A diachronic study on the Mandarin complex directional complement guòlái ‘come over’ from the macro-event perspective

4.1 Corpus and periodization

The analyses are based on the 2014 online version of the Center for Chinese Linguistics Corpus, Peking University (abbreviated as CCL)^[2] (Zhan et al. 2003, 2019), a widely used corpus of Chinese containing over 700 million characters. Specifically, the data for this study were drawn from its Classical Chinese sub-corpus, which comprises 163.7 million characters, 41.05 % of which are categorized by dynasties. This dataset spans from the Western Zhou dynasty (1046 BCE) to the early Republican period (1920s), covering a total of 15 dynasties. Following the traditional periodization in Chinese linguistics (e.g., Wang 1957; Li 2004; Guo 2013) and the chronological segmentation used in the CCL, this study clusters the 15 dynasties into five historical stages, following the periodization used by Du and Li (2021) (see Table 4). This periodization reflects characteristics of Mandarin’s phonology, grammar and morphology in each stage and aligns with broader societal changes, providing a structured framework for examining its diachronic evolution.

4.2 Data collection

Firstly, to closely examine the diachronic development of guòlái ‘come over’, we focused on how the two independent verbs, guò ‘pass’ and lái ‘come’, compounded to form a single lexical unit that further grammaticalized into a complex directional complement. To ensure a maximally inclusive search, we used the pattern “过$10来” (guò $10 lái), which retrieved sentences where the two target words appeared within a span of fewer than 10 characters. In this way, how the two words were combined and further developed could be observed. Finally, 14,524 occurrences were obtained. Given the relatively large number of instances from the Ming Dynasty (N = 2586), Qing Dynasty (N = 8631), and the Early Republican Period (N = 2752), we randomly selected 1,000 occurrences from each period using Python. The resulting dataset was then manually processed to remove semantically unrelated instances. After this refinement, 2,126 sentences were identified as valid data for further analysis (see Table 5).

Secondly, to investigate the evolutionary order of the macro-events represented by the V guòlái construction, data points of this form were retrieved. Since the corpus data have not yet been segmented or annotated, the present data utilized the query of “过来” (guòlái) and manually processed the extracted occurrences, totaling 7,467.

During manual processing, the syntactic structures of V guòlái with objects proved to be a point of debate. As summarized by Zhang (1991a, 1991b), four syntactic forms of “C_3-11 C_1-2 (= the complex directional complements)” (see Table 1) involving objects have been identified:

A: V C_3-11 C_1-2 O (e.g., ná guòlái tā ‘bring it over’)

B: V C_3-11 O C_1-2 (e.g., ná guò tā lái ‘bring it over’)

C: V O C_3-11 C_1-2 (e.g., ná tā guòlái ‘bring it over’)

D: bǎ O V C_3-11 C_1-2 (e.g., bǎ tā ná guòlái ‘bring it over’)

To ensure the accuracy, we adopt Yang’s (2005) claim that these syntactic structures reflect different cognitive patterns and adhere to the principle of temporal sequence (Dai and Huang 1988). This principle states that the relative order of two syntactic units is determined by the temporal sequence of states they represent in the conceptual domain. In Yang’s analysis, only in the V C_3-11C_1-2 O structure is the complex directional complement fully grammaticalized, as it conceptualizes V C_3-11C_1-2 as a unified whole. For example, in ná guòlái tā ‘bring it over’, ná guòlái ‘bring over’ represents the entire movement process, while tā ‘it’ serves as the theme of this movement. In contrast, in other structures, V and C_3-11 C_1-2 are interpreted as separate cognitive units, with C_3-11C_1-2 functioning as a directional verb rather than a grammaticalized complement. Therefore, these structures are better classified as serial verb constructions, where V and C_3-11C₂ operate as two independent verbs with no grammatical dependency between them. Specifically, V O C_3-11C_1-2 is considered the most typical serial verb construction. In the expression ná tā guòlái ‘bring it over’, the action of ná tā ‘take it’ occurs first, followed by the movement guòlái ‘come over’, making it a clear example of a serial verb construction. Similarly, V C_3-11 O C_1-2 is also classified as a serial verb construction (see Wang [1989]), as V C_3-11 O (ná guò tā ‘take it’) and (O) C_1-2 ([tā] lái ‘it comes’) represent two sequential actions, though this structure is less typical. Finally, the cognitive pattern of bǎ O V C_3-11C_1-2, falling under disposal constructions, is similar to that of V O C_3-11 C_1-2, as exemplified by Guō Yīng bǎ Wáng Dǐng huóqín guòlái ‘Guo Ying has captured Wang Ding alive and brought him over’. In this example, bǎ Wáng Dǐng huóqín ‘captured Wang Ding alive’, which can be paraphrased as the VO order huóqín Wáng Dǐng, represents the initial action, while guòlái ‘come over’ follows as the subsequent event. This sequential structure aligns with the cognitive pattern of V O C_3-11 C_1-2, where C_3-11C_1-2 functions as a directional verb.

