Mfalen: an attempted initial sketch

[Circeus]

This thread will serve as an attempt for me to jot down my ideas although they might not be as well-defined as I wish at the time. This first post is just a preliminary but of info because I need to start somewhere and I tend to get overwhelmed fast whenever I try to start writing actual sections of the grammar. Questions as to how aspects of the grammar are approached are welcome.

Concept

With this language I'm going for a polysynthetic language on a level closer to Nahuatl than say, Algonquian or Inuit languages. Intended key features are

• A phonology with some amount of unusual initial clusters and a tendency to move consonants to the beginning of the following syllables, but relatively straightforward rules to deal with illegal clusters.
• A derivational morphology that is noun-based instead of verb based. That is, there are no verb-to-noun nominalizers other than zero-grade nouning of the verbal root.
• Adjectives are not a separate workd class and may be noun, verbs, or in some cases incorporated affixes.
• No cases, but nouns do bear several grammatical affixes (at least definiteness, pluralization—including a dual—and possession are possible).
• The major verbal categorization is subject animacy. Valence is unconnected to the root. Instead it is determined by a verbal affix (almost certainly forming a system with tense or mood so it cannot seep back into the root). Pronouns are expressed.
• Pragmatic (e.g. males more likely to use polite speech to females than the other way) and vocabulary (there are more distinctions for relatives of a female than those of a male) elements marking a society egalitarian verging on matriarchal.
• An underlying society based roughly on ramages but in a developed society on the cusp of industrialization.

Basically the "sketch" will built as I add posts to this. I Intend for a summary of the phonology tomorrow, but It might well be superseded by a social sketch so I can free my mind of all the ideas about that XD.

[Frislander]

I know you haven't posted anything yet but already I'm intruiged!

A phonology with some amount of unusual initial clusters and a tendency to move consonants to the beginning of the following syllables, but relatively straightforward rules to deal with illegal clusters.

Interesting. Makes me think of Adyghe actually.

A derivational morphology that is noun-based instead of verb based. That is, there are no verb-to-noun nominalizers other than zero-grade nouning of the verbal root.

Now that's an interesting way to go about things. So how would your language distinguish, say, agent versus patient nominalisations? Just a zero-derivation of a root plus valence/voice affix?

Adjectives are not a separate workd class and may be noun, verbs, or in some cases incorporated affixes.

Promising! I like to see people do interesting things with their "adjectives" (though I feel I should point out that adjectival affixes is a very Algonquian feature, since you bring that up as an example of what you're no going for).

No cases, but nouns do bear several grammatical affixes (at least definiteness, pluralization—including a dual—and possession are possible).

Again nice, honestly I feel that sometimes conlangers often have a bit of a thing for case, and that it therefore turns up in conlangs more often than in natlangs. Maybe that's just because pretty much all of the most prominent/famous conlangs have had case systems.

The major verbal categorization is subject animacy.

Again, if you mean on the lexical level then this is also very much an Algonquian feature. Not that's bad in th slightest, just again to contrast with what you said earlier.

Valence is unconnected to the root. Instead it is determined by a verbal affix (almost certainly forming a system with tense or mood so it cannot seep back into the root).

Interesting! I'd love to see how this turns out. I'd guess this would manifest itself as overt transitivity marking such as that found in Munda and I think also Salishan?

[Salmoneus]

[*] A derivational morphology that is noun-based instead of verb based. That is, there are no verb-to-noun nominalizers other than zero-grade nouning of the verbal root.
[*] No cases, but nouns do bear several grammatical affixes (at least definiteness, pluralization—including a dual—and possession are possible).

I'm curious how those go together. If case isn't marked on the noun, and isn't marked on the verb either (since it's 'noun-based instead of verb-based'), do they just use strict word order instead? It's unusual, I'd have thought - not being an expert! - to see a heavy reliance on word order in a 'polysynthetic' language.

[*] Pragmatic (e.g. males more likely to use polite speech to females than the other way) and vocabulary (there are more distinctions for relatives of a female than those of a male) elements marking a society egalitarian verging on matriarchal.

I'm not sure that follows. Politeness toward women, after all - the sense that they are more important than men, and more dignified and deserving of respect - is often an indicator of patriarchy and misogyny.* Similarly, since distinctions in relations indicate social obligations, having more distinctions for relations of women may indicate that men are socially less constrained, while women are more tied down by complicated familial duties.

