Professional Documents
Culture Documents
mqwertyuiopasdfghjklzxcv
bnmqwertyuiopasdfghjklzx
The origin of tones in the
cvbnmqwertyuiopasdfghjkl
languages of New Caledonia
zxcvbnmqwertyuiopasdfghj
klzxcvbnmqwertyuiopasdfg
hjklzxcvbnmqwertyuiopasd
fghjklzxcvbnmqwertyuiopa
sdfghjklzxcvbnmqwertyuio
pasdfghjklzxcvbnmqwertyu
iopasdfghjklzxcvbnmqwert
yuiopasdfghjklzxcvbnmqwe
rtyuiopasdfghjklzxcvbnmq
wertyuiopasdfghjklzxcvbn
mqwertyuiopasdfghjklzxcv
bnmqwertyuiopasdfghjklzx
David Valls
Content
1.
Introduction.................................................................................................. 2
2.
3.
4.
5.
4.1.
The Lapita.............................................................................................. 6
4.2.
Oceanic-Papuan contact.........................................................................8
4.3.
4.4.
4.5.
Tonogenesis.......................................................................................... 14
5.2.
6.
Conclusions................................................................................................. 21
7.
References.................................................................................................. 23
1. Introduction
In New Caledonia there are five languages with contrastive tones: Pac,
Cmuhi, Num, Drubea and Kwnyii. Ethnologue (SIL International, 2013) and
WALS (Matthew S. Dryer, 2013) classify Kwnyii as a dialect of Num.
Ethnologue also states that for some it is considered a separated language.
Other resources, such as Blust (2013, p. 659) and Rivierre (1993, p. 155) state
Kwnyii as separated language as for this essay this is not relevant at all.
I will work in two hypotheses on why these New Caledonian languages would
have developed tones. The first hypothesis is language contact. These five
languages could have developed tone due to language contact either by the
surrounding languages or because its speakers came from somewhere else
where they got in contact with tonal languages, and had brought the tones with
them to New Caledonia. The second hypothesis would be that these languages
would have developed tone by themselves, with no external influence. After
exposing the facts, in the conclusions Ill value which of the two hypotheses is
more plausible according to my personal view, but also according the general
view of the linguistic community exposed.
I must say that the references from these languages are scarce and that some
of the books are quite old already and hard or impossible to find anymore.
Malayo-Polynesian,
Eastern
Malayo-Polynesian,
Oceanic,
Central-
Checking several resources, such as Blust (2013), Ross (1988) and Lynch
(2002) is possible to see that theres no evidence for tone in Proto-Oceanic or
even in Proto-Austronesian languages. Therefore tone it is a feature that has
developed in more recent times and, indeed, the rest of the languages from
New Caledonia do not have tone either. When it comes to Austronesian
languages, Remijsen (2001, p. 3) points out that tone is found in the following
3
In general, the languages of this region are quite small. Two of the four largest
languages are spoken in the Loyalties, and two others were introduced by
immigrant laborers from French Polynesia (Wallisian), and Indonesia (New
Caledonian Javanese) (Blust, 2013, p. 125). The largest native New Caledonian
language is thus Pac, one of the tonal languages, with about 7,300 speakers in
2009. Unlike the Solomons, where 93 percent of the population is of Melanesian
descent, or Vanuatu, where this figure reaches 98 percent, only 42.5% of the
216,494 residents of New Caledonia and the Loyalties in July, 2005 were of
Melanesian origin. This gives a native population of about 92,000 with 32
languages, hence an average language size of about 2,875, the smallest in the
Pacific apart from Vanuatu (Blust, 2013, p. 145).
Figure 2: Map of New Caledonia in context. Source: Google Maps.
Number of Speakers
Pac
(year)
7300 (2009)
Cmuhi
2600 (2009)
2180 (2009)
Drubea
2009 (1996)
Figure 3: Area of where these tonal languages are spoken. Source: Google Maps
languages in Oceania, it is not surprising that there has been contact of various
kinds among them. In the western part of the region, there are many languages
with small populations and small territories. An area of just a few square
kilometers may therefore house a number of distinct languages. The area
covered by individual languages in the eastern part of the region is often much
larger, though much of this territory is sea. However, people's seafaring skills
were correspondingly greater in Eastern Oceania, and the sea was more a
vehicle for, rather than a barrier to, inter-language contact. See Figure 1 in
Chapter 2 for the geographical references just mentioned.
