VOWEL HARMONY FOR PROTO-DREM:

 

Vowel harmony is a type of long-distance assimilatory phonological process involving vowels in some languages. In languages with vowel harmony, there are constraints on what vowels may be found near each other. The vowel that causes the vowel assimilation is frequently termed the trigger while the vowels that assimilate (or harmonize) are termed targets. In Proto-Drem, the vowel triggers lies within the root of a word while the affixes added to the roots contain the targets. But note that the words target and trigger might be misleading since the way VH works in Proto-Drem is to go in reverse and start at the end and work your ways backwards towards the root. The type of VH Proto-Drem has is what is called regressive VH.

 

Harmony assimilation may spread either from the beginning of the word to the end or from the end to the beginning. Progressive harmony (a.k.a. left-to-right harmony) proceeds from beginning to end; regressive harmony (a.k.a. right-to-left harmony) proceeds from end to beginning. Languages that have both prefixes and suffixes often have both progressive and regressive harmony. Languages that primarily have prefixes (and no suffixes) usually have only regressive harmony — and vice versa for primarily suffixing languages. Since subjects in Proto-Drem have both, one would expect both progressive and regressive, but only regressive shows up. Also Objects having only suffixes, one would expect a progressive style, but like the rest of the language, the regressive style (R > L) only shows up.

 

The extent of R > L [+ATR] spreading is variable and dependent on rate and style of speech. The more formal the speech, the more limited the VH tends to be, while In casual speech, cases of [+ATR] spreading over several syllables, a minimum of 2 syllables are fairly common. The normal limits are unbounded, although its effect may tend to diminish with distance from the triggering [+ATR] vowel. In contrast to leftward [+ATR] spreading, rightward spreading affects only a single syllable to the right of the trigger, and the target vowel must be high /i, u/, as Preservative VH targets high vowels.

 

With most vowels undergoing VH, neutral vowel /a/ are better at either undergoing or susceptible to anticipatory VH than preservative VH. Some scholars look at Proto-Drem as trying to have both, but the system seems to be overwhelmingly an anticipatory VH system. Anticipatory [±ATR] harmony, causes the mid vowels /e, o, E, O/ to shift from R > L. All mid vowels uninterruptedly adjacent to the right edge of a phonological word are lax. All other mid vowels are tense. Within a word, all vowels belong to a single category.

 

Finally, languages that do have vowel harmony sometimes have words that fail to harmonize. This is known as disharmony. Many loanwords exhibit disharmony, either within a root or in suffixes. Proto-Drem is no exception, although the words are very rare, as the loanwords are very few in number. When looking at Proto-Drem vowels, especially the VCV structure, one must realize that there is a lot of intervention and constraints on the vowels and there’s a ‘left to their own devices” attitude that is usually not taken with these vowels.

 

One final note on Proto-Drem vowel harmony is how we look at VH and Tones and how the anticipatory system started even in Proto-Migration. We look at Tones and the standard LL, LH, HL, HH system and see how the suffixes that followed the FV would take the same tone as the FV. This is the core of the anticipatory system, and works in basically the same way as VH in many languages on the continent today.

 

Observations of Proto-Drem ATR vowel harmony:

 

For instance, rightward [+ATR] in Proto-Drem targets high vowels, and scholars agree that there is always a high vowel trigger. The fact that high vowels trigger rightward harmony, however, follows from pressure in the language to associate [+ATR] as far to the right in the root as possible. Thus, [+high] is an inherent trigger condition in Proto-Drem rightward [+ATR] harmony.

 

Canon VH

• A typical trigger vowel occurs in a prominent syllable (usually a root) with a specific feature that is contrasted somehow.
• A typical target vowel occurs in a non-prominent syllable (usually an affix) with an unspecific feature that is not contrasted somehow.
• Canon Proto-Drem VH is structure-preserving

 

Roots and Affixes

• Roots are good VH triggers on affixes, because affixes are subject to reduction, which is a good resource for VH.
• Roots are bad right-targets due to stem-initial resistance. In Proto-Drem, root initial syllables are very strong and resist most changes, even vowel harmony.
• Suffixes are reasonable VH triggers on roots, because the process is anticipatory. Suffix controlled VH is less frequent than root-control, presumably because roots do not as readily undergo reduction as do affixes.
• Prefixes are poor VH triggers on roots for two reasons: First, roots do not readily undergo reduction; and second, the VH process would not be anticipatory.
• Prefixes are bad left-triggers, especially on subsequent prefixes as well as on stems. Cases are known where a prefix triggers VH on a preceding prefix, it is harder to find examples where it triggers VH on a following one.
• An earlier word is a poor VH trigger on a following word, since neither reduction nor anticipatory directionality is involved.
• VH prefers anticipatory (right-to-left), like most languages on the Dremish Continent.
• VH effects are stronger between roots + suffixes, which form stems vs. prefixes + roots, which combine at the word level.

 

Lexical vs. Post-Lexical

• Lexical right-to-left VH is obligatory and unbounded. The trigger can be a root, suffix, or prefix. The target can be high, mid, or low.
• The trigger and target must both be high vowels /i, u/.
• Post-lexical right-to-left VH is optional. The target must be [+high] and the VH is bounded

 

Types of right-> left VH seen

 

P = prefix

S = suffix

R = root

 

Local VH:

 

P₁ < P₂ (which is a further away prefix towards a nearby prefix)

R < P₁ (which is the following prefix affecting the root)

R < S₁ (this is the following suffix affecting the root)

S₁ < S₂ (this is for the further away suffixes affecting the nearby suffix)

 

Non-local VH (affix transparency):

 

P₂ > R (which is the preceding prefix affecting the root)

R < S₂ (this is the following suffix affecting the root)

 

Root-root compound VH:

 

R₂ < R₁ (Which is the first root affecting the second roots vowels when they form a compound root)

 

Observations of the P, R, S system

• A P₂ prefix should be able to trigger VH on a preceding P₁ more readily than on a following prefix P₃.
• An S₂ suffix should be able to trigger VH on a preceding S₁ as an anticipatory process, but also on a following S₃ suffix which may be in a weaker position.

 

Proto-Drem vowels [±ATR]

 

*note: within the [] are the allophones of vowels. In this case, the allophones of /e, o, a/.

 

[±ATR] Vowel sets

 

+ATR /i, e, o, u, @/

 

XXX
XXX
XXX

 

-ATR /E, O, a/

 

XXX
XXX
XXX

 

 

 

The vowel harmony chart above shows the V₁ and V₂ vowel combinations possible within Proto-Drem. The two sets of vowels, deals with the allophones of /I, e, o, u/ as seen as /I, E, O, U/ and are heard as such. The schwa is part of the 1^st set, while the /a/ is part of the ‘allophonic set’. Note that the schwa is used when the /a/ shows up as the V₂ in this system. Note that this is the ‘default’, and the rules seen below will sometimes over-rule the chart and the sounds will change to a more disharmonious style, yet still within the standards of Proto-Drem VH.

 

Total Root Harmony:

 

In Proto-Drem, this is the simplest way to use VH in Proto-Drem. From the simplest CV or CVC-V or CVCV root to a long three, or even four syllable compound root, Total VH is seen fairly commonly in this language. The key thing to remember is that the vowels MUST be the same or allophones of the root vowel. Total root harmony only affects the root and compound word. The harmony does not spread in either direction and this harmony is considered fairly common to see with a lot of the roots having the same vowel or allophones in the roots. This style of harmony does not affect any of the affixes surrounding the root, the domain of this type of harmony is root-stem only. Some scholars try to deny that TRVH is really harmony at all, but one thing is clear, the vowels within the word are of the same [± ATR] and are technically harmonious to each other.

 

Note that with the process known as vowel coalescence, Proto-Drem still keeps the TRVH, in these cases over a long vowel (V:). Vowel Coalescence is a rarely used process used with compound roots and certain affixes to create long vowels. In this way, the vowels that are long are the same vowel and therefore enjoy TRVH under all circumstances. The preceding and following vowels though do not have to be part of TRVH and are usually not, and therefore, a stem that shows TRVH is exceedingly rare indeed.

 

Nouns with total VH

 

wembe                          n. eating utensil (from v. to eat + n. knife)
ladamba                        n. Seat, chair
mb’omo                         n. Servant

 

Nouns with a- prefix

 

Since the a- prefix is a null syllable, the prefix is outside of all total root vowel harmony considerations. These words are then treated just like other TRVH roots.

 

amb’omo
a+servant
One who exercises servancy (one who ‘is’ a servant)

 

anab’ana

a+silence

One who exercises silence.

 

ab’amandawә

a+to flatter

One who exercises flattery (a flatterer)

 

Nouns with bo- suffix

 

mb’omobo
servant+DOER
servant

 

TRVH due to Vowel Coalescence:

 

The rule states that the canon structure is CV.CV:.NCV and In this example, the tone-sequence <áà> is shown as <â> for ease of use

 

kàkâmbà
XXX
XXX

The Main VH rules within Proto-Drem:

 

There are 3 main styles of VH in Proto-Drem. The one thing to note is they are all bi-directional. What this means simply is that VH pretty much is triggered within the root, and spreads from there. The root spreads one way thru the prefixes, and spreads the other way thru the suffixes. Now note, that only the Subject uses bi-directional, as the verb only has prefixes and the object only has suffixes, so those two only use mono-directional VH.

 

The three main styles are the ATR/RTR harmony seen in a lot of languages on the Drem continent, the next two are a fronting VH and the last is where rounding takes place. The ATR is for all vowels depending on the triggering ATR vowel. Note that ATR harmony across word boundaries is strictly right-to-left:

 

ATR Harmony w/ Prefix:

 

XXX
XXX
XXX

 

ATR Harmony w/ Suffix:

 

Vowels that are already [+ATR]

XXX
XXX
XXX

 

Vowels that aren’t [+ATR]

XXX
XXX
XXX

 

Interesting vowel changes:

 

Changes of the vowels show in the examples two main changes.

 

/e, o/ > /E, O/

 

XXX
XXX
XXX

 

/a/ > /@/ > /e/

 

XXX
XXX
XXX

 

Changes with -/a/- as a prefix or suffix

 

Changes for /a/ when –a is a suffix to the root. The changes are seen depending on what vowel the root trigger vowel is. Note that when the /a/ > /e/, it goes thru ATR first, and then it fronts, so that fronting is a secondary process in Proto-Drem VH. When the /a/ > /o, O/ is changes thru ATR and rounding VH

 

When a high-front vowel precedes / V₁ = /i/

/a/ > /e/

 

XXX
XXX
XXX

 

When a high-front vowel precedes / V₁ = /u/

/a/ > /e/

 

XXX
XXX
XXX

 

When a mid-front vowel precedes / V₁ = /e/

/a/ > /e/

 

XXX
XXX
XXX

 

When a mid-front vowel precedes / V₁ = /E/

/a/ > /E/

 

XXX
XXX
XXX

 

When a mid-round vowel precedes / V₁ = /o/

/a/ > /o/

 

XXX
XXX
XXX

 

When a mid-round vowel precedes / V₁ = /O/

/a/ > /O/

 

XXX
XXX
XXX

 

When a low vowel precedes / V₁ = /a/

* note, no changes occur

 

XXX
XXX
XXX

 

Note: The above process of ATR, Fronting and Rounding is bi-directional, so that the V₁ vowel of a root will also affect prefixes as well the same as the V₂ vowel affects suffixes.

