Cham, A Preliminary Sketch of Phan Rang (Thurgood) PDF
Cham, A Preliminary Sketch of Phan Rang (Thurgood) PDF
Cham, A Preliminary Sketch of Phan Rang (Thurgood) PDF
Graham Thurgood
California State University, Chico
1.0 Introduction
The extraordinary French scholar Coeds noted that Cham is the earliest
attested Austronesian language. Coeds dated the Cham inscription found at Tra-
kiu near the old Cham capital of Indrapura as being from the middle of the
fourth century, describing the inscription as the oldest text, presently known,
written in a Malayo-Polynesian dialect. The language of the text is associated
with the once flourishing kingdom of Champa, a kingdom first mentioned by
the Chinese around 190 to 193. Champa reached its zenith about the sixth
century, continuing to flourish until the Vietnamese push to the South in the
tenth century began its slow demise. At the time of the first inscriptions, the
Chamic languages were still a largely undifferentiated dialect continuum, but in
the subsequent fifteen hundred or so years of change, realignments in patterns of
affiliation and language contact restructured stretches of the original dialect chain
into distinct languages and distributed the speakers over a much wider area. No
longer functioning as the lingua franca of the kingdom of Champa, Chamic lives
on in its modern descendants: the Tsat spoken on Hainan, the Rade, Jarai, Haroi,
Chru, and Roglai spoken in the southern Vietnam highlands, the Phan Rang
Cham spoken in Vietnam, the various Western Cham communities of
Cambodia, and the Acehnese of north Sumatra.
Quite correctly, the literature simply assumes that the mainland Chamic
languages form a subgroup, but there have been minor questions about the
relationship of Acehnese with the mainland Chamic languages. Niemann reached
the correct solution as early as 1891, first subgrouping Acehnese and Cham
together on the basis of similarities in the verbal morphology, the treatment of
inherited vowels, and in various instances of apparent lexical agreement, and then
positing a migration of Chams to Aceh (cf. Thurgood 1999 for extended
discussion).
Phan Rang Cham (or, Eastern Cham) is a Chamic language spoken in
southern Vietnam by 35,000 to 50,000 people in the area around the towns of
Phan Rang and Phan Ri. It is closely enough related to Western Cham for the
two to be considered dialects of one another. Baumgartner (1998:1) notes that
the differences between the two are primarily matters of pronunciation and
vocabulary with the grammars being almost identical. As for number of
speakers, he notes that Western Cham is the numerically larger of the two, with
300,000 to 350,000 speakers in Cambodia, and another 35,000 or so speakers in
the Mekong delta region of Vietnam, particularly around Chau Doc, Tay Ninh,
and Saigon.
century that scholarship would return to Crawfurds position. Toward the end of
the nineteenth century and for the first half of the twentieth, the classification of
the Chamic languages was controversial due to the belief of scholars like tienne
Aymonier, who thought, along with many of his contemporaries, that the
Austronesians had migrated to the islands from this part of the Southeast Asian
mainland. Failure to distinguish between borrowed Mon-Khmer elements and
inherited Austronesian elements led Aymonier to write (1889) that Cham formed
a kind of transitional language genetically intermediate between Khmer and
Malay. Schmidt (1906), influenced by Chams Mon-Khmer influenced
typological characteristics and its numerous Mon-Khmer borrowings, described
Cham as an Austroasiatic mixed language (Austroasiatic is Mon-Khmer plus
Munda). In fact, as late as 1942 Thomas Sebeok was to misclassify Chamic
languages as Austroasiatic.
From a modern perspective, it is evident the Chams reached the
mainland from a site probably somewhere in West Borneo some two thousand
years ago. The Chamic languages are far too closely-related and far too easily
reconstructed to date back much over two thousand years, let alone the six
thousand or so that would be required to account for Austronesian. The borrowed
Mon-Khmer elements can readily be distinguished from the inherited
Austronesian elements, leaving a clearly Western Malayo-Polynesian language
behind, and in the process providing a magnificent venue for studying the effects
of language contact on language change.
