An analysis of register variances between China and the US in international communications

Abstract

Translation of international communications plays an important role in facilitating the spread of a country’s voice and shaping its national image. This study employs multidimensional analysis to analyze China’s interpretation of international communications and compares it with the speeches from the White House press conferences to uncover the register variances between the two. The findings reveal that China and the U.S. differ in 4 dimensions and 46 individual linguistic features. The former is more informational with explicit written features and can be categorized into the register of learned exposition; the latter is more interactive and persuasive with prominent features of colloquialism, belonging to the register of involved persuasion. Through comparison with the source language, we find that the differences between the two are influenced by multiple factors such as source language, interpreting process, and political considerations. In conclusion, the interpretation of China’s international communications is a “third language code” with distinctive features from the language used by the U.S. in the political scenario, reflecting a meticulous, responsible but less interactive national image. These findings could shed light on the interpreting strategies adopted by government interpreters and further improve the effectiveness of international communications.

Keywords:

• Translation of international communications
• register features
• multidimensional analysis
• national image shaping

REVIEWING EDITOR:

• Maria del Mar Sanchez Ramos, Modern Philology, University of Alcala de Henares: Universidad de Alcala de Henares, Madrid, Spain

SUBJECTS:

• Political Communication
• Language & Linguistics
• Literature & Translation

1. Introduction

To some extent, the national image of a country is closely intertwined with how it presents itself to the world. In this era of globalization, China, as a non-English-speaking country, has actively engaged with the international community, striving to project a peaceful and responsible national image. However, comprehending how China is perceived globally is a multifaceted task. Due to the language barrier, media in the English world has wielded significant influence in shaping its global perception. Today, the translation of international communications guided by the Chinese government serves as a crucial avenue for the country to convey its message to the world. In a more direct, instantaneous and interactive way, interpretation plays a vital role in shaping the national image of China.

The interpretation of international communications resides at the intersection of translation studies and communication studies. The convergence of the two disciplines offers valuable insights into the formation of China’s and other countries’ national images. While extensive comparative analysis has explored the translation of government documents, the language used in interpretation has, regrettably, often been overlooked. As an interpreted language, China’s interpretation of international communications is bound to present linguistic features that are different from those of native English. To bridge this gap, this paper conducts a comparative analysis between the interpreted language used by the interpreters of the Chinese government and the original language used by the White House spokesperson based on Douglas Biber’s (1988) framework of Multidimensional Analysis (MDA). It aims to address three research questions: (1) What distinctions exist between China and the U.S. in their international communications concerning microscopic linguistic features and macroscopic register features? (2) What types of national images are being constructed? (3) What are the potential sources of these linguistic differences between China and the U.S.? The primary objective is to elucidate the interpreting strategies utilized by government interpreters and improve the effectiveness of international communications.

2. Multidimensional analysis method

The term “register” was first introduced by Reid (1956: 32), referring to speakers “on different occasions speak or write differently according to what may roughly be described as different social situations”. “Early register studies accounted for the fact that language use varies in response to underlying relationships between linguistic and functional parameters” (Goulart et al., 2020), but there was still no systematic framework for register analysis. Subsequent theories proposed by Halliday (1978); Brown and Fraser (1979) laid a strong foundation for Douglas Biber who proposed the ground-breaking multi-dimensional analysis (MDA) framework in 1988. According to Biber & Conrad (2019: 79), register analysis involves three steps: describing situational characteristics, identifying distinctive linguistic characteristics, and showing how the situational and linguistic descriptions are related.

Biber’s MDA was initially proposed to investigate register variations across speech and writing, in which he extracted 481 texts from London-Lund and LOB corpora and identified 67 linguistic features for microscopic analysis. As outlined by Biber (1988), he quantified the occurrences of these features in the texts, and further classified them into seven factors through factor analysis. Factor analysis reduces a large number of original variables (linguistic features) to a small set of underlying variables—the factors or dimensions of variation. Each dimension represents a group of linguistic features that tend to co-occur in texts (Biber & Conrad, 2019: 270). The 7 factors are interpreted as different “dimensions” as illustrated in Table 1 based on their functions, assuming that co-occurring linguistic features in texts serve shared communicative functions (Biber 2006: 179). “Each linguistic feature’s relation to each factor is represented by factor loading” (Biber, 1988: 111) with stronger loadings indicating a greater influence on the interpretation of each factor and its function. The function of individual linguistic features is scrutinized by evaluating the positive or negative "factor loading" assigned to them. Factor loading serves as an assessment of the contribution each language feature makes to the overall factor or dimension. A positive factor loading indicates a positive correlation, while a negative factor loading implies a negative correlation. Since the loadings of linguistic features in dimension 7 are not so strong for a firm interpretation, it is not considered in his study. Biber’s MDA approach, functioning as a systematic framework for register analysis, empowers researchers to attain a deeper understanding of the intricacies of language use by establishing connections between speech situations, linguistic features, and their respective functions.