Based on the criteria, the present study considers only Form A as valid and excluded data from the other three syntactic structures to avoid ambiguity. Occurrences of “jiāng O V guòlái” were also removed, as this pattern can be seen as an alternate of Form D (Liu and Cuyckens 2023). Moreover, instances of “V dé/bù guòlái” were excluded to prevent ambiguity, as this structure serves a different semantic and functional role from V guòlái. Specifically, V dé/bù guòlái expresses the notion of ability or inability, as in ná dé/bù guòlái ‘have/not have the ability to bring over’, rather than simply denoting directional movement. Finally, 1,406 occurrences remained, as seen in Table 6.

4.3 Data analysis

This study adopts a data-driven approach to explore the diachronic syntactic development of guòlái ‘come over’. It traces the syntactic forms relevant to guòlái in terms of frequency and distribution across five stages. Furthermore, by calculating the occurrence frequency of the macro-event types represented by V guòlái in each stage, this study seeks to identify their evolutionary order.

5.1 The syntactic development of guòlái ‘come over’

This section focuses on the diachronic development of the syntactic structures associated with the varied usages of guòlái ‘come over’ in each stage. Since only two relevant occurrences were found in the first two stages, it is reasonable to argue that guòlái, or even its earliest form, had not yet become entrenched. From Stage III onward, the frequency of occurrences increased, marking the starting point of our analysis. This observation also corresponds to the findings of Li (2019).

5.1.3 Stage IV

This stage accounts for the majority of the data, warranting a thorough examination. According to Chinese Linguistics (Li 2004), the language from this period is considered foundational for modern Chinese. The surge in data can be attributed to the rise of colloquialism during this era. Table 10 presents the frequency and proportion of all syntactic structures, revealing the emergence of 10 additional types compared to Stage III. A higher degree of grammaticalization and entrenchment of guòlái is evident at this stage, as guòlái and V guòlái can accept various intervening linguistic elements, including tense markers such as zhē (the progressive marker) (see Type N) and lē (the perfect marker) (see Type I, M), negation marker bù (see Type K, P, G), and auxiliaries dé (see Type F, J) and jiāng (see Type S).

Table 10:

The distribution of different structures in Stage IV.

