The Logodaidalos

'There is no great invention, from fire to flying, which has not been hailed as an insult to some god.'

[logodaidalos]

Wow. It's been years, hasn't it.

[logodaidalos]

Morphology -- Nouns (I)

Unlike verbs, nouns embody a more static approach to thought expression in a language: this dichotomy permeates most layers of morphology of both verbs and nouns, and introduces several additional details.

Nouns are, for the most part, references to verb arguments that further specify what the argument is. Trollian verbs may stand alone (as in a previous post, with <krazda> "I see the two of them/you"), whereas nouns mostly serve as complements to verbs.

To facilitate easier communication, information may end up encoded in many ways such as word order and morphological marking on words. As the general feel of Trollian I'm aiming at is very strongly inflectional and is mostly word-order-agnostic, it implies that nouns would get about as marked as verbs.

Trollian nouns receive marking for case and number, but don't differentiate between noun genders or classes on their own. They're built upon the noun root, usually taking the form of one or two syllables, extended by derivational affixes into stems, that change the semantic meaning of the root appropriately, and only then inflected for case, number and whatever else.

This utilisation of derivational affixes -- probably extensive due to the polysynthetic nature of the language -- would also mean that noun and verb roots would overlap significantly, as both noun and verb roots could thus end up being extended into a verb or a noun, respectively. To pre-emptively name this process, I'll introduce the terms cross-derivation and direct derivation, as well as their corresponding verbs and adjectives. Cross-derivation would be the process of deriving a verb from a noun root, and direct derivation its inverse.

Even though most nouns would take derivational suffixes to derive nuances in meaning, some would of course have to be basic and underived; body parts would, for the most part, be in this category. For illustrative purposes, I'll introduce some such bodypart terms:

<i:j> - horn

<sak> - finger

<k'ulo> - hair

To preserve stem shape homogeny (as stems will all end in vowels), roots that end in consonants but do not get specialised derivation get an anaptyxic vowel, usually <-a/o->, but can also be an echo vowel, repeating the last vowel in the root sans length.

Without actually introducing the suffixes, I want to point out that there's going to be a difference between noun and verb suffix orders in that nouns will have a rigidly fixed order of suffixes: the number suffix will always precede the case markings.

—Nexvs

[logodaidalos]

Morphology -- Verbs (III) [Person Agreement II]

Having gone through with the basics of morphosyntax and did some preliminary work on person agreement with the absolutive singular, we've come to the point where we can express some rudimentary concepts. It doesn't take too much thought to come to the conclusion that that's basically practically useless!

For this we need to introduce something other than the singular absolutive, of course, and that necessitates at the very least the rest of the absolutive and agreement with the ergative.

As for grammatical number, in addition to English's singular and plural, Trollian also has a dual, representing exactly two instances.

As for the absolutive, the morphemes are:

Dual

First person: <-sa->

Second/third person: <-da->

Plural

First person: <-ʐa->

Second/third person: <-ma->

The second and third persons dual and plural are marked with the same morpheme. If we want to disambiguate, we'll have to introduce some information outside the verb.

As for the ergative, I tentatively propose the following singular-plural morpheme pairs:

First person: <-s-/-v->

Second person: <-k-/-i->

Third person close: <-n->

Third person far: <-r->

It distinguishes only the singular from the plural (anything twofold is plural in the ergative), but doesn't distinguish number in the third person.

A bit of toying with the new morphemes gives verbs such as <ʂriʐa> "you (pl) eat us" and <krazda> (from *krasda) "I see the two of you/them".

—Nexvs

[logodaidalos]

Nasal allophony disambiguation

It's come to my attention that in the last post I used a phoneme that doesn't actually exist in the language. To this end, to prevent deletion, and since I believe the negative morpheme is pretty good with <-ɲ->, I'm proposing a pair of complementary allophonic shifts involving <n>. The first is the merger "nj > ɲ" that, among other potential things, produces the negative morpheme, and the second is "n > ɳ / _j" that functions as a disambiguative change.