Table 1. Names of dimensions.

Dimension
Dimension 1: Involved versus informational production
Dimension 2: Narrative versus non-narrative concerns
Dimension 3: Explicit versus situation-dependent reference
Dimension 4: Overt expression of persuasion
Dimension 5: Abstract versus non-abstract information
Dimension 6: Online informational elaboration
Dimension 7: Academic hedging

According to Brown and Fraser (1979: 38–39) “a focus on isolated markers without considering their systemic variances can be difficult and even misleading.” Over the past three decades, multidimensional analysis (MDA) has expanded from studies of register variation across speech and writing to more detailed fields, including: register variation studies (Grieve et al., 2011; Pinto, 2014), variation in academic discourse (Atkinson, 2001; Conrad, 2001; Gardner et al., 2019; Gray, 2015), variations in dialect and gender language (Biber & Burges, 2001; Grieve, 2014; Rey, 2001), diachronic register variation studies (Biber & Finegan, 2001; Souza, 2014; Westin & Geisler, 2002), and cross-language variation studies (Biber, 1995; Sardinha, Kauffmann & Acunzo, 2014) (Pan, 2022: 26–30). Besides, MDA is also increasingly applied in translation studies.

As this paper delves into the interpreted language utilized by the Chinese government through the application of MDA, it becomes imperative to undertake a comprehensive review of MDA within the Chinese context. Most of the studies applying the MDA framework are introductory articles (Wu, 2001; Rong, 2007) or focus on academic genres (Lei & Yang, 2001; Wu, 2004; Liu & Hui, 2010; Pan, 2012; Xiao & Cao, 2014; Cao & Xiao, 2014; Cao & Xiao, 2015). Only a few researchers designed corpora to investigate specialized fields like business English (Jiang & Xu, 2015), museum English (Hu & Song, 2022), and letters to shareholders (Hu & Tan, 2020). While MDA is not a novel approach among Chinese researchers, the attention dedicated to interpreted language within existing studies remains notably inadequate. A few studies have investigated register features in a translated language such as Hu (2010), Pang (2020), and Hu & Song (2022). Among them, Li (2014) is the only researcher who has analyzed register features in diplomatic interpretation, comparing them with other types of discourse such as face-to-face communication and interviews. Hence, there exists an unexplored research gap in the utilization of the MDA framework within the context of interpreted languages for international communications.

3. Date and research tool

3.1 Corpus profile

Table 2 presents a description of the three corpora analyzed in this study. To ensure data representativeness, this study focused on texts from two sources: the Chinese government website (www.gov.cn) and the White House official website, covering the period of 2018–2022. These two websites are considered highly formal channels for information dissemination in both countries, sharing similar characteristics in format, content, timing, and formality. The data were collected from speeches delivered by high-ranking officials such as the President, Premier and ministers and other government representatives over the past five years. The Corpus of Interpreting for the International Communications of China (CIICC) comprises 92,640 tokens of interpreted texts, while the Corpus of Original Chinese for the International Communications of China (CIICC) contains 134,320 characters of Chinese texts. Additionally, the White House Press Conference Corpus (CWHPC) consists of 50 texts with 147,483 tokens in total. The situational parameters of the texts in the corpora remain consistent, and as a result, situational analysis is not covered in the following sections.

Table 2. Composition of corpora.

