means of / from, so it and -long can be used to mark possession / genitive things.

All good.

The -i- in numbers is not purely orthographic; it is pronounced. Furthermore, the - is necessary to distinguish powers of ten. Each - is a step down. Without it, reading numbers becomes very confusing, as there is just a big jumble of nobzeznirzeznobzeznobfim. Also, there is a word for 100: zezbog! I made it for exactly that reason.

Yeah, I knew it was pronounced, I just wanted to see if we could do without it. It's fine, though; I just wouldn't say it expresses "multiplication" just because I don't think languages work like that - would mezivor mean twelve? It seems better to use it just to mark multiples of ten: nobizez, mezizez, vorizez, etc.
I just wanted to avoid orthographic markers if at all possible, for the reason (which I'm sure you're tired of hearing) that, IMO, things need to be understandable by sound, not sight. Orthography can help clarify written things, but it can't be the sole source of distinction. I think the hyphens could (and probably should) be kept in, as they do help clarify, and in spoken Dwarven there'd still be little/no confusion: we readily say "one thousand two hundred fifty three" and understand it well.
I meant a word for 100 that wasn't related to the word for 10 (zez), like "hundred" in English. Just for clarity's sake.

Why are nirt and zezt unacceptable? Why can there be to t before dastot? Just enunciate; finish saying the t, then say the d. I think the variable -i- is definitely too confusing. Swedish does something like that with its numbers: trettio = 30, trettioett = 31. In trettio, the o is pronounced, while in 31 it is not pronounced... sometimes. It is inconsistent and confusing.

Phonological reasons, again. Nirtdastot produces a three-consonant cluster, -rtd-, and we know that the only three-consonant clusters Dwarven allows are -st+C- or -C+st-, because -st is the only true cluster allowed. Cf. selsten "bother" and thestkig "conflagration" (Unless we count ng as a cluster: shungmag "disgust"). Yes, you and I can enunciate -rtd-, but it would seem that the Dwarves can't.
And yes, the variable -i- is confusing; I only suggested it for fun.

Also, tel is a superlative marker, so telustos does not mean "very angry", but rather "angriest / most angry".

Superlatives can mean "very" sometimes, like in Latin, especially if they're not used in comparisons. Besides, I didn't want to come up with another word for "very" on the fly. I think it works fine.

I do not like your deliverately difficult stress-test because it breaks the syntax. All the extra information belongs at the end of the sentence, after the object.

Again, I wasn't using your syntax; in fact I was trying to show that you didn't need all the "extra information" to come at the end in order for the sentence to be understandable. I think it's extremely useful to have parts of the sentence - prepositional phrases, phrases of time, etc. - appear in different places depending on emphasis or somesuch. In my suggestion Telemendatandastot'esh "with a very strong iron sword" and kodor'avor "before dawn" could be shuffled around to emphasize different aspects:

     Kodor'avor Urist vagúsh-geth Omerlur Telemendatandastot'esh.
     Here the emphasis is on when the action occurred ("before dawn").

     Telemendatandastot'esh Urist vagúsh-geth Omerlur Kodor'avor.
     And here, on the manner or by what means the action occurred ("with a very strong iron sword").

     Kodor'avor Telemendatandastot'esh Urist vagúsh-geth Omerlur.
     And here, on both.

All of these example, I submit, are just as understandable as Urist vagúsh-udiz Omerlur Kodor'avor Telemendatandastot'esh, following your syntax, because the prepositional phrases are clearly marked (with 'avor and 'esh); they cannot be confused with the subject (Urist) or object (Omerlur), which are unmarked; and these cannot be confused with each other, so long as they stay on either side of the (marked) verb. Appositives, being likewise unmarked, would also not cause confusion.

Again, I stress, I was not using your syntax; I was trying to demonstrate a different (and I think better) way that sentences could be arranged without causing confusion. This really doesn't differ from your suggestion all that much: I have merely used your post-positions to free up parts of the syntax I found needlessly restrictive. As long as something's function is marked morphologically, there's no reason for it to be marked syntactically as well.

Now, one might say that "proper" Dwarven prefers such "extra info" at the end of a sentence; that's okay. But it doesn't need to be there.

Also, imperative forms necessarily can only be second-person, so using it as "must" seems weird to me. I think must should be its own word. My "havt" is arbitrary, though. I think we could develop something better than that.

Again, I was adjusting your suggestions to meet my own needs. It occurred to me that, rather than an imperative "form," Dwarven could have a form expressing necessity or obligation - hence "must" - that, when used with an unstated "you" could function as an imperative:
     Deb-zilir Shokmug! "Eat cheese!"
With no other stated subject, this sentence wouldn't be confused with any other form.

However, it's not important. I was just making an off-the-cuff example, and rather than come up with a new word for "must" I just used your imperative. I'm not attached to the above usage, I just thought it up on the spur of the moment. Now that I think about it it has some issues, so probably a distinct imperative is in order.

(and havt couldn't work for, again, phonological reasons; plus, /h/ isn't in the Dwarven consonant inventory. kast might do.)


@Sanctume: while I love infixes myself, they don't work if we take Dwarven to be an agglutinating, mostly analytic language. Also, most forumites (presumably English-speakers) would find them strange, and the idea was to keep Dwarven approachable for everyone. A good suggestion nonetheless.

[Solitarian]

I'm glad that we agree on so much. I don't think I need to respond to a lot of it, given that I agree.

I still just don't understand the consonant cluster thing, but I trust that you know what you mean. At the boundaries between the components of compound words can be almost any combination, so Dwarves could clearly pronounce these clusters. There are plenty of funky in-game names, after all. The "ark" in arkim can be pronounced at the beginning, so why not at the end? I don't mean to say that you are wrong; I mean that I don't understand. The ún prefix seems ok, though. However, I think many people would not know how to type ú.