As an aside, although Moken is sometimes classified as Chamic,
careful comparative examination makes it clear that, despite certain areally-
expected typological similarities and their common membership in
Austronesian, there is no special subgrouping relationship between Moken and
Chamic (however, see Larish [this volume] for an opposite view).
The breakup of ancient Cham into various modern languages followed
an almost predictable pattern. For roughly the first millennium, the Chamic
languages were a largely undifferentiated dialect chain that certainly extended
along the coast of Vietnam and may even have stretched as far south as the east
coast of the Malaysian peninsula (although Chamic-like features now found
along the east coast of Malaysia certainly date from a much later Acehnese influ-
ence). The dialect chain along the coast of Vietnam broke into clearly distinct
languages when the Vietnamese moved south down the coastline, a movement
that pushed many Chamic speakers up into the highlands and destroyed much of
the rich interactional network between the coastal communities. New sets of
language networks developed for almost all Chamic speakers, with some like the
Haroi eventually coming to be part of a Mon-Khmer social network, while
others like the Phan Rang Cham eventually came to be part of the lowlands
society dominated by the Vietnamese.
As for Cham itself, it mostly began its split into Western and Phan
Rang Cham around the end of the fifteenth century with the fall of the southern
capital to the Vietnamese.
Cham has its own literary tradition, one that dates back to the first
inscriptions written in an Indic script in the middle of the fourth century.
Various texts and inscriptions have been gathered, primarily through the work of
various French scholars. However, much of the epigraphic work remains to be
done and until then much of the early history of Champa and of Cham will
remain beyond our reach.
1.2 Sources
on heavily. Several other works have dealt with aspects of the historical
phonology (Blood 1962, Thurgood 1996, 1999) but more remains to be done.
No detailed grammars exist, although there is a good grammatical
sketch of Western Cham by Baumgartner (1998). For Phan Rang Cham, there
are a handful of articles, including David Blood (1977) on Cham sentences and
Doris Blood (1977) on Cham clause and sentence-final particles. The discussion
of discourse data in this chapter comes largely from Doris Blood (1978), which
gives three Cham texts along with a sophisticated and insightful analysis. The
numbers following the cited sentences below refer to these Blood sources; the
citations are obvious, except for Blood 1978, which contains three separate texts;
for this source the three texts are coded as 6.1, 6.2, and 6.3, plus a line number.
, particular Blood . Two other sources of Cham texts exist: an extensive French
collection of Cham manuscripts catalogued in Pierre-Bernard Lafont, Po Dharma,
and Nara Vija (1977) and six reels of microfilms of Cham documents in the
Echols Collection of the Kroch Library at the University of Cornell.
Several good dictionaries exist: Aymonier and Cabaton (1906), written
in French, has its idiosyncracies and naturally is somewhat dated, it is still quite
usable. Moussay (1971) is also useful; in addition to citation in the Cham, it
has two additional representations, one a transliteration of the Chamic script, and
the other a modified transliteration intended to approximate the contemporary
pronunciation. Still another is a Cham-Vietnamese dictionary by Bui Khanh
The (1995), which seems to take most of its entries from Aymonier and Cabaton
(1906); this dictionary uses the same script as Aymonier and Cabaton, but
replaces their transliteration with one that is more transparent.
Overall, more historical work than synchronic description seems to
have been published on Chamic and Cham. Thurgood (1999) presents an
overview that incorporates most of the available historical work.
This grammatical sketch is text-based, a necessary but not ideal
constraint. Despite this limitation the attempt has been made to give a precise
and succinct statement regarding the polysemy of various, often historically-
related formatives, although at times there was not enough data to do this.
2.0 Phonology
The phonotactics of the word are the key to the phonology of Cham (Blood
1967). Like the surrounding Mon-Khmer languages, most words are disyllabic
and iambic. The rare trisyllabic word is often morphologically complex and, as
Blood notes, frequently collapses into a disyllable, following interesting but still
unclear paths of reduction.
Contemporary Phan Rang Cham is rapidly going from disyllabic to
monosyllabic. A small number of words of course were always monosyllabic,
but under the internal influence of final stress and the external influence of
Vietnamese now even the typically disyllabic words are increasingly becoming
monosyllabic and motivating much of the phonological variation within Cham.