Name	Language	Size	Number of texts	Content source
CIICC	English	92,640	30	China Government Network
COICC	Chinese	134,320	30	China Government Network
CWHPC	English	147,483	50	White House Official Website

3.2 Research tool

The present study applied Biber’s MDA framework and utilized the Multidimensional Analysis Tagger 1.3 (MAT), a specialized tagging tool developed by Nini (available at: http://sites.google.com/site/ multidimensionaltagger) that incorporates 67 linguistic features and 6 functional dimensions, as selected by Biber (1988). It streamlines the analysis process by integrating the Stanford POS Tagger and enables automated lexical tagging and data computation. Finally, MAT can report statistics such as dimension scores, frequency of linguistic features and the closest genre, etc. All linguistic features on each dimension are identified based on the classification of Biber (1988). According to the statistics shown in the manual of MAT, it has been tested for reliability on the LOB and the Brown corpora. The results demonstrate that MAT is largely successful in replicating Biber’s (1988) analysis. Since MAT is designed to process English texts, this study compares the register and 67 linguistic features of the two English corpora in 6 dimensions, and independent sample t-tests were conducted using SPSS to identify the register and linguistic features that show significant differences (p < 0.05). In addition, the frequency of linguistic features with significant differences was further examined in COICC in 4.3 to trace the source of these disparities.

4. Results and discussion

4.1 Comparison of CIICC and CWHPC’s dimension scores

The MAT generates two sets of dimensional scores after processing the CIICC and CWHPC, which have been standardized. The results of the t-test in Table 3 show that CIICC and CWHPC are significantly different (p < 0.05) in dimension 1, dimension 3, dimension 4, and dimension 6, and do not show significant differences in dimension 2 (t = -1.930, p = 0.057 > 0.05) and dimension 5 (t = 1.130, p = 0.262 > 0.05).

Table 3. t-Test for dimensional differences between CIICC and CWHPC.

Dimension	CIICC	CWHOC	t	df	Sig	MAD
Dimension 1: involved versus. informational production	−18.56	1.49	−1.053	78	0.000	1.995
Dimension 2: narrative versus non-narrative concerns	−2.99	−2.25	−1.930	78	0.057	0.381
Dimension 3: explicit versus situation dependent reference	13.43	2.64	14.070	78	0.000	0.766
Dimension 4: overt expression of persuasion	2.98	1.07	3.246	78	0.002	0.586
Dimension 5: abstract versus non-abstract information	−0.48	−0.88	1.130	78	0.262	0.359
Dimension 6: online information elaboration organization at the moment	−1.56	0.32	−6.434	78	0.000	0.292

Figure 1 illustrated the differences between CIICC and CWHPC at the macro level across six dimensions. The most substantial difference was observed in dimension 1, with a negative value of −18.56 for CIICC and a positive value of 1.49 for CWHPC (t = −10.053, p < 0.001). A positive value on this dimension indicates a focus on interactivity, while a negative one indicates a focus on delivering information. Thus, CCIIC is more informative but less interactive, while CWHPC exhibits more interactivity but is less informative. Another significant difference was found in dimension 3, where the scores for both CIICC and CWHPC are positive, but the score for CIICC (13.43) was much higher than that for CWHOC (2.64) (t = −1.930, p < 0.001). This suggests that both are non-situational dependent, but the former is more explicit in comparison. In dimension 4, both CIICC and CWHPC had positive values (2.98 and 1.07, respectively), meaning that both were persuasive, but the former was more persuasive in comparison (t = 3.246, p < .05). In dimension 6, CIICC had a negative value of -1.56, while CWHPC had a positive value of 0.32, with a significant difference between the two (t = −6.434, p < 0.001), indicating that CWHPC was more improvisational and more interactive. In addition, there were no significant differences between the two in dimension 2, "narrative vs. non-narrative concerns" (t = −1.930, p > 0.05) and dimension 5, "degree of abstraction vs. specificity of information" (t = 1.130, p > .05). Finally, the report by MAT showed that CIICC belongs to the register of learned exposition with a focus on conveying information, while CWHPC belongs to the register of involved persuasion which emphasizes interactions.

