13.1 Introduction

The Chinese language had no aspect suffixes attached to the verb stem before the tenth century AD. In his seminal work The History of Chinese Grammar, Reference WangWang (1989: 99) stated that the emergences of the aspect suffixes le (perfective) and zhe (progressive) were epoch-marking events in the history of Chinese grammar and that their development processes should be seriously studied. To explain the reasons why they emerged at the time, two closely related questions need to be answered: what factors in the earlier grammatical system blocked aspect suffixes from grammaticalizing? What changes in the grammar at the time made it possible for them to come into existence? When studying the motivation and mechanism for their development, we must separate two issues: (a) the overall change of the grammatical system at the time and (b) the concrete grammaticalization process of each aspect suffix.

The emergence of the aspectual system actually created a new grammatical category, namely the verb suffix, which was expressed by auxiliary verbs before this time (we will return to this point in the subsequent section). In the evolution of Chinese grammar, grammatical changes can be classified into the following five types:

1. (a) The grammatical category simply died out, which meant that there was no replacement in the grammar at a later stage, such as the co-ordinate conjunctions for verbs ér and qiè in Old Chinese, which fell out of use in the first half of the Medieval Chinese period and have had no equivalents in the grammar since then (for details, see Chapter 4).

2. (b) The grammatical morpheme was replaced by a later form; for example, the disposal jiāng occurred approximately 200 years earlier than the disposal bǎ, and after more than a millennium of competition, the disposal bǎ entirely replaced its predecessor jiāng (for details, see Section 9.9).

3. (c) Some grammatical apparatus was entirely innovative in terms of the grammatical category but not in terms of the word class of its markings. For instance, the disposal morphemes belong to this type because they mark the disposal construction, which did not exist before the sixth century AD, but the markings were prepositions that existed in the language from the beginning (for details see Chapter 9).

4. (d) The grammatical elements were completely new in terms of grammatical category and the word class of marking, such as classifiers, the plural marker -men, and the diminutive marker -er (see Chapter 19 for details).

5. (e) The grammatical category has always remained in the language but the word class representing this category has changed. This type of change includes the emergence of the aspect suffixes. As will be discussed below, the same grammatical category was represented by auxiliary verbs in Old Chinese, where they preceded the verb, but in Modern Mandarin the category is represented as suffix as a kind of clitic which immediately follows the verb.

Some clarification is needed here. Aspect is easily confused with tense because they are often intertwined in use. Reference WangWang (1989: 91) correctly pointed out that the suffixes in question stand for aspects rather than tenses. Aspect refers primarily to the stage at which the action denoted by the verb takes place, focusing on the internal development of an action: whether it has reached completion (with or without current relevance) or is in progress. Despite the same label being used for a type of aspect, such as “perfective,” the meanings and functions of a particular aspect vary greatly from language to language, depending mainly on the division of labor of the aspect categories that the language grammaticalizes. For example, English has two aspects: progressive (e.g. I am eating) and perfect (e.g. I have eaten), but Chinese has a three-way contrast in aspect: perfective, progressive, and experiential. Consequently, the perfective aspect in English must have characteristics that differ from those of its counterparts in Chinese. For convenience of discussion, let us first consider the definitions of the three primary aspect suffixes (for a full discussion, see Reference Li and ThompsonLi and Thompson 1981: 184‒237, Reference ShiShi 2010: 229‒247).

1. (a) Perfective -le, referring primarily to an action in the past but with current relevance or denoting a present state resulting from a past situation. When occurring after a VO phrase or in sentence-final position, it indicates a change of state, which may be labeled le₂ in the literature. This perfective aspect focuses on the current situation after the action has happened. This term is labeled “perfective” in Reference Li and ThompsonLi and Thompson (1981: 185) and describes an action viewed as a simple whole. Their definition accurately characterizes the function of the suffix -le, as the following examples show its contrast with the suffix -guo. Thus, in our reanalysis, we use the label “perfective aspect” to refer to -le, including both le₁ and le₂. Historically, the two perfective markers evolved from the same lexical source.

2. (b) Experiential -guo, referring primarily to an action in the past, without the implication of current relevance, focusing on the end point of the action. The contrast between -le and -guo is evident in the following examples.

If there is no time word in a sentence, the default time reference is the point at which the speech action takes place. Thus the aspect markers -le and -guo often indicate an activity in the past, which makes them appear to be a past tense, but they can also refer to an action in the future if the time reference is set in the future, as in the following examples.

1. (3)

   明年這個時候我就拿到了畢業證。 (現代漢語)
   Míngnián zhège shíhòu	wǒ	jiù	nádào-le	bìyè zhèng.
   Next-year this-CL time	I	then	obtain-PERF	graduate diploma
   “I will have obtained my graduation certificate at this time next year.”

Note that -le and -guo often co-occur within a verb phrase, as illustrated above, and the order is always V-guo-le, where -guo highlights the action in the past and -le refers to the current relevance.

