| { |
| "paper_id": "Y03-1036", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T13:34:24.992850Z" |
| }, |
| "title": "On the Sentence Category Transfer of Action-effect Sentences in Chinese-English Machine Translation", |
| "authors": [ |
| { |
| "first": "Keliang", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Chinese Academy of Sciences No", |
| "location": { |
| "addrLine": "21 West Beisihuan Rd", |
| "postCode": "100080", |
| "settlement": "Beijing", |
| "country": "China" |
| } |
| }, |
| "email": "idiang99@sina.com" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "Of the 57 basic sentence categories (SC) defined in the HNC theory, action-effect SC is an important one with distinctive features. In the light of the HNC conceptual network, action-effect sentences in the Chinese language arise directly from causative verbs and compelling verbs, and indirectly from general acting verbs, i.e. via the use of the \" de(4)\" construction. In Chinese-English machine translation, action-effect sentences follow different SC and SF (sentence format) transfer rules. Therefore, different transfer frames should be adopted so as to ensure the generation of TL sentences with proper syntactico-semantic structures. Experiments show that the rules underlying the SC-SF transfer of action-effect sentences from Chinese to English can cover 90.3% of all the sentences in question.", |
| "pdf_parse": { |
| "paper_id": "Y03-1036", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "Of the 57 basic sentence categories (SC) defined in the HNC theory, action-effect SC is an important one with distinctive features. In the light of the HNC conceptual network, action-effect sentences in the Chinese language arise directly from causative verbs and compelling verbs, and indirectly from general acting verbs, i.e. via the use of the \" de(4)\" construction. In Chinese-English machine translation, action-effect sentences follow different SC and SF (sentence format) transfer rules. Therefore, different transfer frames should be adopted so as to ensure the generation of TL sentences with proper syntactico-semantic structures. Experiments show that the rules underlying the SC-SF transfer of action-effect sentences from Chinese to English can cover 90.3% of all the sentences in question.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "The HNC (short for Hierarchical Network of Concepts) theory (Huang, 1998) put forward the idea of action-effect chain (AEC), which is composed of the following six links: action (X), process (P), transfer (T), effect (Y), relation (R) and state (5). It advocates that AEC can be used to reflect the maximum commonness of the world and to depict the fundamental laws underlying the existence and development of everything in the universe (Huang, 1998:29) . Besides AEC, HNC expounds human thinking via the concept of judgment (D). As judgment is a kind of response made by the rational subjective world to the objective world and the emotional subjective world, judgment and AEC combine to describe the relationship between subjectivity and objectivity, between rationality and emotionality (Zhang Keliang, 2002) . AEC+judgment can be called \"the generalized action-effect chain\" (GAEC), which serves as the basis both for the classification of principal primitive concepts (which constitute one of the eight HNC semantic networks) and for the semantic categorization of sentences. The semantic category of sentences is referred to in the HNC theory as sentence category (SC). The HNC SC system includes 57 basic SCs and, in theory, 3192 composite SCs and over 10 millions of compound SCs (Huang, 2001) . Basic SCs are those that have no more than one eigen chunk (EK) which depicts just one of the seven links of GAEC . Composite SCs have only one EK, but their EK originates from the semantic blending of the EKs of basic SCs and depicts two or more of the links of GAEC. As to compound SCs, they come from the compounding of basic SCs and/or composite SCs, with two or more EKs that contain information related to corresponding links of GAEC.", |
| "cite_spans": [ |
| { |
| "start": 60, |
| "end": 73, |
| "text": "(Huang, 1998)", |
| "ref_id": "BIBREF5" |
| }, |
| { |
| "start": 437, |
| "end": 453, |
| "text": "(Huang, 1998:29)", |
| "ref_id": null |
| }, |
| { |
| "start": 790, |
| "end": 811, |
| "text": "(Zhang Keliang, 2002)", |
| "ref_id": "BIBREF15" |
| }, |
| { |
| "start": 1288, |
| "end": 1301, |
| "text": "(Huang, 2001)", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "The 57 basic SCs fall into 7 main types: action, effect, process, transfer, relation, state and judgment. This paper focuses on a special subtype of action SC, i.e. action-effect SC, and investigates the rules underlying the SC transfer of sentences of this sort from Chinese to English. By this study we expect to gain some insights into the relationship between the syntactico-semantic structure of a Chinese It should be pointed out that some basic SCs may not have any EK as far as a specific language is concerned. To Chinese, for example, there are four EK-free SCs (referential comparison-jugment jD021J/jD022J, restrictive comparison-judgment jDO1J/jD011J/jD012J, concise tendency jD2J, concise state SO4J) and a quasi EK-free SC (concise judgment jDOJ). sentence and that of its English translation and to facilitate the structural transfer of the HNC-based Chinese-English MT engine under construction.