*think of patronising old Englishmen saying "You're wrong, Humphrey, and I'll tell you why..." vs "I'm sure you must be correct, my dear lady" (and then having the argument with her husband when she's out of the room). He's more polite to the woman - perhaps out of genuine concern and respect - but he's also excluding her, idealising her and infantilising her. Patriarchy rarely works by people flat-out saying that women are less important than men - it usually works by saying women are more important than men (and thus must be protected from harm). To give a topical example: in Russia at the moment, there's a list of occupations it's forbidden for women to hold, on the grounds of potential danger (things like 'lorry driver', 'factory worker', 'fisherman', etc).

[hwhatting]

Patriarchy rarely works by people flat-out saying that women are less important than men - it usually works by saying women are more important than men (and thus must be protected from harm). To give a topical example: in Russia at the moment, there's a list of occupations it's forbidden for women to hold, on the grounds of potential danger (things like 'lorry driver', 'factory worker', 'fisherman', etc).

Well, "usually" is probably not quite correct. E.g. Islamic patriarchy clearly states that women are worth less than men (as witnesses, when inheriting. etc.) The justification "women are so special and important that they need to be protected (even from themselves)" is a specific feature of 19th century Western patriarchy (and of some of its modern continuations).

[Salmoneus]

Patriarchy rarely works by people flat-out saying that women are less important than men - it usually works by saying women are more important than men (and thus must be protected from harm). To give a topical example: in Russia at the moment, there's a list of occupations it's forbidden for women to hold, on the grounds of potential danger (things like 'lorry driver', 'factory worker', 'fisherman', etc).

Well, "usually" is probably not quite correct. E.g. Islamic patriarchy clearly states that women are worth less than men (as witnesses, when inheriting. etc.) The justification "women are so special and important that they need to be protected (even from themselves)" is a specific feature of 19th century Western patriarchy (and of some of its modern continuations).

Sorry, but I just don't think this is right. It has been a long trait of European culture, going back certainly to the 17th century at least, and I think long before that - the idea of courtly love in Chaucer certainly seems to at least suggest this dynamic.

Regarding Islamd, I'm not an expert, and perhaps what you say is true; but I'm skeptical. The difference in inheritence law can be explained economically (women did not have to pay dowries, and were expected to be maintained by a husband, so daughters did not need as much money as sons); the rules on witness are explicitly to safeguard against a woman making a mistake in giving evidence. That implies that women are thought of as more prone to error, particularly in intimidating legal surroundings, but thinking that women are silly little creatures who heads explode if they try to think too hard is not the same as thinking they are worth less - indeed, historically in Europe, making that equation would have been seen as misogynist (because it equates worth with what was believed to be the 'male' virtue of reasoning, rather than the perceived female virtues). Indeed, you yourself accept 19th century Europe as an example of protective patriarchy, yet these two views - that women were more prone to intellectual error, and that women naturally ought not to have the same inheritance rights as men - are exactly two of the most stereotypically patriarchal views of that paternalist era.

But maybe this needs its own thread...

[Zaarin]

It's unusual, I'd have thought - not being an expert! - to see a heavy reliance on word order in a 'polysynthetic' language.

Unusual but not unheard of. Some polysynthetic languages, like Mohawk, have very free word order, but some of the polysynthetic languages of the PNW are much stricter.

Sorry, but I just don't think this is right. It has been a long trait of European culture, going back certainly to the 17th century at least, and I think long before that - the idea of courtly love in Chaucer certainly seems to at least suggest this dynamic.

Prior to the 11th or 12th century in Europe you definitely have a general atmosphere wherein a person's value is related to their ability to fight; since women didn't fight, they had less value. After that time, you see a trend towards glorification of women promoted by the Church (Marian theology), the nobility (Eleanor and Louise of Aquitaine and their "Courts of Love"), and above all by the troubadours (who generally attributed it to either Arthur or Charlemagne). In some ways this did indeed improve the status of women; in others, it made them "the fairer sex" who must be sheltered.

[Circeus]

Gosh, I didn't expect any answer at all, much less this amount!

The egalitarian-to-matriarchal thing

I have some amount of conworlding done on this, but probably not as much historical/biological background for it as might be wished, owing in part to it having never been written down. This section is basically my first attempt to organise a lot of distinct ideas and concept into a coherent whole. Please with me!