Within the last several generations initial syllables were first omitted in
informal, colloquial speech and now seem to have been dropped entirely by some
speakers. Doris Blood (1962:11) gives a vivid instance, citing the following
variants of the word new: perw ~ prw ~ phirw ~ phrw ~ firw ~
frw. She notes that the scholars tend to maintain the full forms in speech, but,
typically, non-scholars modify the first syllable, reducing its vocalism,
subjecting it to assimilation, or loosing it entirely.
The segment inventories correlate directly with the syllable structure
and its iambic stress pattern. The preliminary syllables (using terminology
introduced by Blood 1967) are unstressed, shorter in duration, and typically CV,
although occasionally a final -N occurs that may assimilate to the following
4
consonant. Blood (1967) notes that as the vowel disappears, the preliminary
syllable is restructured, with the consonantal onset sometimes becoming syllabic
and sometimes, where the phonetics are compatible, becoming the onset for a
main syllable cluster (see the examples in the preceding paragraph).
vl. stops p- t- c- k- /-
aspirated stops p- t- c- k-
implosives - -
fricatives s- h-
nasals m- n- n-
semivowels/liquids w- l-, r- y-
vl. stops p- t- c- k- /-
aspirated stops p- t-
fricatives s- h-
nasals m- n-
liquids l-, r-
The main syllable codas (Table 3) are not only even more limited than
preliminary syllable onsets, but they also have special characteristics (Blood
1967).
vl. stops -p -t -k -/
fricative -h
nasals -m -n -N
liquids -l, -r
2.2 Tones
Note that it was the breathy voiced quality induced by the voiced
obstruents, not the voicing per se, that resulted in the low tones. The historical
developments are complicated slightly by the spreading of voice quality
differences from the preliminary to the main syllable but the conditions are still
transparent. In fact, the overall transparency of the process makes Phan Rang
Cham invaluable for its insights into the mechanisms of tonogenesis.
2.3 Reduplication
Note that the last example (Table 7) fits neither pattern particularly
well. In any case, the function of reduplication seems to be to provide some sort
of distributive meaning. That is, it functions much as do various matched pairs
of words, both with the same meaning: hya cO/ cry, with both morphemes
meaning cry and hua/ baN eat, with both morphemes individually also
meaning eat.
Cham morphemes never consist of less than a syllable, and they always
consist of whole syllables. Overwhelmingly, the morphemes are phonologically
separate words, with the possible exception of the causative prefix pa- in Modern
Cham, which may still be marginally productive.
Word classes are defined distributionally. Verbs can be negated, nouns
can be pluralized and can occur with classifiers, classifiers can occur with
numerals, and so on. The problematic cases are locative nouns and co-verbs, for
which see sections 3.4 and 4.5.
Until recently Cham served for a long time as a lingua franca first along
the coast and then in the highlands of southern Vietnam. This fact, combined
with its relative lack of morphology, probably accounts for the fact that it is a
configurational language with a fairly rigid word order. The staple of Cham
clausal syntax is the verb-centered, basic declarative sentence. Most of the other
clause types are essentially extensions of the declarative clause modified by
sentential particles or by deletions required by interclausal cohesion. Within the
basic clause, the number of arguments is determined by the semantics of the
verb. The two core arguments, the subject and the object, are only marked by
word order, with the subject preceding the verb and the object following it.
Cham thus is a SVO language. Indirect objects never seem to appear as a third
unmarked core argument, rather they are always marked as such by a preposition.
Stative intransitive clauses consist of a subject plus stative verb.
This pattern is used for descriptive adjectives; equational sentences follow the
topic-comment pattern discussed below. As is obvious from example (2) and
8
others elsewhere, the stomach, the liver, or some major organ are seen as the
seat of the emotions, and the metaphor of an emotion going into the stomach or
liver is common in the languages in the area.
In some instances, Cham uses a topicalization construction in which
the topic is marked with the distal demonstrative nan. This construction is the
typical way for marking equative sentencesthe apparent origin of the
constructionand is also used widely for marking other kinds of topics. In her
brief discussion of topic-comment sentences, Doris Blood (1977:63-64) presents
the example in (3) with its double occurrence of nan, the first functioning as
adnominal modifier, the second as a topic marker.
