The Windowing and Shifting of Attention of Motion Verbs in the Path Event Frame: A Corpus-Based Contrastive Study between 去 (qù, go) and “go” in Chinese and English ()
1. Introduction
去 (qù, go) and “go” as a pair of typical path verbs, imply clear and discernible paths in their meanings (Deng, 2018), which have always received extensive attention and heated discussion (Wen, 2007; Fan, 2014; Zhao, 2024). The initial studies try to make a comparison of the commonalities and discrepancies between the following two groups of verbs, 来 (lái, come) and “come”, 去 (qù, go) and “go” from aspects like tense, aspect and mood (Gao, 2008), part of speech, word meaning and collocation (Jiang, 2008), and semantic and syntactical meaning (He & Deng, 2019). Chen (2016) constructs a Windowing-Prominence-Metaphor (WPM) model applying the Windowing of Attention Theory as one of its foundations, based on which he conducts a quantitative and qualitative analysis of the modes of windowing and cognitive motivations of 10 high-frequency motion verbs in English. Zhao (2024) conducts a comparative study of the path types and modes of windowing of 来 and “come” in the path event frame, and also analyzes the cognitive motivations behind these phenomena. Most of the studies above focus on qualitative comparative analyses of the semantic, syntactic, and pragmatic dimensions of motion verbs in English and Chinese (Goddard, 1997; Wen, 2007; Liu, 2011; Fan, 2014), but a few studies apply a corpus-based method to conduct a cross-linguistic empirical study on motion verbs from the perspective of the Windowing of Attention Theory (Nicolle, 2009; Cai & Shen, 2018; Ding, 2021; Zhang, 2023). Although individual studies have summarized factors coercing attention windowing in general (Hu, 2019), not so many have focused on a specific pair of motion verbs in the context of authentic language data to make a summary of the factors adjusting attention windowing.
Discrete linguistic segmentation of the continuous reality brings about Talmy’s (2000a) instructive classification of 5 different types of event frames: the path event frame, the causal-chain event frame, the cycle event frame, the participant-interaction event frame, and the interrelationship event frame. Among them, the path event frame is the most typical event frame that characterizes the path along which an entity moves, and it is also the one with the most clearly observable structure (Deng, 2018). Moreover, the path event frame involves both the classification of physical motions and fictive motions of path verbs, which is more comprehensive to examine the path verbs represented in both physical space and mental space. The Windowing of Attention theory is constructed based on the path event frame to deal with different situations of windowing. Therefore, this study concerns itself with a corpus-based endeavor to detect 去 (qù, go) and “go” under the rubric of the path event frame and the Windowing of Attention Theory, and the study focuses on the questions below: 1) What are the path selections of 去 (qù, go) and “go”? 2) What is the distribution of modes of attention windowing? 3) What are the factors shifting attention windowing? By discussing the three questions above, the research is expected to reveal the universal cognitive pattern behind 去 (qù, go) and “go”, validate the Windowing of Attention Theory across two languages, and complement the introspective research with a corpus-based method.
2. Theoretical Background
2.1. The Windowing of Attention Theory
The cognitive process of the windowing of attention is set forth by Leonard Talmy (2000a) to account for the distribution of attention in languages. In this process, some portions of a coherent referent situation are picked out into the foreground of attention by specification in languages (windowed portions) while the remainder of it is ensconced in the background of attention through omission in linguistic expressions (gapped portions) (Talmy, 2010). Divided by different windowed portions, the windowing of attention falls into 3 modes, namely the Initial windowing (abbreviated as I), the Medial windowing (M), and the Final windowing (F).
The overall course and elements of motion involved can be illustrated in Figure 1. The Ground is an indispensable background in which the whole event can happen. The arc curving like a parabola denotes the path along which the figure in motion selects different modes of windowing of attention. Three squares respectively situated in 3 end points indicate 3 single modes of windowing of attention. In addition, mode complexes, the combination of different single modes of windowing of attention, are ubiquitous in linguistic expressions. By different positions and strength of attention, there appear to be 7 disparate patterns of windowing of attention in total (Talmy, 2006), which are abbreviated as I, M, F, I + M, I + F, M + F, and I + M + F.