4.2 Interpretation of the differences in specific linguistic features between CIICC and CWHPC

“A cluster of features co-occurs frequently in texts because they are serving some common function in those texts. At this point, micro-analyses of linguistic features become crucially important” (Biber, 1988). To further distinguish the differences between CIICC and CWHPC in terms of micro-linguistic features, we conducted independent samples t-tests on the standardized scores. The results showed that 46 out of 67 features were significantly different (p <  0.05), accounting for approximately 68.66%. According to the classification of Biber (1988), 23 linguistic features in dimension 1, 5 in dimension 2, 7 in dimension 3, 5 in dimension 4, 2 in dimension 5, and 4 in dimension 6 were significantly different. Since there was no significant difference between the two corpora in dimensions 2 and 5, the following analysis of linguistic features only involved the remaining four dimensions.

As shown in Table 4, a total of 23 linguistic features showed significant differences in dimension 1, which accounted for 41.07% of the overall different features. In terms of dimension score, CIICC is found to be more informative, while CWHPC is more interactive. In CIICC, these differences are reflected in various features such as longer average word length, more attributive adjectives, more prepositional phrases, fewer abbreviations, indicative pronouns, first and second person pronouns, pronouns, private verbs and possibility modals. All the above features demonstrate the accuracy of the language and a deliberate effort to reduce subjectivity in the language of CIICC. As a result, CIICC is more formal and carries features of written documents and are stronger in terms of informativeness. For instance, the usage of first-person pronouns (I, we, me, and us) and private verbs (such as believe, decide, observe, and suppose) is reduced to decrease subjectivity. On the other hand, more attributive adjectives and prepositional phrases are used to enhance specificity and accuracy. In example 1, the speaker used several attributive adjectives (multilateral, common, differentiated, and respective) to modify corresponding nouns, in an attempt to make the expression more explicit and specific. In such an open and formal scenario, any ambiguity should be eliminated by using certain linguistic features.

Table 4. Linguistic features with significant differences in dimension 1.

Individual features	Code	CIICC	CWHPC	Sig	MAD
Independent clause coordination	ANDC	−0.559	−0.275	0.000	0.284
Average word length	AWL	2.109	0.354	0.000	1.755
Be as main verb	BEMA	−1.951	−1.279	0.000	0.672
Contraction	CONT	−0.701	0.312	0.000	1.014
Demonstrative pronouns	DEMP	−0.441	0.883	0.000	1.324
Discourse particles	DPAR	−0.484	−0.019	0.000	0.465
Emphatics	EMPH	−0.407	0.633	0.000	1.040
First person pronouns	FPP1	0.017	0.710	0.000	0.693
Attributive adjectives	JJ	2.645	−0.246	0.000	2.891
Possibility modals	POMD	−1.030	−0.469	0.000	0.561
Private verb	PRIV	−1.019	−0.239	0.000	0.780
Pro-verb do	PROD	−0.684	−0.128	0.000	0.556
Second person pronouns	SPP2	−0.545	0.369	0.000	0.914
Stranded preposition	STPR	−0.652	−0.232	0.000	0.420
Subordinator that deletion	THATD	−0.616	−0.110	0.000	0.506
Type-token ratio	TTR	0.812	−.097	0.000	0.909
Present tense	VPRT	−1.103	−0.370	0.000	0.733
WH-clauses	WHCL	−0.540	0.304	0.000	0.844
Analytic negation	XX0	−1.059	−0.530	0.000	0.529
Causative adverbial subordinators	CAUS	−0.561	0.291	0.001	0.852
Pronoun it	PIT	−0.611	−0.032	0.001	0.580
Present participial WHIZ deletion relatives	WZPRES	0.592	−0.004	0.001	0.596
Total prepositional phrases	PIN	−0.131	−0.491	0.002	0.360

Example 1

我们将不断加强绿色合作, 支持和践行多边主义, 坚持共同但有区别的责任原则、公平原则和各自能力原则, 积极履行应对气候变化《巴黎协定》。

Interpretation: We will continue to support and follow a multilateral approach, stick to the principles of common but differentiated responsibilities, equity and respective capabilities, and actively implement the Paris Agreement on climate change.

from CIICC

In contrast, the usage of the present tense in CWHPC is significantly higher than in CIICC, accompanied by more frequent use of first-person pronouns, second-person pronouns and the pronoun it in the discourse. In example 2, the use of the present tense and second-person pronouns highlights stronger narrative and interactive features, thereby giving listeners a stronger sense of involvement.

