8ull. Nat'l Taiwan Nor. Univ. Vol 40 pp.261-294 1995
師大學報
第 40 期第 261-294頁民國 84年
The syntactic structure of W ulai and
Mayrinax Ataya
l:
a comparison
Li
llian M. Huang
Abstract
Atayal, one of the Formosan languages, consists of two major dialect grou阱，
namely, Squliq and C'uli'. Tsuchida (1980) proposes three linguistic criteria --phonological, lexical and morphological-- for c1assi可ing Squliq and C'uli' as distinct Atayal dialects. The present paper attempts to provide some syntactic evidence for such c1assification.
The Squliq dialect investigated here is spoken in Wul祉， Taipei Prefecture (hereafter Wu), and the C'uli' dialect under disωssion is that of Mayrin缸， Miaoli Prefecture (hereafter Ma). Below are some .structural characteristics of the two Atayal dialects:
(l) pronominal 可stems: Whi1e Wu has four sets of pronomina1 forms, i.e. Bound Nomi-native pronouns, Bound Genitive pronouns, Free 1ρcative pronouns and Free Nomi-native pronouns, Ma has only three, lacking the Locative pronouns. Also, the distri-bution and semanti臼 of these pronouns are different.
(2kasejprenominal markers: While Ma requires the presen臼 of c部，e markers 個 utter anc郎， Wu seldom requires these prenominal markers. Th帥， word order plays a ve可
important role in Wu.
(3hense-aspect systems: Whi1e Ma often uti1izes afflxes orjand reduplication to illustrate
di旺erent tensesjaspects, Wu uses affixes orjand auxi1iari郎， but not reduplication. (4)negative constructions: Both Ma and Wu have two negators in dec1arative sentences,
but in imperative constructions, Ma has two negators whi1e there is only one in Wu.
As shown above, Wulai Atayal and Mayrinax Atayal share some syntactic features and display some differences as well. Such syntactic simi1arities and di旺erences further support that they are two different dialects of the same language.
l.
Introduction
1Atayal is one of the more wide spread of the Formosan languages, ranging from Ilan county in the northeast to Taipei and Taoyuan counties in the north and southward through Hsinchu, Miaoli and Taichung counties to Nantou in the central portion of the Taiwan island. The language seems to consist of two major dialect groups, namely, Squliq and C'uli' (Ts'0Ie).2 Tsuchida (1980), as cited in Li (1985a:712), proposes three linguistic criteria for classi句ring Squliq and C' uli ' as distinct Atayal dialects: (1) the correspondence of Squliq /r/ to C'uli' /s/; (2) personal pronominal forms, i.e. Squliq /saku' / or
/ku' /
vs. C' uli' /cu/ or /ci/ '1'; (3) lexical differen臼s. 四le p肥sent paper attempts to ∞mp訂e the two dialects with respect to some of their structural
characteristi臼: their (1) personal pronominal systems; (2) case/prenominal markers; (3) tense/aspect systems; and (4) negative constructions.
The Squliq dialect investigated here is spoken in Wulai Hsiang, Taipei Prefecture; and the C' uli' dialect under discussion is that of Mayrin缸， spoken in Chinshui Village, Taian Hsiang, Miaoli Prefecture.3 Li (1980b, 1982c, 1983a) remarks that Mayrinax makes distinction between male and female speech. Since my major Mayr扭扭 infor
mant is Mr. Tang Ching-fa, all the Mayrinax data presented here are those of male's. Before we start to compare the two dialects, let 凶∞nsider their phonological
sys-tem. Both dialects have five vowels and nineteen consonants. The vowel symbols used here are/i, e, a, 0, u/:
烏來和汶水泰雅語語法比較 黃美金
268
High Mid Low Front 工 e Central a Back U OFigure la. Vowels in Wulai and Mayrinax Atayal
The consonant symbols are /p, t, k, q, " b, g, s, Z, x, h, c, m, n，嗯， 1, r, w, y/, where /' /
stands for a glottal stop, /b / for a voi臼d bilabial 剖臼tive， / g/ for a voi臼d vel訂 fri臼﹒
tive, and /ngt for a velar nasal. The inventory of Wulai and Mayrinax consonants is provided below: 5
Bi- Al- Alveo- Palatal Velar 白凡斗ar Glottal
labial veolar Palatal
Stop v1 P t k q Fricati ve v1 S x h Vd b Z q Affricate v1 C b姐sal rn n 11<主 Liquid l 主 Glide W y
Figure
Ib.
Consonants in Wulai and Mayrinax AtayalLike most, if not all, of the Formosan languages, Atayal, either Wulai or Mayrinax dialect, is basically a verb initial language. Below let us start to examine some
s甘uc-tural characteristics of the two dialects.