 

VH Spread within Proto-Drem

 

An easier way to show spread is to use tones. Dremish tones using a simple contrast and to see it with /L-H/ and /H-L/ being realized L-LH and H-HL (i.e. with contouring of the second tone). Scholars agree that tone is more prone to contouring than syllabic vowel features. Where occurring, cases of /aCi/ becoming [aiCi] or [aCai] are frequently followed by loss of trigger vowel, ultimately deriving [aiC] or [Cai]. Cf. changes of /iCa/ to [iaCa] or [iCia] and then to [iaC] and [Cia], respectively. So we could show them as possible SC’s showing a distinct palatalization that has seemingly occurred in several languages of the region.

 

CaCi > CaiCI > CaiC

 

CaCi > CaCai > CCai

 

CiCa > CiaCi > CiaC > C_jaC

 

CiCa > CiCia > CCia > CC_ja

 

VH affecting Proto-Drem Vowel Coalescence

 

Proto-Drem uses vowel coalescence in a narrow way for realistically only a few roots, but it does provide a small window into what the language does with these vowel sequences. The main issue with vowel coalescing is that a bi-syllabic root compresses into an onset with a long vowel, usually different vowels that have combined instead of creating a diphthong. Therefore, the vowels need to work out the vowel harmonies of each sequence so that the vowels harmonize. Note that as said above, TRVH is usually just on the compressed root, and so for the initial vowel and following vowel, all parts of the sequence must harmonize. Which means that with [±ATR], which the 4 vowels align together as if it were 1 stem or root? We will see below to see.

 

Front and Rounding where /a/ turns into either /i/ or /u/ will occur for either the first or last vowel, depending on the vowel sequence between those two vowels, as the middle two vowels will be considered transparent to VH between those two. The example below shows V1, V2, V3 and V4. V2 and V3 coalesce into the long vowel /u:/ which makes the /a/ to be the only vowel disharmonized in this compound. V1 and V4 then look at each other to see how VH will work, and in this case, the /a/ changes to a /o/ to become part of the other vowel set according to [±ATR]

 

Examples:

NkabOgumbo > Nkabu:mbo > Nkobu:mbo

 

Miscellaneous VH Rules in Proto-Drem:

 

VH in Proto-Drem as seen above isn’t as complicated as some would think, even though it has a Right > Left anticipatory ATR Harmony. The additional rules give Proto-Drem some style and a few things to watch out for, and as above, the V is bi-directional, which means that both prefixes and suffixes are affected by these rules.

 

Reduplicated syllables and final vowels

 

Mid-Vowel issues

 

 

 

FV Harmony when the trigger is a reduplicated VC.

 

The VH harmony changes presented below show how specific the changes are for 3 sounds /a, i, u/ and how the allophone /@/ shows up due to changes towards /a/. The major changes as seen above show how the FV (which is the V₂ of the root here) affects the suffix. The VC technically is a reduplicated segment as the root then truly becomes CVC-VC-FV as the FV keeps the root a canon ‘open syllable’. Seeing below, one can see (as explained above) where the /a/ is affected by ATR first, then fronted or rounded, in these cases to /i/ or /u/.

 

The demise of the final /a/ and the rise of the schwa /@/

 

/CVC-iC-a/ > CVC-iC-ə (FV a > [ə])

/CVC-uC-a/ > CVC-uC-ə (FV a > [ə])

/CVC-aC-a/ > CVC-əC-ə (FV a > [ə])

 

The fronting for /i/

 

/CVC-iC-i/ > CVC-iC-i

/CVC-uC-i/ > CVC-uC-i

/CVC-aC-i/ > CVC-iC-i (FV a > i / __ i) (a > ə > i)

 

The rounding for /u/ and some changes for the /a/

 

/CVC-iC-u/ > CVC-uC-u (FV i > u / __ u)

/CVC-uC-u/ > CVC-uC-u

/CVC-aC-u/ > CVC-uC-u (FV a > u / __ u) (a > ə > u)

 

Mid-vowel [ATR] issues

 

For words containing only mid vowels, the vowels must have the same ATR value; that is, [+ATR] mid vowels /e, o/ cannot occur in the same word as [−ATR] mid vowels /E, O/.

 

XXX
XXX
XXX

 

Right-to-left ATR harmony in Proto-Drem:

 

 

The examples are triggered by the suffix /-i/. As indicated, /a/ is either undergoes or is invisible to VH.

 

/i/ > [i]

 

XXX
XXX
XXX

 

/u/ > [u]

 

XXX
XXX
XXX

 

/I/ > [i]

 

XXX
XXX
XXX

 

/U/ > [u]

 

XXX
XXX
XXX

 

/E/ > [e]

 

XXX
XXX
XXX

/O/ > [o]

 

XXX
XXX
XXX

 

/a/ > /a/

 

XXX
XXX
XXX

Left-to-right ATR VH targets only high vowels in Proto-Drem:

 

/i/

XXX
XXX
XXX

 

/u/

XXX
XXX
XXX

 

The styles of ATR for compound roots:

 

Four combinations of [±ATR] and non-ATR words exist in Proto-Drem. The styles are fairly simple and remind scholars of the 4-way tone system of Proto-Migration. Below are the 4 differing ways [±ATR] compounds can form. The most common used way is when [±ATR] + [±ATR] roots combine to form a compound, since VH is generally pervasive in Proto-Drem.

 

[-ATR] - [-ATR]:

 

XXX
XXX
XXX

[+ATR] - [+ATR]:

 

XXX
XXX
XXX

[+ATR] - [-ATR]:

 

XXX
XXX
XXX

[-ATR] - [+ATR]:

 

XXX
XXX
XXX

 

General properties of Proto-Drem phrase-level ATR assimilation:

 

• Post-lexical ATR harmony applies across (some) word boundaries;
• is unidirectional (right-to-left)
• is subject to certain vowel height constraints (the last vowel of the first word must be [+high])
• is optional, but generally used.
• Is optionally unbounded, potentially targeting more than one [+high] vowel in successive syllables.
• In addition, /a/ is capable of being used in VH and blocking VH.

 

When root-control is not involved, anticipatory VH will have exclusive or greater effects than preservative, where in the prefix /i-/ spreads its ATR feature onto following high vowels.

 

XXX
XXX
XXX

 

XXX
XXX
XXX

 

Proto-Drem VH: Stems and Affixes:

 

I will now consider harmony at the level of affixes. Proto-Drem stems are generally quite short, and words often have a large number of affixes. For this reason, the use of harmony and affixation provides valuable information. The vowel-containing prefixes and suffixes of Proto-Drem are shown below. I will consider each affix in turn.

 

Verbal prefixes:

 

Pronouns (C)V- [+high], [±ATR]

 

This prefix is normally seen before nouns, but can be used as a noun, prefixed to the verb. The pronouns are seen in the pronoun charts in the noun section, yet here they will change the vowel due to [±ATR] concerns.

 

unj’ò
2P+fly
(s)he flies

XXX
XXX
XXX

Non-Past Tense bV- [mid], [±ATR], [±round]

 

This affix on the verb always shows a future act and future completion. Here, the rounding and fronting only keep the vowel to a mid-vowel range /e, E, o, O/, due to fronting/rounding the path is /a/ > /@/ > /e, o/.

 

ubonj’ò
2P+TAM+to fly
(s)he will fly

 

XXX
XXX
XXX

Progressive Aspect rV- [+high], [±ATR], [±round]

 

This affix that attaches onto the verb is always seen as the “-ing’ in English. The prefix also shows the act as being done at that moment, and so is not complete until at some future point. With the fronting/rounding, the path is again /a/ > /@/ > /i, u/. Since the vowel is [+high], they are seen as /i, I, u, U/. Since the vowel is [+high], the vowel will always be either a trigger or a target for VH.

 

uronj’ò
2P+TAM+fly
(s)he is flying

 

XXX
XXX
XXX

Perfect Aspect a-

 

Used as a prefix to the verb. This prefix does not change with [ATR] or [round]. Note that this can then cause disharmony.

 

U anj’ò > wanj’ò
2P+TAM+to fly
(s)he flew

XXX
XXX
XXX

 

Ingressive: bV- [±ATR], [±round]

 

This prefix to the verb is for an action towards the speaker and is inherently in the progressive aspect. The action is not complete, but being done at that moment. The prefix is nearly all the time a mid-vowel /e, E, o, O/, although, it can be /i, I, u, U/ as per the /a/ > /@/ >/i, u/ fronting/rounding.

 

ubonj’ò
2P+TAM+To fly
(s)he (is) coming (here) to fly

XXX
XXX
XXX

 

Egressive: ŋkV- [±ATR], [±round]

 

This prefix to the verb is for an action away from the speaker and is inherently in the progressive aspect. The action is not complete, but being done at that moment. The prefix is nearly all the time a mid-vowel /e, E, o, O/, although, it can be /i, I, u, U/ as per the /a/ > /@/ >/i, u/ fronting/rounding.

 

uŋkonj’ò
2P+TAM+To fly
(s)he went/left/going away to fly

XXX
XXX
XXX

 

 

Nominal prefixes:

 

Singular: -V [±ATR]

 

This is a strange prefix for nouns. The Proto-Drem noun class system is varied and vibrant, as linguists discuss this strange seemingly fossilized prefix. It acts as a noun class but is a prefix like during the time of Proto-West. Note that this prefix doesn’t agree with the standard noun class system, but this vowel prefix does change and agree with the [±ATR] of the noun it attaches to.

 

XXX
Sg+fish
a fish

XXX
Sg+person
a person

 

Verbal suffixes:

 

Past Tense –i [±ATR]

 

The suffix is either an /i/ or an /I/ depending on [±ATR] requirements within the verb itself.

 

XXX
XXX
XXX

XXX
to fall+IPAST
just fell

 

De-verbalizer -i /i, I/

 

This suffix, attached to verbs causes a verb to become a noun. The suffix is either an /i/ or an /I/ depending on [±ATR] requirements within the verb itself.

 

XXX
to finish+DEVRB
The end

XXX
to fall+DEVRB
West

 

Nominal suffixes:

 

Nominal –V [±ATR], [±round]

This suffix is noted because it does change and acts almost like a FV for subjects. This vowel also strictly adheres to the root vowel for [±ATR] and [±round]

 

XXX
XXX
XXX

XXX
XXX
XXX

 

“Person” -ni

This suffix attaches to a noun (more likely a subject as a complement) and gives the noun a ‘personal meaning’. The affix does not change, but forces the preceding vowel to change [±ATR] or [±round]

 

XXX
Bear+PER
a clan-member of the bear clan

XXX
mate+PER
a rich man

 

Diminutive –wa

 

This suffix on nouns is used on both subjects and objects. The diminutive is part of the degree affix system in Proto-Drem and as such, does not change due to VH concerns. So note that stems can and do disharmonize. Scholars feel that this affix originally was a -Cua loanword from a remnant language during contact sometime during the Old Coastal era and changed thru sound changes to what was used in Proto-Drem.