Since both of these are optional, there can be many variants of a sequence of two morphemes. I am thinking that the merger into the palatal nasal effectively be commonplace (if not obligatory) with the negative morpheme, and the disambiguative change would apply when there isn't a negative morpheme in a word but a <-nj-> sequence still happens. If the word contains a negative morpheme, the rest of the elements of the word that contain <-nj-> can then be free to merge into a <-ɲ->.

—Nexvs

[logodaidalos]

Morphology -- Verbs (II) [Negation I]

It's all good to know who did what and to whom, but one very important element of language is to distinguish events that happen from those that do not. For this we introduce polarity, basically the category of "yes/no". English deals with this primarily by sticking "not" somewhere into the sentence, and some languages have truly baroque ways of expressing the same concept. In the spirit of Trollian we introduce some verb inflections to handle just that.

The first, and the only covered here, would be the general negator of the verb. For this I propose <-ɲ-> which would come in some phonetically "easiest" place in the verb, but in general pretty free to move around. As the palatal nasal is a bit messy in word-final and pre-consonantal position, the negative morpheme could also optionally acquire an anaptyctic <-a> as a bit of phonological grease so to say.

This would give some verb forms such as <k'anaɲa> "s/he doesn't shout". Oddly familiar, hey! Just transcribe it as English would and get somebody's name *cough cough*.

That aside, the morpheme doesn't seem to be causing any significant issues.

—Nexvs

[logodaidalos]

Morphophonology (I)

Morphophonology is basically just alternations of phones or phonemes based on morphological patterns. In Trollian, they would probably be something along the line of cluster management.

They can be divided into two groups: assimilatory or dissimilatory.

Assimilatory morphophonological processes deal with morphemes becoming more similar in pronunciation, or in the extreme merging into one phoneme.

The most common of these is definitely nasal assimilation, where nasals change their place of articulation based on a neighbouring consonant. In the case of Trollian, these would be both progressive (left-to-right) and regressive (right-to-left) assimilations. The processes would be:

n > m -- next to a labial non-nasal consonant (*npa > mpa; but mna > mna)

n > ɳ -- next to a retroflex non-nasal, or /r/ (*rna > *rɳa)

{m n} > ŋ -- next to a velar non-nasal (*mka *nka > ŋka)

Another assimilatory process is the merger of a denti-alveolar or retroflex stop and fricative sequence into an affricate. This affricate would have the place of articulation, voicing and ejectivity of the fricative (so that *ɖs'a > ts'a). For this reason, a tie-bar isn't obligatory and affricates can simply be written as a sequence of a plosive an a fricative.

Dissimilatory processes make consonants more different from one another. The most prominent one in Trollian is the transformation of /r n ɳ/ into fricatives. The simplest shift is of: *rra > sra. When either of the two nasals come together, the first shifts into a fricative that matches the place of articulation of the second nasal (f.e. *nɳa > ʂna) and when they precede /ŋ/ both shift to /s/ (f.e. *ɳŋa > sŋa). Several other minor dissimilatory shifts are also extant, such as that of: *ɭɭ > rɭ.

Another important process is that of obstruent voicing changes: all obstruent sequences must agree in voicing with the last member of the sequence. To this end they usually either acquire its voicing, or receive an anaptyctic vowel which serves to break up the sequence.

—Nexvs

[logodaidalos]

Morphology - Verbs (I)

So, verbs! They're probably going to be the most fun bit of Trollian (temp. name), and where we'll spend the most time.

(For reference, from now on I will be using a simplified transcription scheme, where /ɑ(:)/ will be represented by <a(:)> and /ə(:)/ will be represented by <o(:)>)

Trollian verbs will mostly be built out of either general stems or bare roots. Roots in general will be very verb-y in nature, and verbs would be the primary thing derived from them. Nouns are additional derivations or just re-interpreted verbs.