1. (c) Progressive -zhe, referring primarily to an ongoing action at a temporal point or span. This aspect roughly corresponds to the progressive in English, as illustrated in (5):

1. (5)

   我喫着午飯呢。		(現代漢語)
   Wǒ	chī-zhe	wǔfàn	ne.
   I	eat-PROG	lunch	PRT
   “I am eating my lunch now.”

Owing to the same motivation and mechanism as the emergence of the above aspect suffixes, and during the same period in history, the language developed other minor aspect markers such as the inchoative -qǐlái, referring to the start of action, and the so-called delimitative aspect – the reduplication verb form, which will also be discussed in subsequent sections.

In Contemporary Chinese, the aspect suffixes can be freely combined with verbs and adjectives. Note that, as discussed repeatedly in this book, in Chinese grammar the verb and the adjective are syntactically alike; hence a cover term, wèicí “predicate word,” has been coined to label both of them (Reference ZhuZhu 1982: 40). Specifically, adjectives in Chinese behave like verbs; for instance, they can be directly used as the predicate without a copula such as be in English to link it with the subject. Moreover, the three aspects have been cliticized and undergone phonological reduction due to the effect of grammaticalization; all have lost their tonal values, and their vowels have become a schwa [ǝ]. Their phonological forms in Medieval Chinese and in Contemporary Chinese are listed in Table 13.1.Footnote ¹

The emergence of these aspect suffixes creates a distinction between finite and nonfinite verbs in the grammar, and a sentence is structured around the center of the finite verb. If two or more verbs refer to events that happen at the same time, only one of them can be suffixed with an aspect marker, or another verb may be suffixed with the progressive -zhe (similar to I saw him playing basketball in English), as illustrated in (6):

As exemplified above, the emergence of the finite clause was a major event in the evolution of Chinese grammar. As a consequence, the language has acquired a morphological method for distinguishing prepositions from verbs. All prepositions in Chinese developed out of ordinary verbs, but before the Modern Chinese period it was difficult to tell prepositions from verbs because of the lack of overt morphological forms. In Contemporary Chinese, some prepositions, such as the disposal bǎ and the passive bèi, have permanently lost their verb status, but some of them, such as the dative gěi, whose verb status “give” is still in use, can still be used as a verb in different contexts. Morphological markings are helpful in structuring a sentence by reducing potential ambiguity, and in this sense this change may be regarded as a progression of the grammar.

13.2 Conditions for the Emergence of Aspect Suffixes

In fact, the emergence of aspect suffixes was a by-product of the development of the resultative construction. In other words, it can also be said that the aspect suffices virtually belong to a subtype of resultatives that are just a small group of resultatives with general meanings suitable for describing the internal structure of an action. For instance, the perfective -le grammaticalized from the verb “finish,” the same change found in many languages (Reference Heine and KutevaHeine and Kuteva 2002: 334). The combined force of the reanalysis of numerous resultative cases created a new syntactic position, labeled R in the square brackets in the following schema, between the verb and the object, where R could be intransitive and might bear no direct semantic relation to the object (for an in-depth discussion, see Chapter 6).

1. (7) V [R] O

The creation of the R position between the verb and the object was necessary for the emergence of many major grammatical devices in Modern Chinese, including the resultative construction, the three aspect suffixes, and the reduplication of verbs. The structure depicted in (7) was achieved through two critical changes in the grammar that are outlined as follows.

The first change was the disappearance of the verb conjunction ér, an important grammatical morpheme in Old Chinese (for details, see Chapter 4). At the time, instances of the serial verbal construction were ruled out because they were ill-formed in the grammar (Reference WangWang 1989: 255‒261), and the grammar then required the conjunction -ér to link any two verbs, adjectives, or even adverbs within a co-ordinate construction. In Old Chinese, the verb and the resultative were also separated by this conjunction, as illustrated below:

In the above examples, the underlined words refer to the resultatives of the verbs, which were separated by the conjunction -ér. Because the verb and the resultative could not occur adjacently, no reanalysis could happen to make them a compound. Due to the disappearance of the conjunction between the first century BC and the fourth century AD, the serial verb construction became the paradigm structure, which enabled the reanalysis of the verb and the resultative to take place. Then, the original boundary between the verb and the resultative weakened and they eventually fused into a single constituent consisting of “verb + suffix.”

