Dreamlandic grammar

The empire of Dreamland spoke a large variety of languages due to the many small, isolated settlements along the coast whose people had little contact with other Dreamers. However, all of these languages were closely related and shared many characteristics. Most languages belonged to the Dreamlandic family, but there were some longstanding settlements in Dreamland populated by speakers of other language families such as Hipatal.

Background

Lenian people began settling what is now Dreamland from the west around the year 1320 AD. They founded a major colony at about 26°N, 25°W. From here, they quickly divided into two groups of people: those who lived along the coast and survived by fishing the sea, and those who moved inland and survived by hunting and gathering, as well as fishing in lakes and rivers. There was no middle ground and both groups were migratory. In this way, the early Dreamer state strongly resembled Paba.

Unlike Paba, however, the hunting tribes and the fishing tribes were both of the same ethnicity: the blonde blue-eyed Lenian people. Therefore, members of one group could quickly assimilate into the other group, and the tribes maintained friendly relations even as they became more physically isolated.

For the most part, the hunting tribes found success in the hot, humid climates of the southern side of the peninsula, and formed the Sepesi branch of the family as they moved eastward along the coast towards the ancient city of Baeba Swamp. Soon, they signed a pact with Baeba Swamp and promised to pursue pacifism.

The fishing tribes lived in smaller, more concentrated colonies, and spent less time outside their territory. Therefore they divided into many small branches early on, with each one later becoming just a single state in the empire of Dreamland. However, the people furthest to the west did not join Dreamland and in later times came to consider themselves superior to the Dreamers.

Major languages

See Lenian languages for sound changes.

The most prolific branch of the family was North Dreamlandic, whose speakers split off from the others in the year 3370, and then followed the coast northward and reached 46°N within just a few hundred years. Their northern fishing settlements were poorly defended and they came to live as minorities within other empires, but further south, the new cities of Posensene and Enoneta appeared and came to dominate the empire of Dreamland.

Minor North Dreamlandic languages

These languages started with a syllable inventory of

  a    i    u    ya   wa   yi   wu
 pa   pi   pu   pya  pwa  pyi  pwu
 ma   mi   mu   mya  mwa  myi  mwu
 ta  (ci   cu)  tya       tyi
 na   ni   nu   nya       nyi
 sa   si   su   sya       syi
 la   li   lu   lya       lyi
 ra   ri   ru   rya       ryi
 ka   ki   ku        kwa       kwu
 ŋa   ŋi   ŋu        ŋwa       ŋwu
 ha   hi   hu        hwa       hwu
 fa   fi   fu
mpa  mpi  mpu  mpya mpwa mpyi mpwu
nta (nsi  nsu  nsya      nsyi)
nka  nki  nku       nkwa      nkwu
ppa  ppi  ppu  ppya ppwa ppyi ppwu
kka  kki  kku  kkya kkwa kkyi kkwu

Proto-Dreamlandic (1495) to Tasasina

This language remains with three vowels longer than the others. This is located immediately north of the Dolphin Rider birthplace and therefore may be intimately associated with the Dolphin Rider language, Neamaki.

1. The sequences kwa kwu ŋwa ŋwu shifted to pwa pwu mwa mwu.
2. All long vowels became short.
3. The sequences nsi nsya nsyi shifted to nyi nya nyi.
4. The sequences si sya syi shifted to si sa si.
5. Then, the sequences ha hi hu hwa hwu became a yi bu bwa bwu.
6. The sequences lya rya merged as ya. Then lyi ryi merged as yi.
7. The sequences nya nyi (including from earlier /nsya nsyi/) shifted to na ni.
8. The sequences fa fi fu became ba bi bu.
9. Remaining h disappeared to Ø.
10. The sequences pw bw mw shifted to plain labials p b m.
11. All remaining s shifted to h.
12. The sequences pya pyi mya myi shifted to ta ti na ni.
13. The sequences tya tyi kkya kkyi shifted to sa si ssa ssi.
14. The sequences ci cu became si su.
15. Remaining ns shifted to nz.

The syllable inventory at this point consisted of 61 glyphs:

  a    i    u        
 wa   wi   wu
 pa   pi   pu      
 ba   bi   bu
 ma   mi   mu   
 ta   ti   
 sa   si   su
 na   ni   nu    
 la   li   lu    
 ra   ri   ru
 ya   yi   yu
 ka   ki   ku      
 ŋa   ŋi   ŋu      
 ha   hi   hu 
mpa  mpi  mpu   
nta  nti      
nza       nzu
nka  nki  nku             
ppa  ppi  ppu   
tta  tti
ssa  ssi  
kka  kki  kku    

The sequence /nzu/ was rare. There was also a rare syllable-final /n/. The consonant inventory can be analyzed as /p b m t s n l r k ŋ h/.