Some elementary verbs would probably be minimal in size, expressing some rudimentary concepts. A suggested few, for illustrative purposes:

<ʂr> - eat

<ɖʐ> - please

<k'> - shout

<kr> - see

Some other, more complex roots, would embody more complex concepts, such as:

<ne> - accept

<sa:xt> - write

In general, roots cannot stand alone as a verb. Due to grammatical limitations as to what constitutes a sentence, a verb must at least be inflected for the absolutive. I'll tentatively propose the following series of morphemes:

<-ja-> - first person singular absolutive

<-ra-> - second person singular absolutive

<-na-> - third person singular absolutive

Since Trollian verbs are going to become massively inflected, I've been considering a free-er order of affixes, something resembling Greenlandic but not quite like it: affixes do not have a fixed position in the verb and instead are free to move around, causing lots of allomorphy.

Giving a freer order to affixes in the verb gives us travesties such as <ʂrra> and <sa:xtna>, which do not feel like they belong in a language I'm looking to make smooth and flowing. We need to introduce some restrictions and processes.

Primarily, I'm thinking about anaptyctic vowels i.e. vowels inserted to break up unwieldy consonant clusters. As all the vowels in that chart are phonemic (except for the [ə:] which I'll have to elaborate upon later), it is basically a choice of using any of the short vowels. Instead of going with <-o->, I'm thinking of using <-a-> as it's a fairly neutral and pleasant vowel.

Inserting the anaptyctic vowel would be also variable. One could insert it in any desired location, though some clusters might obviously be obligatorily broken up. As it's a morphologically null-valued affix (meaningless morpheme), it could also potentially be infixed into root. This would give verbs such as <sa:xatna> or <sa:xtana>, and <ʂrara>.

A limitation I'd propose is to make all ejectives always come before vowels. This would mean that a hypothetical "he is shouted" would have to be translated as <k'ara> or <rak'a> and definitely neither <k'ra> or <rak'> as it would go against the limitation. Might be arbitrary, but word-final ejectives don't feel good to me.

Another limitation would probably be that clusters with ascending sonorancy (such as <ʂr>, with a low-sonorancy fricative followed by a high-sonorancy approximant) would definitely be disallowed from being syllable codas. This would mean that sonorancy decreases as one moves farther away from the vowel, and it would also disallow a verb such as <jaʂr> because it would have a bad sonorancy order at a coda, and would instead force it to be something like <jaʂra> or <ʂraja>, or even <ʂarja>. 

I think this free-er verb building actually gives a sort of unique feel and flow to the language. If anyone wants to do some work on this, while keeping the morphosyntactic alignment in mind, try translating the following sentences in as many ways as you can, and send here if you've got answers!

"You are pleased"

"I am accepted"

"S/he is seen"

"I am eaten"

—Nexvs

[logodaidalos]

Basics of Grammar - Typological Considerations

In general, human languages can be grouped according to many things, one of which is their morphological typology: they can be, among other things, divided into analytic and synthetic languages, or rather placed on the analytic-synthetic continuum. On the analytic end, languages have approximately one morpheme per word, and on the synthetic end they have multiple morphemes per word. Languages such as Mandarin are very analytic, and those such as Latin are moderately synthetic.

At the far end of the synthetic part of the continuum, we find polysynthetic languages. Polysynthetic languages are basically languages with highly inflected words where pretty much everything revolves around verbs.

I think polysynthesis is where we could put trollish.

To elaborate more on the topic of polysynthetic languages, they are languages have verbs that possess a very high level of inflection, noun incorporation and polypersonal agreement.

Those last two are the most important.

Polypersonal agreement is just a term for having verbs agree with multiple arguments at once. While in English "sings" only tells us that the subject of this verb would be in the third person singular, sentences such as "I gave her it" require multiple words. Trolls would deal with this by inflecting the verb for all three arguments -- the agent, patient and recipient -- all at once. This would also mean that, as in f.e. Latin's "cantō" being translated as "I sing", pronouns for arguments would not be needed and would be used only for emphasis.