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Action-effect Sentences Action-effect sentences are sentences that belong to the semantic subtype of action SC. They are represented as follows: XYJ=A+XY+B+YC, YC=!31ErJ In the representation above, XYJ is the SC code for action-effect sentences, while A stands for agent chunk, XY for action-effect EK, B for object chunk, and YC for content chunk of effect. Moreover, YC is derived from the extension of a chunk (hence extended chunk or EC) and is essentially a sentence (called EC sentence and represented as ErJ) in the sentence format (SF) of !31. !31 is a code for elliptical SF that covers sentences omitting the first generalized object chunk (JK1 r. The omitted JK1 is in fact the previously adjoining B. In other words, B+YC is a semantically (and also syntactically to Chinese at least) complete sentence. Moreover, the EK of the EC sentence is also the ErK of the action-effect sentence. For example, (1c)9\u00b0( le) Weekend gales!! forcedly himll [#to changell his plans#]. Note: (i) sentence (*e) hereafter is the reference English translation of the Chinese sentence (*c); (ii) for readers' convenience, all the examples given in this paper are tagged with the following symbols: \"II\" (for chunks in independent sentences and EC sentences), \"^\" (for logical chunk markers), and \"[#...#]\" (for EC sentences); (iii) attached in the parentheses, if any, is the SC-SF expression of the sentence in question. In sentence (lc), there exists the following mapping: \" JiT j tg (weekend gales)\"-A, \"17,aft (forced)\"-XY(Ep), \"1t (him)\"-B, \"atlettato change his plans)\"-YC. And B+YC, i.e. \"it,E30 itt(him to change his plans)\" is an action-effect composite sentence XYO*22J(SC-SF expression: XYO*22J=A+XY0+B), whose EK \"E3t(change)\" plays the role of the Er in the action-effect sentence.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2", |
| "sec_num": null |
| }, |
| { |
| "text": "Action-effect SC is closely related to basic action SC whose representation is XJ=A+X+B, B=XB+YB+YC. When the B chunk in XJ is extended into a sentence (EC sentence), XJ becomes XYJ. Here are some of the distinctive features of action-effect sentences:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "(1) XYJ usually has two EKs: the former is called premise EK (or Ep), i.e. XY, while the latter is called result EK (or Er) which is in fact the EK of the embedded EC sentence ErJ. Since Ep and Er depict respectively the action link and the generalized effect link of AEC, XYJ is in nature a compound SC. Nevertheless, considering that XYJ covers a large part of any natural language, it is purposely defined as a basic SC.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "(2) The relationship between the two EKs in XYJ is not a global-local relationship (Eg-E1 relationship), but a cause-effect or sequential relationship (Ep-Er relationship). In inflectional languages like English, Ep functions as a finite verb or verbal cluster while Er functions as an infinite verb or verbal cluster. In sentence (1 e), for instance, its Ep is the finite verb \"force\" while its Er is the infinite verb \"to change\".", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "(3) As one of the HNC-defined SCs with an embedded EC sentence, XYJ uses the coupled Ep and Er for the complete depiction of the action link and the effect link of AEC. In other words, XYJ's YC is indispensable to the syntactico-semantic integrality of the sentence. For example, sentence (2c) and its translation sentence (2e) would neither stand on its own without their YC \" .:14 -r\" and \"go mad\" respectively.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "(2c) EPAJAVAng iii0i*VNIMOiW 1147#i 0 E Refer to Zhang Keliang (2001) for details about HNC sentence formats.", |
| "cite_spans": [ |
| { |
| "start": 55, |
| "end": 69, |
| "text": "Keliang (2001)", |
| "ref_id": "BIBREF14" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "\u2022 For a more detailed explanation of these and other symbols used in HNC tagging, please refer to Huang (2002a) .", |
| "cite_spans": [ |
| { |
| "start": 98, |
| "end": 111, |
| "text": "Huang (2002a)", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "(2e) The news that Beijing succeeded in bidding for the 2008 OlympiadII madell the whole city) [#go mad#]. (4) XYJ's YC must be an EC sentence in !31 SF. And the absent JK1 must be the previously adjoining B, which also serves as the object of XY. Take sentence (3c) for example.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "(3c) K1tli \u00a7 1 7.11 11AffillUrSEffillfil \u00b0( 3e) The mobsters) forced) the hostagesil [#to hand over!' their cell phones, pagers and any other communications devices#].", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "In this XYJ-type sentence, the YC \"-Se.ffia \u2022 sPf=a-iii,\"A. (to hand over their cell phones, pagers and any other communications devices)\" is itself an embedded sentence with T21J (physical transfer) as its SC and !31 as its SF. \"AIN (the hostages)\" functions both as the absent TA in the EC sentence T21J and as B, the object of XY \"3M (forced)\", in the SC sentence XYJ. (5) Despite the fact that XYJ is closely related to XJ, they differ from each other in terms of the A chunk: the focus of the A in XJ cannot take content chunk C whereas the A in XYJ can. The A containing C as its focus in XYJ often takes the form of sentence ecdysis (SE), which is a chunk in function but a degraded or transformed sentence in nature. Take the following sentences for example. (6c) it hOzfff1=11E. R*bi' , 4 4 11-tVZitilfAMIPA--3-)Et*Ire (6e) Local governments and enterprises take part in education, which!' can also drivelj colleges and universitiesil [#to further reform) their management systems#]. In sentence (4c) the focus of A, i.e. \"IT(conditions)\", is content-free, as is the case with XJ. In sentence (5c), however, A's focus \"It IlEti(the liberal attitude)\" is characteristic of content, which is not allowed in the case of XJ. As to sentence (6c), its A \"it...)3*(Local ... in education)\" takes the form of SE that never occurs in XJ.