Types	Structures	Dynasty	Frequency	Sum	Percentage
O	V guò N lái	Yuan	20	408	27.76 %
		Ming	270
		Qing	118
H	V guòlái	Yuan	10	351	23.88 %
		Ming	150
		Qing	191
E	guòlái	Yuan	9	313	21.29 %
		Ming	62
		Qing	242
L	V N guòlái	Yuan	20	138	9.39 %
		Ming	72
		Qing	46
S	V jiāng guòlái	Yuan	1	58	3.95 %
		Ming	37
		Qing	20
Q	bǎ N V guòlái	Yuan	0	51	3.47 %
		Ming	10
		Qing	41
I	V le guòlái	Yuan	0	30	2.04 %
		Ming	7
		Qing	23
B	guò N (ér) lái	Yuan	2	23	1.56 %
		Ming	17
		Qing	4
D	guò (N)…lái V/N	Yuan	1	16	1.09 %
		Ming	9
		Qing	6
M	V le N guòlái	Yuan	0	15	1.02 %
		Ming	4
		Qing	11
K	V bù guòlái	Yuan	0	14	0.95 %
		Ming	6
		Qing	8
C	guò N lái (N) V	Yuan	0	12	0.82 %
		Ming	8
		Qing	4
R	jiāng N V guòlái	Yuan	0	8	0.54 %
		Ming	3
		Qing	5
F	guò dé (N) lái	Yuan	0	7	0.48 %
		Ming	6
		Qing	1
J	V dé guòlái	Yuan	1	6	0.41 %
		Ming	1
		Qing	4
N	V zhe N guòlái	Yuan	2	6	0.41 %
		Ming	3
		Qing	1
A	guò (N)…(bù) lái V/N	Yuan	0	5	0.34 %
		Ming	3
		Qing	2
T	guòlái dè/(zhī)rén	Yuan	2	5	0.34 %
		Ming	1
		Qing	2
P	V bù guò N lái	Yuan	0	3	0.20 %
		Ming	0
		Qing	3
G	guò bù lái	Yuan	0	1	0.07 %
		Ming	0
		Qing	1
				1470	100.00 %

To better illustrate this trend, the frequency and distribution of data at Stage IV are visualized in Figure 1.

Figure 1:

The frequency and cumulative proportion of each type in Stage IV.

According to Figure 1, three types – Type O, V guò N lái (cf. [4]), Type H, V guòlái (cf. [3]), and Type E, guòlái (cf. [2]) – exhibit higher frequencies, collectively accounting for nearly 75 % of the data in this stage. Notably, V guò N lái emerges as the most prevalent usage during this period, reaching its peak in the Ming Dynasty. As discussed in Section 4.2, this form is considered one of the derivations of V guòlái that involves the object, further supporting the higher syntactic tightness between guò and lái. For example, in (4), qǔ guò shū represents Event 1, denoting the action of “taking the book”, where the displacement applies to the book rather than the agent. Event 2, marked by lái, represents the agent’s movement of “carrying the book to the reference point”. Rather than being expressed in two separate clauses, these two events are integrated into a single clause, validating the syntactic continuum proposed in the Macro-event Hypothesis.

(4)

Jīng	Gōng	qǔ	guò	shū	lái.
Jing	Gong	take	come	book	over
‘Jing Gong took the book over.’
(Yuan Dai Hua Ben Xuan Ji, the Yuan Dynasties)

Unlike V guò N lái, both V guòlái and guòlái exhibit a consistent upward trend over time, reflecting their increasing entrenchment and conventionalization. Furthermore, the proportion of V guòlái surpasses that of guòlái, indicating the gradual establishment of guòlái as a complex directional complement rather than as an independent verb.

5.1.4 Stage V

Although this stage covers a relatively short period, the data closely resemble modern Chinese. In addition, the dataset contains over five hundred instances, suggesting the widespread use of guòlái. The frequency and distribution of each syntactic structure are shown in Table 11 and visualized in Figure 2.

Table 11:

The distribution of different structures in Stage V.

Types	Structures	Frequency	Percentage
H	V guòlái	204	36.17 %
O	V guò N lái	100	17.73 %
E	guòlái	78	13.83 %
L	V N guòlái	47	8.33 %
I	V le guòlái	42	7.45 %
S	V jiāng guòlái	19	3.37 %
Q	bǎ N V guòlái	15	2.66 %
K	V bù guòlái	10	1.77 %
R	jiāng N V guòlái	10	1.77 %
D	V guò (N)…lái V N	9	1.60 %
M	V le N guòlái	8	1.42 %
T	guòlái dè/rén (adjective)	8	1.42 %
P	V bù guò N lái	5	0.89 %
A	guò (N)…(bù) lái V/N	3	0.53 %
N	V zhe N guòlái	3	0.53 %
B	guò N (ér) lái	2	0.35 %
F	guò dé (N) lái	1	0.18 %
		564	100.00 %

Figure 2:

The frequency and cumulative proportion of each type in Stage V.