 

mbuwa
water+DEG
Little water

XXX
XXX
XXX

 

Personal Plural –ndU [-ATR]

 

This affix which is a simple plural for ‘personal’ items, works only as a [-ATR] vowel. In ways it works with the “person” suffix –ni, and would pluralize the same things that the –ni prefix makes ‘personal’. Since this suffix only deals with an [-ATR], it could cause disharmony issues

 

XXXnindU
Bear+PER+PLU
clan-members of the bear clan

XXX
XXX
XXX

 

Kinship Plural: -nUm [-ATR]

 

This suffix, like the personal plural is always a [-ATR] vowel, and cannot change. Thus, this vowel can also cause disharmony issues. Note that since this suffix deals with kinship issues, the root pluralized must be living relatives of the speaker only.

 

njabinUm
mother+KINP
mothers

XXX
XXX
XXX

 

The “Medial Vowel Puzzle”

 

"The medial vowel puzzle" is one that has aggravated Scholars of Dremish languages for years. The main idea is that the root should normally have TRVH within itself, yet there are some disharmonious roots, and especially disharmonious stems due to grammatical affixes. The middle vowel is the V₂ whereby V₁ and V₃ might be from the same VH set, while the V₂ is in “limbo” and needs to be tweaked to become harmonious with the rest of the stem. A common “word” form would be CV-CVCV for a prefix, or CVC-VC-V for a reduplicated root, or even CVCV-CV for a suffix. The main issue with the R > L is that VH starts at the end and works backwards, so Scholars still debate if the “middle vowel” is affix controlled, or stem-controlled. Below, I will try to show data on how Proto-Drem handles this issue.

 

Stem Control:

 

A vowel between V₁ vowel and V₃ agrees with the stem vowel (V₃) under stem control.

 

The medial vowel under stem control:

/CiC –oC –aC/ > CiCiCaC or CiCaCaC

 

Affix Control:

 

The vowel between the rightmost affix vowel (V₁) and the stem vowel (V₃) agrees with the affix vowel in affix control.

 

The medial vowel under affix control:

/CiC –oC –aC/ > CiCaCaC or CiCiCaC

 

Consonantal Constraints on Proto-Drem VH

 

The three main reasons for stopping VH are in some ways highly unusual since 2 of the 3 reasons deal with nasal vowels and labial-velars. It is known that basically, consonants have very little to do with VH, yet here in Proto-Drem and other languages, it is the same. The first case is the low vowel /a/ and with its reason for blocking VH is similar to the above cases of why the “medial vowel puzzle” also exists since the blocking also deals with 3 syllables and therefore a ‘medial’ vowel.

 

• Blocking by /a/
• Blocking by nasals
• Blocking by consonant clusters (seen primarily in Proto-Drem with /Ngb)/, /gb)/, /Nk/).

 

Blocking vowel harmony by /a/

 

In vowel harmony languages, under circumstances where a non-alternating vowel occurs between the target vowel and the trigger, the harmony span of the triggering vowel is blocked. Hence these non-alternating vowels are called opaque vowels. A very common vowel is /a/. So In the examples below, /a/ occurs word-medially and there is no agreement with the [+ATR] value of the triggering vowel /i/ on the right hand side.

 

 

Proto-Drem tri-syllables with medial /a/ and final /i/

 

XXX
XXX
XXX

 

XXX
XXX
XXX

Nasals blocking harmony in Proto-Drem

 

Vowel harmony is sometimes blocked by intervening nasal consonants. In the data below vowel harmony is blocked by an intervening nasal consonant. Note that the intervening nasal consonant must be within the root, and not a part of the stem. Also, the triggering vowel must be preceded by the nasal. Scholars show that vowel height harmony and nasal harmony in Proto-Migration has its origin in a late Proto-Migration sound change of nasalized Vowel Lowering. *[O] N[i/u] blocks the spread of the feature [+ATR]. It is evident that satisfying the constraint *[O] N[i/u] is more important than sticking to the *[-ATR][+ATR] constraint.

 

 

 

XXX
XXX
XXX

 

XXX
XXX
XXX

 

All the nasals /n/, /m/ and /N/ block harmony in the examples above: a nasal immediately preceding the potentially triggering vowel always blocks harmony (i.e. if the nasal is in the onset position of a syllable containing /i/ or /u/, vowel harmony will not take place; whereas, whereas a nasal somewhere else in the word does not

function as a blocker. Note that when the onset nasal is a blocker when in C₁ position and harmony goes thru prefixes. Also note that when the nasal is in C₂ position, the V₁ vowel cannot be used for VH when looking towards suffixes.

 

XXX
XXX
XXX

 

Consonant Clusters blocking harmony in Proto-Drem

 

Vowel harmony is sometimes blocked by intervening consonant clusters. The clusters in Proto-Drem as separate sounds found as a cluster, namely the labial-velars found in Proto-Drem. The 2 sounds /Ngb)/ and /gb)/ are also in company with a third, /Nk/. Since there are no geminates in Proto-Drem, these sounds when found V_V do block vowel harmony as the ‘area’ between the vowels seem to be too large for the vowels to assimilate. Why this is the case, research must still be done.

 

XXX
XXX
XXX

 

XXX
XXX
XXX

 

Exceptions to Vowel Harmony

 

Disharmony and Loanwords

 

Disharmony caused by the Proto-Drem VH system

 

Differences between Tone and VH:

 

Most people think that Tone and Vowel Harmony are strange, complex and mysterious. Well, they are right, as scholars are only starting to get into detailed research on how the systems work and those same scholars have also noted similarities, and some surprising differences between them.

 

• Partial tone assimilation is asymmetrically sensitive to featural content in a way that VH appears not to be
• Full tone assimilation (tone spreading) is asymmetrically preservative, whereas full (or partial) vowel assimilation may be anticipatory or preservative
• Tone triggers occur freely, whereas VH is frequently root- or stress-controlled
• Tone assimilations can be bounded or unbounded; VH is almost always unbounded
• Tone assimilations can be lexical or post-lexical; VH is almost always lexical

 

Tone height verses VH height

 

Tone                 Anticipated        Preservative

L-H                   M-H                  L-M

H-L                   M-H*                 H-M*

 

* Not-attested/sparsely attested

 

VH

aCi                   əCi                   aCe

iCa                   eCa                  iCə

 

So the height assimilation for tome shows a rise in either the L-tone or the lowering of the H-tone (which is seen in Proto-Drem downshift), while the next group with the H-L sequence doesn’t follow the ‘normal’ rules and is sparse or not attested at all. The VH sequences show the vowels rose or fall or a combination of the two. So in some ways a tone and VH height assimilation can take place, but to say there is one is quite a stretch. Also note that anticipatory tone spread is very sparse while preservative tone-spreading (as seen below) is quite well attested in most language families on the planet.

 

Tone Spread vs. Vowel Harmony

 

Tone height verses VH height

 

Tone                 Anticipated        Preservative

L-H                   H-H*                 L-L

H-L                   L-L*                  H-H

 

* Not-attested/sparsely attested

 

VH

aCi                   iCi                    aCa

iCa                   aCa                  iCi

 

The third difference between Tone and VH is that vowel harmony is hardly ever restricted depending on root or stressed vowels, since that is the core area of VH ‘work’. Tone on the other hand can seemingly be restricted on both root and stressed vowels, but the debate still continue. Vowel Harmony also isn’t restricted by affix or unstressed vowels and have freedom in those areas as well. Tone as the debate continues shows some arguments for being restricted by affix and unstressed vowels.

 

For vowel harmony, the trigger of VH is the root and the target of VH is affixes and unstressed vowels. The term as some linguists use is called “root-controlled”. Tones are not restricted like this in any way, since there is not trigger or target of a tone. Scholars are fairly unanimous in that there are no languages which force affix tones to change to the tones of the root, and in this way, tones act the opposite of VH.

 

The fourth main difference between Tones and Vowel Harmony is if it is bound or unbounded. There is a definition dealing with vowel assimilation in that if only one vowel is targeted and changed, that is vowel assimilation and not vowel harmony. SO according to the definition, VH MUST target multiple vowels and change them all, if able to. Tones thus act the opposite of vowel harmony in that one or more can be changed. SO in this way, VH is totally unbounded while Tones are both bound and unbound.

 

The last major area is the area of focus or what is known as ‘domain’. For VH, the domain is the stem and word only, while tonal systems are very well known for stems, words, and phrases. Thus the term, Phrasal-tones are seen.

 

Stem                Word                Phrase

Tone                 Y                      Y                      Y

VH                    Y                      Y                      N

 

Oddballs

 

Chameleonic vowels:

 

In Proto-Drem, there is a strange fusion of VH and grammatical tones. The vowels do form total vowel harmony, which was be described above. The odd thing about these is the vowels are more specific in their lexical issues and forming an obvious tone shift, thus showing them as being grammatical tones.  The chameleonic vowels will be talked about more in detail below in the tones chapter.

 

2^nd Sing.                       3^rd Sing.                       Root                 Gloss

XXX                               XXX                               XXX                   XXX
XXX                               XXX                               XXX                   XXX
XXX                               XXX                               XXX                   XXX

 

TONES IN PROTO-DREM:

 

Proto-Drem is a register tone system, whereby the tones are level pitches or sound, and each pitch can give a certain meaning to the vowel that pitch is assigned to. One thing to note is that in Proto-Drem and nearly all Dremish languages that have tone, the contour tones are not a single pitch, but a sequence of pitches. Proto-Drem has a fairly simple register tone system that shows a high regard for the way it spreads, works the rules, constraints and repairs itself to keep away unwanted aspects. One notable thing is how Proto-Drem started with the HTS, a long range spread whereby a tone language worries about the adjacent syllable, an accent system worries long term and along the whole phrase. Also how Proto-Drem looks at its tones for each morpheme while tone systems usually deal with a tone per syllable. So there are 2 contrasts which show Proto-Drem already leaving the realm of a tone system and falling into an Aspect system.

 

Due to HTS (High-tone spread), Proto-Drem and many of the later languages of the family were able to shift tones and displace tones towards the penult syllable and thereby grow a sort of accentual system that took it away from a purely tonal system. Scholars even today acknowledge that the Dremish languages are all more accentual in ways than tonal, yet the later languages definitely grew accentual and eventually lost their tone system totally. Thus from Proto-Drem, Dremish languages would become more like accentual languages in having long distance tone rules as the result of the attraction of high tones to the accented penult.

