A corpus-based study on semantic and cognitive features of bei sentences in Mandarin Chinese
-
Yonghui Xie
and
Haitao Liu
Abstract
Bei sentences in Mandarin Chinese with SOV word order have attracted extensive interest. However, their semantic features lacked quantitative evidence and their cognitive features received insufficient attention. Therefore, the current study aims to quantitatively investigate the semantic and cognitive features through the analysis of nine annotated factors in a corpus. The results regarding bei sentences show that (i) subjects exhibit a tendency to be definite and animate; non-adversative verbs have gained popularity over time, and intransitive verbs are capable of taking objects; (ii) subject relations tend to be long, implying heavy cognitive load, whereas the dependencies governed by subjects are often short, suggesting light cognitive load; and (iii) certain semantic factors significantly impact cognitive factors; for instance, animate subjects tend to govern shorter dependencies. Overall, our study provides empirical support for the semantic features of bei sentences and reveals their cognitive features using dependency distance.
1 Introduction
The bei sentence is a unique and significant grammatical construction in Mandarin Chinese. In contrast to the widely recognized SVO word order of active sentences, as illustrated in (1) (Cheung 1973), bei sentences adopt SOV word order to express passive voice, as exemplified in (2). Moreover, there are frequent syntactic alternations between these two constructions, such as (1) and (2a).
| 高俅 | 陷害 | 林冲。 |
| Gāo Qiú | xiàn-hài | Lín Chōng |
| Gao Qiu | frame | Lin Chong |
| Subject + | Verb + | Object |
| ‘Gao Qiu framed Lin Chong’ | ||
| 林冲 | 被 | 高俅 | 陷害。 |
| Lín Chōng | Bèi | Gāo Qiú | xiàn-hài |
| Lin Chong | BEI | Gao Qiu | frame |
| Subject + | bei + | Object + | Verb |
| ‘Lin Chong was framed by Gao Qiu.’ | |||
| 淮安市 | 淮安区 | 被 | 誉为 | 美食 | 之 | 乡。 |
| Huái’ān-shì | Huái’ān-qū | bèi | yù-wéi | měi-shí | zhī | xiāng. |
| Huai’an City | Huai’an District | BEI | known as | gourmet food | DE | hometown |
| Subject + | bei + | Verb | ||||
| “Huai’an District in Huai’an City is known as the hometown of gourmet food.” | ||||||
As shown in the glosses, (2a) has an object while (2b) does not. Thus, the standard bei sentence can be formulated as Subject + bei + (Object) + Verb, where each syntactic component has basic features. The subject of a bei sentence, acting as the patient of an action, corresponds to the object of an active sentence (Siewierska 2013), like Lín Chōng 林冲 ‘Lin Chong’ in (2a) and (1). The foregrounded patient is one of the outstanding features of the bei sentence, and indeed, of many passive constructions in the world’s languages (Keenan and Dryer 1985). Conversely, the object of a bei sentence, serving as the agent of an action, corresponds to the subject of an active sentence, such as Gāo Qiú 高俅 ‘Gao Qiu’ in (2a) and (1). The backgrounded agent in a bei sentence may not even be mentioned (as in 2b), with a frequency of approximately 60 % (Xiao et al. 2006). Bei (bèi 被 ‘BEI’), as a passive marker, functions to introduce an agent (i.e., an object). Additionally, a verb is generally considered the most difficult component to process within a sentence (Hammarström 2016; Levin and Hovav 2005), with no exception in the bei sentence. In summary, the subject and verb of the bei sentence have more prominent features and involve more cognitive activities than the object and marker bei, thereby becoming our primary concern.
Bei sentences have been extensively studied from various perspectives, including their structural characteristics (Hashimoto 1988; Huang 1999; Tang 2001), semantic features (Chappell 1986; Li 1980), historical evolution (Wang 1957; Zhang 1994), and comparisons with English passive constructions (Kong 2014; Wang 1983). Particularly, some studies have focused on the semantic features of subjects and verbs in bei sentences, such as definiteness, animacy, adversity and transitivity.
Specifically, the subjects of bei sentences are typically regarded as definite (Huang and Liao 2002; Li 1994). A referent is considered to be definite if the speaker assumes that the addressee is able to identify the particular entity in question among other entities of the same or different class within the context (Chen 2004). The subjects also tend to be animate in bei sentences (Xiong and Wang 2002). Animacy refers to the extent to which the referent of a noun is perceived as sentient or alive (Santazilia 2022). In (2a), the subject Lín Chōng 林冲 ‘Lin Chong’ as a personal name is typically viewed as definite and animate.
Adversity or affectedness of the verb in bei sentences, in contrast to the passive constructions of other languages, is a unique linguistic phenomenon (Li and Thompson 1981; Wang 1984). The adversative verb often imposes misfortune or unpleasantness on the patient, implying an unfortunate or unhappy event. However, after the profound language revolution sparked by the May Fourth Movement in 1919, bei sentences have gradually embraced non-adversative usages (He 2008). It is also notable that transitive verbs are usually used in bei sentences to convey disposal or causative meanings (Li and Thompson 1981). Nonetheless, with the frequent occurrence of the construction bèi zì-shā 被 自杀 ‘BEI suicide’ in 2008, intransitive verbs have gradually been incorporated into the bei sentences (Peng and Gan 2010; Wang 2009). In (2a), the verb xiàn-hài 陷害 ‘frame’ is adversative and transitive, causing misfortune to its patient Lín Chōng 林冲 ‘Lin Chong’.
Previous studies have provided us with a relatively clear understanding of these four semantic factors. However, they largely relied on qualitative analyses and, as a result, lacked quantitative evidence. This limitation restricts our ability to accurately measure the semantic factors, potentially hindering the development of robust linguistic theories. The limitation may also raise questions regarding the validity of the conclusions, particularly when faced with arguments similarly grounded in introspective methods.
For instance, there is a tendency for the subject to be definite but the object to be indefinite in Chinese (Chao 1979). In a bei sentence, the subject corresponds to the object of an active sentence, which may cause debates regarding the definiteness of such subjects. In general, the direction of an action is from the animate agent to the inanimate patient in a sentence. After the foregrounding operation, it is unknown whether the patient in a bei sentence occupying the subject position retains its inanimate property or not. Additionally, it remains uncertain whether linguistic evolution will introduce new changes to the adversity and transitivity of verbs. Despite several quantitative studies (Li 2012; Xiao et al. 2006; Zhu and Hu 2014), these endeavors have not yet sufficiently addressed the issues at hand.
Language is the fundamental expression of human cognition and is intrinsically linked to cognitive activities (Croft and Cruse 2004; Pinker 2007). As such, the processing of bei sentences is ultimately constrained by cognitive resources (i.e., working memory). However, little is known concerning the cognitive features of subjects and verbs in bei sentences, primarily due to the challenge of conducting psychological experiments in this particular area (Zeng et al. 2020). This lacuna is not conducive to a deeper understanding of bei sentences. Fortunately, recent developments in computational cognitive sciences have proposed data-based metrics (e.g., dependency distance, Liu 2008), which effectively reflect the cognitive load during sentence processing. These metrics offer convenient methods for measuring the cognitive features of bei sentences.
Dependency distance refers to the linear distance between two syntactically related words in a sentence. In (2a), the dependency distance between the word Lín Chōng 林冲 ‘Lin Chong’ (the dependent of the subject relation) and the word xiàn-hài 陷害 ‘frame’ (the governor of the subject relation) is 3 (see Section 2 for more details). According to memory-based processing models (Gibson 1998, 2000), the longer a dependency distance is, the greater the cognitive load is. Specifically, language processing is entailed in continuously activating elements related to new words from existing structures and then forming syntactic relations; and the activation level of a new word declines with time (measured by linear distances) due to the limited working memory of human beings (Nicenboim et al. 2015; Warren and Gibson 2002). In psycholinguistics, dependency distance has received empirical supports as a metric of syntactic difficulty and cognitive load, owing to its validity in predicating reading times (Grodner and Gibson 2005; Niu and Liu 2022).
