The Path of Grammaticalization for Coordinate Conjunctions and its Universality

From the study of polysemy of coordinate conjunctions, it can be found that there are grammaticalization processes from coordination to transition and from coordination to succession. It is a general evolutionary process in connective structures to go from non-specific semantic relationship to specific semantic relationship. This was caused by the cognitive tendency of approximately substituting typical functions for atypical functions.

conjunctions om 3 in Li language (Zhuang-Dong group of Sino-Tibetan), mà in Vietnamese (Mon-Khmer group in the Southeast Asia), et in French, und in German and да in Russian, all signify coordination as well as transition. The conjunctions serta, dan in Indonesian (Austronesian family), pak in Czech (Slavic group of Indo-European family), and и in Russian indicate coordination as well as succession. Given such apparent universality in the polysemy of these conjunctions, a question to ask is in what order do changes occur?
In theory, conjunctions have the following sources: ◎internal changes (changes in relevant meanings without changing ◎of the sources of motivation sought the parts of speech) within the relevant domains (changed from other relevant meanings) ◎external changes (changes in relevant meaning accomapanied by change of parts of speech) ◎of the sources of motivation sought without the relevant domains (changed from irrelevant meanings) ① It is argued here that the path of grammaticalization proceeds in order from coordination > transition, and from coordination > succession, and not the reverse, nor any other path. The sequence of change can be summed up as coordination>succession in internal changes: a generalization that the semantic relations denoted move from a general relationship, then change to a special one with a joined structure. This process of change results from the fact that people tend to cognize the non-prototypical function with the aid of the prototypical function.
The motivation resources outside of the relevant domains sometimes are regarded as a helpful means to see clearly the sequence of internal changes. The external changes involve those in correlative sense and parts of speech, with an increase of complexity ② beyond the scope of this paper. Beyond conjunctions, the verbal markers of coordination, e.g. relative adverbs, auxiliaries, and recurrent forms of correlation (e.g.一边(yì biān)…… 一边(yì biān) (meaning "on the one hand…and on the other"), as well as paired linking forms (e.g."既(jì )……又(yòu)…(meaning "both…and…") are not discussed either.
1. Analyses in form 1.1. The tendency to change from word-coordinate conjunctions to sentence conjunctions Conjunctions are classified into two kinds: word-linking and clause-linking, depending on where the joining takes place: within a phrase or between clauses of a complex sentence (or a compound sentence). For coordinate conjunctions that have relatively broader ① For example, words signifying the category of littleness have changed into transitional conjunctions. See " §2. Semantic Analyses". ② For example, the conjunctionhood of "乃(nǎi)", "又(yòu)" resulted from relative adverbs. syntactic applicability, there is a tendency to change from word-linking to sentence-linking, as evidenced below.
1.1.1. The semantic relations between words Polysemic relation is the epitome of the semantic relation between words, where the word-coordinate conjunctions are the oldest, the most essential of all conjunctions. Most of the languages which have conjunctions have word-coordinate ones. Chen Mengjia (1988) said too: "Most of the conjunctions in Oracle inscriptions in the Yin Ruins linked noun(s) with noun(s), whereas only a few linked two clauses." Even the most conjunction-lacking language, e.g. Tu language (a language of Altaic-Mongolian group) "only have one word-coordinate conjunction: da (and)" (Junast,1981[1]). Lhoba (Bengni-Bokr) (a language of Tibeto-Burman group, Sino-Tibetan), which has only a few conjunctions, has not only word-linking conjunctions but also fine differentiation between two kinds of conjunctions: one joining the substantivals and another for the adjectivals. The languages which borrowed conjunctions extensively from other languages have sometimes only word-coordinate conjunctions of their own, e.g. Cun (a language of Zhuang-Dong group, Sino-Tibetan family) has only one conjunction: the coordinate conjunction nam 1 . Compare the counterparts in other language of Zhuang-Dong group: Lingao (Tai group) hem 1 Bouyei in Guiyang (Zhuang-Dai branch) tiam 1 In addition to the word-coordinate conjunctions in Cun language, there are other common conjunctions which were borrowed obviously from Chinese, for example, hk 1 (or), huan 4 si 5 (or), za 5 si 5 (or), zi 5 ku 5 (if), ti 2 zen 4 (although), dan 3 si 5 (but), zin 2 vi 3 (because), s 5 zi 5 (therefore), and roughly so in Yanghuang and other languages (Bo Wenze 1997). Botibunia, a Russian comparative linguist in 19th century, on the basis of a large corpus, pointed out that zero conjunction appeared earlier than conjunctions themselves. Speaking of coordinate and subordinate linkage, Botibunia"s contemporary (and compatriot) Kozuhi believed that the appearance of the former was earlier than the latter. In China, Hu Menghao(1980) agreed, "this statement is all correct." 1. The observations about the contexts in which coordinate conjunctions can be syntactically applied are summed up in the following table. In theory, B can be subdivided into B 1 (not to be used directly as a predicate) and B 2 (to be used directly as a predicate); C into C 1 (the same subjects) and C 2 (not the same subjects ). Little of the present literature available classifies B and C into subgroups, so the following No sentence-coordinate conjunction has so far been found which is only used in [C]. In contrast, for some languages, for example, Hani, Naxi, Yanghuang, Aching, "their juxtaposed phrases do not have to be joined by conjunctions, but it is most common to put a conjunction between them." Moreover, sentence-coordinate conjunctions ④ which can link two clauses in C pattern, can, generally speaking, join first of all two predicates in B pattern, e.g. "而且(ãrqiě)" (which came into being at the latest in the Yuan Dynasty and Ming Dynasty), with a common practice to link two one-syllable adjectives (sometimes verbs) and to function as word-coordinate conjunctions (e.g. (1)-(3)) (B pattern). A few examples of the use of "而且" as clause-linking conjunction existed (e.g.(4)) (C pattern ), and its usage indicating progression did not prevail till the Qing Dynasty.