 

Proto-Drem started in future daughters and later languages hence that morphologically conditioned shift of H tones to the penult, rather than purely phonologically motivated penult shift as in other Dremish languages (including today’s’ Vakomo). Due to this change, languages with shorter stems (two to three syllables) kept the same system generally that Proto-Drem had and that Vakomo has. What caused the shift in local single syllable H-tone shifts to an unbounded multi-syllable shift is what really got the ball rolling? The shift shows in Proto-Drem how is can do a local single syllable shift keeping the shift ‘bound’ and also, it had the capacity to use unbounded shifts and shift as far as it could across the syllables. The shift from the antepenult to the penult was a long process that Proto-Drem generally is considered to start. Therefore, based on shorter stems, speakers could reanalyze the penult as the target edge of the tone domain for all stems, leading to a change from a tonal language with local tone shift to an accentual language with long distance tone shift.

 

 

General characteristics of a Tone System:

 

On the Dremish continent a few thousand years ago, the languages that had tone were fairly limited just due to the number of languages known then, yet, their influence upon today’s languages is immense. Their tonal systems literally carried over for the next several thousand years, and today, nearly all of them are tonal. But, what makes a tonal language verses one that say uses stress to convey information. This small section will bring to the fore, Why Proto-Drem is a tonal language compared to say a stress system or even a non-tonal language like today’s Kobembe which is spoken by a small group high in the Grey Mountains, and some scholars try to claim it is an isolate instead of attached to a language family.

 

Is Proto-Drem a Tone or Stress language?

 

Proto-Drem from seeing the information below is definitely a Tonal system, rather than one dealing with stress. Since Proto-Drem deals with Pitch, and those changes in pitch (called tones) affect others like them, and an important way these interact with grammar is thru the morpheme (rather than the phrase), makes Proto-Drem an obvious tonal language. Tonal languages are quite common on the Dremish continent, and in fact today’s languages, nearly all of them are register-tone system languages, rather than contour-tone system or non-tone system languages.

 

Tone vs. Stress Characteristics:

 

Characteristic                            Tone                                         Stress

Distribution                                Free                                         Cumulative

Lexical Domain                          Morpheme                               Word

Function                                    Distinctive                                Demarcative

Realization                                Pitch                                        complex

Effect on Phonology                   Tones affect tones                   affects many other features

Effect from phonology                 affect by Consonant types          affected by syllable weight

Interact w/ grammar                   morphemic                              phrasal structure

Rule type                                  segmental features                  no assimilation

 

Proto-Drem Tonal Type:

 

The tone/stress types for languages are spread over 4 styles, where the contrasts are [±tone] and [±stress]. Proto-Drem as seen above is a highly tone heavy system with a bit of stress system built into it due to the influence of syllable weight, and therefore the vowel harmony system. On the Dremish continent today, nearly all languages are [+tone], [-stress] which makes Proto-Drem fairly close to that style as commonly heard in today’s languages.

 

Type 1              [+tone][-stress]                          Hævu, Óngani

Type 2              [-tone][+stress]                          Wo Khambu, Ŋambu

Type 3              [-tone][-stress]                           Dhuŋu, Kobembe

Type 4              [+tone][+stress]                        Vakomo, Mbole

 

Is Proto-Drem a register-tone or a contour-tone system?

 

From the information we will see below, Proto-Drem is definitely a register-tone system, which is quite common on the Dremish continent. There are a few languages in the region that might have claimed to be a contour tone language, but with the persistence of floating tones, tone-spread, downstep, and multi-syllabic words, a true contour tone system would have a very hard time of being established on the continent.

 

A. “Contour tone systems”                                           B. “Register tone systems”

Fewer level tones than contours                                      more level tones than contours

Contour tones = units                                                     Contour tones = sequences (clusters)

Contour tones have free distribution within the utterance    Contour tones (clusters) are often limited to the last syllable

Dissimilation of contour + contour                                    Dissimilation of contour tones = rare

Metathesis of features within a contour                            Metathesis of contour tones = rare

No downstep                                                                 Downstep

Floating tones = rare                                                      Floating tones = frequent

Tone spreading = rare                                                     Tone spreading = frequent

Function of tone = lexical                                                Function = lexical and/or grammatical

Words are monosyllabic                                                 Words come in various sizes

Tones are restricted by syllable type                                Tones may occur on any syllable type

 

Phonetic Opposition       Phonological Opposition Example Languages

[H] vs. [M] vs. [L]           /H/ vs. /Ø/ vs. /L/                        Hævu, Vakomo

 

Proto-Drem Tone Inventory:

 

 

Tone Contrasts and Tone Inventory:

 

In terms of tone, words fall into two broad categories: those for which tone is an autonomous part of the word’s phonological make-up and those for which tone is an critical part of the morphology. The latter category comprises all verbs and certain other word classes. The former comprises most words in categories other than verbs. Tone marking throughout this grammar is as follows:

 

Acute accent (´) = H(igh):

 

ré                                 n. left; loc. (to the) left
wé                                v. To eat
má                                AUXVERB – future time marker

Grave accent mark (`) = L(ow):

 

bà                                 MODAL/AUXVERB – may, might
nà                                 MODAL/AUXVERB – tend to, usually
g’ù                                v. to work

Allotone Inventory in Proto-Drem:

 

All tones have ‘variants’ that have to be accounted for. These variants are called allotones and act like allophones to the regular sounds of the language. Allotones in Proto-Drem are always seen within certain narrow circumstances, usually dealing with the end of phrases or the penultimate syllable. Proto-Drem is no different from other register tone systems as it’s allotones are varied and even have a few “contour tones” that are very rarely seen in the language. Below are the known allotones for Proto-Drem and the circumstances they are seen.

 

High-tone allotones

 

Extra-High /_T/: Found after another High-tone not in Penultimate syllable

 

XXX
XXX
XXX

 

High-tone /_H/: Found in Bodies of words

 

XXX
XXX
XXX

 

Mid-tone /_M/: Found at end of Phrase

 

XXX
XXX
XXX

 

High-Falling /_H_F/: Found in Penultimate syllable of phrase before a low-tone.

 

XXX
XXX
XXX

High-Mid /_H_M/: Found in Penultimate syllable in phrase before a high-tone.

 

XXX
XXX
XXX

 

Low-tone allotones

 

Extra-Low /_B/: Found following another low-tone

 

XXX
XXX
XXX

Low-tone /_L/: Found in bodies of words, as noun prefixes and finally after high-tone

 

XXX
XXX
XXX

Low-Falling /_L_F/: Found in Penultimate syllable

 

XXX
XXX
XXX

 

Minimal Pairs & Homophones in Proto-Drem Tone:

 

 

Minimal pairs are often ways to see just how different tones can be for words “spelled” exactly the same. So a word might be phonologically the same, due to the tones, their meanings are entirely different. Homophones are words that share the exact same tonal patterns, yet have also entirely differing meanings. The main issue to show is that tones are a critical part of a word and the meaning of that word. Note that not all words have tone, as some words are inherently toneless, yet due to affixes, will acquire tone and thus become tonal.

 

Minimal Pairs:

 

XXX
XXX
XXX

XXX
XXX
XXX

 

Homophones:

 

XXX
XXX
XXX

XXX
XXX
XXX

 

General Observations of Tone in Proto-Drem:

 

Proto-Drem has a highly unusual tonal system. The main fact is that the tone system works a lot like the vowel harmony system. Why is this strange? The main idea is that the VH and Tone system are both anticipatory, and therefore regressive. A VH system that is anticipatory is rare enough, and some scholars still debate if an anticipatory tone system is still theoretical and doesn’t actually exist. But speeches and papers, even some written as this grammar is being written all tell of the existence, albeit extreme rarity, of an anticipatory tonal system. Nearly all tonal systems are called “progressive” in that they spread and shift rightward, but Proto-Drem, like the VH system all act backwards and go against the trend. Also in downstep and assimilation, Proto-Drem takes the seemingly “un-natural” path and goes the reverse of what nearly all other languages do. There are currently languages like Vakomo which have major tendencies to be considered an Anticipatory tone system, yet, Proto-Drem seems to be the poster child of this tendency.

 

Main Observations

• bounded HTA + delinking;
• unbounded HTA (±delinking);
• attraction to strong position (accent);
• attraction to a weak (e.g. unstressed) position (SOW): prosodically weak vowels can anticipate features from the final vowel.
• Pressure from the right edge (usually due to erosion of final vowel etc.)
• regressive downshift H-L > L-L
• Regressive Vertical Assimilation: (Anticipatory)
• Regressive Horizontal Assimilation: (Anticipatory)

 

Other observations:

 

• /H/ spreads to preceding vowel and delinks:
• /H/ is anticipated a second time within “macrostem”:
• /H/ is anticipated a second time across words:

 

Tone anticipation

 

Generalities of HTA:

 

HTA is a mysterious and often un-understood aspect of Proto-Drem tone. Since the tonal system is anticipatory, HTA will need to be quickly mentioned to show the reader some very basic aspects of some of the characteristics. The generalities look at how HTA can bend a few rules normally associated with tones. The main thing is how the tone domains of “stem” can be expanded via affixes to create a “macrostem” and therefore, HTA can expand its domain and extend out technically over word boundaries all the way to affecting the whole phrase. The key aspects below will show that anticipatory systems work backwards, just like the VH system described above. One is always looking at the preceding vowels and syllables.

 

• /H/ spreads to preceding vowel and delinks only over a single domain. In this case a word.
• /H/ is anticipated a second time within “macrostem”: the process is over 2 domains, the stem and the “macrostem” being the added affixes.
• /H/ is anticipated a second time across words: This one works with a ‘double-anticipation’ as it shifts and spreads the H-tone regressively in the same process.

 

HTA is usually a process of how it develops over time. In this case, from most likely Proto-Migration when the tonal systems started to change to where they were by the time Proto-Drem was spoken. The process is noted in three stages, each one slightly different as HTA pushes the boundaries and extends the domains it can work. HTA can not apply unless it can cross a word boundary. Boundary tone effects start with larger domains, phrasal usually and work themselves down to smaller domains, such as the stem and even the root in some cases.

 

• The first step is that HTA originates from intonational H-tones.
• The second step is where anticipated H-tones can occur across words.
• The third step is where unbounded HTA occurs within words as well.

 

Note Proto-Drem does not shift a high-tone to where H-tone is anticipated onto the penult of a preceding word. This would be expected only if the preceding word were metrically strong and the host word weak, but this is a trait that doesn’t show in Dremish languages, even today.

 

How did HTA begin?

 

One question then arises that some scholars still debate over. Just how did HTA begin? For a quick study of how the tones changed over the thousands of years from Proto-Migration until Proto-Drem, we need to look at tones and the direct links to the processes and how those processes changed the tones to what they are in Proto-Drem. Below are two sequence series, as each one will end up in the same spot with an H-L inversion of what Proto-Migration used to have with its L-H tone sequence. In between are the changes and the tonal reason behind each change.

 

*L-H > LH-H > H-H > H-HL > H-L

• Attraction of the H-tone to the penult, creating a LH-H sequence, where the penultimate vowel is also lengthened.
• contour simplification (H-H),
• final contouring with a low-tone (H-HL),
• Anticipatory absorption (H-L).