Figure 1. Windowing schema (Ungerer & Schmid, 2008: p. 222).
2.2. The Path Event Frame
The cognitive process mentioned above should take place in a limited environment where language users can distinguish what is the central identifying core in it and what lies peripheral or outside of the environment (Talmy, 2010). The so-called environment demarcated here is defined as an event frame by Talmy (2000a). Based on diverse referent situations, he delimits 5 generic types of event frames, namely the path event frame, the causal-chain event frame, the cycle event frame, the participant-interaction event frame, and the interrelationship event frame (Talmy, 2000a). Consequently, 5 types of windowing of attention are yielded with respect to the 5 types of event frames, namely path windowing, causal-chain windowing, phase windowing, participant-interaction windowing, and interrelationship windowing. Among them, the path event frame contains the most clear and discernible structure in which we can detect the correlations between formal structures in language and the relevant configuration in perception (Talmy, 2000b).
The windowing of attention process can be extended in terms of the path event frame; hence it falls into 3 categories—open path, closed path, and fictive path (Talmy, 2000a).
An open path here constructs a path of motion that can be conceptualized as integrity by cognitive subjects. To be identified as an open path, two conditions must be met, one of which is that the path along which an object is moving is supposed to possess a starting point and an ending point though the two points do not necessarily appear together in linguistic expressions, another of which is that the two points have to be situated in different spatial locations. Talmy’s (2000a) example The crate that was in the aircraft’s cargo bay fell out of the plane through the air into the ocean exhibits maximal windowing of the whole path, which windows the initial, medial, and final portion in the path.
A closed path has a similar connotation to the open path but with only one difference that the starting point and the ending point of a closed path coincide so that the entire path finally comes full circle. Talmy’s (2000a) example Go get it out of the refrigerator and bring it here respectively foregrounds the departure, the away, and the return portion of the path, which is a typical demonstration of full windowing in a closed path.
The fictive path involves the rendering of the nonveridical phenomenon abounding in both the visual field and linguistics forms. It is concerned with the motion of attention focus of cognitive subjects but there is no physical occurrence in reality (Talmy, 2010). A sentence like This fence goes from the plateau to the valley (Talmy, 2000a) portrays a motion with no physical occurrence but depicts a fictive path along which the focus of attention moves. The patterns of windowing of attention in a fictive path resemble the corresponding patterns in the open and closed path. Fictive motion embraces 6 types in Talmy’s (2000a) classification, which are emanation, pattern paths, frame-relative motion, advent paths, access paths, and coextension paths.
3. Research Design
3.1. Source and Reasons for Corpus Selection
Authenticity and naturalness feature highly in the spoken corpus, for its production relies more on specific cultural backgrounds and communication scenarios (Liu & Lin, 2018). The panorama of a certain language can be revealed to the full extent in spoken corpora than in written corpora. Talmy (2000b) also argues that the choice of a corpus should be more inclined to the spoken language.
The Chinese corpora applied in this research is a specialized spoken corpus designed by a team from Beijing Foreign Studies University led by professor Xu Jiajin, which is called Diversified Spoken Chinese Uttered in Social Setting (acronym as DiSCUSS corpus). The DiSCUSS corpus contains 300 texts and 1,002,538 words, which is broadly divided into two categories: dialogue and monologue, and further into four subcategories: conversation, public discussion, spontaneous, and prepared (Sun, 2022). Different categories represent diverse speaking scenarios or contexts. The DiSCUSS corpus is well represented and balanced so it is a proper choice for this study.
The English corpus adopted in this research is the Corpus of Contemporary American English (acronym as COCA). COCA is a contemporary English corpus developed by scholars at Brigham Young University led by Mark Davis, which is primarily designed for language learners and linguists to investigate the frequency of words, phrases, and sentence structures. With 130 million words, the spoken part of the COCA corpus (as of March 2020) embraces one of the largest spoken corpora at present with millions of language materials in multiple genres.