Drawing insights from the data, it is strikingly evident that Type H, V guòlái, now dominates the overall distribution. Its proportion has steadily increased over time, rising from 12.36 % to 23.88 % and then to 36.17 %, further reinforcing the higher degree of grammaticalization of guòlái as a complex directional complement. In contrast, Type E, guòlái, exhibits the opposite trend, with its proportion decreasing from Stage III to V (28.09 % to 21.29 % to 13.83 %), indicating a shift in guòlái’s function from an independent verb to a complement. Remarkably, Type B, from which guòlái originated, now appears in only two instances, which supports the hypothesis that its syntactic evolution has progressed from separate clauses to a single clause rather than the reverse. Finally, with its increased degree of grammaticalization, guòlái exhibits greater flexibility and versatility in usage. Particularly noteworthy is the prevalence of object-involving constructions derived from V guòlái, such as Type O and Type L, as evidenced by their increasing frequency and proportion.

5.2 The semantic development of the V guòlái construction: the evolutionary order of the macro-events

This section delves deeper into the development of the V guòlái construction from the macro-event perspective. While its structural composition has remained stable since its emergence, the onset of grammaticalization marked a notable expansion in its semantic scope. Over time, V guòlái began to convey a broader range of meanings, initiating its grammaticalization journey. These semantic developments can be roughly categorized into five types of the macro-events.

By tracing the evolutionary sequence of the macro-events represented by this construction, this study aims to provide a cognitive framework that contributes, to some extent, to our understanding of the evolution of Mandarin. Given the absence of valid data from Stage I and II, we tentatively posit the emergence of guòlái as a directional complement at Stage III. This finding aligns with previous research in Chinese linguistics such as Wang (2005: 186), which identifies the Song Dynasty as the period of its emergence.

5.2.2 Stage IV

During this stage, the usage of the V guòlái construction surged. The corresponding distribution is presented in Table 13, with a visual representation in Figure 3.

Table 13:

The distribution of the macro-event types in Stage IV.

Types	Yuan	Ming	Qing	Sum
Motion	8	325	307	640
Temporal	0	0	0	0
State Change	1	16	98	115
Correlating	0	0	4	4
Realization	1	1	6	8
Others	0	0	7	7
Total	10	342	422	774

Figure 3:

The comparison of distribution in each dynasty in Stage IV.

It is noteworthy that the macro-event types represented by V guòlái became increasingly diverse over time, as illustrated in Figure 3 by the growing variety of colors, confirming its expanding semantic scope. In addition, this construction started to represent events that do not fit into the established five macro-event types, a trend particularly evident during the Qing dynasty. These instances have been collectively categorized as “Others”, encompassing expressions such as huíxiǎng guòlái ‘reminiscing about something’, fǎn guòlái ‘in turn’. A more detailed discussion of these cases is provided in Section 6.1.

Despite this diversification, Motion Events remain the most prevalent category across all dynasties, which aligns with the inherent directional nature of V guòlái. Notably, the Event of State Change (cf. [6]) has witnessed a consistent rise in proportion, as illustrated by the expanding grey segments in Figure 3. This trend suggests that guòlái has developed the capacity to denote changes beyond mere physical displacement, such as marking a transition from the state of being asleep to being awake, as illustrated in (6). It is pertinent to observe that, as of this analysis, the Temporal Contouring Events have yet to appear in the dataset.

(6)

Tiān	Shī	xǐng	guòlái.
Tian	Shi	awake	come-over
‘Tian Shi woke up.’
(San Bao Tai Jian Xi Yang Ji, the Ming Dynasty)

6.1 The syntactic development of guòlái ‘come over’

From a syntactic perspective, the findings indicate that guòlái, originally conceived as a motion verb, derives from the pattern guò N (ér) lái, where guò and lái functioned as two separate motion verbs. Over time, it underwent grammaticalization, ultimately becoming a directional complement in the form of V guòlái. With a higher degree of grammaticalization, V guòlái has developed the ability to co-occur with various intervening elements, including the Chinese tense marker zhē and lē, the negation marker bù, and the auxiliaries dé and jiāng. It also allows for intervening objects, as seen in structures such as V N guòlái and V guò N lái. Furthermore, guòlái can appear in disposal constructions, including those introduced by bǎ and jiāng.