 

*L-H > L-HL > LH-HL > H-HL > H-L

• Final contouring with a low-tone, creating an L-HL sequence.
• creating an anticipatory assist (LH-HL),
• contour simplification (H-HL)
• Anticipatory absorption (H-L).

 

All of the stages in are attested in Dremish languages, which show the developments of the tone changes from Proto-Migration to Proto-Drem. The Process is also known as Dremish Tone Inversion.

 

Voiced stops depress tones H > L:

 

Depressor Consonants are in nearly all languages that use tone on the Dremish continent. These consonants provide a dampener due to their slight lowering of pitch which is a process called Depressor Induced H-Tone Shift (DIHTS) for short. Note that DIHTS can block HTS. This effect is also known as consonant-tone (CT) interaction. The effect is two fold in Proto-Drem. The first effect is that for high-tones, the consonant changes it into a LH contour-tone and if the tone is a low-tone or a null-tone, the final realized tone is a L. The real issue is that the high-tone gets delinked from the original TBU and gets shifted rightward to the penult syllable. One thing to note is the consonants which depress the following vowel. This group is voiced stops, and 2 voiced fricatives. /b, d, g, v, z/.

 

H-Tone

 

XXX
XXX
XXX

 

Non H-Tone

 

XXX
XXX
XXX

 

DIHTS Blocking HTS:

 

As spoken before, that this process can block HTS and the spread of High tones. The blocking of HTS acts just like grammatical Low-tones. With the implication of spread, the high-tones are made into a LH sequence, and other tones become a straight low-tone. Note that DIHTS can be affected by WFC and repairs to the contour tones created. In this way, grammatical low-tones are stronger than DIHTS. See below for the effects of high DIHTS works

 

DIHTS with no need for WFC

 

ga_L.ma_H.Ngb)e_H.b@_L [H-H-L]

DIHTS with a need for WFC

 

ga_L.va_H.Ngb)e_H.b@_H > ga_L.va_LH.Ngb)e_H.b@_L = ga_L.va_L.Ngb)e_HH.b@_L

note: DIHTS still works with the standard melody sequences available and after DIHTS took effect, but what speakers would used is for the LH sequence to have been split as the high-tone would be delinked and become a floating H-Tone and leave the final syllable a low-tone. That is why the high-tone shifts to the right and in this case off the word and become a floating high-tone.

 

The blocking of DIHTS:

 

What would block DIHTS, well the answer is fairly simple. If the antepenult syllable has a depressor consonant, then the original depressor cannot send shift the high-tone over usually and the high-tone gets ‘trapped’ in between the depressor consonants. In this case, the tones coalesce and form a LH sequence

 

XXX
XXX
XXX

 

Implosives raise tone L > H:

 

In Proto-Drem, due to the implosives being at the other end of the scale, they are the ones to raise the pitch of the following vowels. They are called the Anti-Depressor Consonants in Proto-Drem due to their opposite effect and the inherent rise in tone. It might be strange to see implosives being voiced being so unlike their regular voiced stops. But, the main thing is the constricted glottis that provides the force for the rise in pitch. And therefore, implosives raise the tones of the following vowel. Looking at the examples we see that the implosives are also divided between tone types. In this case, the low-tone and non-low tones are the two areas. The low-tone is changed into a HL tone sequence, and the non-low tones are changed into a straight high-tone. Unlike the examples above, there is no IILTS, but note that the ADC (anti-depressor consonants) is generally weaker at raising the tones, than the Depressors are at lowering it. The HL sequence is fairly weak and is easily changed via WFC by repair strategies.

 

L-Tone

 

XXX
XXX
XXX

 

Non H-Tone

 

XXX
XXX
XXX

 

Post-lexical rules and constraints in Proto-Drem:

 

These are the real guts of the rule-book shall we say. The Post-lexical rules deal with what we see and for this grammar in order of how it’s done. The order of how its done is key, since a person needs to know just how to deal with tone-sequences in order to do them correctly, and know why they are doing them. Below are the rules and possible “obstructions”, known as constraints they will face when dealing with tones.

 

Tonal Rule ordering in Proto-Drem:

 

Most tonal rules in Proto-Drem reveal to be intrinsically ordered. Post-lexical rules apply in the following order:

 

• Spreading
• H-delinking following underlying Low-tone (Anticipatory)
• WFCmax > WFCmin (Default Low-tone insertion or Non-initial Spreading)
• Floating-Tone Deletion
• HReplacement

 

Tone spreading

 

At the phrase and sentence level, Leftward since Proto-Drem is an anticipatory-tone system where tone spreading also occurs, but in slightly different and more complex conditions. The main ways of tone-spread are seen here, especially with bound and unbound spread, but also the second way of tones-changing is due to a process known as assimilation which will be discussed below as well. Below are some major observations of the Proto-Drem tone-spread system, and scholars still debate the first 2 views, but all agree that Proto-Drem is heavily on the HTS side.

 

• HTS will apply to /H-L-L/ base-tones (also known as ‘input’)
• HTS will apply to /H-L/ if a floating H follows.
• HTS will apply to /H-L/ if the L syllable is stressed
• +target delinking for complete tone spreading
• +trigger delinking for tone spreading vs. shifting
• non-iterative
• Domains are the root, stem and phrase as most tonal systems are.
• Directionality is still spoken of as a right-to-left due to the anticipatory nature of the tone system.
• Proto-Drem is both bounded and unbounded
• inhibitors are the famous depressor consonants (voiced stops)

 

Anticipatory Tone Spread:

 

The main thing with tone spread is that the spread like the assimilation above heads towards the right edge yet looks towards the preceding tones as to where to spread. In these cases, there are 3 main areas one must look at spread and how it works in Proto-Drem. Due to HTA, a null-tone preceding a high-tone will change as seen here. /Ø-H/ becomes [H-H]

 

a. in partial spreading, the trigger combines with the target to create a contour tone. Now let’s see below how in the 1^st example below, the low-tone is anticipated. Due to this and other examples, some scholars state that Proto-Drem must have had LTA (Low-tone anticipation) in addition to HTA, but the debate is still raging.

(H-L → HL-L, L-H → LH-H)

 

H-L → HL-L

XXX
XXX
XXX

L-H → LH-H
XXX
XXX
XXX

b. in complete spreading, the trigger causes de-linking of the target tone. The de-linking causes a floating tone to form and floating tones are usually spread rightwards towards the next syllable to create a contour tone. If the floating tone heads toward a null-tone, then that morpheme then acquires a tone. Note that the first morpheme becomes a contour tone while the floating tone is sent rightwards towards the next TBU (Tone bearing unit).

 

(H-L → HL>L, L-H → LH>H)

 

XXX
XXX
XXX

XXX
XXX
XXX

c. spreading may also be to a preceding toneless TBU

(Ø-H → H-H, Ø-L → L-L)

 

Ø-H → H-H

XXX
XXX
XXX

Ø-L → L-L
XXX
XXX
XXX

 

High Tone Spread (HTS) in Proto-Drem:

High-Tone Spread is very common in languages of the region and in fact, even today HTS is very common. LTS and even rarer is MTS.

 

Bound vs. Unbound

The main variety of HTS is seen simply as bound vs. unbound. The main issue in binding is the distance that the tone can spread. With bound, the tones can only spread at most to the adjacent morpheme. With unbounded tone spread, the tones can spread all the way to the antepenultimate syllable, making for some long-distance spread sometimes. Note that since the tone system is anticipatory, the tones look to the preceding tone, yet spread to the right edge.

 

bounded, e.g. Hævu

 

XXX

XXX

XXX

 

unbounded, e.g. Vakomo -> antepenult

 

XXX

XXX

XXX

 

H tone shift/displacement

One variation on HTS is referred to as tone shift or displacement. Here with tone shifting, the High-tone doesn’t actually spread, it just shifts its position further along the targeted morphemes, and with unbounded, all the way to the antepenult syllable. Note that since the tone actually shifts off of the original TBU, there is de-linking occurring, and the tone docks onto the new TBU.

 

bounded, e.g.

 

XXX

XXX
XXX

unbounded HTS + delinking, e.g. Vakomo → antepenult 

 

XXX
XXX
XXX

 

Vertical and Horizontal Assimilation:

 

In Proto-Drem, these technically aren’t considered Spread, but they are postlexical affects that affect the tones around them. The two main ones dealt with in Proto-Drem are vertical assimilation, which is also known by a very important term known as downstep. The other important type of assimilation is nearly like tone spread, but is distinctly different in its ways. The horizontal assimilation is also a well known and studied tone-change known as HTA, which most languages in the region use compared to the rarer LTA or Low-tone anticipation.

 

Regressive vertical assimilation:

H-L → M-L (Downstep)

 

The corresponding change shows a very important and fairly common tonal change in Proto-Drem. We see the following low-tone change the high-tone and step it down to an allophonic mid-tone /_M/, as it would be seen as a downward arrow to show the high-tone being stepped down.

 

XXX
XXX
XXX

 

Regressive Horizontal assimilation

 

L-H → LH-H (High-tone anticipation HTA)

 

Horizontal assimilation is generally a fancy term for tone spread. In Anticipatory tone spread, the tones still assimilate towards the right, but act in reverse of what most scholars would expect. Instead of the L-H becoming an L-LH like in most languages of the region, it goes the reverse as seen above.

 

L-H → LH-H

XXX
XXX
XXX

 

Downshift in Proto-Drem

 

Note that some people, even scholars get the terms down-step and downdrift confused, some even using the term for an unintended meaning. This small note should make clear that there is a distinct difference between the two and there should be no more confusion on what downdrift is compared to the above examples bringing up the importance of downstep.

 

Downdrift, where the absolute pitch (not tones) of the speaker's voice is gradually decreased as the sentence continues (often resulting in initial low tones being pronounced at a higher pitch than final high tones), is a feature during natural speech. Basically, a high tone immediately following a low tone is pronounced at a slightly lower frequency than a previous high tone.

 

Additionally, a slightly more dramatic lowering of pitch (a downstep) may occur between certain syllables. In Proto-Drem, the downstep (indicated with a ↓) naturally occurs with high-tones preceding a low-tone where the low-tone causes the high-tone to fall in pitch to become a mid-tone.

 

 

Delinking rules

 

Tone-delinking rules apply in two cases:

 

• As a repair strategy applying when a “well-formedness condition” (also known as WFC) is violated. (see below)
• When multiply linked H are preceded by L. delinking applies to all association lines of a multi-linked H-tone except for the final one (which will be maintained) whenever this particular High-tone is preceded by a (floating or associated) underlying Low-tone, a tonally non-specified morpheme or an utterance-initial boundary.

 

For the delinking rules, the second one can cause some trouble for people, as in some ways it seems like a drastic change to the high-tone. But the diagram below will show how this process is done.