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Features of action-effect sentences", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "The generalized AEC grounds both the classification of principal primitive concepts and the semantic categorization of sentences (Huang, 2001 ). Since the SC of a sentence is predominantly determined by the core of its EK, and the core of an EK is invariably a verb, the syntactico-semantic classification of sentences is closely linked with the conceptual/notional classification of verbs by the HNC conceptual/semantic network-.", |
| "cite_spans": [ |
| { |
| "start": 129, |
| "end": 141, |
| "text": "(Huang, 2001", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "As a special subtype of action sentences, action-effect sentences strongly correlate with the 00-node concepts, i.e. basic action concepts. Verbs that express concepts subordinate to this node can all, in principle, be used in action-effect sentences. Under node 00 are the following four son nodes: node 003 (action that is invariably accompanied by its effect or result), node 008 (macro physical action), node 009 (micro chemical action) and node OOa (biological action); and node 003 has its own son nodes: node 0038 (causative action) and node 0039 (compelling action). Verbs expressing concepts subordinate to node 0038 and node 0039 are called causative verbs and compelling verbs respectively, examples of which are to be given in a later part. Verbs expressing concepts subordinate to node 008 (like \"I (bite)\", \"La(inhale)\", \"(drag)\", \"IT(beat)\"), node 009 (like \"Mburn)\", \" ,, melt)\", \"gr s' (dissolve)\") and node OOa (like \"it(eat by moth)\", \"Mt(digest)\", \"ft/6(excrete)\") are all called general acting verbs.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "We can thus draw some inferences about verbs commonly used in action-effect sentences. Firstly, causative verbs such as \"ft\" , , \"1:14\" , , \"tre4\" in Chinese and \"make\", \"get\", \"cause\", \"enable\" in English directly result in action-effect sentences, as shown in sentences (7c)/(7e) and (8c)/(8e). Secondly, compelling verbs such as \"a\" , \"a\" , \"wiz\" , \"siz\" , \"aLI\" , 3311\" , \" \u00a7$1. 1\" , \"LIft\" , \"far , \"icy\" in Chinese and \"force\", \"compel\", \"coerce\", \"drive\", \"impel\", \"press\", \"urge\" in English also result in action-effect sentences in a direct way, as shown in sentences (9c)/(9e) and (10c)/(10e).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "(9c) \u2022 -iliAllXit4;r1I-litilitlIC#1/11Iff#1 \u00b0( 9e) Lack of funds'' will eventually force!' himil [#to look foil' work#].", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "( 100 ffg111 222.11a11111171-VA* (10e) The circumstances(' impelled'' usil [#to take action at that time]. Thirdly, as far as the Chinese language is concerned, general acting verbs can engender action-effect sentences in an indirect way, i.e. via the use of \"f4(de)\" construction. The auxiliary word \" V4 (de)\" is used in such a way that two independent clauses are bridged to form a cause-effect relationship. The predicate verb and the auxiliary 14(de)\" combine to function as the EK (i.e. XY) of the sentence. Action-effect sentences of this sort are ubiquitous in Chinese, and are inclined to take ! 113 SF and ! 123 SF, two variant forms of regular SF. Take sentences (11c)/(1 1 e) and (12c)/(12e) for example.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "(11c) litTligf41011[ 0: T*1#] \u00b0( 11 e) The mosquitoes'' bitil me ++[so that] III could not fall asleep.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "(12c) /1\"(A-11101e1111-t411Viii*_ \u00b0(", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "12e) The little girlil was beaten black and blue'' by her stepmother. In sentence (11c), the auxiliary \"'(de)\" is used to connect two otherwise independent clauses: \"413Z-T-I R(The mosquitoes bit me)\" and \"Iv: Tvto could not fall asleep)\". The predicate verb \"I (bit)\" and the auxiliary \"Mder combine to serve as the XY of this XYJ sentence. Similarly, sentence (12c) is formed, via the use of \"M de)\", by bridging the following two independent clauses: \"ga l-pi.. TA (The stepmother beat the little girl)\" and \"/J\\tai<af2k, :,if(The little girl was black and blue)\", and its XY is composed of the predicate verb \"nibeat)\" and the auxiliary word \"'(de)\". In addition, sentence (12c) takes ! 123 as its SF which, as well as the !113 SF, is the usual case with action-effect sentences embedded with a \"de\" construction.