Example 2

These election deniers are not only trying to deny you your right to vote. They’re trying to deny your right to have your vote counted.

from CWHPC

Several factors may contribute to this disparity. First, occasions with interpreters in China are typically formal scenarios with clear themes; second, Chinese interpreters all have received professional training and are familiar with political documents, which likely contributes to the more informative and formal language used in CIICC. As a result, China’s interpretation of international communications is more informative and formal; it also reflects that the interpretated language tends to be less interactive than that of the White House press conferences. In sum, the Chinese government places greater emphasis on releasing authoritative information and is more meticulous in its use of language, while the U.S. government is more interactive and more relaxed in its communication style.

Regarding dimension 3, both CIICC and CWHPC have positive dimension scores, indicating that neither is situation dependent. However, CWHPC appears to be more situation dependent than CIICC based on their dimension scores. This is evidenced by the frequency of TIME and PLACE shown in Table 5, which according to Biber (1988:110) are considered negative linguistic features as they require external information for understanding. Thus, the higher the score on these features the more situation dependent the language is. CWHPC received higher scores on these features. On the other hand, WHSUB is a positive linguistic feature, indicating that higher scores correspond to better performance in explicitness. CWHPC scored higher than CIICC on this feature, but both were below average. Among the other positive linguistic features, CIICC scored higher than CWHPC in PHC and NOMZ. PHC refers to any and that is preceded or followed by words with the same part of speech, like nouns, verbs, adjectives, and adverbs. NOMZ refers to noun structures ending in -tion, -ment, -ness, or -ity. As Biber(1988:110) pointed out “the co-occurrence of phrasal coordination and nominalizations with these relativization features indicates that referentially explicit discourse also tends to be integrated and informational.” In general, both CIICC and CWHPC are explicit in their language, but the CWHPC is relatively more situation dependent compared to CIICC.

Table 5. Linguistic features with significant differences in dimension 3.

Individual features	Code	CIICC	CWHPC	Sig	MAD
Phrasal coordination	PHC	8.368	1.985	0.000	6.383
Nominalization	NOMZ	3.246	0.830	0.000	2.416
Total adverbs	RB	−2.447	−1.683	0.000	0.764
Time adverbials	TIME	−0.923	0.440	0.000	1.363
WH relative clauses on subject position	WHSUB	−0.868	−0.384	0.000	0.484
Place adverbials	PLACE	−0.128	0.302	0.001	0.430
WH relative clauses on object position	WHOBJ	−0.814	−0.726	0.011	0.089

In dimension 4, overt expression of persuasion, the scores of both corpora are positive, revealing that all texts are persuasive. This is reflected in several individual linguistic features in Table 6. For instance, the frequency of NEMD is significantly higher in CIICC than in CWHPC (p < 0.001), demonstrating that the interpretated language is more assertive, with more words that reflect determination and the willingness to take actions, such as should, must, and ought. See examples 3 and 4.

Table 6. Linguistic features with significant differences in dimension 4.

Individual features	Code	CIICC	CWHPC	Sig	MAD
Necessity modals	NEMD	1.765	−0.460	0.000	2.225
Conditional adverbial subordinators	COND	−0.964	−0.379	0.000	0.585
Infinitives	TO	1.375	2.117	0.005	0.742
Split auxiliaries	SPAU	−0.453	−0.860	0.013	0.407
Predictive modals	PRMD	1.166	0.534	0.028	0.633

Example 3

发挥主管部门合作机制作用, 妥善应对毒品贩运、跨国有组织犯罪等非传统安全威胁, 为各国经济社会发展提供保障。

Interpretation: Our cooperation mechanisms between the competent authorities must live up to their role in tackling unconventional security threats, including Our cooperation mechanisms between the competent authorities must live up to their role in tackling unconventional security threats, including drug trafficking and cross-border organized crimes, to provide safeguards for our economic and social development.

from CIICC

Example 4

中国和东盟国家坚持由直接当事国磋商谈判和平解决争端, 管控分歧, 推进共同开发, 为地区和平稳定做出了积极贡献。

from CIICC

Interpretation: China and ASEAN countries have maintained that the disputes should be peacefully settled, differences managed and joint development advanced, China and ASEAN countries have maintained that the disputes should be peacefully settled, differences managed and joint development advanced, through consultation and negotiation, by the countries directly concerned, and have made positive contributions to peace and stability in the region.