3.2. Research Method and Design
In general, this piece of research combines the strategies of quantitative research and qualitative research to have an umbrella inspection of 去 (qù, go) and “go” from the perspective of Cognitive Semantics and to testify to the universality of the Windowing of Attention Theory across the two languages. A corpus-based approach is applied in this study to complement the introspective research lacking evidence from authentic language data. Cognitive Linguistics embraces the usage-based research diagram of language and agrees that language knowledge derives from language usage (Langacker, 1987; Wen & Yang, 2024). Therefore, the study will be more cogent if it is supported by large quantities of language data obtained from authentic contexts.
The incipient stage of the research is to collect corpora of 去 (qù, go) and “go”. The operation in the DiSCUSS corpus is simple because it only involves the input of 去 (qù, go) in the searching box and pressing “start query” to find all the related contexts of 去 (qù, go). Comparatively, the search for “go” in COCA requires a special form “[go]” in the searching box to find out all the possible inflectional forms. The next step is to tick the spoken section below the search box and press “find matching strings” to get all results. The DiSCUSS corpus yields 3807 pieces of language data with 去 (qù, go) while COCA displays 8 distinct forms of “go”, namely “going”, “go”, “went”, “goes”, “gone”, “gonna”, “goin” and “gos” with 780,672 hits in total.
The preliminary search results need to be filtered to conform to the purpose of this study. The central tenet for filtration is that both 去 (qù, go) and “go” should be used as verbs denoting a kind of motion or indicating a path in their connotations. Specifically, the following cases are excluded from the study. The “remove” meaning of 去 (qù, go) is abandoned because it deviates at all from the typical path meaning of 去 (qù, go). In addition, 去 (qù, go) when appearing as a dirty word in 我去 (wǒ qù, damn) in spoken language will not be included in this study because it completely loses the motion meaning. Moreover, 去 (qù, go) used in voice-over to explain the background information is eliminated as well. As for English, the mere case of “go” used as a noun and “going” in the “be going to” construction are not qualified for they do not function as verbs, which are incompatible with the purpose of this study. The above mismatching cases are removed in both English and Chinese corpora to stick to the purpose of this research. For the convenience and clarity of analysis, following the filtration of linguistic data is the operation of drawing 300 samples in the two corpora because the amount of linguistic data between the two corpora is extraordinarily unbalanced.
Secondly, according to Talmy’s (2000a) classification of path types in the path event frame, the linguistic data after filtration and sampling will be annotated and distributed into different categories. Furthermore, based on the 7 different windowing modes drawn out from the Windowing of Attention Theory, different windowing modes of English and Chinese path verbs are sorted out.
The last step is to calculate the amount of each type of information under different frameworks and tabulate what has been calculated.
4. Results
4.1. Detection of Path Selections of 去 (qù, go) and “go”
Table 1. The path selections of去 (qù, go) and “go”.
Path types |
去 (qù, go) |
go |
|
Frequency |
% |
Frequency |
% |
LL score |
Sig. level |
Open path |
248 |
82.6 |
244 |
81.3 |
44.37 |
p < 0.0001 |
Closed path |
34 |
11.3 |
4 |
1.3 |
|
|
Fictive path |
18 |
6 |
52 |
17.3 |
|
|
Total |
300 |
100 |
300 |
100 |
|
|
Table 1 shows the distribution of path types of 去 (qù, go) and “go” in the two languages. Statistically, the difference in the distribution of path types between the DiSCUSS corpus and COCA is significant (chi-square = 40.23, p < 0.0001) with a small effect (Cramér’s V = 0.2598). For 去 (qù, go) alone in Chinese, the open path is the most common type in samples extracted from the DiSCUSS corpus which accounts for 82.6%, while the closed path and fictive path take up a relatively smaller proportion in the overall Chinese data with the number of closed paths narrowly outstripping the amount of fictive path. Similarly, “go” in English data shows the highest percentage 81.3% in the open path in COCA samples, but the status of the closed path and the fictive path is inverted compared to the counterpart in Chinese data with the number of fictive paths (52 out of 300) far exceeding the amount of closed path (4 out of 300).
From the horizontal point of view across the two languages, both 去 (qù, go) and “go” present a tendency toward the open path in the two corpora with 去 (qù, go) accounting for 82.6% in the DiSCUSS corpus and “go” for 81.3% in COCA.
