2.1 Theoretical Assumptions
2.2.3 Atelic Reduplication
11 Another similar form with the 2nd person possessive suffix –hu (i-ra-rik’d-ən-hu) is categorized as manifesting progressive aspect (Adelaar 2000: 42).
12 Na is a partitive marker.
13 Tu means ‘during’; -ən is an undergoer-oriented marker; nein is a third person plural genitive marker.
14 -ian is a first person plural exclusive genitive marker.
(8) atelicity in Siraya
a. k‹m›a-kiĭm ‘‹actor.oriented.of.class.3.verb›reduplication-seek’ (=T3e)
The atelic reduplication is attributed to (9).
(9) atelicity for Siraya in S&C
Σ (string in the active workspace): verb root ϛ (initiator of Search): ϛi → a → ϛj
γ (target of Search): γi = first X; γj = first X δ (direction of Search): δi = R; δj = R β (beginning point of Search): βi = #; βj = #
in prose: For a verb root, add a precedence relation for the first X-slot to precede /a/ and another precedence relation for /a/ to precede the first X-slot.
In the previous section, when the initial timing slot is linked to a vowel, I have made the hiatus process responsible for either the missing fixed segment /a/ or the coalescence of the fixed segment with the preceding or following vowel. However, it seems that the presence of the fixed segmentism is masked whenever a verb root (for both the deverbal and atelic reduplication) is vowel-initial. Examples (8b-c) may nonetheless decrease this uncertainty because the timing slot for the fixed segmentism is speculated to be present.
(8) atelicity in Siraya
b. ä-äu-äux-aw-mau-kow ‘comitative-reduplication-follow-subjuctive.undergoer.oriented-first.person.singular.agent-second.person.singular.topic’ (=T3w)
c. [m]ä-äw-äwx-a ‘agent.oriented-reduplication-follow-subjective’ (=T3x)
I consider that the first [u] in (8b) is where the fixed segment /a/ sits and that the first [w] in (8c) is the same as well. That is to say, (8b-c) go through the intermediate representations (10a-b), respectively.
(10) inferred intermediate representations for (8b-c) after atelic reduplication a. äa-äux >>> (some unknown process(es)) >>> äu-äux
b. äa-äwx >>> (some unknown process(es)) >>> äw-äwx
The fixed segment /a/ perhaps is further modified by a process of labialization (/u/ in (10a) and /w/ in (10b)). The speculated process of labialization seems to bleed the application of the hiatus process. It is worth repeating that the nature of this discussion is speculative, but the speculation affords me an analysis of the atelic reduplication (and the deverbal reduplication as well) which indiscriminately copies the initial X-slot with the fixed segmentism.
Table 3 Siraya atelicity data (A00 = Adelaar 2000; A11 = Adelaar 2011)15 I. progressive aspect
(plain) reduplicated
gloss source
a. ba-varŭx-a16 ‘reduplication-revile-subjunctive’ A00: 43
b. da-diri-ən ‘reduplication-sow-undergoer.oriented’ A00: 43
c. ma-da-dĭlux ‘actor.oriented.of.class.1.verb-reduplication-lead’ A11: 113 d. pa-u-ka-kua-n
‘transitive-motion-reduplication-go-undergoer.oriented’
A00: 43 e. k‹m›a-kiĭm ‘‹actor.oriented.of.class.3.verb›reduplication-seek’ A11: 112 f. ka-kĭta-n(=umi)
h. m-i-ra-rkŭd ‘actor.oriented-locative-reduplication-fast’ A11: 113, A00: 45
15 Most examples do not have their unreduplicative counterparts.
16 Recall the b-v conjecture from table two.
17 The second vowel appears simply as a in Adelaar (2000: 47).
j. ta-tatul ‘reduplication-grow’ A00: 43 k. ni-ma-ta-täutäux ‘past.tense-agent.oriented-reduplication-teach’ A00: 43 l. mä-xa-xawey-a
‘actor.oriented.of.class.1.verb-reduplication-look.like-subjunctive.undergoer.oriented’
A11: 113 m
.
m-i-irung ‘actor.oriented-reduplication-sit’ A11: 113,
A00: 42
reduplicated gloss source
p. ma-da-dĭlux ‘agent.oriented-reduplication-lead’
v. tä-tätäx-ən-ato19 ‘reduplication-hew-undergoer.oriented-emphatic’
w. ä-äu-äux-aw-mau-kow
reduplicated gloss source
z. pa-i-da-rie20 ‘transitive-locative-reduplication-old’
A00: 44-aa. ma-la-liko ‘stative.intransitive.verb-reduplication-lie.down’ 45
bb. ää-la-likux ‘anticipating.sequence-reduplication-again’
18 This item displays the uncertainty for the word ‘to arrive’ (Adelaar 2011: 69-70): the former two forms suggest the verb root to be irua; the latter one says it is rua.
19
cc. mä-xa-xawey-a ‘agent.oriented-reduplication-look.like-subjunctive’
dd. mä-i-ä-äpoo-kamu ‘agent.oriented-locative-reduplication-need-second.person.plural.topic’
2.3 Tao
Spoken on Orchid Island (蘭嶼), the language classification of Tao (formerly called Yami) is:
Austronesian, Malayo-Polynesian, Philippine, Bashiic, Yami (Lewis 2009: 528). The island is administratively governed by the Lanyu Township (蘭嶼鄉) of Taitung County (台東縣), Taiwan.
I present patterns of reduplication from three constructions drawn from Rau and Dong (2005, 2006) and Rau et al. (2012).1 The comparative and collective reduplications are together discussed in 2.3.1; in 2.3.2, the plural reduplication is discussed. Some relevant explanations are needed for us to read the data sets:
a. ng stands for //, ’ for //, j for /dʒ/, and y for /j/;
b. stress is unpredictable and falls on one of the last two syllables—if the penultimate vowel is stressed, it is marked with an acute accent (ˊ); if the final vowel is stressed, there is no mark;
c. According to Rau and Dong (2005: 59, 2006: 80, 82), ay, aw, oy, and iw are underlying diphthongs, and the syllable template is (C)V(C). Instances like bayo [no gloss provided in the sources], apaw ‘light (weight)’, sáray ‘joyful, happy’ are then ambiguous whether the pertinent parts are diphthongs or not. In the following discussion, I will assume that
1
they are a combination of monophthong plus coda/onset glide. If they turn out to be diphthongs, because I assume that only vowels (glides excluded) are [+vocalic] (Nevins and Chitoran 2008: 1981), the head of a diphthong, which is vocalic, is distinguished from the dependent of a diphthong, which is non-vocalic. This means that all the following analyses which are based on monophthong assumption are maintained when the assumption is changed to diphthongs: what is constantly referred to is vowel (i.e.
[+vocalic]). In other words, the glide portion of a diphthong is not referred to because it bears [-vocalic];
d. The following five pairs of vowels are modified when the members of the pair are adjacent: eo >>> ow, ew; ao >>> ow; ai >>> ey; oa >>> wa; oo >>> ow (Rau and Dong 2005: 61, 2006: 85). For additional phonological details, please refer to Rau and Dong (2006: 79-83).