The second change was that the co-ordinate verb construction ceased to work. Roughly before the fifth century AD, the co-ordinate verb construction required that all verbs be transitive and each of them was combined with the object to form a “verb–object” relation (for a fuller discussion, see Chapter 4), which can be formalized as follows:

1. (10)

   (V1 + V2) + O = (V1 + O) + (V2 + O)

   An example is provided below:

Under the operation of the rule of verb co-ordination, if the second verb element was intransitive, the constituent order could only be “V_tr O V_intr,” where the resultative element was separated from the first verb by the object. Reference Ohta, Shaoyu and TianhuaOhta (1987: 360‒367) was the first to find the contrastive distribution of the two synonymous verbs – shā “kill” and sǐ “die” – when used as a resultative: the former could occur only between the verb and the object because it was transitive, and the latter could only follow the whole VO phrase because it was intransitive. This could be viewed as a rule derived from the general principle regulating the co-ordinate verb construction. This was a rigorous rule by at least the fifth century AD (for details, see Section 6.4), as illustrated in (12) and (13):

The structures determined by the two grammatical rules discussed above channeled the path and process of the developments of the resultative construction and the aspect suffixes. Clearly, when the rule of the above verb co-ordination was still in operation, the VRO structure was ruled out by the grammar because resultatives were usually intransitive and did not syntactically govern the object argument. Under this condition, it was impossible for aspect suffixes to emerge.

Thus far, many studies on the emergence of the aspect suffixes have been conducted in the circle of Chinese historical linguistics. The above phenomenon of the grammar at that time is helpful in understanding the shortcomings of previous studies on the subject of the development of the aspect suffixes. Regarding the perfect marker -le, superficially, its emergence involved a change in the following constituent order:

1. (14) V O liǎo > V-le O, where liǎo is the former verb form of the perfective aspect -le.

Only when -le could occur between the verb and the object can we say that it had been grammaticalized into an aspect suffix. How could the verb liǎo jump across the object in history? Many researchers have proposed that this change happened on analogy with other resultatives such as de “modal suffix,” què “complete,” and zháo “attach,” because this group of intransitive words originally followed the VO phrase, and they moved to the position between the verb and the object at a later stage. However, this explanation fails to answer the ultimate question: what factors triggered the position change of these intransitive verbs? Furthermore, as Reference JiangJiang (2005b: 136‒154) correctly pointed out, the position changes of the modal suffix de happened even after the perfective -le had changed its position, and so it was impossible for the former to affect the latter.

At face value, the emergence of the “V [R] O” structure violated the grammatical rule of the co-ordinate verb construction “(V_tr-1 + V_tr-2) + O” as analyzed above. Historically, these two constructions were indeed incompatible with each other. The establishment of the “V [R] O” structure occurred at the expense of the earlier co-ordinate verb construction. After the Late Medieval Chinese period, the following construction, which was extremely common in Old Chinese, became impossible (for details, see Chapter 4).

In Contemporary Chinese, there is no conjunction word such as ér to co-ordinate two verbs or adjectival phrases. Additionally, the structure in which two or more verbs share an object, e.g. mǎi chī fàn “buy eat food,” is ungrammatical.

Here, a crucial question arises: how could a new structure enter a language by violating an existing grammatical rule? As we discussed in Chapter 5, the earliest VRO instances came into existence through compounding, and at the beginning they belonged only to disyllabic compound verbs. Under the influence of the disyllabification tendency, pairs of verb and resultative that were both monosyllabic with a high frequency of co-occurrence first formed disyllabic compound verbs. The compounding happened in the context where no object occurred between the verb and the resultative. At the early stage, every verb–resultative phrase (including the aspect suffixes) had to undergo this process of compounding, and once the number of these compound verbs had reached a certain point, a new paradigmatic structural schema was innovated that was highly productive and influenced other old usages via analogy. At this point, new members occurring in the R position of the VRO form no longer had to undergo their own compounding processes. As a productive new paradigm, the resultative construction was finally established around the eleventh century AD.

13.3 The Emergence of Aspect Suffixes

13.4 Inchoative and Continuous Aspects

As mentioned above, the creation of the syntactic position between the verb and the object that could contain an intransitive element to describe the situation resulting from the action of the preceding verb was one of the most important changes after Late Medieval Chinese. This change exerted a far-reaching effect on the development of Chinese grammar. The position could be occupied by numerous lexical items that may be regarded as some kind of resultative that describes the state resulting from the action of the matrix verb but bears no action–patient relation to the following object. Some lexical items, such as téng “painful” in xiào-téng dùzi “laugh-painful stomach” (the stomach hurts because of laughing too much), were only sporadically used in this position, but some resultatives, such as the aspect suffixes and the modal suffix discussed above, might have developed into stable grammatical devices, due to their semantic suitability in relation to the verb and the frequency of their uses as resultatives. In addition, two minor aspects are also worth mentioning: (a) inchoative aspect -qǐlái (lit. “get up”) and (b) continuous aspect -xiàqù (lit. “go down”). The former means the start of an action, and the latter refers to the continuation of an action; both were introduced into the language after the fifteenth century AD, when the resultative construction and the aspect system were already well developed. First, let us consider two early uses of the inchoative aspect:

The earliest instances of the continuous aspect -xiàqù were attested in the nineteenth century AD, nearly five centuries later than the inchoative aspect and almost a millennium later than the perfective aspect. Let us consider one of the earliest examples of the continuous aspect.