(4a) 我 说 我 前 面 的 主 任 已 经 调 到 台 北 去了。
wǒ shuō wǒ qián miàn de zhǔ rèn yǐ jīng diào dào tái běi qù le.
I said my former DE director already went to Taipei go PRF.
(“I said the former director had gone to Taipei.”)
(4b) It was a straight-faced report from the FBI agent, saying they’d followed me to Philadelphia and I’d driven to South Broad Street and I had gone into this three-story building…
The open path depicts a conceptually panoramic unity which has a starting point and an ending point and the two points occupy different spatial locations. Amid the conversation in example (4a), the information about the former director’s occupation mobility is revealed. The speaker’s words portray a path of the former director from an unclear starting point to a definite destination Taipei, which suggests that the two points are in different spots. Though the starting point of the mobility is not instantiated, the listener can still infer from the contexts about it with no effort and conceptualize the whole open path. Therefore, the establishment of an open path does not necessarily involve the foregrounding of all the information to listeners.
Similarly, “go” in example (4b) directs the attention of listeners along the open path it depicts. In an interview about his experience of being followed and bugged, the speaker narrates the motion from Philadelphia along South Broad Street to a three-story building which is regarded as the destination of the path. Even though the starting point where the speaker takes off is unknown, listeners unconsciously generate a fuzzy starting point which is not instantiated in language. Therefore, the identification of an open path does not necessarily require all the information to be placed in the foreground. The interaction can proceed fluently when the important information is foregrounded while known-to-all knowledge and peripheral messages are backgrounded.
Aside from open path, 去 (qù, go) and “go” have diverse preferences for closed path and fictive path. The frequency of 去 (qù, go) indicating a closed path (34 out of 300) surpasses the amount of closed path instantiated by “go” from COCA (4 out of 300). Conversely, “go” in COCA shows an inclination for the fictive path and the percentage of it (fictive path of “go” = 17.3%) is much higher than the counterpart in Chinese data of 去 (qù, go) (fictive path of 去 (qù, go) = 6%).
Chinese is more inclined to portray a closed path using 去 (qù, go) compared with English using “go”. 去 (qù, go) and the verbs following after it often stick together as a construction so it is easy for Chinese to produce the meaning of closing a path in constructions like 去接回来 (qù jiē huí lái, go pick it back) or 去拿回来 (qù ná huí lái, go bring it here) where simple combination of several linguistic chunks can be at work to form a closed path while in English, detailed description involving at least where to go and come along the path can be effective to hint at a closed path.
(5a) 儿子！去 给 饼 干 拿 过 来 给 阿姨 吃！
ér zi! qù gěi bǐng gān ná guò lái gěi ā yí chī!
Son! Go give cookies bring here give auntie eat!
(“Son! Go get the cookies here for auntie!”)
(5b) He forms the intent to kidnap Riley. He leaves the house to go get his car, comes back into the house…
The closed path differs from the open path in that finally the starting point and the ending point of the path are situated at the same spatial locations, that is, the moving entity experiences a path A→B→A. The conversation in example (5a) happens at home between a mom and her son entertaining their guest with cookies. The son is asked to go to bring the cookies and get back to the starting point. The starting point and the ending point occupy the same location so the son’s path to bring the cookies back forms a loop that can be conceptualized as a closed path in whole. In English, the example of “go” involves a detailed and clear description of the path. According to the context, the man leaves the house, goes get his car, and finally comes back to the house. From his motion track, it can be discovered that the starting point and the ending point coincide, so the path the man moves along comes full circle and can be identified as a typical closed path.
Compared from the perspective of the fictive path, English and Chinese have different preferences according to statistics in Table 1 as mentioned above. The amount of 去 (qù, go) leading a fictive path is listed at the last place in the overall path types in Chinese data while “go” in English by contrast is more apt to depict fictive paths.
(6a) 他 的 情 绪 又 向 抑 郁 走 去。
tā de qíng xù yòu xiàng yì yù zǒu qù.
his DE emotion again to depression walk go.
(“He has gone for depression again.”)
(6b) Well, the highway going south, the Seward Highway, there’s been a number of rock slides. So that highway is closed.