The basic gender relation is that a woman, culturally and legally, is not treated any different from a man. This stem from childbirth being much less a health risk to Mfalen females than it is to human ones. Since Mfalen as a race (the Mfalen are mammalian but nonhuman, the basic concept is that they are humanoid that evolved from what we would describe as a mid- to small-sized feline) have less body strength difference between genders and youngs that are helpless for significantly shorter length than humans (I'm thinking it may not be possible for a female to become pregnant for a certain length of time after birth, similar to lions), the original tendency to treat women as a resource and not individuals was less present. Furthermore, the males are unusually susceptible to a form or early-onset dementia (i.e. basically Alzheimer), the first symptoms of which may appear as early as the equivalent of the mid-to-late thirties. As such, experienced leaders are more likely to be females.

Except for things like lactation and childbirth, a women genuinely can do anything a man can and there are virtually no profession that are gender dominated. In marriage, either spouse may join the other's lineage (family law is based off a ramage system). A servant hired to take care of a weaned child is just as likely to be male than female. Leadership is generally based off consensus: a leader is expected to express the will of their governed less they lose their "job", but they are given the benefit of the doubt either way. Passing laws and setting policy is kinda similar with the presidential system in that regard: a leader has a limited amount of political capital to put forth ideas that requires the approval of others (i.e. unpopular or controversial ones), and they need to choose how to expend that capital.

To give a specific pragmatic example: while it is common for a man to use polite pronouns for their female relatives (especially the older ones), a woman that does so a lot is perceived as either sarcastically mocking, or exaggeratedly servile. A man will usually talk/refer to to his mother using female pronouns all his life, but most women stop doing so when they reach marriageable age, and they hardly ever do so when talking about siblings.

Women have a Sudanese-like kinship system. Men's relatives are closer to an Inuit system, with only nuclear family members being distinguished by gender. The parents and siblings of a married male are not part of that nucleus, and as such a married and unmarried male do not refer to these relatives the same way (a married male refers to his parents/siblings with the same gender-neutral term as he does his aunts/uncles and cousins).

Answers

Interesting. Makes me think of Adyghe actually.

Actually, as you'll see, it's a lot closer to Russian, with a (C)(C)V(C) structure that generally disallows vowel hiatuses. I wanted something that was definitely a little out there compared to Standard Western European phonology, but which I could still read aloud with reasonable ease.

Promising! I like to see people do interesting things with their "adjectives" (though I feel I should point out that adjectival affixes is a very Algonquian feature, since you bring that up as an example of what you're no going for).

Incorporated "affixes" was probably the wrong term. I'm really thinking of word incorporation/agglutination, here with the incorporated word being an otherwise nominal root that is a color word. The base root tlīlli in Nahuatl works that way, e.g. tlīlpa "paint black", tlīlpoyāhua "mark with black spots". Otherwise a language with no separate adjective class is not all that unusual, especially where polysynthesis is present (Nahuatl is one; its handful of adjectives are seemingly derived from lost verbs).

Now that's an interesting way to go about things. So how would your language distinguish, say, agent versus patient nominalisations? Just a zero-derivation of a root plus valence/voice affix?

I had not even noticed this issue! Now that I think of it, I may use an affix that would have been used for direct/inverse marking in Old Mfalen.

Again, if you mean on the lexical level then this is also very much an Algonquian feature. Not that's bad in th slightest, just again to contrast with what you said earlier.

Again, my short description may have been unclear. Animacy does not affect the conjugation. It affects which verbs can be used with a given subject. For example, in Mfalen, "to rain" is an animate verbs. So you do say "I rain", "you rain" (the meaning is actually "it rains on me") and "New York is raining" (locatives are animate), but you cannot say e.g. "my house rains" (i.e. "it rains at my house") because possessives cannot be made into locatives. Similarly, you falling from tripping and a branch falling on you require different verbs. Think of it (very awkward simile incoming!) as if a split-S language required different verbs instead of different cases.

Interesting! I'd love to see how this turns out. I'd guess this would manifest itself as overt transitivity marking such as that found in Munda and I think also Salishan?

That's the idea, although (as I mention) the transitive marker wouldn't be independent but part and parcel or another of the conjugation paradigms.

[*] A derivational morphology that is noun-based instead of verb based. That is, there are no verb-to-noun nominalizers other than zero-grade nouning of the verbal root.
[*] No cases, but nouns do bear several grammatical affixes (at least definiteness, pluralization—including a dual—and possession are possible).