In example 3, the Chinese side clearly states its stance on the international issues under discussion by using NEMD such as must and should in its interpretation. The word our is also used to modify cooperation mechanisms, trying to convey a message that the speaker and the audience are on the same side. In example 4, the word should also reflects China’s stance of supporting the peaceful resolution of disputes through consultation and negotiation. This reflects the strong persuasive nature of CIICC.

At the same time, there is a significant difference in the use of PRMD (p < 0.05). CIICC uses more PRMD (such as will, would, shall) to envisage a promising future. In contrast, CWHPC uses more COND and POMD (such as could, might, may), indicating a strong sense of possibilities, which are significantly different from CIICC (p < 0.001). Adverbial clause of condition, using if, unless, whether, etc. to give the circumstances under which something will take place, is paired with infinitives to depict a bright vision, and it is more moderate in tone but also with strong persuasive implications. Compared with CIICC, these linguistic features serve to bridge the gap between spokespersons of the U.S. government and their intended audience.

Dimension 6 refers to on-line informational elaboration. High scores in this dimension signify that the text possesses an informational nature but was created within specific time constraints. CWHPC had a positive dimension score (0.32) and CIICC had a negative one (−1.56), with significant differences between the two (p < 0.001). This discrepancy is supported by four linguistic features in Table 7. The positive dimension score shows that CWHPC is more impromptu and interactive, while the negative dimension score shows that CIICC is not impromptu and interactive enough. According to Biber (1988:113), that complements to verbs that complements to adjectives, and that relative clauses on object positions all have positive weight in Dimension 6, while the only feature with a salient negative weight is phrasal coordination. From Table 7, we can see that CWHPC is higher than CIICC in all four linguistic features with positive weight, which means that texts from CWHPC tend to be done under real-time conditions. In example 5, the use of TOBJ conveys the speaker’s personal stance and adds more personal emotion into the discourse.

Table 7. Linguistic features with significant differences in dimension 6.

Individual features	Code	CIICC	CWHPC	Sig	MAD
Demonstratives	DEMO	−0.977	0.592	0.000	1.569
That relative clauses on object position	TOBJ	−0.380	1.087	0.000	1.467
That relative clauses on subject position	TSUB	0.567	2.551	0.000	1.983
That verb complements	THVC	−0.581	−0.270	0.028	0.311

Example 5

And the thing about Kathy Hochul is she has shown you who she is. She has shown you what she cares about. And I believe that when you know what you stand for, you know what to fight for.

from CWHPC

Biber (1988:114) points out that “cohesion in unplanned informational discourse relies heavily on demonstratives”. In this regard, CWHPC has a better performance. The weak performance of CIICC in this dimension could be attributed to various factors, including that many speeches delivered by Chinese officials are not impromptu, as well as the style of interpreters and the interpreting strategies they adopted. It is worth noting that Biber did not consider filler words and re-phrasing corrections as linguistic features in dimension 6, although they are important markers of impromptu speech. This is exemplified in example 6 extracted from CWHPC.

Example 6

The President has – as I just mentioned at the top, as you just said, Phil – a series of meeting with – he ‘s going to have a series of meetings with leadership in Congress to talk about in a – in a – you know, a range of issues that matter to the American people. that matter to the American people.

from CWHPC

4.3 Comparison of linguistic features among corpora

To investigate the source of the linguistic disparities between CIICC and CWHPC, this study selected four linguistic features coexisting in English and Chinese as examples, namely adjectives, prepositions, first-person pronouns and second-person pronouns in consideration of language comparability. The results for each linguistic facture are presented respectively by box plots in Figure 2.

Figure 1. Corpus dimension scores.

Figure 2. Comparison of the proportion of four individual linguistic features.