Regarding the words in the X position of the VXO construction, there was a continuum from concrete lexical items to highly grammaticalized markers. This diachronic evidence shows that any boundary between lexical items and grammatical morphemes is by nature arbitrary.

As we already know, under the influence of the disyllabification tendency, the resultative construction was able to emerge and develop through the compounding of the verb and the resultative (for details, see Chapter 5). During the process of the development of the resultative construction, the active and productive rule was still verb co-ordination, which regulated the pathways of each “verb + resultative” pair that took place, including the aspect suffixes. In the period from the fifth century AD to the twelfth, the rule worked at sentence level, while the compounding of verb–resultative phrases happened at the word formation level. The VRO construction as a new paradigmatic structure that entirely replaced the verb co-ordination rule was quite a late event, approximately after the twelfth century AD. After that, the co-ordinate verb construction became ungrammatical and was eventually ruled out by the new grammatical principle, and the VRO pattern became a highly productive paradigm that leveled irregular and less frequent verb–resultative forms. For instance, as mentioned earlier, the verb bì “complete” could occur after the VO phrase to refer to the completion of the action represented by the verb, a usage attested as early as the fifth century AD. Since the verb liǎo became dominant in expressing this function, bì was restricted to the written language. In texts composed around the sixth century AD, however, bì could also occur between the verb and the object, as illustrated in (63). This use of bì did not come out of its own development but clearly was due to analogy with the perfective -le and the like.

The verb bì “finish” was only a lexical item used in the written register and never developed into an aspect suffix.

The establishment of the “VXO” structure, where X could be a suffix of the verb, also made it possible to borrow the schema of word formation from other languages. One of the representative examples in this regard is the suffix -ize and other verb suffixes of English, which function to turn nouns or adjectives into verbs, e.g. modern ~ modernize, nasal ~ nasalize, fossil ~ fossilize. In the past century, Chinese has created a suffix, huà, that is extremely productive in making a verb out of a noun or adjective, and some of the derived verbs even have no counterparts in English, as shown below:

The derived verbs above behave like transitive verbs, which can be followed by an object noun. In our view, only after the establishment of the aspect patterns could this sort of schema-borrowing of word formation become possible. In other words, in the grammar of Old Chinese, no suffixes were allowed to be attached to the verb so that the structure “V-huà + O” was ill-formed and thus ruled out by the old grammatical system. As far as grammatical schema are concerned, language borrowing cannot be allowed if it violates the grammatical rule of the receiving language.

13.5 Interaction between Aspect Suffixes

Cross-linguistically, aspect markers vary from language to language, and within a language the meaning and function of a particular aspect marker depends on the contrast with other aspect markers. In Chinese, aspect suffixes emerged after the tenth century AD, but they did not all come into existence at the same time. As discussed above, the perfective -le became grammaticalized around the tenth century AD, and the experiential -guo and the progressive -zhe were introduced into the language approximately 200 years later. According to Reference WangWang (1989: 102‒121), it took approximately eight centuries to set up the division among these aspect suffixes. He observed that there was some confusion of the uses of the perfect and progressive aspects; for example, in (64) the perfective -le was used as the progressive suffix, and in (65) the progressive -zhe was used as the perfective aspect:

Additionally, there were some confusing usages of the perfective -le and the experiential -guo. This phenomenon has not been reported in the literature, probably because the difference between these two aspects was too subtle for their mixtures in history to be noticeable. Both referred to the completion of an action in the past, and their characteristics were that the perfective -le emphasized the current relevance and the experiential -guo focused on the end point of the action. In certain contexts even native speakers have difficulty telling one from the other. However, these two aspects were indeed confused during their development. For example, the experiential -guo in (66) is used in an imperative sentence whose action has not yet started, where only the perfective -le can be used according to the grammar of Contemporary Chinese:

Furthermore, in Contemporary Chinese, psychological verbs, such as rènshí “know,” xuéhuì “learn” and dǒngdé “understand,” can be suffixed only with the perfective -le and not the experiential -guo because they lack the semantic characteristic – a clear end point that the experiential aspect requires. As exemplified below, however, in texts composed in the nineteenth century AD, the verb rènshí “know” was suffixed by the experiential -guo, where the perfective -le should be used now:

If we consider the situation in Chinese dialects, the picture appears extremely messy. Grammatical morphemes typically result from lexical items, and the lexical item must satisfy two conditions: possessing semantic suitability and frequently appearing in particular contexts. However, these conditions are applied only to those that really undergo a grammaticalization process. For this sort of aspect suffix, there are always transparent semantic relations between the functions of the aspect suffixes and the meanings of their lexical sources (see Table 13.5).

