Icecap

Icecap Moonshine is a highly divergent language spoken around the year 6843 in cold climates famous for its small root vocabulary, compact morphology, and wide gap between male and female speech registers. When men are allowed to speak at all, they use a much more difficult speech register than women do, and when women speak to men, they use a speech register that omits crucial information, so men have to listen closely and think quickly whenever a woman gives them a new chore to do.

The first Moonshine speakers arose in the year 3948, and committed the Great Conspiracy, forever abolishing all male social power structures and spreading their revolution to foreign nations as well. The Moonshines prospered in their radical new society for about 150 years, whereupon a traditional male army invaded and crushed the Moonshine empire. Nevertheless, the winners of the war were unable to occupy Moonshine territory, and the Moonshines became even more feministic as they retracted into supreme isolation for the next three thousand years.

Moonshine women are much taller than their men, and it soon became unnecessary to apply social pressure to force men down to the bottom of the society; female superiority was seen as the only possible natural order, and few men even contemplated fighting back.

Although there have been other societies in which female power was even more unfairly stacked against men, the Icecap Moonshine language is notable even by comparison to these other societies for the great extent to which the social way of life has become entrenched in their language.

Scratchpad

New genitive infix

16:19, 2 March 2022 (PST)

The new sound change list opens up the possibility for a genitive infix appearing around 4800 AD, such that e.g. ŋàsiḳa "door" and its genitive ŋasiḳas could appear as ŋàsʲk and ŋàsʲikˠ. There may need to be a way to stop the palatalization from staining the following consonant in the resulting cluster. The stabilization of the accent on the root is not a problem however.

Possible use of Play block scripts

16:19, 2 March 2022 (PST)

The Police in Play territory wanted to write Moonshine using the Play block script or an adaptation of it, at least for pleasure. They admired the small children they saw writing messages in a script that their own adults could scarcely make sense of, let alone read, but which the Play children all seemingly understood without a moment's pause. Moonshine had too many phonemes to use the Play script directly, they realized, but the Police hoped that they could make a new version of it, whether it be more complex to handle the larger phonology of Moonshine or less complex to show their admiration for the superiority of the Play language's adaptability to artistic means of writing.

ther

equivalent forms

ăpiz "pregnant womb" (not *ăpʲiz) is equivalent to puz "pregnant womb", and the two are freely interchangeable, as the /a/ means "body, core, whole" and the second part is genitive in the longer word but nominative in the shorter word. Nonetheless this forms the basis of more interchanges.

Possible expanded consonant inventory

It's not clear that this actually adds any new consonants to the store ... it's just a matter of analysis, particularly with consonant + /w/ sequences.

Consonants

                      VOICELESS              VOICELESS      LIQUIDS AND
                        STOPS         NASALS   FRICS         LATERALS
Rounded labials:      ¦     pʷ ¦ bʷ ¦     mʷ ¦ fʷ     ¦     ¦     w  ¦    ¦    ¦    ¦
Neutral labials:      ¦ p      ¦ b  ¦ m   ṁ  ¦ f   ṗ  ¦     ¦        ¦    ¦    ¦    ¦
Spread labials:       ¦ pʲ     ¦    ¦ mʲ     ¦        ¦     ¦        ¦    ¦    ¦    ¦
Lower dentals:        ¦     ṭʷ ¦    ¦     ṇʷ ¦        ¦     ¦        ¦    ¦    ¦    ¦
Upper dentals:        ¦     ṭ  ¦    ¦     ṇ  ¦     ṣ  ¦     ¦     ḷ  ¦    ¦    ¦    ¦
Alveolars:            ¦ t      ¦ d  ¦ n      ¦ s      ¦  z  ¦ l      ¦ r  ¦ c  ¦ ʒ  ¦
Postalveolars:        ¦ č      ¦ ǯ  ¦ ň      ¦ š      ¦  ž  ¦ ʎ      ¦ ř  ¦    ¦    ¦
Palatals:             ¦ ć      ¦    ¦ ń      ¦ ś      ¦     ¦ y      ¦    ¦    ¦    ¦
Neutral velars:       ¦     k  ¦ ġ  ¦     ŋ  ¦        ¦     ¦ r̄      ¦    ¦    ¦    ¦
Labiovelars:          ¦ kʷ     ¦ ġʷ ¦ ŋʷ     ¦ hʷ     ¦  gʷ ¦        ¦    ¦    ¦    ¦

Alternate glyphs available for certain sounds

1. b: ƀ ƃ ɓ
2. p: ᵽ ᵱ (these display differently in different fonts)
3. k: ʞ ꝁ ꝃ ꝅ

The true palatals ć ń ś can also be represented as kʲ ŋʲ hʲ since they alternate primarily with velars in the morphology.

NOTE Dec 13 2021: I no longer remember why /bʲ ġʲ/ are not included here. It is possible that they never arose because the voiceless counterparts are assumed to have arisen from reversed forms like /jp jk/, and the palatal glide would not have crossed the clusters /mp ŋk/ that later became /b ġ/.

analytic possessives

ćan use oblique forms of the words for man and woman (and boy, girl, dolphin, etc) to indicate possession, even if the old way also hangs on in certain constreutions. it happens that the nominative and oblique are oftne alike: -taps for men (thus, -ʲidaps) and -puš for women. note that under the old sound change padagirm, it would actually be *paš for women, because /o/ always surfaces as [a], but under the new snd change parafdigm that /o/ is shifted to /u/ early on.

sound change errata and notes

remember that kʷ[V]ʲ at the end of a verb leads to /c/, because this vowel was always deleted.

past tense

past tense can be indicated by nasalization of the first consonant .... so /p t č k/ > / b d ǯ ġ/ .... and not by an infix. For fricatives, the nasal would simply disappear, meaning that for example serial verbs beginning in š- could not be marked for tense. Thus Moonshine resembles Poswa, but for a different reason.

true palatals

it is possible that hik > hʲk > kʲ is bad, since this would always come from earlier /hʷikʷ/.

Also, ph th > pʷ tʷ could make sense, since it would "explain" the gap of */kh/.

Specific ergativity

As promised, Moonshine marks masculine agents in sentences where the patient is feminine (animates only), showing which woman gave the man permission to act. At least a two-way distinction is present grammatically: -ṗa when the permissor is the same woman as the patient, and -ṗi when a different woman is involved. Note that these appear identical to the reflexive and feminine 3f verb endings, though only the latter is actually the same morpheme originally.

However, these markers are in fact an open class .... the /-i/ that marks a generic third party can be replaced by any word indicating a female patient, accounting for sound changes and sandhi. For example, -ṗat means "his wife"; the lack of final -i is due to the early formation of the compound. It is likely that there are also first and second person forms, unless this is somehow marked by possession as in Poswa.

The verb in such a sentence is marked with -a regardless of the permission marker, since the patient is the focus of the sentence, and the verb is historically reflexive. Note that this does not imply anything regarding volition.

Verb endings

The 3m->3f verb ending is the same as the 3f reflexive verb ending, and does not include the final masculine /-n/. all 3m->3f verbs can indeed be analyzed as passive verbs with the 3f as the subject (because subj=patient)

Model palatalization

The verb prefix hʲ- "above" (always from /hʷi/) can be added to transitive verbs almost trivially, indicating metaphorical superiority, acting from authority, and so on, and the resulting clusters like /śt/ will be pronounced somewhat like [sʲtʲ], leading to loss of the first element and thus a phonemic /tʲ/ and other consonants even in classical Moonshine. Note, though, that /tʲ/ could also be described as the aspirated counterpart of /č/. On the other hand, because these would contrast almost exclusively before /a/, it may just be that /tʲ/ is /č/.

The same is true of š- "below", but although they sound similar, the palatalization of /š/ cannot transpose across a following consonant.

[1] convenience link to sound changes

copy of consonant inventoru

Helicopter Consonants

Rounded labials:        p   b   m   ḟ   w
Spread labials:         pʲ  bʲ  mʲ  f              (ṗ   ṁ)
Dentals:                            ṣ       ḷ      (ṭ   ṇ)
Alveolars:              t   d   n   s   z   l   r  (c   ʒ)
Postalveolars:          č   ǯ   ň   š   ž   ʎ   ř
Palatals:               kʲ  ġʲ  ń   ś   y
Velars:                 k   ġ   ŋ   h   g   r̄

There are three kinds of consonants: palatalized, labialized, and velarized. Voiced stops may be all velarized, and velars may be all labialized; these are thus not marked. The rounded labials can also be considered labialized.

The velarization is perhaps not classic, as it includes /ǯ/ unless its effects disappear. But it seems it may shift a preceding unstressed i to u (ORTHOGRAPHICALLY). This means it is not elided but also not palatalized.

It is possible that the velar row is in fact labiovelar, and that an incomplete row of plain velars is also present. Thus the inventory would even more closely resemble a helicopter.

Gĭri (2021)

Gĭri is the children's sociolect. It may be descended from Middlesex's own children's speech register, and therefore its vocabulary will be mostly or entirely loans.

Jun 21, 2021

All Moonshine nouns have a gĭri equivalent, derived in formulaic way. The formulas are based on the inherited vocabulary and thus gĭri is immune to all sound changes. Verbs may or may not have direct one-to-one matches in g$iri.

Early June 2021

Children's speech in Moonshine is based on the Andanese model rather than the Poswa model, and thus does not resemble Poswa or Pabappa in having large sequences of labial consonants.

It may be that those few foreigners who are allowed in marginal areas of the Moonshine Empire (e.g. where habitats overlap) are allowed to speak the Giri idiolect, since it is in a sense a children's IAL, meaning that foreigners must bring small children with them, and communicate through them. This situation is likely created as a deliberate inconvenience, however, as any area in which foreigners are allowed to mingle with Moonshines is an area of relaxed speech mannerisms already.

Grammar is head-initial ... that is, "X Y" expresses "X of Y". Children learn to reverse the order of words when they reach the age of five and begin learning the full language. There is rarely any confusion because not only are the words different between the two registers, but the standard language also uses noun inflections to mark the genitive case, whereas Gĭri uses unmarked forms only.

Importantly, the second morpheme in a Gĭri compound is not represented in its standalone form; rather, the consonants harmonize with those of the head word, and vowels to some extent also change (probably just /i/ and /u/ harmonizing). This is in line with Late Andanese, but behaves differently because consonants and vowels harmonize separately whereas Late Andanese instead had syllable harmony. This harmony process is one reason why adults do not make use of a similar compounding mechanism in the standard speech registers.

Word roots that are composed of two identical syllables can perhaps be compounded with each other to produce new bisyllabic roots, as their basic forms are assumed to simply be monosyllables. Thus for example /baba/ + /pipi/ = /bapi/. Note, though, that this would therefore be an exception to the rules of consonant harmony.

Possible Gĭri phonology

Labials:       p   m   b   
Coronals:      t   n   l

The vowels were /a i u/ with no length or tone distinctions, and all syllables were open. Wide variation in pronunciation was allowed, and the syllables were perceived as being a subset of those of Late Andanese. This is why the /l/ sound was considered the proper representation of the coronal approximant; many children instead produced [w], [j], or silence. This is why Gĭri did not allow hiatus.

Because there are 18 syllables corresponding to the thirty of Late Andanese, it may be that Moonshines considered the true Andanese inventory to be 36, accepting the digraphs wa wi wu ya yi yu as part of the syllable inventory. However, little would be gained by making this connection. Alternatively, it could merely be that there were only five consonants in Giri, with a unified /b/~/l/ phoneme, and that this was the consonant that could optionally surface as hiatus.

notes o verbs

feb 28, 2021

verbs are not able to incorporate objects in any way, which stands MS out from all the surrounding languages .... they can only behave as if incorporating an object when "augmented", e.g. po-sal-t- from sal-t- "change clothes".

Singularities

nOV 11, 2020

no noun stem ends in -k or anothger velar consonant.

the only labials posttonically are /b/ and /w/

Sonority hierarchy

Oct 7, 2020

Initial clusters could be distinguished based on whether a vowel preceded or not. For example, /ask-/ would pull through as /sk/ but simple /sk-/ would have changed to /th-/.

Nov 10,2020

Additionally, there will be constrasts such as /čk/ > /šk/ but /čsk/ > /čk/.

Pàd

Aug 29, 2020

The IMS word for menstrual spirit is pàd ... Im noting for my own sake that this is not a pun, and that although it is derived from the word for blood (pà) it is not identical with that root nor with the inflected form meaning "bloodied". The stem is pàn, with a suffix -t.

Note that the Angry Womb spirit, list, is more powerful than menstrual spirits and probably should not be categorized as one of them.

Note also that it is a coincidence that this word's root, /lis/, is the same as the root of the pre-canonical name "Lispeb"; the original morpheme was lì kìha. Lispeb is an avatar of a different goddess who does not have menstrual spirits or an angry womb spirit.

The -t is the same as the -t on pàd, because this refers to a wider class of spirits that includes both the "low" menstrual spirits and the higher womb spirits.

Verb structure

aug30

clinic n-......for passive verbs. Also clinic c for some transitive . Pi+n+pi, etc

Aug 27, 2020

Perhaps final -ž on verbs could be analyzed as /Ø/, but this would contrast with true vowel-final stems due to morphological behavior.

Root-initial clusters

Aug 26, 2020

Many Moonshine words will begin with the clusters ps pš. This they share with contemporary Poswa (6700s AD); in later Poswa these clusters both shifted to /p/ while in Pabappa they were handled with conditional shifts, but never ended up as /p/. It is possible that at least one Moonshine daughter language will follow Poswa in showing /p/ as the reflex in initial position, but most would not.

These clusters form part of a set: ps pš c č ks kš. The clusters with /k/ are rare, however, because they can only arise from highly conditional changes. /ks/, at least, is common across morpheme boundaries. Note that the coronals are spelled with single consonants because they can occur word-finally, whereas the others cannot.

Assorted gender and noun class markers

Marking plurals

Aug 27, 2020

Plural marking might be defective for masculine and neuter nouns ... present sometimes, but not always. There may be suppletives for commonly used nouns such as "man/men". The ergative suffix -oh ~ -og, at least, is unmarked for number. The verbal infix -p- "to act on many" will also help, though it is likely that Moonshine is SOV like its neighbors.

Permission

Aug 26, 2020

Whenever a masculine or neuter agent takes a female patient, the agent must be marked with an ergative morpheme that additionally distinguishes which female gave them permission to act. This marker is -oh if the woman in charge is also the patient of the sentence (making it a reflexive verb), and -og if a different woman has control. Since /-g/ disappears before certain other consonants, it is possible that instead of just using /-og/ there will be different forms for each specific female party.

Note that the -oh/-og/etc morphemes are indifferent to number, and replace any -az/-as/etc stems, if they occur. They replace -er only sometimes because there is more than one -er morpheme.

Pondice

Aug 21, 2020

• saz "ice" ~ sog "pondice" ~ sagī "icy pond". Note that the /o/ is, as usual, an underlying sequence of /a/ + /ʷ/.
• foz "water" ~ fog "pond". This time the /o/ is independent.

The ice words may have a prefix /p-/.

Another feminine property suffix

Aug 21, 2020

Possibly e.g. fog "(the) lake" > fož "her lake", even though this merges with foz > fož "her water". Meaning that palatalization turns nouns feminine even when the final consonant is velar instead of just when it is coronal.

This may lead to a mixing of forms with final -k and -t in nouns for furniture, undersea objects, etc. Likewise with -z and -g.

Diminutive for small animals

Aug 21, 2020

The word for frog is k; this becomes ken "tadpole". This suffix thus is homophonous with the masculine property suffix, and thus male property incorporates a diminutive. (e.g. imagine if Spanish had su balón "her ball" vs su baloncito "his ball".)

Masculine gender markers

Aug 20, 2020

The masculine gender can be marked explicitly by an infix -(a)ř- whenever it would otherwise be identical in form with the corresponding feminine. This is the same morpheme as /-eř/ but was generalized from a stage before the shift of /aʲ/ > /e/.

One use of this is the singular indefinite form of some words, such as agent nouns, where the masculine is /-ři/ and the feminine is /-i/. Both morphemes previously ended in /-n/, and this nasal sound appears in some further extended constructions.

Subsyllabic head morphemes

Aug 15, 2020

A common trait of Leaper and PMS was the use of subsyllabic morphemes as the head of a compound noun, or sometimes an important part that is not the head. Thus, for example, in Leaper, lŏri meant flower and kʷŏri was a particular type of flower. Icecap Moonshine does this to an even greater extent, meaning that in some cases, the most audibly distinct part of a word is not the most syntactically meaningful. But Moonshine has evolved vowel alternations where Leaper and PMS only featured consonant alternations, mitigating this problem.