The great bulk of the Austronesian languages of the New Guinea region belong
to the Oceanic subgroup; indeed, all languages east of Cenderawasih 2 Bay
belong to it. Most of the remaining Austronesian languages of West Papua
belong to the South Halmahera-West New Guinea group. As the name implies,
it links most closely to the Austronesian languages of the southern half of the
island of Halmahera. The exception is a small set of languages in the Bomberai
Peninsula, which forms a subgroup with the language of the Tanimbar Islands of
Maluku and which belongs with them to the Central Malayo-Polynesian
subgroup (Blust, 2013, p. 132). As said before, Proto-Oceanic may have been
spoken somewhere along the north coast of New Britain (Foley, 2006, p. 361)
by the Lapita (Foley, 2006, p. 362). Other researchers dispute the claim that
the Oceanic speakers were the originators of this culture complex and
maintained instead that it predated their arrival in New Britain (Foley, 2006, p.
362). New Britain, the possible homeland of the Oceanic languages, is situated
far to the east of Cenderawasih Bay, the location of the closest South
Halmahera-West New Guinea languages, which are the closest living relatives
2 Although theres a map, it is impossible to show with detail all these names
as it would require filling this essay with maps. The given map is just a
reference, for more accurate information I recommend to take a map a part.
9
of Oceanic languages. This suggests either the unlikely scenario of a direct onestep migration to New Britain from much farther to the west or, more likely, a
step-by-step migration along the north coast of New Guinea, with all traces
having been erased by a later east-to-west back-migration along this coast by
Oceanic speakers.
From the New Britain region, ancestral speakers of Oceanic languages spread
rapidly out into the more remote southwest Pacific. Within the New Britain
region there are three subgroups of Oceanic languages, among the nine or so
now fairly well established Oceanic subgroups in the Pacific (Foley, 2006, p.
364). The Admiralties cluster, found in and around Manus Island; the Saint
Matthias group, in the islands of the same name north of New Ireland; and West
Oceanic, which includes all the remaining Oceanic languages of the region.
Western Oceanic, in turn, is composed of three very large and diverse groups:
the Papuan Tip cluster, along the southeast coast of New Guinea and the
adjoining islands groups; the Meso-Melanesian cluster, covering New Ireland,
northerly coastal areas of New Britain, and the northwest half of the Solomon
Islands, including Bougainville; and finally, the North New Guinea cluster, found
in the west and on the south coast of New Britain and spreading from east to
west along the north New Guinea coast from the Huon Gulf to the Sarmi coast
in West Papua (Foley, 2006, p. 364). The central position of New Britain in the
distribution of these subgroups is evidence for its claim to be a central
dispersal point of migration, a possible homeland of Proto-Oceanic. Much less is
known of the other two groups of Austronesian languages found in the far west
of the New Guinea region. The South Halmahera-West New Guinea languages
are the closest relatives of the Oceanic languages, both genetically and
geographically. They spread from the languages of the islands Biak and Yapen
and from the coastal mainland of Cenderawasih Bay, along the north coast of
the Birds Head, through the Raja Empat Islands off the western tip of West
Papua, and on to the southern half of Halmahera. They are thus restricted to
the far northwest of the New Guinea region. The southwest appears to be
occupied by the Central Malayo-Polynesian subgroup. On mainland New
Guinea, only a small number of languages along the south coast of the Birds
Head of West Papua belong to the Central Malayo-Polynesian subgroup. They in
10
turn are included with the languages of the island groups south of West PapuaKei, Taninber, and Aru, forming a subsubgroup within Central MalayoPolynesian, which suggests a migration into West Papua from one of the
islands, probably Tanimbar (Blust, 2013, p. 10), the most distant of the three.
In coastal areas, for example, we find Oceanic languages interspersed with
Papuan, and although Oceanic languages predominate in the islands, there are
nevertheless many insular Papuan languages as well. Given that many Western
Oceanic languages are geographically contiguous with Papuan languages, and
that most languages of both groups are spoken by small populations, the
potential for sociolinguistic contact between Oceanic and Papuan languages is
considerable. Lexical copying, for example, occurs frequently, and in both
directions. In a number of parts of the region, however, contact has resulted in
widespread bilingualism (Lynch, et al., 2002, p. 15).
In the following map we can see New Britain, where supposedly Oceanic
languages had originated and spread out from. White dots show none tonal
languages, red dots show complex tone languages and pink dots simple tone
languages. As we can see, there are some tonal languages around, but going
south, towards New Caledonia, all are none tonal.
Figure 5: Adapted map of the origin of the Oceanic languages and tonal languages
around. Source: (Matthew S. Dryer, 2013)
11
One more doubtless fact is that Oceanic languages spread out from north to
south, form Papua to New Zealand, including New Caledonia in the middle of
12
their way. However, none of the authors Blust (2013), Foley (2006), Lynch
(2002) and Ross (1988) claim tonal borrowing, not even clear lexical borrowing
from Papuan languages to the New Caledonian ones. Since the origin of the
Oceanic languages seems to be in New Britain, it might have been possible
that some of these languages had just spread out way down south with few or
no contact with tonal Papuan languages.