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "Lastly, two distinct colloquial Chinese words, \"a\" and \"W-\", deserves much of our attention. They can be called vague verbs, verbs with an obscured denotative meaning for action. In fact these two frequently used words are characteristic of both causative verbs and general acting verbs. Conceptually, they belong to node 0038 and therefore are causative verbs. What distinguishes them from other causative verbs, however, is the fact that they are grammatically more like general acting verbs. For example, together with a word expressing concepts of generalized effect, they can form a compound word which expresses the blended concept of action and its effect or result, such as ttf (do...well)\"and \"J:(ruin, spoil)\". What's more, they can only be used in an action-effect sentence via the use of \"f4(de)\" construction, which is characteristic of general acting verbs. Take the following two sentences for example. SC transfer is referred to as the change of SC from one into another within a specific language, whereas interlingual SC transfer is used of the change of SC from one into another between two different languages. This paper focuses on interlingual SC transfer.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "Interlingual SC transfer, which plays a pivotal role in the translation engine of any HNC-based bilingual MT systems, can be divided into three types: bare transfer, obligatory transfer and optional transfer. Bare SC transfer applies to cases where no SC transfer occurs when a SL sentence is translated into its corresponding TL equivalent. Obligatory SC transfer applies to cases where SC transfer is a must to the successful translation of a SL sentence into a TL sentence. Optional SC transfer, somewhere between the two extremes, applies to cases where the SC of the SL sentence may or may not change in the process of translation. With respect to the second and the third types, SC transfer is further described as deterministic and non-deterministic. Deterministic SC transfer denotes the single-valued SC mapping relationship where the SC of a SL sentence is definitely transferred to the SC of its corresponding TL sentences. In contrast, non-deterministic SC transfer denotes the multi-valued SC mapping relationship where the SC of a SL sentence can be transferred to different SCs of its translations. For examples that illustrate the preceding terms, please refer to Mang Keliang (2002).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A conceptual analysis of verbs commonly used in action-effect sentences", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "Before proceeding to the actual analysis of the SC transfer of action-effect sentences from Chinese to English, we need to define a formal approach to describe SC transfer, whether intralingual or interlingual. Therefor, SC transfer frame (or TransFrame in short) is proposed, and the TransFrame of sentence J to sentence J' is defined as TransFrame(J', J)=[(E'J, EJ), (i', i), (K', K)]. In the above expression, ET, i and K represent respectively the SC, SF and semantic chunk (including supplementary chunk 1K) of the source sentence J, while E'J, i' and K' represent those of the target sentence J'. EJ and E'J, i and i', K and K' correspond to their respectively defined range of values. It should be noted that to a given SC transfer pair (E'J, EJ), chunk transfer pair (K', K) is repeatable. Alternatively, (K', K) can be rewritten as ((K1', K1), (K2', K2),...). And the maximum number of chunk transfer pairs is assumedly no more than four, something obvious according to the HNC theory.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "SC TransFrame", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "In the following sections, the SC transfer of action-effect sentences in the context of Chinese-English machine translation is investigated in terms of the verbs commonly used in this type of sentences.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "SC transfer of action-effect sentences from Chinese to English", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Causative verbs, forming a closed set in presumably any language, express 0038-node concepts and can definitely form action-effect sentences, as explicated in 2.2. Generally, when the Er of a Chinese action-effect sentence of this sort pertains to a generalized effect concept, the syntactico-semantic structure of the corresponding English translation is to be generated by way of bare SC transfer. In other words, the SC and the SF as well of the English sentence remain the same as those of the Chinese sentence. Take the following sentence for example.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Causative verbs", |
| "sec_num": "4.1" |
| }, |
| { |
| "text": "A XY B YC Sentence (15c) is an action-effect sentence engendered by the typical Chinese causative verb \"ft\"; its B and YC combine to form a complete effect sentence YJ. When being translated into English, it complies with bare SC transfer. Accordingly, its TransFrame is [(XYJ, XYJ), (!0, !0), ((A, A), (XY, XY), (B, B), (YC, YC))], and the default form is [(XYJ, XYJ), (!0, !0)]. This TransFrame determines the number and types of the main chunks as well as the order of their occurrence (hence the syntactico-semantic structure) of the corresponding English sentence, as is shown in the reference translation sentence (15e) below. XYJ B+YC=[#SO4J#] A XY B YC Like sentence (15c), sentence (16c) is also an action-effect sentence engendered by the causative verb \" ft\", and its B and YC combine to form a concise state sentence SO4J. According to the default TransFrame for sentence (16c), i.e. [(XYJ, XYJ) , 00, !OA, the syntactico-semantic structure of its corresponding English sentence can be well determined, as shown by the following reference translation sentence (16e):", |
| "cite_spans": [ |
| { |
| "start": 633, |
| "end": 650, |
| "text": "XYJ B+YC=[#SO4J#]", |