2.Personal pronominal systems
Wulai Atayal has four sets of personal pronominal forms, i.e. Bound Nominative pronouns, Bound Genitive pronou郎， Free 1ρcative pronouns and Free Neu甘al pro. nouns,6 as shown below (Huang 1993:17):
Person 拋岫岫N4岫岫恤α叩吼[l山L Number
Bound Free
Locative I Neutral
lS s副主u'/ku' rnaku'/mu/ku' knan kuzing/kun
2S su' su' sunan 工 su'
3S nya' hiyan hiya'
1PI 七a' ta' itan ita'
1PE sarru myan Sffilnan sarr江
2P SJlTlU mamu smunan SJlTlU
3P 吋泡， hg誼1 hga'
lS.BG十 2S.BN → misu'
Figure 2. Personal pronouns in Wulai Atayal
Unlike Wulai Atayal, Mayrinax has only three se臼 of pronouns, lacking the Locative set, as presented in the following figure (Huang MS.a):
Bound
Free
Person Nαninative Genitive Neutral
Number
lS ∞/ci' mu/mi' kuing
2S su' /si' su' isu'
3S nia' hiya'
1P 工 ta' /ti' ta' /ti' ita'
1PE carru nlam caml
2P CJlTlU mamu CJlTlU
3P nha' nha'
lS.BG+ 2S. 四J → misu'
Figure 3. Personal pronouns in Mayrinax Atayal
Considering the personal pronominal systems presented in the above two figures, we find that, like most other Formosan languages, the personal pronouns in Wulai and Mayrinax Atayal (a) have no gender distinction; i.e. no masculine-feminine contrast, (b)
烏來和汶水泰雅語語法比較黃美金 265
make a distinction between 'inc1usive' (speaker
+
addressee) and 'exculsive' (addressee not included) first person plural pronouns, and (c) have no Bound Nominative pro-nouns for third person, either singular or plural. As for indi討dualsets of pronouns, we notice that both dialects have Bound Nominative and Genitive pronouns and that they function similarly in the two dialects. Consider that following senten臼s with Bound Nominative pronouns first:(1) Wulai a. m -ihiy =主注 laqi' [AF-beat 三 lS.BN childl '工 beat a child' '工 am beating a child' 'd -114 q4.l ah' lcg m N B 二
w-1
胸
一一二 d mWM 一一 C t VAae )也扭扭 ←」 TL' bc. bhiy -un 且 laqi'
[beat-NAF 二 lS.BN childl
'The child will beat me' (2H但yrinax
a. t
<
um> uting二主立 cu' 'ulaqi' [beat<
AF>
beat 三 lS.BN Acc.Nrf childl,
I beat a child''工 am beating a child'
b. tuting-un 二五半 叫<u' 'ulaqi'
[beat -NAF 二 lS.BN Gen.Rf childl
'The child beat me'
c. ta -tuting-un 二旦旦 吋<u' 'ulaqi' [Red-beat -NAF= lS.BN Gen.Rf childl
'The child will beat me'
Bound Nominative pronouns in both dialects serve as grammatical subject, showing agreement with the verbal affIXes. That is, when the Bound Nominative pronoun mani-fests the agent participant, the verbal afflX is 旦﹒ (sometimes it can be -m-j φin Wulai
dialect， or 空空-/-旦旦-/主 in Mayrinax), as given in (1a) and (2a). This is Agent Focus (AF) construction, one of the common structures in Austronesian languages. When the Bound Nominative pronoun represents the patient participant, the verbal affIx is then
-旦旦 or --旦旦， as examples (lb-c) and (2b-c) show. This is Non-Agent Focus
(NAF)ωn-structíon.
As for Bound Genitive pronouns, they may either designate possessive relationship (e.g. [3a] and [4a]) or manifest the agent participant in NAF constructions (e忌 [3b] and [4b]): 門U B U-Qu
m--一一一 'FL , ild 司 A ﹒工 1-4 a 、n ﹒工 工 ch C 'r 弘 L aoiV4 9Tm NS 'Fι-l a-V4que --i3H a'hU 『 1 ， 14 恥 ι 可 db. bhiy -un = 1旦~ laq土，
[beat-NAF= lS .BG child]
'士 will beat the child'
(4 )I'但yrinax
a. 凡 hiya' ga' 'ulaqi' 三 l且
[Neu 3S.FN Top child 三 lS.BG] , He is my child'
b. ta -tuting-un 三五旦 ku' 'ulaqi' [Red-beat -NAF= lS.BG NαTI.Rf child]
'工 will beat the child'
However, the free pronouns in the two dialects iIlustrate di旺érent cases. First of aII, Mayrinax does not have 1ρcative pronouns, which often, though not always, manifest the patient participant in Wulai Atayal, as illustrted below:
(5) Wulai
a. m -ihiy 主且全旦 tali' [AF-beat lS.FL Tali'] 'Tali' beat me'
鳥來和汶水泰雅語語法比較黃美金 267
b. m < n > wah m -ita' knan tali' [AF< Past> cαne AF-see lS.FL Tali'] 'Tali' came to see me'
的 the English gloss designates, the Locative pronoun 坦哩 represents the partient of the action 'beating' j'seeing'. AIso, Locative pronouns may designate a pla臼 where
an action ends, as exemplified in the following sentences: (6) Wulai
a. k < m> ayan tali' squ' 主旦旦旦 krry缸
[say < AF> say Tali' Loc lS.FL every:day] 'Tali' talks to me every day'
b. n -byiq 七ali' qutux kopu' te k且全旦 hira' [Past-give Tali' one cup Loc lS.FL yesterday] 'Tali' gave a cup to me yesterday'
The Locative pronoun 扭且 in (6a七) represents a place where the action 'talking' j
'giving' ends. In either case, the Locative pronoun expresses a nonvolitional partici-pant, one that has less control over the events in question. This postulation is ∞n
firmed by the following pair of examples:
(7) Wulai
a. musa' =豆豆垃 m -aniq kira' [Asp
=
lS. BN AF-eat later] '1 am going to eat later'(it is my will; 1 want to) b. musa' 主且主且 m -ariiq kira'
[Asp lS.FL AF-eat later]
'1t will be my turn to eat later'
(whether 工 like it or not)
Comparing the English gloss in (7a) and (7b), we observe that the Locative pronoun expresses a participant lacking volition, and thus the English gloss 'it will be my turn to ...' in (7b).
they may serve as non-subject (or semantic patient, e.g. [8a-b]) or Topic (e.g. [8c]):
(8)Wulai
a. sayun m 一起間， kuzing/kun [Sayun AF-wash lS.FN 'Sayun is washing me'
叫
Z
nw
叫
-MW
1.4 na y 句 at' ← LR 凸可 -4 a dp 句 qJTt qd-N -m-J 血 z-qu u-1 工 、K 一『 L' CDue to the absence of 1ρcative pronouns, the Free pronouns in Mayrin缸 have to play dual roles, equal to 1ρcative and Free Neutral pronouns in Wulai Atayal. Consider the following sentences:
(9) 地yrinax
a. tuting-un si' kuing [beat -NAF= 2S.BG lS.FN] , You beat me'
b. tayhok ki' kuing 'i' yurr吐n
[arrive Loc lS.FN NomYm世n]
'Yumin arrived at my place'
c. binas-un ki' kuing ni' baicu' ku' imuwaag= nia' [sell-NAF Loc lS.FN Gen Baicu' Nαn.Rf house 三 3S.BG]
,
Baicu' sold his house to me'd. '土 kuing ga' tawqi' na' itaa工
[Neu lS.Fn Top chief Gen.Nrf Atayal]
'1 am the Atayal chief' (self introducing)
Like 1ρcative pronouns in Wulai atayal, Mayrinax Free pronouns may designate a pa-tient (e.g. [9a]), a location (e.g. [9a吋)， or serve as Topic (e.g. [9d]).