Dependency distance has been applied in various linguistic fields, including L2 writing proficiency (Ouyang et al. 2022), English structural features (Dai et al. 2023) and interpreting (Liang et al. 2017). In particular, Xu and Liu (2015) and Li (2020) employed this metric to analyze general Chinese sentences that takes SVO word order. Fang and Liu (2018, 2021, 2022) utilized it to observe Chinese ba sentences, a unique Chinese construction with SOV word order. Based on these previous studies, we propose five specific factors related to dependency distance to delve into the cognitive features of subjects and verbs in bei sentences. One factor is the dependency distance of the subject relation, which measures the cognitive load of establishing this relation. The cognitive load of both the subject and the verb, when acting as the governor respectively, can also be quantified by calculating all the dependency distances pointing from each of them to their respective dependents. In light of this concept, the four other factors we consider are: the mean and furthest dependency distances pointing from a subject to its dependents, gauging the average and maximum cognitive load of the subject as a governor in a bei sentence; the mean and furthest dependency distances pointing from a verb to its dependents, quantifying the average and maximum cognitive load of the verb as a governor.
The first element of science is research objects that are defined by their features and the interrelations among them. Likewise, language, as a complex adaptive system comprising interconnected components (Liu 2018), necessitates an exploration of both the features of individual components and their interrelations. Consequently, investigating the interplay between semantic and cognitive features in bei sentences holds particular significance. While this area remains understudied, insights gained from prior research on related language constructions can be invaluable. For example, Xu and Liu (2015) found that contextually given subjects in Chinese tend to govern longer subject relations compared to contextually new subjects. Fang and Liu (2018) uncovered that the givenness of subjects in Chinese ba sentences also influences dependency distance. Moreover, Sharma et al. (2020) discovered that animate subjects in Hindi can override the constraint of dependency distance minimization. In brief, there is a research tradition to explore the impact of semantic features on cognitive features, which we can follow and build upon.
Given the above background, the current study aims to quantitatively examine the semantic and cognitive features of bei sentences, as well as their interrelations. The three specific research questions are as follows:
What semantic features do subjects and verbs of bei sentences exhibit, based on new data and quantitative analyses?
What cognitive features do the subjects and verbs possess, using the metric of dependency distance?
What is the impact of the semantic features on the cognitive features?
To address these questions, we first collect bei sentences to build a corpus, and then annotate and quantify nine specific semantic and cognitive factors. Finally, we discuss the statistical results to reveal the semantic and cognitive features. The layout of this paper is as follows. The details of data collection and factor annotation are introduced in Section 2. Section 3 presents the statistical methods and results. Section 4 analyzes the results and discusses the answers to the three research questions. In Section 5, this paper is ended with conclusions.
2 Data and factors
2.1 Corpus
Our self-built corpus consists of 600 bei sentences, which follow the form of Subject + bei + (Object) + Verb and were produced from 2009 to 2019. The primary data source is the ToRCH (Texts of Recent Chinese) family of corpora,[1] developed by Beijing Foreign Studies University, which includes ToRCH2009 (Xu 2014), ToRCH2014 and ToRCH2019. The corpora are designed based on the sampling frame of the Brown Corpus and comprise 15 genres, such as press reportage, science, biographies, reports and adventure fiction. Although the ToRCH corpora contain a large number of raw texts totaling approximately five million Chinese characters, the occurrence of bei sentences within these texts, as we observed, is very limited.[2] Therefore, it becomes imperative to collect bei sentences from additional data sources, including the People’s Daily published from 2011 to 2015, Memoirs of Literature (Mu and Chen 2013), and Records of Pastoral Gods created by the internet writer Zhu Zhai from 2017 to 2019.
To extract bei sentences from these raw texts, Python scripts were developed. Firstly, the Stanford CoreNLP tool (Manning et al. 2014) was used to annotate all the raw texts for sentence splitting, word tokenization, part-of-speech (POS) tagging and dependency parsing. Subsequently, bei sentences were identified based on specific dependency tags. Specifically, the tag nsubjpass represents the passive nominal subject, auxpass denotes the passive auxiliary, and ROOT represents the root of a sentence, typically a finite verb. A sentence can be classified as a bei sentence if it contains a word labeled as nsubjpass and its governor labeled as ROOT. To ensure accuracy, the annotations were finally proofread by two linguistics professionals.
To illustrate, Table 1 provides the dependency parsing of (2b). The tag nsubjpass indicates that the word Huái’ān-qū 淮安区 “Huai’an District” is the passive nominal subject of the verb yù-wéi 誉为 ‘known as’. This verb is tagged as ROOT. Therefore, (2b) should be added to our bei sentence corpus.
Dependency parsing of (2b).
| Word | Word order | POS | Word order of governor | Dependency relation |
|---|---|---|---|---|
| Huái’ān-shì | 1 | NR | 2 | compound:nn |
| Huái’ān-qū | 2 | NR | 4 | nsubjpass |
| bèi | 3 | SB | 4 | auxpass |
| yù-wéi | 4 | VV | 0 | ROOT |
| měi-shí | 5 | NN | 7 | nmod:assmod |
| zhī | 6 | DEG | 5 | case |
| xiāng | 7 | NN | 4 | dobj |
2.2 Factors
The nine specific semantic and cognitive factors under investigation are listed in Table 2. This section introduces how they were coded.
Nine semantic and cognitive factors.
| Factor | Code | |
|---|---|---|
| Semantic features | Definiteness of subjects | Definiteness |
| Animacy of subjects | Animacy | |
| Adversity of verbs | Adversity | |
| Transitivity of verbs | Transitivity | |
| Cognitive features | Dependency distance of subject relations | DD |
| Mean dependency distance pointing from subjects | Sub_MDD | |
| Furthest dependency distance pointing from subjects | Sub_FDD | |
| Mean dependency distance pointing from verbs | Verb_MDD | |
| Furthest dependency distance pointing from verbs | Verb_FDD |
Definiteness possesses a binary feature of being definite and indefinite. In some Indo-European languages, definiteness can be judged by grammatical categories. For example, constituents in English that contain a proper noun or pronoun as the head, or those starting with a definite article or demonstrative, are tagged as definite (Garretson 2004). Nonetheless, Mandarin Chinese lacks these grammatical categories. Drawing upon the classification of Chinese nominal elements (Chen 2004) and its quantitative application (Fang and Liu 2021), our study coded the five formal definitions as definite subjects: personal pronouns, proper nouns, zhè/nà/zhè-xiē/nà-xiē 这/那/这些/那些 ‘this/that/these/those’ + (quantifier) + nouns, possessive pronoun + nouns, and given nouns (the words that have occurred in the preceding ten sentences). Otherwise, the subject was marked as indefinite.
Animacy is usually marked as a binary property: animate or inanimate (Bresnan et al. 2007; Wolk et al. 2013). According to Li (2012), we coded subjects related to people, animals, bodies, and collectives[3] as animate, like nǚ-rén 女人 ‘women’, māo 猫 ‘cat’, shǒu 手 ‘hand’ and guó-jiā 国家 ‘country’. Moreover, we coded natural forces, concrete things and abstract things as inanimate, such as fēng 风 ‘wind’, shū 书 ‘book’ and găn-qíng 感情 ‘emotion’.