I'm curious how those go together. If case isn't marked on the noun, and isn't marked on the verb either (since it's 'noun-based instead of verb-based'), do they just use strict word order instead? It's unusual, I'd have thought - not being an expert! - to see a heavy reliance on word order in a 'polysynthetic' language.

As I explicitly stated (or so I though. My Asperger and the simplicity of the description means I may well not have been clear at all!), I'm not looking for as heavy a level of polysynthesis as in Algonquian or Inuit languages, but rather something halfway between them and, say, Quechua or German, with a result close to Nahuatl (i.e. "oligosynthetic" as some have called it), which does have some rather complex verb and a high level of Germanic-type nominal agglutination, but also a lot of syntactic particles and no case markings.

[Circeus]

And the phonology as promised. Hopefully this is relatively understandable. It is by far the most developed aspect of the language at this point, although I have shortened the explanations and not listed any examples.

Phonology

Phonology is relatively simple, with only stops and affricates showing a voicing contrast. As far as I'm aware, the most divergent feature (if it can even be called that) is the lack of /w/.

• Stops and affricates: /p b/ /t d/ /k g/ /ts dz/
• Nasals: /m n/
• Fricatives: /f s h/
• "Approximants": /l ɾ y/

All consonants geminate except /h j/. A geminated /ɾ/ is a thrill.

Vowels: /a ɛ i u ɒ iɒ au ɛi/
I have some written material about vowel variation which is mostly just ass-pulling, so I won't include it here.

Allowed clusters are as follow:

• /m/+fricative or liquid, /n/+liquid /ms/, /mf/, /mh/, /ml/, /mɾ/, /nl/, /nɾ/
• Voiceless stop, affricates and nasals+/j/ /pj/, /tj/, /kj/, /ʦj/, /mj/, /nj/
• fricative or affricate+/ɾ/ /sɾ/, /fɾ/, /hɾ/, /ʦɾ/, /ʣɾ/
• /p/+voiceless stop or affricate /pt/, /pk/, /pʦ/
• /ps/, /fs/, /hs/

As I mentioned above, these are not all that exotics, but to me they feel so, while remaining reasonably pronounceable. I have a vague theory that Proto-Mfalen had a /p b p^h/ distinction, that became /p f b/, with no aspirate-initial clusters, but that makes the presence of pairs like /ps hs/ hard to justify.

Stress and phonotactics
Mfalen stress is on the penultimate syllable, adding suffixes do displace stress.

Syllables are in the form C₁V(C₂), with C₁ any consonant or legal cluster, and C₂ any consonant other than /h j ts dz/. Syllables without onset only occur at the beginning of words. Nonnasal C₂ should a marked tendency to move to the beginning of the following syllable, all the more so if that syllable is stressed. Nasals will move only if the resulting cluster would be legal.

Double vowel hiatus are illegal, as are hiatuses involving long vowels and diphthongs. That is, only a single pair of simple vowels may be in hiatus. If an illegal hiatus would occur, an epenthetic /n/ is inserted (/l/ if the root contains a liquid).

Ungeminated /l/, /n/ and /ɾ/ (i.e. alveolar sonorants) are elided between identical vowels unless the hiatus would be illegal. I haven't yet quite figured out whether the result of manat -> ma'at is /ma.at/ or just /māt/. in either cases, there no other sources of long vowels.

Haplology occur when two identical CV syllables with short vowels occur one after the other. This results in a geminated consonant. All consonants other than affricates and /h j n l/ can geminate in this way (/n l/ are elided /ɾ/ becomes a thrill). A pair of affricates may not collapse together, but a stop followed by the matching affricate will do so (i.e. pocci[/i ]is from potici, not an hypothetical pocici).

Geminates (including those generated by haplology) do not occur word-initially or within cluster. If a geminate would become part of a cluster, it simplifies to a single consonant.

The rules for cluster simplification are as follow:

• The sequences /tpt kpk tpʦ/ reduce to geminates.
• In other sequences of stops+<pt pk pc>, the first consonant is dropped unless it's a nasal (which remains in coda position).
• In all other instances, the middle consonant is dropped. If this result in a /ɾl/ or /lɾ/ cluster, they assimilate to /r/.
• If the middle consonant of a cluster reduced to /r/ was /h/ or a nasal, the resulting consonant is nasalised or aspirated as approriate.