Above is a brief comparison between Wulai and Mayrinax personal pronouns. From the ∞mparison， we observe the simil訂ities and differen臼s between the personal
pro-鳥來和汶水泰雅語語法比較黃美金 269
nouns of the two dialects. Below we will examine the casejprenominal markers in the two dialects.
3.Casefprenominal markers
Mayrinax Atayal requires the presen臼 of case markers in utterances. A1so, Iike some other Formosan languages (e.g. Amis [cf. Huang 1995]), Mayrin缸 has two sets of case markers: one for common nouns and the other for proper nouns, kinship terms and Free pronouns. The figure below presents the ∞mplete sets of case markers in Mavinax: aymax: 7
Cases N叩1 Acc Dat Gen Ben Cαn6 Iρc 工ns
Nouns
cαmnon Rf ku' cu' na' i' na'
Nrf a' cku' cku' 叫主u' 吋心， cku' 叫心，
Proper , i' , i' '土， 'ni' 'ni' 'ki' 'ki' Figure 4. Case markers in Mayrinax Atayal
Now examine the following senten臼s with case markers for common nouns: (10) Mayrin缸
a. ukas 且一 'ulaqi' 二 rnu
[not:exist Nαn.Nrf child = lS .BG]
'工 have no child'
a' .ukas 主主，ulaqi' = rnu [not:exist Nαn.Rf child 二 lS.BGl
'My child is not in'
b. pa -' agaal= ci' cu' pila'
[Fut-take 二 lS.BN Acc.Nrf rnoney]
'工 want rnoney'
b' .pa -' agaal 三 ci' 主包 pila' ka' hani
Neu
[Fut-taJ但二 lS.BN Acc.Rf money Lin thisl
'1 wnat the money'
c. m < in > aiq 二 ci' cu' pila' 旦控 'ulaqi'
[AF< Past> gi ve 二 lS.BN Acc.Nrf money Dat.Rf childl
'1 gave money to the child'
d. ta -tuting-un ∞且:心(u' 'ulaqi'
=
nha'[Red-beat -NAF 二 lS.BN Gen.Nrf/Gen.Rf child
=
3P.BGl 'Their child will beat me'e. m -hahapuy 且區 'ulaqi' ' i ' yaya' [AF-cook Ben.Rf child Nαn motherl
'Mo仁her cooks for children'
f. kaa qilaap 主 paga'= mu [Neg sleep Loc bed
=
lS .BGl 'Don't sleep on my bed!' f' .kaa usa' 主 uray[Neg go Loc Wulai 1 'Don't go to Wulai! 使
f".tayhok 且包 tawqi' ' i ' yumin [arrive Loc.Rf chief Nαn Yur吐n]
'Yumin arrived at the chief's place'
g. bahiy-un 且且 hawku' ni' ba'unay ，土， watan
[beat-NAF 工ns.Nrf stick Gen Ba'unay Acc Wat誼1]
'Ba'unay beat Watan with a stick'
g' .bahiy-un 三 mu 且區， hawku' 'i' watan
[beat-NAF 二 lS.BG Ins.Rf stick Acc Watanl
'工 beat Watan with the stick'
As the English gloss in (10a-g) show, the noun following a 臼se marker ∞ntaining
19l
,(i.e. 坦， vs. 笠，也， vs. 里， and 也， vs. 且~) designates a more referential participant. According to Givon (1978:293), 'referentiality' is defined as follows:
It involves, roughly, the speaker' s intent to ' refer to' or ' mean' a nominal expression to have non-empty referen臼S ﹒i.e. to ' exist ' - within a p訂ticul訂閱iverse of dis∞urse
-鳥來和汶水泰雅語語法比較黃美金 271
. If a nominal is 'non-referential' or'generic' the speaker does not have a commit-ment to its existence within the relevant universe of discourse. Rather, in the latter, the speaker is engaged in discussing the genus or its properties, but does not commit himjherself to the existence of any specific individual member of that genus.
Following Givon's definition, we find that the notion of 'referentiality' can better ac-count for the differences between the two sets of case markers for common nouns in
Mayrin血﹒ 9
Next, examine sentences with case markers for proper nouns or personal pronouns: (l1 H但yrinax
a. m -itaal cku' 'ulaqi' 斗， watan [AF-see Acc.Rf child Nαn Watan] 'Watan is looking at the child'
b. i:r阻， ku' s <
un>
iuwal ' i' baicu' [who Nαn.Rf like< AF> like Acc Baiαl' ],
Who likes Bai∞， ?'c. m < in > aiq ci' cu' pila' 斗， wat血
[AF< Past> give = 18.BN Acc.Nrf money Dat Watan]
,
1 gave money to Watan'd. tuting-un ni' ba'unay 'i' watan [beat -NAF Gen Ba' unay N叩1 Watan] 'Watan was beaten by Ba'unay'
e. m 出hapuy 且主 kuing
,
i' watan(AF-cook Ben 18. Fn Nan Watan] , Watan cooks for me'
f. rna -tuting= c研吐且，恤'ay
[Rec-beat = 1PE. BN Cαn Ba' ay]
'工 fought with Ba'ay'
g. 地址 'an 主主 kuing
,
i' watan[exist Loc lS.FN Nαn Watan] 'Watan is in my place'
g'. m < in > uw品主主 watan 吐，恤'unay cu' 悅目，
[AF< Past> cαne Loc Watan Nt叩1 Ba'unay Part yesterday] 'Ba'unay came to Watan's place yesterday'
g".binas-un 三 mu 主立 watan ku' situ'ing
[sell-NAF= lS.BG Loc Watan Nαn.Rf clothes] 'I sold the clothes to Watan'
g'" .b < in > ainay= mu 主立 ba'unay ku' [buy< NAF.Past> buy = lS.BG Loc Ba'unay N叩1.Rf
situ'ing ka' hani [clothes Lin this]
'I bought the clothes frαn Ba' unay'
h. 三主 baicu' 'i' yaba' 二 mu
[Neu Baicu' Nt個1 father 二 lS.BG] 'Baicu' is my father'
Notice that in (l1h), the case marker 立三 is named 'Neutral case marker' and is only used for proper nouns. To name it 'Neutral 個se marker' is because its form is identi-cal to the Nominative case marker for proper nouns, i.e. 立~， and appears in the pr叫l ]
cate position, i.e. sentence initial position, while the sentence also has a Nominative case marker, either 立三，主 or 坦If each sentence can only have one focussed noun marked with the Nominative case marker in a focus language, then the marker 主 10
sentence initial position can not be a Nominative case marker. It may be used simply to mark a proper noun without showing any case relationship and so it is optional; thus the term 'Neutral case marker' is used.IO As expected, such a ∞nstruction is an equa-tiona
If
nominal sentence, as shown in (l1h). Also, compared with case markers for common nouns, the 臼se markers for proper nouns are relatively fewer, because proper nouns seem to designate more specific participants and thus this 田t does not have∞r-responding markers for nonspecific participants.