Religious terminology

Religious terminology is concise, as in the original 1994 draft of Moonshine. For example, the word for pray is /z~ž/, meaning effectively that it is a zero-stem, as this consonant is automatically inserted before nearly all vowel-initial words. And this word will actually be used in some cases. But Moonshines for the most part use more precise words, which are often one full syllable long when other languages would translate them as longer words or as phrases.

It is theoretically possible that this zero-stem word for pray could attach to the names of the three goddesses to make verbs for "pray to Malamen", etc, consisting of the verbal person markers attached directly to the names of the goddesses in the genitive or even the nominative case. This would likely use the clitic form of the genitive (s~š) instead of the bare form (which is also /s/ in most cases but not all).

Noun clitics

Aug 13, 2020

A dummy clitic could be used to preserve the noun case endings, which otherwise survive only in absolute final position. They would not come through in the same form, however, because the clitic would behave as though it were a vowel-initial word. Thus, where in final position one finds ` n s for accusative, circumstantial, and genitive, with the clitic one finds č ň s~š.

Children's speech

Aug 13, 2020

Adult use of children's speech varies by social status, and perhaps also geographically. Since many words are taken from Poswa, it may be that the areas closer to Poswob territory use more children's vocabulary. Alternatively there may be simply a "cute" territory where people talk that way regardless of their geographical location. It is possible that the children's sociolect was called /gĭri/ in early PMS and that this is also the name of the territory where it is spoken. The IMS reflex of this name would be ziř but it would need some sort of suffix.

Possibly add a dinminutive /-a/ to all children's words that ot herwise end in consonants, ... this is from Poswa influence and could spread to the main language as well.

June 13, 2020

Some Moonshine scholars may consider the entire Poswa language to be part of Icecap Moonshine's fifth register, spoken by children to other children. Since children at this stage cannot read, and almost never hear the proper Poswa language spoken to or around them, adults are required to bring Poswa words into Moonshine, and the body of children's slang changes constantly as it draws from its very large but finite pool of Poswa words.

Because of the way Poswa's grammar works, making every clause a word and every word a clause, IMS only needs to borrow the vocabulary to effectively borrow the whole language, at least according to the scholars. Children on the ground do not see the patterns in the Poswa words and merely think of them as ordinary words that for whatever reason are never used by adults.

It is possible that the children's words are deliberately spelled wrong in order to emphasize their nonstandard nature, or that they are spelled in Poswa's letters, or that they are represented with symbols. Or they may be spelled properly after all, if speakers rely on their sound to set them apart from the rest of the language. As children who speak the childish register in its pure form are too young to read, this choice matters only to adults representing the speech of the children they hear.

Words from the childish register are not used in adult speech to denote children's things. For example, pižupa is Poswa for school (in general), and small children will use this word for the schools they attend beginning at the age of five, only learning to say fič as they move up the grades. However, adults and older children do not say *pižupa for "children's school" or "elementary school". Instead, a teenager saying pižupa in any context other than quoting a child would in fact be signalling that she felt overwhelmed, naive, helpless, or in some other manner unfit for the school she was in. Because children use this register primarily when speaking to other children rather than to adults, when an adult uses it they are often seeking to connect with someone in a similar situation rather than with an authority figure.

Comment, it is possible that the kids' words are used in both senses, since e.g. an elementary school *is* a pižupa from the point of view of its students, who are referents in most sentences involving elementary schools. But if this is true, complex rules must exist to piece apart the context.

Rarely, children will have a word for something that has no precise adult equivalent. Many of these are related to social play, particularly for activities involving running and jumping that adults rarely engage in. Others are more utilitarian; for example, puč "to use a toilet sitting down", used by small children who are aware of their need to use the toilet but are not sure of what they will need to do when they get there.

Cultural aspects

The Poswob people are somewhat smaller than Moonshines, particularly those of Lenian ancestry, though not so much that they seem childlike. It is the speech and culture of the Poswobs that seems most childish by comparison to Moonshine, and Moonshines see the Poswobs not so much as a socviety of children, but as a society of adult babies. When a Moonshine speaker above the age of five uses the childish register in any situation where a more mature speech register is available, it is taken as a conscious decision on the speaker's part.

Vocabulary

The childish register encompasses Poswa but is not confined to it. Many childish words are native Moonshine vocabulary items that have come to be seen as unfit for adults for various reasons. Many are compounds for which adults prefer to use atomic roots. For example a small child might describe a swamp as a "wet forest".

As above, adults using the childish register are describing themselves, not the objects they are naming. The word blila "bottle" does not mean a small bottle, a cute bottle, or a baby's milk bottle, but may sometimes be used to describe a very big bottle, something so heavy that an adult has difficulty carrying it in their hands and feels unnaturally small by comparison. Childish words can also be used to describe strong people with a loving or protective role. Because childish words are nearly always longer than their standard Moonshine equivalents, they feel heavy and tend to dominate any sentence in which they occur.

When only a single word in an adult's speech is from the childish register, the intent is to mark out that word specifically and draw attention to it. When an adult uses more words, or even an entire sentence, in the childish register, they are consciously imitating a child, and this often expresses their emotions, whether positive or negative. Adults in a playful mood may produce many sentences like this, but adults who are frightened by the world around them will do so even more. Even here, subtle differences exist: adults in a happy mood tend to use the childish register in its most pure form, whereas those under stress will not struggle to remember all of the words and grammatical constructs they had used when they were young.

Like every other language, Poswa has words for adult concepts that children cannot understand. However, these words are not typically borrowed into Moonshine. When children need to use words for such concepts, they use descriptive phrases made from native Moonshine words.

Phonology

The children's speech register uses the same phonology as the standard language, and children are expected to pronounce words properly once they are physically able to do so. Thus a toddler who pronounces all of /p ṗ f f̣/ as [p] will be recognized as if they had pronounced the sounds properly, but an older child who does so will only be seen as mocking small children, which is strongly discouraged.

Children's verbs

If IMS has first person -o (present) and -i (past) in any of its many conjugations, this conjugation will be the first one taught to children and children will be expected to use it even when it is incorrect so that they can use the Poswa loanwords that require these endings. The other person markers may or may not match up, but this is of little importance, as it is expected that small children will talk almost entirely about themselves and their needs, mentioning other people only in regards to their ability to please the child. This is because children in Moonshine culture are allowed and expected to be very selfish.

Alternatively, children use only 1st person verbs, or use 1st person verbs in at least some contexts where adults would use other person markers... e.g. with evidentials "i see (someone) climbing a fence" vs "someone is climbing a fence".

Possible additional sound changes

Aug 10, 2020

It is possible that /s z t d n/ will all palatalize before any front vowels, neutralizing the distinction with the postalveolars. /ř/ will not, and it is not clear what to do with /l/. Consider Korean.

The vowel /o/ often occurs before a palatal consonant in a closed syllable. It is possible that this could shift to /e/ if a means can be found where this does not "over-palatalize" the final consonant and thus shift it to alveolar later on. If not, /o/ will remain.

Note also that palatal consonants rarely occur before front vowels, because they were "over-palatalized" into alveolars. Many words begin with /si/, but few begin with /ši/. However this could be repaired with massive compounding using the oligosynthetic C_C shells, or even just single-consonant prefixes, provided that shifts like /šs/ > /š/ are introducted.

Word-initial vowels

Very few inherited words have word-initial vowels, and they will be inherited with the low tone. In theory it is possible for other tones to arise from compounds, but these would never have single-glyph symbols because they were compounds. Thus, the alphabet has symbols for all of /ža žà žā žá/ but not for all of /a à ā á/. Thus sequences like /lā/ are spelled l-žā, etc. It may be that initial vowels and hiatus need ot be indicated in a spewcial way in the script.

It may or may not be that a sound change created new word-initial vowels due to effects at word boundaries. For example, /gi/ > /zi/ normally, but it could be just /i/ when following a word ending in almost any consonant. Perhaps this becomes generalized. It will not create hiatus however.

Mathematics

See Minor_languages_of_Teppala#Mathematics. This certainly does not still exist in the new revision of Moonshine, but it is likely that at least some use of the letter symbols appears in mathematics. Probably, separate numeric symbols will exist, however, eliminating the need to use letters as numbers.

Pronouns

Aug 2, 2020

Consider trying to move this offsite so that sections can be automatically sorted by date.

Verbs again

Aug 13, 2020

Use /p ~ ʷ/ as a plural verb patient marker. This could be represented in spelling by the same glyph regardless of its pronunciation. For example, there could be an alternation between ṛižāb ~ ṛipāb "hunt" depending on whether one animal or many is being targeted, but the spelling of the latter form would just use /žp/ where the glyph for /p/ is confined to this use and thus effectively logographic.

The fact that /ž/ > /p/, whereas most other consonants just velarize, is another argument for analyzing /z~ž~g/ as hiatus.

Aug 9, 2020

Feminist verb alignment

A possible shared feature with Palli is that bare-form nouns end in a gender marker, and when they are used as the agent of a verb, that verb ends in the same gender marker. Thus the particle could be analyzed as a clitic instead of an inflection. In Palli, the gender marker appears only once and effectively marks the end of the clause ... that is, in Palli an SOV clause is identical in outward form to just a noun. In Moonshine the same may be true as well, although perhaps Moonshine differs by also using a reduced form of the gender marker on the noun when it occurs internally.

In Moonshine, at least the post-verbal form of the clitic also marks tense and number, and therefore behaves like a normal verb marker. It also marks person since this is considered part of the gender system. In Palli, the morpheme was "defective" and simply stayed the same regardless of its referents.

Target

Aug 4, 2020

Ideal target verb inventory:

LIKELY INHERITED FORMS

• 1st person singular, with a transitive vs reflexive distinction and no gender marking (discourse markers handle this)
• 2nd person singular, with a transitive vs reflexive distinction and probably no gender marking (discourse markers mark both 1st and 2nd person, but it is possible that 2nd person verbs will provide a finer distinction)

LIKELY INNOVATED FORMS

• 1st person plural exclusive, with a three-way transitive/reciprocal/reflexive distinction. Probably no internal gender marking because all plural forms were epicene in Gold, but could have innovated gender markers descended from "pronouns" that mark out the non-1p participants. (i.e. a woman would say effeectively "we men" to mean "the men and I", etc.)
• 1st person plural inclusive, same orientation as above.

• 2nd person plural. Possibly uses third person agreement, unless 1st person and 3rd person are the same. Gender marking is external.

• 3rd person singular masculine.
• 3rd person singular feminine.
• 3rd person singular neuter. Sometimes used for small children, but mostly used for animals. Unlike Gold, neuter is considered animate, so there are no inanimate objects here .... inanimate agents can only be used with passive verbs, and passive verbs probably do not mark the agent.

• 3rd person plural epicene. Gold had no gender marking in the plural, so this could possibly be inherited, but it is likely that the inherited form would have simply used zero-marking, and thus could not be the direct ancestor of the Moonshine form.

Gold marked either the agent or the patient but never both. Moonshine might find a way to mark both at least in the third person.

Aug 2, 2020

There may be three conjugations instead of so many as below. Of these, only one will be truly productive.

First person active endings could be à ì ù for the three conjugations. Second person active endings could be Ø ʲ ʷ for the three conjugations. First person passive endings could be ā~o ʷi~ʷj ū for the three conjugations. Second person passive endings could be an in un for the three conjugations, but the final consonant would only persist if the verb was placed in absolute final position.

The gluon verb markers

It may be that Gold lacked pronouns and person marking on verbs, using only tense, as Japanese. However, the distinction between transitive and reflexive~passive is robust, as is the distinciton between "feel" and "see" evidentials. The 2x2 matrix of these two markers is sufficient to generate all ten person combinations (1:1 1:2 1:3, 2:1 2:2 2:3, 3:1 3:2, 3:4 4:3) because the number of arguments to the verb makes clear which party is which.

These markers could perhaps be called gluons because of the way they are able to make ten (or nine) distinctions from just two parts.

There is also a reciprocal /-s/, which is not needed but exists anyway. This is probably merged with the passive/reflexive suffix /-k/ in Moonshine fairly early on, but its meaning may merge into that of the reflexive.

May 24, 2020

The Poswa strategy of attaching person markers directly to the genitive case of a noun to indicate "I, a ..." is due to Play's unique sound change /sb/ > /s/, the presence of three evidential morphemes that began with /b/, and a later sound change of /sb/ > /b/ that grammaticalized the first one. None of these were present in Gold and therefore none of them are likely present in Moonshine either, unless an astonishing coincidence took place.

It is possible that Gold and its non-Play descendants will have to use the essive case for this function even though the original intent all along was to use the genitive. ("The genitive and member endings are the same.") Try to find a way to use the genitive so that the later use of /-l/ instead of /-s/ for genitives can be more easily justified.

Note that if the essive case were used, the 1st and 2nd person forms (that is, "one of my ..." and "one of your ...") would have been /-ča-/ and /-hia-/ in Gold, and thus still distinct, but would have merged into /-ša-/ before the proto-Moonshine stage. (NOTE: But they were still distinct in Leaper!!) Only the 2nd person form is productive in Moonshine, but if the two had merged before the breakoff of Moonshine, Moonshine must have gotten its 2nd person form from somewhere else.

May 23, 2020

Female speech in IMS lacks pronouns entirely, just as its parent language did. Only men use pronouns, and there are quite a lot of them to memorize, each tuned to a specific social situation.

Men's pronouns for themselves are often grammatically 2nd person, e.g. "your listener", when addressing a woman. Then, the obedience markers stack on top of this, and determine whether the verb in the sentence will use 1st, 2nd, or 3rd person agreement. (It may even be that men cannot use 1st person at all.)

This explains the confusing statement below that male speakers say /oḍ/ for /as/ "whenever the /as/ is 2nd person" ... because sometimes for men it can be 1st person.

• ṭu-š-as ~ ṭuš(a)- "your husband", from Gold /tuhʷuʕ gahas/. This form is identical to its own genitive, unless analogy takes place. This is also the case with ordinary 2p feminine nouns.

Note that according to writeups below, men cannot ever actually use this pronoun in bare form because it implies #Involuntary obedience, and therefore its existence is just a construct of the grammar.

TO DO: Need to figure out how to say things like "your admirer", "your listener", etc in Gold. The Poswa formula cannot be used exactly, but something similar may work.

Possibly just "I listen to you" with a genitive marker on? That or essive?

Padded accusatives

High-tone CV verbs attach to nouns and replace the accusative endings. e.g. là "buy/sell" appears as -l(a)č- on nouns of the handheld class whenever the verb relates to exchange. These high tones were treated as if they had always ended in /č/ and thus came to be seen as infixes with the accusative at the end; then, they replaced the accusatives.

Note, though, that there is no tradition of converting high tones on verbs into /-č/, and that the window of time in which this could have evolved in PMS is fairly small.

Reanalysis and unique grammatical forms

Aug 6, 2020

The locative of feminine nouns ending in -ēy is -ey, which is pronounced the same (both are just /ē/) but spelled differently in the script. The locative may be the base for many constructions such as illative, allative, etc as in Poswa.

July 23, 2020

Masculine nominative -oh, generalized from the inherited nouns /noh/ and /n/, both meaning "man". Thus, analyze /n/ as the basic masculine morpheme, instead of /noh/. Masculine nouns thus cannot have an unmarked free-form nominative. No other Gold language has a marked nominative form, so this cannot be placed neatly into the inherited Gold noun declension system. It stands to reason that Moonshine will need a table layout all its own, not patterned after those of Gold, Leaper, Poswa, etc.

The masculiune nominative /-oh/ changes to -od ~ -oġ-, effectively an ergative, whenever there is a feminine patient, even an inanimate one. The vowel after the /ġ/ corresponds to the patient. This is exactly like Poswa's passive verb system, and it could be said that in IMS all sentences with masculine agents and feminine patients are passive. (But there may be a secondary true passive verb paradigm, such that Moonshines simply learn the -oh/-od/-oġ alternation and give it little thought.)

Fusional plurals

July 23, 2020

Possibly evolve deep fusional plurals like IE. This would set apart Moonshine from most or perhaps all other Gold languages, as the plural was always segmentable even when it affected surrounding affixes. Rather than affecting nouns directly, they will attach to the inherited classifier suffixes and some elements such as -č~c that came to behave like classifier suffixes. Thus all nouns are "weak nouns"; not even the very most common nouns have stem ablaut to show plurals. This is because the stem vowels are already changing for other reasons and the two patterns would interfere. There may be a few suppletive plurals, however.