To determine whether a verb is adversative or not, we followed the principle proposed by Wang (1985). If there is an animate subject, the attitudinal meaning of the verb is directed to that subject. As illustrated in (3a) below, the verb xià 吓 ‘startle’ makes the animate subject mǔ-qīn 母亲 ‘mother’ have a negative feeling, so this verb can be marked as adversative. For inanimate subjects, however, the attitudinal meaning of the verb is directed to the speakers. As in (3b), the verb fēng-zhù 封住 ‘seal’ probably poses a danger to the people at the fire scene (i.e., the potential speakers); thus, this verb can be labeled as adversative.
| 母亲 | 被 | 吓 | 得 | 打 | 了 | 一 | 个 | 冷战。 |
| Mǔ-qīn | bèi | xià | dé | dǎ | le | yī | gè | lěng-zhàn. |
| Mother | BEI | startle | DE | do | LE | a | CLF | shiver |
| ‘The mother was startled so much that she shivered.’ | ||||||||
| 由于 | 火势 | 凶猛 , | 门窗 | 被 | 封住 , |
| Yóu-yú | huǒ-shì | xiōng-měng , | mén-chuāng | bèi | fēng-zhù , |
| Due to | fire | ferocity | door and window | BEI | seal |
| 烟熏火燎 | 容易 | 迷失 | 方向。 | ||
| yān-xūn-huǒ-liáo | róng-yì | mí-shī | fang-xiàng. | ||
| smoke and flame | easy | lose | way | ||
| ‘Due to the ferocity of the fire, the doors and windows were sealed, making it easy for people to lose their way in the smoke and flames.’ | |||||
The transitivity of verbs was coded in accordance with the standards set by Chao (1979) and Liu et al. (2019). Transitive verbs, such as xiàn-hài 陷害 ‘frame’, are capable of taking objects. On the contrary, intransitive verbs, like zì-shā 自杀 ‘suicide’, typically cannot carry objects. However, there are three types of objects that intransitive verbs can take: the location of the action, such as shàng shān 上 山 ‘go up the mountain’; the tool used to perform the action, such as guò shāi-zi 过 筛子 ‘pass through the sieve’; and something that exists, appears, or disappears, such as lái le yí-gè rén 来 了 一个 人 ‘a person comes’.
Within the framework of dependency grammar, dependency distance can be gauged by subtracting the linear word order of a governor and its dependent (Liu et al. 2009). As illustrated in Figure 1, arrowed arcs represent dependency relations and point from governors to their dependents. The DD between the subject Huái’ān-qū 淮安区 ‘Huaian District’ and the verb yù-wéi 誉为 ‘known as’ is |4 – 2| = 2. There is only one dependency governed by this subject, namely from Huái’ān-qū 淮安区 ‘Huaian District’ to Huái’ān-shì 淮安市 ‘Huaian City’, so Sub_MDD is |2 – 1|/1 = 1 and Sub_FDD is |2 – 1| = 1. Excluding punctuation, there are three dependencies governed by the verb yù-wéi 誉为 ‘known as’, and thus Verb_MDD is (|4 – 2| + |4 – 3| + |4 – 7|)/3 = 2. Among the three dependencies, the dependency distance between the verb and its direct object (dobj) xiāng 乡 ‘hometown’ is the furthest, resulting in Verb_FDD of |4 – 7| = 3. The linear word order used to calculate dependency distance are readily available in our dependency-annotated corpus. As shown in Table 1, dependency distance is calculated as the absolute difference between the “Word order of governor” and “Word order” columns.
Dependency structure of (2b).
Among the nine factors, three – animacy, adversity and transitivity – were manually annotated by two linguistics professionals. Their annotations exhibit high consistency upon manual comparison. The remaining six factors, however, were automatically annotated based on a Python environment.
3 Methods and results
Utilizing our bei sentence corpus, three statistical analyses were conducted to address the three research questions (see Section 1). Firstly, considering the categorical nature of the four semantic factors, we calculated their respective percentages in our corpus. As depicted in Figure 2, the results show that definite subjects constitute 63.83 % of all subjects, outnumbering the percentage of indefinite subjects (36.17 %); animate subjects account for 58.33 % of all subjects, slightly exceeding the percentage of inanimate subjects (41.67 %); the percentage of adversative verbs stands at 42.33 %, marginally below that of non-adversative verbs (57.67 %); and the percentage of transitive verbs is remarkably high, reaching 95.17 %, surpassing that of intransitive verbs (4.83 %).
Percentages of the semantic factors.
The One-sample Chi-squared test was then performed to assess whether there was a significant difference between the two categories of each semantic factor or not. The results presented in Table 3 suggest that the percentage distributions of the four factors differ significantly from the expected 50 % distribution (all p < 0.001). Therefore, the differences observed in Figure 2 are statistically significant.
Significance of differences in semantic factors.
| Definite ∼ indefinite | Animate ∼ inanimate | Adversative ∼ non-adversative | Transitive ∼ intransitive | |
|---|---|---|---|---|
| P-value | <0.001*** | <0.001*** | <0.001*** | <0.001*** |
-
*p ＜ 0.05; **p ＜ 0.01; ***p ＜ 0.001.
Secondly, since the five cognitive factors are continuous variables, we computed their respective values in each bei sentence and then plotted their distributions using a boxplot, as shown in Figure 3. The five data labels on the boxplot represent the mean values of the five cognitive factors. Specifically, the mean values of DD, Sub_MDD, Sub_FDD, Verb_MDD, and Verb_FDD are 2.922, 1.108, 1.363, 5.098, and 13.220, respectively.
Distributions of the cognitive factors.
The third analysis investigated the impact of the semantic factors on the cognitive factors. On the one hand, we measured the variation in each cognitive factor across the binary categories of each semantic factor. Figure 4 presents the results, where, for instance, ▵DD = 0.15 signifies an absolute difference in the mean values of DD between the definite and indefinite subjects. Overall, the higher a data label is positioned in the figure, the greater the influence of a semantic factor on a given cognitive factor is.
Variations in cognitive factors across semantic factors.
On the other hand, we assessed the statistical significance of these variations in Figure 4. The Shapiro-Wilk test and Q-Q plots indicated that only one pair of cognitive factors, namely Verb_FDD across the transitivity of verbs, followed the normal distribution.[4] Subsequently, the Independent Samples Test was employed to examine whether there was a significant difference in Verb_FDD between the transitive and intransitive verbs. The Mann–Whitney U test was used to determine whether significant differences could be traced in the remaining nineteen pairs of cognitive factors. The results are summarized in Table 4. Both Sub_MDD and Sub_FDD exhibit significant variations across the different categories of definiteness and animacy (all p < 0.001). Similarly, both Verb_MDD and Verb_FDD display significant variations across the binary categories of definiteness and transitivity (all p < 0.05).
4 Discussion
This section provides a comprehensive discussion of the semantic and cognitive features of bei sentences, based on the above statistical results.
4.1 Semantic features
4.1.1 Definite subjects
Bei sentences exhibit a tendency to use definite subjects (63.83 %, Figure 2), though their subjects would be indefinite objects in active sentences. This finding empirically supports previous qualitative analyses (Huang and Liao 2002; Li 1994), which argued that the subject of a bei sentence is often treated as definite. While quantitative investigations into definiteness are quite limited, it should be noted that one such study, namely Li (2012), reported the percentage of definite subjects was 97.23 %, surpassing our results. This discrepancy is primarily attributed to the different definitions of definiteness: Li (2012) used qualitative definitions (see Section 1), while our study utilized formal definitions (see Section 2.2). Notably, the latter has the potential to enhance the replicability of the results and thereby promote consensus within the academic community. In this sense, our finding provides a more reliable estimate of definiteness.
Three possible explanations for the use of definite subjects are analyzed. To begin with, it may be related to the old-before-new preference (Rochemont 2013; Velnić 2019), which states that definite and given information is typically presented first in a sentence, compared to indefinite and new information. This is because such information is more accessible, easier to produce and understand, and more likely to engage the attention of the speakers or listeners. Consequently, it is the position at the beginning of the bei sentence that makes the subject definite.
Secondly, the use of definite subjects may be caused by the patient role. When a sentence highlights the result of an action, the patient, which undergoes a state change during force transmission, needs to be definite and specific. On the contrary, the action cannot be executed upon indefinite or ambiguous entities. In a bei sentence, the subject as the patient bears the result of the action (Li 1994; Wang 1983), which explains why it tends to be definite.
The third motivation probably lies in the thematic continuity in discourse. If the patient holds greater thematic relevance to the context than the agent, it is considered to have stronger thematic continuity. In such cases, a passive sentence with a foregrounded patient is more likely to be used than an active sentence (Liu 2011; Thompson 1987). Consequently, the subject in a bei sentence, serving as the patient of an action and the theme of a text, often exhibits a definite and known state.