We suppose that all the above grammaticalization events genuinely happened in standard Mandarin (i.e. the northern family of dialects), but not all the dialects actually underwent these developments; instead, they simply borrowed the aspect system from the standard language. As a result, there are many mismatched aspect suffixes with regard to lexical sources. As we know, Chinese has eight dialect families and hundreds of subdialects. For the sake of simplicity, we focus only on how the perfective aspect is expressed in different dialects, and, as we see below, all six types of different aspect device are used to express the perfective aspect, as a stable grammatical device in those dialects.

1. (a) Perfective aspect expressed by the progressive -zhe, e.g. Cantonese, as illustrated below:

1. (b) Perfective aspect expressed by the modal suffix -de, e.g. Guiyang dialect, as illustrated below:

1. (c) Perfective aspect expressed by the inchoative -qǐqù, e.g. Sichuan dialect, as illustrated below:

Quite surprisingly, verb reduplication is also recruited to express the perfective aspect in some dialects. As discussed in the subsequent section, only quite recently (around the thirteenth century AD) was verb reduplication introduced into the language, defined as a delimitative aspect, expressing the short duration of an action. However, in quite a few dialects, particularly those of the Wu family, such as the Zhejiang dialect, the reduplicated verb form can function to express the meaning of the perfective aspect, as illustrated below:

Some dialects even adapt the diminutive marker -er to express the meaning of the perfective aspect. As we will discuss in Chapter 19, the word ér “child” started its grammaticalization process toward a diminutive marker, an apparatus attached primarily to nouns, after the sixth century AD. Interestingly, it has been used as a perfective aspect in some dialects, such as the Haiyang and Muping dialects (Reference HuangHuang 1996: 175), as exemplified below:

It seems that these dialects do not care where the forms come from and what their original functions were; they simply adopt a form to perform the task of expressing the meaning of the perfective aspect, a grammatical category that was developed in the northern dialect and borrowed by many other dialects. Grammatical borrowing occurs much more easily across dialects than across languages, creating many interesting phenomena that are theoretically significant.

13.6 Diverse Forms across Dialects

It is well known that a grammaticalization process often proceeds along a cline (Reference Hopper and TraugottHopper and Traugott 2003: 6):

1. (73) Content item > grammatical word > clitic > inflectional affix > zero

However, this cline should be considered with caution because it reflects only one type of change that has undergone a complete development process. According to Reference Harris and CampbellHarris and Campbell (1995: 50), there are only three basic mechanisms of syntactic change: reanalysis, extension, and borrowing. If a language or a dialect simply borrowed a grammatical category from another, what would happen to the phonological form? Apparently, the borrowed item does not have to go through all the stages of the grammaticalization process. Because of the vast geographic area and the extreme diversity of dialects in China, grammatical borrowing across dialects, similar to loanwords, happened often. Grammatical borrowing is different from contact-induced change because it does not have to happen between geographically adjacent dialects. In this section, we deal with various forms of grammatical borrowing across dialects.

We often find that for a given grammatical category, some dialects employ just a sound feature to express its function, whereas standard Mandarin makes use of a function word; e.g. the plural is expressed by the morpheme -men in standard Mandarin but by a high-pitched tone in Cantonese, by reduplication in the Chongqing dialect, and so forth. If we quickly jump to the conclusion that the plural markers in these dialects are more grammaticalized than those in standard Mandarin simply on the basis of their phonological representation, we actually ignore the many possible ways that a language community can acquire a grammatical device. In addition to grammaticalization, a grammatical device can be obtained through phonological derivation (e.g. the emergence of distal demonstratives; see Chapter 18 for details), borrowing from other dialects, or developing it by itself. In what follows, this point is illustrated by the rich phonological representations of aspect markings. Some dialects seem to rely purely on phonological form to express the function of the perfective aspect, such as the addition of a sound, a change in tonal value, or a vowel change.

1. (a) In the Hakka dialect, the vowel [e] is added to the matrix verb to express the completion of the action with current relevance (Reference HuangHuang 1996: 175):

1. (b) Some dialects express the function of the perfective aspect by changing the tonal value of the matrix verb. According to Reference HuangHuang (1996: 175), in the Shangxian dialect (spoken in northwestern China), there are four tones, each of which has a lengthened tone with the following tonal contours:Footnote ³ 21~3231, 24~2141, 53~5231, and 55~551. The verb bearing the long tone is equivalent to the verb plus the meaning of the perfective -le in standard Mandarin; thus no other perfect suffix is needed. In the following example, the tonal value of the verb tʂ‘ǝ²¹ “eat” can be lengthened to tʂ‘ǝ³²³¹ to express the completion of the action.

A similar strategy to express the perfective aspect is found in Cantonese (Reference HuangHuang 1996: 175), a southern dialect that is far from the Shangxian dialect. Specifically, the original tone of the matrix verb is changed to a high-pitched tone value of 35 to express the function of the perfective aspect, as illustrated in (76):

1. (c) Altering vowels to express the perfective aspect. According to Reference HeHe (1989: 58), in the Huojia dialect (belonging to the Jin family), the vowel or the rhyme of the syllable of the verb is altered to express the function of the perfective aspect. This dialect has a phonological rule to derive the form of the perfective aspect from its base syllable. Here we provide only two examples to illustrate this point. In the following two pairs of examples, the first example has the base form of the verb, which refers to an action that has not happened yet, and the second example has the derived forms of the verb, which refer to the completion of the action with current relevance.