1n Wulai Atayal, on the other hand, thought nouns manifesting participants can be case marked, these prenominal m缸kers are seldom required in daily conversation,
espe-烏來和汶水泰雅語語法比較黃美金 273
cially not in the young generation' s speech. However, some of the 臼se marker百 some
times, though very rarely, do appear in senior speakers' utterance, especially in longer discourse. Consider the following examples:
(12}Wulai
a. nanu'
,
n臼1副主 qu' ngarux uzi[we工 1 special Nαn bear too]
'Wel工 I this bear was special too'
門， at 位 sa-n B-ht qA+ 」 o
md
開几
E aan squu '--na m-KS -t ra e-n rhIt a •L 可 mh 3[,
bb' . nyux hkngy-un 平豆， tali' sayun
[Asp seek -NAF Gen Tali' Sayur汀
'Tali' is looking for Sayun'
b" . s < m > ka' inlungan 二 nya' 19a'
,
nyt以 mtnak[split< AF> split heart 二 3S.BG LGA' Asp resemble inl ungan ngu' ngarux lmaga
[heart Gen bear I.MAGA]
'His heart split into two; one half resembled the bear's heart' c. m -kayan三 S品吐 主主 tali' zyaw qani
[AF-say 二 lPE.BN Cαn Tali' thing this]
'Tali' and 工 discussed on this matter' d. k < m > ayan 二 s品u' 主旦 sunan krryax
[say< AF> say
=
lS.BN Loc 2S.FL every:day]'工 talk to you every day'
d' .m < in > qbaq 三 S拭u' 主立 sunan hira' [AF< Past > learn二二 lS.BN Loc 2S.FL yesterday]
'工 learned frαn you yesterday'
d".nanu' s -panga' 工 nya' lru
,
ras -un 二 nya' 5qu'[50 NAF-back:carry二 3S.BG LRU carry-NAF 三 3S.BG Loc
ska' hlahuy [middle forestJ
d'lt .musa' = ku' m -tzyaw 乏主 ulay
[Asp = 13.BN AF-work Loc Wulai]
'工 will work in Wulai'
e. m -iru' 二 saku' 且ζenpit qani [AF-write= 13.BN 工ns pen this] ' I write with this pen'
The case markers in Wulai Atayal can be summarized 品 follows:
Cases b位rkers N叩l qu' Gen na' nqu' Cαn ki' Iρc te squ' sa Figure 5. Case markers in Wulai Atayal
工ns na'
Comparing Figures 4 and 5, we obseIVe that while the case markers in Mayrinax are obligatory and ve可 complex， the case markers in Wulai Atayal are optional and rela-tively simple. Further, since the case markers in Wulai Atayal rarely appear in daily conversation, word order thus plays an important role in the Wulai dialect.11
4.Tensefaspect systems
In this section, let 凶 examine the teQSejaspect systems in the two dialects. Both di-alects have two ways to illustrate different tensesjaspects; one is through focus affIXes,
and the other through aspectual affixes or j and auxiliaries. Consider the former way first. In the absence of temporaljaspectual affixes or modals, focus afflXes then play
dual roles, carrying focus and temporaljaspectual information simultaneously. Below let us examine some sentences from Wulai Atayal:
(13) Wulai
[AF-eat 二 lS.BN] '工 eat (often)' 'I ate' 'I arn eating' b. p -qaniq= ku' [AF-eat 三 lS.BN] 'I wi11 eat' (14) Wulai a. nig -旦旦三 mu qulih [eat-NAF= lS.BG fish] 'I will eat the fish' b. niq -:全旦二 mu qulih
[eat-NAF 三 lS.BG fish]
'工 ate the fish'
鳥來和汶水泰雅語語法比較黃美金 275
Notice that without cooccurring tensejaspect markers, (13a) with the Agent Focus afflx
旦 gets simple present, past and present progressive readings. ln other words，旦 refers to situations having either occuηed or actually taking place; i.e. 且 gives realis readings. 12
(13b), on the other hand, presents a future (or irrealis) reading because of the presence of Agent Focus affix Q:. (14a) with Non-Agent focus afflX﹒旦旦 gives a future reading,
and (14b) with 空空 a past reading.
