Project Tumbleweed

A forum for all topics related to constructed languages
Post Reply
User avatar
Posts: 490
Joined: 18 Nov 2019 07:35

Project Tumbleweed

Post by VaptuantaDoi »

I've decided to go for something a bit different (for me at least) and make an Australian-inspired language family with big consonant inventories and plenty of agglutination. I'm calling it Tumbleweed because I think it sounds cool and I can't think of anything else to call it.


1. Phonology
Tumbleweed is something different for me, inspired by Australian languages. It has a lot of consonants, not many vowels and strict restrictions on word formation. Other distinctive features include three series of stops with no plain voiceless ones, eight pre-stopped consonants (inspired by Kaurna) and uvular nasals.

1.1 Consonant inventory
Proto-Tumbleweed (PT) had 46 consonants

/d ɖ ɟ ɡ ɡʷ ɢ ɢʷ/
/n͡t ɳ͡ʈ ɲ͡c ŋ͡k ŋ͡kʷ ɴ͡q ɴ͡qʷ/
/n͡d ɳ͡ɖ ɲ͡ɟ ŋ͡ɡ ŋ͡ɡʷ ɴ͡ɢ ɴ͡ɢʷ/
/d͡l ɖ͡ɭ ɟ͡ʎ ɡ͡ʟ/
/d͡n ɖ͡ɳ ɟ͡ɲ ɡ͡ŋ/
/l ɭ (ʎ)/
/n ɳ ɲ ŋ ŋʷ ɴ ɴʷ/
/ɾ ɻ j ɰ w ʁ ʁʷ/

/ʎ/ only occurred in suffixes as a result of laminal harmony. The three series of coronals were laminal dental, apical retroflex and laminal palatal.

1.2 Vowel inventory
PT only had two vowels

/i ɑ/

These likely had a lot of allophonic variation, with /i/ being [i~e~ɛ] and /ɑ/ being [a~ɑ~ʌ]. There were no rounded vowels. Although there were no diphthongs, there was the sequence /ɑji/ which may have become monosyllabic [ɑe̯] allophonically.

1.3 Word structure
Word structure was quite restrictive.

1.3.1 Syllable structure
Word-initially, only approximants (/ɾ ɻ j ɰ w ʁ ʁʷ/) or stop-initial segments (/d ɖ ɟ ɡ ɡʷ ɢ ɢʷ d͡l ɖ͡ɭ ɟ͡ʎ ɡ͡ʟ d͡n ɖ͡ɳ ɟ͡ɲ ɡ͡ŋ/) could occur; word-internally any consonant could occur. Every syllable was CV with an obligatory onset.

1.3.2 Word length restrictions
All words had to be an even number of syllables. This meant that all nouns were even-syllable roots, and all verbs were odd-syllabled roots with obligatory odd-syllabled affixes.

1.3.3 Laminal harmony
In a word, only consonants from one series of laminals (i.e. dental or palatal) could occur; this caused assimilation of affixes which is the only source of /ʎ/.

1.3.4 Vowel harmony
All vowels other than the first vowel in a word had to be the same, including suffixes. This means that all suffixes are unspecified for vowel quality and just assimilate to the second vowel of the root.

1.3.5 Word structure strings
Word structure can be defined as CaVaCbVb(CcVbCdVb... CeVbCfVb), such that Ca is any permissible word-initial consonant.

1.4 Prosody
There was probably some form of strict prosodic pattern in PT which incentivised the even syllable word length, but I’m not sure what it was yet. I’ll have to read up on Yidiny.

1.5 Orthography
PT is written with an Australianist-based orthography. Dentals are written Ch, palatals Cy and uvulars rC.

⟨dh d dy g gw rg rgw⟩
⟨nth nt nty ngk ngkw rnq rnqw⟩
⟨ndh nd ndy ngg nggw rng rngw⟩
⟨dlh dl dly gl⟩
⟨dnh dn dny gng⟩
⟨lh l ly⟩
⟨nh n ny ng ngw rn rnw⟩
⟨rh r y gh w rr rw⟩

The vowels are obviously written ⟨i a⟩.