The syntactic development of guòlái aligns with the syntactic continuum observed in the transition from a double-clause to a single-clause representation. This progression supports the grammaticalization continuum proposed in the Macro-event Hypothesis (Li 2023: 139). For instance, during Stage III, guò and lái functioned as two distinct independent verbs (cf. [1]). By Stage IV, the combined usage of guòlái had significantly increased, predominantly conveying the meaning of “come over”. Importantly, the concept of “passing through a location and arriving at a reference point” became encapsulated within a single clause through the use of guòlái, as exemplified in (7):

Ultimately, when combined with another verb signifying the manner of movement, guòlái undergoes grammaticalization, transforming into a directional complement, as illustrated in (8):

The examples above demonstrate that, during the development of V guòlái, two distinct events – “passing by a place” and “moving closer to the reference point” – have been diachronically conceptualized as a single, integrated event: “come over.” This development is further reinforced by the incorporation of manner verbs, all encapsulated within a single clause. This observation aligns with the syntactic continuum proposed in the Macro-event Hypothesis.

Drawing upon the developmental sequence of guòlái, a broader schema for the syntactic evolution of all complex directional complements can be proposed (cf. Figure 4), providing support for the syntactic and grammaticalization continua outlined in the Macro-event Hypothesis.

Figure 4:

The syntactic evolutionary sequence of Mandarin complex directional complements.

6.2 The semantic evolution of V guòlái construction

Concerning the macro-event types represented by V guòlái over time, this study hypothesizes its evolutionary order as Motion Event > Event of Realization > Event of State Change / Event of Action Correlating (> Event of Temporal Contouring). The sequence corresponds partially to Zhang (1991b), who posits that the “directional meaning” (Motion Event) of complex directional complements has bleached to resultant meaning (Event of Realization), and to state meaning (Event of State Change) throughout the process of linguistic evolution (p.185). However, it is slightly different from that raised by Li (2019), which illustrates the development path as Motion Event > Event of State Change > Event of Temporal Contouring > Event of Realization by examining the 11 simplex directional complements. The differences can be explained as follows. Firstly, the present study only explores one complex directional complement, guòlái, and the final development process may largely result from its own unique characteristics. Secondly, presenting the Event of Realization is the second step of the evolution, which may be due to the meaning of accomplishment expressed by guò ‘pass’. In addition, since data is limited, it is hard to decide the appearance order of the Event of State Change and the Event of Action Correlating. Therefore, the hypothesized sequence is only tentative and needs support from future studies.

Remarkably, the Event of Temporal Contouring did not appear throughout the whole process. Even though we identified relevant occurrences in the contemporary Chinese corpus included by CCL that are beyond the scope of our discussion, these instances are few in comparison to other types of macro-events and they are not typical examples compared to those proposed by Talmy (2000b). Most of them denote a Temporal Contouring Event with a specific focus on the progression from the past to the present, rather than from the present to the future. This can be explained by the basic directional meaning of guòlái. Liu (1998) argues that guòlái indicates an object moving closer to the standpoint of the speaker. Zhang (2010) and Wang (2005) also demonstrated that guòlái expresses the direction of the move, i.e., to the speaker. Therefore, metaphorically, it refers to a target moving from the past to now (the speaker), rather than from now to the future, such as in xiě guòlái “have/has been written”. It is possible that the event of “write” continues, but it is not the pivot. However, the activating process of the Event of Temporal Contouring emphasizes the target’s “progression through time” (Talmy 2000b: 231), which is in nature opposite to the event presented by V guòlái. Thus, Talmy’s Temporal Contouring Event should be revised to explain this language phenomenon.