In example (3), the topic then is oN nan nan. The comment consists of the
classifier uraaN person followed by the head noun toy guest.
The topic constituent is not limited to nouns but can involve more
complex phrases and even clauses, as shown by the following two examples in
which the topical constituent is underlined:
(4) ptaaw Taluy/ Na/ nan oh /jOw/ ptaaw pya/ o. (Blood 1977:64)
king Taluch do DIST NEG correct king real NEG.
King Taluchs behavior was not that of a true king.
(5) Na/ kn)i nan pyh /wa/ Na/ yaaN, (6.2.34-5)
make kanhi DIST in.order rub make spirits
(They) make the kanhi (a kind of instrument) to appease the spirits
The verb hu have; get can also be used as a main verb. It then is preceded by
the subject, as in
(9) crO/ mEE/ MKaam Na/ ___ rilO baN plO~h. (6.1.37)
charok mom Kam make ___ meat eat finish
Your charok, Kams mother has already made (it) into food.
(10) kn) uraaN oh /wa/ ___ m/in (6.2.33)
kanhi person NEG rub ___ play
The kanhi (is an instrument) that people do not play ___ for fun.
(11) mEE/ MKaam hmit n/ rimOON baN ____, cO/ hya (6.1.188)
mother Kam hear child tiger eat ____ cry cry
Kams mother, hearing her daughter had been eaten by a tiger, wept
Notice that the final particle y has increased the directness of the
command, but kaay, n), and ah have taken away some of the bluntness.
Like many Southeast Asian languages, Cham has a non-compositional
negative imperative co~y dont. (19) co~y di/ co~y ah! (Blood 1977:47)
dont climb dont PTCL Better not climb (it)!In this example, the
force of the imperative has been moderated by the use of the final particle ah.
10
Note that the negative imperative may occur before the verb, clause-finally, or in
both positions. However, in final position, it ispreceded by a pause and is
spoken on a higher level of intonation with a rather sharp fall (Doris Blood
1977:47), a description suggesting a right-dislocated element more than a fully-
incorporated final particle.
Questions follow the same word order as the corresponding declarative
sentences. Questions answerable with a yes or no typically are signaled with
nothing more than a rise in intonation on the last element in the sentence (Doris
Blood 1977:42). Less commonly a yes/no question may be signaled by the sen-
tence-final particle laay Q.
Some particles are more tightly incorporated than the one just
discussed. Blood (1978:43) mentions hu laay, which has the nuance of
possibility from the hu possible; able; get; have.
Content questions, like yes/no questions, use the same word order as
the corresponding declarative sentences, but with the question word inserted in
place of the questioned item. Examples for question words are pa where<
pa/ hlaay at which?, taaw where?, kE~/ what?, thipa~al why?, etc.
The final particle kay/ here, Blood notes, seems to demand a specific
answer.
(25) MKaam tO~O/ ti~ la tuh pi~h ikaan MHl/ tm tnEh (6.1.11)
Kam stay LOC down pour all fish Hlok into basket
Kam stayed below and poured all of Hloks fish into her basket
PC PR Cham PC PR Cham
The table includes one of the two plural markers: the plural form meaning ka~w/
*gp group; other, which shows up in various plural pronouns, for instance the
second person plural, which simply adds ka~w/ to the singular.. The other, found
in the combination khOl n)u, is a pluralizer for the third singular pronoun n)u; the
combination is identical to the combination chng n meaning they in
Vietnamese.
Nouns characterizing social roles such as grandmother, elder sister, step-
mother, king, and such are frequently used in the place of pronouns. In the texts
examined, the third person pronoun is only used for humans.
3.2.2 Demonstratives
case they follow the constituent they modify: Only the minimal two-way
distinction between a distal and a proximal demonstrative is made. Finer
semantic distinctions are found in the directional coverbs (see section 4.3). The
distal demonstrative nan is the one that has taken up the duties involved in
marking NPs as anaphoric; the proximal ni is only used when a proximal
meaning is emphasized.