Ah, I see what you are saying about the passive voice now. Yes, using ver like a tense marker makes more sense to me. Let's do it that way. Urist deb-geth'ver Shokmug. Urist became eaten by cheese. Perfect! I did not realize you were using syntax outside of my rules. Now that you mention it, I suppose these other constructions make sense, and I like your idea that they would be possible, but not "proper". Much like how "The dog a ball I gave" is intelligible but strange. Such constructions would be reserved for poetry or orators, I think. Changing word order for emphasis seems like something a poet would do anyway, even if the syntax were rigid.

I guess we could make a word not based on ten for "hundred". After all, the English word hundred has a pretty funky etymology. Suggestions?

Sure, let's remove havt and make some other way of expressing "must". I feel like we could either make it its own verb or a tense marker. Maybe all the modal verbs could be tense markers instead of verbs, though I think that might be too confusing.

[Loam]

Okay, I'll try to explain the consonant clusters in more depth:

Dwarven syllables are of two types: VC and CVC, where C and V represent a single consonant/vowel each. We can see these in monosyllabic words: ím "bread", cog "boot".
The only time we find monosyllables that begin or end with two consonants, those two consonants are st: stal "bust", ùst "breach". Thus, we can say that C can also be st, but not any other cluster - because no others are found in the lexicon (again, this depends how you analyze ng; I take it as a single consonant).

Of words with two syllables we find these types: VC-VC, VC-CVC, CVC-VC, and CVC-CVC. These are all the possible combinations of the two attested syllable types. Thus ok-on "burden", eb-sas "candy", sal-ul "skin", min-baz "enchant". Note that sal-ul, not *sa-lul, because *CV is unattested.
Now, at first glance a word like minbaz may seem to contain a cluster: -nb-. But look at the syllables: min-baz = CVC-CVC. This kind of clustering happens only at the syllable boundary. You can look through the lexicon and see for yourself. The ark- of arkim is the same way: ar-kim.
So, Dwarven allows any cluster (of two consonants) at syllable boundaries, but nowhere else. As we said, st can be used for C in either of our syllable types: so we can even have est-stek "flea" with four consonants in a row, but that's as much as we get.

The position of a cluster is important. Think about English: a word like expose contains a medial -xp- cluster. And we can pronounce that no problem. But we don't have words like *xpas or *laxp; and maybe we can pronounce these, but we find it hard, hard enough to keep such clusters, in initial or final position, out of our language. French and Spanish really didn't like initial st-, sp-, or sc-, hence words like escuela for "school", Esteban for "Stephen", and our borrowed especial for "special". The e- was added to change the syllable boundaries, making these clusters easier to pronounce. And other languages are extremely intolerant of clusters: in Japanese the only syllable types are CV or CVn; you might get a -nts- cluster medially, but no Japanese word starts or ends with that.

So yes, Dwarves can pronounce -rk-, and many, many other clusters, but only at the syllable boundary. In initial or final position -- that is, wherever there's a C in the above syllable diagrams -- only one cluster is allowed: st. This can all be observed in the lexicon, I made none of this up. It seems weird to us, but we're not Dwarves.

[Solitarian]

I never would have noticed any of this, given my lack of knowledge about phonetics. I recognize that languages have patterns, and I wil try to heed the Dwarven patterns when making words. Still, there are many words in real languages that don't match the patterns. German has so very many anglicisms that do not follow German phonetic patterns, yet are still in the langauge. The "th" sound does not exist in German words, yet it is still spoken in German. I think phonetic boundaries are more "soft" restrictions than "hard" restrictions, in that they can be exceeded, though it is recognizably foreign. Anyway, I'll think about this when making words, though I probably will still accidentally break the rules occasionally.

[Loimulohi]

Huh. I found a lot of Swedish but I didn't find the Finnish despite it being my mother tongue. That's weird

[Miuramir]

...
The only time we find monosyllables that begin or end with two consonants, those two consonants are st: stal "bust", ùst "breach". Thus, we can say that C can also be st, but not any other cluster - because no others are found in the lexicon (again, this depends how you analyze ng; I take it as a single consonant).

Again, not an expert, but does that imply that in some form of Dwarven, "st" was / is a ligature, digraph, or additional character? 

[Solitarian]

If we assume that every sound has its own letter, but we don't really know if that's the case.

[Loam]

Again, not an expert, but does that imply that in some form of Dwarven, "st" was / is a ligature, digraph, or additional character? 

Not necessarily, it's just a phonemic cluster: it's not actually one sound. It's not even an affricate, like ts or ch (tsh), but rather two distinct sounds pronounced together, just like it is in English.
My explanation above was extempore and not very accurate: technically we should say Dwarven allows a syllable (C)(C)VC(C) -- but, since the only onset/coda (i.e. not at the syllable boundary) cluster allowed is <st>, it seemed easier to paraphrase, to avoid confusing anyone.

As far as orthography: we can't know for certain, having no samples, but it does seem odd that Dwarven allows just this one cluster. Perhaps it had some phonological significance, enough to be written with a single character. But that's all conjecture.

Another question is: how do we analyze <st> when it appears at the syllable boundary? E.g. in dastot "sword" -- is that dast-ot (with cluster) or das-tot (without)? This affects our understanding of the frequency of individual phonemes.

[therahedwig]

I am pretty sure the only reason the languages are agglutinative right now is because it is the easiest to code, so just keep that in mind when trying to decipher/construct an extension for Dwarven language.