younger sister, who is pretty <and> virtuous. (the romance of the three kingdoms) All the word-coordinate conjunctions joining two arguments (B pattern) cannot be raised to be the sentence-coordinate conjunction (C pattern). In (5)a, (6)a (Modern Chinese), juxtaposed phrases can only be the lower-level members of the clause subject-predicate, not the upper-level predicate. So, (5)b-(6)b cannot be regarded as single sentences.

. Positions
Conjunctions are termed "middle-position words", which can be used in a balanced way between the former-item and the latter-item but classified neither into the former-item nor into the latter-item. Depending on whether the clause in which the markers appeared is the former item or the latter item in a compound sentence, the sentence-relevant markers are classified as the former-item mark and latter-item mark. It can be seen that the middle-position words playing the word-coordinate role for sentence juxtaposition often move on to the latter items. Compare (7) with (8). "而"(er) in (8) is an example of conjunction having become attached to the latter item, similar to the case with English "and" . In fact, a word-coordinator tends to move on, and mostly it appears before the last element of three items in juxtaposition, e.g. in Chinese "小 张、 小王和小李" (Little Zhang, Little Wang and Little Li) , and likewise with Primi (Qiang branch, Tibeto-Burman group, Sino-Tibetan Family), Hani (Yi branch, Tibeto-Burman group, Sino-Tibetan Family), Indonesian (Austronesian family), English (Germanic group, Indo-European family), and Portuguese (Latin group, Indo-European family). This usage shows a strong universal tendency. In some languages, however, it tends to be attached to the former item, e.g. li 55 in Bai (Yi branch, Tibeto-Burman group) as a word-coordinate conjunction, is a middle-position word but when indicating clause transition it is attached to the former item. Compare: (9)   <  >   (I and she) (10)     ĩ   4  1   <  >,   ĩ   4           (Literally, we today will not finished though, tomorrow mill a while more. / i.e. Though we shall not finish milling today, we can do it tomorrow by milling a while more) Whether they are attached to the former item or latter item, they have all resulted from the increased semantic load of the two items giving rise to a pause in intonation and restricted understanding, therefore a break in structure. This phenomenon is generally understood to be the consequence of word-coordinators changing into sentence-conjunctions. On the contrary, sentence-conjunctions cannot be changed into middle-position words by getting degraded to be used as word-linking elements, because their structural subordination has not changed. For example, " 但 "(d à n ， but) in Chinese, a sentence-transition word expressing contrast, is still attached to the latter item; "由于" (yóuyú， because of) , a causal conjunction, while joining clauses, belongs to the causal item, and remains so while joining substantivals. Compare: (11) bú piàoliang <dàn> gèxìng shízú de nǚrén (不漂亮/<但>个性十足的女人 / a not pretty/<but> individualistic woman) (12) tì nàxie <yóuyú> nǐ cái dūn jiānyù de xiōngdì bàochóu (替那些<由于>你/才蹲监狱的兄弟报 仇 / revenge those who went to jail <because of> you) As for the joining function of a conjunction, the path of its change from word-linking to sentence-linking exhibits a break in structure, its principle being that the word-linking relation was copied as a clausal relation. In a word-conjunction linking framework where clauses rather than words were filled in, the structural load of word-conjunctions swelled, thus giving rise to the appearance of sentence-conjunctions.
It follows from the above-analysis that for coordinate conjunctions with broader application in syntax there is a tendency of change: "word-conjunction > sentence-conjunction". By this logic, it is inferred that while there is frequently polysemy among "word-coordinate conjunction ◇ sentence-coordinate conjunction ◇ sentence-transition conjunction◇sentence-succession conjunction," it is common that the earliest usage is as a word coordinator, and then, through internal change, it comes to be used as a sentence coordinator, and finally as a conjunction of sentence-transition or sentence-succession.
