The Formation Mechanism - The Formation of Completive Use

Chapter 4 The Formation of Completive Use

4.3 The Formation Mechanism

We have learned that the completive qi-lai should be a derived use. In chapter 3, we know that qi-lai does not simply require a change point, a following duration, but also the degree feature or the scalar quantity feature. Thus, qi-lai is not compatible with achievements or with accomplishments, which take a quantized object/theme or resultative morpheme. All the incompatible VPs above are telic types.

Qi-lai excludes telic situation types and selects atelic situation types. Thus, how to

derive completive (telic) reading for a qi-lai sentence is our focus.

The last section introduced Van Hout’s and Kratzer’s theories on the

interaction between telicity and the characteristics of VPs. On the basis of their ideas, I will also provide an account for the derivation of the completive (telic) qi-lai

construction.

I take theme into consideration as well as object. The mechanism is the interaction between syntax and the characteristics of VPs. Since qi-lai selects atelic VPs taking bare plural object/theme or definite object/theme, a precisely quantized object/theme is out of our concern. Before deriving the mechanism, we first see the attachment of the bound morpheme qi-lai and the phonological representation of

qi-lai sentences.

Take the event of shou shu (put away books) as an illustration. Qi-lai, as a bound morpheme, cannot stand alone. It must be attached to the verb shou (put away) and form a new VP as the following structures show.

(62)

qi-lai VP

Shou (put away)

shu (book)

(63)

VP shou qi-lai

shu

The above structure is the syntactic representation. However, the actual phonetic or phonological performance is not exactly the same as its syntactic linear order.

indicates the tone of the syllable, and the bracketed tone is the sandhi ones.

Here we briefly review Liu’s claim on the phonological representation. Liu (2006) contrasts the following two sentences, and argues that qi-lai is a single bound morpheme and cannot form a prosodic foot. In the following sentences, the marker # represents the boundary of tone sandhi, the number attached behind each syllable

(64) a. gou[2] yao3# xiao[2] gou3# inent as a free morpheme like pao ‘run’, so it cannot form an independent ot.

‘ ther,

er says that dogs always bite exch other cruelly.’

Liu (2006:5)

ooks’ for ‘The dog bites the small dog.’

b. *gou[2] pao3# qi3 lai2# hen3 kuai4#

dog run qi lai very fast

According to Shih (1986), the prosodic foot is bisyllabic. Chinese speak would string unpaired syllables into binary feet. Base on Yip (1980), we know a prosodic foot has to possess a prominent syllable. Qi-lai, as a bound morpheme, is as prom

(65) a. 狗咬起狗來很狠

gou3# yao[2]-qi3# gou3 lai2# hen[2] hen3#

dog bite qi dog lai very cruelly

Based on his/her personal experience of watching dogs’ biting ezch o the external speak

b. *狗咬起來狗很狠

*gou[2] yao3# qi3-lai2# gou[2] hen[2] hen3#

dog bite qilai dog very cruelly

Hence, in order to prevent qi-lai from forming an independent foot in phonology, the object/theme has to occur between qi and lai if the object/theme does not serve as the topic. We continue to take the event of shou shu ‘ put away b

example. The phonological form after the attachment of qi-lai is as follows.

(66) ph

He put away qi book lai le

equire that the event culminate, but just states what it is like if the event c

oks’

. The word in the bottom of the structure is the phonological representation.

onological representation 他收起書來了 Ta shou qi shu lai le

Next, we examine the relation between the event and the object/theme. As section 4.1 states, the verb has to affect the object/theme in a way of changing th position or the appearance of the object/theme and the object/theme has to be a definite NP in order that the event has the potential to culminate. Culmination condition does not r

ulminates.

Following Kratzer’s ideas, we also utilize an operator [telic] to induce completive (telic) interpretation. VPs with culmination condition are said to be able be imposed with [telic], which operates over the whole V-qi-lai predicate and turns atelic VPs into telic ones. I continue to take the event of ‘putting away those bo for example and show how the completive (telic) construction is derived. The

following VP structure (67) is a simplified form of structure (63), which is the new VP after the combination of the bound morpheme qi-lai and the original VP

string

-In the structure, the external argument is taken to be irrelevant. The object/theme is

‘those books’, which is a definite NP. In addition, the event of ‘putting away’ affects the definite object/theme in the way of incrementally changing its position or

appearance, and the event has the potential to affect all the relevant parts of the referent of ‘those books’. That is, the event has a culmination condition. Therefore, the whole predicate is able to be imposed with [telic], which is illustrated as follows.