It happened from time to time in history that when the language developed to a certain point where the conditions became mature or certain pragmatic devices were ready, the language community then recruited those semantically suitable lexical items to develop the grammatical morpheme that was favored or needed by the grammatical system at the time. The change first occurred in the central dialect of the language, namely standard Mandarin, which was spoken in the vast majority of the country, at a particular period. This situation was something like a general call, with every dialect having to respond to the new grammatical innovation. For example, there were a set of grammatical innovations that did not exist before the fifth century AD, such as the classifier system, the plural, the diminutive, and the system of the aspect suffixes. No dialect could resist these changes and remain in the former state of its own grammar, but the methods of responding to these innovations varied greatly from dialect to dialect. Now, let us summarize how many methods standard Mandarin used to innovate grammatical apparatuses and how the other dialects reacted to represent the innovated grammatical devices.

1. (a) Grammaticalization. This is a major channel for a language to acquire an innovative device. For example, the disposal bǎ grammaticalized from a verb meaning “take”; the structural particle de (marking relative clause, genitive, and associative) from a demonstrative; the plural men from a classifier for family; and the diminutive er from the concept “child.”

2. (b) Phonological derivation. This method is typically applied to two closely related grammatical categories within a grammatical domain. For instance, the proximal and distal demonstratives are crucial for the daily communication of any language community. It is unlikely that each side grammaticalized from different lexical sources, undergoing independent development and finally reaching completion at different times. As we will see in Chapters 17 and 18, the proximal demonstratives zhè in the northern areas and gè in the southern areas developed out of a general classifier in Late Medieval Chinese, and they acquired their distal demonstrative usages simply through a phonological rule.

3. (c) Conventionalization of the function and the phonological form. This method is different from phonological derivation, termed “grammatical borrowing across dialects,” and it is like coining a new word whose meaning and phonological form are purely conventional or entirely arbitrary. This often happened to dialects in isolation from other dialects in ancient times. These dialects simply chose a phonological form to express the function of the newly invented grammatical categories in standard Mandarin. For instance, as we saw in related sections, the plural is expressed by a nasal consonant, the diminutive by a high-pitched tone, the perfective aspect by a vowel shift, and so forth.

4. (d) Adapting the form from the major dialect. This represents a special kind of language borrowing across dialects that often results in a mixture of function and form. For example, the perfective aspect is expressed by verb reduplication, by the modal suffix de, and even by the diminutive marker -er in different dialects. The language communities in these dialects have treated the grammatical categories as loanwords and have often chosen different phonological forms to encode the function of the borrowed grammatical categories.

5. (e) Regional grammaticalization. For a given grammatical category, some dialects choose their own lexical sources to develop a grammatical morpheme. As we saw above, for example, the progressive marker -zhe grammaticalized from the preposition “in” or “to” in Medieval Chinese. Similarly, many dialects have chosen other prepositions, either zài “in/at” or dào “at/to,” to express the function of the progressive aspect. For example, the preposition dào developed into a progressive marker in the Hubei, Sichuan, Guiyang, and Yunnan dialects, as illustrated in (79), and the preposition zài “in/at” grammaticalized into a progressive aspect in the Chaoxian, Hefei, Huoqiu, and Southern Min dialects and is still a free word rather than a suffix because it can be separated by the object, as illustrated in (80) (for a detailed description, see Reference HuangHuang 1996: 175):

According to Reference HuangHuang (1996: 204), the progressive use of the preposition can be dated to Late Medieval Chinese, for example:

That is, the two synonymous prepositions zháo and zài were competing for the progressive aspect at the time. The former eventually superseded the latter in northern areas, but the latter was still preserved in some southern dialects.

The above factors are largely responsible for the great variations and diversity of forms in expressing a given grammatical category across dialects. According to Reference CaoCao (2008: 95), for instance, there are at least sixty-nine different markers for passive structures. In addition to the grammaticalization process, there are many other channels for a dialect (including Mandarin Chinese) to acquire a new grammatical device, which makes the picture extremely complex.

13.7 Future Tenses

Both aspect and tense belong to the major grammatical categories of verbs across languages, but the Chinese language developed aspect suffixes only, without acquiring tense markings during the development process. Why was Chinese unable to develop tense markers? Now we are in a position to answer this question. Tense and aspect are closely related and often expressed by the same devices in some languages. For example, in English, tense and aspect markers all take the form of verb suffixes, e.g. the past tense -ed, the past participle -ed and the progressive aspect -ing. Thus, why Chinese did not acquire verb suffixes for these two grammatical categories at the time is a legitimate question. The answer lies in the historical contexts in which the aspect suffixes were created.