The sequences /ia ua/ had emerged from earlier /iha uha/, but there was no new bare /i/ or bare /u/.

Proto-Dreamlandic (1900) to Tata-A

Makes long vowels from heavy syllables.

Proto-Dreamlandic (1900) to Tata-D

Treats geminates as single consonants.

Proto-Dreamlandic (1495) to Tata-B

This may be spoken in Tata, but is too early to have been a Matrix language.

1. The alveolar affricate č shifted to š.
2. The labials pʲ mʲ fʲ f became h m h h.
3. The labiovelars kʷ ŋʷ shifted to k ŋ.

Sepesian languages

These are the languages of Sepesi on the south coast of Dreamland. The speakers here are in intimate contact with those of Baeba Swamp and parts of Kuhilani.

This branch might keep or even elaborate on the classifier prefix system, but probably still does not reach the levels of the Andanese or Subumpamese inventories.

Proto-Dreamlandic (1495) to Suds (~3550)

1. Labialization was defeated, leaving no effects on the vowels.
2. The sequence ya shifted to ye.
3. The sequence mp shifted to mm.
4. Any singleton p shifted to Ø. (This includes /pu pi/ > /w y/.) Any preceding vowels became long. (But note that īy ūw were equivalent to ī ū.)
5. The geminates pp čč kk shifted to p č k.
6. Before a vowel, the sequences ki ti pi hi fi merged as s. Then tu shifted to su. (This also implies /č/ > /š/ > /s/, as in Sessi.)
7. Any remaining ki ti pi merged as the consonant geminator ʔ, which shifted to /Ø/ in absolute final position; any remaining hi fi lengthened the preceding vowel (ː) and then disappeared to Ø.
8. Long vowels before a geminate consonant became short.
9. The sequences ll rr shifted to l r.
10. The sequences mi ni ŋi li ri shifted to n n n i i. Sequences like /miwa/ > /nwa/ came to be spelled /nua/, etc. This shift did not affect the long ī, which was equivalent to /iy/.

    A yer-like process must appear here to prevent the syllables from bunching up on each other. Since ī was rarely word-final, it is likely that word final -i will survive here and that therefore /mimi/ > /mmi/, etc. Word-initial geminates had already existed in the parent language.

11. The sequences lu ru shifted to u.
12. The sequences nfa nfu nha nhu changed to mpa mpu na nu.
13. The sequences fa fu shifted to ha hu.
14. The sequences kī tī pī merged as si. Then mī nī ŋī merged as ni. Bare ī shifted to i.
15. The sequences ui ūi merged as i.

    This means old kui, pui, kūi, etc all become consonant + /i/.

16. The sequences ūa ūu became ua uu . This also caused /uwa/ > /ua/.

    Note that nearly all /ū/ had been before another vowel even in proto-Dreamlandic.

17. The sequences yi wu shifted to i u. (The surface realization of the two merged, and either can be thought of as the basic form.)

Baeba Swamp crushed PDP territory here and the population switched to speaking Khulls.

At this point the alphabet consisted of a very small number of glyphs:

  a    i    u
 wa         
 pa   pi   pu
 ma   mi   mu
 ta      
 na   ni   nu    ne
 sa   si   su    se
 la    
 ra    
 ya              ye
 ka   ki   ku
 ŋa   ŋi   ŋu
 ha   hi   hu

However, unlike neighboring languages, syllables could end in /n t/, which assimilated Palli-style to a following consonant.

Common vowel nuclei were /a i u ā ai au āi āu ua uu ia ii iu/. There were also /e ei eu/, but these could only appear after /n s y/ and never in bare form. Also, there was no */ea/.

The sequences /ia ua/ can be considered equivalent to the bare syllables /ya wa/ because /ya wa/ never occurred after consonants.

The alveolars /t l r/ only occurred before /a/, but this /a/ could precede /i/ or /u/, and also could be long, so a five-way contrast of /a ai au āi āu/ was maintained.

Distribution

This was the language of the Soap party, also known as Suds. They were pacifists who lost a war against the Crystals and therefore their language became a substratum of the Crystal language. The Crystals opposed slavery, and therefore the Suds were allowed to move out, but many stayed and resigned themselves to living as an underclass in Baeba Swamp. This was because, as pacifists, they opposed both sides of the war that had cost them their homeland, and thus considered themselves equally at home among their enemies as among their former allies.