Edit: Switched around everything to make it only even-syllabled

I've got a couple of ideas for descendants, for example:
  • A split based on how the stop series evolve (kinda like IE)
  • A split based on where the vowel inventory will be retained as /i (e) ɑ/, or expanded to /i ʉ (e) ɔ a/
  • Innovation of a prestopped approximant series /d͡ɾ ɖ͡ɻ ɟ͡j ɡ͡ɰ (ɡʷ͡w ɢ͡ʁ ɢʷ͡ʁʷ)/
  • A shift of /d̻ ɖ ɟ ɡ ɢ/ → /d̪ ɖ d̠ ɟ ɡ/
  • Syncope
  • Non-contrastive pitch-accent based on how prosody evolved alongside syncope
Coming next, some grammar!
Last edited by VaptuantaDoi on 04 Oct 2021 05:11, edited 1 time in total.
Posts: 192
Joined: 16 Dec 2016 03:31

Re: Project Tumbleweed

Post by DV82LECM »

Yay, I liked this one.
User avatar
Posts: 490
Joined: 18 Nov 2019 07:35

Re: Project Tumbleweed

Post by VaptuantaDoi »

PT is going to be polysynthetic. And I mean really polysynthetic. I haven't finished all the verbal morphology, but here's a taste of what's to come:

“Have they recently finished breaking the big trees by the lake?”

The split-S alignment is doing my head in so I'm gonna leave that for tomorrow.
User avatar
Posts: 490
Joined: 18 Nov 2019 07:35

Re: Project Tumbleweed

Post by VaptuantaDoi »


PT – and all of its descendants – was highly polysynthetic; it had an extensive TAM marking system, polypersonal agreement, noun incorporation and an open class of proclitics. Almost all the marking was on verbs; while a verb could theoretically contain up to about fourteen morphemes (it's hard to reconstruct based on the inconsistencies surrounding clitics in descendants), free nouns were restricted to only a few suffixes. I'll get to noun morphology later because it's not as interesting. Sources of inspiration are of course this thread from the old ZBB by Whimemsz, and several natlangs including Berik and Bininj Gun-Wok.

Verbal morphology

To start with, I'll break down the verb I posted a few days ago. Actually I got that one wrong; the original would read “have the things recently finished breaking those big tree-people by the lake?”. Here's the correct version:
“Have they recently finished breaking the big trees by the lake?”
This can be divided into eleven morphemes; two lexical roots, two clitics and seven suffixes.


The order of these affixes is determined by a slot-based template:
–1.5 adverbial clitics
–1 adjectival clitics
0 root
+1 valency
+2 S/O incorporation
+3 tense
+4 aspect
+5 mood
+6 S/O agreement
+6.5 S/O gender
+7 S/A agreement
+7.5 S/A gender
I'm not gonna discuss these in a lot of detail, because that would make this post absurdly long. I'll leave the specifics of noun incorporation and the proclitics for another post as well.

Morphosyntactic Alignment
PT had a complex split-S alignment system. The complexity revolved around two splits; one in the NP and one in the verb. Verbally PT displayed a tense-based split and nominally a split based on animacy and pronounness. I'll be using Dixon's notation; S for intransitive subject, A for transitive subject and O for object.

Verbal alignment
The verbal alignment affected the whole of a conjugated verb, including incorporated nouns (which had no morphology of their own anyway). Alignment was shown in two ways in verbs; through agreement and through incorporation. Firstly, verbs showed polypersonal agreement; one O slot and one A slot. Depending on tense, one of these was extended to cover the S role. In the two past tenses, PT showed an ergative alignment and thus conflated S with O; in the two non-past tenses (or when tense was unmarked) S was conflated with A – a nominative-accusative alignment. For example, here's two sentences (without NI) to demonstrate:
"He used to be yellow"

"He's yellow"
Noun incorporation of the O role reduced the valency of verbs to become intransitive, the O markers were used for the now-intransitive subject in past tense.

Noun incorporation of a noun in the S role was also only allowed in the past tenses.
"The beach used to be yellow"

Dnyinggi dangkwinyidnyiyi
dnyinggi-Ø dangkwinyi-DNHi-Ø-RHi
beach-ABS be.yellow-STAT-3SG.A-INAN.A
"The beach is yellow"
Nominal alignment
When not incorporated into verbs, nouns and pronouns showed three alignment patterns; tripartite, erg/abs and nom/acc. The tripartite was just an overlap of the two other systems, based on two rules.