1.2. Zero sentence-coordinate conjunctions Whether a word-coordinate conjunction can be raised to a sentence-coordinate conjunction or not depends on the degree of the need for sentence-coordinate conjunctions, as demanded by the cohesion of a compound sentence. The need for a sentence-coordinate conjunction in turn depends on the cohesion of the clauses juxtaposed in compound sentences of various languages. For example, the loosest type of coordinate clauses in modern Chinese lacks an explicit conjunction (e.g. (13)), whereas in English, sentence cohesion is achieved by the use of "and" in similar cases (Lü Shuxiang 1980). Although both conjunctions and zero conjunctions can be used, conjunctions are used far more frequently for word coordination than for sentence coordination. It is so because word coordination reflects a low-level structural relation in syntax, as the basis on which the high-level structures are formed, and the higher-level structural units need its constituents to be marked clearly. Sentence coordination, on the other hand, is not conditioned as word coordination is, since the former is on a level nearer actual communication, relatively independent of structure. Hence, on that level, the use of a lexical item to mark the sentence-coordinate relation might become superfluous. For example in Primi, Aching, Hani, Lhoba (Bengni-Bokr), between "word-coordinate phrases is often put a conjunction," and "between two clauses of a compound sentence, a conjunction is generally not used." In Zuona Monba, for word coordination a conjunction is optional, but between coordinate clauses in a compound sentence, a conjunction is generally not used. Whether the word-coordination function of a conjunction could, in the course of semantic change, be upgraded to that of sentence-coordination, is constrained by its competition with other means of sentence-coordination (e.g. the relative adverbs, which indicates coordination, etc.). For example, Miao language (Miao-Yao group, Sino-Tibetan family) has   (a word-coordinate conjunction), no  (a sentence-transition conjunction), and the adverb which indicates sentence coordination, but Miao has no sentence coordinators.
Since so few explicit forms mark sentence coordination, and since word transition and word succession have not much need for marking, the conjunctions that remain are often polysemous between word coordination, sentence transition, and sentence succession. Take "word coordination ◇sentence transition" for example first. Mawo Qiang (a language in the Sino-Burman group) has the word-coordinate conjunction , and the sentence-transition conjunction , but no clause-coordinate conjunctions; and Derung language has word-coordinate conjunction   and clause-transition conjunction     、ǎ    but no clause-coordinate conjunctions ⑤ . Li (Zhuang-Dong group, Sino-Tibetan family) uses   as both word-coordinate conjunction and clause-transition conjunction. The coordinate conjunction  of Ewenki (Manchu-Tungusic group, Altai family) is often used between two adjectives and used in a compound sentence to express transition, and "the latter case is relatively more frequent." The word-coordinate conjunction şi in Romanian (Latin group, Indo-European family) has the basic meaning "and", with the function of clause transition in folk speech. There is likewise the polysemy of "word coordination◇sentence succession". For example, n 1 htom 2 、n 1 na 2 in Jingpo (Enkun) (belonging to the Tibeto-Burman group, Sino-Tibetan family), are word coordinators linking two adjectival elements and word-succession conjunctions linking two verbal elements (thus distinguishing their sequence), and both at the same time middle-position words standing between two parallel items; if used as clause-succession words, they shift to stand at the end of the former-item clause. It is seen from the above observations that in the conjunctions" polysemy "word coordination◇sentence transition", "word coordination◇sentence succession" or "word coordination◇sentence transition◇ sentence succession", the change takes place in a certain order: "coordination>succession: transition." 1.3. The changes of adjunction and linkage A case-marking element (CM), such as a preposition or a case auxiliary, can change into a conjunction, and a conjunction can change into an adjunct to the predicate verb. Examples of this kind of internally-motivated change have provided, from various angles, the evidence that conjunctions travel the path "juxtaposition>succession: transition".
1.3.1. The change from a CM to a conjunction Due to their semantic similarity, the coordinate conjunctions have a natural and universal relation with the CMs signifying togetherness. In Chinese, for example, "与(yǔ)", "同 (tóng)", " 跟 (gēn)"and" 和 (hé)" are all simultaneously coordinate conjunctions and prepositions or case markers for togetherness. It is a focus of discussion as to which of them came into being first, which of them later. This problem can be answered against the broad background of the etymological relation between the conjunction and the CM.
Early on, most scholars agreed that first a verb developed into a CM, a process of incomplete grammaticalization, and then into a conjunction, a process of full grammaticalization. Jiang Jicheng (1997), a proponent of this view, said, for instance, that conjunctions "共(gòng), 将(jiāng), 和(hé ) and 同(tóng), etc." in early modern Chinese had resulted from the prepositions expressing the meaning of "togetherness". Chen Yongzheng (1986) stated, "The conjunctions in the earlier time were changed into prepositions from ⑤ Derung, nevertheless, has clause-coordinate adverbs.
verbs through grammaticalization, and then further from prepositions into conjunctions." Chen noted cautiously in his paper that at the onset of the development of a verb into a preposition, the function was almost something between a conjunction and a preposition, whose usage was hard to distinguish from that of a conjunction.