Animate plurals may be fixed as the historical partitive plus -d. Thus /tad/ "man" > /tadīd/ "men", and so on. Case markers would go outside the plural marker rather than fusing entirely with them as in IE. /dč/ > /ǯ/. The /ī/ reflects the partitive of /a/, but could be generalized to other nouns by assuming a dummy suffix ending in /-a/, as occurs elsewhere. Alternatively, the choice of vowel could reflect something else.

The other two partitive vowels, if they are used at all, would be /ʷ/ and /e/. But use of /ʷ/ will lead in many cases to deletion of a preceding consonant (e.g. tʷkṅdu > kṅdu, as if the /t/ and /ʷ/ had never been there).

The choice of tones is nearly arbitrary due to analogy ... ī above could just as easily pull through as í or i, and possibly even ì.

Collective nouns in -ž can form counting plurals in -nec.

Overview of morphonology

Although superficially it may seem that both Icecap Moonshine and Poswa have greatly changed from the Gold language, in fact Moonshine changed much more rapidly. Poswa's grammar is very similar to Gold, and much of what has changed has been repurposed rather than discarded. For example, Poswa's person markers are derived from evidentials.

Also note that Poswa has a longer maturation time (1700 AD to 8773) versus Icecap Moonshine (1700 AD to 6843).

Tone is lexicalized to a much greater extent in IMS than in Leaper or PMS. a à ā á are seen as four separate vowels, not one vowel realized on four tones. Tone-changing alternations are common, but vowel-changing alternations also exist, and the Moonshines considered these morphological changes to be two of a kind.

Sound changes

Aug 2, 2020

Possibly /aʲk/ > /ek/ fails to occur, since it would leave a gap or near-gap of /*ak/. Alternatively, the shift happens and then reverses, meaning that the much rarer inherited /ek/ would also shift.

Jul 18, 2020

The voiced dentals ḍ ẓ might just be ṇ, since ṇ was post-stopped all along and in this position a fricative would not have been distinct from a stop. Likewise, in most or possibly even all cases, ẓ would have required a preceding nasal to come into being.

The same applies to ḅ v.

Compound mutations

The existence of noh "man" as the second element of a compound means that ALL inanimate nouns must undergo prenasal mutation, such as /ces/ "sword" > /cé/, making the final consonant unrecoverable. In bareform it will either be generalized to the tone only or take a unified consonant such as making the acute tone always -s, grave always -c, and long always -l./

Stamp script

Scholars write Moonshine in a script resembling something on a scale between modern emojis and ancient Egyptian hieroglyphs. Phonetically, it is much like Egyptian, since few people use emojis phonetically, but also uses many ideas that Rgytpian never did. For example, vowels, tones, and consonants are all obligatory.... nothing about a word's pronunciation can be inferred from context. But IMS resembles Egyptian in many ways:

1. IMS uses redundant spellings, such as marking tone on both the vowel and the following consonant (teachers sometimes describe tones as being sequences of a vowel plus a tone morpheme, but the vowel always comes first). This can be called a fertile orthography.
2. IMS has many choices of symbol for each sound. This actually harkens back more to Japanese hentaigana than to hieroglyphs, though it reminds me of hieroglyphs too because Egytpian had pairs such as /k~ḳ/ that seem homophonous from an English standpoint. Moonshine's homophonous glyphs will probably not be derived from previously existing contrasts. e.g. in pre-PMS there was a contrast between /k/ and /ḳ/ just like in Egyptian, but these did not retain separate glyphs in IMS. Rather, the separate glyphs in IMS represent newly evolved distinctions.
3. IMS sometimes uses grammar to choose which glyph to spell a sound with; this is most common at the end of a word.

Introduction

There may be an ornate script resembling Dreamland's syllabary, the difference being that the symbols here stand for consonants, vowels, and consonant pairs (including C_C shells) rather than for syllables. The single-consonant signs could perhaps be analyzed as doubles in which one member is the silent consonant /ʕ/, and where a preceding or following vowel is really between the two "consonants".

It is likely that morphemes will play an important role in determining a word's spelling. Codas recognized as freestanding classifiers will be spelled with separate glyphs rather than using a C_C shell glyph. But in words where the initial consonant is seen as belonging to the classifier, they will be spelled with one glyph. For example, n_ň right now is listed as meaning "pocket" where the medial vowel indicates where on the body the pocket is.

If a word contains no classifier at all, it would be spelled with a C_C shell glyph combined with a vowel. For example, sap "fish" would be spelled with s_p and a.

It is possible that certain VC sequences could be spelled with individual glyphs as well, on the basis that, for example, the common masculine suffix /-en/ attaches to words ending in /-a/ and can be thought of as /ʲ/ + /n/, or even as freestanding /y/ + /n/. (Historically /ay/ > /e/, preserving tone and length, though there were also shifts that caused lengthening.)

Spelling may be highly redundant, with tone marked both on the vowel and the final consonant, if signs like ʔ-p (grave tone with /p/ coda) exist. These would be considered two-consonant signs. More complex signs would be available, such as _́-y-n, that is, an /n/ following a high tone whose vowel is grammatically fronted, but these would still be considered two-consonant signs.

These symbols would be written with "stamps" since there would be hundreds of them and they would have complex shapes.

Objects depicted

Like emojis, there are depictions of both objects and people, and they are of similar size, meaning that for the most part the objects appear comically oversized next to the people they are depicted with. Sometimes, though, a glyph depicting a human can be replaced with just that person's head.

Glyphs depicting humans might be used only for consonants, and will mostly occur right after the tonic vowel, where they indicate grammatical functions. Rarely if ever will a glyph depicting a human stand in for a content word or even for part of a content word, even if that word means "woman", "man", etc .... one possible exception to this is single-consonant standalone morphemes such as t "son", if they do still exist.

Sources of difficult glyphs

Because Gold short monosyllables typically had voiced initials, there were few or no words such as *pă, *tŏ, etc which would have generated IMS standalone glyphs for voiceless stops like /p/ and /t/. These thus must either come from the pharyngeal words (tâ, pâ, etc) or be analogized from other words. Note that the pharyngealized words would have mostly appeared at the end of a phrase, not the beginning.

Likewise, most of the vowel glyphs will need to arise by elimination of preceding consonants. For example, there can be no source of pure ō ó because words like Gold /ʕuk/ "flower blossom" would have reflexed into PMS short /o/ which later became IMS short /a/. And ò can only come from à followed by a labialized consonant.

It is likely that IMS simply has no glyphs for /ō ó/, and instead uses its glyphs for /ā á/ followed by a labializing consonant or a standalone /ʷ/ glyph (which is likely the same glyph as /ṛ/).

A few words might begin with /ař-/ and /oř-/, corresponding to Leaper /âr-/ and /ôr-/.

No words began with /ř/ in PMS, and it will be a rare sound word-initially in IMS. But it is likely still present in at least a few words, and those words would be short ones because they would only have arisen from loss of initial vowels. Even so, it's possible that all of the /ř/ glyphs are just repurposed /l/ glyphs.

Some other difficult glyphs are:

• The alveolar affricates c ʒ (since they typically occur before front vowels, and these cannot be used to make standalone signs)
• The dentals ṭ ḍ ṣ ẓ ṇ, either with or without a preceding /ʷ/. These are difficult to generate only because they arise from rare sequences. However, they could all arise from disyllables like /tăpa/, /săma/, etc.
• There may not be a k_ʔ glyph. (That is, /k/ before a high tone.)

Relationship with alphabet

Moonshine uses an alphabet, but there may be hundreds of superfluous symbols similar to the hentaigana of Early Modern Japanese. As well, there may be ligatures that stand for consonant clusters.

There may be two symbols for every consonant, depending on whether or not a vowel immediately follows. This would allow a grammatical distinction between /Vn/ and /nV/ ... that is, monoliteral roots fall into two classes, depending on whether their accompanying vowel goes before or after the root consonant. These can also be analyzed as ʕn ~ nʕ, since they behave as if they were biconsonantal roots with one of the consonants being silent. This is how Moonshine will spell them if the two sets of consonant letters are not employed instead.

It is not clear how roots should be categorized. Etymology may not always help, as for example the roots for pine tree, apple, and curtain are all just ž and all can be equally well characterized as either ž_ʕ or ʕ_ž since they were already single phonemes in pre-PMS.

When such homophones pile up, one will always be considered the categorical word, and the others "homophones of ___", etc. The categorical word for each single phoneme may determine the name of the letter as well, if a longer word with similar meaning exists.

END OF STAMP SCRIPT INFO

Saying please and thank you

June 13, 2020

IMS retains the PMS/Leaper dual imperative setup, where ṅ means "please" and ṅt indicates the standard imperative. Thus in Leaper and PMS (and possibly even in Gold) the polite imperative was actually shorter than the standard imperative. IMS inherits this system, in the sense that its standard imperative is probably a simple suffix -d in all environments, and although it has a highly complex system of polite imperatives, these build on a suffix -n, not *-d. (These may both lengthen any preceding vowel.)

The polite forms of the "please" formula are extremely complex for both men and women, and there are many situations in which women are required *not* to use a polite form. Small children probably use a simple form, possibly consisting of just the inherited suffix /-n/. In most speech registers, there is no standalone word that means "please", as this is expressed with a variety of verb inflections. In very polite speech, both the inflections and a series of standalone words are used.

The Moonshine "thank you" is a sentence by itself, although like English it can be attached to another sentence in various ways. The forms of "thank you" are also extremely complicated, depending on the gender and social status of both partners, whether the action has been completed yet or not, and on what that action is. Small children are expected to be polite here, unlike in many other aspects of their speech, because the expression of thanks always involves addressing another person, usually an adult, and therefore the simple, self-focused children's register is not appropriate. Children therefore use different forms of address depending on whether they are talking to a man, a woman, or a child, and also different forms for close relatives, distant relatives, and strangers.

Moonshine culture expects the recipient of a favor to always offer some form of thanks, even in the extreme cases where a woman of high status is being helped by a man she does not know and would ordinary say few or no words to. In such cases the expression of thanks may serve primarily to let the man know that the woman is satisfied and does not require him to perform any additional tasks for her. Between social equals, the expression of thanks is used more freely and can have a wide scope of emotionally significant meanings.

Honorifics

July 20, 2020

see wikipedia:Nyōbō kotoba

June 11, 2020

In a radical departure from the grammar of other Gold languages, perhaps IMS's honorifics could be prefixes. This would require them to have formed early on as PMS free words.

This seems likely now. They are derived from the genitives and accusatives of freestanding nouns.

June 7, 2020

From a dream I had on June 6, I figured Moonshine could use honorifics "to show how great something is". These would pattern like the expressives of Poswa, showing the speaker's love and admiration for something, meaning that they are in theory all optional, but some speakers .... probably men .... are expected to use them in every single instance when referring to certain nouns .... probably their wives and any other females close to them. They would not become *truly* universal, however, since that would mean they'd lose their meaning ... if this does happen, a second layer of honorifics will need to be generated to replace the first.

Unlike the Japanese, Moonshine speakers are not particularly humble, so there is no prohibition against a speaker using an honorific to denote themselves, although since Moonshine's feminine register lacks pronouns, it might be restricted to formulations such as "we great council members" where the speaker is part of a larger group.

The choice of how to structure the honorifics depends on what is chosen below in the #Early changes section.

Speech registers

All four adult speech registers use honorifics. Even men using the typically vulgar male/male speech register use honorifics for women, but they use them freely instead of obligatorily, and also have some pejorative gradations. Essentially they use the honorifics to express their emotions and opinions of the women around them.

Women also use honorifics, because, as above, the Moonshines are far from humble and their women see no problem describing themselves as "great", "beautiful", "powerful", and so on. Women also use honorifics of a different type to describe their husbands and other males among their friends and family. It could even be that IMS has no word for husband, simply using the word for "man" with a diminutive suffix on. Pejorative suffixes for men also exist, but are rarely encountered, as women see all words for males as being pejorative by nature unless padded with a hypocoristic affix.

Early changes

June 6, 2020

Between Gold and Play, the original three options for how person markers attach to roots were likely lost and then rebuilt. Play has a three-way distinction between "I have an ___", "I am an ____" and "I use an ___", all without using any of its noun cases ... it instead uses different stems of each root, which in Poswa evolved into the B-, C-, and D- stems. If Gold similarly had a three-way distinction, it must have been formed in a different manner, because the C and D stems did not exist in Gold and its B stem could not perform all three functions. Neither could the bare root be used here because it often ends in a consonant and the Gold person markers were also consonants.

It's worth noting, however, that the Play formulas are built on noun cases .... "I have a" comes from the locative (sic, not circumstantial); "I am a" comes from the genitive with an additional Play-specific padding morpheme; and "I use a" comes from the bare root but with another Play-specific padding morpheme.

Note that PMS has an "I have a" form that is built on the locative case; this may need to change to circumstantial. The locative case could have been chosen due to reanalysis of the accented /g/ as being the produce of /gg/, though, so perhaps it is valid after all. If this is chosen, then "have" would need to have been originallt an intransitive verbm, since it has no accusative marker, merely the oblique.

Weak verb conjugations

Strong verbs in Icecap Moonshine each have their own declension, and therefore the name of each declension is simply the name of the verb. There are only at most a few dozen strong verbs in the language, and probably less than that. The weak verb conjugations are just as difficult and unpredictable as the strong ones, but it is much easier to tell which weak conjugation a given verb belongs to.

Womb

The womb conjugation contains verbs whose stems end in -š, which can mean "womb" or a wide variety of other things. Nonetheless, the womb meaning predominates, and therefore this conjugation can only take a feminine agent. Even in a passive conjugation, there cannot be a masculine or neuter agent. This is however a syncretic conjugation, combining elements of the word that originally meant womb with other morphemes that also collapsed into the single consonant š.

The womb elements provide the 1p present and past tense morphemes, which are p and f respectively when occurring after a vowel. That is, the stem /š/ becomes one of /p f/. When occurring after a consonant, however, both merge into /w₂/ ... that is, the sound is not pronounced, it merely mutates the preceding consonant, and in some cases it has no effect even on that. Unlike the more common /w₁/, this sound does not bleed onto the preceding vowel either.

Womb verbs include verbs for childbirth, menstruation, and so on .... actions that require an agent with a womb. They also include all other inalienable feminine nouns, however, so a woman who uses her hands, legs, and so on to perform an action still uses a womb verb. Notably, even though it refers mostly to inalienably possessed objects such as body parts, this inflection paradigm does not include a reflexive marker, so the literal semantic meaning of the affix is "with a womb", not "with her womb", and so on.

This group also include some alienable feminine nouns, mostly for words in which an earlier feminine /-ž/ suffix came into contact with a voiceless consonant and became voiceless itself. The proper /ž/ conjugation is nonetheless quite similar to this one.

Womb words also include all verbs derived straightforwardly from nouns ending in /-š/, regardless of the meaning of that /š/. As it happens, most such words describe feminine verbs such as "give birth", and so on. But there is also the subtype including verbs such as leš "set on fire" and liš "bite into", which have been reanalyzed as feminine verbs as though they involved the use of the agent's womb just like childbirth. Because these words have a vowel before the /š/, they conjugate in the simple pattern. Thus "I (a woman) did set it on fire" translates as lef.

Bite

The bite conjugation consists of verbs whose stems end in -z. They typically take neuter agents, and typically describe actions that only animals can do, such as biting through solid objects, swimming deep underwater, and flying through the sky. Moonshine does not assign human agents to actions like these even metaphorically, although the feminizing palatal morpheme ž can be suffixed to the verb stem (and /zž/ > /ž/) to create derived forms which can take feminine agents. These are not grammatically considered to be the same verb because the stem has changed from ending in /z/ to ending in /ž/, but the meanings are typically closely related.

Pine

This category describes verbs involving feminine property that is not part of the womb conjugation. It is typically inalienable property. Stems end in /-ž/. This /ž/ can either occur alone or be suffixed to a preceding consonant. It is the oldest specifically feminine category in the language, preceding even the womb verbs.

The canonical name for this verb class is actually "vagina", but Moonshine teachers introduce it with a different name when teaching overseas so as not to alarm students who have not yet learned of the Moonshine women's plainspoken manners.

Sap

Used for verbs derived from male property; largely the masculine counterpart of the Pine conjugation. Stems usually end in -en, and always end in -n. First person, which is always masculine, is perhaps marked by a shift of /n/ > /ġ/. (This is due to nni > nəči > ŋʷši > nkʷi > ʲnġʷ > ʲġ. The ʲ is already part of the /e/.)