 

kàŋkámá --> kàŋkámá

-|-----\----| -à -|-----|----|

-L-------H- à -L----fL--H-

L-H-H -----à L-L-H

 

As one can see, the High-tone is linked to multiple TBU’s and therefore capable of being shoved off. The delinking process shoved the multi-linked high-tone off all the links except its original TBU. The then empty TBU’s then get occupied by floating low-tones and the delinking process is done. Note that delinking is only used when lots of HTS is used, but normally only is used for 1 or 2 TBU’s, as an unbounded HTS usually doesn’t get delinked.

 

 

The asymmetry of H and L

 

In scholarly language “it is obvious that underlying H and L (along with the absence of tonal specification) behave asymmetrically. A High-tone can delink a Low-tone initially linked to the segmental tier through a single association line, but things will not occur vice versa.” What this means in laymen’s terms is that high and low tones don’t act the same, as they are different from what we’ve seen before, and what you will see. A high-tone can delink a low-tone from the TBU that the low-tone is attached to, but a low-tone cannot readily delink a high-tone from its TBU (except in terms of a repair strategy or other “help”). This means for speakers, a high-tone has more prominence on its own, but thru repair strategies, a high tone can be easily moved around.

 

The Well-formedness Condition and repair strategies

 

Another important aspect is the language specific Well-formedness Condition (WFC) according to which in Proto-Drem not more and not less than one tone should be associated to each TBU. This WFC, which applies to items at both the lexical and the postlexical level, implies an efficient way of avoiding contour tones and downstep in Proto-Drem. The implication is nice in general use, but there are times when even a repair cannot take effect, as allotones and postlexical issues will arise where a contour tone and downstep will occur.

 

To show how the repair strategies when all used can even repair a seeming sequence destined for trouble. The below shows a double contour tone and end up with a simple sequence. Note how the strategies can be used slightly differently, yet both come to the same final tone-sequence. Of the two strategies used, the bottom of the two sequences is most likely to be used.

 

LH-LH → H-LH               H-LH → L-HL     L-HL → HL-L     HL-L → H-L

LH-LH → HL-HL             HL-HL → L-L     L-L → LH-L        LH-L → H-L

 

Contour change             levelling             dissimilation      absorption

 

Repairing contour tones thru Tonal Absorption:

 

In Proto-Drem, regulating and controlling contour tones is a big issue, just like “plateaus” are a repair strategy to control downstep and ‘troughs’. Here with contour tones, even though they do exist as allophones, they are kept under strict control, and so Tonal Absorption is used to do that job. Like most operations in Proto-Drem, the job is anticipatory and thus works backwards as it works right-to-left unlike most other tonal languages out there.

 

Within Dremish languages, contour tones are frequently restricted to long (bimoraic) vowels, to long or short vowels in penultimate position, or to final vowels which are typically monomoraic. In addition, contours may arise from cases of vowel coalescence, which occur both word-internally and across words. In many cases the surface realization of a contour will depend on the surrounding tones. Anticipatory absorption is seemingly less natural due to its rarity, as is anticipatory spreading in general. But even though it’s very rare, it is attested even today on the continent.

 

Typical tonal absorption:

L-LH → L-H

H-HL → H-L

 

If a contour is not followed but is preceded by a like tone, regressive absorption will apply. The repair strategy here is to allow a L-H-L sequence, which below in the section on “plateau”, a H-L-H is not allowed. Why this L-H-L sequence is allowed and the H-L-H must be changed is still debated by scholars.

 

L-LH-L → L-H-L

 

Summary of contour simplification in Proto-Drem:

 

In Proto-Drem, the tone sequences usually seen with contour tones are quite needed, as the sequences must be broken down into 3 differing groups, depending on where the sequence is in the word. The three areas are word initial, word internal and word final, as each have their own distinct affects. The word initial contour tones usually simplified the contour tones and created a single tone where the contour tone was. Of the tone changes below, any with parenthesis ( ), are considered to be floating tones, and are usually sent to the right to dock onto the next words first available TBU.

 

Note how most of the floating tones are high-tones, while a few are low-tones. This is consistent with Proto-Drem HTS and floating H-tones. Most of the time, when tones change, one sees how some tones are just dropped and not even formed as floating tones. One can see this tone-dropping in the word-internal processes especially as both low and high tones are dropped, especially from the contour tone in between single tones. Note how in Dremish languages, an H-L-H are made into an H-H-H via a plateau effect, while a L-H-L sequence is allowed.

 

Word initial tone sequences

# LH-L → LH-L

# HL-L → H-L (L)

# LH-H → L-H (H)

# HL-H → HL-H

 

Word Internal tone sequences

L-LH-L → L-H-L

L-HL-L → L-H-L

L-LH-H → L-L-H (L)

HL-H → L-H

H-LH-L → H-H-L

H-HL-L → H-H-L

H-LH-H → H-L-H (H) → H-H-H (Plateau effect)

HL-H → L-H

 

Word final tone sequences

H LH # → H-L (H)

HL # → H (L)

L LH # → L-L (H)

HL # → L (H)

 

Proto-Drem does not allow contours in like neighbor contexts. So in that contour tones must not follow like tone (*L-LH, *H-HL). Also note that HL appears only when preceded by L [HL-L].

 

H-Ø-H / H-L-H Sequences are not allowed

 

The sequence is usually called a trough, as it shows a dip in between the high-tones. There are two ways a Proto-Drem verb deals with this issue. The two ways all deal with if one is within a verb stem or outside of it. Note that if a stem requires and is grammatically stated to have a low-tone, then the preceding H-Tone will be lowered in a low-tone spreading fashion. This change is one of the few exceptions where scholars do state that Proto-Drem does use Low-tone spread. And this way does show that a grammatical L-Tone can block the Plateau affect.

 

High-Tone Plateaus in Proto-Drem stems

 

Another Proto-Drem repair strategy actually seems to block OCP; yet, the infamous “Plateau” is a common occurrence in the regions languages. Class prefixes usually are the trigger of tone changes as the tone spreads back into the stem. If both the target and trigger are H-tones, it can usually create what is called a “plateau” which is a HHH and is not broken. Why is this plateau created in the first place? A trough made of H-Ø-H or an H-L-H is seen, and the middle toneless or Low-tone are raised to the status of a high-tone to create the plateau. Note that the toneless roots having a high-toned prefix and suffix could actually make a trough to create the dramatic tone-rising. The rules comes out as H-Ø-H > HHH and H-L-H > HHH. These examples are for mono-syllabic verb roots only, as subjects are affected by differing rules. The examples below will use single syllable verb roots to show the interaction with object roots clearly.

 

Toneless Plateau (using a toneless verb root)

 

káŋkamá --> káŋkámá

-|------|----| -à -|-----\----|

-H----Ø---H à H------H-

 

Low-tone Plateau (using a low-tone verb root)

 

káŋkàmá --> káŋkámá

-|------|----| -à -|-----\----|

-H----L---H à H------H-

 

Non Plateau Repair strategies for H-L-H troughs

 

The stems and affixes outside of a verb stem are treated much the same way with a H-L-H trough or even a H-Ø-H trough in that regressive HTS voids the middle tone and replaces it with a high-tone. The process is the same in that the plateau just changes the middle tone and spreads no farther and thus creating the famous H-H-H block we see. The next example is where a Low-tone is inserted due to grammatical concerns and is usually a TAM or other immutable tone. Thus the process of HTS or Plateau cannot be worked and thus blocks HTS. The low-tone regressively links and spreads with the preceding TBU and lowers the H₁ and creates a small low-tone block.

 

Plateau outside of a verb stem

H-L-H > H-H-H

 

káŋkàmá --> káŋkámá

-|------|----| -à -|-----\----|

-H----L---H à H------H-

 

Rule due to grammatically forced L-Tone [Anticipatory LTS]

H-L-H > L-L-H

 

káŋkàmá --> kàŋkàmá

-|------|----| -à --\----|----|

-H----L---H à -----L----H-

 

Floating tones

 

In order to account for specific tonal phenomena, we also need to look at floating tones, and how lexically marked floating tones can affect tones. We will use two examples, the first is a base high-tone, and the other is a base low-tone. With the base tones, we will add a numeral ‘one’ to it which is a high-tone and see how this will work. Lastly, we will take a demonstrative pronoun ‘this’ and add it to the base root to see what happens.

 

Rat

H-H

 

One rat

H-H-H -> H-M-H

 

This rat

H-H-L

 

Hand

L-H

 

One hand

L-H-H

 

This hand

H-L-H -> L-L-H

 

We see from the above examples that in the phrase ‘one rat’, the natural plateau normally doesn’t exist, and downstep takes a toll on the preceding high-tone, in this case the middle one. Essentially the H-H-H becomes an H-M-H and so keeps away from a repair process. In the second example with ‘this hand’ we see a root low-tone and instead of a plateau forming, we have the low-block formation. Why the low-tone block forms here is that the root contains a low-tone which normally isn’t changed, as a root tone is usually the start from where tone changes occur.

 

Note that in Proto-Drem, a null-tone Ø cannot be made into a floating tone, as each TBU must have a specified tone docked onto it. Thus a TBU will have a H. L or contour tone docked only.

 

H-replacement in sentence final position

 

H-replacement occurs in final position for declarative sentences. Every H-tone in this position will be systematically replaced by a Low-tone. Note this L-tone must be a floating boundary tone that spreads to the left and consecutively delinks the H associated to the final syllable. This delinking causes the high-tone to be null and void and thus not become a floating H.

 

H-H-H # -> H-H-L

 

L-Tone Displacement rule

 

Tone Displacement rule: Its application seems to be restricted to the associative construction, and is triggered by the devocalization or the deletion of the vowel segment of the agreement marker when the latter is prefixed to the associative morpheme _   _ . (Devocalization in case of a [+low] vowel and deletion in case of a [–low] vowel.) The associative morpheme must be analyzed as being underlyingly low whereas the agreement marker is undoubtedly underlyingly high, except for noun classes 1 and 9. This Tone Displacement rule, which applies before the initial spreading rule, is exemplified in (25) and (26). In both examples, the noun class 14 agreement marker (_ _  _ _ ) is being used.

 

Nouns in Proto-Drem Tonology

 

Proto-Drem verbs are fairly well studied and tested, thru a lot of theoretical work, especially in the last 30 years to get us to the point that the language carries very few mysteries anymore. Here is the noun tonology, and what to expect with this language. Note that the areas looked at are well known and very common in the region, even with today’s languages, especially Vakomo. Noun tonology is fairly similar yet distinct from verbal tonology which for most languages is very complex. Proto-Drem noun tones are usually simple and rarely have to be repaired due to a trough or grammatical tone, yet it does happen.

 

Common noun tonal sequences:

 

The tone sequences here in Proto-Drem, we will just show the high-tones, as the null-morphemes are for the actually nulls or low-tones that do show up on roots, stems and compounds. Note that a compound might actually have the tone sequence changed, as the 1^st root might have tones, yet as a compound, might have those tones nullified and lowered depending on the tone sequence of the compound. In this respect, the tones are separated by group, as each group deals with a specific group of objects. Note also that animals are usually treated as animate objects, but due to their special grouping, they have their own tonal group.