| "ref_id": null |
| }, |
| { |
| "start": 896, |
| "end": 907, |
| "text": "[(XYJ, XYJ)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "(16e) Exchange of unequal valued makesll the poor countriesli [#poorer#] . A XY B YC", |
| "cite_spans": [ |
| { |
| "start": 62, |
| "end": 72, |
| "text": "[#poorer#]", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "There are two points worthy of note. For one, words like \"ft\", \"-S\", \"rig\", \"m\" in Chinese, as well as words like \"make\", \"get\", \"put\" in English, are all multi-SC verbs, and that sentences containing any of these verbs may not necessarily be action-effect sentences. Sentences of different SCs have different transfer frames and should be translated into English using different SC transfer strategies. Listed below are some typical examples: For the other, optional SC transfer is applied instead when the Er of a Chinese action-effect sentence expresses the concepts subordinate to nodes 710 (mental reaction) and 713 (emotive expression). That is, the English translation may or may not inherit the SC of the source sentence, despite the fact that it does, more often than not, take a different one. Obviously, the SC transfer for this kind of action-effect sentences is non-deterministic, and they may choose for the SC of their corresponding English translations from among the following four options: XYJ, X2J, X22J and X22S*22J. Let's take sentence (21c).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "(21c) itn-fijgilii1011[#\u00b1ni---M#] (XYJ, YOr-[#!31X2OS*11J#]) (1) If XYJ is chosen as the SC of its English translation, then the TransFrame for sentence (21c) will be [(XYJ, XYJ), (!0, !0), ((A, A), (XY, XY), (B, B), (YC, YC))]. Guided by the TransFrame, the sentence could be translated into an action-effect sentence like (21e1).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "(21e1) His behavioril madell melt [Wed) very much surprisedn (XYJ, YC=[#!31X2OS*11J#]) In sentence (21e1), B and YC combine to form a complete reaction-state sentence that is semantically independent (yet grammatically dependent), i.e. B+YC=UX2OS*11M.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "(2) If X2J is chosen, the TransFrame will be [(X2J, XYJ), (!0, !0), ((XAC, A), (X2, XY+YC//X2S), (XBC, B))], by which sentence (21c) could be translated into an action-reaction sentence like (21e2).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "(21e2) His behavioril greatly surprised / startledll me. (X2J=XAC+X2+X2BC) Action-reaction sentences are meant in the HNC theory to describe the AEC from the end of action. They are widely seen in English but not in Chinese.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "(3) If X22J is picked, the TransFrame for sentence (21c) will be [(X22J, XYJ), (!0, !0), ((X2B, B), (X2, YC//X2S), (XAC, A))]. Under its guidance a passive reaction sentence like (21e3) could be generated.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "(21e3) Ill was greatly surprised!' by his behavior. (X22J=X2B+X22+XAC)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "In opposition to action-reaction sentences, passive reaction sentences are meant to describe the AEC from the end of effect with focus on the entity affected by the action (the patient). Additionally, a logical chunk marker is needed to introduce the doer of the action (the agent) represented by XAC. As far as the English language is concerned, the required chunk marker is exclusively the preposition \"by\". (4) If X22S*22J is chosen, the TransFrame for sentence (21c) will be [(X22S*22J, XYJ), !0), ((X2B, B) , (X2S, YC//X2S), (XAC, AM and the Chinese sentence could be put into a passive reaction-state sentence like (21e4). (21e4) III was very surprised at his behavior. (X22S*22J=X2B+X2S+XAC) Like passive reaction sentences, passive reaction-state sentences are also meant in the HNC theory to describe the AEC from the end of effect. But they differ from each other in that the latter focus on the state of the bearer of the action while the former concentrate on the bearing per se of the action. Another differentiating feature is that passive reaction-state sentences prefer to use \"at\" or \"with\" as the required logical chunk marker preceding XAC while passive reaction sentences invariably use \"by\".", |
| "cite_spans": [ |
| { |
| "start": 502, |
| "end": 511, |
| "text": "((X2B, B)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "(15c) It mig MinteElitiftniluwAvzoroti (XYJ, B+YC=[#YJ#])", |
| "sec_num": null |
| }, |
| { |
| "text": "When translated from Chinese into English, action-effect sentences engendered by compelling verbs follow a simple SC transfer strategy: unconditional bare SC transfer. Sentence (22c), for example, is an action-effect sentence derived from the Chinese compelling verb \" \u00a717131\".", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Compelling verbs", |
| "sec_num": "4.2" |
| }, |
| { |
| "text": "(22c) Ri[111Tft g2L 1 \u00b0 (XYJ, YC= 31T1J#]) Its B and YC combine to form a reception sentence T1J. This is an EC sentence and could otherwise be independent both syntactically and semantically in the Chinese language. The default TransFrame for sentence (22c) is [(XYJ, XYJ), !0)], by which the corresponding English translation like (22e) could be generated. ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Compelling verbs", |
| "sec_num": "4.2" |
| }, |
| { |