4.1.2 Animate subjects
Bei sentences have a slight tendency to use the animate subjects (58.33 %, Figure 2), even though their subjects (i.e., agents) would function as the inanimate patients of active sentences. This tendency conforms to a principle of sentence production: animate nouns usually appear at the beginning of sentences (Branigan et al. 2008; Gennari et al. 2012). This principle assumes that animate nouns enjoy greater accessibility, thus giving the speakers more time to organize new information and making the utterance more fluent.
Since the subjects of bei sentences tend to be both definite and animate, the relationship between definiteness and animacy deserves study. Given their categorical nature, the Phi coefficient (r φ ) was employed, and as a result, a slightly positive correlation between animacy and definiteness was observed (r φ = 0.257, p < 0.001). This correlation could be explained from two perspectives: in psycholinguistics, it has been shown that both animate and definite entities enjoy high accessibility (Jaeger and Tily 2011); in our statistical analyses, many linguistic cases can be annotated as both definite and animate, such as personal pronouns, personal names and the names of collectives.
4.1.3 Non-adversative verbs
Bei sentences do not show a significant tendency to use adversative verbs (42.33 %, Figure 2), despite adversity being their typical feature. Previous studies have favored the utilization of news texts to conduct statistical analyses on adversity. To achieve a diachronic comparison, we additionally calculated the percentage of adversative verbs in 200 bei sentences derived from news texts in our corpus. The results show that the percentage of adversative verbs, over time, displays a decreasing trend, from 51.5 % in the early 1990s (Xiao et al. 2006) to 45 % in 2006 and 2007 (Zhu and Hu 2014), and ultimately down to only 35 % in our corpus spanning from 2009 to 2019. In other words, the non-adversative verbs in bei sentences appear to grow prevalent with time. This shift could be attributed to the ongoing influence of Indo-European languages (He 2004) as well as Chinese grammaticalization (Guo and Han 2012).
In addition, previous quantitative analyses have revealed the varying usage of adversative verbs across genres. Utilizing the 1990s corpus, Li (2004) and Xiao et al. (2006) observed that the percentage of adversative verbs in fiction texts is significantly higher than in news texts. To extend this observation, we calculated the percentage of adversative verbs in 200 bei sentences sourced from fiction texts in our corpus from 2009 to 2019. The results show that the percentage of adversative verbs is notably higher in fiction texts (58 %) than in news texts (35 %). The findings suggest that the fiction genre has exhibited a greater preference for adversity in different historical periods compared to the news genre.
The variation between these two genres may be related to their inherent properties. The news genre has the features of written language, whereas the fiction genre possesses the features of both spoken and written language (Biber 1988). Historically, the contact between Chinese and Indo-European languages, particularly English, has been primarily indirect, through written translations. Therefore, the influence of the non-adversative usage of English passive sentences occurs predominantly in written language (He 2004). Moreover, Wang (1980) argued that the centuries-old customs of spoken language are less prone to significant alterations within a short period of time. For example, our corpus shows that some non-adversative verbs, such as píng-wéi 评为 ‘rate’, chēng-wéi 称为 ‘call’ and yù-wéi 誉为 ‘acclaim’, are more frequently used in news texts than in fiction texts. These Chinese verbs are likely to be influenced by their English counterparts, which are often used in English passive sentences (Liu 2011). This instance is a vivid illustration of the different degrees of influence of English on Chinese across genres.
4.1.4 Transitive verbs
Bei sentences display a strong tendency to use transitive verbs (95.17 %, Figure 2), which can take affected objects. Notably, among the 29 intransitive verbs in our corpus, such as pài-wăng 派往 ‘dispatch to’, qiān-huí 迁回 ‘move back’ and sòng-wăng 送往 ‘sent to’, we observed that they can take objects indicating the location of the action. As illustrated in (4) below, the intransitive verb pài-wăng 派往 ‘dispatch to’ (in bold) can introduce its destination Yán’ān Lŭ-yì 延安 鲁艺 “Luyi School in Yan’an City” (underline). Moreover, these intransitive verbs, being compound words, contain an implied morpheme that serves both as the object of the first verb morpheme and as the subject of the second, fulfilling dual roles. In (4), the first morpheme pài 派 ‘dispatch’ of the verb pài-wăng 派往 ‘dispatch to’ takes an implied object Liú Xīlín 刘西林 ‘Liu Xilin’ (underline), which concurrently functions as the subject of the second morpheme wăng 往 ‘to’.
| 1942年, | 刘西林 | 被 | 派往 | 延安 | 鲁艺 | 学习。 |
| 1942 nián | Liú Xīlín | bèi | pài-wăng | Yán’ān | Lŭ-yì | xué-xí. |
| 1942 | Liu Xilin | BEI | dispatch to | Yan’an City | Luyi School | study |
| ‘In 1942, Liu Xilin was dispatched to Luyi School in Yan’an for his study.’ | ||||||
The feature of intransitive verbs in our corpus contrasts with previous analyses, which used different corpora. For instance, Li (2012) found that 0.44 % of verbs were intransitive in 2055 bei sentences from 2002 to 2003, and characterized them as non-actional and non-autonomous. Furthermore, these verbs were categorized into three pragmatic meanings: mental states, such as táo-zuì 陶醉 ‘intoxicate’; physiological phenomena, like zhì-xī 窒息 ‘suffocate’; and changes in state, such as péng-zhàng 膨胀 ‘swell’ in (5).
| 胸 | 中 | 渐渐 | 被 | 欣喜 | 膨胀 | 起来。 |
| Xiōng | zhōng | jiàn-jiàn | bèi | xīn-xǐ | péng-zhàng | qǐ-lái. |
| Chest | In | gradually | BEI | joy | swell | up |
| ‘The chest gradually swelled up with joy.’ | ||||||
In addition, numerous studies on the construction bei + intransitive verb have focused on catchwords (e.g., Peng and Gan 2010; Xiong and Zeng 2011), such as bèi zì-shā 被 自杀 ‘BEI suicide’. More specifically, Xue and Zhang (2021) gathered 80 such catchwords from 2009 to 2021, and classified them into three pragmatic meanings: inauthenticity, such as bèi jiù-yè 被 就业 ‘BEI get a job’; involuntary, like bèi jiā-bān 被 加班 ‘BEI work overtime’; and suffering, such as bèi yú-lè 被 娱乐 ‘BEI play a trick’. All in all, our analysis provides novel insights for the study of bei sentences concerning intransitivity.
4.2 Cognitive features
4.2.1 Long dependency distances of subject relations
The mean value of DD (dependency distance of subject relations) in our bei sentence corpus is 2.922 (Figure 3). By comparing the result with that in general Chinese sentences and ba sentences, the potential universalities and peculiarities of bei sentences in terms of syntactic and cognitive features could be identified.
The first comparison is with general Chinese sentences. As reported by Xu and Liu (2015) and Li (2020), the mean DD in such sentences was about 2.65, which is lower than that in bei sentences (2.922). This discrepancy indicates that the bei sentence possesses a comparatively long subject relation and potentially requires the heavy cognitive load in processing it. This phenomenon is attributed to SOV word order of bei sentences.
From a cognitive perspective, since language processing is constrained by limited cognitive resources, the factors that mitigate the cognitive load invoked by this long dependency in bei sentences are worth investigating. According to Givón (1992) and Chafe (1994), a subject with high accessibility is somewhat insensitive to time-invoked memory decay, and facilitates its later retrieval in a long dependency. In our study, the bei sentences tend to use definite and animate subjects that have high accessibility; therefore, it is plausible to hypothesize that such subjects allow long subject relations in bei sentences. However, the results presented in Figure 4 and Table 4 fail to support the hypothesis, as there are no significant differences in the mean DD when considering the definiteness or animacy of subjects (both p > 0.05). It would be worthwhile to delve into alternative explanations, such as the priming effect of the marker bei (Pickering et al. 2013) and the influence of short information structures (Wasow 1997).