 

*note: The sequences below are usually in animate/inanimate focus. This has led some scholars to deduce that the Animacy issue is an old fossilized aspect of Proto-Drem that thousands of years before, was much more prevalent in the language and the grammar.

 

Small Roots

CV                                H (Tone group 1 – usually seen as an animate object)

CV                                Ø (Tone group 2 – usually seen as an inanimate object)

 

Roots

CVCV                           H-Ø (Tone group 1 – usually seen as an animate object)

CVCV                           Ø-H (Tone group 2 – usually seen as an inanimate object)

 

Small Compounds 1a

 

CV-CVCV                      Ø-H-Ø (Tone group 1 – usually seen as an animate object)

CV-CVCV                      Ø-Ø-H (Tone group 2 – usually seen as an inanimate object)

 

Small Compounds 1b

 

CVCV-CV                      Ø-H-Ø (Tone group 1 – usually seen as an animate object)

CVCV-CV                      Ø-Ø-H (Tone group 2 – usually seen as an inanimate object)

 

Compound roots

CVCV-CVCV                 Ø-Ø-H-Ø (Tone group 1 – usually seen as an animate object)

CVCV-CVCV                 Ø-Ø-Ø-H (Tone group 2 – usually seen as an inanimate object)

CVCV-CVCV                 Ø-H-H-Ø (Tone group 3 – usually seen as an animal)

 

Note for compound roots, if the original root had tones, the compound will have tones as well, but might have them changed to low-tones due to the tone-group sequence.

 

Tonal Coalescence in Proto-Drem nouns:

 

Tones change due to vowel coalescence, and therefore combine into un-reparable tone sequences, that are unaffected by WFC. The Tone sequences as seen in the Phonology section will be seen here as well. Note that the long vowels are considered two morphemes and so the tone-sequence is still a sequence of tones instead of a unit like a falling/rising tone as some scholars assert.  Also below is an interesting process whereby the high and low tones simplify and spread instead of just stay as two separate morphemes. Why this occurs is still under debate, but scholars, including myself, think it is a fossilized WFC (repair strategy) used in the early stages of the language that was by then no longer used in Proto-Drem.

 

General Process of Tonal Coalescence:

 

HL sequence / F-tone

 

With a HL- sequence, the following tone is important, as if the tone is a low-tone, then an allowable HL-L sequence occurs, while if a high-tone follows, then a HL-H sequence occurs which normally would simplify to a L-H sequence, but due to the need to keep the HL sequence, the following high-tone is de-linked and replaced with a low-tone as seen above in the repairs section. Also one must look to the preceding tone and see if the sequence holds true as well. In the case of the HL sequence, it is not affected by either a preceding Low or high tone.

 

L-HL-L > no tonal changes

H-HL-L > no tonal changes

 

L-HL-H > L-HL-L (H-Replacement repair strategy)

H-HL-H > H-HL-L (H-Replacement repair strategy)

LH Sequence / R-Tone

 

Like the above sequences, this one works the same with a few quirks instead. The LH sequence usually presents problems for only the T₁ tone of this sequence. Remember from above that an L-H-L sequence is allowed, while an H-L-H sequence is not. But also realize that the L in either HL or LH is considered grammatical in priority and therefore cannot be changed by a repair strategy such as “plateau”. So while the repair strategy of “Plateau” is unusable, the repair strategy of “H-replacement” works just fine.

 

L-LH-L > no tonal changes

H-LH-L > L-LH-L (H-Replacement repair strategy)

 

L-LH-H > no tonal changes

H-LH-H > L-LH-H (H-Replacement repair strategy)

 

HH > H Sequences

 

A combination of a HH just simplifies and coalesces into a simple High-tone that is spread over two morphemes as the high-tone spreads regressively over both morphemes. Below shows the long syllable with the initial two high-tones, then coalescing the tones into a single high-tone, then watching the high-tone regressively spread to the preceding TBU.

 

gàà > gaà > gàà

--|-| à -=-| -> -\-|

-HH -> Ø-H -> H

 

LL > L Sequences

 

A combination of a HH just simplifies and coalesces into a simple High-tone that is spread over two morphemes as the high-tone spreads regressively over both morphemes. Below shows the long syllable with the initial two high-tones, then coalescing the tones into a single high-tone, then watching the high-tone regressively spread to the preceding TBU. Once again, some scholars point to Proto-Drem using an LTS process.

 

gáá > gaá > gáá

--|-| à -=-| -> -\-|

-LL -> Ø-L -> -L

 

Specific locations of when to use:

 

Pronouns:

 

Pronouns are very common in Dremish speech. They usually go ahead of nearly anything in the noun phrase, except for a few concords, and negation. The Pronouns are a mix of toned and toneless vowels, and so troughs can be expected. Note in the first example, the pronoun is considered a grammatical low-tone which instead of being raised, will force the negation to switch to a low-tone. Note in the second example, that the pronoun is a grammatical toneless-vowel. Therefore, it cannot be changed to a high-tone normally due to HTS and the high-tone regressively spreading onto the preceding toneless TBU.

 

NEG+PRO

 

Nà-úbàbà [H-L-H-H] > Ná-úbàbà [L-L-H-H]

NEG+2P+man

She is not a man

 

PRO+NC

 

Gumbàbà [_-H-H]

2P+NC1a+man

It is a man

 

CONC+PRO

 

Nèlagu [H-_-_]

CONC+2P

It is from …

 

Noun Classes

 

Noun classes usually deal with troughs and unusual tone patterns with bi-syllabic noun roots since the noun classes always precede the nouns.

 

NC+1 syllable NOUN

 

mbàbà [H-H]

NC1a+man

man

 

NC+2 syllable NOUN

 

múbàbà [L-H-H]

NC1b+man

men

 

Noun Concords

 

Noun concords come in several flavors, nearly all of which come before the pronoun or noun class that precedes the noun itself. The one thing that usually always precedes a noun concord is the negation particle. Note the second example is a HL sequence which falls to a low level and does not conflict with the next vowels low-tone. Note in the first example, that the second vowel in the number is toneless; since it is not a grammatical tone; it is subject to spreading by the following tone, in this case a low-tone. The toneless vowel therefore becomes a low-tone due to LTS.

 

NUM+CONC

 

Ŋèlomímúbàbà [H-_-L-L-H-H] > Ŋèlómímúbàbà [H-L-L-L-H-H]

NUM+CONC+NC+man

Three men

 

CONC+NC

 

Jêmúbàbà [HL-L-H-H]

CONC-NC+man

Men who…

 

Negation

 

Since negation is the usage of the nà- particle, and the negation always goes as a sentence initial particle, its effect on nouns, specifically the subject is to be noted. The negation can easily be used before pronouns and noun classes, so the high-tone might be followed by a toneless vowel, or a tone. Note in the second example that the toneless vowel presents a ‘trough’ appearance, but since the toneless vowel would be considered a ‘mid-tone’ in essence, there is no tone-rising. Note in the second example, that the pronoun is a grammatical toneless-vowel. Therefore, it cannot be changed to a high-tone normally due to HTS and the high-tone regressively spreading onto the preceding toneless TBU.

 

NEG+PRO

 

Nàguja [H-_-_]

NEG+2P+person

It is not a person

 

NEG+NC

 

Nàgumbàbà [H-_-H-H]

NEG+2P+NC1a+man

It is not a man

 

NEG+CONC

 

Nànèlagu [H-H-_-_]

NEG+CONC+2P

It is not from …

 

Verbs in Proto-Drem Tonology:

 

Proto-Drem verbs are fairly well studied and tested, thru a lot of theoretical work, especially in the last 30 years to get us to the point that the language carries very few mysteries anymore. Here is the verbal tonology, and what to expect with this language. Note that the areas looked at are well known and very common in the region, even with today’s languages, especially Vakomo. Verb tonology deals with complex tense, mood and aspect interactions and so the verb root usually has tones shifted and sent off to float towards the right. This section will detail those issues that often happen to verbs.

 

Common Verb sequences:

 

Unlike the nouns above, Proto-Drem verbs reduplicate and are always at least 2 syllables. So verbs are distinct compared to nouns. Reduplication issues as well as TAM issues present their own challenges which must be met and understood.

 

The tone sequences here in Proto-Drem, we will just show the high-tones, as the null-morphemes are for the actually nulls or low-tones that do show up on roots, stems and compounds. Note that a compound might actually have the tone sequence changed, as the 1^st root might have tones, yet as a compound, might have those tones nullified and lowered depending on the tone sequence of the compound. In this respect, the tones are separated by group, as each group deals with a specific group of objects. Note also that animals are usually treated as animate objects, but due to their special grouping, they have their own tonal group.

 

Roots

CVCV                           H-Ø (Tone group 1 – usually seen as action done by/to an animate object)

CVCV                           Ø-H (Tone group 2 – usually seen as action done by/to an inanimate object)

 

Redup roots

CVC-VC-V                     Ø-H-Ø (Tone group 1 – usually seen as action done by/to an animate object)

CVC-VC-V                     Ø-Ø-H (Tone group 2 – usually seen as action done by/to an inanimate object)

CVC-VC-V                     H-H-Ø (Tone group 3 – usually seen as action done by/to an animal)

 

Compound roots

CVCV-CVCV                 Ø-Ø-H-Ø (Tone group 1 – usually seen as action done by/to an animate object)

CVCV-CVCV                 Ø-Ø-Ø-H (Tone group 2 – usually seen as action done by/to an inanimate object)

CVCV-CVCV                 Ø-H-H-Ø (Tone group 3 – usually seen as action done by/to an animal)

 

Redup compounds

CVCV-CVC-VC-V           Ø-Ø-H-H-Ø (Tone group 1 – usually seen as action done by/to an animate object)  

CVCV-CVC-VC-V           Ø-Ø-Ø-H-H (Tone group 2 – usually seen as action done by/to an inanimate object)

CVCV-CVC-VC-V           Ø-H-H-H-Ø (Tone group 3 – usually seen as action done by/to an animal)

 

Note for compound roots, if the original root had tones, the compound will have tones as well, but might have them changed to low-tones due to the tone-group sequence.

 

Twin Sets of Tones for Proto-Drem

 

Like nearly all languages of the region, verbs have 2 sets of tones, those 2 sets are for the infinitive and the other is for the toneless. Since the Object root is directly attached to the verb, the object is considered a suffix to the verb. The form is basically seen as Prefix-Verb-Obj-Suffix

 

Verbs – Infinitive – H Tone

 

XXX
XXX
XXX

 

Verbs – Toneless

 

XXX
XXX
XXX

 

Verbs – Hi Tone – with Object

 

XXX
XXX
XXX

 

Verbs – Toneless – with object

 

XXX
XXX
XXX

 

Initial H-tone deletion on verbs

 

A strange occurrence happens in Proto-Drem that is unusual since the verb root is the last thing of a verbal phrase. The odd thing is that even if the initial tone is a contour, such as a HL, the H is taken off therefore leaving the L-tone. Scholars argue that on verbs, the initial tone is a low-tone at the core, no matter what tone is manifested over it, so due to the rule, the underlying low-tone shows. If the morpheme is actually a null-tone, then the null-tone shows and the morpheme become toneless.