Regarding adjectives, CIICC exhibits the highest proportion, followed by COICC and CWHPC. This suggests that the frequency of adjective in the original Chinese has been further increased in the interpreted language. In terms of prepositions, CIICC shows a significantly higher proportion compared with COICC, and has surpassed the proportion in CWHPC. That is consistent with the linguistic scores reported by MAT and could be partly attributed to some political considerations aiming at improving the accuracy of their speeches. Besides, COICC is the lowest in the use of first-person pronouns and second-person pronouns, which is also consistent with the reported scores with the aim to reduce subjectivity in international communications. While both features increase in CIICC, they still differ significantly from those in CWHPC. To test for significant differences among the three corpora, this paper conducted a Kruskal-Wallis test on the four linguistic features, followed by a Mann-Whitney U test for multiple comparisons. The results are presented in Tables 8 and 9, respectively:

Table 8. Kruskal–Wallis test for four individual features.

Individual features	Mean	SD	F	Sig
Adjective	8.024	2.724	H = 82.194	0.000
Preposition	8.269	3.291	H = 69.633	0.000
First person pronoun	3.223	1.812	H = 66.757	0.000
Second person pronoun	0.759	0.954	H = 65.346	0.000

Table 9. Mann–Whitney U test for four individual linguistic features.

Individual Features	Corpus	Mean rank	Z	Sig
Adjective	COICC vs. CIICC	20.42 vs.40.57	−4.465	0.000
	CIICC vs. CWHPC	65.50 vs.25.50	−7.454	0.000
	COICC vs. CWHPC	63.60 vs.26.64	−6.887	0.000
Preposition	COICC vs. CIICC	15.50 vs.45.50	−6.653	0.000
	CIICC vs. CWHPC	50.55 vs.34.47	−2.997	0.003
	COICC vs. CWHPC	15.50 vs.55.50	−7.454	0.000
First person Pronoun	COICC vs. CIICC	18.48 vs.42.52	−5.204	0.000
	CIICC vs. CWHPC	22.67 vs.51.20	−5.527	0.000
	COICC vs. CWHPC	17.33 vs.54.40	−6.907	0.000
Second person Pronoun	COICC vs. CIICC	18.77 vs.42.23	−5.432	0.000
	CIICC vs. CWHPC	21.97 vs.51.62	−5.341	0.000
	COICC vs. CWHPC	18.43 vs.53.74	−6.659	0.000

According to the Kruskal–Wallis test, the three corpora differed significantly (p < 0.001) in all four linguistic features, and the Mann–Whitney U test confirmed a significant difference (p < 0.01) in all pairwise comparisons. These findings are consistent with Frawley’s (2012: 257) "third code" theory, which suggests that “the translation itself is essentially a third code which arises out of the bilateral consideration of the matrix and target codes: it is, in a sense, a subcode of each of the codes involved”. The box plots show that while CIICC is closer to COICC in terms of first and second person pronouns, it leans more towards CWHPC in terms of the preposition proportion usage. To further clarify the register of COICC and provide more accurate explanation of the observed differences, future research can construct a balanced corpus which contains both spoken and written languages covering multiple topics and analyzes the date using factor analysis to establish a multidimensional analysis framework of Chinese.

5. Conclusion

This study compares the interpretation of international communications of China with the speeches by the U.S. officials. The findings suggest that the text type of CIICC belongs to the register of learned exposition which focuses on conveying information through formal language. In contrast, CWHPC belongs to the register of involved persuasion, characterized by colloquial language, frequent interactivity, and moderate tone. These differences are influenced by multiple factors, including source language, interpreting process, as well as political considerations. In conclusion, the findings support Frawley’s (2012) “third code” theory and reveal distinctive features in the interpretation of China’s international communications, reflecting a meticulous and responsible but less interactive national image. In contrast, the United States’ national image in international communications is characterized by a strong emphasis on interactivity and involvement, contributing to a more approachable and interactive international image. These findings could shed light on the interpreting strategies adopted by government interpreters and further improve the effectiveness of international communications. Since this study uses the 67 linguistic features selected by Biber without corresponding adjustments according to the political context, the results are somewhat limited. Future research could expand the corpus to include more texts from both China and other English-spoken countries and target more linguistic features, so that the persuasiveness of the findings could be enhanced.

Acknowledgments

My thanks go to my friend Mengfei Li for his advice in the data analysis of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Guo Zhen

Guo Zhen is currently a Ph.D. candidate at Beijing Foreign Studies University, China, specializing in translation studies. His research focuses on corpus translation studies and translation training. Prior to his doctoral studies, he earned a master’s degree from BFSU in 2019, majoring in translation and interpretation.