BUT NOTE THAT PMS MAY NOT HAVE SHIFTED /čʷ/ > /kʷ/ AS DID LEAPER. THAT WOULD MAKE THIS INTO ʒ.'

Noun declensions

Icecap Moonshine likely uses declensions, similar to Future Poswa but unlike classical Poswa. The declensions may cut across gender lines, though perhaps masculine nouns will need to take a suffix that unites them into a single declension of their own. Even if this is true, it is possible that some very common irregular nouns remain outside this declension.

Many nouns form their accusative by adding /č/. These can perhaps be all grouped together into a "trivial" declension regardless of what their root-final consonant is, though it would be best to see if they also have similar ways of forming their other stems.

For example, sùsa "female farmer" adds /-č-/ for its accusative and /-ň-/ for its circumstantial.

Declension 1

š "feminine property". Used for female inalienables, women's clothes, and certain other objects considered to be inherently feminine. Accusative k-, circumstantial ġ-. There is also a dative, either /h-/ or /f-/, which could serve as the genitive. The inherited genitive would probably merge with this in EITHER case. (The question is whether pre-PMS /xʷ/ or /hʷ/ dominates in a cluster comprised of both.) A second accusative form in /-p/ may be used for when the agent and patient are both third person feminine.

If agent is male, the accusative remains /-š/, because it is not etymologically an accusative at all.

Declension 2A

-ž "feminine property". Often follows a noun that already ends in a consonant, and therefore changes that noun's declension class. Expands to /-g-/ when a vowel follows. It is likely that when a padding consonant precedes, the accusative of this is -k-, circumstantial is -ŋ-, and there is also a commonly used instrumental and a commonly used /s/ genitive. The instrumental is probably just /g/ again, therefore indistinguishable from the nominative. The genitive is probably /h/. Thus all four are velars.

Declension 2B

-Ø "feminine property". Used in only a few inherited nouns; etymologically identical to 2A but found in words where the final consonant was eliminated by sound change.

Declension 2C

It's possible that a third pattern exists: ž/č/ň, which would make it almost identical to the zero suffix. This is more likely to be the third person feminine "named possessor" form, however, and it is never used with inalienables.

Declension 3

noh "man", alternates with nó. Accusative nok-, circumstantial noh-. Thus, this could be patternized as h/k/h. This particular noun is masculine, but there is no masculine suffix on this word, so it shares the declension with neuter and likely also feminine nouns.

As with declension 1, the accusative is /-p/ if the agent is 3rd person feminine, and possibly also for 3rd person in general. It may be /d~ʒ/ for the circumstantial.

Declension 4

lař "boy". Accusative řak-, circumstantial řaŋ-. Stem contracts as lař > lř > ř. Possibly grouped as a subset of Declension 2. Gender-neutral but possibly most used for masculine nouns.

Declension 5

ṛòt "boy". Accusative ṛòč-, circumstantial ṛóň-. Possibly part of the trivial declension as tč > č in all environments and /ň/ may come to be seen as the generic circumstantial ending.

Declension 6

-en "male property; made for men". May need to be broken into three subtypes. Primary form has accusative -eʒ and circumstantial -en. Note the lack of palatalization.

Because of Moonshine's gender asymmetry, this has a feminine version, likely marked by palatalizing each form so that they appear as -eň, -eǯ, -eň, etc. The -en affix does not mean "belonging to a male" but merely "made for a male", meaning that a woman can also own such an object. By contrast the feminine inalienable affix -š is restricted to feminine possessors, meaning it has no masculine or neuter (or free) forms.

New vocabulary

May 31, 2020

Leaper gṅĭ "penis" is likely just a masculine inalienable possession marker in PMS. Leaper's feminine version of this, gʷî, could mean clitoris, urethra, or even breast in Leaper, or it could be unused, but in PMS it would likely appear at least somewhere as an inalienable feminine marker. It seems that it would merge to /ž/ or even /z/. Both of these are derived from ṗ "eye" and thus are euphemisms in Leaper, which is why their meanings are so broad in PMS.

The Icecap reflexes of these morphemes would likely depend on environment .... /gṅĭ/ might be just the familiar "high rising n" or it might need to retain its final vowel, while the feminine version might be /g/ or /r̄/ or it might retain its final vowel as well.

May 30, 2020

Leaper's word for moon could be kʷô (tone??) instead of having /ġʷ/, by adding /hʷ/ "sleep" to the onset. This could happen in PMS too, leading to IMS having just a k for this root.

Romanization

May 29, 2020

Try to learn to use one of r̂ ṛ r̄ ṙ to represent the uvular fricative so that the plain r can be freed up for the much more common alveolar trill (currently using ř). Another possibility is one of ħ ḥ, since it is voiceless as often as it is voiced.

Categorization of nouns by gender is not the same as nominal possession

May 23, 2020

The feminine marker /-ž/ (from /gə/) is not a 3rd person feminine possessive marker, it merely means "of women". Thus a new 3rd person form must be found. This /ž/ also appears as palatalization of a noun whose stem ends in /s/. There is a masculine form, descended from Gold /ŋùni/, which became B-ʷ-nni at some early stage of the language and thus could appear as -ʷ-n (although apparently the /ʷ/ is suppressed too, therefore /en ~ oni-/) in IMS, through it would be "volatile" because of the suppressed syllable. If there is still a masculine possessive marker /-t/, it goes on the outside of this morpheme, and therefore the suppressed vowel reappears. This means that, contrary to below, there may be more than one masculine stem in the language after all ... though there may be still only two masculine morphemes. Females can still take possession of objects with the masculine category marker on, but men cannot take possession of categorically feminine objects (and indeed there would be no place to put the masculine marker /-t/).

Nouns categorized as masculine include male body parts (Gold /ḳaunni/ "penis" > cen; /halaunni/ "testicles" > hen ). Possibly, non-distinctive body parts also change form, so that e.g. the word for thigh could derive from Gold /dă/ in bare form, from /dă gə/ for a woman, and from /dă ŋùni/ for a man. The proper Gold forms of these would be dă dāʕ daunni, and in Icecap Moonshine they would evolve to l(a) l(a) len. Probably, the feminine form merges with the unmarked form in a great many forms, perhaps the majority. However, because it requires an accent shift in Gold in order to hold the superheavy syllable, CVCV forms will typically not merge.

This development implies that the words for body parts that derive from CVCV roots will come to begin with consonant clusters in IMS much of the time. For example, kidney will become psi for a female (the male form in Gold would have been /patənni/, which would thus develop a syllabic /ṅ/, and would thus turn into an irregular such as /pén ~ páni-/ in IMS if it is not repaired with a dummy morpheme).

The Leaper word for eye is simply ṗ; this would be preserved into PMS and IMS alike as p and could appear in both its literal usage and metaphorically, similar to the /-v-/ evidentials of Poswa.

It is possible that the body part suffix /-m/ will be needed here to prevent collapse of the many single-consonant body part words; this would allow an additional vowel to be inserted. This is likely at least for words used in their bare form. However, as below, body part words are sometimes used after verbs to indicate the means by which an action was performed ,either literally or metaphorically.

The sequence -en-t- will probably always resolve to -ed- regardless of whether the vowel rotates to /o/ or not in the unappended forms.

Possible gender-neutral instrumental verb markers

Gold /gakū/ > -čo- "using my" Gold /gahʷū/ > -h(ū)- "using your"

The 1st person might contract to /k/ before a vowel just as the 2nd person contracts to /h/. Thus, in a roundabout way, the pre-Gold person markers /k h/ reappeared.

These can only "reliably" be used with nouns whose stems end in a vowel. But it is likely that analogy will lead to many nouns being used as if they had zero-morph suffixes. For example, body parts. There was probably a dummy noun such as /gà/ that contracted to /Ø/ before the suffixes so that they can be used in IMS without having to think up what the historical oblique forms of each noun would be. Thus for example, /tā/ or even just /ta/ can serve for "telepathy" instead of having to work back to the Gold oblique form /taha-/. The word for telepathy is basic, and the word for brain is not, so the word for telepathy may end up being the word used to refer to actions that came about because of detailed planning.

Masculines might need to insert -ʷni- here, on the basis that the feminine marker collapses to Ø but the masculine marker must retain at least a consonant (/n/) and possibly also a vowel. It may even cause vowel lengthening of the preceding stem unless this is analogized out.

Remember that these person markers do not need to match the person markers of the verb itself. These are equivalent to Poswa's -ibo/-ube/-oba instrumentals. On the other hand if they become integrated into the stem, then they will match and the alternation will be seen as part of the stem-changing properties of the person markers rather than as part of the person markers themselves.

Feminine after all

Body parts are likely to appear commonly here in metaphorical meanings. For example, though they may seem vulgar, /gì/ "vagina" and /xʷ/ "womb" are likely to be common verbal suffixes which create feminine verbs. They are highly polysemic; e.g. /gì/ also means "flask; canteen" and this is not a euphemism.

These may pattern like the /z/-stem neuters, but at least womb already contracts to just a /Ø/ in the 1st and 2nd person (3rd would collide with 2nd unless it is restructured). And vagina is just a /ʲ/.

If there are three feminine types, they may be called breast/belly/womb, and will encompass mergers of many more morphemes, perhaps one for every consonant in the language, into three patterns. "vagina" is ʲk/ʲh for 1p&2p, but perhaps even the ʲ is to be reanalyzed as part of the stem and this thus merges with womb in at least 1p&2p.

I dont quite remember what I was doing here, or how I managed to derive a /ʲk/ morpheme, .... perhaps it is just /ʲ/ + 1st person /k/.

If there are at least two patterns, they could be m/ġ/f̣ "breast" and š/k/h "womb". Note that these are free/1p/2p, and 3p is not listed here. The so-called breast verbs might be described as the "forward conjugation", because the root morpheme for breast appears in words for frontward motion. Following this pattern the womb verbs might be thus described as the "central" or "interior" conjugation.

Other body parts can be used like this, but most do not have three distinct forms. For example, the word for eye appears as /p/ always, and cannot be used to distinguish between different persons.

Neuter verb markers

May 22, 2020

The neuter verb marker -z- mentioned below is probably still valid, but it may not really belong to a separate conjugation.... it depends on analysis. Such verbs are rarely used with human agents, but it may not be so far evolved as to be ungrammatical. This means that neuters can only be the agents of verbs that end in vowels and perhaps a few consonants, to which the /-z-/ is added. This puts them in the same category as men, though perhaps a bit higher.

Masculine verb markers

May 22, 2020

Gold /dagə/ > IMS -řg- (-rž when final) "to use a man"; likely also the masculine verb marker. Must come before a vowel, but not necessarily after one. Note that we already have ṭ for this exact function and they may be suppletive.

If all male agents must use this morpheme, this means that males can only be the agents of verbs that end in vowels, and are potentially even lower on the animacy hierarchy than neuters.

The feminine verb marker is Ø, but there may be some suffixes anyway for formulas such as "she uses her ____ to ____". In other words, from IMS's perspective a man is just one more piece of feminine property.

Another possible feminine property marker is Gold /gʷū/ > IMS -gū-, which apparently turns into just a /g/ by some kind of analogy. ( Perhaps gʷū ~ gʷuʕ- ~ gʷug- ?) This would mean that effectively the masculine verb marker can be analyzed as /-ř-g-/, consisting of a unique suffix before the feminine property marker, again implying that adult males are merely one particular type of feminine property.

An alternative analysis, saying that all of the suffixes belong to the stem, is that males can only use verbs that happen to end in /-rž/.

Feminization

Nouns can palatalize or go under other mutations to specifically mark out that they are feminine. Neuters ending in /s/, for example, might shift of /š/ to become feminines. The masculine formation would likely just add -t.

Note that the fact that it is palatalization is essentially just down to the fact that most of the stems end with alveolars ... there was never a /ʲ/ in the feminizing morpheme, and in fact it was originally a morpheme that alternated between velar and labiovelar. Put another way, stem final /-s/ is to /-š/ as /-Ø/ is to /-g ~ -ž/.

Think of this as a derivational morpheme, not an inflectional one, even though inflections attach to it.

Attempt to sketch up a "neat" ordered verb structure

There will need to be one of these for nouns too. Noun morphology may be even more complex than verb morphology.

Derivational morphology

VERB STEM + (GENDER 1) + POSITION + inflections

Verb stem is malleable.

The GENDER 1 marker is ř for a masculine agent, possibly z for a neuter (unless this marker appears somewhere else), and Ø for a feminine agent.

The POSITION marker determines which conjugation the verb belongs to. "Forward" verbs are marked with /ġ/ in the 1st person and /f̣/ in the 2nd person; "interior" verbs are marked with /k/ in the first person and /h/ in the 2nd person. It is possible that this marker is restricted to appearing only whrn the GENDER 1 marker does not, since they both occupy the same place in the word, and could be considered to be a single morpheme after all. However, the /ř/ is present in constructions like /-řg-/ and appears transparent. In either analysis, female agents are the only ones with full access to the POSITION markers and thus to the wide array of conjugations.

Note that in the analysis below, the /z/ of neuters determines the conjugation, which is why it is unclear.

Inflectional morphonology

derivations + OBEDIENCE + PERSON/TENSE + MOOD + EXTERNAL AGENT + DISCOURSE MARKERS

The derivational morphology is covered above.

The OBEDIENCE marker, if it exists, probably goes here, as it does in Poswa. However, there is no cognacy between the two obedience systems. And therefore they may appear in different places. Obedience is already marked on the noun, and most of the semantic load would be there, but the verb would at least agree with the noun. This could also help distinguish between two different third persons as it does in Poswa, but note that IMS distinguishes three genders in its third person verb markers, and may also distinguish number.

The PERSON/TENSE marker has been fused for thousands of years and has never been pulled apart by analogy. Only a few of the paradigms have separate markers for all of the possibly conjugations. The tense distinction is between past and present; ther may or may not be also marking frog aspect. It is not clear where number is marked on the verb, if it is at all. It is possible that IMS retains the original lack of number distinction in 3rd person and that the person markers are sufficient on their own to clarify the category of number for 1st & 2nd person.

Imperative is probably retained as a distinct tense.

The MOOD marker can be indicative, subjunctive, resultative, and probably others. Resultative is a mood, not an aspect, just like in Poswa. This is inherited from the original language; the resultative is the inverse of the subjunctive. Likewise, imperative is probably a tense, not a mood, because it has separate forms for all three persons and is formed analogously to the present and past tenses.

The EXTERNAL AGENT marker may be vestigial, but it serves to prevent phonetic álly impossibly cvlusters.

The DISCOURSE markers are only used by males and by women talking to men; see below.

Diachronics and other information

Feminist Compact Imperial (3948) to Icecap Moonshine (~6800)

The expansive inherited phonology simplified quickly during the settlement period as the proto-Moonshine speakers passed through territory inhabited by speakers of Play and other languages with similarly small inventories.

Early shifts (Feminist Compact Imperial to Police)

The term Police is used instead of proto-Moonshine here, as it could be unclear whether the proto-Moonshine language was spoken in 3958 (when it was identical with Leaper) or around 4300. Properly the proto-Moonshine language should be considered to begin in 3958 because even then they considered it a separate language from Leaper for political reasons.

1. The ejective stops ṗʷ ṗ ṭ ḳ ḳʷ ḳṗ shifted to the voiceless aspirates pʷ p t k kʷ kp. Thus aspiration became nondistinctive.

   Note that PMS did not have voiced stops either.

2. All high rising tones became ordinary long tones.
3. All pharyngealized vowels became ordinary low (mid) tones. The stress became weak.
4. The rare labialized glottal fricative hʷ (sometimes spelled ħʷ for distinctness) shifted to a voiceless bilabial fricative f.
5. The common velar fricatives x xʷ came to be spelled h hʷ. Note that /hʷ/ contrasts with /f/.
6. Labialization was lost in the syllable coda; pʷ mʷ kʷ ŋʷ hʷ gʷ became p m k ŋ h g. It thus became possible to analyze labialized consonants as clusters, so long as those consonants which were both palatalized and labialized were still analyzed as /čw/ etc rather than /kyw/. The words that consisted of just these consonsnts (e.g. Leaper /xʷ/ "womb") likely continued to be labialized, but there was no distinction at this point.

VOWEL SHIFTS

The changes below are not assigned to any particular time, and may have occurred before the breakup of the parent language, but because the words in the lexicon are written with the Leaper vowels, some changes neeed t be reversed.