(68)

The definite object/theme NP is assumed to contain a nominal [acc] feature. I also suppose that [telic] takes index 1 as its feature and the nominal [acc] also bears index 1 as its feature. The following is the illustration.

Due to feature agreement, the index 1 of [telic] attracts the object/theme to the specifier position of [telic]. The index 1 of the object/theme NP is left in its base position within VP and the rest of the object/theme is moved up to the specifier position of [telic]. Then the uninterpretable nominal [acc] is deleted as shown in the tree below.

(70)

VP [telic]1

shou (put away)

-qi Na xie shu

(Those books) [acc]

1 -lai

Bare plurals or mass nouns, which are indefinite in syntax, can be interpreted either as specific (semantically definite) or as non-specific indefinite. If the superficial bare plurals or mass nouns are interpreted as specific, they are regarded as similar to definite NPs. Hence if the NP is specific, it can also enter the mechanism to move from the position within VP to the specifier position of [telic]. In the sentences below, the superficial bare plurals are interpreted to be specific and are said to be similar to definite NPs, so the specific NPs can be attracted from the position within VP to the position before the whole V-qi-lai predicate. Hence, the sentences below obtain telic reading as examples.

(71) a. 書收起來了 shu shou qi lai le books put away qi lai le

‘The books are put away.’

b. 稻草綁起來了 daotzao bang qi lai le straw tie qi lai le

‘ The straw are tied up.’

c. 箱子封起來了 xiang tze feng qi lai le boxes seal qi lai le

‘The boxes are sealed up.’

d. 門窗關起來了

men chuang guan qi lai le door window close qi lai le

‘The doors and windows are closed up.’

By the way, it is not necessary that VPs taking a definite or specific NP have to be imposed with [telic] operator and then force the specific NPs to move. The application of [telic] is optional. If we choose not to impose [telic], the definite or specific NP will not be attracted out and stay in the position within VP. The sentence will denote atelicity (inchoative interpretation) instead of telic (completive)

interpretation.

4.4 Summary

The culmination condition offered by definite object NPs along with suitable

measuring rods for the event makes the whole VP have the potential to be imposed with a head [telic]; then, [telic] attracts the definite object/theme to its specifier position through feature agreement between the [telic] and the object/theme NP.

Chapter 5 Characteristics of The Mechanism and qi-lai Sentences

This chapter is a comparison between the mechanisms of Van Hout and Ktrazer, and that of Mandarin completive qi-lai. Furthermore, we will briefly discuss the function of sentence-final ‘le.’

5.1 Differences between the mechanisms

Van Hout’s and Kratzer’s mechanisms allow achievements and

accomplishments while in Mandarin Chinese qi-lai sentences, achivements and accomplishments are first precluded by the requirements of qi-lai.

In the data discussed in the mechanisms of Van Hout and Ktrazer, an

accomplishment VP taking a quantized indefinite object can be imposed with a [telic]

operator to induce telic reading. Yet in Chinese, a quantized indefinite object NP never enters the mechanism. Actually, a quantized indefinite object NP, which might provide an endpoint to the event, is not permitted to appear either in inchoative (atelic) construction or in completive (telic) construction. This is due to the requirements of

qi-lai that preclude telic event types. Thus, a precisely quantized object NP which

provides an endpoint to the event never shows up in qi-lai sentences.

However, if a definite determiner is inserted in front of the quantized indefinite NP, the qi-lai sentences containing it will become acceptable as (72) and (73) demonstrate.

(72) a.他封起那五個箱子來了

ta feng qi na wu ge xiangtze lai le he seal qi that five CL boxes lai le

‘He begins to seal those five boxes.’

b.他收起那八本書來了

ta shou qi na ba ben shu lai le he put away that eight CL book lai le

‘He begins to put away those eight books.’