[hwhatting]

@Sal: I'm no expert on Islamic culture either; but my impression from reading about other cultures (Islamic, Far Eastern, India, Antiquity) is that seeing women as more important and therefore in need of protection is something that is typical only for a limited period and region (Western Culture, and we can quibble on when exactly that view started to become an important justification for denying rights to women). Maybe it's a bit more farspread, but I doubt that justification is the usual one in other cultures and periods. I don't have the time now to search for all those writings from different cultures and periods where women are described as less valuable, weaker, less intelligent, morally inferior etc., but they're out there (and it's of course a strand of thinking which never totally went away even in the Western justifications of patriarchy).

[Vijay]

At 17:22 in this documentary, a Pakistani lady (the narrator throughout said documentary) asks the leader of the Muttahida Majlis-e-Amal (I think Qazi Hussain Ahmad) in English why two women have to be witnesses to a crime and only one man, and he says, "This is in order to lessen the burden of this on the women because as the- according to the Islamic principles, the women should give more attention to their homes. We give respect to the women of...We...We prefer them. We give them preferable..."

[Circeus]

(One day I will be able to start threads on here that don't get suddenly hijacked over a trivial matter)

[Vijay]

I'm sorry, when did it "get suddenly hijacked"? Did you not bring up egalitarian and matriarchal societies? Is it that surprising that a discussion of patriarchy might follow from that?

[Circeus]

I'm sorry, when did it "get suddenly hijacked"? Did you not bring up egalitarian and matriarchal societies? Is it that surprising that a discussion of patriarchy might follow from that?

Let's just say that I think a debate on what real world societies were or weren't misogynist patriarchies and their reasoning for it is pretty remote from what was really just a list of linguistic features,

[Frislander]

In a bid to get things back on track...

Interesting. Makes me think of Adyghe actually.

Actually, as you'll see, it's a lot closer to Russian, with a (C)(C)V(C) structure that generally disallows vowel hiatuses. I wanted something that was definitely a little out there compared to Standard Western European phonology, but which I could still read aloud with reasonable ease.

Well this initial description could apply to Adyghe just as well as Russian if not better, however, the actual list of CC clusters you give subsequently is definitely more Russian, yes.

Promising! I like to see people do interesting things with their "adjectives" (though I feel I should point out that adjectival affixes is a very Algonquian feature, since you bring that up as an example of what you're no going for).

Incorporated "affixes" was probably the wrong term. I'm really thinking of word incorporation/agglutination, here with the incorporated word being an otherwise nominal root that is a color word. The base root tlīlli in Nahuatl works that way, e.g. tlīlpa "paint black", tlīlpoyāhua "mark with black spots". Otherwise a language with no separate adjective class is not all that unusual, especially where polysynthesis is present (Nahuatl is one; its handful of adjectives are seemingly derived from lost verbs).

So more akin to noun-compounding then? Also I'm perfectly ware of languages which have little or no "adjectives" as such, in fact they're my preferred type when it comes to conlangs.

Again, if you mean on the lexical level then this is also very much an Algonquian feature. Not that's bad in th slightest, just again to contrast with what you said earlier.

Again, my short description may have been unclear. Animacy does not affect the conjugation. It affects which verbs can be used with a given subject. For example, in Mfalen, "to rain" is an animate verbs. So you do say "I rain", "you rain" (the meaning is actually "it rains on me") and "New York is raining" (locatives are animate), but you cannot say e.g. "my house rains" (i.e. "it rains at my house") because possessives cannot be made into locatives. Similarly, you falling from tripping and a branch falling on you require different verbs. Think of it (very awkward simile incoming!) as if a split-S language required different verbs instead of different cases.

That's still sounds very much like Algonquian stem alternation (its little talked about but it's a given in Algonquian languages that "he is cooked" and "it is cooked" will use different verb stems, and the same with the transitive pair "I see him" vs. "I see it"). But no matter.

Additionally your coice of the verb "to rain" was perhaps a poor choice, since it does not (by your characterisation) appear to have a typical subject at all and instead takes a locative argument, akin to the "dative subject" phenomenon. Also I am most confused by your characterisation of locatives as animate; I myself would have thought it more natural for them to be inanimate instead. And "possessives cannot be made into locatives"? If by this you mean that possessed nouns cannot take the locative (as your example would indicate) then 1. it's not a "possessive" as such, 2. that seems like a weird constraint, like how would you say "at my house"? and 3. what kind of language is this that "I am rained on" and "it is raining in New York" are expressible but "it is raining at my house" is not?