Example (6a) shows the speaker’s introduction to a famous composer’s life story. It is in our common sense that emotion is an abstract and stationary representation in our mind so it does not possess the ability to move. The literal meaning of (6b) ascribes motion to the normally believed stationary emotion. The emotion or depression is a factive existence though it cannot be touched while the motion ascribed by the path verb “go” makes it more fictive. The representation of an entity contains both factive parts and fictive parts. The production of a fictive path is the integration of factive existence and factive representation.
In English, the motion verb “going” in (6b) indicates a metaphorical motion (Shan & Aunga, 2018) of the highway and thus evokes out a fictive path of it. What is factive here is the actual existence of the stationary highway whereas the fictive part here is the moving of the highway along the imagined whole path. The fictive path directed by “going” further belongs to coextension paths. It depicts the “form, orientation, or location of a spatially extended object in terms of a path over the object’s extent” (Talmy, 2000a). The highway in example (6b) is imagined as a dynamic route extending from a certain point to the south along which the attention focuses of speakers and listeners follow to proceed.
4.2. Detection of Windowing Modes of 去 (qù, go) and “go”
Table 2 provides a statistical overview of the distribution of path windowing for 去 (qù, go) and “go”. In general, both 去 (qù, go) and “go” are prone to windowing over the final portion of the path respectively accounting for 81.33% and 79.67%, which implies that English and Chinese tend to give prominence to destinations of the whole path. 去 (qù, go) and “go” present a much lower frequency of initial windowing with only 2 hits in Chinese for 去 (qù, go) and 3 hits in English for “go”, which suggests that both English and Chinese seldom select the initial windowing in the path. Compared with the extraordinary tendency toward final windowing of 去 (qù, go) and “go”, it can be deduced that the starting point of the motion is not as important as the ending point in both English and Chinese so it is often put into the background of the path event frame. And because of the overwhelming number of final windowing, the mode complexity of windowing containing final windowing like M + F, I + F, and I + M + F hold a moderate position in terms of quantity.
Table 2. Modes of path windowing of去 (qù, go).
Modes of path windowing |
去 (qù, go) |
go |
|
|
Frequency |
% |
Frequency |
% |
chi-square |
Sig. level |
I |
2 |
0.67 |
3 |
1 |
57.64 |
p < 0.0001 |
M |
1 |
0.33 |
34 |
11.33 |
|
|
F |
239 |
79.67 |
244 |
81.33 |
|
|
I + M |
19 |
6.33 |
0 |
0 |
|
|
M + F |
11 |
3.67 |
6 |
2 |
|
|
I + F |
24 |
8.00 |
12 |
4.00 |
|
|
I + M + F |
4 |
1.33 |
1 |
0.33 |
|
|
Total |
300 |
100 |
300 |
100 |
|
|
A difference lying in the medial windowing between 去 (qù, go) and “go” is that the path led by “go” is more inclined to get the medial portions windowed than 去 (qù, go) with 11.33% of English being greater than 0.33% in Chinese. One possible explanation for this discrepancy may be that English has a propensity for satellites after motion verbs. According to the qualitative detection of the data, satellites like “through”, “past”, or “over” often follow the verb “go” to cause the sentence to window the medial portion. This can be exemplified in Eventually this storm will turn and go right over the top of Shanghai with only windowing the medial part of the whole path.
The mode of I + M windowing in English is zero in number whereas there are 16 instances of I + M windowing in Chinese. The discrepancy probably stems from a deeper cause that the difference in the number of closed paths between English and Chinese may result in the gap in the amount of I + M windowing. One significant point in the statistic is that both 去 (qù, go) and “go” only have several cases of full windowing. This is an indication in perception that cognitive subjects do not intend to get all portions of the path windowed. On the contrary, it is more common that one or two portions along the path will be put into the foreground while most elements are ensconced background because the entire path can be inferred from the existing windowed portions.
(7a) 我 刚 刚 离开 台 南 的 家 里，到 台 北 去 生 活。
wǒ gāng gāng lí kāi tái nán de jiā lǐ, dào tái běi qù shēng huó.
I just left Tainan DE home to Taipei go live.
(“I had just left my home in Tainan at that time and went to live in Taipei.”)
(7b) We went way out of Ridge Road and backed up the edge of a big hill.