Furthermore, there are instances that cannot be categorized into the five types of macro-events. For example, the above example of xiǎng guòlái ‘reminiscing about sth.’ or huíxiǎng guòlái ‘reminiscing about’, as well as fǎn guòlái ‘in turn’ or dào guòlái ‘in turn’ in its idiomatic usage, meaning ‘reversely’ or ‘in contrary’, as shown by (9). They cannot be classified into any discernible type, which impedes the refinement of the existing macro-event types.

Moreover, through data annotation, it is doubtful that whether the core schema of the Event of State Change is expressed by the verbs or by the satellite guòlái. For example, (1) verbs indicating ‘awake’, such as xǐng, jīngxǐng, qīngxǐng; (2) verbs indicating ‘understand’ including wù, juéwù, míngbái, lǐjǐe; (3) verbs referring to ‘change’ and ‘reverse’, such as biàn, fǎn, dào; (4) verbs indicating ‘marry’ consisting of jià and qǔ. In Chinese, these verbs inherently convey the notion of change from one stage to another, which aligns with Talmy’s (2000b: 238) argument regarding the combination of transition types with states. Consequently, the core schema of the Event of State Change is possibly encoded within the verb itself rather than the satellite. However, Talmy believed that Mandarin is a satellite-framed language, with the core schema encoded by the satellite guòlái. This presents a paradox, as he also argues that the core schema of the State-Change Events is represented by the main verb in verb-framed languages but by satellites in satellite-framed ones. The specific location of the core schema in this type of construction remains an issue for future study.

Similarly, the Event of Action Correlating faces a similar dilemma, as the only collocated verb gēn ‘follow’ in Chinese inherently implies two participants. Therefore, it becomes evident once more that Mandarin Chinese does not adhere strictly to the characteristics of a typical satellite-framed language, which contributes valuable insights to language typology debates.

6.3 A qualitative examination on the macro-event types represented by the 28 Mandarin directional complements

The meticulous examination of the diachronic development of guòlái ‘come over’ led us to question whether all Mandarin directional complements undergo a similar developmental process to represent five types of the macro-events. To address this inquiry, this research proceeded by qualitatively analyzing the diachronic data of all 28 directional complements (cf. Table 1). Data retrieval employed the same corpus and periodization as that of guòlái, albeit not using a systematic quantitative statistical method. Rather, once we identified more than one occurrence encoding a specific macro-event type, we assumed the directional complement construction’s capability to represent that type of the macro-event. This approach aligns with Flach’s (2021) perspective that once textual evidence of a certain construction is discovered in corpora, it is reasonable to assume its conventional usage in publishing or correspondence, even within a limited subsection of the speech community. Finally, a tentative development path of all 28 Mandarin directional complements is delineated and visualized in Figure 5. The color blocks, categorized into five types, correspond to five distinct types of the macro-event. The appearance of a color block at a stage signifies the complements’ capacity to represent that macro-event type.

Figure 5:

The development of the macro-event types represented by the 28 Mandarin directional complement constructions.

From Figure 5, several conclusions can be drawn:

1. Mandarin complex directional complement constructions emerge significantly later, primarily surfacing at Stage III (618–1279), whereas most simplex directional complement constructions appear predominantly in Stage I (1046 BCE-220). Exceptions include huí ‘back’ and kāi ‘away’. This finding aligns with broader research in Chinese linguistics (e.g., Liang 2007; Song 2012).

2. Directional complement constructions involving kāi ‘away’ and kāilái ‘away’ initially do not represent Motion Events, possibly due to the distinct semantics of kāi, as discussed in Wang (2005), Wang and Gong (2015).

3. Generally, the evolutionary sequence of the macro-event types represented by all directional complement constructions can be summarized as follows: Motion Event > Event of Realization/State Change > Event of Temporal Contouring. The Event of Action Correlating is excluded, since it tends to be encoded within the semantics of the main verb rather than in the directional complements (see Section 6.1). Moreover, since complex directional complements consist of a simplex directional complement (i.e. C₃ – C₁₁) combined with lái ‘come’ or qù ‘go’ to indicate reference points, their meanings are more specific and therefore more constrained than those of simplex complements. Consequently, most complex directional complements lack the ability to represent the Event of Temporal Contouring, further validating Liang’s (2007) findings.