[mèþru]

(One day I will be able to start threads on here that don't get suddenly hijacked over a trivial matter)

My only response:

More: show

ha ha ha Ha Ha Ha HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA Ha Ha Ha ha ha ha

Keep up the good work! But thread drift is inevitable.

[Circeus]

In a bid to get things back on track...

Interesting. Makes me think of Adyghe actually.

Actually, as you'll see, it's a lot closer to Russian, with a (C)(C)V(C) structure that generally disallows vowel hiatuses. I wanted something that was definitely a little out there compared to Standard Western European phonology, but which I could still read aloud with reasonable ease.

Well this initial description could apply to Adyghe just as well as Russian if not better, however, the actual list of CC clusters you give subsequently is definitely more Russian, yes.

Again, if you mean on the lexical level then this is also very much an Algonquian feature. Not that's bad in th slightest, just again to contrast with what you said earlier.

Again, my short description may have been unclear. Animacy does not affect the conjugation. It affects which verbs can be used with a given subject. For example, in Mfalen, "to rain" is an animate verbs. So you do say "I rain", "you rain" (the meaning is actually "it rains on me") and "New York is raining" (locatives are animate), but you cannot say e.g. "my house rains" (i.e. "it rains at my house") because possessives cannot be made into locatives. Similarly, you falling from tripping and a branch falling on you require different verbs. Think of it (very awkward simile incoming!) as if a split-S language required different verbs instead of different cases.

That's still sounds very much like Algonquian stem alternation (its little talked about but it's a given in Algonquian languages that "he is cooked" and "it is cooked" will use different verb stems, and the same with the transitive pair "I see him" vs. "I see it"). But no matter.

I guess this reveals more about my lack of knowledge of specifics in Adyghe and Algonquian languages than about Mfalen XD

Additionally your choice of the verb "to rain" was perhaps a poor choice, since it does not (by your characterisation) appear to have a typical subject at all and instead takes a locative argument, akin to the "dative subject" phenomenon. Also I am most confused by your characterisation of locatives as animate; I myself would have thought it more natural for them to be inanimate instead. And "possessives cannot be made into locatives"? If by this you mean that possessed nouns cannot take the locative (as your example would indicate) then 1. it's not a "possessive" as such, 2. that seems like a weird constraint, like how would you say "at my house"? and 3. what kind of language is this that "I am rained on" and "it is raining in New York" are expressible but "it is raining at my house" is not?

It can be expressed easily enough, just not with using the locative as the subject, again my terminology is awful. I should never have used "Locative" to describe the proper noun of a place. Placenames in Mfalen are generally animate (as would, indeed, be locatives derived from them), but a locative derives from a noun+possessive affix is not normally a placename (which brings the question whether I want separate series of locatives for animates and inanimates, mmhmmm...).

I can think of several solutions to the "it's raining at my house" issue. Mfalen will have a deagentive affix performing double duty as an impersonal and a passive (only in the semantic sense is it a passive, since it deletes the verbal subject without promoting the object.). Since your locative at-my-house can still act as an "adverb" without being an argument (indeed, I haven't explored locatives much, so they may well end up not being available as verbal arguments at all), you get what looks like the same phrase in a pro-drop language, except it's marked for the impersonal.

In practice, I figure people would tend to say "I rained at-my-house" regardless of whether they were there at the time.

[Circeus]

A few tidbits I wanted to show off/jot down:

The three meanings of verbs

An example:

Base verb: EAT (deverbal root: food/meat, the Mfalen assimilate most plant matter very poorly)
Intransitive: have a meal
Transitive: eat something ("you gotta eat to live" would is a deobjective)
Bitransitive: eat something with someone (the bitransitive is frequently comitative or benefactive in meaning, I have not decided if it's considered normal for several options to be possible or if it's more set in stone)

It is a common usage in the more formal registers to coin new usages by using verbs in a valency that is not usual

Base verb: DREAM
Intransitive: sleep (as a result the original verb for sleep refers to other forms of unconsciousness in modern Mfalen)
Transitive: dream about something
Bitransitive: [missing]

An example of a literary bitransitive that became normal usage is "reveal/expose". The bitransitive of that verb came to mean "reveal something that was hidden by something else".