| "text": "As explored in 2.2, action-effect sentences engendered by general acting verbs contain a \"de\" construction and are prone to !113 SF and !123 SF. Generally, they take three different approaches to decide on the SC of the corresponding English translations. Firstly, if a given Chinese action-effect sentence is in basic format (!0 SF), the basic SC (XYJ, that is) will be transferred to a compound SC by putting the Chinese sentence embedded with the \"de\" construction into an English complex sentence with an adverbial clause of result. For example, sentence (25e) is an action-effect sentence in basic SF that is engendered by the general acting verb \"r5Z\". Its reference translation, sentence (25e), is a complex sentence composed of a main cause, whose SC is XJ, and a subordinate cause, whose SC is SOJ.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "General acting verbs", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "(25c) ta-T-11W4IIRII[Mr:: TWkitt] XYJ, YC=[#!31SN A XY B YC (25e) The mosquitoes'' bite me ++[so that] Ill could not fall!' asleep. (XJ++SOJ) A X B SB SO SC Secondly, if the Chinese action-effect sentence has a regular SF of !113, the basic SC (XYJ) will be transferred to a composite SC by blending the core of the Ep and the Er of XYJ into the core of the EK of the English translation, and the SF will be transferred from regular (!113) to basic (!0). In addition, the EK will be represented by an English verb or verbal phrase that expresses the blended concept of action and effect. Take the following sentence for example.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "General acting verbs", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "(26c) ii\\d`rrgASMRIIRRE&Tiirif4411[#4rES,7,0] 0113XYJ, B+YC=UNN) This Chinese sentence is an action-effect sentence in the regular SF of !113. According to the rule, it can be translated into English like sentence (26e). (26e) The small Greek army!' routed!! the Italians. (XYO*22J=X+XY0+B) Sentence (26e) is an action-effect composite sentence in basic SF. Its EK is represented by the word \"rout\" (meaning \"cause to flee or retreat in a disorderly way\"), a typical English verb that expresses the blended concept of action and effect.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "General acting verbs", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "Similarly, if the Chinese action-effect sentence has a regular SF of ! 123, the basic SC (XYJ) will be transferred to a composite SC and the regular SF(!123) to basic SF(!01). The EK of the English sentence generated is derived from the blending of the core of the Ep and the Er of the XYJ of the source sentence and represented by a verb or verbal phrase expressing the blended concept of action and effect. Sentence (27c), for example, is an action-effect sentence in the regular SF of ! 123 and can be properly translated into sentence (27e). The English translation is an action-effect composite sentence in basic form (!O1), its EK derived from the blending of the core of the Ep(Iff\") and the Er(\"Mir) of the XYJ of sentence (27c), and represented by the verbal phrase \"blow ... to smithereens\", a phrase that expresses both action ad its effect or result.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "General acting verbs", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "(27c)TG Rig giN-91-EmEit pawl -.1.3maiglerNallimi4iluteim", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "General acting verbs", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "(! 123XYJ, B+YC=[#Y9J#]) (27e) This wise strategy of the Fuhreril was blown to smithereensll by Pearl Harbor. 01XY9*22J=B+XY9+AA) Additionally, when it comes to action-effect sentences engendered by vague verbs like \"Wand \" fr-\", the preceding three approaches for the case of general acting verbs can be referred to for the SC and SF transfer, as shown in the following three pairs of Chinese sentences and their corresponding English translations. Based on the conceptual classification of the verbs that engender Chinese action-effect sentences, we have explored in this study the SC and SF transfer of this kind of sentences from Chinese to English and discovered some underlying rules. The rules have been applied to a collection of 145 Chinese action-effect sentences drawn from an aligned bilingual corpus of more than 8,000 Chinese-English sentence pairs. The findings show that these transfer rules can be successfully applied to 131 sentences, i.e. they can cover 90.3 percent of all the sentences in question.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "General acting verbs", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "Interlingual SC transfer is very complex, though. Apart from action-effect sentences, HNC-based Chinese-English MT engine research should also focus on concise state sentences, basic judgment sentences, comparative judgment sentences, bearing sentences, unilateral relation sentences and some subtypes of effect sentences, which all together cover 90 per cent and over of the real-text sentences that need to be transferred in terms of SC and SF when translated from Chinese into English. And these sentences will be dealt with in our future work.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "General acting verbs", |