(73)

a. 那五個箱子封起來了

na wo ge xiangtze feng qi lai le those five CL boxes seal qi lai le

‘Those five boxes are sealed.’

b. 那八本書收起來了 na ba ben shu shou qi lai le

those eight CL books put away qi lai le

‘Those eight books are put away.’

This does not violate the analyses so far. The VPs in (72) and (73) take an object NP with a precise number. So it seems that such VPs cannot appear in qi-lai sentences. Yet, the object/theme NPs preceded by a definite determiner are actually definite NPs. According to Lin (2006), a definite object brings out atelic interpretation for the event. Hence, this does not depart from the analysis in chapter 3 that qi-lai selects atelic VPs. The grammaticality of the sentences above is thus well explained.

Another point is that in Kratzer’s and Van Hout’s mechanisms, the amount of the referent of the direct object can be changed whereas in the mechanism of

Mandarin Chinese qi-lai completive construction, the substance or amount of the

referent of the object/theme cannot decrease or increase from null. Besides, in the mechanism of completive qi-lai constructions, the measuring rod for the culmination condition of an event is not necessarily the part structure of the referent of the object itself. The process of changing the position or appearance of the object might be a more suitable measuring rod. The process should be definite so that the event has a culmination condition. Hence, the culmination condition is not only determined by the definiteness of the object/theme NP, but also by the definiteness of the changing process. The definiteness of the measuring rod (the changing process) can be inferred from context or some cognitive factors. This needs more study in the future.

To sum up, the event does not change the amount of the referent of the object/theme, but changes the position or appearance of the referent. The changing process must be definite. That is, in the event structure of a verb taking a definite object, there should be a potential culmination point¹⁰, which can be inferred from the

“V+O” VP. An event must contain the above three factors so that it can form a completive construction.

5.2 The role of sentence-final ‘le.’

Some say that sentence-final le possesses the meaning of inchoativity. And it seems that all qi-lai sentences require sentence-final le. This gives rise to a question about whether the inchoative meaning comes from qi-lai or from sentence-final le.

However, in previous literatures, we see that some qi-lai sentences do not require

10 Take the event tui zhe bu che “push the cart” for example. When someone pushes the cart, the amount or quality of the cart is intact. The position or appearance of the cart is indeed changed. The VP still cannot form a completive qi-lai construction. This is because we do not know when the event culminates. From the verb and the object alone, we cannot infer a potential endpoint of the event of

“pushing the cart.” The process of changing the position or appearance of the object is not definite.

Hence, the following sentence is unacceptable.

i. *這部車推起來了

*zhe bu che tui qi lai le this CL cart push qi lai le

sentential le¹¹. Examples like (20b) and (25a) are repeated below.

We also find other acceptable qi-lai sentences without sentential ‘le.’

(76)a. 昨天半夜，隔壁的小嬰兒突然哭起來，把我們都驚醒了

11 In some qi-lai sentences, sentential le is obligatory.

i. 他笑起來了

Actually, some other simple sentences also require sentential le.

ii. 他笑了

The obligatory appearance of sentential le involves many respects and other sentence patterns. The issue is unique to the use of qi-lai. We here focus on the uses of qi-lai and cannot explain the issue in a limited space. This is left to future study.

tzuotian ban yie, gebi de xiao yinger turan ku qi lai, ba women dou jing xing le.

yesterday mid- night, nextdoor de little baby suddenly cry qi lai, ba we all frighten awake le

‘Yesterday at midnight, the little baby nextdoor suddenly cried, and we were all awakened and frightened.

b. 喜愛音樂的張三時常即興地唱起歌來

xiai yinyue de Zhangsan shichang jixingdi chang qi ge lai like music de Zhangsan often impromptu sing qi song lai

‘Zhangsan who likes music often sing songs impromptu’

The sentences above do not contain sentential le, but they are still grammatically correct and have inchoative interpretation. Therefore, sentential le does not denote inchoativity but it possesses other functions.

In completive (telic) use, sentential le does not necessarily occur, either.

(77) a. 錢趕緊藏起來，不要被別人發現

Qian ganjin tsang qi-lai, buyao bei bieren faxian Money hurry hide qi-lai, don’t bei others find

‘Hurry, hide the money. Don’t let others see that.’

b. 張三把書收起來之後便出門了

Zhangsan BA shu shou qi-lai zhehou bian chu men le Zhangsan BA book collect qi-lai later then out door le.