Significance of variations in cognitive factors across semantic factors.
| ΔDD | ΔSub_MDD | ΔSub_FDD | ΔVerb_MDD | ΔVerb_FDD | |
|---|---|---|---|---|---|
| Definiteness | 0.176 | <0.001*** | <0.001*** | 0.008** | 0.029* |
| Animacy | 0.228 | <0.001*** | <0.001*** | 0.612 | 0.596 |
| Adversity | 0.560 | 0.129 | 0.203 | 0.785 | 0.964 |
| Transitivity | 0.172 | 0.718 | 0.563 | 0.030* | 0.017* |
-
*p ＜ 0.05; **p ＜ 0.01; ***p ＜ 0.001.
For ba sentences, Fang and Liu (2018) reported the mean DD of approximately 5.953, which is higher than the mean DD of 2.922 observed in our bei sentences corpus. This discrepancy suggests that the subject relations in bei sentences impose lighter cognitive burden, despite both constructions following SOV word order. The variance could be attributed to the omission of the intervening components within the subject relations of bei sentences.
In a bei sentence, the main intervening component is bei + object, and the object is frequently omitted with a frequency of approximately 60 % (Xiao et al. 2006). On the contrary, the main intervening component in a ba sentence, ba + object, cannot be easily omitted because the object functions as the patient of a causative verb (Wang 1985) or a disposal verb (Ziegeler 2000). To illustrate, (6) presents a ba sentence that conveys the same semantics as (1) and (2a). In this example, bǎ Lín Chōng 把 林冲 ‘BA Lin Chong’ is the intervening component of the subject relation, and its omission would result in the absence of the patient of the verb xiàn-hài 陷害 ‘frame’, thereby rendering the sentence ungrammatical.
| 高俅 | 把 | 林冲 | 陷害 | 了。 |
| Gāo Qiú | bǎ | Lín Chōng | xiàn-hài | le |
| Gao Qiu | BA | Lin Chong | frame | LE |
| Subject + | ba | Obejct + | Verb | |
| ‘Gao Qiu framed Lin Chong.’ | ||||
4.2.2 Short dependency distances pointing from subjects
The mean values of Sub_MDD (mean dependency distance pointing from subjects) and Sub_FDD (furthest dependency distance pointing from subjects) in bei sentences are 1.108 and 1.363, respectively (Figure 3). This section compares the result with that in general Chinese sentences to analyze the feature of bei sentences.
While previous research has not examined all the dependencies governed by the subject in a general Chinese sentence, Wang (2018) calculated the mean length of noun phrases (including subjects) across multiple texts, typically ranging from three to five words. To ensure an efficient comparison, it is necessary to measure the length of subjects within bei sentences based on the current statistical results. Therefore, we calculated the percentages of different values of Sub_MDD and Sub_FDD in the 600 bei sentences. The results indicate that the cumulative percentages of these two factors are 88 % and 82 %, respectively, when their values are no more than 2. Consequently, the length of subjects in bei sentences (i.e., Sub_MDD/Sub_FDD plus one) generally does not exceed three words, which is shorter than that observed in general Chinese sentences. All in all, the subjects of bei sentences tend to be short and concise, which may imply the light cognitive load during processing.
The use of short subjects can be expounded from the following three perspectives. Firstly, it may be influenced by the semantic features of subjects. Detailed statistical analyses (cf. Figure 4 and Table 4) demonstrate that when subjects are definite, the mean Sub_MDD (0.807) and mean Sub_FDD (1.010) are significantly lower than those for indefinite subjects (1.641 and 1.990, respectively). Similarly, animate subjects exhibit significantly lower mean Sub_MDD (0.793) and mean Sub_FDD (0.980) than inanimate subjects (1.550 and 1.900, respectively). Moreover, definite and animate subjects are frequently used in bei sentences. It can be concluded that subjects with high accessibility contribute to the low values of Sub_MDD and Sub_FDD (i.e., short subjects). Additionally, previous research remains controversial about whether there is an independent or interactional relationship between animacy and length (e.g., Kirby 1999; Rosenbach 2005). Our study provides new evidence for their relationship.
Secondly, the short subjects are adopted in order to follow the short-before-long preference (Hawkins 1994), a universal principle in sentence production that has been verified in languages such as English and Russian (Arnold et al. 2000; Kizach 2014). The cognitive motivation for this principle is that short linguistic units are easier to access than longer ones. In other words, subjects tend to be short because they are put at the beginning of the bei sentence.
The third account is couched in the principle of least effort (Zipf 1949), which states that the least amount of effort should be expended to accomplish a task. According to the percentage distributions of Sub_MDD and Sub_FDD, approximately 45 % of subjects are unmodified (Sub_MDD = 0 and Sub_FDD = 0), while 18.2 % of subjects have a single modifier (Sub_MDD = 1 and Sub_FDD = 1). The frequent occurrence of bare nouns and adjacent dependencies may be a strategy employed by the speakers to minimize effort expended, following the principle of least effort in sentence production.
Based on the third reason, it might be assumed that the short subjects serve to mitigate the cognitive load caused by long subject relations in bei sentences (see Section 4.2.1). To test this hypothesis, we performed the normality tests and correlation tests for the three factors－Sub_MDD, Sub_FDD and DD. The Shapiro–Wilk test and Q–Q plots indicated that their distributions were non-normal, and consequently, the Spearman’s correlation coefficient (r s ) was utilized to assess their correlations. However, the results do not uncover a significant correlation between Sub_MDD and DD (r s = 0.023, p > 0.05), as well as between Sub_FDD and DD (r s = 0.021, p > 0.05), thus refuting the mitigation effects.
4.2.3 Varied dependency distances pointing from verbs
The mean values of Verb_MDD (mean dependency distance pointing from verbs) and Verb_FDD (furthest dependency distance pointing from verbs) in our corpus are 5.098 and 13.220, respectively (Figure 3). Previous studies have placed less emphasis on investigating all the dependencies governed by the verb in different language constructions. Consequently, in this section, we aim to analyze the syntactic and cognitive features of the verbs from various perspectives, thereby facilitating a more comprehensive understanding of bei sentences.
Firstly, the variation in dependency distances pointing from verbs was observed. The mean Verb_FDD (13.220) is obviously higher than the mean Verb_MDD (5.098), and this difference is greater compared to that between the mean Sub_FDD (1.363) and the mean Sub_MDD (1.108). To enhance the rigor of this comparison, the correlation coefficients of each pair of cognitive factors were calculated. Given the non-normal distributions of the above four factors, we adopted the Spearman’s correlation coefficient. The results show that the correlation between Verb_MDD and Verb_FDD (r s = 0.959, p < 0.001) is slightly weaker than that between Sub_MDD and Sub_FDD (r s = 0.992, p < 0.001), supporting that the dependency distances pointing from verbs are more varied. The phenomenon arises because, as a root of a sentence, the verb governs a greater diversity of dependency relations than the subject, spanning a wider range of dependency distances. This, in turn, leads to more varied values of Verb_MDD and Verb_FDD. In summary, the cognitive mechanism involved in processing the verbs of bei sentences is more complex than that involved in processing their subjects.
Additionally, two semantic factors related to subjects and verbs have a significant impact on both Verb_MDD and Verb_FDD, as illustrated in Figure 4 and Table 4. The first factor is the transitivity of verbs. Transitive verbs tend to exhibit the higher mean Verb_MDD (5.164) and mean Verb_FDD (13.450) than their intransitive counterparts (3.791 and 8.660, respectively). This observation may be attributed to the inherent properties of transitive verbs, which typically take more kinds of objects, potentially increasing dependency distances. The second factor is the definiteness of subjects. When a subject is indefinite, the mean Verb_MDD (5.473) and mean Verb_FDD (14.080) tend to be higher compared to those for definite subjects (4.885 and 12.730, respectively). This may be because indefinite subjects require additional words to facilitate readers’ comprehension, potentially causing long dependency distances. However, the results do not demonstrate the significant influence of adversity and animacy on these two cognitive factors (all p > 0.05).