| "sec_num": "4.3" |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "This paper is based on work supported by the National 973 Project under Grant No. 61998030506 and by the IACAS President's Fund for Selected Projects under Grant No. Sl3SJJ01, to which we would like to express our thanks. We're also indebted to Prof. Zhiwei Feng from the Institute of Applied Linguistics, Ministry of Education of China for his invaluable comments on earlier versions of this paper. Thanks also go to our colleagues for their contribution to the collection, alignment and tagging of the bilingual materials on which this study is based.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acknowledgements", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Chinese Sentence Patterns", |
| "authors": [ |
| { |
| "first": "Xiao", |
| "middle": [], |
| "last": "Fan", |
| "suffix": "" |
| } |
| ], |
| "year": 1998, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Fan, Xiao. 1998. Chinese Sentence Patterns. Shuhai Publishing House, Taiyuan.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Computer Processing of Natural Languages", |
| "authors": [ |
| { |
| "first": "Zhiwei", |
| "middle": [], |
| "last": "Feng", |
| "suffix": "" |
| } |
| ], |
| "year": 1996, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Feng, Zhiwei. 1996. Computer Processing of Natural Languages. Shanghai Foreign Language Education Press, Shanghai.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "The Case for Case", |
| "authors": [ |
| { |
| "first": "C", |
| "middle": [ |
| "J" |
| ], |
| "last": "Fillmore", |
| "suffix": "" |
| } |
| ], |
| "year": 1968, |
| "venue": "Universals in Linguistics Theory. Holt, Rinehart and Winston", |
| "volume": "", |
| "issue": "", |
| "pages": "1--88", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Fillmore, C.J. 1968. The Case for Case. In Emmon Bach. & Robert Harms eds.: Universals in Linguistics Theory. Holt, Rinehart and Winston, New York. 1-88", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Some Problems for Case Grammar", |
| "authors": [ |
| { |
| "first": "C", |
| "middle": [ |
| "J" |
| ], |
| "last": "Fillmore", |
| "suffix": "" |
| } |
| ], |
| "year": 1971, |
| "venue": "Georgetown University Roundtable on Languages and Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "35--56", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Fillmore, C.J. 1971. Some Problems for Case Grammar. In Georgetown University Roundtable on Languages and Linguistics 1971. Georgetown University Press, Washington D.C. 35-56", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "A Survey of the HNC Theory", |
| "authors": [ |
| { |
| "first": "Zengyang", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| } |
| ], |
| "year": 1997, |
| "venue": "Journal of Chinese Information Processing", |
| "volume": "", |
| "issue": "4", |
| "pages": "11--20", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Huang, Zengyang. 1997. A Survey of the HNC Theory. Journal of Chinese Information Processing, No.4:11-20.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "The HNC (Hierarchical Network of Concepts) Theory-A New Approach to Computer Understanding of Natural Languages", |
| "authors": [ |
| { |
| "first": "Zengyang", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| } |
| ], |
| "year": 1998, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Huang, Zengyang. 1998. The HNC (Hierarchical Network of Concepts) Theory-A New Approach to Computer Understanding of Natural Languages. Tsinghua University Press, Beijing.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "HNC Theory and the Understanding of Natural Language Sentences", |
| "authors": [ |
| { |
| "first": "Zengyang", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| } |
| ], |
| "year": 1999, |
| "venue": "China Fundamental Sciences", |
| "volume": "", |
| "issue": "2-4", |
| "pages": "83--88", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Huang, Zengyang. 1999. HNC Theory and the Understanding of Natural Language Sentences. China Fundamental Sciences, No.2-4:83-88.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "The Development and Future of HNC", |
| "authors": [ |
| { |
| "first": "Zengyang", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "HNC and Linguistic Research", |
| "volume": "", |
| "issue": "", |
| "pages": "53--68", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Huang, Zengyang. 2001. The Development and Future of HNC. In Quan Zhang & Guozheng Xiao eds.: HNC and Linguistic Research. Wuhan University of Technology Press, Wuhan. 53-68.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Temporary Regulations for HNC Tagging of Bilingual Corpora", |
| "authors": [ |
| { |
| "first": "Zengyang", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Huang, Zengyang. 2002a. Temporary Regulations for HNC Tagging of Bilingual Corpora. Unpublished document.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "HNC Conceptual Network", |
| "authors": [ |
| { |
| "first": "Zengyang", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Huang, Zengyang. 2002b. HNC Conceptual Network (updated version). http://www.hncnlp.com", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "English Verb Classes and Alternations", |
| "authors": [ |
| { |
| "first": "Beth", |
| "middle": [], |
| "last": "Levin", |
| "suffix": "" |
| } |
| ], |
| "year": 1993, |
| "venue": "CAS", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Levin, Beth. 1993. English Verb Classes and Alternations. The University of Chicago Press, Chicago. Li, Yaoyong. 2001. HNC-based Machine Translation. Postdoctoral report, Institute of Acoustics, CAS.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "A Study of HNC Sentence Category Knowledge", |
| "authors": [ |
| { |
| "first": "Chuanjiang", |
| "middle": [], |