Next, consider sentences with verbal afflXes and auxiliaries indicating different tenses j aspects:
(15) Wulai
a. m <主立 > aziy qutux ngaSill1 y品a' = mu
[AF< Past> buy one house father二 lS.BG]
'時 father bought a house'
b. n -aziy qutux ngasan yaba' 二 mun
[Past-buy one house father 二 lS.BG]
'My father bought a house'
c. suxan ga' m <主且 > aniq= ku' 19a' m -wah = ku'
1a [Part]
'Tαnorrow after 工 eat ，工 will cαne then'
d. b <主且 > hiy -an saku' = nya' hira' [beat< past> beat-NAF= lS .BN= 3S .BG yesterday] 'He beat me yesterday'
e. n -wah -an 三 m副凹， m -cisan u1ay 1a
[Past-cαne-NAF 二 lS.BG AF-p1ay Wu1ai Part]
'工 have visited Wu1ai a1ready'
(16) 趴llai
a. musa' 二 ku' m-ir阻， ta1i'
[Asp 二 lS.BN AF-wash Ta1i']
'I wi11 wash Ta1i"
b. nyux= ku' m -irna' ta1i'
[Asp 三 lS.BN AF-wash Ta工土，]
'I am washing Ta1i" ﹒工 -1 1414 aa ←」巾斗， ltby 戶恆 C 』 14 ma--吋 z mE 呵 附叫址 Es
wn
闖 一一三自 x-PLW Wd-bI C-rsL bc. cyux m -irna' ta1i'sa訊m
[Asp AF-wash Tali Sayun]
'Sayun is washing Ta1i' (there), c' . nYl心{ m -缸悶， ta1i' sayun
[Asp AF-wash Ta1i' Sayun]
'Sayun is washing Ta1i
,
(right here),
d. wan 三 ku' m -irna' ta1i'
[Asp三 lS.BN AF-wash Ta1i']
'I have washed Ta1i' (a1ready),
With the affix -垃- or !!-, sentences (15a-e) give Past readings, or to be exact, as shown by (15c), an anterior reading; i.e. the affix -坦- or 空﹒ designates Relative Anteriority in-stead of Absolute Past. Sentences (1缸-b) 的th auxiliaries 旦旦空旦旦，巨型 and 空空空
鳥來和汶水泰雅語語法比較黃美金 277
using 監控 and 旦旦 to designate on-going activities lies on (a) whether the agent is 1st/2nd person or 3rd person, i.e. 旦旦些 is used for 1st/2nd person and 空空空 for 3rd per-son; (b) whether the event is happening now or was happening in the past. In other words, when spatially or temporally close to the speaker，且也些 is used; otherwise，里墮 is
used.
Like Wulai Atayal, Mayrinax also uses focus affixes to illustrate different tenses/aspects, as exemplified below:
(17 ).1'但yrinax a. 且一anlq二 cu [AF-eat = lS. BN] '工 eat (often)' '工 ate' '工 am eating' b. pa -qaniq= cu [AF-eat 二 lS.BN] '工 wi11 eat' (18) 他yrin缸
a. niq -un 三 mu ku' iyok
[eat-NAF 二 lS.BG Nαn.Rf orange]
'工 ate the orange'
'1 am eating the orange'
-mM
-mkm叫一啊加以
拉出
While AF afflXes 哩- (and sometimes ﹒旦旦﹒) and ~- in Mayrinax also indicate realis and irrealis contra哎， the NAF affix ﹒旦旦 gives past and present progressive readings instead of future. To designate a future (or irreslis) activity, Mayrin阻 utilizes reduplication in NAF construction, as given below:
(19)!'但yrinax
a. r旦-旦iq-un = mu ku' iyok
[Red-eat-NAF= lS.BG N叩1. Rf orange]
'工 will eat the or部ge'
b. t.主弋t.uting-un nia' ku' 'ulaqi'
[Red-beat -NAF= 3G.BG Ncrn.Rf child ] 'He will beat the child'
Notice that the added syllable is formed by reduplicating the initial ∞nsonant and the
vowel 呈. As for the afflX -且 in Mayrin缸， it does not designate a p品t activity either,
although it is also one of the focus a扭扭s in the language, as shown below: 14 (20)!'但yrinax
a. hihipτ呈旦 ni' yaya' ku' 'ulaqi'
[kiss-NAF Gen mother Ncrn.Rf child] 'Mother is kissing the child'
b. s'wa' 弋呈旦=mi' limuy cu' balay
[like-NAF 二 lS.BG Limuy Part very]
'工 like Limuy very much'
Mayrin缸， Iike Wulai Atayal, also utilizes verbal affIXes and auxiliaries to indicate different tensesjaspects. For instance:
(21)!'但yrinax
a. m <主且 > ingilis ku' 'ulaqi' [AF< Past> cry N叩1.Rf child ]
'The child cried (before)'
b.m <扭> aniq二 ci' ∞ iyok
[AF< Past> eat 二 lS.BN Acc.Nrf orange] , 1 ate an orange'
c. m <扭> aniq ，土， y，品a' ga' m -a'usa' la [AF< Past> eat N咽1 father Top AF-go Part]
烏來和汶水泰雅語語法比較黃美金 279
d. k <主旦 > tal-an
=
mi' yumin ki' bai∞'[see< Past > see-N虹、二 lS.BG Yumin Cαn Baicu' J
'工 saw Yumin and Baicu"
e. b <拉> aiq -an 二 ci' ∞ qutux ∞ pinatiq
[give< Past> give-NAF 二 lS.BN Acc.Nrf one Part letter ni' lirnuy
[Gen Lirnuy J
'Lirnuy sent a letter to me'
(22) 地yrin缸
a. 主主主 'i' m -ingilis ku' 'ulaqi'
[Asp Lin AF-cru Nαn.Rf child J
'The child is crying'
b. 挂盟主旦旦 ci' m -aniq ∞ iyok
[Asp
=
lS.BN AF-eat Acc.Nrf orangeJ'工 am eating an orange'
c. han立且二 mi' tuting-un ku' 'ulaqi'
[Asp 二 lS.BG beat -NAF Nαn.Rf child J
'工 am beating the child'
Like the afflX -in- in Wulai Atayal, Mayrin缸 h部﹒垣- designating anterior activities 部
well. As for the auxiliaries indicating tensejaspect, unlike Wulai Atayal, Mayrinax h品
no auxiliaries to designate future or past acti吭ies. The dialect only h品區主 and 坦巫笠旦
for on-going activities. And the di旺erence between 啞巴 and 坦豆，an is similar to that
of 旦旦 and~ in Wulai Atayal.