Similarly, in the polysemy between the conjunction and the case auxiliary, the case auxiliary came into being first, and the conjunction later, because the former was the result of the direct grammaticalization of a notional word. Matsumura Akira (1971) said that "the etymology shows that: to(と), a linking word in Japanese (i.e. our so-called conjunction), is generally considered to have come from the case auxiliary to(と)". Tohti· Litip (2001) pointed out that "Uygur bilän (a coordinate conjunction) came from a postposition (i.e. case auxiliary in our terminology). Historical documents also indicate that the usage of a conjunction could have come into being at a later stage. According to A Dictionary of the Turkic Languages, which was completed in the eleventh century, birlä, a postposition in ancient Turkic, was both a marker of instrument case and that of togetherness case, but was not used as a conjunction at that time (Zhao Mingming 2001).
A contrasting opinion is that the conjunctions came into being first, and then developed into the CMs such as prepositions. This view is novel, but the evidence for it is somewhat scanty. For example, Liu Jian et al. (1992) said that "和(he)," which was a verb in the Tang Dynasty, developed into a conjunction at the same time, and into a preposition in the Song Dynasty ⑥ . However, Dong Zhiqiao & Cai Jinghao (1994), through inter-textual analysis, offered a counter-opinion, that the use of the word "和 (hã， and) " as a conjunction resulted from the preposition with the meaning of "together". Yu Guangzhong et al.(1999), having analyzed most of the same materials in Liu Jian et al. (1992), proved that the path of change was "[verb (to mix)]→[preposition (together)→[coordinate conjunction]." The change of a case auxiliary into a conjunction underwent not only the process of continuous grammaticalization semantically, but also syntactically, in that it changed from postposition to a middle-position word ⑦ .
Each of these two processes can be explained by the other. If conjunctions should develop towards case auxiliaries, the origins of conjunctions themselves were the harder to explain. Due to the preponderance of evidence, the first opinion is adopted here, i.e. whenever there is polysemy of "togetherness case marker◇coordinate conjunction", the conjunction has developed from the case marker. This opinion follows the general logic of grammaticalization, i.e. from notional words to functional words (or from less functional to ⑥ See C.J. Fillmore (1980) for similar opinion. ⑦ The changing course of position perhaps takes the reoccurrence of the same item as its transitional phase: "ABx→AxBx→AxB", e.g. Japanese "A は B と→A と B と→A と B". fully functional). The changes from prepositions and case auxiliaries to conjunctions, then, go from CMs to conjunctions or from adjunction to linkage.
Furthermore, based on the direct etymological relation between coordinate conjunctions and togetherness CM, it is inferred that whenever the polysemy consists of "togetherness CM ◇ coordinate conjunction ◇ transition conjunction ◇ succession conjunction", a conjunction"s usage for transition and succession generally develops internally out of usage for coordinate linkage. See §2. Semantic Analysis for more details on the relation between CMs and coordinate conjunctions.
1.3.2. The expansion of a linker to an adjunct of the predicate verb The coordinate conjunction linking predicate verbs can change gradually into adjuncts to them. They can indicate, to a certain degree, a time relation analogous to the succession relation. Some examples are provided by two languages in the Tibeto-Burman group of the Sino-Tibetan family. In the Yi language, the word-coordinate conjunction     linked two substantivals, and when linking a verb and a subject-predicate phrase or a verb, becomes a resultant-complement marker expressing sequence. Chen Kang et al. (1998)， through literal translation of (14) and idiomatic translation of (15), showed how it changed from indicating juxtaposition to indicating succession: Chen Shilin et al. said (1985), "the use of     as a resultant complement marker is an extension of that of     as a conjunction." A second example comes from Jingpo (Daiva).   is a word-coordinate conjunction that links substantive or predicative phrases. When it links two successive acts, it indicates succession, not simultaneous coordination. Xu Xijian et al. (1984), using example (16) below, made it clear that the internal sense-change of the Jingpo conjunction   "coordination>succession" followed in fact the logic for sense-change similar to that of     in Yi language.

The origin of the coordinate sense items
The coordinate sense-items often result from the direct change of the non-coordinate words meaning togetherness, togetherness, sameness, and nearness (collectively-called x). ⑧ Here, the usage of the conjunction e ?55 in Yi language is the same with that of English "and" but different from Chinese usage.
The path of this external change is more or less universal. The following are some examples for each type.
2.1.1. Words meaning togetherness Togetherness CMs/verbs with their accompanying meanings were the most general source of coordinate conjunctions. Instrumental mark/agency verb are also universally turned into coordinate conjunction. It is in fact the metaphor of togetherness or accompanying sense, because instruments are used together with agents.