Noun incorporation is a process, usually valency-reducing, that transfers whole arguments from the sentence as a whole to specifically inside the verb. We can show this in English by comparing "She is picking berries" and "She is berry-picking", where "berries" becomes incorporated into the verb, which then becomes intransitive. Polysynthetic languages basically do this, but only on a much larger scale.

Noun incorporation is divided into four classes:

Patient incorporation for the purposes of narrowing down the function of the verb

Core argument incorporation as to make way for non-core (oblique) arguments to be emphasised

Incorporation of "old information"

Incorporation of semantically broad noun-like morphemes to use as verb classifiers and as regular inflection

To give examples for all four classes of incorporation:

"I tree-chopped today"

"I tree-chopped my field today"

"A whale came to us, and they whale-attacked it"

"A whale they fish-saw, and then they fish-attacked it until it fish-died"

According to current analyses of polysynthesis, all languages that have one type of incorporation will always have all other types of incorporation under it.

As for trollish, I'm thinking the language wouldn't work all that smoothly and elegantly with class four incorporation; second or third level incorporation seems quite good enough!

—Nexvs

[logodaidalos]

Terminology guide

Just a small notification: the Logodaidalos has a terminology page set up and the page is continuously expanded as more content is added. Don't understand something? The terminology page might be able to sort it out, but even if that fails you could send an ask!

[logodaidalos]

Basics of Grammar - Morphosyntactic Alignment

Morphosyntactic alignment, or in brief MSA, is the relationship between the sole argument of an intransitive sentence (the "subject") and the two arguments of transitive sentences (the "agent" and "patient"). English deals with these arguments by grouping the subject and agent together and differentiating them from the patient (f.e. "I eat" and "I hit him" & "He hits me"), but I've thought that the trolls could do with something called an ergative-absolutive alignment, where the subject and patient are grouped into the "absolutive" case and the agent into the "ergative".

What this implies for the language is, besides the noun marking, some structural differences from f.e. English. As all languages treat the case/category that includes the subject with preference over other categories, it would mean that all normal intransitive sentences in the language would work just like equivalent English passive sentences; what English would express with "I am eaten", the trolls would say as simply "I eat". Transitive sentences would remain the same, mostly.

An additional artefact of this is that we now need to introduce something that would be the mirror opposite image of the English passive: in linguistics it is generally called the "antipassive". This grammatical voice behaves by deleting the absolutive and "promoting" the ergative to the absolutive, just like the normal passive works by deleting the nominative and promoting the accusative to a new nominative ("I killed a bear" > "A bear is killed").

These antipassive sentences would work and mean the same as normal English intransitive sentences, mimicking what happened with the troll intransitive sentences and English passive.

Furthermore, this focus on the absolutive means more focus on the objects of verbs, which is reflected in the grammar: instead of using passive participles (like English "killed" where the action is carried out onto something), we would use antipassive participles (such as the English active participle "killing") more often, and deriving new words would probably go in the opposite direction: where English has a really productive <-er> morpheme that produces verb agents, trolls would use a corresponding and equivalent morpheme that would produce verb patients more often than not. English has the <-ee> morpheme that is infrequent and not quite as productive, and trolls might even lack such an equivalent and would rather use the participle.

Additionally, when leaving out certain arguments of a verb, it would be much more frequent for trolls to leave out the patient rather than the agent/subject. English drops the agent, as seen in the example sentence "I saw the bear and killed it"; trolls would rather say something equivalent to "I saw the bear and I killed" which, while totally wrong in English, would be just fine for them.

This property that the language of the trolls has is called "ergativity" and it is somewhat frequent even among human languages, but almost never to this extent.

—Nexvs

[logodaidalos]

Basics of Grammar - Castes (I)

On Alternia, and in the Alternian Empire, the trolls are the main (if not the only) component. Biologically, trolls are theoretically sorted into one of twelve blood castes, though limebloods, as far as we know, are not an extant caste.