1. All instances of the vowels i u, on all tones, shifted to ʲi ʷu of the same tone. At least the /u/'s were probably already all pre-labialized because the Moonshine branch never did /əu/ > /ū/ as did Leaper.
2. The vowel sequence əi əu shifted to ī ū, without staining a preceding consonant. These sequences never occurred in a closed syllable, even stressed, so all such syllables were open. (The collapse of syllabic consonants had not happened yet.) There were a few syllables containing /ʷi/ and /ʲu/ in the Moonshine branch, whereas Leaper only had /ʷi/.

At this stage, reached by about 4300 AD, the Police language had a consonant inventory of:

Labials:                p   m   f       w     
Alveolars:              t   n   s   r   l    
Palataloids:            č   ň   š   ž   y    
Velars:                 k   ŋ   h   g  (Ø)

Vowels were

High tone     à  ì  ù  ə̀  
Low tone      a  i  u  ə
Long          ā  ī  ū  ə̄

The PMS /ə/ vowel corresponds to Leaper /o/ and the two were written with the same symbol rather than PMS reviving the early Gold schwa glyph. The script also had a row of symbols for /e/, but this /e/ could be analyzed as /ai/. It just happened that there were no other falling diphthongs in the language. Unlike Khulls, the palatal glide /y/ could occur after labialized consonants, and it did not stain a following vowel. Thus all four vowels could occur after the /y/.

Importantly, the vowel and diphthong inventory was very similar to that of contemporary Play, which also had /a i u ə/ and the diphthongs /ai au əi əu/. The differences were that Moonshine also had a long schwa /ə̄/ but lacked Play's distinction between long vowels and vowel sequences. That is, Play had both /aa/ and /ā/, but Moonshine allowed only /ā/.

Prenasals existed in word-initial position, also unlike Khulls. e.g. /mpʷà/ "house" vs Khulls pà.

Police (c. 4300) to Ice Cream in a Bowl (c. 5547)

1. The sequence kp shifted to pˠ.
2. The clusters mm nn ŋŋ shifted to m n ŋ and lengthened the preceding vowel.

   this may lead to unstressed longs, unknown in Khulls, which could survive vowel deletion.

   Alternatively these could be fortis.

3. Syllabic consonants bordered by vowels became normal.

   This means that the vowel + g + consonant declension simply becomes vowel + consonant.,

4. Unstressed syllable-final s shifted to h. This has repercussions later on because it leads some vowels to disappear before /s/ and not when in an open syllable.
5. All remaining syllabic consonants (bounded by consonants) became normal.
6. The postalveolar affricates š ž became fortis ʂ ʐ. (This is a spelling convention to avoid stacking diacritics on other diacritics.)

   NOTE, THIS SHIFT IS SOMETIMES WIPED AWAY IF THE PRECEDING VOWEL IS /i/. (because iž ~ iy)

7. The clusters hʷy gʷy became ʂʷ ʐʷ.
8. The consonant w in the coda (corresponding to Leaper /ʕʷ/) here became an ordinary u, forming a diphthong. Note that this consonant could occur after diphthongs itself, but only in newly formed compounds, and these words may have been "rescued" by changing the preceding diphthong elements from /j w/ to /ž gʷ/.
9. Before a word-final unstressed vowel, the free palatal glide y shifted to ž. (This is probably due to analogy rather than sound change, and since there were only a few such morphemes, they can be considered as substitutions. Essentially /y/ was assumed to have come from /gy/ whether it had or not.)

   THIS CHANGE MAY NEED TO BE BURIED IN THE MIDDLE OF THE VOWEL CHANGES because it relies on them.

Loss of unstressed vowels

All four vowels /a i u ɜ/ disappeared in unstressed position, except when propped up by certain conditions explained below. The list speaks of two (perhaps three) dummy vowels.

1. In unstressed position, the vowels a ɜ disappeared to Ø when bordering one of /g h gʷ hʷ/ in either direction. This created clusters.

   This shift may have happened simultaneously on both sides of the consonant, so that for example, /-taga/ > /tg/, or it may have preserved the final vowel, which nonetheless disappears later. The distinction will only be meaningful if words appear in which there is in fact a difference in final outcome.

2. Where the voiced velar fricative g collided with any consonant from either side, it made that consonant into a fortis consonant, marked by Cˠ (perhaps /Cʰ/ for easier typing and display). At the immediately following stage, the entire inventory of consonants could be fortified this way, and thus they were not considered new phonemes but still as clusters. The same thing happened with gʷ except that the consonant was labialized in addition.

   Note that this shift even applies to sequences like /ng/, which became /nˠ/.

3. Where the voiceless velar fricative h collided, it made the preceding consonant voiceless and aspirated. This meant nothing for stops; other consonants are dealt with below. These consonants also became fortis, and can be spelled Cˠ or as Cʰ.
4. All LABIALIZED consonants became FORTIS in addition, while retaining labialization.
5. Any unaccented ɜ shifted to Ø .
6. The diphthongs ai au (of any origin) now became āʲ āʷ, with the coarticulations cleaving onto a following consonant, if any. In open syllables they continued to be spelled with the symbols for inherited /ai au/.

   This may be a problem, however, as it suggests that even the plain vowels /i u/ should stain following consonants.

7. In stressed position, the vowels ɜ̆ ɜ̀ ɜ̄ shifted to the color of the vowel in the next syllable, and to ă à ā in monosyllables. They did not acquire preceding glides /ʲ ʷ/.
8. The diphthongs ɜi ɜu became āʲ āʷ when an /a/ followed in the next syllable, and ī ū otherwise (even in monosyllables). Any iu (that is, /i/ + /ʕʷ/) also became ū.
9. Unaccented WORD-FINAL a ʲi ʷu (not just unstressed) shifted to Ø ʲ ʷ unconditionally, creating more clusters. This happened even after fortis consonants. Thus for example /-taga/ > /-tˠ/, with both vowels deleted.

   Now, every word ended with a consonant unless the final syllable was stressed. These shifts meant that for example, /-kas -ka/ had become /-kˠ -k/.

10. Unstressed ʲi u (and /i/ if it existed) now shifted to ʲ ʷ unless they occurred in one of Cˠ_ _Cˠ C_Cʲ where clusters also count as fortis consonants.
11. Then, unaccented a (and any other vowels) became Ø (sometimes spelled ʕ), unless it occurred in one of the following environments: Cˠ_ _Cˠ C_Cʲ where clusters also count as fortis consonants. The dummy vowels ɛ ɔ can be used here to spell /aCʲ aCʷ/, and ɐ to spell /CʷaC/ and /CCa/. These were mere allophones of /a/, however.

At this stage the language had a problem with unstressed closed syllables, and deletion of /ž žʲ/ (which were distinct) would not solve this problem because Play requires a distinction that would have forced Leaper and Moonshine to have the same sound for the instances in which it would be most distinct. That is, Moonshine *-az is a mistake, as Moonshine only ever had /žʲ/ in that position (a separate, much rarer plain /ž/ existed in other morphemes). Nonetheless, deletion of /ž žʲ/ could happen for its own sake, without solving the closed syllable problem.

Monosyllabic stage

Now, the Moonshine language had only three vowels, but an abundance of consonants, including a lenis:fortis distinction and consonants that were palatalized, labialized, or both. All labialized consonants were fortis, but not all fortis consonants were labialized, so there were five possible modes of consonant: plain, palatalized, fortis, fortis labialized, and palatalized labialized. The fortis articulation was velarization for the frontal consonants but not for the velars.

1. Before a vowel, the labial approximant w shifted to a uvular approximant r̄.
2. Doubled consonants (and fortis) simplified to singles and caused the tone of the preceding vowel to become high (à or á).
3. Any consonant before a nasal disappeared and lengthened the preceding vowel. If the sound had been voiceless, it caused the tone of the preceding vowel to become high rising (á). If it had been voiced, it caused the tone of the preceding vowel to become high falling (ā).

   Possibly /ntn/ > /nn/. Or ʔnn.

4. The clusters nlh nlk shifted to ŋh ŋk.
5. The palatalized labials pʲ mʲ fʲ became the labiodentals fˠ ṃˠ f in all positions. Then, a shift from ṃˠ to a plain bilabial mˠ spread throughout the territory, slowly driving out the new phoneme.
6. The alveolar flap r became l before any stop.

Ice Cream in a Bowl (c. 5547) to Classical Cartwheel (~6800)

1. The palatalized velars kʲ hʲ ŋʲ merged with the palatalized coronals tʲ sʲ nʲ into č š ň. This also includes all such consonants before front vowels, since these were always preceded by /ʲ/. If /gʲ/ existed, it became /ž/.
2. A single, non-fortis labial, when following any posttonic consonant and before a vowel, generated /ʷ/ and then disappeared.
3. The labialized alveolars tʷ sʷ nʷ became the dentals ṭ ṣ ṇ in all positions. This shift is distinct from a similar one that happens later when new labialized alveolars are created.
4. All high rising tones (á etc) merged into either ā or à when not before a nasal.
5. Then lʲ rʲ became y r.

   This may mean that rž and rʲ have distinct reflexes. Consider whether the cluster /rž/ would behave uniquely here or whether it would simply behave like its constituent parts.

   Consider also that there will briefly exist a /lʲʷ/.

6. Unstressed a in a closed syllable became ʲi (or possibly just /i/) before a palatal consonant. That is, /ɛ/ > /ʲi/. There was no corresponding change before a labialized coda.
7. All coda palatalization was lost (and they became alveolars, not velars.)

New palatalization

1. The labialized postalveolar consonants čʷ ǯʷ šʷ žʷ ňʷ delabialized.
2. The palatal consonants č ǯ š ž ň became c ʒ s z n in all positions.
3. The labialized alveolar approximant lʷ shifted to w. Possibly /rʷ/ (the "dark" one) also shifted.
4. Nasals disappeared before a fricative.
5. All nasal+stop combinations become voiced; prenasals mp nt nc ňč ŋk shifted to plain voiced stops b d ʒ ǯ ġ. Here, the coarticulations are omitted for conciseness. ALL VOICED STOPS WERE FORTIS.
6. Posttonic low short a became Ø if an adjacent syllable had a long vowel (even if that long vowel was not the tonic vowel).
7. The velar stops k ġ became labialized to kʷ ġʷ before any labial consonant. Possibly a few other clusters like /ks/ also survived. It is possible also that kl > kʷl > kw, distinct from kʷ.
8. The velars k ġ h g ŋ were fronted to postalveolars č ǯ š ž ň. These behaved as palatalized consonants, cʲ ʒʲ sʲ zʲ nʲ, and any remaining fortification was lost. It is possible that the fortis velars had all become labiovelars by this point.

Second unstressed vowel deletion

Assuming the distinction of lenis/fortis in absolute final position holds on from the beginning, it becomes important here as it determines whether an unstressed vowel is lost or not. Thus for example /nk/ could alternate with /nak/ in some noun to form the genitive.

Also note that new prenasal sequences are created here, e.g. /map/ > /mp/, and these do not become voiced stops or even voiced prenasals. It is possible that the prenasals in the earlier shift never lost their nasal element, and thus there would be /mb mp/ but no free /b/, but this would be unusual.

1. All unstressed vowels weakened: a i u all disappeared to Ø, unless protected by a fortis consonant. Unlike the similar shift about 2,000 years earlier, there was no protection afforded by bordering another collapsing vowel. As before, most unstressed /i/ was really /ʲi/ and thus this created new palatalized consonants. For the most part, the consonants did not stain each other.
2. Nevertheless, palatalization sometimes skipped across a syllable boundary. For example, /hip/ > /hʲp/ > /pʲ/ is possible. It may be that /h/ was the only consonant in the entire language that violated the rule against consonants staining each other.
3. All unstressed syllables with "tone" now became plain. These had arisen from secondary stress in compounds. Thus unstressed ā ī ū à ì ù became a i i a i i.
4. Any remaining fortis consonants not touching the tonic syllable became plain.

Sequence changes

1. The sequences ph th shifted to pw tw.
2. Coronals followed by the labial glide /w/, including those generated above, became dental. Thus for example the verb ending nw became ṇw.

Icecap Moonshine phoneme inventory

Three-vowel analysis ("Kick Jam")

If the vowels are just /a i u/, then every other acoustic distinction between the vowels must be analyzed with the consonants. The resulting consonant inventory thus has pairs distinguished by palatalization for any consonant that can occur before (or sometimes after) a surface-level vowel pair of [a]~[e]. Likewise any [o] must be analyzed as either /a/ (when before a velar or velarized consonant) or as /u/ (in freestanding position or before any other consonant).

                   PLAIN                                  PALATALIZED
Bilabials:       p   b   m   ḟ   w                      pʲ  bʲ  mʲ
Labiodentals:   (ṗ       ṃ)  f   
Dentals:        (ṭ       ṇ)  ṣ       ḷ
Alveolars:       t   d   n   s   z   l   r   c   ʒ      č   ǯ   ň   š   ž   ʎ   ř
Velars:          k   ġ   ŋ   h   g   r̄                  kʲ  ġʲ  ń   ś   y

If the labiodentals ṗ ṃ, which are indifferent to palatalization, are analyzed as clusters ending in f (which is also indifferent to palatalization), the consonant table can be rearranged into a traditional shape:

Rounded labials:        p   b   m   ḟ   w
Spread labials:         pʲ  bʲ  mʲ  f              (ṗ   ṁ)
Dentals:                            ṣ       ḷ      (ṭ   ṇ)
Alveolars:              t   d   n   s   z   l   r  (c   ʒ)
Postalveolars:          č   ǯ   ň   š   ž   ʎ   ř
Palatals:               kʲ  ġʲ  ń   ś   y
Velars:                 k   ġ   ŋ   h   g   r̄

But note that this analysis groups the labiodental f with the bilabials pʲ bʲ mʲ, which have little in common and do not have the same effects on surrounding vowels.

Five-vowel analysis

The final consonant inventory was

Bilabials:          p   b   m   ḟ   w
Labiodentals:      (ṗ       ṃ)  f    
Dentals:           (ṭ       ṇ)  ṣ       ḷ
Alveolars:          t   d   n   s   z   l   ř   c   ʒ
Postalveolars:              ň   š   ž           č   ǯ
Palatals:                       ś   y
Velars:             k   ġ   ŋ   h   g   r

And vowels were /a e i o u/ on all four tones, though the long /o/ tones only occurred over morpheme boundaries and were not represented in the native script.

Minimalist consonant inventory

The fish shaped inventory includes affricates as single consonants. The first four listed, ṗ ṃ ṭ ṇ, are taught as single consonants in Moonshine classrooms because they can occur in positions where other clusters cannot, and because they sometimes appear as the reflexes of PMS single consonants. (Note that this sound shift is directly related to Poswa's shift of /pʲ mʲ tʲ nʲ/ > /f v f v/, but Moonshine's shift had both a different starting point and different results.) However, these affricates cannot contrast with clusters, and therefore under a strict analysis they can be analyzed as the sequences /pf mf tṣ nṣ/.

Likewise, the voiced stops b d ʒ ǯ ġ can be removed from the inventory because they only appear as the reflex of earlier clusters /mp nt nč ŋk ŋkʷ/, or in a very small number of words from other clusters involving nasals. The voiceless fricatives f̣ ś can also be ignored, because although they contrast with the more common /f š/, they do so only in positions where clusters can occur, and can thus be analyzed as clusters with /h/.

Even in a minimalist analysis, the voiceless affricate /c/ remains, as do the postalveolar consonants / ň š ž č/. This is because they can contrast with clusters over syllable boundaries; that is, /tš/ will always be split when it straddles a syllable boundary, whereas /č/ will always go with the second syllable. This affects the pronunciation of the preceding vowel, as well.

The consonants /w y/ cannot be considered vowels because they cannot take tones. Thus an absolute minimum consonant inventory for Icecap Moonshine would have 23 consonants:

Labials:         p   m   f       w
Dentals:                 ṣ       ḷ
Alveolars:       t   n   s   z   l   ř   c
Palataloids:         ň   š   ž   y       č
Velars:          k   ŋ   h   g   r

In theory a minimalist approach could analyze the four tones as sequences of a vowel followed by a silent consonant, but the Moonshine grammar behaves such that it is far more convenient to treat each tone as inseparable from its vowel, meaning that even in a minimalist analysis there are five vowels and four tones, for a total of 20 syllabic nuclei.

Note that if there are no voiced stops, then the initial clusters like /mp nt nč ŋk/ are like /pf tṣ/ in that they can violate the sonority h ierarchy.