In the texts, the demonstratives also do not seem to occur referring to humans or
animates; whether this is an absolute restriction or just a strong tendency is not
clear. In addition, and as mentioned in section 3.1, the demonstrative nan occurs
as part of the topicalization construction.
3.2.3 Classifiers
The analysis of examples like (30) is problematic as pO~h fruit; CLF for fruit can
be analysed both as a head noun and as a classifier. Similarly, phun functions
both as the word for tree and as the CLF for trees.
parallel to the way that nouns characterizing social roles such as grandmother,
elder sister, stepmother, king, and such are used in the place of pronouns. In
fact, in the texts examined, pronouns are never used to track nonhumans;
instead, the tracking is done with a combination of zeros (safely ignored in the
discussion to follow), CLFs, and various (other) noun substitutes. Compare the
following segments drawn from the same text.
Note that in this example it could also be argued that pO~h functions as a noun
rather than a CLF, in particular in 32 and 33]
Here, the CLF with tha one is used to first introduce the fruit to the listener
tha pO~h prON yOm d c~lu/ (6.1.125); that this is an initial introduction is made
particularly obvious by the use of the presentational hu have. Several lines
later (6.1.128) the fruit is again introduced but this time to an old woman with
tha pO~h mkya prON kca~an, again marked with tha one indicating that it is
indefinite from the old womans viewpoint. The next overt reference (6.1.131)
pO~h mkya nan that ebony fruit occurs without a numeral; here, the CLF is
being used solely for reference tracking, with the deictic nan that making the
definite reference obvious. Two lines later (6.1.133) the form pO~h mkya
occurs, again without tha one and again being used pronominally. Twenty
lines later (6.1.153) the form pO~h mkya occurs once more, again without tha
one and again being used pronominally.
As for the function of classifier constructions in counting, it is a
decidedly minor one: Classifiers can and do co-occur with numbers, of course,
but more often than not they occur without numbers and, with a great frequency,
numbers occur without classifiers.
clause. They also involve words for time and place, as illustrated by the
following two examples.
These are simple juxtapositions, typically composed of the two most salient
members or characteristics of a class, an activity, or whatever, used as a general
noun. These are exceedingly common in Southeast Asia.
Cham does not mark tense, but it does mark basic aspectual
distinctions. The two items most frequently used to mark aspectual distinctions
are the forms plO~h finish and tO~O/ which as a main verb means live; stay but
also may mean still. . In their aspectual uses these two forms distinguish
ongoing states and activities (PROGRESSIVE) from completed states and activities
(COMPLETIVE).
The differences in the meaning of tO~O/ correlate with its placement in
an overall construction: in (41) tO~O/ live; stay is used as a main verb; in (42)
where it precedes an attributive predicate tO~O/ indicates the continuance of a state
and is best rendered by still; and in (9) where it precedes an activity
predicate tO~O/ indicates the continuance of an activity.
15
(41) MKaam tO~O/ ti~ la tuh pi~h ikaan MHl/ tm tnEh, (6.1.11)
Kam stay LOC down pour all fish Hlok into basket
Kam stayed below and poured all of Hloks fish into her basket and
(42) mEE/ MHl/ mtaay tu/ MHl/ tO~O/ siit (6.1.2)
mother Hlok die hour Hlok stay small
Hloks mother died when she was still small.
(43) MHl/ tO~O/ m/in oh m/ hu ikaan. (6.1.14)
Hlok stay play NEG get have fish
Hlok is still playing and she didnt get any fish.
Example (44) shows still another use of tO~O/. Here it occurs at the beginning of
a clause, marking an overlap of the activities in the ensuing paragraph and the
activities in the previous paragraph.
Similarly, the different functions of plO~h finish also correlate with its
distribution: as a main verb plO~h means finish; finished; sentence-initially as
in (43), plO~h marks the next significant event (a meaning roughly translated here
as then), and clause-finally, as (44), it marks one activity as completed with
respect to the next one.
These forms are transparently related to the basic meaning finish; finished with
differences in scope related to differences in placement.