There also exists a dichotomy between seadwellers and landdwelling trolls that permeates all layers of the caste mindset (as seen with Equius' contempt of Eridan). An additional feature of the hemospectrum (the blood caste system) is that there occasionally may be glitches or mutations, which can (as with Karkat) produce trolls of an abnormal blood colour, or give (as with Rufioh) physical mutations of an undesirable kind.

This division into blood castes is so prevalent that it even seeps into Karkat's own speech, even if he is against the hemospectrum. He identifies three different caste types:

When talking to Terezi, he identifies her as a 'blue blood' indirectly via his use of 'your blue blooded vernacular', but when he explains to Dave the nuances of romance in a trashy romance novel he uses the terms 'highblood' and 'lowblood'. Equius additionally uses 'sea dweller' as a label for Eridan's caste:

As Equius identifies Eridan as a sea dweller, but also as a highblood ("his blood is even purplier"), that kind of brings about that there might be multiple caste groupings involved.

I propose the following:

Lowbloods > [Red, Bronze, Yellow, (Lime)]

Midbloods > [Olive, Jade, Teal]

Highbloods > [Teal, Cerulean, Indigo, Purlple, Violet, Tyrian]

Additionally, I propose another division level that deals specifically with more general groupings by colour terms which excludes the lowbloods whose blood colours differ significantly:

Greenbloods > [(Lime), Olive, Jade, Teal]

Bluebloods > [Teal, Cerulean, Indigo, Purple]

Seadwellers > [Violet, Tyrian]

In both groupings, teal acts as a sort of a boundary, meaning that it can be used as both a midblood/greenblood colour or as a highblood/blueblood colour when appropriate in context.

Further, due to the massive impact on troll society and culture that the hemospectrum and castes have, a formality system based both on social status and blood colour would definitely be present. In form and function this system would likely resemble the formality system of Japanese, with honorific speech and extensive formality in the grammar itself.

On the other hand, I don't think that castes would be marked on the nouns themselves (which is what the conlang on nowordsforsnow did), but rather it seems better to mark the nouns with a sort of an animacy system. In that case, nouns would belong to one of three classes: terrestrial (land-based), aquatic (sea-based) and inanimate. Twelve noun classes might be a bit overkill.

—Nexvs

[logodaidalos]

A Masterpost -- Nexvs Logodædalī

So, the Logodaidalos has set up a masterpost, called the Nexvs Logodædalī that will contain every relevant post categorised and listed! Check it out!

Additionally, all posts will have the small linked "Nexvs" at their bottom so you can go and take a look there whenever you'd wish.

--Nexvs

[logodaidalos]

Phoneme Inventory

The first step towards making a conlang would probably be figuring out what contrastive sounds make up its phoneme inventory. The phonemes are the theoretical contrastive sound units that function as building blocks of morphemes, not the actual sounds that a mouth produces; the real sounds are phones and are dealt with under allophony and general phonetics.

For the purposes of inspiration gathering, the Logodaidalos always turns towards canon sources, but these are pretty non-existent (at the present time), so we're forced to be a bit more innovative. What struck me as a good parallel was the caste system of Alternia versus the caste systems of some human societies -- in this case, Tulu castes, which are arguably more complex than Alternian ones --  and so it might be a good parallel to draw between the phonology of a casted people and of Alternian trolls.

The phonology of Tulu is very Indian in its nature, as well as characteristically Dravidian. It contrasts retroflex consonants with denti-alveolar ones, and has a palatal series that is undergoing a shift to a postalveolar affricate series. It has a voicing contrast, but no aspiration or breathiness contrast present in a lot of other Indian languages. It also has a length contrast on six vowel qualities and one back unrounded vowel.