To ∞nclude the present discussion, the following figure summarizes the tensejaspect systems of the two dialects:
AF constructions NAF constructions
realis irrealis realis irrealis
future future
simple progres- past present
S 工ve
Wulai m一 m- p- -an -un
一 ln- 一
ln-'''-an
wan nyux musa' wan nyux musa'
cyux cy山〈
b生3.yrln由主 m- m- pa一 -un -un
一 ln一 -ln一
Reduplica-"'-an tion
hani'an har吐，叩
kia' kia'
Figure 6. Tensejaspect systems of Wulai and Mayrinax Atayal
5.Negative constructions
There are three kinds of negative constructions in Wulai and Mayrinax Atayal; narnely, negative existential construction, negative declarative (excluding existential) construction and negative irnperatives. The Wulai dialect uses 啞巴t (or i旦旦!) in nega-tive existential construction while Mayrinax uses 些塾， as the following exarnples show:
(23) Wulai
a. ungat laqi' 三 nya'
[Neg child= 3S.BG] 'He has no child'
烏來和汶水泰雅語語法比較黃美金 281 ' y e n o m -JO 'v-n ae 114ns -lo--pm -e t-r a-qde qd-eh n-NT ﹒工-「 L' b (24) 已但yrinax
a. ukas a' 'ulaqi' = nia' [Neg Nom.Nrf child 三 3S.BG] 'He has no child'
'VAVA aee 『inn -工 OO Pmm 乎 4o rn N -S 'mi aNm s-qe a-eh 斌一仰叮 b
Both Wulai and Mayrinax Atayal have two negators in declarative sentences, ini' and
旦! (or 控型) in Wulai At可al and 垣立 and 扭扭笠 in Mayrinax. The negators in both Wulai and Mayrinax Atayal declarative sentences seem to resemble eaCh other. Struc-turally, only the rootjimperative form of the verbs may follow 坐立 in both dialects，部 illustrated in (25) and (26): d -14 '，中 ﹒工﹒工 t --41--s aa tT 仁 ydn
a--rp&S UU-l qdt ns' .1,
.qJl a--ea in-NT a ﹒工一『 L ， 14. 恥 ι Rd 弓，』b.á且，三 ku' qaniq qulih
[Neg 三 lS.BN eat fish]
'I didn't eat fish'
c. wan 二 ku' 主且， qniq qulih
[Asp= lS.BN Neg eat fish]
'工 still haven't eaten fish'
d. musa' 二 ku' 主且主， q臼liq qulih
[Asp 二 lS.BN Neg eat fish]
'工 will not eat fish'
e. n訊江三 ku' 主且， qaniq qulih
[Asp三 lS.BN Neg eat fish]
'工司叮lot eating fish'
uu cc' iid
」山區，恥
mt -los 'N qao--n
此 ds qJ ﹒工 -l 血可' qAtu SC -qd ﹒工 -m-hh -l 一『 L' a b. 些主二∞ 事niq [Neg= lS. BN eat '工 didn't eat' '1 am not eating'However, following 坦! (ín Wulai Atayal) and 盟主哩t (in Mayrinax) there can be ei-ther nouns or verbs with tense / aspect afflX白， as shown in (27) and (28):
(27) Wulai
a. yat tayan tali' [Neg Atayal Tali']
, Tali' is not Atayal' b. yat tali' nguray
[Neg Tali' Stl耳pid]
'lt's not Tali' (but sαneone else) who is Stl耳pid'
c. yat 二 ku' P -q.缸üq 年llih
[Neg三 lS.BN Fut-eat fish]
'1 will not eat fish'
d. yat = ku' m
<
in >血iq qulih [Neg= lS.BN AF< Past> eat fish] 'lhaven't eaten fish'(28) 1-'位yrinax
a. yakaat itaal 吐， baicu' [Neg Atayal N叩1 Bai凹，) , Baicu' is not Atayal'
b. yakaat ' i ' baicu' ku' mangotiq. ' i ' kuing [[Neg Neu Baiαl' Nαn.Rf stupid Neu lS.FN]
'工 t's not Baicu' who is stupid; it's me'
﹒呵泣 ne q
•
一凹， pm 削此 Ee u 門 b 二 c1d 二三 n 卡」 -14 a-1i a-qd ﹒工 此 -hw y-cI C d. :t.型盟主=∞ m<
in > aniq烏來和汶水泰雅語語法比較 黃美金 288
[Neg 二 18.BN AF< Past> eat] 'I still haven't eaten'
Functional旬， the two forms of declarative negation in both dialects seem to be distin-guished by how they relate to the event. The negator 垣主 in the two dialects focuses directly upon the event and denies only its performance. By contrast, ~ (in Wulai Atayal) or 坦坦型 (in Mayrinax) does not directly deny the performance of the event, and the point of its application may be diffused throught the proposition and the non-language context. 的
In imperative constructions, M可rin揖 has two negators 垣i and 包里， whereas there is only one，區!， in Wulai Atayal. Let us examine the two negators in Mayrin缸 first:
(29) 1'但yrinax
a. 捏立 nubuwag cu' 中.waw
[Negdri叫< Acc.Nrf wine]
'Don't drink wine!'
﹒工， l yaa a--U 恥恥 C OO --Lt Jdmv so uqt EqJn a-eo a-ND E[' b
c. 主豆豆 tuting-i 'i' ba'ay
[Neg beat -NAF Nαn Ba' ay] 'Don't beat Ba'ay!'
d. 坦主 ras -i ku' qusia'
[Neg bring-NAF N叩\.Rf water] 'Don't bring the water!' (30) 1'位yrinax
a. 1盟主 ku' m -nubuwag cu' quwaw
[Neg N叩\ AF-dri吐 Acc.Nrf wine] 'Don't drink wine!'
b. 1.盟i ku' m -usa' i' uray [Neg Nαn.Rf AF-go Loc Wu1ai]
'Don't go to Wulai!'
[Neg Nαn. Rf Rec-beat
'Don't fight against each other!'