5 Conclusions
The current study conducted a corpus-based investigation into the semantic and cognitive features of subjects and verbs in Chinese bei sentences, using nine annotated factors. Regarding the semantic features, our findings are as follows: subjects tend to be definite (63.83 %), which may be explained by the old-before-new preference, patient role, and thematic continuity; subjects have a slight tendency to be animate (58.33 %), primarily due to their high accessibility; the usage of non-adversative verbs in the news genre shows an increasing trend over time, which may be influenced by Indo-European languages and Chinese grammaticalization; and intransitive verbs (4.83 %) can take objects indicating the location.
As for the cognitive features, our findings are summarized below. In bei sentences, the dependency distances of subject relations tend to be long, implying the heavy cognitive load in processing these relations, which is due to SOV word order. The dependency distances pointing from subjects tend to be short, suggesting the light cognitive load in processing the subjects, which is perhaps driven by the definite and animate subjects, short-before-long preference, as well as least effort principle. Additionally, the dependency distances pointing from verbs vary significantly, and they are likely to be influenced by both transitivity and definiteness.
Our findings offer quantitative evidence for the semantic features of bei sentences, facilitating the resolution of the controversies triggered by different introspective methods. Meanwhile, our findings, derived from one of the first dependency-based studies focused on bei sentences, provide new insights into the cognitive features of such sentences. At last, the influence of the semantic features on the cognitive features contributes to the understanding of the intricate interrelations between language subsystems.
Funding source: Innovative Practice Project of Beijing Language and Culture University Education Foundation
Funding source: MOE Project of Key Research Institute of Humanities and Social Sciences at Universities in China
Award Identifier / Grant number: 22JJD740018
Funding source: Science Foundation of Beijing Language and Culture University
Award Identifier / Grant number: 22YBB28
Acknowledgments
The authors would like to thank four anonymous reviewers and the editors for their invaluable comments and suggestions. All errors in the writing belong to the authors.
-
Research funding: This work was partly supported by the MOE Project of Key Research Institute of Humanities and Social Sciences at Universities in China (No. 22JJD740018), the Science Foundation of Beijing Language and Culture University (supported by “the Fundamental Research Funds for the Central Universities”) (No. 22YBB28), and the Innovative Practice Project of Beijing Language and Culture University Education Foundation.
References
Arnold, Jennifer E., Anthony Losongco, Thomas Wasow & Ryan Ginstrom. 2000. Heaviness vs. newness: The effects of structural complexity and discourse status on constituent ordering. Language 76(1). 28–55. https://doi.org/10.2307/417392.Search in Google Scholar
Biber, Douglas. 1988. Variation across speech and writing. London: Cambridge University Press.10.1017/CBO9780511621024Search in Google Scholar
Branigan, Holly P., Martin J. Pickering & Mikihiro Tanaka. 2008. Contributions of animacy to grammatical function assignment and word order during production. Lingua 118(2). 172–189. https://doi.org/10.1016/j.lingua.2007.02.003.Search in Google Scholar
Bresnan, Joan, Anna Cueni, Tatiana Nikitina & R. Harald Baayen. 2007. Predicting the dative alternation. In Gerlof Bouma, Irene Kramer & Joost Zwarts (eds.), Cognitive foundations of interpretation, 69–94. Amsterdam: Royal Netherlands Academy of Arts and Sciences.Search in Google Scholar
Chafe, Wallace. 1994. Discourse, consciousness, and time: The flow and displacement of conscious experience in speaking and writing. Chicago: University of Chicago Press.Search in Google Scholar
Chao, Yuan Ren. 1979. Xiandai Hanyu Kouyu Yufa [A Grammar of Spoken Chinese]. Beijing: The Commercial Press.Search in Google Scholar
Chappell, Hilary. 1986. Formal and colloquial adversity passives in standard Chinese. Linguistics 24(6). 1025–1052. https://doi.org/10.1515/ling.1986.24.6.1025.Search in Google Scholar
Chen, Ping. 2004. Identifiability and definiteness in Chinese. Linguistics 42(6). 1129–1184. https://doi.org/10.1515/ling.2004.42.6.1129.Search in Google Scholar
Cheung, Hung-nin Samuel. 1973. A comparative study in Chinese grammars: The ba-construction. Journal of Chinese Linguistics 1(3). 343–382.Search in Google Scholar
Croft, William & D. Alan Cruse. 2004. Cognitive linguistics. London: Cambridge University Press.10.1017/CBO9780511803864Search in Google Scholar
Dai, Zheyuan, Haitao Liu & Jianwei Yan. 2023. Revisiting English written VP-ellipsis and VP-substitution: A dependency-based analysis. Linguistics Vanguard 9(1). 13–23. https://doi.org/10.1515/lingvan-2022-0088.Search in Google Scholar
Fang, Yu & Haitao Liu. 2018. What factors are associated with dependency distances to ensure easy comprehension? A case study of ba sentences in Mandarin Chinese. Language Sciences 67. 33–45. https://doi.org/10.1016/j.langsci.2018.04.005.Search in Google Scholar
Fang, Yu & Haitao Liu. 2021. Predicting syntactic choice in Mandarin Chinese: A corpus-based analysis of ba sentences and SVO sentences. Cognitive Linguistics 32(2). 219–250. https://doi.org/10.1515/cog-2020-0005.Search in Google Scholar
Fang, Yu & Haitao Liu. 2022. Locality effects and predictability in sentence construction: An investigation of ba sentences and SVO sentences in Mandarin Chinese. Journal of Chinese Linguistics 50(1). 103–139. https://doi.org/10.1353/jcl.2022.0005.Search in Google Scholar
Garretson, Gregory. 2004. Coding practices used in the project optimal typology of determiner phrases. Unpublished manuscript. Boston: Boston University.Search in Google Scholar
Gennari, Silvia P., Jelena Mirković & Maryellen C. Macdonald. 2012. Animacy and competition in relative clause production: A cross-linguistic investigation. Cognitive Psychology 65(2). 141–176. https://doi.org/10.1016/j.cogpsych.2012.03.002.Search in Google Scholar
Gibson, Edward. 1998. Linguistic complexity: Locality of syntactic dependencies. Cognition 68(1). 1–76. https://doi.org/10.1016/s0010-0277(98)00034-1.Search in Google Scholar
Gibson, Edward. 2000. The dependency locality theory: A distance-based theory of linguistic complexity. In Alec P. Marantz, Yasushi Miyashita & Wayne O’Neil (eds.), Image, language, brain: Papers from the first mind articulation project symposium, 94–126. Cambridge: The MIT Press.10.7551/mitpress/3654.003.0008Search in Google Scholar
Givón, Talmy. 1992. The grammar of referential coherence as mental processing instructions. Linguistics 30(1). 5–55. https://doi.org/10.1515/ling.1992.30.1.5.Search in Google Scholar
Grodner, Daniel & Edward Gibson. 2005. Consequences of the serial nature of linguistic input for sentenial complexity. Cognitive Science 29(2). 261–290. https://doi.org/10.1207/s15516709cog0000_7.Search in Google Scholar
Guo, Hongjie & Hong Han. 2012. Yuliaoku qudong de yinghan yuyan jiechu yanjiu–yi “bei”ziju wei li [A corpus-driven study of English-Chinese language contact: A case of “bei” passive]. Waiyu Jiaoxue yu Yanjiu [Foreign Language Teaching and Research] 44(3). 359–370.Search in Google Scholar