3.3.2 Negation
Cham reflexives are marked with tra~y \trE~y\ self, a word that derives
from the still extant word for body. Note the idiomatic usage in (52).
Joint action can be marked explicitly with ca~a/ ka~w/, which in these
texts is a construction always marked by a pronoun followed by ca~a/ mutual
which, in turn, is followed by ka~w/ other.
Reciprocal action is marked by using ka~w/ other after the verb, which
in certain contexts is extended to mean together or each other.
17
Note that all three of the examples include twa~ two used to indicate
that the pair of entities that follow form a group.
Cham marks sources with m)N from and goals with tal arrive. The
word tal marks, as the gloss indicates, arrival at the goal. Movement toward a
goal is marked either by the co-verb naaw go or the preposition ti~ to; at,
which will be discussed shortly. As for tal, unlike the other prepositions
discussed here, it has a clearly verbal origin; it is neither an inherited
preposition, nor does it originate in an earlier noun.
m)N and tal occur before a full range of noun phrases, including gerunds
and sentential complements, marking literal and metaphoric spatial and temporal
movement.
The most general and thus least marked preposition is ti~ to; at which
seems to have a generalized locative function. It also occurs with locative nouns
converting them into phrasal prepositions as in (67). It often codes a goal,
including addressees and recipients.
Other prepositions include thON with; and, pa/ at, yaw like.
As is true for a number of the languages of the area, locative nouns
supplement the work that is done by prepositions in English. Locative noun
phrases are genitive constructions consisting of two juxtaposed nouns. The head
noun (the locative noun) designates a place and the following noun designates
whatever is located with reference to that place. Usually, but not always, the
locative noun is preceded by a true preposition, often the very general ti~. In
(38) and (39), the distinct words for put notwithstanding, no preposition
precedes the locative noun.
(75) n)u boh n/ crO/ pa-doN tra~y tko~/ ti~ /ya. (6.1.23)
she see child charok CAU-rise self up LOC water
She saw the baby fish rise to the top of the water.(i.e. raise itself)
(76) muu/ tkr kra/ /y/ thay uraaN Na/ lithay ka muu/. (6.1.47)
grandmother want spy see who person make rice for grandmother
The woman wanted to discover who had made rice for her.
(77) MHl/ m/ Na/ raw rilo. (6.1.34)
20
(78) mEE/ MKaam ca~aN pra~y aaw khan kOON krah (6.1.66)
mom Kam also give clothes bracelet ring
Kams mother also gave clothes, bracelets and rings
(79) min mEE/ MKaam oh pra~y naaw (6.1.70)
but Mom Kam NEG give go
but Kams mother would not let her go,
(80) pra~y ka t~hla/ likaw /y/ bOO/. (6.1.76)
give for me beg look cheek
Please let me see (her) cheek (i.e., see her face).
The most straightforward clausal complements in Phan Rang Cham are found
with verbs of perception and ordering.
(81) pdal MHl/ koy/ m/ tlaaN ikaan pa~ ta~l ti~ la tnN dih (6.1.39-40)
order [Hlok dig get bone fish take bury LOC underneath bed sleep]
ordered Hlok to dig up the bones and bury them beneath the bed.
(82) n)u snN Po Te~pi~ta trun tO~N nu, (6.1.85)
she think [God down help her]
She felt that God had come down to help her.
(83) n/ mEE/ MKaam boh pya~ Hl/ maay thaaN (6.1.102)
child mom Kam see [queen Hlok come house]
Kam and her mother saw Queen Hlok come to the house,
Notice that in these three examples, beyond being embedded in the matrix sen-
tence, nothing special marks these as complements.
Wish- and want-clauses are more distinctly marked: The embedded
complement immediately follows the matrix verb, but here the subject of the
complement clause is marked by ka, labeled for as it occurs elsewhere as a
benefactive preposition.
(84) min cON [ka pO~h mkya nan lE/ trun t~lam li/i muu/]. (6.1.131)
but wish [for CLF ebony DIST fall down into basket grandmother]
She wished that the fruit would fall down inside her basket.