The development of the resultative construction directly motivated the emergence of the aspect suffixes. From another angle, it can be said that the aspect suffixes virtually belonged to a subtype of resultative element. Under the influence of the resultative construction, the principle of action–resultative ordering, which governed the information structure of the predicate, came into effect, requiring that only resultative constituents could remain in postverbal position and other non-resultative elements had to occur in preverbal position. By definition, aspect markers refer to the state of the action, whether in progression or in completion, and since all these features can be seen only after conducting the action, they belong to a kind of resultative of the verb. Therefore the related resultative words could occur in postverbal position, where they had the potential to develop into verb aspect suffixes. In contrast, however, tense markers mark the time when the action takes place relative to the time of speaking or other deictic center, which is independent of the action. Thus they are non-resultatives of the verb and can appear only in preverbal position according to the principle of information structure, where they might stand a chance of developing a prefix but not a suffix. In other words, the lexical items for developing aspect suffixes correspond with the words of either the number of repetitions or the duration of an action. As discussed in Section 7.9, words of this type could occur only in postverbal position. Meanwhile, those words that specify the time of an action, including those that express tense information, were limited to preverbal position, which blocked their development into verb suffixes. In Chinese, the information about the time of an action is expressed by time nouns (e.g. zuótiān “yesterday,” qùnián “last year”) and temporal adverbs (e.g. céng-jīng “ever,” cóng-qián “before”), all of which are limited to preverbal position, but verb suffixes must be grammaticalized in postverbal position. This is the real reason why the Chinese language could not acquire tense markings during the development of the resultative construction.

Nevertheless, the Chinese language does have certain grammatical devices for expressing tenses, though they are not verb suffixes. Now, we briefly discuss the linguistic forms expressing futurity, which has been represented by two means throughout history: adverbs and auxiliary verbs. The former could freely occur either before or between the subject and the predicate, but the latter had to appear immediately before the predicate.

Typological investigation reveals that there are twelve lexical items that tend to develop into future markers across languages (for details, see Reference Heine and KutevaHeine and Kuteva 2002: 331), and four of them have been attested in Chinese. Moreover, Contemporary Chinese has contributed a new type of lexical item for developing a future marker, huí-tóu “come back,” which expresses an immediate future in the Pekingese dialect.

1. (a) Verbs denoting the movements of legs. The verbs in this group are the most common lexical sources for developing future-tense markers. Among a sample of more than 500 languages investigated by Reference Bybee, Perkins and PagliucaBybee et al. (1994) and Reference Heine and KutevaHeine and Kuteva (2002), there are fifty-seven languages whose future markers originated from this group of verbs, e.g. be going to in English, referring to the near future. This grammaticalization pathway is based on a common pattern of cognition, mapping a spatial distance onto a temporal distance (Reference LangackerLangacker 1991: 219–220). There were always quite a few verbs of this type that developed into future markers in the history of the Chinese language, as attested in the earliest texts. For example, the verb xíng “go” in Shi Jing was already used as a future marker; (82) illustrates its verb xíng “go” and (83) its use as an auxiliary for expressing futurity:

Around the fifth century BC, another verb, jiāng, which was synonymous with xíng, also developed as a future marker. In Reference Heine and KutevaHeine and Ketuva (2002: 331), it is classified as a verb meaning “take” which developed into a future marker, but this classification is problematic. The verb jiāng was polysemous at the time, meaning “go,” “go along with” or “take,” and it referred mostly to the movements of legs, as illustrated in (84). It was also a common future marker in Old Chinese, as illustrated in (85):

The future marker jiāng has survived to the present day, forming a compound word with the verb lái “come,” jiāng-lái “future,” which is the canonical marker for the distant future in Contemporary Chinese. However, the future marker xíng fell out of use after Old Chinese, although it was still sporadically seen in the compound xíng-jiāng in Medieval Chinese.

The verbs qù “go” and lái “come” can also refer to the immediate future in certain contexts, but they have not developed into a stable device for the expression of the future tense. Interestingly, the verb compound huí-tóu (lit. “return head”) “come back” has recently developed into a future marker in standard Mandarin, especially in Pekingese, in which it is often reduced to huí-er by omitting the noun and coloring the preceding rhyme with a retroflex feature, as illustrated in (86).