[Togetherness CM/verb of accompanying sense>coordinate conjunction] "与(yǔ)、暨(j ì) ⑨ " in Archaic Chinese, the togetherness case preposition   、 1 in Li language, the togetherness preposition mai in Wa language, the togetherness case auxiliary   in Tibetan language, are all word-coordinate conjunctions concurrently. The verb "比(bǐ)" in the oracle inscriptions, which indicated togetherness, was also used as a coordinate conjunction. The meaning of "并(bì ng)", whose archaic form was like two men standing together, was extended from its original meaning to a coordinate conjunction (without progression sense) (Zhao Cheng, 1988).
[The instrumental CM/verb with the performative sense>coordinate conjunction] the  in Jingpo(Enkun) is an instrumental case auxiliary, but is used also as word-and clause-coordinate conjunction ⑩ ; menen in Kirgiz, an instrumental case auxiliary, is used also as word-coordinate conjunction (Hu Zhenhua 1986); bilen in Tuva, an instrumental case auxiliary, and togetherness case auxiliary, is also used as a coordinate conjunction; bilän in Uygur, an instrumental case auxiliary, is also used as a coordinate conjunction. "将"(jiāng) in Chinese was used as an instrumental CM in the Pre-chin period (see (17)), as in Yang Jing"s note on Xun Zi: "将，it means "with"." And its use as a coordinate conjunction came late, as in (18) (Yang Bojun et al. 1992). Zhao Cheng"s opinion (1986) that "以(yǐ)" which was used as a coordinate conjunction in the oracles seems a sound-loan word should be considered incorrect, since it ignores the etymological relation between the instrumental CM and the coordinate conjunction. ⑨ It also indicated "到(dà o) 、及(jí)" (meaning "to; to get, to reach," e.g. "上求不暨，是为外利也(国 语) " "The king did no get what he asked for, it was called interest outside of his own" ( 《States' Annals》 ). ⑩ The sound transcription by Liu Lu (1984) is adopted here, but Dai Qingxia et al. (1992) transcribed it as "the ① ". ⑪ "He" means a crane, and "luan" is a kind of bird similar to a crane in imagcry, and both stand for male 2.1.2. Words with the sense of sameness Words expressing sameness sometimes are changed into those expressing coordinative linking. This kind of change is weakly universal.
[Words with the sense of sameness>coordinate conjunction] " 同 (tóng)" denoting sameness in Early Modern and Modern Chinese changed to become a preposition indicating sameness at first, and then a coordinate conjunction (Jiang Jicheng, 1997). "为"(wéi) in the oracle and bronze inscriptions indicated "to do, to be", and in Archaic Chinese was changed into a coordinate conjunction (see below). Two copulas,  and  in Lhoba language also are concurrently coordinate conjunctions.

Words with the sense of nearness
These words which could change into coordinate conjunctions share, in the main, two types of meaning: likeness (类同) indicating the nearness in features, and proximity (比及) indicating the nearness (either in space or time).
[Words with the sense of proximity>coordinate conjunction] Shuowen Jiezi: 及(jí ) means to reach. In the inscriptions on oracles and bronze, its meaning was indicated by a hand having reached a man. It is also used as a coordinate conjunction (以及[yǐjí]).
Near (proximal) deictic ("这"(zhè)、"这样"(zhè yang), and near (2 nd person) reference ("你"nǐ) in Chinese are the result of grammaticalized substantivals with the sense of approaching, and their realization in pronouns. Quite a few coordinate conjunctions in various languages also have senses of near deictic and near reference; their close birds, whereas "diao" and "chan" implies female. So, the line of poem reads roughly as "If met with men, could one talk it as women?" relationship provides further evidence proving the change of "words with approaching senses > coordinate conjunctions", and near deictic had also a important effect on the appearance of the successive senses (see below).
These various sensesreaching, likeness, instrument, togethernessare psychologically related, giving rise to polysemous, synonymous, or cognate relations among words with those denotations. A word with "instrument-togetherness" senses, e.g. birlä in Old Turkic was both an instrumental CM and togetherness CM. Similarly, a word with "instrument-togetherness-reaching" senses, e.g. 将 (ji ā ng) in Chinese is used as an instrumental CM, and has the sense of carrying things and that of reaching, as described in Guangya: Words Explanation. Another example of "likeness◇reaching," from Chinese, is the word "如(rú)," literally denoting likeness, but in some contexts, such as"不"如"你 (bùrú nǐ, "not like you"), it means "not reaching your level." A word with the sense of "togetherness◇likeness", e.g.  in Kangjia (a language belonging to Mongolian group) originally meant "to follow", then changed to mean "being like". Likewise, "像……一 样 " ( xià ng……yīyà ng, A"s image B the same ……) is equivalent to " 跟 …… 一 样"(gēn……yīyà ng, "A following B the same"); both mean "A is like B." "似(sì )" in Chinese, which indicates likeness, is cognate with "以"(yǐ，and), and was an instrumental CM. "是(shì )" in Chinese, which expressed the sense of near deictic in the Ancient Time but is a copula in Modern Time, could indicate metaphor. The meaning expressed by various sense-items of x can all be summed up as "identification". In general, only the elements with the same or near senses can form coordinated constructions. It seems that the motivation for the use of x expressing "identification" as coordinate conjunctions is based on the recognition of the essence of coordinate relations.