Because the language lost most /p/ early on and also lost some /m/, it does not have the same character as the northern languages, DPR and Baywatch. But the restricted distribution of the coronals /t l r/ made the language stand out from all of its neighbors even so, and when the Crystals took control of the Soapies' homeland, the Soap language rapidly evolved a more phonologically stable inventory.

Proto-Dreamlandic (1495) to Sesesi (~2700)

The end date might be earlier, because this originally contained additional sound changes under influence from Baeba Swamp.

1. Labialization was defeated, leaving no effects on the vowels.
2. The sequence ya shifted to ye.
3. Before a vowel, the voiceless sequences pu p pi shifted to w Ø y. (This includes mp > mm > m). Any preceding vowels became long. (But note that īy ūw were equivalent to ī ū.)
4. The geminates pp čč kk shifted to p č k.
5. The sequences hi fi ti tu shifted to si si si su. Then all č became š.
6. Any remaining t shifted to r.
7. The fricatives f š shifted to h s.
8. Before a vowel, the sequences pi mi ni li ri (including from earlier /mbi/) shifted to s n n l r .

   added /ni/ because there was no /ñ/ in the target phonology.

9. The prenasalized sequences nr ns ŋk shifted to n s k. Adds vowel length?
10. The sequences aw ew iw uw shifted to ō ō ū ū. Thus a new vowel, /ō/, was created; however there was no short counterpart. This shift also included tautosyllabic /au/, /eu/, etc.
11. The sequences ay ey iy uy shifted to ē ē ī ī.
12. The double vowel sequence aa shifted to ā.
13. Long vowels followed by another vowel became short. Thus the new sequences like /ōa/ became /oa/, and the reflexes of earlier /ipia/ merged with /ipa/, and so on.

Thus the alphabet consisted of 39 syllables:

  a    i    u    
 wa        wu   
 pa   pi   pu         
 ma   mi   mu       
 na   ni   nu   ne
 sa   si   su   se
 la   li   lu   le   
 ra   ri   ru   re  
 ya   yi        ye 
 ka   ki   ku        
 ŋa   ŋi   ŋu    
 ha   hi   hu   

The short vowels were /a e i u/ and the long vowels were /ā ē ī ō ū/.

This branch divides less rapidly than the others. There will be /s h/ > /0/, and may be /uk ku/ > /up pu/ and /ik ki/ > /it ti/ (through k > ć > ṭ), but only if these consonants can be separated from this context by later vowel shifts.

Sesesi to Valley of the Minds

This language is spoken in the northern part of Dreamland, by people who made it to the north before the speakers of North Dreamlandic did. It may have survived in a valley with one major city. Also, these sound changes might apply to other languages of the family as well, since VoM did not break off from near the root.

1. When not before a vowel, the sequences se ne le re shifted to the consonants s n l r. Thus closed syllables were created.
2. The long vowels ē ō decomposed to ay aw.

Western languages

The western branch broke off around 1300 AD, and then lost contact with the others.

Changes unique to Western Fojy

1. The voiceless non-sibilant fricatives f fʲ h hʷ shifted to Ø Ø Ø w.
2. The palatalized labials pʲ mʲ shifted to p m.
3. The bare vowel u shifted to ʉ. Then wa shifted to wo. Neither of these shifts were phonemic.

The language at this stage had exactly 100 syllables, if the clusters pp kk mp nt nk are counted as single onsets. This makes WF one of two related languages to evolve a 100-character syllabary early in their history, the other being the very early stages of the Gold language.

However, the 100 syllable analysis ignores the rare independent final nasal /n/. There were many vowel sequences, such as /ie/, as well as inherited long vowels which were now written as doubles.

The syllabary consisted of signs for

  a    i    ʉ    ya   wo   yi   wu   ye
 pa   pi   pʉ   __   pwo  __   pwu   pe
 ma   mi   mʉ        mwo       mwu   me
 ta  (ci   cʉ)  tya       tyi       tye
 na   ni   nʉ   nya       nyi       nye
 sa   si   sʉ   sya       syi       sye
 la   li   lʉ   lya       lyi       lye
 ra   ri   rʉ   rya       ryi       rye
 ka   ki   kʉ   kya  kwo  kyi  kwu  kye
 ŋa   ŋi   ŋʉ   ŋya  ŋwo  ŋyi  ŋwu  ŋye
ppa  ppi  ppʉ       ppwo      ppwu  ppe
kka  kki  kkʉ  kkya kkwo kkyi kkwu kkye
mpa  mpi  mpʉ       mpwo      mpwu  mpe
nta (nci  ncʉ) ntya      ntyi      ntye
ŋka  ŋki  ŋkʉ  ŋkya ŋkwo ŋkyi ŋkwu ŋkye

In the year 3958, one of the languages of Western Fojy (here "WF", or Wildfire) took over all of Dreamland, only to be swept out by the Dolphin Rider language in 4108.