Verbal incorporation

Mfalen does not (oddly enough) allow the incorporation of nouns within a verb, but it does allow incorporation of other verbs. An incorporated verb does not usually bring arguments with it, but rather expresses a mean or a result of the action. While the verb must share subject animacy, the semantic relations between them may vary greatly

lay-punch &rarr; punch to the ground
lay-push &rarr; shove down/to the ground, tackle
sleep-punch &rarr; punch unconscious
open-the-arms--reveal &rarr; reveal by throwing curtains open (i.e. inaugurate)
scrape-reveal &rarr; reveal by scraping something else away

etc.

In-laws

Traditional Mfalen society doesn't have a concept of in-laws because one of the two spouse is essentially adopted into the other's clan (i.e. come sunder the exclusive authority of that clan's leader for various matters). As such, they refer to their spouse's relatives as though they were their own, whereas the link to their prior relatives is considered to be severed. If there was really a need to refer to these relations, one would talk about one's or one's spouse's former relatives.

However, there has been a rise in popularity of previously rare nonadoptive marriages, and language is still in flux about it. A common way to refer to in-laws in such a marriage is to either refer to them as the relatives of one's spouse, but females in particular find it awkward to use the limited vocabulary available to males. Expressions like "fake cousins" or "new mother" are widely used.

[Zaarin]

In-laws

Traditional Mfalen society doesn't have a concept of in-laws because one of the two spouse is essentially adopted into the other's clan (i.e. come sunder the exclusive authority of that clan's leader for various matters). As such, they refer to their spouse's relatives as though they were their own, whereas the link to their prior relatives is considered to be severed. If there was really a need to refer to these relations, one would talk about one's or one's spouse's former relatives.

However, there has been a rise in popularity of previously rare nonadoptive marriages, and language is still in flux about it. A common way to refer to in-laws in such a marriage is to either refer to them as the relatives of one's spouse, but females in particular find it awkward to use the limited vocabulary available to males. Expressions like "fake cousins" or "new mother" are widely used.

Wouldn't that trigger most cultures' incest taboos? In most cultures with matrilineal clans, the children inherit their mother's clan and typically have a more father-child relationship with their maternal uncle(s) than with their father, who is not part of their clan. (See, for example, the Iroquois, Haida, or Tlingit, among others.)

[Soap]

Verbal incorporation

Mfalen does not (oddly enough) allow the incorporation of nouns within a verb, but it does allow incorporation of other verbs. An incorporated verb does not usually bring arguments with it, but rather expresses a mean or a result of the action. While the verb must share subject animacy, the semantic relations between them may vary greatly

lay-punch &rarr; punch to the ground
lay-push &rarr; shove down/to the ground, tackle
sleep-punch &rarr; punch unconscious
open-the-arms--reveal &rarr; reveal by throwing curtains open (i.e. inaugurate)
scrape-reveal &rarr; reveal by scraping something else away

etc.

I like this, but wouldnt the "reveal" ones be better ordered the other way around? Revealing the object is the result of a physical motion there, which in the other three verbs comes at the end of the word.

It is a common usage in the more formal registers to coin new usages by using verbs in a valency that is not usual

Base verb: DREAM
Intransitive: sleep (as a result the original verb for sleep refers to other forms of unconsciousness in modern Mfalen)
Transitive: dream about something
Bitransitive: (missing)

Is the ditransitive* marked morphologically such that it's distinct even on its own? i.e. could I say I dreamt about you, using the ditransitive form of the verb, and just not mention the other argument, or would that result in a sentence that's exactly the same as the ordinary transitive voice?

--------------
*I know, I know, it "should" be bi- because its Latin, but thats just the way it is.

[Circeus]

I like this, but wouldn't the "reveal" ones be better ordered the other way around? Revealing the object is the result of a physical motion there, which in the other three verbs comes at the end of the word.

In the case of reveal, the usage started as one of those cases of high-literary ditransitive that became a normal usage. The verbal incorporation came later as the construction was more broadly appropriate by language users. As a result, just like the formula goes "black and white" and not the other way around, "scrape" does not get verbal incorporation in that way (I'm pretty sure the default oblique object in Mfalen scrape is an instrumental).