Analyses show that a coordinate conjunction was derived from more substantival (not fully grammaticalized) forms in x. It is inferred that when "words with x◇coordinate conjunctions◇transitional conjunctions◇successive conjunctions" constitute a polysemic relation, the transition or succession use of a conjunction is generally derived from the coordinate use through internal changes. Below are several examples from Archaic Chinese.
[用(yòng)] Its original image was a handled-bucket and its extended meaning is "to use" (Yu Xingwu，1978). During the Western Zhou Dynasty, it was grammaticalized into a preposition, and came to be used as a conjunction (Chen Yongzheng,1996). According to Guan Xiechu (1987), in the inscriptions on bronze, there was one example of the use of "用" to link substantivals indicating coordination, one example to link adjectival predicates indicating coordination, and 160 examples to link clauses indicating succession. Here, in cases where there is polysemy between "instrumental CM/words of the performative sense ◇ coordinate conjunction ◇ successive conjunction", the successive meaning of a conjunction appears to have arisen from the coordinate conjunction through internal change. Although the successive use was overwhelmingly greater than other uses in frequency, this did not prove that it enjoyed priority over others in etymology. There is no correspondence between high frequency and original meaning, hence the purpose of differentiation of the original meaning from the basic meaning.
[以( yǐ) ] "以" in the inscriptions on oracles was pictured as a plough, a kind of farm tool. Shuowen Jiezi notes: "以" means to use. During that period, it was used as an instrumental CM (preposition). "以"indicates near deictic (e.g. (22)) and is used as a coordinate conjunction (e.g.(23)) (Zhao Cheng 1986). According to Guan Xiechu (1981) there were later 4 examples in the inscriptions on the bronze of the Western Zhou in which "以" linked substantivals indicating juxtaposition (word-coordination), and 4 examples in which it linked clauses indicating succession (clause-succession). We see that the latter then developed into the usage of the transition conjunction (e.g. (24)). Here, when the polysemy is among "instrumental CM or verb with the sense of performance / word with the sense of approaching◇coordinate conjunction◇succession conjunction◇transition conjunction", the above evidence shows that the coordinate sense of a conjunction came first, and the succession sense and transition sense came into being on the basis of that. [为(wãi)] In the inscriptions on oracles and bronze, "为" meant "to do, to be" and was changed to be a coordinate conjunction in Archaic Chinese (e.g. (25)), and a succession conjunction (e.g. (26)). When the polysemy consisted of "verb with the sense of equivalence◇coordinate conjunction◇succession conjunction", the succession usage of a conjunction came into being after its coordinate use. [若(ruò)] In ancient Chinese, it meant "the same"(see above), and "to reach"(e.g.(27) below). It served as the proximal deictic (this, so) (e.g..(28)), and as the "near" reference (you) (e.g. (29)), and it was a conjunction (see above), including the sense of succession (e.g. (30) and (31)). In cases where the polysemy consists of "Words with the sense of the sameness / words with the sense of approaching◇coordinate conjunction◇succession conjunction", the succession usage of a conjunction resulted from the coordinate usage through internal change. [且(qiě)] According to A collection of Notes on the Function Words in Ancient Books by Pei Xuehai: "且 meant approaching." The notes by Hu Sansheng on Zi Zhi Tong Jian(Shima Guang"s Chronicles) read:"且，a word expressing nearly reaching". "且" indicated near deictic"(e.g. (32)); it also linked adjectives or verbs indicating juxtaposition (e.g. (33)), and it linked clauses indicating succession (e.g.(34)), and it also expressed a transition (e.g. (35)). However, Yan Bojun and He Leshi (1992) believe that its usage of transition "was very rare". When the polysemy consists of "words with the sense of approaching◇coordinate conjunction◇succession conjunction◇transition conjunction", the successive, transitional usage of a conjunction came from the coordinate usage through internal changes. [而(ér )]"而"denoted similarity (as if, like) (e.g.(36)). In ancient time, "而" and"和" were often exchangeable; as Shi Wen (释文， Explanative Notes on the Text in ancient time) stated: in one text the words "如……如……" in (37) were both replaced by "而". The word also expressed near reference (you: both singular and plural) (e.g. (38)); "而" as a conjunction was not seen in inscriptions on oracles and the bronze during the Western Zhou Dynasty, and there were only two examples with clear indications in the literature of the Spring and Autumn Times (Chen Yongzheng，1986), one of which denoted juxtaposition, another sub-succession. After that period, such examples of "而" used as a conjunction began to appear frequently (Cui Yongdong，1984), and an example of its word-transition function was found in (39). Thus, when the polysemy consists of "words with the sense of the sameness/words with the sense of approaching◇coordinate conjunction◇succession conjunction◇transition conjunction", the successive, transitional usage of a word came from its coordinate usage through internal change. The source for the usage of 而 as a coordinate conjunction, which many people consider to be unknown, is in fact clear. Remember the case of "and", an English coordinate conjunction with the polysemy of "a word with the sense of the sameness◇coordinate conjunction◇succession conjunction◇ transition conjunction", and the case of Japanese to(と), with the polysemy of "a word with the sense of the sameness  /with the sense of likeness/togetherness CM◇coordinate conjunction◇succession conjunction". From these two as well as the examples in the paragraphs above, it is inferred that their coordinate usage came first, and the succession and transition usage later.