Comparing the above sentences, we notice that following the negator 主豆豆 is the rootjimperative form of the verbs in Agent Focus construction (e.g. [29a-b]) or the verbs with Non-Agent Focus affi血 -i in Non-Agent Foωs construction (e.g. [29c-d]), and that a nominal clause (or phrase) follows the negator laxi, as pr臼en臼 of the Nomina-tive case marker k史 shows and also as confirmed by the following sentences: 16
(31) r-位yrinax
a. lax -i ku' xuil [give:up-NAF N叩l.RF dog]
'Don't want the dog!' ('Give up the dog!')
b. .1且 -1 ku' quwaw
[give:up-NAF N咽l.Rf wine]
'Don't want the wine!' ('Give up the wine!')
The two imperative negators of Mayrinax seem to merge into one in Wulai Atayal. Consider: (32) Wulai a. laxi usa' [Neg go] 'Don't go!'
扭扭
Notice that there is only one negator 且豆 in Wulai Atayal and that only the rootjimperative form of verbs can follow this negator. In other words, the imperative negation in Wulai seems to be a combination of the Mayrinax negator form 世 and
烏來和汶水泰雅語語法比較黃美金 285
6.Concluding remarks
As shown above, Wulai Atayal and Mayrinax Atayal share some syntactic features and display some differen臼S 品 well. Such syntactic similarties and di旺eren臼s， in addi-tion to the three criteria--phonological, lexical and morphological--as suggested by Tsuchida(1980), futher support that they are two different dialects of the same language.
Notes
1.The earlier version of this paper was presented at the Seventh International ConÎ er-ence on Austronesian Linguisti白， Leiden, August 22-27, 1994. Here, I would like to thank the National Science Council, Republic of China, for generous financial support of the work on Atayal which is represented in part here, and which aid was provided by through grants NSC 77-0301-HOO3-14, NSC 78-0301-HOO3-19, NSC
80-0301-H∞3-01 and NSC 83-0301-H∞13-017.
2.While the Squliq dialect has been studied by many linguists, such 部 Egerod (1965a, 1965b, 1966a, 1966b, 1969, 1974, 1980, 1993), Rau (1992) and Huang (1989, 1991, 1993), the C'uli' dialect does not get investigated as much, except by Li, who has been working on both dialects (1980a, 1980b, 1981, 1982a, 1982b, 1982c, 1983a, 1983b, 1985a, 1985b, 1986, 1994), Mei (1994) and Chang (1994).
3.The data on Wulai Atayal are mainly given by Pastor You Jin-chuan and those on Mayrinax Atayal are provided by Mr. Tang Ching-fa. I am grateful to them both for sharing generously their knowledge of the Atayal language and culture with me, and for their understanding and patien∞.
4.Notice that while /ng/ stands for velar nasal in this paper, /n _g/ indicates a sequen臼
of the alveolar nasal /n/ and the velar fricative /g/
5.The abbreviations and 可mbols used in this paper 訂e as follows: lS.BN: 1st Person Singular Bound Nominative
lS.BG: 1st Person Singular Bound Genitive 1 S.FL: 1st Person Singular Free Locative lS.FN: 1st Person Singular Free Neutral lPI : 1st Person Plural IncIusive
鳥來和汶水泰雅語語法比較 黃美金 287
1PE : 1st Person Plural Exclusive
Acc : Accusative AF : Agent Focus
Asp : Aspect Ben : Benefactive Com : Comitative Gen : Genitive Ins : Instrumental LIn : Linker
Loc : Locative NAF : Non-Agent Focus Neg : Negation Neu : Neutral
Nom : Nominative Nrf : Non-referential Part : Particle P1 : Plural
Rec : Reciprocal Red : Reduplicate
Rf : Referential Vd : Voiced V1 : Voiceless
三: The equal sign 二 indicates the following pronoun is a clitic
<
>:
Angle brackets<
>
enclose infIxes and their glosses6.The term 'Free Neutral pronouns' used to be 'Free Nominative pronouns', which is used in Huang (1989 & 1993) and which m可 be misleading because this set of pro-nouns can not only serve as grammatical subject which Bound Nominative propro-nouns designate, but also non-subject. Thus, the term 'Free Neutral pronoun' is adopted here instead.
7.Mei (1994) presents a case system of Mayrinax which is different from the one given here. Below 1 list his analysis (Mei 1994:6) for readers' reference:
Nαninative Accusative JJ' 自』 ev u-l qd 七 -l ﹒工 14n ∞侮 Dative cαrnnon nouns ku' a' cu' 叫<u' na' [-hurnan] cku' proper nouns , i' , i' ni' ni' ca' Allative/ Cαnitative ki' 工nstrumental 工。cative na' [+hurnan] i'
8.Mayrin且 does not have a Comitative case marker for common nouns; the conjunction
間， is used in conjoining two or more nouns (either proper or common nouns) and clauses, as sho.vn below:
(1) a. ngiaw 玉u' xuil [cat and dog] , a cat and a dog'
q]
心自間
叫一副削
叫
mm
b c. baicu' 玉立，往立 yumin[Baicu' and/Cαn Yur吐n]
, Baicu' and Yur吐n'
(2) rna -quwas cu' quwas na' itaal 王注:
[AF-sing Acc.Nrf song Gen.Nrf Atayal and si -papung= nia'
,
i' kuing[NAF-listen= 3S.BG N叩1 lS.FN]
,
He sang an aboriginal song to me'9.For further iIlustration of this point, see Huang MS.b.
10.Amis also has Neutral case markers. For details, see Huang 1995.
l 1.For a detaHed discussion on the role of word order in Wulai Atayal, please refer to Huang (1993:10-15).
12.For further discussion on the realis-irrealis contrast in some Formosan language, please see Zeitoun and Huang 1994.
13.In 旦哩!y， n- is the morphophonemic representation of m-in-.
14.The choi臼 ofNAF affix ﹒盟 or ﹒旦旦 actually depends on verbal semanti白， which 吶11
not be further discussed here.
15.For a detailed discussion on such IMMEDIATE - REMOTE contrast of negation, p!ease see Huang 1993:74-87.