Maximalist consonant inventory

A maximalist analysis would accept all 34 of the consonants in the first table above as phonemic, and also add new phonemes for historical coarticulations such as /pʷ kʷ sʲ/ and the like, which no longer contrast audibly with their corresponding plain forms, but affect the pronunciation of any preceding vowel. Note that, for example, the only occurrences of long /o/ in Icecap Moonshine are from long /a/ followed by a historically labialized consonant. These cannot be analyzed as sequences such as /aw/ either because they occur on all four tones whereas sequences like /aw/ are always long. The Moonshine script follows this maximalist approach, but because there is no acoustic distinction between these pairs of consonants, it is not taught as such in advanced classrooms.

A glottal stop is inserted after a high tone followed by another vowel or any of a certain set of consonants, and could thus be added to the consonant inventory as well, but it cannot contrast with silence because the tone would also change.

A maximalist approach might also expand the vowel inventory by not only considering tones as inseparable from their vowels, but also certain sonorant consonants such as /l ř r w y/.

Overview

Icecap Moonshine is highly derived, in the sense that it scarcely resembles the proto-Moonshine language spoken 3,000 years earlier. It is one of the few fusional languages in which morphemes can delete preceding morphemes or trigger other phonemic shifts such as /a/ shifting to /e/ or /o/.

The rapid turnover of vocabulary and grammar led Moonshine scholars to place the language in a category of its own even as they knew that it had evolved from the same branch of the family that had led to Leaper. These scholars were more interested in shared traits than in genetics, and saw that Moonshine was so unlike all other languages that it could not be linked to either closely related languages such as Khulls or distantly related but superficially more familiar languages such as Poswa.

Gender and cultural interactions

NOTE, this section is currently undergoing upheaval.

All of the traits that made Moonshine famous arose during the period of political isolation after the Feminist Compact lost their war against the all-male Matrix army and retreated into the world's coldest habitats. There were some influences from Repilian languages, but by the time of the creation of the Moonshine state, Repilian languages had been in decline for thousands of years, and Moonshine quickly drove out the languages that had earlier driven out the Repilian languages. Some distinct characteristics of the language are:

1. When men address women, they must add evidential morphemes to most nouns and verbs (all except those of the 1st person) explaining how sure they are of what they're describing, and if they are sure, which woman is the one who pointed it out to them. There are three sets of evidentials, and each set is itself a table of forms that vary depending on the gender of the noun (or verb) and in some cases other things.

   My plan is to add this function to the language without deciding whether it is optional or mandatory. At the very least, it will be grammatically possible for men to speak normally, rather than, for example, requiring male speakers to produce one of three possible forms for each word with none of them being identifiable as the default.

2. Men are not allowed to use the everyday words for female body parts and many other feminine nouns, even those only distantly related to gender, such as the words for certain flowers. A separate word acceptable for men to say must therefore be learned for every such object in the lexicon, and many concepts have four such words: one each for female and male speakers and listeners.

   This is probably no longer true, as I am no longer assigning gender to words based on the Gold principle of "rightmost gendered morpheme". Instead, gender is simply determined by the gender marker, if present.

3. Some men's words are the same as women's words but with an additional honorific morpheme expressing power, beauty, and other feminine traits.

   For example, the circumfix ho-...-i.

4. When men quote women's speech they do so exactly, even though men are ordinarily not allowed to say certain words out loud.
5. The semantic scope of the register for male speakers and male listeners is very limited, and makes communication difficult. This register is called the vulgar register. Thus even when men speak to men their vocabulary is limited by the rules women push on them. This is in part because women ensure that men will not come to prefer conversations with men to those with women.
6. There is only one inherently masculine noun in the entire language, t "son; boy; man", although neuter nouns become masculine when they are possessed by males.

   This is probably no longer true.

7. The common word for "man" as an agent, however, is not t, but rather lem, ending in a suffix indicating feminine property. Thus males may act as agents and possessors on neuter nouns such as inanimate objects, but not on their own selves. Any male agent requires an implicit female external agent considered to be the true subject of the sentence.
8. Furthermore, male agents cannot be the owners of grammatically feminine objects, because one piece of property cannot own or act upon another piece of property.

   This is still true in the formal register (when women are present), but likely not true when men speak amongst themselves.

9. A separate verbal inflection paradigm exists for actions in which a woman forces an inanimate object or a man to perform the action, but there is no matching counterpart for other gender matchups.

   This is still true. This arises from the conflation of the masculine noun class with that of handheld tools, which is likely the only group that can be used to derive instrumental verbs.

10. Every content word in the sentence takes yet another marker, the discourse marker, which agrees with the gender of the speaker and listener of the sentence, except that when both are female, the discourse marker is a null morpheme. There is not just one discourse marker however; they are different for different nouns and verbs. This requires the man to keep track of such things as whether an object is visible or invisible, tangible or intangible, and so on, whereas women speaking to women (and likely also to men) use the words freely.

    If this is still true, the discourse marker is conflated with other morphemes wherever possible, so that even men in the most formal situations will not need to stack two so-called "humiliation morphemes" together.

11. Male and female agents often have entirely separate verbs for common actions; in most cases, a verb that is not derived from a noun will be inherently either masculine or feminine, and cannot be used with the opposite gender.

    Any instances of this will be due to stress shifts and the masculine stem will be also available to feminine agents. Thus it could rather be said that there is a group of defective verbs which can only take feminine agents.

12. The feminine verb conjugations are mostly strong verbs with generally more compact morphology, while the masculine verb conjugations are weak. However, there is also one feminine weak conjugation.

    This is probably no longer true, as it is not clear that verb conjugations really even exist. The morphemes that I had planned to develop into conjugation markers are separated from the gender markers now.

13. A male or inanimate agent acting on any patient other than themselves or another inanimate object requires a set of obedience morphemes marking out which external female agent gave them permission to act. These are marked on the noun, not the verb, and changes the verb to the feminine conjugation corresponding to the external forcing agent. The obedience morphemes are built from possession markers compounded with one of four morphemes indicating the degree to which the obedience is voluntary.

    Something like this will still be true, but it is more likely to be marked on the verb after all.

Outside perspectives

Poswa

NOTE: Because the Moonshine language is far from finished, I can only write about this indirectly. I am using Poswa because it is my best developed language. If I ever get far enough, I will use actual Moonshine instead of an intermediary language.

The difficulties encountered by men trying to make themselves understood in Moonshine were frequently mocked even by Poswob scholars, who typically considered their own scholarly accomplishments inferior to the Moonshines', but were just as proud of Poswa as the Moonshines were of Moonshine. Male Moonshine characters in Poswob stories needed no parody, because any literal translation of the speech of a man addressing a woman in Moonshine was so long-winded that further exaggeration actually weakened the comedic effect.

For example, a Poswob woman asking a man to choose one of two flowers could ask

Fabumbope, pipop tammavape?

Do you (want) a tulip, or do you (want) a rose?

This sentence is grammatical, easy to understand, and makes use of the common four-syllable speech tempo that dominates Poswa. The man's reply, even if he were a scholar, would simply be

Fabumbopo.

I want the tulip.

Or

Tammavapo.

I want the rose.

Both of which are simply vowel-rotated forms of the two content words in the original question.

But in a Poswa play, a Moonshine woman asking a Moonshine man to pick between a rose and a tulip would say

Fi pu waba?

Tulip or rose?

Using obscure onomastic words with ambiguous meanings and thus requiring the man to do the work of figuring out the question on his own, and an incorrect word for or. The man would then reply with something such as

Paefimpose, nubevwope, mapembe twuppupopo wembabofafo fupie wataežos.

You, guardian over me, I see what you show. Your heart allows me a beautiful apple petal flower to carry and I happily follow you by accepting the choice you offer me.

This reply requires the man to recognize the woman's superior status, to specifically call out that he indeed sees the flowers in front of him, that he needs her permission to choose one of the flowers, that it will not become his property, that the flowers are beautiful, that he is accepting the gift of the flower willingly, and yet that he is also obeying her by doing so, all while avoiding actually saying the names of the flowers because the names of many flowers are derived from grammatically feminine words considered too beautiful to come from the speech of a man.

The actual Moonshine words involved here are not nearly as long as the Poswa words or the English gloss because such structures are and have been part of Moonshine for thousands of years, but the literal translations are accurate. The word for tulip occurs in the middle of the Poswa sentence, surrounded by three sequences of important politeness morphemes on each side, just as it does in the original Moonshine.

Some Moonshine women instruct their husbands to ignore some of the politeness rules above so they can speak more rapidly. These men thus have separate speech registers for talking to different women, and must remember which rules apply in each given situation. Other Moonshine women would avoid the situation up above by instructing their husbands to point to the flower they want.

Other unusual characteristics

Though Icecap Moonshine is primarily known for its gender system, it also stands out from neighboring languages in many other ways, and Moonshine linguistic scholars have much more to talk about than how gender shapes their discourse. For example, Icecap Moonshine is at once the only Cosmopolitan language with circumfixes, the only one with a fusionally overlapping inflection system (that is, the borders between morphemes are impossible to define), and the language with the highest number of permitted syllables.

1. Most basic content words in Moonshine are monosyllabic, and even these monosyllabic roots are often segmentable into a vowel padded by a consonantal circumfix. There are many homophones, and scholars take pride in writing long paragraphs in which many of the words can be interpreted in more than one way on their own but make sense in only one way when the piece is parsed as a whole.
2. Icecap Moonshine has a large phoneme inventory, but is written with an alphabet that makes it seem larger still by marking silent letters and having several ways of writing each phoneme. Though the writing system is alphabetic in nature, with each glyph standing for a single phoneme, scholars have introduced many logograms over time, which are also pronounced as single phonemes (mostly consonants) and serve to indicate the precise meaning of the word they occur in without the writer needing to use a longer word. All of the logograms are graphically simple, similar to the inherited alphabetic glyphs.
3. Morphemes can delete or merge with preceding morphemes, and sometimes a single phoneme can represent two or three morphemes. Since the boundaries between these morphemes are no longer segmentable, Moonshine teachers usually do not speak of them as separate morphemes, and instead teach students to memorize tables of fused morphemes.

Phonology

Moonshine has a large phonology with with about 40 consonants, 5 vowels, and a strong tone system with contrasts on every syllable and weak tonal sandhi. Counting tones as a feature of syllables, Icecap Moonshine has the largest permissible syllable inventory in the world.

Consonants

The consonants /c ʒ/ are in IPA /ts dz/, and are considered phonemic only because they would otherwise violate the sonority hierarchy because they can occur at the ends of words where one would otherwise expect just /t d/. The stops /ṗ ḅ ṭ ḍ k ġ/ are not distinguished from affricates /ṗf ḅv ṭṣ ḍẓ kh ġg/ at all, however, so given that /č ǯ/ exist without homorganic stops it could be said that /c ʒ/ are just as basic to the phonology as /t d/ are. (The true bilabial stops are indeed distinguished from affricates, but only because the bilabial fricatives have [w] as an allophone after a stop.)

The palatal approximant is placed with the postalveolar row by tradition, but is a true palatal.

Voiced stops are prenasalized when preceded by vowels. But fricatives are not.

Romanization of consonants

Note that the dot diacritic has several unrelated meanings: it can indicate a (labio)dental pronunciation, as with ṗ ḅ ṭ ḍ ṣ ẓ ṇ ḷ; a bilabial one, as with ḟ ṿ, or a simple stop as opposed to a fricative, as with ġ. Additionally, although the caron marks a postalveolar pronunciation on š ž č ǯ (and ň if this spelling is substituted for ñ), it marks an alveolar trill when used on ř.

Laryngeal consonants

The Moonshine alphabet contains two more consonant symbols: /ʔ/ and /ʕ/, which are both silent. However, /ʔ/ makes the previous consonant voiceless; thus Tòdʔřóm (the name of a state) is pronounced as if spelled Tòtřóm. The /ʕ/ is silent and has no effect at all on surrounding consonants, but both symbols mark places where vowels used to be and sometimes reappear in conjugations.

Approximants

The sound [w] is an allophone of the voiceless bilabial fricatives /f̣ ṿ/ after another consonant. A bare /w/ does not occur in native words but can be spelled ʕṿ, where the silent /ʕ/ shows that the following ṿ is using its post-consonantal allophone.

Teachers use two additional letters, ʲ and ʷ, which are not found in the traditional alphabet because they primarily surface as grammatical alternations in which preceding vowels are brought further frontward or backward. In the rare case where one of these phonemes appears between two consonants, they are pronounced /e/ and /o/ respectively, and spelled as such in ordinary writing.

NOTE, these could be spelled with fullsize j and w so long as the other /w/ is spelled as ṿ. However, this would lead to writing clusters like /kw/ as kṿ, etc.

Clusters

There are many clusters found in no other language, such as /th/ (IPA [tx]), /fl/, etc., but the sonority hierarchy is strictly observed. This is why the affricates are considered single consonants. That is, /ts/ can occur at the end of a word because it behaves as a single consonant c, but /ps/ cannot occur at the end of a word. Stems ending in clusters like /-ps/ are always followed by vowels.

Vowels

The vowels are cardinal IPA /a e i o u/. They become more centralized ("lax") when in a closed syllable, and because the ` tone adds a glottal stop after the vowel, all ` vowels are closed syllables and therefore lax. This even applies to cases in which a vowel immediately follows the grave-tone vowel.

There are no diphthongs or vowel sequences; written sequences like ài are separated by a glottal stop because the first vowel is a grave tone.

Syllable structure

Teachers analyze Icecap Moonshine's syllable structure as CCVC, with CCVCC roots permissible only because they belong to a category of words which can only appear when followed by a suffix beginning in a vowel or with one of the consonants that can follow other consonants. All syllables obey a strict sonority hierarchy.

However, the silent letters ʕ ʔ ʲ ʷ complicate the syllable structure considerably. Though descended etymologically from vowels, teachers analyze these as consonants because they can affect the pronunciation of adjacent consonants. For example, any /dʔ/ is pronounced [t]. These silent letters do not count as consonants in determining the syllable structure, as, for example, /pʔlàt/ is a valid word even though it begins with three written consonants. This is because there is no situation where the silent letters have a consonantal realization; in the rare cases where a silent letter (usually one of /ʲ ʷ/) is trapped between consonants at a syllable edge, it is pronounced as a weak short vowel instead.

Gender

IMS is notable for its extreme feminine bias, in that women and feminine objects are associated with power and success whereas males are ranked lower than some inanimate objects. It is often necessary to introduce a feminine subject simply to complete a sentence. The Moonshine people have lived in an extremely feministic society for 5,000 years and this has shaped the language to a degree found nowhere else. Many of the traits today found in Moonshine were part of Repilian languages such as the Owl family for thousands of years, but Moonshine expresses these traits to a greater extent than any Repilian language ever did, and Moonshine has also evolved similar traits of its own not shared by any Repilian language.

Gender can be marked up to five times on a single word: for the word itself (even if it is a verb), for the owner, for the agent (even if it is a noun), for the speaker, and for the listener.

It is common to list the possession markers in the order 1f 2f 3f 3m, because 1m & 2m can be derived from the rest. E.g. the clothes ending in ň are ǯā ǯas ǯ nen.

Nominal possession markers

NOUN CLASS             GENDER          FREE  1F    2F    3F   3M

Clothes                neuter         -ň    -ǯā   -ǯas  -ǯ   -nen
Furniture              neuter         -č    -čā   -čas  -č*  -cen 

Places/Female Body     feminine       -m    -žā   -š    -ž  (-šten)

Males                  masculine**    -t    -tā   -tas  -ta  -ten

*Dialectal.
**Reorients to feminine when free.

All neuter nouns inherit the gender of their possessor. Forms in parentheses are present only for a subset of the nouns in the category.

Masculine 1st and 2nd possessive forms are not recognized because all nouns inflect for the gender of the speaker and listener, and therefore the masculines are derived from underlying feminine forms with these discourse markers added on. When a case marker occurs between the possession marker and the discourse marker, the underlying feminine form is inflected, because the discourse marker must be the outermost morpheme on any noun.

Nouns outside the class system often end in vowels, and these must take a linking consonant, usually -č- or -m-, whereupon they come to behave identically to the other nouns ending in those consonants.

Neuter nouns

Most nouns in IMS are neuter. IMS retains most of the neuter nouns inherited from the Gold language, whereas in the other descendants a large number of neuter nouns were shifted into the feminine and masculine genders. Gender in IMS is more closely tied to semantics than in most related languages, and there are very few nouns semantically excluded from the neuter gender because they are the category that encompasses all nouns not in the other two categories.