(85) myah uraaN tkr [ka kn)i mn)i syaam sap] (6.2.30)
if person want [for kanhi call good sound]
If they want the instrument to sound good,
(86) muu/ doy/ kol kmay t~ra nan dom lay/ tcO ly,.. (6.1.158-59)
grandmother run hug woman young DIST speak say g.child hey
The woman ran, hugged the young girl and said, Hey grandchild.
21
In (87) the quotative lay/ say occurs immediately after dom speak, another
verb of saying. This marker occurs with both direct and indirect quotations:
(87) n)u pathaw lay/ pru~/ nan n)u Na/ plO~h paj. (Blood 1977:55)
he inform QUOT work DIST he do finish already
He stated that he had completed that work already.
(88) dom plE/ MHl/ ti~ mEE/ lay/ (6.1.13)
speak lie Hlok to mom QUOT
(she) lied about Hlok to her mother, saying,
MHl/ tO~O/ m/in oh m/ hu ikaan. (6.1.14)
Hlok stay play NEG get have fish
Hlok is still playing and she didnt get any fish.
Other typical clausal conjunctions include min but, mjah (jah) if,
since, kaal when, kayua (yua) because, and Nan or (cf. David Blood 1977).
Clausal expression of a purpose is also often done simply by
juxtaposing two clauses with the purpose reading inferable from the context
(91). If a more explicit purpose clause is needed, it is indicated with pyh in
order to, which appears to be used when the speaker thinks the purpose reading
might otherwise not be obvious to the listener as in (92).
Purpose clauses usually share a subject with the prior clause and thus
occur without an explicit subject, but if an overt subject occurs it is marked with
the preposition ka for; BENEFACTIVE, as in (91).
Such sequences of clauses (or verb phrases) are completely transparent and do not
involve any specific constructional properties other than the omission of the
subject expression. However, in some juxtapositions of verb phrases certain
conventionalized syntactic patterns have developed with semantics not fully
predictable from the sum of the parts. These are often termed serial verb
constructions, which differ primarily from other clausal coordination patterns
through their conventionalization of a specific through reference to a series of
events conceptualized in some way as constituting a single conceptual unit.
Essentially one of the verbal elements takes on what is an adpositional character,
and, as this characterization suggests, the deverbal adposition can no longer be
independently negated or marked for aspect.
As in many other languages in the area, there are in particular two
domains in Cham where one finds serial verb constructions: directionals and co-
verbs, i.e. markers for peripheral (non-core) arguments.
The directionals are commonly used motion verbs that have occurred so
frequently in a semantically secondary role in clausal concatenations that they
have come to have the semantic properties more characteristic of prepositions or
verbal particles. The two most common co-verbs are the two from the least
marked motion verbs naaw go > motion away from the center of action and
maay come > motion towards the center of action.
(98) mEE/ MKaam chO~/ pa~ tko~/ ca~n kO/ Na/ baN. (6.1.29)
Mom Kam scoop take up chop head make eat
Kams mother scooped it up, cut off its head, and made it into food.
(99) pO~h mkya truh trun tpa/ li/i. (6.1.133)
CLF ebony fall down straight basket
the ebony fruit fell down into her basket.
As with go and come these directionals co-exist with homophonous fully ver-
bal forms.
23
Example (101) involves both the literal meaning and the change of state
meaning in that this example literally means something like Hlok came to have
a lot of fish.
Note that m/ is only used with stative expressions when a change of state
occurs. Thus, in (102), which indicates not a change of state but simply the
existence of a state, m/ does not occur. However, in (103), which records a
change of state, m/ does occur. And, in (104), which indicates a caused change
of state using Na/ make as the causative marker, m/ also occurs.
The verb tuy followmay also mean according to and with, thus
functioning as a co-verb.
Other co-verbs most likely exist, but the line between a verb and a co-
verb is often difficult to pinpoint, especially where the metaphor involved is
(cross-culturally) transparent and the iconic sequence of events is still retained.
Thus, take in (107), for example, could be analysed as expressing
accompaniment.
Acknowledgements
24
The examples and a significant part of the analyses come from the work
of David and Doris Blood done roughly a quarter of a century ago.
References