2.2. The origin of transition senses Some grammar books group all the transition sentences as subordinate sentences (for example, Hu Yushu, 1984:404) and some mainly as compound sentences (for example, Li Jinxi (1933)). Actually, there are two types of transition sentences, i.e. "parallel transition" and "subordinate transition (e.g. concession) sentences, which belong respectively to the compound and subordinate categories. Ma Qinghua (2005), based on the relation of meanin g changes, distinguished two types of transition, i.e. parallel transition and series transition.
Take the former, it can be the same constructions in juxtaposition (e.g (40b) is the same construction as (40a)), those in opposition  " (e.g (41b) is the opposite of (41a)) and those compact constructions in contrasting juxtaposition  (e.g. (42b) is the opposite of (42a): The same constructions were changed by shifting. The compact constructions were achieved by the addition of relevant elements or assisted by their affirmative or negative transformation. It is characterized by controlled shift. The opposition relation was changed by either shifting or controlled shift. We therefore conclude that the transition relation appeared later than the coordination.
The transformation of the same constructions takes place between the typical and atypical coordinate relations. Before a new connective function was lexicalized, it often appeared first in the capacity of an atypical usage. Almost all coordinate constructions in modern Chinese (for example, "既 A 也 B","既 A 又 B"(both A and B), "一面 A 一面 B","又 A 又 B"(A on the one hand and B on the other), etc.) can imply the sense of transition (Xing Fuyi，1998). Compare (43) with (44): (43) Wǒ <jì> bù xiǎng zànměi zhè zhǒng yīnyuán, <yě> bú yuàn zài cǐkè yòng yáncípòhuài tāde xīnjìng. (我<既>不想赞美这种姻缘,<也>不愿在此刻用言辞破坏她的心境｡ / I will<neither> speak highly of this conjugal felicity, nor disturb her mood by speech.) (typical juxtaposition )  The full name of "the opposite(对式)" is"the opposition construction (对立式) ". The term was coined with reference to the contradictories in N. Smith & D. Wilson (1979).  It can also be a progressive-compatible construction and the compatible construction is another means of addition. Compare:"他听着，这姑娘的形容很奇怪，更奇怪的是她形容得挺准确(He is listening, thinking this girl"s description very strange, and furthermore, very accurate (progression))~他听着，这 姑娘的形容很奇怪， (但)更奇怪的是她形容得挺准确(张承志《北方的河》 (He is listening, thinking the girl"s description very strange, (but)what"s more, very accurate (Zhang Chengzi《The Northern River》) (parallel transition))". This shows that in transformation, the transition relation came later that than other joint relation. Coordination is typical of compound sentences and its characteristics are weak. Once the former or the latter item means implicit transition, that sense easily shows from behind the coordination. Hence, the insertion of "但(dàn)", a transition word, at the beginning of the second clause in (44) does not change the meaning. Only when the veil effect of coordination wore out in the atypical coordinate usage implying transition, could the new sense-item expressing transitional function be established. Besides, in the atypically coordination implying transition, only a pure coordinate conjunction can have compatibility with transition conjunction (compare (45a) with (45b)), but not vice versa (compare (46a) with (46b)). This is the basis for the sequence "coordinate>transition" instead of the reverse process. The lexicalization of this sequence of structure resulted in this path of grammaticalization. Although semantic contradiction between the former item and the latter item is the origin giving rise to transition function derived from coordinate conjunctions, transitional function motivated directly from outside of the relevant domains can hardly cause in return the production of the coordinate function by semantic means. The transition conjunctions, for example,"但(但是) (but) ，只是，不过(however, nevertheless)" in Chinese, "but" in English, "damu" in Manchu, etc., which transformed from a few category words ( originating by the principle of euphemism)，do not have the coordination usage.