 

H > Ø

 

XXX
XXX
XXX

 

Words and stems with all low-tones are forbidden:

 

On verbs, there must be a high-tone somewhere, even if one has to be inserted via an affix. On base toneless roots, an affix will usually put on a tone, thereby giving the stem a tonal value. If the tonal value is a low-tone, therefore, a high-tone is “epenthetically” inserted to give the word the required high-tone. Note that a bi-syllabic root (CVCV) can be an L-L sequence, but with affixes, a High tone will be attached or inserted.

 

Attached via affix

Ø > H

 

XXX
XXX
XXX

 

Attached epenthetically

Ø > H

 

XXX
XXX
XXX

 

How Tones are affected by TAMV in Proto-Drem:

 

In Proto-Drem, tenses, moods, and aspect affect how tone sequences work. Most are straight-forward, others are repaired, while a few work as exceptions to the ‘normal rules’ and are also seen. There three main areas of TAM to look at with Proto-Drem. The first are moods alone with the verb, which uncommon can be used. The verb root is usually left alone, but changes can occur, especially if a ‘trough’ occurs which is usually flattened and a high-tone ‘plateau’ is created. The second form is tenses with the verb root. These are more commonly used and one thing is noticed right away as the verb root has the initial syllable replaced by the tense marker, and so the tone on that syllable becomes a floating tone and heads right to find an available TBU to dock to. The third form is generally considered the most common form used, as it combines the tense and modal, leaving the verb root to be fairly small and the verb root might not even have a tone, especially if the final vowel was replaced by the modal. The combination of the two markers can present its own challenges as exceptions can occur as do some careful consideration of repair strategies as these markers are grammatical and usually ignore tone changes.

 

Tones and Moods:

 

To see clearly how moods will affect a verb, we will use the modals that Proto-Drem has available and see how a verb root is affected. The below examples will use the verb root ‘b’èmbá’ which means to freeze, be concentrated, to get stuck to. Note that some of the tones, especially on the addition of the FV changes the last tone and turns it into a toneless TBU. Also with the addition of tense particles, additional information and further tone changes can easily occur. Note that a good number of examples below require a repair strategy, especially to change the first tone in the root to a high-tone and create a ‘plateau’ of high-tones. Then the next tone change to happen is for the last sentence tone to change from a high-tone to a low-tone, which affects a fair number of the examples below. Those that will require an additional object or pronoun will be seen with parenthesis.

 

Note* always apply repair strategies before finishing with high-tone lowering on sentence final vowel.

 

Optative                        -wâb’èmbá( )                  … wish/desire to freeze …

Indicative                       -gèb’èmbe                     … is frozen.

Imperative                      -síb’èmbe                      … freeze!

Injunctive/Advisory          -sàb’èmbe( )                  … (you) freeze …

Ru particle                     -rùb’èmbà                      … *is* frozen (see!)

Habitual/Iterative            -fáb’émbà                      … frequently freezes

Decisional                     -gbib’èmbá( )                 … decided to freeze …

Libetic                           -líbéb’émbà                    … will freeze (to death)

Hortative                        -sèb’èmbe( )                  … let’s freeze …

Conditional - realis         -ŋáb’émbá( )                  … if frozen …

Conditional - Irrealis         -ŋáb’élimbá( )                 … frozen …

Validational                   -ŋkéb’émbà                  … froze (I swear!)

Dubitative                      -régûb’èmbá( )               … is doubtful (that) … froze                   

Progressive                   -ázâmbà                        … is freezing

Benefactive                    varied

 

Tones and Tense:

 

One of the most common things used in Proto-Drem are tense markers. Indicating time, they can easily tell the listener when an event happened. We will see how tense affects the root below. Note how the examples below are affected by repair strategies such as ‘plateau’ and ‘final high-tone lowering. Due to this tone changes, some ambiguity can arise, so care must be taken with context when proper tone use is required.

 

Affirmative tense

 

Past                            

-immediate past             dáŋúmbà                       just froze

-recent past                   dáŋémbà                       just recently froze                     

-remote past                  dáŋómbà                       is frozen for a long time

 

Anterior Past

–immediate                   dálúmbà                        before (it) just froze

–recent                         dálémbà                        before (it) recently froze

–remote                        dálómbà                        before (it) froze a very long time ago

 

Non-Past

-immediate future           dáŋùmbà                       will very soon freeze

-near future                    dáŋèmbà                       will soon freeze

-remote future                dáŋàmbà                       will freeze in a long time

 

Anterior Non-Past

–immediate                   dálùmbà                        before (it) will freeze very soon

–recent                         dálèmbà                        before (it) will freeze soon

–remote                        dálòmbà                        before (it) will freeze in a very long time

 

Now we will see the tones sequences with the negative version of the tenses. Note that the past and anterior past are considered grammatical tone and so do not shift their tones. This would normally cause an exception due to the final vowel tone-lowering. But unlike a H-L-H trough, a L-H-L peak is acceptable and so the final vowel actually is not an exception to a need of raising the following tone.

 

Negative tenses

 

Past                            

-immediate past             nàŋúmbà                       just didn’t freeze

-recent past                   nàŋémbà                       just didn’t recently freeze                       

-remote past                  nàŋómbà                       isn’t frozen for a long time

 

Anterior Past

–immediate                   nàlúmbà                        not before (it) just froze

–recent                         nàlémbà                        not before (it) recently froze

–remote                        nàlómbà                        not before (it) froze a very long time ago

 

Non-Past

-immediate future           nàŋùmbà                       won’t very soon freeze

-near future                    nàŋèmbà                       won’t soon freeze

-remote future                nàŋàmbà                       won’t freeze in a long time

 

Anterior Non-Past

–immediate                   nàlùmbà                        not before (it) will freeze very soon

–recent                         nàlèmbà                        not before (it) will freeze soon

–remote                        nàlòmbà                        not before (it) will freeze in a very long time

 

Tense and Mood:

 

To see clearly how moods will affect a verb, we will use the modals that Proto-Drem has available and see how a verb root is affected. The below examples will use the verb root ‘b’èmbá’ which means to freeze, be concentrated, to get stuck to. Note that the examples above will be used, as well as the tense marker ‘nàŋú’ which is for the negative immediate future. Note that for some of the examples, if the verb root vowel –á is the last vowel of the sentence, then due to repair strategies, the tone would change to a low tone as high-tones are dropped to a low-tone at sentence final position. The toneless vowels would not be modified by that tone rule. The examples that will most likely show this will be changed to show the surface tone changes. The sentences that will most likely need an addition (such as an object or pronoun) will be seen with a parenthesis.

 

Immediate non-past

 

Optative                        -wânàŋùmbá( )               … have no wish/desire to yet freeze …

Indicative                       -gènàŋùmbe                  … is not yet frozen. [It can also be said “… is almost frozen.”]

Imperative                      -sínàŋùmbe                   … don’t yet freeze!

Injunctive/Advisory          -sànàŋùmbe                  … (you) not freeze …

Ru particle                     -rùnàŋùmbà                   … *isn’t* yet frozen

Habitual/Iterative            -fánàŋùmbà                   … doesn’t frequently freeze yet.

Decisional                     -gbinàŋùmbá( )               … decided to not yet freeze …

Libetic                           -líbénàŋùmbà                 … will not yet freeze (to death)

Hortative                        -sènàŋùmbe                  … let’s not freeze … just yet.

Conditional - realis         -ŋánàŋùmbá( )               … if not yet frozen …

Conditional - Irrealis         -ŋánàŋùlimbá( )              … if not yet frozen …

Validational                   -ŋkénàŋùmbà                 … not yet frozen (I swear!)

Dubitative                      -régûnàŋùmbá( )             it is not doubtful (that) … yet frozen                    

Progressive                   -ázânàŋùmbà                 … isn’t freezing yet

Benefactive                    varied

 

Now we will see the tones sequences with an immediate past tense, and see how the tone sequences change.

 

Immediate past

 

Optative                        -wânàŋúmbá( )               … have no wish/desire to yet frozen …

Indicative                       -gènàŋúmbe                  … is not just frozen. [saying in a strange way that it is frozen longer than just very recently]

Imperative                      -sínàŋúmbe                   … don’t yet freeze!

Injunctive/Advisory          -sànàŋúmbe                  … (you) not freeze …

Ru particle                     -rùnàŋúmbà                   … *isn’t* yet frozen

Habitual/Iterative            -fánàŋúmbà                   … doesn’t frequently freeze yet.

Decisional                     -gbinàŋúmbá( )               … decided to not yet freeze …

Libetic                           -líbénàŋúmbà                 … will not yet freeze (to death)

Hortative                        -sènàŋúmbe                  … let’s not freeze … just yet.

Conditional - realis         -ŋánàŋúmbá( )               … if not yet frozen …

Conditional - Irrealis         -ŋánàŋúlimbá( )              … if not yet frozen …

Validational                   -ŋkénàŋúmbà                 … not yet frozen (I swear!)

Dubitative                      -régûnàŋúmbá( )             it is not doubtful (that) … yet frozen                    

Progressive                   -ázânàŋúmbà                 … isn’t freezing yet

Benefactive                    varied

 

Causatives and Tones in Proto-Drem:

 

Causatives are a fairly important thing in Proto-Drem, and fairly common as well. The three causatives to be seen here are the main causatives and not the ‘specifying-verb’ type causatives that are also known. The three forms are ‘to make/create’, ‘to let/have’, and ‘to give’. The first style is well known with the use of bò and mbòwə. These two verbs which form the ‘make/create’ causative and will be seen. The second form is nèbè and nèbə. The third form we will use is the ‘give’ form and uses gbare. The examples will generally use b’èmbá and ndòndà like above to show how causatives can be used.

 

Note that the causative, like verbal extensions forms a verb “stem”, the causative marker will go after the modal marker if the modal does replace to verb roots final vowel. The whole TAM+Verb complex is considered a “macrostem” for tone spread outside of the stem itself when used in clauses and sentences.

 

To make/create:

 

Bò

 

b’èmbàbò                      caused to freeze
gèb’èmbebò                   caused … to freeze.
gènàŋùmbebò                caused … to become frozen just a bit ago.          

 

mbòwə

 

b’èmbàmbòwə               caused to freeze           
gèb’èmbembòwə            caused … to freeze.
gènàŋùmbembòwə         caused … to become frozen just a bit ago.

 

To let/have:

 

Nèbè

 

ndòndànèbè                   to show (to a person)
gèndòndenèbè               caused to be shown (to a person)
gènàŋùndenèbè             caused to be recently shown (to a person)

 

nèbə

 

ndòndànèbà                   to show (to a spirit/animal)
gèndòndenèbà               caused to be shown (to a spirit/animal)
gènàŋùndenèbà             caused to be just shown (to a spirit/animal)

 

To give:

 

Gbare

 

wègbare                        to feed
wègbaregè                     caused to be fed
wègbaregènàŋù              caused to be just fed