More inspiration the Logodaidalos found in the Caucasus. The Caucasus Sprachraum (language zone) is made up of multiple language families that have acquired characteristics of one another and developed some share features, namely ejective consonants (Ossetian, an Iranian language, has developed them from being in the zone) as well as a very high level of agglutination bordering on polysynthesis (as exemplified by Georgian and Tsez).

An additional, now thematic, inspiration the Logodaidalos found in the Nāhua people, of whom the Aztecs are/were the most notable members. Classical Nāhuatl is extremely polysynthetic, but with a relatively simple phonology that doesn't distinguish voicing in consonants and with four vowel qualities. It is also characterised by its prevalent use of the lateral affricate, /t͡ɬ/ tl, as a non-possessedness marker, and by widespread reduplication.

Mixing and matching some notable features of these interesting and different languages, we arrive at something that has properties of each group:

This table has many features many languages above described have, and is made up of some elements that are more-less unremarkable on their own but together form an unique blend. It has a whole set of ejectives, from all three MOA (method of articulation) classes. It's quite interesting in that it distinguishes ejective and tenuis affricates and fricatives, along with having gemination only on non-sonorant obstruents.

For vowels it is a bit more tame than what can be found on the Caucasus, incorporating some Dravidian and Nāhuatl elements:

The schwa may act as a sort of a reduced vowel that is inserted in predetermined positions and that allophonically lengthens under stress. The exact mechanism of its action is as-of-yet undetermined.

Stress in the language would probably be non-phonemic and uncontrastive, most likely determined by vowel length and syllable position, with a tendency to fall on the penultimate or antepenultimate.

Allophony for the language has yet to be worked out, but if anyone is willing to think about it, feel free to write on it and notify the Logodaidalos!

--Nexvs

[logodaidalos]

A Summary of Previous Attempts

This is definitely not the first such troll conlang attempt. I know of only one that panned out significantly (i.e. beyond the phonetics jab), and that is over on nowordsforsnow. The blog has some extensive morphology laid out, along with basic grammatical concepts (blood-castes dictating noun classes, for one) and some varied and topical lexica. The conlang there builds on the phonetics I wrote about in a previous post. That blog has quite a following for this sort of thing (which let's hope we'll manage here!), but, more importantly, turns up first when "troll conlang" is searched on tumblr. The blog has been fairly inactive for more than a year and a half.

Looking tags up like "homestuck linguistics" gives one additional post, and looking up "homestuck conlang" gives only the Logodaidalos' posts. So far it seems like a moderately empty field with lots of creative wiggling room. Some inspiration might be found in the nowordsforsnow's conlang, of course, but most of the ideas here I'm hoping will be original to as great an extent as I will be able to muster.

There is probably a lot to do, starting from an organisational and logistics point of view (what topic needs the most time allocated to it?) and ending with the questions of realisation (how to sort and present a dictionary f.e.), but it can be done. If anyone knows of any other troll conlang attempt, send a message and it'll be taken a look at!

--Nexvs

[logodaidalos]

Oral Apparatus

The first question that needs to be resolved is probably that of the vocal apparatus i.e. how the trolls would make sounds that make up their language. There are some older posts on tumblr that speculate that the trolls have a radically different mouth from the human one. I am not quite sure if this is really the case, as I have my own theory on this.

That image above is taken from a xenobiology post by isozyme and I disagree with it significantly. It has further spawned some offshoot posts, such as this one on consonants and its sister post on vowels. While the idea itself is quite good (not to mention the reasonably-ish accurate but otherwise really nice art!), I would rather like that we stuck to what precious little canon information there is.

Even if it is painfully scarce, canon actually tells us a whole lot about what the troll mouth looks like!

For one, we know that trolls have teeth (of course) and a tongue with taste buds (say, Kanaya). We know that they can inhale from both the nose (Terezi) and the mouth (say, Cronus), meaning that they would have a way to close off air supply to their mouths when needed (possibly in the form of a glottis?). We also know that they have a uvula-like oral organ (Karkat) which would imply that they also have a soft palate and a closable nasal cavity.