Neuter nouns acquire the gender of their possessor, however, so while the word for seashell is neuter in isolation, it becomes masculine or feminine if it refers to the personal property of a man or a woman.

Feminine nouns

Most feminine nouns in IMS are semantically related to the female anatomy, although the chain of relation can be very long, as any newly derived feminine noun behaves the same as the original, and many new words have been coined during the 5,000 years since the split of Moonshine from its relatives. For example, tulips are feminine because they resemble skirts, and skirts are feminine because they loop around the wearer's womb. But daisies are neuter because there is no such connection.

Triangular objects are mostly feminine because they resemble the shape of an empty womb. Round objects are mostly feminine because they resemble breasts. This includes objects that are round in only one view, such as tubes and rings.

Because of its extreme rate of sound change, most Moonshine nouns are historically compounds. IMS follows the inherited rule that the rightmost gendered (that is, non-neuter) morpheme in a noun determines the gender of the noun. However, the etymology of most nouns is opaque and in many cases reanalysis has taken over.

A second category of feminine nouns contains words for objects that have no semantic relation, even a remote one, to female anatomy, but have come to be feminine because their stems originally ended in low-tone vowels and therefore they acquired a final -m during the rapid overturn and reform of the inflection system. This includes words like lim "clock". There is no semantic pattern to which nouns became feminine and which did not.

Masculine nouns

By contrast, masculine nouns obey a simple rule: they end in the grammatical masculine marker -t. Most masculine nouns are dynamically constructed by adding this suffix to a neuter or feminine noun. The most common word for man, le, is in fact grammatically feminine, because men are considered to be female property. Thus, merely to be the agent of a verb, a man must use the suffix indicating a borrowed noun. Other masculine nouns follow similar patterns: the word for king, used to describe foreign monarchs, is pó "queen", plus a suffix indicating semantic similarity, plus the masculine suffix. Thus the word for king means "a man that is like a queen".

If the stem of the word ends in a primordial /n/, this merges with the /t/ to become /d/.

There is no shape-based analogy creating masculine nouns because men are not seen as having any distinct anatomy apart from the penis, which is considered to be feminine.

Compound gender stacking

Icecap Moonshine requires nouns, and some verbs, to be marked for not only their inherent gender, but also the gender of their possessor, the agent of the sentence, and the speaker and listener. Thus it is common to see five gender affixes on a single word, though these are in all cases fused.

Inherent gender

Each Moonshine root, whether it be a single consonant or a sequence, has an inherent gender. In compound words, the rightmost gendered morpheme determines the gender of the word. The only masculine root in the language is t "son"; there are dozens of feminine roots, but the vast majority of roots are neuter. But any word with a feminine morpheme in it is feminine itself unless it ends in t.

Because of the many single-phoneme roots, whether the gender is analyzed as an ending or as a separate root is a matter of principle. Note that t means "son" and can occur elsewhere within a word, and also can mean "man, boy" when not used with a possession marker, since all men are sons.

Possessor gender

All neuter nouns can be possessed by neuters, males, and females. All masculine nouns can be possessed by males and females. All feminine nouns can be possessed by females. All three genders of nouns can also be unpossessed. These markers always fuse to the final morpheme in the word, and therefore there are many forms for the fused inherent+possessor gender marker (IP). The IP marker also indicates the person of the possessor, if there is one. The total number of forms is eight for neuter nouns (free, neuter, and 3 each for masculine and feminine possessors), seven for masculine nouns (one free and six gendered possessors), and three for feminine nouns. There are thus 18 total forms for the IP-marked form of each phoneme at the end of a word.

The nominal case marker occurs after the IP marker, and these fuse somewhat to the IP marker as well, though most combinations are still transparently segmentable. The most commonly used forms are often the most compressed: for example, -ṭ- signifies a female agent acting on a male patient with no possessor in the instrumental case; that is, a woman using a man.

Thus one can say

Ṭač néʒa šāḍù.

I had a man clean the house.

It is also possible to put the ṭač word at the end, treating it as if it were a verb ending, and omitting the ḍù. This requires also changing the word for house, however.

Agent gender

After the case marker comes an agent marker which can be either feminine or masculine. This, too, also marks person, and therefore there are six possible forms.

A padding morpheme occurs after the agent marker in many words to separate it from the following markers. In most cases, this derives from an earlier evidential marker, but some padding morphemes were copied from other verb forms and have no etymological meaning.

The practice of marking the agent of the sentence on a noun is inherited from Gold, where it was necessary due to Gold not having pronouns. Moonshine developed new pronouns, and thus the agent markers are no longer necessary and some have coalesced with others.

Speaker and listener gender

The speaker and listener gender markers, known as discourse markers, are completely fused with no transparent segments. These lists are full of gaps, because there are many words that males cannot say, and some words that women omit when speaking to men. Additionally, there are gaps corresponding to other gender markers; for example, a verb conjugated with a 1st person masculine agent marker cannot have a feminine speaker.

The maximal possible list of forms is four, as neuters are by definition not involved in either speaking or being spoken to, and because the speaker is by definition 1st person and the listener is by definition 2nd person, so person is not marked.

A further complication is that the speaker-listener morphemes are not static, but take separate forms depending on the preceding morphemes. For example, the ♂→♀ marker has a different form for describing feminine objects vs others, and separate for describing 2nd person feminine objects (that is, the listener's belongings) vs 3rd person feminine objects.

Surface analysis

Note that the nominal case marker is squeezed between two sets of gender markers. Thus, the outermost morpheme on any noun is not its case marking, but the morphemes indicating who is speaking and who is listening. However, for the situation with a female speaking to another female, these two outer morphemes are both null (Ø), and so the original structure of the word is preserved. When women speak to men, the forms of the outer gender markers vary. It is only when men speak that the words all have the same ending.

Speech registers

There are four speech registers: the speaker can be female or male, as can the listener. Children in the nursery do not acquire these speech registers until they start school. In school, boys learning proper grammar are humiliated as they realize that the grammar requires them to use separate, longer forms for nearly every word, while by contrast women's speech leaves out various details whenever the listener is a man.

Thus, not only are the words for men and men's items longer than those of women, but even these lengthened forms are further extended whenever the speaker is a man, and if the listener is a woman, the man must also use evidentiality morphemes to indicate that he is uncertain of what he is seeing, unless a female speaker has mentioned that noun earlier in the conversation, in which case the man must use a different morpheme indicating that he recognizes that she has done so. Speaking the language is so difficult for men that men in the presence of women typically stay silent until they are spoken to.

Direct register

The direct register is used between females, and by females addressing mixed groups. It does not entail any additional marking for the speaker and listener. It is the only register used in writing and is considered the only true descendant of the parent language.

Direct register vocabulary

There are no obscene words in the direct register, and the only word taboos are situational, and dependent on religious and superstitious beliefs. For example, it is taboo to use words for abortion in the presence of a pregnant woman, even for other women.

Deferential register

The deferential register is used by men addressing women and mixed groups, and also whenever a female is present even if the speaker is addressing another male. The character of the register changes depending on the listener, but certain rules must be followed at all times.

When addressing a female, content words of the 2nd and 3rd person must take evidential morphemes explaining why the man thinks he knows what he is talking about. These evidentials are suffixed to the otherwise obsolete interior person markers -č -š -Ø. When addressing a male, the structure reverts to the form used in the vulgar register.

Inflections are of mixed origins: some are suffixed forms of morphemes in the direct register, others are from the vulgar register, and others are unique to the deferential register. For example, endings for 1st person verbs are derived by adding -í to the vulgar register form because this is the only other source of 1st person masculine verbs.

Limits on vocabulary

Polite words are required at all times, whether addressing a male or a female audience. Males must avoid words referring to the female anatomy, even as elements of compounds. The forbidden words include not just terms for sex organs but also any distinctive female body part, such as the womb, breasts, and even long hair. Some words for female body parts consist of just a single vowel, and therefore appear within a large number of other words. For example, the most common word for shield is čáň, but because this word contains the letter á, which means "womb", men avoid it in favor of čàpoň, which has no transparent internal morpheme structure. This situation came about even though the original word for womb was not etymologically involved in the creation of the word for shield. Thus, all words in the standard language containing the letter á are forbidden for men, and men must learn special forms of each of these many words to use while also learning the standard forms so they can understand the speech of women. These special forms are not predictable from the structure of the original word. For example, hád is the standard word for rose, but men cannot use a word such as *hàpod to refer to roses because /àpo/ has no meaning of its own and cannot merely substitute for any /á/.

And this same situation repeats for other forbidden terms such as m "milk", n "moon" (due to the association with menstruation), ū "wide hips", l "egg; vagina" and for CV sequences such as zà and bà, both meaning "breast", and zì, meaning "vagina". Thus, all words in the standard language containing any of /m n á ū/, plus certain other words, are unusable for men, and must be replaced with alternate forms. However, the phonemes themselves are not forbidden, because a few of the men's replacement words happen to use the same sounds just by chance; because women do not use these replacement words, they are not considered to be in violation of the prohibition against mentioning female anatomy.

There is no imperative mood.

NOTE, the word for moon is actually /ʕn/, so this might not apply to most /n/. Note that /n/ is very common in the grammar.

On the other hand, true feminine /n/ arises from other sources. It is possible that the prohibitions could exclude stem-final instances of the sounds, because, for example, while a word for shield could plausibly incorporate a word for womb, it would not be the final morpheme in that word, since then the word would refer to a type of womb rather than an object which protects the wearers' wombs.

Also note that all voiced stops originate from nasal+stop clusters, and so the entire voiced stop inventory would also be forbidden except possibly /ġ/.

Deferential grammar

Males add one of several morphemes to all nouns and most verbs.

First person

Any verb with either a first person patient or a first person agent — that is, any verb involving the man speaking — takes a suffix that varies among -Vʲs ~ -v́ʲ ~ -ʲs ~ ́-ʲ . That is, there are three elements, /V/ + /ʲ/ + /s/, and at least two of the three must appear. If the /s/ does not appear, the preceding vowel (whatever its origin) must switch to the acute tone. Which of the four forms is chosen depends on the shape of the word and on context. Forms involving a second-person agent tend to use the /s/. Forms construed as first person tend not to; thus the first person masculine present tense verb ending is etymologically /óʲ/ but is pronounced and spelled /é/.

First person nouns sometimes use this and sometimes use -V:lé. This is a vowel lengthener (from earlier /n/), plus /l/, plus the 1st person verb suffix -o, which is run through the gradation /o/ > /é/ as above. This obviates the need for the 1st person inner possessive marker, so men addressing women simply say the equivalent of "collar, thigh, tray", and so on, rather than specifying "my collar, my thigh, my tray". But these nouns are no longer first person nouns when acted upon by an outside agent, so for example the word for thigh in a sentence meaning "the fly bit my thigh" is quite different from the word for thigh in "with my thigh i pushed open the door". Thus, for example, a man speaking to a woman would say sonālé for "my tray" in the nominative case, but soǯ + .... + las for the accusative, with no individual phoneme corresponding to the accusative case. This is because it is no longer a 1st person nouns.

Second person

Any verb with a second person agent or patient but not a first-person one will take a suffix that varies between -Vlas ~ -las. (This ending is sometimes also used for third person.)

These might be using different person markers.

There is also -(V)ʷḍ, which is semantically equivalent to the above but has a different meaning. This was caused by /nʷs/ > /nʷh/ > /ʷṇṭh/ > /ʷṇṭ/ > /wḍ/. This can also be used for the third person.

When adding this suffix to a word endin g in the 2nd person final /-s/, it disappears if the preceding sounds are /ta/ (that is, /tas + ʷḍ/ > /toḍ/), and if it is /ǯa/, ift becomes /ǯo/, so that, e.g. /ǯas/ + /ʷḍ/ > /ǯoḍ/. If it is /ča/, it becomes /čo/, so that /ča/ + /ʷḍ/ > /čoḍ/. Thus, the consonants never change.

THEREFORE, the rule /as/ > /oḍ/ generates the deferential form of any 2nd person /-as/ if that /-as/ means a 2nd person feminine noun or verb.

Second person nouns also use this suffix if the object is visible. Note that this includes nouns with second person agents, even if the possessor is first or third person. If the object is invisible, the suffix is -(V)ʲs. If the object is invisible and has a second-person possessor and no agent, this suffix is attached directly to the root, with no oblique form. This often causes an epenthetic /e/ to arise, from the normally suppressed /a/ followed by the /ʲ/.

Because the speaker and listener are encoded in the outermost morpheme of a word, the verb markers are bipersonal for this situation, if the verbs are transitive. That is, any -é, if transitive, means "I, a man, act on you, a woman". Any -(V)ses means "you, a woman, act on me, a man". However, these same forms would also indicate intransitives and verbs of other subjects so cannot disambiguate between situations involving more than one person. For example, the same é ending would be used if the patient was a man, if it was a 3rd person patient.

Third person

For verbs where both the agent and the patient (if present) are third-person, and likewise for nouns, males add one of three suffixes when addressing a woman. Two of these three incorporate the /o/ > /é/ shift as described above in the first person section. As above, these markers all appear outside the nominal case markers. When describing free objects, they are attached directly to the root with no intervening vowel, except for the shift of /ʲ/ > /e/ in some environments.

• The suffix -(V):ʔʲmé indicates an object or action the man is aware of but feels that his female listener may have a better view of the situation. Note that /:ʔ/ indicates an acute tone, and is equivalent to ´ but is morphologically distinct.
• The suffix -(V)hū indicates a situation the man is unsure of.

• This entails /n/ > /hū/ because of the shift /nlh/ > /nh/ > /ŋh/ > /h/.

• The suffix -(V)tàlé (analysed as tà + lé) indicates a situation the man believes is true because of prior thought but typically refers to invisible objects.

The suffix /-(V):ʔʲmé/ is impermissible for males in most public situations, and so is replaced with -(V)ʔʲkwé. However, the /tàlé/ word is permissible, even though it contains the forbidden phoneme /l/, because it arose as a substitute for the first word.

Context

Men continue to use the deferential register even to address their wives and other female relatives. However, women often instruct their men to use vocabulary from the direct register even while using the grammar of the deferential register, as some women find the vocabulary substitutions offensive, as if implying that Moonshine women are little more than menstruation machines.

Terse register

The terse register is used by women addressing trained animals and men. Many grammatical categories merge, and the information is simplified to minimize the length of the sentence. There is a polite imperative and a direct imperative. There are some defective verbs; for example, there is no 1st person form of the verb "to apologize".

Grammar

Many noun cases merge in the terse register. This is a continuation of a process that has affected the entire language, even in the most formal contexts, but women merge noun cases much more often when talking to men.

Because sentences in the terse register often omit information, men sometimes have difficulty understanding women who speak to them, but Moonshine boys learn early on that they need to listen closely and think hard whenever a woman speaks to them so that they can confidently understand what is required of them.

Context

Women continue to address their husbands and other male relatives in the terse register, as the direct register lacks terms of address for males. Because all verbs and many nouns in Moonshine inflect for the gender of the speaker and listener, any woman wishing to speak to a man in the direct register would need to instead direct her sentence to an imaginary woman.

Vulgar register

The vulgar register is used between males when the absence of females makes it safe to do so.

Grammar

The grammar is simpler than that of the direct register and shows many resemblances with that of the terse register, as men learn to imitate the speech of women that address them. Although many men dislike the imprecise vocabulary and grammar, Moonshine teachers keep the vulgar register the way it is because they do not want men to prefer conversations with men over conversations with women.

Nominal case marking

Scholars analyze Icecap Moonshine as preserving eight noun cases: nominative, locative, genitive, accusative, circumstantial, dative, essive, and instrumental. Very few words have distinct forms for all eight cases, and many have only two cases, but because different classes of nouns merge the cases in different ways, the eight-case analysis is convenient.

Obedience

All masculine and most neuter nouns are conjugated for obedience, meaning that their nominative forms are padded by person-marked feminine morphemes indicating which woman give them permission to act. Information here applies to masculine nouns only, which reliably end in -t and are thus much easier to describe than neuters.

There are four degrees of obedience; three of these result in a verb with feminine agreement, while the fourth offers a choice between masculine/neuter ("inferior") and feminine ("superior") verb endings. The obedience morphemes always imply a feminine external forcing agent. If third person, this behaves like an ordinary person marker and thus refers to the most recently mentioned third-person female agent in the conversation.