Hammarström, Harald. 2016. Linguistic diversity and language evolution. Journal of Language Evolution 1(1). 19–29. https://doi.org/10.1093/jole/lzw002.Search in Google Scholar
Hashimoto, Mantaro J. 1988. The structure and typology of the Chinese passive construction. In Masayoshi Shibatani (ed.), Passive and voice, 329–354. Amsterdam: John Benjamins Publishing. https://benjamins.com/catalog/tsl.16.11has.10.1075/tsl.16.11hasSearch in Google Scholar
Hawkins, John A. 1994. A performance theory of order and constituency. New York: Cambridge University Press.10.1017/CBO9780511554285Search in Google Scholar
He, Yang. 2004. Cong xiandai hanyu jieci zhong de ouhua xianxiang kan jianjie yuyan jiechu [A study of the phenomenon of Europeanized preposition in Modern Chinese]. Yuyan Wenzi Yingyong [Applied Linguistics] (4). 82–89.Search in Google Scholar
He, Yang. 2008. Xiandai hanyu ouhua yufa xianxiang yanjiu [A study of the phenomenon of Europeanization in Modern Chinese grammar]. Shijie Hanyu Jiaoxue [Chinese Teaching in the World] (4). 16–31+2.Search in Google Scholar
Huang, Borong & Xudong Liao. 2002. Xiandai Hanyu Xia [Modern Chinese, Vol. 2]. Beijing: Higher Education Press.Search in Google Scholar
Huang, C.-T. James. 1999. Chinese passives in comparative perspective. Tsing Hua Journal of Chinese Studies 29(4). 423–509.Search in Google Scholar
Jaeger, T. Florian & Harry Tily. 2011. On language ‘utility’: Processing complexity and communicative efficiency. Wiley Interdisciplinary Reviews: Cognitive Science 2(3). 323–335. https://doi.org/10.1002/wcs.126.Search in Google Scholar
Keenan, Edward L. & Matthew S. Dryer. 1985. Passive in the world’s languages. In Timothy, Shopen (ed.), Language typology and syntactic description, 1, 243–281. Cambridge: Cambridge University Press. https://doi.org/10.1017/cbo9780511619427.006.Search in Google Scholar
Kirby, Simon. 1999. Function, selection, and innateness: The emergence of language universals. Oxford: Oxford University Press.10.1093/oso/9780198238119.001.0001Search in Google Scholar
Kizach, Johannes. 2014. A multifactorial analysis of the Russian adversity impersonal construction. Russian Linguistics 38(2). 205–211. https://doi.org/10.1007/s11185-014-9128-z.Search in Google Scholar
Kong, Stano. 2014. Long and short passives in English speakers’ L2 Chinese. Journal of Chinese Linguistics 42(2). 400–430.Search in Google Scholar
Levin, Beth & Malka Rappaport Hovav. 2005. Argument realization. Cambridge: Cambridge University Press.10.1017/CBO9780511610479Search in Google Scholar
Li, Charles N. & Sandra A. Thompson. 1981. Mandarin Chinese: A functional reference grammar. California: University of California Press.10.1525/9780520352858Search in Google Scholar
Li, Jinlian. 2012. Rihan Beidongju Duibi Yanjiu [A Comparative Study of Japanese and Chinese Passive Sentences]. Shandong: Shandong University Press.Search in Google Scholar
Li, Linding. 1980. “Bei”ziju [The “bei” construction]. Zhongguo Yuwen [Studies of the Chinese Language] (6). 401–412.Search in Google Scholar
Li, Shan. 1994. Xiandai Hanyu “Bei”ziju Yanjiu [The Study of “Bei” Sentences in Mandarin Chinese]. Beijing: Beijing University Press.Search in Google Scholar
Li, Wenwen. 2020. Jiyu yicun shuku de yinghanyu zhuyu jufa jiliang tezheng de duibi yanjiu [A contrastive study of quantitative syntactic properties of Chinese and English subjects based on Dependency Treebank]. Yuyan Yanjiu Jikan [Bulletin of Linguistic Studies] 25(1). 173–191+380.Search in Google Scholar
Li, Zongjiang. 2004. Hanyu beidongju de yuyi tezheng jiqi renzhi jieshi [The semantic property of Chinese passives and its cognitive explanation]. Jiefangjun Waiguoyu Xueyuan Xuebao [Journal of PLA University of Foreign Languages] 27(6). 7–11.Search in Google Scholar
Liang, Junying, Yuanyuan Fang & Haitao Liu. 2017. Dependency distance differences across interpreting types: Implications for cognitive demand. Frontiers in Psychology 8. 2132. https://doi.org/10.3389/fpsyg.2017.02132.Search in Google Scholar
Liu, Feng-Hsi. 2011. The bei passive and its discourse motivations. Chinese Language and Discourse 2(2). 198–231. https://doi.org/10.1075/cld.2.2.03liu.Search in Google Scholar
Liu, Haitao, Richard Hudson & Zhiwei Feng. 2009. Using a Chinese treebank to measure dependency distance. Corpus Linguistics and Linguistic Theory 5(2). 161–174. https://doi.org/10.1515/cllt.2009.007.Search in Google Scholar
Liu, Haitao. 2008. Dependency distance as a metric of language comprehension difficulty. Journal of Cognitive Science 9(2). 159–191. https://doi.org/10.17791/jcs.2008.9.2.159.Search in Google Scholar
Liu, Haitao. 2018. Language as a human-driven complex adaptive system. Physics of Life Reviews 26(27). 149–151. https://doi.org/10.1016/j.plrev.2018.06.006.Search in Google Scholar
Liu, Yuehua, Wenyu Pan & Shuhua Zhao. 2019. Shiyong Xiandai Hanyu Yufa Di San Ban [Practical Grammar of Modern Chinese, 3rd Ed]. Beijing: The Commercial Press.Search in Google Scholar
Manning, Christopher D., Mihai Surdeanu, John Bauer, Jenny Finkel, Steven J. Bethard & David McClosky. 2014. The Stanford CoreNLP natural language processing toolkit. In Kalina Bontcheva & Jingbo Zhu (eds.), Proceedings of 52nd annual meeting of the association for computational linguistics: System demonstrations, 55–60. Baltimore: Association for Computational Linguistics.10.3115/v1/P14-5010Search in Google Scholar
Mu, Xin & Danqing Chen. 2013. Wenxue Huiyilu [Memoirs of Literature]. Guilin: Guangxi Normal University Press.Search in Google Scholar
Nicenboim, Bruno, Shravan Vasishth, Carolina Gattei, Mariano Sigman & Reinhold Kliegl. 2015. Working memory differences in long-distance dependency resolution. Frontiers in Psychology 6. 312. https://doi.org/10.3389/fpsyg.2015.00312.Search in Google Scholar
Niu, Rouchen & Haitao Liu. 2022. Effects of syntactic distance and word order on language processing: An investigation based on a psycholinguistic treebank of English. Journal of Psycholinguistic Research 51(5). 1043–1062. https://doi.org/10.1007/s10936-022-09878-4.Search in Google Scholar
Ouyang, Jinghui, Jingyang Jiang & Haitao Liu. 2022. Dependency distance measures in assessing L2 writing proficiency. Assessing Writing 51. 100603. https://doi.org/10.1016/j.asw.2021.100603.Search in Google Scholar
Peng, Yongmei & Yuen Gan. 2010. “Bei VV”: Yizhong xinxing de beidong geshi [“Bei VV”: A new passive format]. Zhongguo Yuwen [Studies of the Chinese Language] (1). 57–58.Search in Google Scholar
Pickering, Martin J., Janet F. McLean & Holly P. Branigan. 2013. Persistent structural priming and frequency effects during comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition 39(3). 890. https://doi.org/10.1037/a0029181.Search in Google Scholar
Pinker, Steven. 2007. Language as an adaptation by natural selection. Acta Psychologica Sinica 39(3). 431–438.Search in Google Scholar
Rochemont, Michael. 2013. Discourse new, F-marking, and normal stress. Lingua 136. 38–62. https://doi.org/10.1016/j.lingua.2013.07.016.Search in Google Scholar
Rosenbach, Anette. 2005. Animacy versus weight as determinants of grammatical variation in English. Language 81(3). 613–644. https://doi.org/10.1353/lan.2005.0149.Search in Google Scholar