‘After Zhangsan put away the books, he went out.’

Reexamining sentences with sentential le in (48)-(53) repeated in the

following, we find that in these sentences the focused event or state continues to the present time.

(48) a. 他收起書來了 Ta shou qi shu lai le He put away book lai le

‘He begins to put away the books.’

b 書收起來了 shu shou qi lai le

book put away qi lai le

‘ The books had been put away.

(49) a. 他綁起稻草來了 Ta bang qi daotzao lai le He tie qi straw lai le ‘He begins to tie the straw.’

b. 稻草綁起來了 daotzao bang qi lai le straw tie qi lai le

‘ The straw are tied up.’

(50)a 他疊起積木來了

Ta die qi jimu lai le He pile qi building block lai le

‘He begins to pile up the building block.’

b 積木疊起來了

jimu die qi lai le building block pile up qi lai le

‘The building blocks had been piled up.’

(51)a. 他捆起書來了 Ta kuen qi shu lai le He tie qi book lai le ‘He begins to tie the books.’

b. 書捆起來了 shu kuen qi lai le book tie qi lai le

‘The books have been tied.’

(52) a. 他封起箱子來了 Ta feng qi xiangtzi lai le he seal qi box lai le

‘He begins sealing boxes.’

b. 箱子封起來了 xiangtzi feng qi lai le box seal qi lai le

‘ The box has been sealed.’ (‘The boxes have been sealed.’) (53)a. 他藏起珠寶來了

Ta tsang qi zhubao lai le He hide qi jewelry lai le

‘He begins to hide the jewelry.’

b. 珠寶藏起來了 zhubao tsang qi lai le jewelry hide qi lai le

‘The jewelry had been hidden.’

In sentences (a), inchoative (atelic) use, the intuition is that the dynamic event lasts up to the time of utterance; in sentences (b), completive (telic) use, the result state after the endpoint of the event lasts up to the speech time. On the contrary, sentences without le as in sentences (74) to (76) do not necessarily encode the last-to-present meaning. In sentence ( 76a), the topic time of the event of baby’ crying is last night.

The event of the baby’crying does not necessarily hold at speech time. Sentence ( 76b) encodes habitual reading; the time of the singing event does not necessarily hold at speech time. Sentence ( 77a) is an imperative sentence, which has irrealis

interpretation, so the event of hiding money never holds at speech time. As for sentence (77b), it is a past-tense sentence. The event of putting away the book happened before another event. The result state of putting away the book does not necessarily hold at the present time.

Therefore, I think that sentential le provides a function associated with speech time. The function of le makes the focused event or state hold at speech time. This can be accounted for by the analysis of Lin (2005). Lin gives a proposal for the semantics of le as follows:

(78) [[le]]= λP<i, t>λ^Topλt⁰∃ t [P(t) ∧ Istage (t, P)⊆t Top∧ t Top < t0∧ tana

Rstate (t, P)]

⊆

The formula means that the inner stage is included in topic time. The topic time for result state is a time variable which is like an anaphor and must be bound by an overt topic time. If there is no overt temporal adverbial, the topic time for the inner stage is existentially closed, and the time variable is bound by the speech time. Accordingly, le denotes perfectivity in relation to the inner stage, but imperfectivity regarding the result state. As for the result state of the atelic event types, activity and states, Lin

argues that they have a special kind of result state, named empty result state. The empty result state exists in all times after an event happens. The formula can be applied to all kinds of events. Instead of supporting the non-unifying claim that le represents past tense when occurring in activities and denotes inchoativity when combining with stative verbs, achivement verbs or accomplishment verbs, Lin’s analysis unifies the meaning of le in all contexts.

Following Lin’s proposal, we can provide an explanation for the role of

sentence-final le in qi-lai sentences. In inchoative (atelic) sentences (a) in (49) to (54), there is no overt topic time. Since their event type is atelic, they have empty result state, which overlaps with the speech time and includes the speech time according to

在文檔中漢語"起來"的用法 (頁 60-0)