Person markers

The inner part of the obedience marker is simply a possession morpheme. The three markers are -ā -as -a for 1st, 2nd, and 3rd person feminine external actors respectively. The type of possession here does not refer to physical ownership, but simply to a general semantic relationship. Thus, for example, if a man talks to a woman, he becomes *her* man from the viewpoint of the grammar.

The possession morphemes are partly homophonous with the agent markers, but their internal structure is different and it is possible to mark both. Thus there are three situations: Ø-P, P-Ø, and P-P. (The agent markers are -o -as -à for most verbs, but not all.)

It is possible that the agent & poss markers will in fact merge, because they were only ever distinguished by presence or absence of an /h/ that disappeared after all masculine nouns since /th/>/t/. This would then have been analogized to neuters. The distinction between /o as à/ and /ā as a/ is actually secondary.

Male speakers make frequent use of the 2nd person form of the inner morpheme when addressing women, but it is not merely a fossil morpheme because they still will use the 3rd person form to indicate that they have done something without the permission of the woman they are addressing. When speaking to males, by definition the 3rd person is the only form available.

Note, check tjhis paragraph, it doesnt make sense unless it applies only to the 1st person pronoun

Degrees of obedience

The outer part of the obedience marker may be one of four morphemes. It may be Ø (absent), or it may be one of three morphemes that delete the final /t/ of the verb stem and then replace the possession markers with other morphemes. Thus, the morphemes are fused and not transparently segmentable. The four morphemes thus are:

Involuntary obedience

• Ø, which retains the possession markers -ā -as -a.

Male speakers do not use either the 1st or 2nd person form here for their pornouns, and female speakers use the 2nd person form only when speaking to other women.

Male speakers replace the 3rd person suffix /a/ with forms such as -ékwé when talking to women.

Primary obedience

• S₁, which deletes the final /t/ and changes the possession markers into /-čas -tas -taš/. This can also be analyzed as retention of /t/ and suffixation of -čas -as -aš, because any /tč/ collapses to /č/. This is the preferred analysis by Moonshine teachers anr scholars. Note that the second person form is identical to the one above.

This series of endings implies that the obedience was voluntary but still dependent on the female actor's permission.

• S₂, which does the same as S₁ but must be followed by a repetition of the final syllable of the verb of the sentence. This syllable cannot be predicted by the form of the noun because there are several sets of verb endings. It may be either masculine or feminine depending on the degree of obedience.

• This is directly cognate to Poswa's B-s-Ø-V verb inflection, and has a similar meaning. It signifies that the man is acting under his own power, without coercion, even while obeying another agent.

Secondary obedience

• J, which deletes the final /t/ and changes the possession markers into /-ča -tī -ʒī/. As above, this can also be analyzed as retention of /t/ and suffixation of -ča -ī -⚲ʒī, where the neuter symbol ⚲ indicates that the masculine marker /t/ must be removed after all.

In most social situations, males are not permitted to use the obedience markers implying that the listener coerced him to do the action mentioned in the sentence. Since the Ø and S₁ forms are merged in the second person, neither form is permissible, and males must use either S₂ or J.

There is no male speaker form of /ča/.

The male speaker form for a female listener of /tī/ is coḍ, from /tianus/.

The male speaker form for a female listener of /ʒī/ is ???. The unusual /ʒ/ here is evidently from PMS /ta-g-V/ collapsing to /tž-V/ and then on to /ǯ/ which shifted to /ʒ/ when before front vowels. This suggests that the masculine speakers' form might be ǯoḍ. However, even this is uncertain as it implies that men would be able to use the same marker for 2nd and 3rd person instead of using /hū/.

Other aspects

Female agents can also take the obedience morphemes, as can all other animate agents. Inanimate agents of active verbs are construed as acting obediently by default, and though the morphemes were historically present, they have fused to the other inflections and are no longer analyzable as such.

Verbs

Most strong verbs have two stems, deriving from the mobile stress of Gold. Weak verbs are historically compounds whose final element was monosyllabic in Gold and therefore had a fixed stress.

Many strong verb stems end in č or c; these are the reflexes of consonants that had already become silent by the time of proto-Moonshine, but reappeared in inflected forms and later came to be analyzed as part of the stem. Even so, some inflections delete these consonants.

Of the nine conjugations, four are restricted to occurring with female agents, three can only occur with males, and one occurs predominantly with neuter agents. The remaining conjugation, the sixth conjugation, is only available for feminine and neuter subjects, and thus there is no verb conjugation that can appear with all three genders. There is no one-to-one correspondence between a given feminine verb and its semantic counterpart in the masculine or neuter classes. Many masculine verbs use a derivational morpheme attached to the basic feminine verb, but the choice of which morpheme to use must be learned with each verb.

Gender marking on verbs

For some classes of verbs, male subjects must take a translation marker, -aḍu- or -es(l)-. These verbs include any verb whose stem contains a feminine morpheme, and most strong verbs regardless of their etymology. For example, the verb ná "to envelop" ends in the morpheme á "womb" and is therefore feminine. (It is also a strong verb.) There is no way to attach the masculine verb endings -č -š -Ø to this stem, because all three would simply disappear after the -č- that links most strong verbs to their person markers. Instead, the feminine oblique stem is used, followed by either -aḍu- or -es(l)- depending on circumstance, followed by a second set of verb endings, which each have special forms for each of the two translation morphemes.

Etymologically, these two morphemes are from the third person feminine present tense marker -a- plus the previously existing morphemes /ḍu/ and /ʲs(l)/.

NOTE, it is possible that ḍu just attaches directly to the verb stem with no /a/, and so this is a bed example. The other one, however, indeed requires a sequence of oblique + č + /es(l)/.

Strong verbs

Strong verbs are very complicated and many individual strong verbs are in a class of their own. In the modern language they are restricted to occurring with feminine agents, though one of the masculine agent infixes arose historically from the second conjugation.

First conjugation

The first conjugation contains verbs for female agents only. It is the source of the familiar -o -as -à person endings that appear on weak verbs. The corresponding past tense forms are -ač -aš -ī. Verbs in the first conjugation are those whose stems in Proto-Moonshine originally ended in a short /o/ (either high or low tone).

Male speakers replace the 2nd and 3rd person -as -à endings with -oḍ -?? when speaking to women, and sometimes use -oḍ for both forms, disambiguating the two by adding pronouns or repeating the word denoting the agent.

The verbs in this class are listed as ending in consonants because any final vowels came to be considered part of the stem. Some are single consonants, such as l "nurture, care for", ž "to smile, love, befriend", and š which means both "to count, study, stare at" and "to mimic, mock". There are also h "to see beauty, to cuddle, admire", č "to cover, stand over, jump over" and many others. There is no way to use these verbs with either a masculine or neuter agent, and they are considered inherently feminine verbs despite mostly lacking grammatically feminine morphemes.

If the first conjugation has an equivalent of the 8th conjugation's "verbs of influence", the linking consonant is /g/.

Second conjugation

The second conjugation contains very few verbs. It is the source of the /ʲs(l)/ infix that appears on some verbs with masculine agents, but has no independent use in the modern language because its vowels came to be the same as those of the first conjugation.

Third conjugation

The third conjugation is another class with very few verbs. Like the second conjugation, its verbs came to rhyme with those of the first conjugation, and therefore nearly all verbs in the third conjugation were transferred to the first conjugation.

Fourth conjugation

This is the first "Class II" conjugation of strong verbs. It is conjugated with CV suffixes rather than just vowels. These suffixes are -bi -mis -mì for the present tense and -be -fe -me for the past tense. These suffixes delete any final /-m/ in the verb stem, and that is why the fourth conjugation is considered strong.

Weak verbs

Fifth conjugation

The fifth conjugation is mostly used for verbs with neuter agents. For neuter agents, the present tense forms are -i -is -i (for 1st 2nd & 3rd person) and the past tense forms are all -e. There is also a linking -z- if the stem of the verb ends in a vowel. Note that 1st and 2nd person neuter exist only because of passive verbs; unlike related languages, IMS does not group small children into the neuter gender. The linking /z/ has become a part of the stem of some verbs that are typically used only with neuter agents.

The fifth conjugation also contains two sets of rarely used endings for feminine agents and one for masculine. The first set of feminine endings are -vi -ʲv -ʲṃ~-ʲv for the present, while the second set are -ṣi -ʲṣ -ʲṣ. The past tense forms are defective. For masculine agents, the present tense forms are -ṭi -cē -ʲc and the past tense forms are again defective.

With a gendered agent, any verb stem ending in a consonant loses that consonant and lengthens its final vowel. Thus, the fifth conjugation destroys the integrity of the verb, unlike other weak classes, and has become rare over time.

Neuter-exclusive verbs

For example, the verb nàz "to pierce with the claws" is in the fifth conjugation, and is only used with a neuter agent, typically an animal. The third person gendered forms would be *nèṣ nèc for feminine and masculine, and while not ungrammatical, would not be understood in ordinary connected speech.

Another neuter-exclusive verb is tàc "to fly", typically used of insects. In this case, adding feminine or masculine endings would result in the verb stem /té/, which few listeners would connect with the root. Humans who fly (metaphorically or physically) use different verbs.

Sixth conjugation

The sixth conjugation is a weak verb class that rhymes with the strong verbs of the first declension, but inserts an -ř- after the verb stem for female agents and -l- for neuter. Masculine agents do not use this class. It is the most common weak verb class for female agents. Some verbs in the sixth conjugation were transferred from the fifth conjugation; in such cases, any final /z/ or /c/ is dropped.

Seventh conjugation

The seventh conjugation contains verbs for male agents only. They are added directly to the verb stem because they are weak verbs.

In the direct register (female speaker, female listener), the present tense is -žřì and the past tense is -žře. The /ž/'s disappear after most stems ending in stops. There are never any epenthetic vowels; verbs in this class thus cannot end in clusters.

Eighth conjugation

The eighth conjugation contains verbs for male agents only. The 1st person present ending is bí if addressing a woman and ḍup if addressing a man. The corresponding past tense forms are ġí and ḍuk. The 2nd person ending is ḍus if the speaker is male. The third person imperative ("let him ...") is ḍo if the speaker and listener are both female. (Note that the /b/ and /ġ/ here are diachronically derived from contractions of /ḍup/ and /ḍuk/.)

In the direct register, the present tense is ḍù and the past tense is naš, but this /naš/ contracts to just /d/ when followed by a vowel of any origin. (These are found in the third person only because the direct register excludes men from being 1st or 2nd person.)

The eighth conjugation is derived from the /-aḍu-/ translation marker above, but it outgrew its original context and came to be used on weak verbs and verbs that did not contain feminine morphemes. These verbs thus have no feminine equivalents. It is the largest of the three men-only verb classes. Most verbs are rough semantic equivalents of verbs used elsewhere in the language but suggest a man doing something in a masculine manner, which in Moonshine culture has a variety of associations. For example, žal "dig" is an 8th conjugation verb and suggests the act of digging into the earth vigorously, with all attention focused on the task at hand. Women who dig simply use other words to describe it, because there is no grammatical way to force a masculine verb to have a feminine agent.

Words in this class mostly entered the language through the vulgar register, and although some words do contain the masculine phoneme /t/, most became eighth conjugation through semantic specialization.

NOTE, 8th conjugation verbs can have female "second-order" agents using -ṭ-, so e.g. "I/you/she influenced him to dig", and lesser-used paradigms for the 1st and 2nd person masculines. However, these constructions should probably also exist for 1st conjugation feminine verbs.

Ninth conjugation

The ninth conjugation contains verbs for male agents only. It is sometimes considered the masculine form of the sixth conjugation, and its verb endings are -lap -lač -lačā in the present tense and -lakī -lakī -laf in the past.

Pronouns

Icecap Moonshine is one of few cold-climate languages to have pronouns. The use of pronouns varies by speech register.

Pronouns used by women

Feminine pronouns used by women

The female pronouns were originally a tenth verb conjugation. Their basic forms are bū bus bas for first, second, and third person feminine. Women also

Masculine pronouns used by women

The masculine pronouns are similarly derived from a verb. Their basic forms are sis zis for 2nd and 3rd person. Note that these are not the pronouns men use for themselves or for other men.

Pronouns used by men

First person masculine pronouns

When addressing women and mixed groups, the commonest 1st person masculine pronoun has the basic forms ǯā ~ das , from the root ʔdà. However, these must put obedience and politeness suffixes on. The obedience suffixes will turn /das/ into das ~ das ~ das-V ~ dī.

When addressing men, different words are used. One is pšā. The stem of this word is /peř/. The instrumental form is pàs.

Second person feminine pronouns

The common root ges appears here, used for women in general, including intimate relations. It alternates to his ~ ʷsis , etc.

Second person masculine pronouns

Men tend to address other males with words that are transparently derived from the word for "listener". This is because there is no way to put masculine suffixes onto the words for "man" and "boy" without mentioning a female third party.

Vocabulary

Moonshine's sound changes are so extreme that many morphemes are a single letter long, and many of these are consonants. Thus Moonshine can be described as an oligosynthetic language. The majority of words in the vocabulary are longer than this, with the commonest form being CVC, but even these CVC roots have often been reanalyzed by the speakers as compounds of the type C + VC, CV + C, or C + V + C. Thus, the number of true oligosynthetic roots is small, but the pattern is common enough that longer words have been adapted to the oligosynthetic grammar and can participate in sound substitutions that are not etymologically justified. An example of this is the root čāl "frame, box, surroundings", which is etymologically a single morpheme, but has been reanalyzed as č-ā-l "building hip rectangle" (hips being the widest part of one's body), and thus generated new variant forms.

Circumfixes

Another type of compounds involve C-C morphemes in which a gap must be filled by inserting a vowel. Sometimes, a consonant can precede this vowel, resulting in a root with the structure CCVC. These are typically analyzed as circumfixes surrounding an atomic morpheme because the circumfixes cannot appear independently or even with a null morpheme.

GENERAL WORD STRUCTURE

This section is not well placed and there is no convenient place to put it.

All words in IMS have a familiar structure, but the morphemes fuse differently in different word classes, and sometimes even within word classes, so it is difficult to summarize with a chart of affixes.

There are probably no true prefixes in the language, but some word roots can be analyzed as circumfixes in which two consonants surround a vowel, and although both have transparent meanings, neither can be used alone because a CC syllable with no vowel is illegal and a vowel-only syllable would not be able to attach to its inflections.

Of Nouns

ROOT + classifier + owner marker + CASE INFLECTION + agent marker + DISCOURSE MARKER

It is very difficult to draw boundaries between these morphemes, but the order is reliable. For example, the root and classifier are often fused, the classifier may fuse with the owner, the owner may fuse with the case marker, the case marker may fuse with the agent marker, and the discourse marker can fuse with the entire preceding syllable.

The most common grouping is

ROOT + [classifier + owner marker] + [case inflection + agent marker] + discourse marker

But some combinations trigger

ROOT + classifier + [owner marker + case inflection + agent marker] + discourse marker

Note that in both cases the discourse marker is depicted as being alone but properly it fuses with the entire syllable.

DUSION WORKS LIKE THIS:

The root can be any shape.

The classifier is always a consonant, though it can sometimes affect the final consonant of the root if there is one.

The owner marker is always a vowel, but can trigger changes in the consonant of the classifier.

The case inflection can be any shape, though case markers are a closed class.

The agent marker is always a vowel, but can merge with the case inflection.

The discourse marker interplays with the preceding syllable, whatever it is.

Dialects and variation

The standard language as of 6843 AD was based on the dialect of Enapded, historically home to the Cartwheel party. This was the scholarly capital of the empire, but had no significance politically. The political and military capital of the empire was Safiz, and the cultural capital was sometimes considered to be Todʔrom. Neither of these two other states contributed much to the standard language. But higher living standards led the relatively small state of Enapded to dominate the empire's centralized education system, and thus the language spread from there.

Scholarly language

The centralized school system based in Enapded drove out the other dialects. Even scholars showed little interest in preserving the old languages, and they continued to teach what they called Classical Moonshine, the state of the language as it had been in 6843 AD, even as the vernacular continued to change rapidly.

Later dialects

See Future Moonshine.

Interaction with Poswa

Because of the speakers' isolation, the only language which had a major influence on IMS was Poswa, and likewise, Moonshine was the only language that had a major influence on Poswa. Yet, the two have little in common. IMS actually cut through the Poswob homeland, but the speakers separated themselves by habitat and neither considered the other to be in violation of any territorial rights. Poswobs lived by rivers and lakes found on plains whereas Moonshines lived on mountains and cold windward retreats.

It is said that even the speech register that Moonshine women use to speak to their men is more polite than the common speech used by all speakers of Poswa.