Base verb: DREAM
Intransitive: sleep (as a result the original verb for sleep refers to other forms of unconsciousness in modern Mfalen)
Transitive: dream about something
Bitransitive: (missing)

Is the ditransitive* marked morphologically such that it's distinct even on its own? i.e. could I say I dreamt about you, using the ditransitive form of the verb, and just not mention the other argument, or would that result in a sentence that's exactly the same as the ordinary transitive voice?

All valencies are explicitly marked (I've not yet figured out if a verbal root without such marking exists at all). You do bring a good point! I was intending to include both a deagentive and deobjective affixes, but never considered whether a deobliquative was needed! By all mean one should exist!

Assuming a deobliquative, what you suggest would grammatical, but probably not semantically sensical as the verb is currently defined (at least ignoring "default" meanings like comitative or benefactive) unless such an usage had been defined on an ad hoc basis within the text.

[Circeus]

I had not noticed this post earlier, so apologies for this incredibly late answer.

In-laws

Traditional Mfalen society doesn't have a concept of in-laws because one of the two spouse is essentially adopted into the other's clan (i.e. come sunder the exclusive authority of that clan's leader for various matters). As such, they refer to their spouse's relatives as though they were their own, whereas the link to their prior relatives is considered to be severed. If there was really a need to refer to these relations, one would talk about one's or one's spouse's former relatives.

However, there has been a rise in popularity of previously rare nonadoptive marriages, and language is still in flux about it. A common way to refer to in-laws in such a marriage is to either refer to them as the relatives of one's spouse, but females in particular find it awkward to use the limited vocabulary available to males. Expressions like "fake cousins" or "new mother" are widely used.

Wouldn't that trigger most cultures' incest taboos? In most cultures with matrilineal clans, the children inherit their mother's clan and typically have a more father-child relationship with their maternal uncle(s) than with their father, who is not part of their clan. (See, for example, the Iroquois, Haida, or Tlingit, among others.)

I'll admit I'm not familiar with these culture. This system actually evolves out of an Hawaiian type system where the distinction was not in relative generation but blood distance (so that an uncle and a grandparent are treated the same). Inheritance is still worked in a similar way. In my mind, taboo doesn't trigger simply out of the fact that the couple themselves are not referred as related to each others. (and it's obviously still possible for people of the same clan, but distantly related, to marry each others)

On a different topic, I've been messing around a little with the verbal structure. Although the language is _meant_ to have a somewhat polysynthetic verb, combining everything I wanted makes me nervous because I'm scared of ending up too kitchen-sinky. Without further ado, a quick summary of the verb structure.

(Relativizers) + (Object) + (Suppressives) + (Evidential) + (Derivations) + Root + (Voice) + Valency + (Aspects) + Plural

1. Relativized verbs take a special affix that replaces suppressives, since the relativized argument is not reflected in the relative clause.
2. I was going to drop the object suffix (used to mark object animacy), then decided they are on their way out instead, evolving into a direct/inverse system in some dialects (with different application depending on the dialect!)
3. Suppressives are used to mark the absence, or more accurately the depersonalisation of an argument.
4. Mfalen evidentials are unusual in that instead of indicating any origin of the information, they indicate how firmly the speaker believes it to be true, resulting in one marker that is essentially a negative marker and often referred to by foreigners as a "sarcasm mark".
5. Derivational affixes are a mixed bag halfway between grammatical and derivational affixes, "modal verbs" (which are incorporated verbs with modal meaning, sometimes with no independent verb existing), the derivational causative, oblique affixes and affixes translating to "un-" or "mis-" fall into that category because the relevant independent noun root may or may not exist.
6. Mfalen can mark three voices: a causative (which perform a different valency operation than the derivational causative), an applicative (swapping a peripheral argument into the oblique position) and a reciprocal-reflexive. Mfalen has no passive: causatives or suppressives perform the equivalent function instead,
7. I've discussed valency markers previously. Were they not separated from the root by voice affixes, they would probably be considered derivational.
8. Mfalen has various aspect, though it does not make a perfect/imperfect distinction. Unmarked verb are gnomic/generic statements, but a specific gnomic affix also exist and can be used in combination with other aspect marks.
9. Having removed the person markers (subjects are mandatory unless suppressed), I may or may not keep a plural marker on the verb. I'm thinking maybe base numbers can be used in that spot in additional to the plural proper.

I currently have a set of derivational markers to allow a verb being used with subjects of conflicting animacy, but will probably repurpose the object suffix instead (one of the possible evolution for it).