1: all pronouns (personal and impersonal) and all animate nouns marked the accusative (marked O, unmarked S and A)
2: all inanimate words marked the ergative (marked A, unmarked S and O)

These two rules overlapped with inanimate impersonal pronouns (both personal pronouns were always animate); these used both the nominal ergative suffix and the pronominal accusative suffix, so they had marked O and marked A, with unmarked S.

PT had two valency-changing operations; an antipassive (ANTIPASS) and a causative (CAUS). The antipassive (marked with -dhV turned transitive verbs into intransitives with an indirect object; this shifted the A slot to an S, and the O to an indirect object X. The main usage of this was to allow noun-incorporation of the subject in the past tense. The causative -wa was a generic causative.

TAM marking
PT distinguished four tenses, six aspects and five moods. The tense markers are only attested in a few descendants; there was a strong tendency for these to be instead shown through the adverbial proclitics.

These were shown through CV suffixes with an unspecified vowel which echoed the previous vowel of the word. Tense suffixes were optional.

distant past (DIST.P) -rngV
recent past (REC.P) -nthV
present (PRES) -rhV
future (FUT) -lhV

Six aspects can be reconstructed; habitual, frequentive, stative, inchoative, terminative and discontinuous.

Habitual (HAB) – referring to an event repeated on several distinct occasions
Gwandyinya gayindyiyirnidirngiwiwa
gwandyi-NHa gayindyi-yirnidi-rngV-wi-wa-Ø
woman-ERG gather-fruit.sp-DIST.P-HABIT-3PL.A-ANIM.A
"The women would gather *yirnidi (a type of fruit)"
Frequentive (FREQ) – an event repeated several times but on a single occasion; -rnqa
Gwandyinya gayindyiyirnidintyirnqawa
gwandyi-NHa gayindyi-yirnidi-NTHV-rnqa-wa-Ø
woman-ERG gather-fruit.sp-REC.P-HABIT-3PL.A-ANIM.A
"The women were gathering *yirnidi (a type of fruit) that day"
Stative (STAT) – something which happens for a period of time without change, or something which is always true; -dnhi.
"They only have sons"
Inchoative (INCHO) – beginning of a new state; -ra
"I'm about to start taking a shit."
Terminative (TERMIN) – end of a state; -dnha.
"I've just finished taking a shit."
Discontinuous (DISCONT)– a state which is no longer true at the point of reference; -di. The difference between this and the terminative is that the TERMIN focuses on the end of an event, while the DISCONT just implies that an event has ended at some point prior to the frame of reference.
Dlintidliyiwadlirngidi dlagwa gana gngawilinha
dlinti=dliyi-wadli-rngV-di-Ø-Ø dlagwa gana-Ø gngawili-NHa
north=make-hut-DIST.P-DISCONT-3SG.A-ANIM.A PREP hill-ABS hand-ERG
"He built a hut on the hill to the north by hand (but it's no longer there)"
PT distinguished 5 moods; realis (REAL), optative (OPT), conditional (COND), imperative (IMPER) and interrogative (INTER). Most descendants merged the optative with the imperative. All modal suffixes were CVCV forms with two identical vowels, apart from realis which was unmarked

OPT -dlarra
COND -ngkidhi
IMPER -nhanha
INTER -gwinti

Polypersonal agreement
The "O" markers were used to agree with an unincorporated direct object, or in the past tenses, an intransitive subject (including the subject of a verb with NI). The "A" markers agreed with all transitive subjects, or outside of the past tenses, with intransitive subjects also. They both agreed for person, number and gender (animate/inanimate).

O/S markers
1SG -nhi
2SG -di
3SG -rha
1PL -dni
2PL -dni
3PL -nqwa

ANIM -rngwi
INAN -∅, -rri

A/S markers
1SG -nhi
2SG -dna
3SG -∅
1PL -dlhi
2PL -rnwa
3PL -wa

INAN -rhi
Post Reply