Figure 7: Possible trip from New Britain to New Caledonia. No tonal languages are
found until New Caledonia. Source: (Matthew S. Dryer, 2013)
14
The relative chronology of these two major phenomena for each language led
to different types of tone systems (Ferlus, 2002, p. 299). If the devoicing of
plosive initials took place after the complete loss of the two laryngeal features
- and -h, then the language developed a six-tone basic system as in northern
Viet-Muong languages (Vietnamese, Muong, Tho). If the devoicing took place
after the change of the laryngeal constriction -, while final spirant -h was still
preserved, then the language developed a four-tone basic system as in
southern Viet-Muong languages (Maleng, Arem, Sch/Ruu C, Thavung) (Ferlus,
2002, p. 299).
As was mentioned above, the origin of Vietnamese tones was clearly explained
by Andre G. Haudricourt in 1954. According to the author, sk-nng tones
derive from an ancient final glottal stop, hi-ngaa tones from an ancient final
spirant while, by contrast, ngang-huyn tones developed within vowel-final
contexts. He reconstructed Vietnamese tonogenesis in terms of three stages:
1. Ancient Vietnamese was a toneless language.
2. The final glottal stop - changed into a rising contour. The final spirant
-h changed into a falling contour. The final - remained at an even pitch. The
result was a three tone system made up of the ancestors of the three pairs,
ngang-huyn (from -), sk-nng (from -) and hi-ngaa (from -h).
3. The confusion of voiced initial obstruents into voiceless split the threetone system into two series. The result was a six-tone system.
The same type of explanation can be extended to all Viet-Muong languages,
including those that show a four-tones system with the final -h preserved, as
well as so-called register languages (Ferlus, 2002, p. 301) .
Another source of tonal contrast is a voicing contrast in obstruents. Voiced
obstruents are known to lower the pitch of the following vowel, and voiceless
obstruents may even raise it (Yip, 2002, p. 34). There are various reasons for
this. Voicing in obstruents is associated with slacker vocal folds, and a lowered
larynx. Both these tend to lower pitch, at least at the start of the following
vowel. Voiceless obstruents seem to have tenser vocal folds, and thus tend to
raise the pitch on the following vowel. If the consonants lose their voicing
16
contrast over time, the pitch difference may persist, and the burden of contrast
is then shifted from a voicing contrast in the consonant to a tone contrast in
the vowel (Yip, 2002, p. 35).
Lets take a look at an example. In Kammu, a Mon-Khmer language of
Cambodia (Yip, 2002, p. 36), the southern dialects have a contrast in onset
voicing on obstruents, but the northern dialects have lost the voiced set and
instead they contrast a H tone where the north has a voiceless onset, and a L
tone where the north has a voiced one:
(1) Kammu dialects
South
North
kla stone
If a language is already tonal from some other source, such as final consonants
in example (2) , this change may double the number of tones. For example, in
Cantonese Chinese we see a six-tone system arising out of a three-tone system
as the result of the influence of the historically voiced onsets, all now voiceless.
Stop-final syllables underwent a secondary split probably related to vowel
tenseness, and are excluded from the following table (Yip, 2002, p. 36) .
(2) Cantonese tonal split
Historically voiceless onsets si53/55
si35
sei44
tshi21
rely on
tshi24
si33
serve
of
laryngeal
distinctions,
particularly
voicing,
in
the
surrounding
consonants (Yip, 2002, p. 38). This does not mean that modern tone languages
cant also have a voicing contrast, and indeed many do. There could be many
reasons for this: for example, the voicing contrast may have re-emerged after
tonogenesis took place, or tones may have arisen from an aspiration contrast
instead (Yip, 2002, p. 38). Sometimes tones developed too long ago that
theres no way how did come out.
either the aspirated consonants or the tones that are found in several
languages of New Caledonia.
I will try to describe the nature of these reduplicative processes which, in fact,
were no longer productive in Proto-New Caledonias Mainland (Rivierre, et al.,
1993, p. 154). Rather, aspirates most probably existed in that Proto-language
and were thus directly the origin of tones that appeared later in languages of
the center and southernmost part of the Mainland.
Cmuhi and Pac, localized in the Center North of Mainland, and Drubea,
Num, and Kwnyii in the Far South apparently had a parallel development of
tone in non-tonal environments (Rivierre, et al., 1993, p. 155).
Rivierre (1993, p. 156) says that in 1968, Haudricourt proposed an explanation
for this tonogenesis in an article entitled , the main points of which are as
follows:
1. There is a regular correspondence between the aspirates of some
northern languages, the voiceless fricatives of others, and high tone of
the tonal languages of the Center. These correspondences relate to oral
and nasal consonants but exclude prenasalized consonants.