Proto-Western Fojy (1900) to Wildfire I (3958)

This language loses its /ʉ/, creating an unbalanced setup where most of the consonants cannot occur before a back vowel.

1. The weak vowel ʉ disappeared to Ø, creating closed syllables.
2. Triple consonant sequences became double.

Baywatch language

Phonology

For diachronics, see Lenian languages.

Bilabials:       p   m           b   
Alveolars:       t   n   s   l   r     

Vowels are /a e i o u/ in both short and long forms. Syllable structure is CVC, but words can begin with geminates and thus the structure could be analyzed as CCVC. All clusters are homorganic, and the only consonant that can occur in absolute final position is /n/.

There are no dorsal consonants; instead, the coronals /t n/ are realized as [k ŋ] before any /u/, and for some speakers, also before any /o/.

DRM particles

• a "in, at" (locative)
• e "using" (instrumental). Mutates to o if the following noun begins with /o/ or /u/

The particles a and e can be thought of as verbs, in the sense that they must immediately precede a noun, and take verb-like inflections. But the copula must come after this if there is no main verb.

• nē verb copula

DRM verb prefixes

Because SOV and SVO word orders were both legal, the verb prefix often alliterated with the object prefix, and this led to analogical removal of whichever occurred last in a given sentence.

This list is arranged in the order the morphemes appear in the sentence.

DRM tense markers

• se past tense marker. Merges with following vowels, such that se-o > sō, etc. Because many verb stems begin with vowels, this mutation is common. Because the tense marker comes before the case particles as well, these case particles merge.

The list of mergers is:

1. e-a > ē, as in sē bama "went inside"
2. e-e > ē, as in sē tare "planned"
3. e-i > ē, as in sēsēmpo "made peace"
4. e-o > ō, as in sōpu "sat on"
5. e-u > ō, as in sōbi "felt"
6. e-ā > ē, as in (no examples)
7. e-ē > ē, as in (no examples)

This morpheme originally began with /ns/, but the /n/ was lost at a stage when the child agentive prefix n- still echoed before the verb.

Person markers

• ne 1st or 3rd person human patients
• mpu reflexive verb prefix

• isolated forms with /mu/ may also exist

DRM definite articles

These are also descended from noun classifier prefixes. They mark both gender and noun case. Since the plural of all animate genders had merged with the epicene by this time, and number had never been distinguished for inanimates, in a sense the articles also mark number.

Intransitive

• mi female experiencer
• si male experiencer
• ni maiden experiencer
• n child experiencer

These are also used for possession of objects.

Agentive

• mi female agent (if patient is male, maiden, or child)
• si male agent (if patient is female, maiden, child, or plural)
• ni maiden agent
• n child agent

The forms above are the same as the intransitives. However, there are more:

• i female agent (if patient is female)
• pi female agent (if patient is plural)
• se male agent (if patient is male). Mutates to so if the following noun begins with /o/ or /u/
• ne variant child agent (if patient is female, male, or child). Mutates to no if the following noun begins with /o/ or /u/
• me variant child agent (if patient is male). Mutates to mo if the following noun begins with /o/ or /u/

• bu plural agent. Shifts to pu if the following noun begins with /p/ or /t/

Patientive

• ni maiden patient
• n child patient

The forms above are the same as the intransitives. However, there are more:

• ne 1st person patient

• mpi female patient
• e male patient. Mutates to o if the following noun begins with /o/ or /u/

  In an early draft, this became Ø if the instrumental particle /e/ appeared anywhere earlier in the sentence.

• ppu plural patient. Shifts to pu if the following noun contains another historical consonant cluster (not always transparent due to subsequent shifts)

Those agentive prefixes that also uniquely specify the gender of the patient require the speaker to omit the patientive prefixes. Thus one can say

Me mposu pempure nsebise.

The duck followed the soldier.