1. (b) Verbs of desire and intention. According to Reference Bybee, Perkins and PagliucaBybee et al. (1994) and Reference Heine and KutevaHeine and Kuteva (2002: 331), the second most common lexical sources for future markers are verbs meaning “want” or “desire”; for instance, the auxiliaries will and shall in English belong to this type. Obviously, future markers in a language may originate from different types of lexical item. The verb yù “desire” or “want” became a future marker around the first century AD, and that usage lasted from then to the tenth century AD. About the same time, the verb yào, with the same meaning as yù, developed into a future marker, as illustrated below:

The future markers, which resulted from different lexical sources, accumulated and entered the language at different times. There were three kinds of development fate: some died out, such as the future markers xíng in Old Chinese and yù in Medieval Chinese; some are still preserved in compounds such as jiāng in jiāng-lái. Those that continue to be used tend to manifest a division of labor among them. For example, the form be going to in English expresses immediate futurity, which distinguishes it from will/shall, which refer to intermediate and distant futures. Likewise, yào, huí-tóu, and jiāng-lái form a three-way division of labor in expressing futurity: yào expresses an immediate future, huí-tóu means an intermediate future, and jiāng-lái refers to a distant future. Thus yào can be modified by adverbs such as lìkè “at once” or mǎshàng “immediately,” but the other two cannot; the future yào in (87) is not interchangeable with the other two future tenses:

The finer divisions among future markers in a synchronic system may reflect the increasing complexity of modern life, with too many things that need to be strictly scheduled. The types of grammatical category are to a certain extent influenced by the social activities of the language community. For example, the rich honorific forms in Japanese reflect the life of Japanese people.

13.8 Auxiliary Verbs of Aspect

Finally, we should mention how the Chinese language expressed the meanings of aspect before the emergence of the aspect suffixes. Old Chinese primarily used axillary verbs to express the category of aspect. In Old Chinese, there were three methods to express the function of the perfective aspect. The first method was the possession verb yǒu “have” plus the matrix verb, a very common pattern across languages that is found in hundreds of languages (for details, see Reference Bybee, Perkins and PagliucaBybee et al. 1994, Reference Heine and KutevaHeine and Kuteva 2002: 334), as illustrated in (90) and (91):

In standard Mandarin, this pattern existed from the eleventh century BC to the fourth century AD. Today, in many southeastern dialects, the above pattern is still the predominant form for expressing the perfective aspect. Because of the lack of historical documents, it is unclear whether these dialects preserved the old patterns or came up with them through independent grammaticalization. The apparent cause for the disappearance of this pattern was the rise of the resultative “VOR” pattern around the fourth century AD, where the resultative elements might be certain intransitive elements to denote the completion of the verb with a function similar to that of the perfective aspect. As discussed above, the intransitive resultative liǎo “complete” developed aspect suffix uses at a later stage.

In Old Chinese, the verb jì “finish” also became an auxiliary to express the meaning of the perfective aspect, as illustrated in (92) and (93):

There was also a sentence-final particle, yǐ “already,” that was often combined with the adverb yǐ to denote a change of state, as illustrated in (94) and (95):

Regarding the experiential aspect, in Old Chinese there was an auxiliary cháng, which grammaticalized from the verb “taste,” that expressed an action in the past, as illustrated in (96) and (97) (for details, see Reference Wang and YuzhiWang and Shi 2019). This grammaticalization path is also found in other languages, such as Wolof (an African language) and Ancient Tibetan (Reference DeLancey and Alexandra AikhenvaldDeLancey 2018, Reference Voisin and VittrantVoisin and Vittrant 2012, Reference Kuteva, Heine, Hong, Long, Narrog and RheeKuteva et al. 2019: 365).

After the establishment of the new set of aspect suffixes, the old forms with similar functions, either auxiliaries or adverbs, were not simply abandoned, and some are still preserved in compound adverbs, such as cháng in céng-jīng “once” in Contemporary Chinese (due to phonological reduction, cháng became céng; for a detailed discussion, see Reference Wang and YuzhiWang and Shi 2019).

Regarding the progressive aspect, Old Chinese seemingly lacked a grammatical device to express the function. The preposition yú “in” was extended to take on the function, but it was attested only in the Shi Jing, a text composed around the tenth century BC, as illustrated in (98). The adverb zhèng “just” or “right,” which originated from an adjective meaning “upright,” could indicate an action in progression, as illustrated in (99):

In Contemporary Chinese, these adverbs often co-occur with the aspect suffixes to express certain meanings, forming the following three patterns:

1. (100)

   (a)

   Perfective aspect: yǐ-jīng “already” + V-le

   (b)

   Experiential aspect: céng-jīng “once” + V-guo

   (c)

   Progressive: zhèng-zài “in the process of” + V-zhe

The above adverbs in preverbal position reflect the layer before the first century BC, and the suffixes in postverbal position represent developments since the tenth century AD.

Old Chinese employed auxiliary verbs or adverbs, both preceding the matrix verb, to express the functions of the three aspects, because there was as yet no resultative construction that provided the specific context for the emergence of the aspect suffixes. The Chinese language as an SVO language allows full verbs in the first verb position to become functional words, the path of development from a full verb to an auxiliary or adverb. This is the reason why VO languages always have the constituent order “AUX + V.” The concrete forms to take the aspect function might vary greatly from time to time, but the three types of aspect remained unchanged over time and have always been manifested throughout the history of Chinese: perfective, experiential, and progressive, though the word class changed from auxiliary to verb suffix.