Santazilia, Ekaitz. 2022. Animacy and inflectional morphology across languages. Netherlands: Brill.10.1163/9789004513068Search in Google Scholar
Sharma, Kartik, Richard Futrell & Samar Husain. 2020. What determines the order of verbal dependents in Hindi? Effects of efficiency in comprehension and production. In Emmanuele, Chersoni, Nora, Hollenstein, Cassandra, Jacobs, Yohei, Oseki, Laurent, Prévot & Enrico, Santus (eds.), Proceedings of the workshop on cognitive modeling and computational linguistics, 1–10. Stroudsburg: Association for Computational Linguistics.10.18653/v1/2020.cmcl-1.1Search in Google Scholar
Siewierska, Anna. 2013. Passive constructions. In Matthew, S. & Martin Haspelmath (eds.), The world atlas of language structures online. Leipzig: Max Planck Institute for Evolutionary Anthropology. https://doi.org/10.5281/zenodo.7385533.Search in Google Scholar
Tang, Sze-Wing. 2001. A complementation approach to Chinese passives and its consequences. Linguistics 39(2). 257–295. https://doi.org/10.1515/ling.2001.011.Search in Google Scholar
Thompson, Sandra A. 1987. The passive in English: A discourse perspective. In Robert Channen & Linda Shockey (eds.), Honor of Ilse Lehiste: Ilse Lehiste Puhendusteos, 497–511. Dordrecht: Foris Publications.10.1515/9783110886078.497Search in Google Scholar
Velnić, Marta. 2019. The influence of animacy, givenness, and focus on object order in Croatian ditransitives. Studia Linguistica 73(1). 175–201. https://doi.org/10.1111/stul.12094.Search in Google Scholar
Wang, Canlong. 2009. “Bei”zi de linglei yongfa: Cong “bei zisha” tan qi [The special usage of the word “bei”: From “bei suicide”]. Yuwen Jianshe [Language Planning] (4). 65–66+2.Search in Google Scholar
Wang, Hua. 2018. Hanying mingci duanyu changdu de jiliang yanjiu [A quantitative study on the length of Chinese and English noun phrases]. In Haitao Liu (ed.), Jiliang Yuyanxue Yanjiu Jinzhan [Advances in Quantitative Linguistics], 268–280. Hangzhou: Zhejiang University Press.Search in Google Scholar
Wang, Huan. 1983. Yingyu he hanyu de beidongju [The passive construction in English and Mandarin Chinese]. Zhongguo Yuwen [Studies of the Chinese Language] (6). 409–418.Search in Google Scholar
Wang, Huan. 1984. “Ba”ziju He “Bei”ziju [The “Ba” Construction and the “Bei” construction]. Shanghai: Shanghai Education Press.Search in Google Scholar
Wang, Li. 1957. Hanyu beidongshi de fazhan [The development of the passive construction]. Yuyanxue Luncong [Essays on Linguistics] 1. 1–16.Search in Google Scholar
Wang, Li. 1980. Hanyu Shigao [The Outline of Chinese History]. Beijing: Zhonghua Book Company.Search in Google Scholar
Wang, Li. 1985. Zhongguo Xiandai Yufa [Modern Chinese Grammar]. Beijing: The Commercial Press.Search in Google Scholar
Warren, Tessa & Edward Gibson. 2002. The influence of referential processing on sentence complexity. Cognition 85(1). 79–112. https://doi.org/10.1016/s0010-0277(02)00087-2.Search in Google Scholar
Wasow, Thomas. 1997. Remarks on grammatical weight. Language Variation and Change 9(1). 81–105. https://doi.org/10.1017/s0954394500001800.Search in Google Scholar
Wolk, Christoph, Joan Bresnan, Anette Rosenbach & Benedikt Szmrecsanyi. 2013. Dative and genitive variability in Late Modern English: Exploring cross-constructional variation and change. Diachronica 30(3). 382–419. https://doi.org/10.1075/dia.30.3.04wol.Search in Google Scholar
Xiao, Richard, Tony McEnery & Yufang Qian. 2006. Passive constructions in English and Chinese: A corpus-based contrastive study. Languages in Contrast 6(1). 109–149. https://doi.org/10.1075/lic.6.1.05xia.Search in Google Scholar
Xiong, Xueliang & Zhijun Wang. 2002. Beidong jushi de yuanxing yanjiu [Research on the prototype of passive sentence pattern]. Waiyu Yanjiu [Foreign Languages Research] (1). 19–23.Search in Google Scholar
Xiong, Yonghong & Rong Zeng. 2011. Cong moyinlun de jiaodu jiedu liuxingyu “bei XX” [Interpreting the catchword “bei XX” from the perspective of memetics]. Waiguo Yuwen [Foreign Language and Literature] 27(2). 70–73.Search in Google Scholar
Xu, Chunshan & Haitao Liu. 2015. Can familiarity lessen the effect of locality? A case study of Mandarin Chinese subjects and the following adverbials. Poznan Studies in Contemporary Linguistics 51(3). 463–485. https://doi.org/10.1515/psicl-2015-0018.Search in Google Scholar
Xu, Dekuan. 2004. Xiandai Hanyu Shuangbin Gouzao Yanjiu [Research on the Double Object Structure of Modern Chinese]. Shanghai: Shanghai Lexicographical Publishing House.Search in Google Scholar
Xu, Jiajin. 2014. “ToRCH2009 xiandai hanyu yuliaoku” jiancheng [Construction of “ToRCH2009 Modern Chinese corpus”]. Yuliaoku Yuyanxue [Corpus Linguistics] 1(1). 103–104.Search in Google Scholar
Xue, Xiaoying & Derong Zhang. 2021. Gainian zhenghe shijiao xia xin “bei XX” goushi de yuyi yanjiu [A semantic analysis of the new “bei” construction from the perspective of the conceptual blending theory]. Wuling Xuekan [Journal of Wuling] 46(4). 109–114.Search in Google Scholar
Zeng, Tao, Yuxi Li & Mingjun Wu. 2020. Syntactic and semantic processing of passive BEI sentences in Mandarin Chinese: Evidence from event-related potentials. Neuroreport 31(13). 979–984. https://doi.org/10.1097/wnr.0000000000001507.Search in Google Scholar
Zhang, Hongming. 1994. The grammaticalization of “bei” in Chinese. Chinese Languages and Linguistics 2. 321–360.Search in Google Scholar
Zhu, Yifan & Kaibao Hu. 2014. “Bei”ziju de yuyi quxiang yu yuyiyun: Jiyu fanyi yu yuanchuang xinwen yuliaoku de duibi yanjiu [The semantic preferences and semantic prosody of “bei” passives: A corpus-based contrastive study]. Waiguo Yu [Journal of Foreign Languages] 37(1). 53–64.Search in Google Scholar
Ziegeler, Debra. 2000. A possessive-based analysis of the ba-construction in Mandarin Chinese. Lingua 110(11). 807–842. https://doi.org/10.1016/s0024-3841(00)00018-8.Search in Google Scholar
Zipf, George Kingsley. 1949. Human behavior and the principle of least effort: An introduction to human ecology. Cambridge: Addison-Wesley Press.Search in Google Scholar
© 2024 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.
Articles in the same Issue
- Frontmatter
- Cognitive and sociolectal constraints on the theme-recipient alternation: evidence from Mandarin
- A corpus-based study on semantic and cognitive features of bei sentences in Mandarin Chinese
- The genitive alternation in German
- Idiosyncratic entrenchment: tracing change in constructional schematicity with nested random effects
- Verb influence on French wh-placement: a parallel corpus study
- The red dress is cute: why subjective adjectives are more often predicative
- The diachronic change of English relativizers: a case study in the State of the Union addresses across two centuries
Articles in the same Issue
- Frontmatter
- Cognitive and sociolectal constraints on the theme-recipient alternation: evidence from Mandarin
- A corpus-based study on semantic and cognitive features of bei sentences in Mandarin Chinese
- The genitive alternation in German
- Idiosyncratic entrenchment: tracing change in constructional schematicity with nested random effects
- Verb influence on French wh-placement: a parallel corpus study
- The red dress is cute: why subjective adjectives are more often predicative
- The diachronic change of English relativizers: a case study in the State of the Union addresses across two centuries