For emphasis, a prefix can be used, but it must be a repetition of the agent's prefix rather than a retention of the patient's, and the emphasis is not on the patient but on the agent's control over the patient. Thus the above sentence can be reworded as

Me mposu me pempure nsebise.

The duck followed the soldier.

The repetition of the prefix implies that the soldier may be an unwilling party whereas in the former sentence there is no such implication. The otherwise expected formula, Me mposu *e pempure nsebise, is not used.

Note that the prefixes for children and maidens are the same in all cases.

Note: try to get out of the habit of using SOV.

Use of noun prefixes

The noun prefix is mandatory in sentence-initial position, and therefore there is no distinction between definite and indefinite subjects. This is common in related languages, even those with very different grammars. Thus one must say

Si peno pu pempā nsōnse.

The (male) goat ate the grass.

A (male) goat ate the grass.

Gender and syntax

Gender marking on humans is strictly syntactic, and becomes more grammatical (non-literal) as the objects described become less animate. The maiden and child genders are retained from the proto-language and remain fully functional, but in some areas of the grammar they merge with each other.

Feminine gender

The feminine gender includes words for women and women's property. It also includes words for girls, but younger girls more often take words belonging to the maiden gender. Higher animals are also described according to literal gender. There are also many inanimate objects in the feminine gender, all deriving from various processes of analogy.

Maiden gender

The maiden gender includes words for girls of pre-marriageable age. It also includes many words for immature female animals, and for females of some small animals regardless of maturity. There are many inanimate objects as well.

Child gender

Sometimes called neuter, the child gender includes words for very young children not capable of living independently. In the high register of the language, some verbs are ungrammatical if the agent is a child, but this rule is ignored in colloquial speech. The child gender also includes many words for animals and inanimate objects, and can be called the neuter gender in this context.

Because the prefix for the child gender in most environments is a simple n-, it cannot occur before a word beginning in a cluster or one of l- r-. Such words thus must use one of the other animate genders even when describing small children and other neuter objects. For example, a student might expect the words liri "goldfish" and mpoli "octopus" to be neuter like other sea life, but in fact they are found in the epicene gender.

This phonological restriction also prohibits the neuter possession marker from occurring before the names of objects beginning with these consonants. Thus, for example, it is impossible to say *n mpempe "the baby's rock", because this would be pronounced with /mmp/, which is illegal. The solution here is to use one of the other animate genders, corresponding to the child's closest semantic gender. This is maidens for girls and men for boys. Because gender is inherited from the possessor, the object then becomes this gender itself even though its possessor is a baby.

Note that classifier prefixes are repeated before the objects, so this can cause the initial word to also change. This causes babies to gain semantic gender and requires the speaker to know the syntactic gender of the baby being referenced. Thus, though Baywatchers are accustomed to saying things such as

N pose n pepepo.

The baby's diaper.

To reference "the baby's clothes", one would need to pick a gender and either say

Si pose si ppe.

The boy baby's clothes.

Or

Ni pose ni ppe.

The girl baby's clothes.

Masculine gender

The masculine gender includes words for men and boys, and for masculine property. This includes many words for tools and weapons. Males of higher animals are also assigned the masculine gender. There are relatively few words in the masculine gender that are not tied to men and men's habits.

Plural gender

Also called epicene, the plural gender encompasses all words for groups, even if the group is internally homogeneous. It also includes all mass nouns and some inanimate objects perceived as indifferent to number.

Pu pperi o ontu nsōnē.

The rain fell on the boy.

Si nano pu mpobe nsenessi.

The king spoke to the people.

Dynamic gender assignment

As in most related languages, inanimate objects take on the gender of their owners when they are used in a possessive construction. Thus one can say

Mi pempi.

Her umbrella.

Si pempi.

His umbrella.

Ni pempi.

Her umbrella. (Owner is a young girl)

N pempi. (pronounced /mpempi/)

The child's umbrella.

Repetition of classifier prefixes for possessed objects can cause mutation of the agent marker. Thus one says

Si ontu.

The boy,

But

So ontu so pensipe ēponsi.

The boy is wearing his mask.

Here, si shifts to so because the sentence is transitive.

Dolphin Rider language

DPR Phonology

For diachronics, see Lenian languages.

Bilabials:       p   m           b   
Alveolars:           n   s   l   r  
Velars:          k   

Vowels are /a e i o u/ in both short and long forms. Syllable structure is CVC, but words can begin with geminates and thus the structure could be analyzed as CCVC. All clusters are homorganic, and the only consonant that can occur in absolute final position is /n/.