2. High tone and aspirated consonants have the same origin: they result
from previous syllable reduplications.
Rivierre (1993, p. 156) says that Haudricourt supports this view with two
observations:
First, syllabic reduplication can produce geminate consonants through the loss
of an unstressed pretonic vowel: C1(V1)C1V1 C1C1V1
The phonemic long consonants which result from this development are indeed
sometimes aspirated, as in certain Polynesian Outliers.
Secondly, Haudricourt notes that a number of commonly reduplicated words in
Melanesia and Polynesia have cognate forms with high tone in the Northern
tonal languages. According to Haudricourt, the correspondences observed in
the Mainland can be explained by the developmental pattern shown in Figure 8.
19
That
is,
syllable
reduplication
produces
long
consonants,
which
then,
20
21
6. Conclusions
In this essay, we have seen two hypotheses about how tones could have
originated in the tonal languages of New Caledonia. The contact hypothesis is
plausible in many other languages, only that in order for one language to
acquire such feature, there must be a prolonged contact in time, and it seems
that this is not the case with these New Caledonian languages, or at least
theres no information about it, besides lexical items from other languages
would be very noticeable in these languages. Therefore, the contact hypothesis
as a way of acquiring tone is not very probable. Anyway, the Lapita people
ancestors of the current inhabitants of New Caledonia spoke Proto-Oceanic,
which it did not have tones.
We have seen of how a language that has no tones can generate tones. And we
had reviewed Haudricourt's hypothesis which explains how the tones were
generated in New Caledonian languages. The process is summarized as
follows:
-
wa
So, according to all the given and examined data, the most plausible
hypothesis
is
tone
self-development
due
to
phonetic
causes.
The
22
languages of the world, and in fact, according to available data 60-70% of the
world's languages are tonal languages (Yip, 2002, p. 1), which is a lot.
In classical Latin, for example, there were many geminate sounds: cc, dd, ll, bb,
mm, nn, rr, ss, tt, pp (Borrell & Mir, 2001), but it did not produce tone, instead
there was a relaxation of sounds in most cases, that made gemination to be
lost. Let's look at some examples from Latin to Catalan:
kppa copa 'cup'
crassum gras 'fat'
gibbus gibus gep hump
tottus totus tot 'all'
Or even modern Catalan has plenty of geminated sounds such as:
batlle [] mayor
allot [ll] boy
immediat [mm] immediate
Which again, tend to be relaxed and even in some dialects some have
disappeared such as [ll].
So going further, we should ask why in some languages this phenomenon leads
to tone development and in some others these geminate sounds just melt into
a simpler sound, but the answer to this question it would be part of another
essay.
23
7. References
Blust, R., 2013. The Austronesian languages. Revised edition ed. Canberra:
Asia-Pacific linguistics.
Borrell, E. & Mir, M., 2001. Gramtica llatina: el llat normatiu i la seva
evoluci fins el catal. Barcelona: Universitat Oberta de Catalunya.
Ferlus, M., 2002. The origin of tones in Viet-Muong. Paris: National Center for
Scientific Research.
Foley, W. A., 2006. The languages of New Guinea. Annual Review of
Anthopology, p. 50.
Gordon, M. & Maddieson, I., 2004. The phonetics of Pac vowels. Oceanic
Linguistics , 43(2), pp. 296 - 310.
Lynch, J., Ross, M. & Crowley, T., 2002. The Oceanic languages. Great Britain:
Curzon Press.
Maddieson, I. & Pang, K.-f., 1993. Tone in Utsat. Honolulu: University of Hawai'i
Press.
Matthew S. Dryer, M. H., 2013. The World Atlas of Language Structures online.
[On line]
Available at: www.wals.info
[Last access: 15 12 2013].
Remijsen, A. C. L., 2001. Word-prosodic systems of Raja Ampat languages.
Leiden: Universiteit Leiden.
Rivierre, J.-C., 1973. Phonologie compare des dialectes de l'extreme-sud de la
Nouvelle Caldonie. Paris: SELAF .
Rivierre, J.-C., Edmondson, J. A. & Gregerson, K. J., 1993. Tonality in
Austronesian languages. Honululu: University of Hawaii Press.
Ross, M., 1988. Proto Oceanic and the Austronesian languages of Western
Melanesia. Canberra: Australian National University .
Sheppard, P. J., 2011. Lapita Colonization across the Near/Remote Oceania
boundary. Current Anthropology, 52(6).
SIL International, 2013. Ethnologue. [On line]
Available at: http://www.ethnologue.com
[Last access: 8 12 2013].
24
25