烏來和汶水泰雅語語法比較黃美金 289
seems to. be the N o.n-Agent fo.cus verb. Co.mpare the fo.Ilo.wing sentences with m星空 and 噩噩笠 in the declarative sentences, and 坦i in the imperative sentence:
(1) a. ma -1缸 ci' cu' 羽uil
[AF-give:up= lS.BN Acc.Nrf dog]
'工 don't want a dog'
b. si -alax
=
mu ku' xuil[NAF-give:up- lS.BG Nαn.Rf dog]
'1 set free the dog'
(2) 1缸 一 i ku' xuil
[give:up-NAF Nom.Rf dog] 'Don't want the dog!'
References
Chang, Yo.ng-li. (1994) A preliminary study o.n the focus system in Mayrin缸 (in Chi-nese). Paper presented to. the Co.nference o.n Austronesian Languages in Taiwan, May 20-22, 1994, Academic Sinica, Taiwan, R.O.
c.
Egerod, Sφren. (1965a) Verb inflexio.n in Atayal. Lingua 15.251-282.
Egerod, Sφren. (1965b) An English-Atayal vo.cabulary. Acta Orientalia 19.203-220. Egerod, Sφren. (1966a) Wo.rd o.rder and wo.rd classes in Atayal. Language 42.346-369.
Egerod, Sφren. (1966b) A statement o.n Atayal pho.no.lo.gy. Artibus Asiae Supplemen-tum XXIII (Felicitatio.n Vo.lume for the 75th Birthday o.f Prof. G. H. Luce) 1. 120-130.
Egerod, Sφren. (1969) The o.rigin o.f headhunting--an Atayal text WÌth vo.cabulary.
Egerod, Sφren. (1974) The tattooing of women. Acta Orientalia (Hafn) 36.285-310. Egerod, Sφren. (1980) At，句'al-English Dictionary ( 二 Scandinavian Institute of Asian
Studies, Monograph Series No. 35). Curzon Press.
Egerod, Sφren. (1993) The main grammatical partic1es in Atayal. In Language--A Doorway between Human Culu1i問 Tribut，凹的 Dr. Otto Chr. Dahl on His Ninetieth
Birthdi哼， 184-2∞. Oslo: Novus For1ag.
Givon, Talmy. (1978) Definiteness and referentiality. In Universals 01 Human Lan-guage, Vol. 4: Synt缸， ed. by C. Ferguson and E. Moravcsik, 291-330. Stanford:
Stanford University Press.
Huang, Li1lian M. (1989) The pronominal system in Atayal. Studies in English Li tera-ture & Linguistics 15.115-133.
Huang, Li11ian M. (1991) The semantics of s- in Atayal. Studies in English Literature & Linguistics 17.37-50.
Huang, Li11ian M. (1993) A Stu砂 01Atayal Syntax. Taipei: The Crane Publishing Co. Huang, Lillian M. (1995) The case markers and pronominal system in Amis. The
Jour-nal o[ NatioJour-nal Chengchi University, Vo1.70.
Huang, Lillian M. MS.a. The pronominal system in Mayrin缸 Atayal. Hua嗯， Li11ia M. MS.b. 四le case marking system in Mayrin且 Atayal.
Li, Paul Jen-kuei. (1980a) The phonological rules of Atayal dic1ects. BIHP 51.2.3
49-405.
Li, Paul Jen-kuei. (1980b) Men's and women's speech in Mayrin缸. InPapers in Hon-or o[ Pro[essor L的 Yu次 'eng on her Seventieth Birthday, ed. by Li, et al. 9-17. Taipei: Wen Shin Publishing Co.
Li, Paul Jen-kuei. (1981) Reconstruction of Proto-Atayalic phonology. BIHP
鳥來和汶水泰雅語語法比較黃美金 29 可
Li, Paul Jen-kueÏ. (1982a) Atayalic final voiced stops. In Papers from the Third Intema-tional Con[erence on Austronesian Linguistics ( 三 Pacific Linguisti臼 C-75)， ed. by
Ha1im, Carrington and Wurm 2.171-185.
Li, Paul J凹-kuei. (1982b) Linguistic variations of different age groups in the Atayalic dialects. Tsing Hua Joumal o[ Chinese Sωdi自 14.167-19 1.
Li, Paul Jen-kuei. (1982c) Male and female forms of speech in the Ataya1ic group.
BIHP 53.1.265-304.
Li, Paul Jen-kuei. (1983a) Sex differentiation in speech and the problem of its origin.
Ta Lu Joumal 67.2.40-46.
Li, Paul Jen-kueÏ. (1983b) Types of lexical derivation of men's speech in Mayrinax.
BIHP 54.3.1-18.
Li, Paul Jen-kuei. (1985a) Linguistic criteria for classi句ring the Atayalic dialect groups.
BIHP 56.4.699-718.
Li, Paul Jen-kuei. (1985b) The position of Atayal in Austronesian family. In Austrone-sian Linguistics at the 15th Pacific Science Congress ( 二 Pacific Linguistics C-88), ed. by Pawley and Carrington, 257-280. Canberra: The Australian National University. Li, Paul Jen-kueÏ. (1986) Linguistic variations of di旺erent age groups in some
For-mosan languages. In FOCAL II: Papers 戶'om the Fourth lntemational Con[erence on Austronesian Linguistics (= Pacific Linguistics C-94), ed. by Paul Geraghty, et al.
33-50.
Mei, Kuang. (1994) Word order, case markers and theta agreement in Mayrinax (in Chinese). Paper presented to the Conference on Austronesian Languages in Tai-wan, May 20-22, 1994, Academia Sinica, Taiwan.
Rau, Der-Hwa Victoria. (1992) A Grammar o[ At，句'al. Taipei: The Crane Publishing Co.
Tsuchida, Shigera. (1980) Linguistic position of Sikikun and Mana呵ran--linguistic bases of subgrouping Sqolyeq and Ts' ole dialects in Atayal.
Zeitoun, Elizabeth and Lillian M. Huang. (1994) Toward a 句pology of tense, aspect and modality in the Formosan languages: a preliminary study. Paper presented to The Fourth International Symposium on Chinese Languages and Linguistics, July 18-20, 1994, Academia Sinica, Taiwan, R.O.
c.
烏來和汶水泰雅語語法比較黃美金 293