The velar stop /k/ is realized as [t] before any /e i/. The phoneme inventory is the same as that of Baywatch, but the primary allophone of the dorsal stop is [k] in Dolphin Rider and [t] in Baywatch, leading to a different surface inventory.

Syllable structure is strictly CV.

DPR verb markers

Tense markers

• re past tense prefix

Voice and aspect

• mu reflexive verb prefix

• Possibly retained as /bu/ in some fossils

DPR particles

• a "in, at" (locative)
• e "using" (instrumental)

These are descended from noun classifier prefixes.

Nu erobi nē a mabo.

The lobster is in the ocean.

Note that the copula is usually found before the object, creating an SVO sentence. This sets Dolphin Rider apart from nearly all other languages on the planet.

DPR definite articles

These are also descended from noun classifier prefixes. They mark both gender and noun case. Since the plural of all animate genders had merged with the epicene by this time, and number had never been distinguished for inanimates, in a sense the articles also mark number.

Unlike those of Baywatch, the patient particles sometimes depend on the agent particles, and thus interactions work in both directions.

Intransitive

Feminine

• mi female experiencer

Masculine

• si male experiencer

Maiden

• ni maiden experiencer

Neuter

• nu neuter experiencer

Plural

• Possibly pu

DPR Agentives

Feminine

• mi female agent (if patient is male, maiden, or neuter)

• mi ... e female ---> male
• mi ... ni female ---> maiden
• mi ... Ø female ---> neuter
• mi ... pu female ---> plural

• e female agent (if patient is female)

• e ... bi female ---> female

Plural/epicene

• u plural agent

• u ... bi plural ---> female
• u ... e plural ---> male (PROBABLY)
• u ... ni plural ---> maiden
• u ... Ø plural ---> neuter
• u ... pu plural ---> plural

Maiden

• ni maiden agent

Masculine

• Probably si

Neuter

• no neuter agent (if patient is female)

• no ... Ø neuter ---> female

• me neuter agent (if patient is male)

• me ... Ø neuter ---> male

• nu neuter agent (if patient is neuter)

• nu ... Ø neuter ---> neuter
• ne variant neuter agent (if patient is male or neuter)

Patientive

• bi female patient

• e bi variant female patient (if agent is male)

• no female patient (if agent is neuter and sentence is OSV or OVS)

• ni maiden patient

• e male patient
• se male patient (if agent is male)
• me male patient (if agent is neuter and sentence is OSV or OVS)

• pu plural patient

• Ø neuter patient
• nu neuter patient (if agent is neuter and sentence is OSV or OVS)

In the formal register, masculine nouns are prohibited from being the agents of verbs with female patients, and thus must take an additional /e/ before the patient's article. This can be analyzed as either creating a null feminine agent or masculinizing the patient, since the female agent prefix and the male patient prefix in this narrow context are both /e/.

Gender and syntax

Syntactically inanimate objects are mostly found in the child gender, but mass nouns and some others are in the plural gender, which is also known as epicene.

Feminine gender

Contains words for women of marriageable age, and a few words for girls.

Maiden gender

Contains words for girls and a few words for women.

Masculine gender

Contains words for men and boys, and masculine property.

Neuter gender

This is DPR's counterpart of DRM's child gender. It contains some words for small children, but most neuter words are for animals and objects. DPR places more higher animals, such as birds, in the neuter gender than does DRM.

Unlike the other genders, the agent prefix for the neuter gender is dependent on the gender of the patient. Thus all sentences with neuter agents can be considered passive.

Sentences with neuter agents are often OSV or OVS. When this happens, the neuter agent prefix remains at the beginning of the clause, since it can also serve to mark the patient. Thus just as one says

Mi bupue bolimo rerasi.

The girl caught an octopus.

One would also say

No bupue bolimo rerasi.

The girl was caught by an octopus.

Or

No bolimo bupue rerasi.

The octopus caught a girl.

If both the patient and the agent are neuter, then the sentence is ambiguous, and context is necessary. The normal word orders are SOV and SVO, but in a connected narrative, a subject may be the patient. Thus one can say

Nu bopusepu siselepi repobo.

The eagle bit a rabbit.

But in a story about a rabbit, the opposite wording could be used:

Nu siselepi bopusepu repobo.

The rabbit was bitten by an eagle.

Combining forms

Possibly fuse definite articles and other morphemes to their roots, such as nu erobi > narobi. This would require analogy with the way words had been spelled 4,000 years earlier, but the analogy could have been made in proto-Dreamlandic and then gradually lost from some branches and not others.