Assuming a sort of a humanoid mouthpiece (as has begun to appear), Equius' hisses would be (to us) an ingressive alveolar fricative, thus implying a hard palate and an alveolar ridge. This all pretty much means that many of the human vocal organs are potentially present in trolls; for now that would mean lips, teeth, hard and soft palates, a uvula and probably a glottis or an equivalent. As for an epiglottis and pharynx, a pharynx is pretty much obligatory but an epiglottis might or might not actually be there.

On the topic of the shape of troll tongues, a lot can be said. From Kanaya's above example and from Terezi we see flat troll tongues, and I'm not sure there is a canonical example of forked troll tongues (please correct me if I'm wrong)

Most of this implies a primarily human way of communication, quite unlike what the xenobiology posts tried to guesstimate based on plausible speculative alien physiologies. This kind of oral apparatus implies humanoid vowels and consonants; there may or may not be room for creativity there.

--Nexvs

[logodaidalos]

The language of Alternia?

It isn't the most frequent concern related to the backstory of Homestuck, but the language of Alternian trolls is something that has been mentioned numerous times inside of Homestuck itself and debated about in fan circles, yet none of it has ever been seen outside of 48 lexical items of dubious linguistic quality (troll names and surnames) and a substitution cipher alphabet (inverted Daedric), yet the various communication quirks (disregarding typing quirks) and phrases and minor details, such as language varieties, suggest that there may be something more to the whole thing.

When Terezi drags Karkat's load gaper out of his house, it (presumably) takes her a while to figure out that what Karkat called a "load gaper" she would have called a "toilet". His retort tells us that that's an artefact of her "blue blooded vernacular", leading us to think there might be more where that came from.

Going further than that, we know that the script the trolls in-comic use is inverted Daedric (we've seen magazines, Ouija boards etc. with it), yet another thing that strikes me is the lack of language change over huge swaths of time, as witnessed with Mindfang's journal:

The journal that Vriska reads was written by Mindfang anywhere between hundreds and thousands of sweeps before Vriska's time, yet the younger troll doesn't seem to have any issue with reading the journal as if it were written yesterday.

While it could be argued that Hussie did this as either not addressing the question of language change or doing it for the plot's good, it could be that the language spoken by Mindfang and by Vriska is actually pretty close, leading us to think that it might be artificially controlled. The presence of a "blue blooded vernacular" could be proof of social stratification that withstands and resists regular language regulation, maybe as a form of insurgency against the central authority.

For this reason one could assume that there might have been one single universal language imposed upon the populace, maybe as a lingua franca (well, Alternica in this case!), or maybe as the only legal language (such as the situation was with Basque in Nationalistic Spain f.e.) with punishments for speakers of other, disapproved languages. Dialects will have arisen with almost maximal certainty (if there's different caste vernaculars, there must be some geographical variation as well).

Seeing that humans and trolls could communicate with ease while together on the asteroid, even though they are two different species that hail from different, practically (for the purposes of linguistics) unconnected universes, it might be more of a plot's-benefit thing. It could mean they have really similar (for unknown reasons) vocal apparati, or could mean that Sburb/Sgrub is translating the language of one group for the benefit of the other (or even both!).

If Karkat and Dave can perfectly understand each other, their communcations could practically be insignificant for the purposes of logodaidaly except for maybe discerning some patterns of Karkat's speech and their significance for his original, "Alternian" methods of expression.

Troll names (both Alternian and Beforan) are really not the best indicator of anything linguistic seeing as they have been picked as symbolic, or by fans whatever their reasons be. Surnames such as "Ampora", "Captor",  and "Pyrope" are straight-up human terms, as are names like "Sollux", "Kankri" and "Cronus". Most of them that aren't are slight modifications of something human. Even the planet names are very much human (Alternia and Beforus).

Basically what all of this means is that, when making languages based on or for the trolls, we're going in blind.

--Nexvs