By inferring from the context, the instance in (7a) depicts an open path starting from Tainan to Taipei. The speaker in this interview intentionally foregrounds the starting point and the destination of this experience to emphasize the movement in his life. The structure 到 + place + 去 (dào + place + qù, to + place + go) gets the final portion windowed and the description of the speaker leaving Tainan causes the initial part of the path to be windowed. Though not all portions of the path are windowed, the initial and the final portions of the path are adequate for cognitive subjects to construct the complete path effortlessly in perception. Example (7b) is a display of a different mode of windowing of“go” from the Chinese instance. In example (7b), “go” depicts an open path of a group of people secretly smuggling whisky. By describing the passing road and destination, the speaker gets the medial and the final portions windowed but the initial portion gapped.
5. Discussion: Linguistic Factors Adjusting the Windowing Conflict of 去 (qù, go) and “go”
5.1. Tense Restriction
Different selections of tense have a strong impact on modes of windowing and even path types. For example, a simple le (“finished”) in Chinese, serving to change the event into a finished state in a sentence, can help to transform the structure of modes of windowing.
(8a) A: 前 天 上 午 干 什 么 了？
qián tiān shàng wǔ gàn shén me le?
The day before yesterday morning do what perfective (hereafter PRF)?
(“What did you do the morning before yesterday?”)
B: 我 弄 牙 去 了。
wǒ nòng yá qù le.
I treated teeth go PRF.
(“I went to the dentist.”)
(8b) B: 我 弄 牙 去。
wǒ nòng yá qù.
I treat teeth go.
(“I’m going to the dentist.”)
The casual talk happens between a sister and a brother about what B did the day before yesterday in the morning. The answer of B selects the past tense to express the finished state of the event. Because of that, the path directed by 去 (qù, go) appears as a closed path and the initial and medial portions are windowed. The causal relationship between the use of past tense and the output of the closed path and I + M mode can be clarified in detail by contrast with the same sentence using a different tense. In (8b), the past tense marker le (“finished”) is moved so the sentence turns to use a future tense. Example (8b), in contrast, portrays an open path that only foregrounds the destination with the final portion windowed. The change of tense has a great impact on the mode of windowing. The same referent scene applying different tenses may produce completely diverse path types and modes of windowing.
5.2. Satellite Coercion
As mentioned before, satellites like “over” used after “go” often serve the function of adjusting the attention focus in a sentence. Modes of windowing are easily transformed by different satellites.
According to figures presented in Table 3, there are 7 disparate structures of “go” over half of which contain satellites after the main verb. The number of “go + satellite” construction is listed at the first place in the overall forms taking a high percentage up to 49.67%. Satellites abound in English. According to a rough detection, over 20 types of satellites appear after the path verb “go” in English samples of this research.
(9a) I just want to tell everybody who is going through a cancer journey, be positive and never give up.
(9b) The glucose now goes into the tumor. And the tumor uses it to grow.
(9a) presents a fictive path depicting a nonveridical motion along a cancer journey. By using “through”, the sentence is coerced into foregrounding only the medial portion of the path so that the medial portion is windowed. The progressive tense used in this instance works to limit the sentence to get the medial portion windowed as well. Comparatively, a different satellite “into” in (9b) is put after the path verb “go” to indicate the destination of the whole path thus the final portion is coerced to be windowed.
Table 3. Forms of “go”.