| "last": "Miao", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "CAS", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Miao, Chuanjiang. 2001. A Study of HNC Sentence Category Knowledge. PhD thesis, Institute of Acoustics, CAS.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Collins COBUILD Grammar Patterns 1: Verbs. Shanghai Foreign Language Education Press", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Sinclair", |
| "suffix": "" |
| } |
| ], |
| "year": 2000, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Sinclair, J. 2000. Collins COBUILD Grammar Patterns 1: Verbs. Shanghai Foreign Language Education Press, Shanghai.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "Linguistic and Computational Techniques in Machine Translation System Design", |
| "authors": [ |
| { |
| "first": "Peter & Kieran", |
| "middle": [], |
| "last": "Whitelock", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Kilby", |
| "suffix": "" |
| } |
| ], |
| "year": 1995, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Whitelock, Peter & Kieran Kilby. 1995. Linguistic and Computational Techniques in Machine Translation System Design (2nd ed.). UCL Press Ltd, London.", |
| "links": null |
| }, |
| "BIBREF14": { |
| "ref_id": "b14", |
| "title": "The Representation of HNC Sentence Formats", |
| "authors": [ |
| { |
| "first": "Keliang", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "HNC and Linguistic Research", |
| "volume": "", |
| "issue": "", |
| "pages": "191--195", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Zhang, Keliang. 2001. The Representation of HNC Sentence Formats. In Quan Zhang & Guozheng Xiao eds.: HNC and Linguistic Research. Wuhan University of Technology Press, Wuhan. 191-195.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "The Sentence Category Transfer of Yes-no Sentences in Chinese-English Machine Translation", |
| "authors": [ |
| { |
| "first": "Keliang", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "Developments in Machine Translation Research-Proceedings of the 2002 National Symposium on Machine Translation", |
| "volume": "", |
| "issue": "", |
| "pages": "172--183", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Zhang, Keliang. 2002. The Sentence Category Transfer of Yes-no Sentences in Chinese-English Machine Translation. In Heyan Huang eds.: Developments in Machine Translation Research-Proceedings of the 2002 National Symposium on Machine Translation. Electronic Industry Press, Beijing. 172-183.", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Transition Processing in English-Chinese Machine Translation", |
| "authors": [ |
| { |
| "first": "Yanhong", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "Natural Language Understanding and Machine Translation", |
| "volume": "", |
| "issue": "", |
| "pages": "312--318", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Zhang, Yanhong. 2001. Transition Processing in English-Chinese Machine Translation. In Changning Huang & Pu Zhang eds.: Natural Language Understanding and Machine Translation. Tsinghua University Press, Beijing. 312-318", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "conditionsll makell onell lazy#]. (50 . '1Anditn Itillikr4lItSr glIW-02)A9Z#1 0 (5e) The liberal attitude of the yound causedil AIDSII [#to spread rampantly#]." |
| }, |
| "FIGREF1": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "13c) zyt*Iiilt (13e) Itil madell him) [#distressed#]." |
| }, |
| "FIGREF2": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "17c) t)K3r11)-ffiViMIIST-A-itlitA-T-\u00b0T ransFrame: [(XJ, XJ), (!0, !0), ((A, A), (X, X), (B, B))] (17e) Most western touristsli cannot usell chopsticks. (18c) Et-EgAllUtf**Wittti TransFrame: [(T3XY*31J, T3XY*31J), !0), ((TA, TA), (T3XY, T3XY), (B, B), (YC, YC))] (18e) The doctorli orderedll here [#to stayll in bed#]. ( 19c) [4111411-F-'NEVIIWAII[itiaVi TransFrame: [(T3XY*31J, T3XY*31J), (!31, !31), ((T3XY, T3XY), (B, B), (YC, YC))] (19e) Askil the next applicant [#to come in], [please]. (20c) 511T-iditift11[#.41.M#1 (X21XY*31J) TransFrame: [(X21XY*31J, X21XY*31J), (!0, !0), ((X2A, X2A), (X21XY, X21XY), (B, B), (YC, YC))] (20e) Father!! did not allow!! here [#to goll to the south#]." |
| }, |
| "FIGREF3": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "(22e) Well cannot forcell them!' [#to adopt!' the Marxist ideology#]. (XYJ, YC=[#!31T1J#]) Similar Chinese sentences and their reference English translations are given below: (23c) sl5R/gaamitili ESI/g&Z#1 \u00b0( 23e) The march of events!! compelledll himll Utto reconsider!' his decision#The allied forcesll would force!' the German armyll [#to surrender#]." |
| }, |
| "FIGREF4": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "28c) fil:5171:11filit'4Iltzitiluttat,A XYJ, B+YC=[#X2OS*11J#] (28e) Your jokell gotil heril [#very angry#]. (XYJ, B+YC=[#X20S*11J#] (29c) gia ...KIIIEffIliiV411[#tirElktff#] !113XYJ, B+YC=[#SO4J/] (29e) Those visitorsji messed upjj the sitting-room. (XS1*22J=A+XS1+B) (30c) A lf`Wil Mtlt , ErYg*Nfriffl.A144-MiYgT igiillAr411PirgiNfifT#1 \u00b0( !123XYJ, B+YC=[#X2OS*11J#]) (30e) Hitler apparently got tamed away by his own playacting and the constant assurances of Ribbentrop. (!01XX20*22J=B+XX20+AA) 5 Conclusion" |
| }, |
| "TABREF0": { |
| "text": "7e) Cruelty and injusticell often makell Pour bloodll boil#]. (8c) wa*Ilitf1411M-f-31-11[#tigr7ffniaAll1bff#] \u00b0( 8e) New techniques'' enablell surgeons'' [#to open and repaid the heart#].", |
| "html": null, |
| "content": "<table/>", |
| "type_str": "table", |
| "num": null |
| } |
| } |
| } |
| } |