2.3. The origin of succession senses Succession can be regarded as a specialized coordination. A coordinate conjunction can link constructions of various kinds, but a succession conjunction generally links verbals or clauses, because only the latter admits of time sequence. When there is a temporal sequence between two actions or events in juxtaposition, the original coordination can naturally reveal a sense of succession. The similarity between the coordination and the succession seems to be greater than that between the coordination and transition. No wonder some linguists have included succession within the coordination. The lexicalization of the sequences of structural changes "coordination > succession" gave rise to the grammaticalization "coordination>succession" in conjunctions.
In Spanish, clause coordinator "y" connects two affirmative clauses in juxtaposition. If the tense of the first clause comes before that of the second one, then succession sense was expressed. The change seems to have been realized by taking the parallel coordinate structure as a constant, and the time sequence as a variable. Then, in cases where one clause is in the affirmative, but the other is negative, a transition comes into being.
(47) Tā bǎ kēng tiánmǎn, <bìngqiě> yònglì zài shàngmian cǎi le jǐ jiǎo. (他把坑填满,<并且>用力在 上面踩了几脚｡/ He filled the pit(with earth), <and> stamped it hard several times.) Although the successive sense of "并且"is not found in dictionaries, it can be said at most that the relation has not been lexicalized yet. The succession conjunction (e.g. 然后 ránhò u，then) motivated directly from outside of the relevant domains can hardly be used as a coordination marker. This is further evidence for the grammaticalization path "coordination>succession" in conjunctions.
The fact that the coordinate sense of a conjunction changed internally to successive sense and transitional sense does not mean the coordination is a stimulant to produce successive sense and transitional sense; rather, it only provides the structural condition for it. Only the concrete relation of senses between the former-item and latter-item, for example, the time sequence or affirmative-negative relation, is the fundamental stimulant. As stated above, a near deictic, as a realization of approaching sense in pronouns, appears in the polysemic system of the coordinate conjunctions. The combination of the near-deictic sense-item (这 [this]) with the structural conditions and the sense-relation between the former-item and the latter-item, could quicken the lexicalization of a succession conjunction. Succession involves the sequence in time, whereas a proximal deictic in itself does not involve sequence. It can, however, express a confirmation of a fact in anaphoric usage, and become a natural marker for time sequence. так in Ukrainian, used as an adverb, means "this way, so"; as a conjunctions, it means "then, so." So is "是(shì )" in Archaic Chinese, cf. (48). Lü Shuxiang (1985) stated, "这(zhâ，this)" means "then" and "like this", with succession function, as seen mainly in idioms "这就(zhè jiù)、 这才(zhè cái)、 这可(zhè kě)"(e.g. (49)) ⑫ . ⑫ A near deitic could also express causes which were the former-latter relations in things and reasons. For example, "此(cǐ)", a near deitic pronoun, indicated the results: "有德此有人，有人此有土，有土此有财， 有财此有用 ( 《礼记· 大学》 ) "(The fact that one is resulted in the fact he has men, then the land, and money, and useful facilities( 《The Book of Rites· Great Learning》))."So" in English is used as an adverb of degree, indicating "by this way"; as a conjunction, indicating "hence". eva, an adverb in Old Sanskrit, indicating "so", and "like this way, hence" too.

Conclusion
Analysis shows that "coordination>succession: transition" is a well-worn path in the grammaticalization of coordinate conjunctions, a path showing the development of specific structural relations out of forms that originally denoted non-specific structural relations. The relative universality of this pattern further supports the claim that in most cases the higher the abstraction of sense classes, the more identical the conceptual motivations are (Ma Qinghua, 2000: 96-121). This path represents only the tendency of change, but not a rule without exception, i.e. it does not mean that coordinate conjunctions must develop into successive or transitional conjunctions, nor vice versa, that the latter must develop from the former, for it is possible for markers of succession and transition to have been motivated from outside of the relevant domains, and in addition there are the ways of external change, etc. In general, the simpler, older, and commoner the conjunction is, the more dominant in colloquial use the connective form is, the closer it is related to prototype, and the more possible the grammaticalization. For example, the expansion of one-syllable conjunctions in Archaic Chinese was more active, but in modern Chinese, owing to its more developed connective means, two-syllable conjunctions are in the state of inertia and single meaning. This has confirmed Li Yingzhe (2001)"s point, that the development of conjunctions could often be divided into two phases of activity: appearance in groups and progressive formation, and the large-scale grammaticalization of conjunctions took place mainly in the former phase (for Chinese, it was respectively in the Spring-Autumn and Warring-States period, 770-221 BC and in the Dynasties of Song and Yuan, 960-1368). The lengths, directions, the degrees of the expansion of conjunctions can vary, i.e. some would change to transition or succession and stop, some continued to expand till they developed into subordinators, but these will be discussed elsewhere.