This developmental process not only corresponds to the grammaticalization cline across the five types of macro-events proposed by Li (2018) but also to the semantic evolution path proposed by Liang (2007) based on a qualitative analysis of shàng ‘up’, xià ‘down’, and qǐlái ‘get up’. Moreover, it aligns with the widely acknowledged evolution of human cognition, which transitions from spatial to temporal concepts, paralleling the evolution of human language. For instance, Lakoff and Johnson (2003) [1980] argue that temporal metaphors in languages often originate from spatial metaphors. Similarly, Evans (2004: 5) suggests that humans commonly conceptualize and discuss time in terms of motion through and location within three-dimensional space, thereby attributing the emergence of temporal concepts to spatial ones. Heine et al. (1991) proposed a general hierarchical grammatical structure: person > object > activity > space > time > quality. Huang and Heish (2008) further refined this framework by introducing the concept of “non-specific space” between “space” and “time”, highlighting the need for a more nuanced categorization. However, these theories may oversimplify the relationship between time and space, treating “time” as a monolithic category rather than recognizing its internal complexity. The evolutionary order of macro-event types represented by Mandarin directional complements suggests that the notion of “time” encompasses multiple dimensions, including culmination and continuation, rather than existing as a singular entity.

Specifically, the evolution follows a continuum from events rooted in spatial configurations to events extending into the future, with events culminating in the present serving as the transitional phase. This transition reflects a gradual shift from concrete spatial relations to temporally abstract meanings, where the boundaries between past, present, and future are relative to the discourse context. Figure 6 illustrates the underlying logic of this continuum, where S represents the starting point, and G denotes the goal of the conceptual trajectory.

Figure 6:

The cognitive explanation for the development of the macro-event types represented by the 28 Mandarin directional complement constructions.

Motion Events are the most tangible, as they occur within the spatial dimension and can therefore be initially represented by most directional constructions. Over time, Events of Realization and State Change begin to incorporate the temporal dimension, as they denote event progression along time, especially emphasizing the endpoint of an action or a change-of-state event. Specifically, Events of Realization highlight the accomplishments of actions at present, whereas Events of State Change mark the endpoints of preceding actions or states and the inception of new ones. However, both only encode culmination, rather than temporal continuation. This tendency may explain why Events of State Change typically follow Events of Realization in most directional complements. Finally, Events of Temporal Contouring, being the most abstract, focus on the continuation of actions or states from past or present into the future. Many directional complement constructions, particularly complex ones including guòlái in this study, struggle to represent this type of macro-events due to the constraints imposed by lái ‘come’ and qù ‘go’ in the second complement position. In conclusion, Mandarin directional complements witness the development of indicating concepts of space > culmination > continuation. This pattern aligns with the semantic extension path of Mandarin directional complements proposed by Chinese scholars such as Liu (1998), Liang (2007), Song (2007), and Wang and Gong (2018). Future studies can further substantiate this sequence through corpus-based quantitative analysis.

Overall, drawing on the syntactic and semantic development of the entire closed set of Mandarin directional complements, this study proposes the Unidirectional Progression Hypothesis: the closed set of Mandarin directional complements has evolved simultaneously in both form and meaning along a unidirectional trajectory. Syntactically, this evolution has moved toward greater structural compactness – shifting from two clauses to a single clause, with increasing degree of grammaticalization within the single-clause structure. Semantically, it has progressed from concrete to more abstract meanings, starting from two independent motion events that merged into a single macro-event of motion, subsequently expanding to other macro-event types, representing culmination and continuation in the temporal dimension. Both trends follow a unidirectional path, with later stages exhibiting stronger syntactic integration and a more abstract semantic relationship between the main verb and the directional complement.^[3]