Forms of “go” |
|
go |
Frequency |
% |
go + satellite |
go in go out go through go up go down … |
149 |
49.67 |
go… + place |
go to + place go into + place go + place go in + place |
96 |
32 |
go |
|
25 |
8.33 |
go… + verb (hereafter V) |
go + v. go into + v. go to + v. |
17 |
5.67 |
go + direction |
|
6 |
2 |
go + adjective (hereafter ADJ) |
|
6 |
2 |
go + noun (hereafter N) |
|
1 |
0.33 |
Total |
|
300 |
100 |
5.3. Verbal Inertia
As mentioned above, 去 (qù, go) tends to give prominence to the final windowing. One possible cause for this phenomenon may be that some entrenched verb serial constructions abounding in data of 去 (qù, go) force the information to be expressed into indicating at least the final windowing in a path. The two serial verbs with no link words in between are defined by Goldberg (2005) as Go VPbare construction (Xing & Liang, 2012). However, how the serial verb construction coerces the final portion to be windowed is not explained. There exists a kind of inertia between the two movements represented by the two successive verbs. This inertia functions to link the two motions and it regulates that the first movement represented by 去 (qù, go) plays an essential role in deciding the happening of the second movement. If one has to go to some place before he or she performs an act, the successive motions necessarily suggest a change of location in the path event frame, which implicitly denotes the destination of the path. Therefore, the inertia that connects the two successive movements plays a vital role in windowing the final portion of a path. Though the destination to perform the act is not ostensibly specified in language, a place strongly associated with the act to be conducted has already been assumed in the listeners’ minds and is thus identified as a final windowing in the path event frame. This is a special case of final windowing.
(10) 我 就 选 择 了 离 开 队里，我 去 上 学。
wǒ jiù xuǎn zé le lí kāi duì lǐ, wǒ qù shàng xué.
I then chose PRF leave team, I went attend school.
(“I then chose to leave the team and I went to attend school.”)
For example, in (10) the inertia between 去 (qù, go) and 上学 (shàng xué, attend school) forces the destination to be foregrounded. To be specific, before attending school, one has to go there, which causes the final portion of the path to be windowed.
Table 4 shows, the verb serial construction prevails in Chinese data with 去 + V (qù + V, go + V) taking up a relatively large proportion in all forms of 去 (qù, go). Most of the 去 + V (qù + V, go + V) constructions are inclined to window the final portion of the path. In addition, 去 + place (qù + place, go + place), explicitly indicating the destination of the path, shares a slightly larger proportion than 去 + V (qù + V, go + V). Therefore, the large number of 去 + V (qù + V, go + V) and 去 + place (qù + place, go + place) which indicate destinations of the path implicitly or explicitly, may contribute to the large amount of final windowing.
Table 4. Forms of 去 (qù, go).
Forms of 去 (qù, go) |
去 (qù, go) |
Frequency |
% |
去 + place (qù + place, go + place) |
127 |
42.33 |
去 + V (qù + V, go + V) |
106 |
35.33 |
到 + place + 去 (dào + place + qù, to + place + go) |
24 |
8 |
去 (qù, go) |
24 |
8 |
V + 去 (V + qù, V + go) |
9 |
3 |
V + place + 去 (V + place + qù, V + place + go) |
7 |
2.33 |
上 + place + 去 (shàng + place + qù, to + place + go) |
3 |
1 |
Total |
300 |
100 |
6. Conclusion
Under the framework of the Windowing of Attention Theory, this study explores the tendency of path selections and modes of windowing of 去 (qù, go) and “go” by making a corpus-based survey of the linguistic data extracted from the DiSCUSS corpus and COCA. This research has the findings below.
1) Both English and Chinese tend to give priority to the type of open path with 去 (qù, go) counting for 82.6% and “go” accounting for 81.3%. Types of closed and fictive path take a comparatively small proportion in the overall path types.
2) Data of 去 (qù, go) and “go” shows a similar tendency toward the windowing of the final portion along the path, which denotes that both languages tend to foreground the destinations in the path. In addition, the modes of I, M, M + F, I + F, and I + M + F are all presented in English data in different numbers except for the mode of I + M. Comparatively, all 7 modes of windowing appear in the Chinese data with just little figures in each mode.
3) The modes of windowing of attention are not just governed by the free will of cognitive subjects. Instead, some factors, observed from the qualitative detection of the data, also feature prominently in adjusting the focus of attention in a linguistic expression. These important factors include tense restriction which can change the modes of windowing and even the type of path employed, satellite coercion which functions to assign different windowed portions with diverse satellites, and verbal inertial between the two verbs in serial verb constructions working to get final portions windowed.
Acknowledgements
The present research is funded by the 2024 university-level research project of Inner Mongolia University entitled “The Path Selection and Windowing of Attention of Motion Verbs in the Path Event Frame: A Corpus-based Contrastive Study between 去 (qù, go) and ‘go’ in Chinese and English” (11200-54220396).